Paleobotany -The Biology and Evolution of Fossil Plants

SAVE OFFLINE

PALEOBOTANY

This page intentionally left blank

PALEOBOTANY The Biology and Evolution of Fossil Plants Second Edition

THOMAS N. TAYLOR

Department of Ecology and Evolutionary Biology and Natural History Museum and Biodiversity Research Center, The University of Kansas, Lawrence, Kansas

EDITH L. TAYLOR Department of Ecology and Evolutionary Biology and Natural History Museum and Biodiversity Research Center, The University of Kansas, Lawrence, Kansas

MICHAEL KRINGS Bayerische Staatssammlung für Paläontologie und Geologie und GeoBio-Center LMU, Munich, Germany

AMSTERDAM • BOSTON • HEIDELBERG • LONDON • NEW YORK • OXFORD PARIS • SAN DIEGO • SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO Academic Press is an imprint of Elsevier

Academic Press is an imprint of Elsevier 30 Corporate Drive, Suite 400, Burlington, MA 01803, USA 32 Jamestown Road, London NW 1 7BY, UK 525 B Street, Suite 1900, San Diego, CA 92101-4495, USA 360 Park Avenue South, New York, NY 10010-1710, USA Copyright © 2009, Elsevier Inc. All rights reserved No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher. Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone (44) (0) 1865 843830; fax (44) (0) 1865 853333; email: [email protected]. Alternatively visit the Science and Technology Books website at www.elsevierdirect.com/rights for further information. Notice No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made. Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN: 978-0-12-373972-8

For information on all Academic Press publications visit our website at www.elsevierdirect.com

Typeset by Charon Tec Ltd., A Macmillan Company. (www.macmillansolutions.com) Printed and bound in the USA 09 10 11

10 9 8 7 6 5 4 3 2 1

CONTENTS Chemical Fossils Ancient DNA Mummification Amber Summary Discussion Palynology Geochronology and Biostratigraphy Paleoecology Absolute Dating Geologic Timescale Biological Correlation Systematics and Classification Nomenclature of Fossil Plants Classification of Organisms Background Reading

Preface xv Acknowledgments xvii About the Authors xxi

CHAPTER

1

Introduction to Paleobotany, How Fossil Plants are Formed 1 What Is Paleobotany? The Objectives of Paleobotany Reconstructing the Plants Evolution of Plant Groups Form and Function in Fossil Plants Biostratigraphy and Correlation Paleoecology: Plants in Their Environment Determining Paleoclimate from Fossil Plants Tree Rings Nearest Living Relative Leaf Physiognomy Stomatal Index Summary Preservation: How Plant Fossils are Formed and Preserved Depositional Environments of Fossil Plants Compressions Cuticle Biofilms and Plant Fossil Preservation Electron Microscopy Confocal Microscopy Maceration and Dégagement Other Techniques Coal and Charcoal Impressions Molds and Casts Cellular Preservation Permineralization Peel Technique Coal Balls Other Permineralizations Petrifaction Unaltered Plant Material

1 2 2 3 4 4 5 6 6 6 7 7 7

CHAPTER

2

Precambrian Life

43

The Origin of Life on Earth Origin of Life: Theory and Biology Earliest Record of Life on Earth Historical Background Earliest Records of Life: Paleoarchean (3.6–3.2 Ga) Geochemistry Microfossils (Body Fossils) Isua Greenstone Belt, Greenland Warrawoona Group, Australia Barberton Greenstone Belt, South Africa Stromatolites Sedimentary Evidence Mesoarchean–Neoarchean Life Conclusions: Archean Life Oxygenation of the Earth (2.45–2.2 Ga) Proterozoic Life Paleoproterozoic Origin of Eukaryotes Mesoproterozoic Earliest Multicellular Life Neoproterozoic Bitter Springs Biota

8 8 10 13 16 17 17 17 18 18 21 22 23 25 25 27 29 30 30

v

32 33 33 33 34 34 36 37 38 39 40 40 41 42 42

44 46 47 47 47 47 49 49 49 51 52 53 54 55 57 59 59 61 64 64 64 65

vi

contents

Stromatolites Other Microfossils Doushantuo Formation Conclusions CHAPTER

66 67 70 70

3

Fungi, Bacteria, and Lichens 71 Fungi Earliest Fossil Fungi Systematics of Fungi Chytridiomycota Zygomycota Glomeromycota Ascomycota Basidiomycota Other Fungal Remains Fungal Life-History Strategies Saprotrophism Parasitism Mutualism Fungi–Animal Interactions Geologic Activities of Fungi Epiphyllous Fungi Fungal Spores Fungal-like Organisms Peronosporomycetes (Oomycota) Eubacteria and Archaea Archaea Eubacteria Cyanobacteria Lichens CHAPTER

Algae

71 73 77 77 82 84 90 93 97 98 98 99 103 105 107 108 111 112 112 112 113 113 115 117

4

121

Chlorophyta (Green Algae) Prasinophyceae Chlorophyceae Volvocales Tetrasporales Chlorococcales Ulvophyceae Dasycladales Receptaculitida and Cyclocrinales Caulerpales Taxa Incertae Sedis Charophyceae Charales Zygnematales Euglenophyta

123 124 126 126 126 127 128 128 130 130 133 133 134 138 138

Dinophyta (Dinoflagellates) Heterokontophyta Bacillariophyceae (Diatoms) Dictyochophyceae (Silicoflagellates) Xanthophyceae (Yellow-Green Algae) Phaeophyceae (Brown Algae) Prymnesiophyta (Haptophytes) Rhodophyta (Red Algae) Solenoporaceans Other Calcified Red Algae Corallinales Uncalcified Red Algae Acritarcha (Acritarchs) CHAPTER

5

Hornworts and Bryophytes

161

Early Fossil Evidence Anthocerotophyta (Hornworts) Bryophyta (Bryophytes) Marchantiophytina (Liverworts or Hepatophytes) Bryophytina (Mosses) CHAPTER

6

The Move to the Land

139 141 141 142 142 143 144 145 146 149 149 150 158

163 165 166 167 174

179

Enigmatic Organisms Nematophytes Prototaxites Nematothallus Nematoplexus Nematasketum diversiforme Pachytheca Spongiophytaceae Spongiophyton Orestovia Other Enigmatic Organisms Protosalvinia Parka Isolated Fragments: Clues to the Transition to Land? Cuticle and Cuticle-Like Material Spores and Spore Tetrads Tubes Land Plant Ancestors The Transition to Land Anchorage and Water Uptake Structural Support and Water Transport Protection Against Desiccation and Radiation Gas Exchange Reproduction on Land Life History Biology

180 180 180 183 183 183 184 185 185 186 186 186 188 189 189 189 192 193 194 194 195 195 195 196 196

Contents

Homologous Theory Antithetic Theory Animals A Fungal Partner Conclusion

CHAPTER

196 196 198 198 199

7

Introduction to Vascular Plant Morphology and Anatomy 201 Plant Organography Cell Types Parenchyma Collenchyma Sclerenchyma Tracheary Elements Tracheids Vessel Elements Sieve Elements Plant Tissues and Primary Growth Xylem Tissue Phloem Tissue Meristems Epidermis Cuticle Stomata Trichomes Anatomy of Stems and Roots Arrangement of Primary Tissues Primary Xylem Maturation Patterns Secondary Development Vascular Cambium Cork Cambium (Phellogen) Secondary Xylem Secondary Phloem Stele Types Primitive Vascular Plants (Vascular Cryptogams) Seed Plants Leaf Morphology and Anatomy Leaf Anatomy Leaf Evolution Further Reading

CHAPTER

202 203 203 203 203 204 204 206 206 207 207 207 208 208 209 209 210 210 210 212 212 212 213 214 216 216 216 219 221 221 222 222

8

Early Land Plants with Conducting Tissue 223 Conducting Elements in Early Land Plants History of Discovery

224 225

Rhyniophytes Rhynie Chert Plants Aglaophyton major Rhynia Gwynne-vaughanii Horneophyton lignieri Asteroxylon mackiei Nothia aphylla Trichopherophyton teuchansii Ventarura lyonii Gametophyte Generation Other Rhyniophytes Discussion: Rhyniophyte Evolution Zosterophyllophytes Zosterophyll Evolution Trimerophytes Trimerophyte Evolution Early Land Plant Evolution CHAPTER

9

Lycophyta

vii 227 228 229 235 237 238 239 241 241 241 246 251 252 259 259 262 263

265

Evolution of the Microphyll Drepanophycales Protolepidodendrales Lepidodendrales Vegetative Features Stem Surface and Leaf Bases Stem Anatomy Cortical Tissues Stem Development Leaves Underground Organs Development of Underground Organs Reproductive Biology Microsporangiate and Bisporangiate Cones Megasporangiate Cones Gametophytes Sigillariaceae Leaf Bases Leaves Stem Structure Underground Organs Reproductive Biology Other Lepidodendrid Genera Lycopodiales Selaginellales Pleuromeiales Isoetales Putative Lycopsids Conclusions

267 268 271 279 282 282 285 286 287 289 289 293 294 295 297 302 303 304 305 305 306 306 307 310 312 316 320 325 326

viii

contents

CHAPTER

10

Sphenophytes

329

Pseudoborniales Sphenophyllales Devonian Sphenophyllales Sphenophyllum Leaves Stem Anatomy Roots Reproductive Biology Other Sphenophyllales Ecology Equisetales Calamitaceae Archaeocalamites Calamites Pith Casts Stem Anatomy Extraxylary Tissues Growth and Development Roots Leaves Other Calamitean Leaves Reproductive Biology Spores Tchernoviaceae and Gondwanostachyaceae Vegetative Body Reproductive Biology Equisetaceae Forms with Uncertain Affinities Sphenophyte Evolution CHAPTER

331 332 333 334 334 335 337 337 338 341 342 343 343 345 349 350 352 352 353 354 357 358 366 368 368 369 371 376 379

11

Ferns and Early Fernlike Plants 383 Evolution of the Megaphyll Cladoxylopsida Pseudosporochnales Calamophyton Plant Iridopteridales Phylogenetic Position of the Cladoxylopsids Early Fernlike Plants Rhacophytales Rhacophyton Other Taxa Systematics of the Rhacophytales Coenopterid Ferns Stauropteridales Zygopteridales

386 387 388 396 398 400 401 401 402 403 404 405 405 408

Zygopterid Evolution Marattiales Psaroniaceae: Vegetative Features Psaronius Plant Other Stem Taxa Psaroniaceae: Reproductive Features Paleozoic Compression Taxa Mesozoic Marattialeans Marattialean Evolution Ophioglossales Leptosporangiate Ferns Osmundales Paleozoic Stem Taxa Guaireaceae Mesozoic and Cenozoic Stem Taxa Sterile and Fertile Foliage Osmundalean Evolution Botryopteridaceae Vegetative Organs Reproductive Organs Other Genera Anachoropteridaceae Kaplanopteridaceae Psalixochlaenaceae Sermayaceae Tedeleaceae Skaaripteridaceae Tempskyaceae Schizaeaceae Hymenophyllaceae Gleicheniaceae Dicksoniaceae Cyatheaceae Matoniaceae Loxsomataceae Dipteridaceae Polypodiales Salviniales Marsileaceae Salviniaceae Conclusions CHAPTER

12

Progymnosperms

417 418 418 418 425 425 431 433 434 435 436 436 437 438 438 440 442 443 443 446 449 449 451 452 453 454 457 457 459 462 462 464 465 466 469 469 470 472 472 473 476

479

Archaeopteridales Archaeopteris Leaves Archaeopterid Reproduction Callixylon Stems Other Archaeopterids

480 481 483 484 487

Contents

Aneurophytales Aneurophyton Tetraxylopteris Triloboxylon Rellimia Other Aneurophytes Protopityales Noeggerathians Progymnosperm Evolution CHAPTER

489 489 489 491 492 494 496 497 501

13

Origin and Evolution of the Seed Habit 503 Homospory, Heterospory, and the Seed Habit Homospory Heterospory Sporangia Endospory Lycopsid Heterospory Seed Habit Evolution of the Integument Evolution of Pollen Capture Pollen Cupules Cupulate Devonian Seeds Reproductive Biology Carboniferous Seeds Pollen Chamber Function Microgametophytes Diversity of Early Seeds Paleozoic Seeds with Embryos CHAPTER

14

Paleozoic Seed Ferns

503 503 504 504 507 508 508 509 510 511 511 511 517 518 523 524 525 526

529

Calamopityales Buteoxylonales Lyginopteridales Lyginopteris Plant Vegetative Organs Reproductive Structures Other Lyginopterids: Vegetative Remains Heterangium Microspermopteris Schopfiastrum Pitys Devonian–Mississippian Taxa Problematic Lyginopterids Other Lyginopterids: Seeds and Cupules Sphaerostoma Salpingostoma

531 539 540 540 540 542 546 547 550 550 551 552 554 555 556 556

Conostoma Coronostoma Physostoma Tyliosperma Calathospermum Gnetopsis Megatheca Other Lyginopterids: Pollen Organs Incertae Sedis Lyginopterid Evolution Medullosales Stems Medullosa Other Stem Taxa Leaves (Fronds) Roots Growth Habit Seeds Pollen organs Pollen Medullosan Evolution Callistophytales Vegetative Organs Reproductive Structures Callistophytalean Evolution Glossopteridales Leaves Glossopteris Gangamopteris Other Leaf Types Stems and Roots Ovulate Reproductive Structures Permineralized Forms Impression–Compression Specimens What is the Glossopterid Ovulate Structure? Pollen Organs Glossopteris Habit and Habitat Phylogenetic Position CHAPTER

15

Mesozoic Seed Ferns Caytoniales Sagenopteris Caytonanthus Caytonia Ruflorinia and Ktalenia Corystospermales Foliage Stems

ix 556 557 557 558 558 559 559 560 563 565 566 566 566 569 570 572 572 573 581 590 591 593 594 595 598 598 599 599 603 603 605 606 606 609 614 616 618 618

621 622 622 623 624 626 627 627 630

x

contents

Pollen Organs Ovulate Structures Petriellales Peltaspermales Foliage Reproductive Organs and Whole-Plant Concepts Conclusions CHAPTER

631 634 637 639 639 643 648

16

Late Paleozoic and Mesozoic Foliage 651 Late Paleozoic Foliage Adiantites Alethopteris Aneimites Aphlebia Alloiopteris Botrychiopsis Callipteridium Cardiopteridium Cardiopteris (Fryopsis) Charliea Cyclopteris Dicksoniites Discopteris Eremopteris Ginkgophytopsis Kankakeea Karinopteris, Mariopteris, and Pseudomariopteris Lesleya Linopteris, Reticulopteris, and Barthelopteris Lobatopteris Lonchopteridium and Lonchopteris Megalopteris Neuropteris sensu lato Laveinopteris Macroneuropteris Margaritopteris Neuralethopteris Neurocallipteris Neurodontopteris Neuropteris sensu stricto Paripteris Sphenoneuropteris Neuropterid Growth Habit Blanzyopteris Nothorhacopteris Odontopteris and Lescuropteris Pecopteris

652 655 656 657 658 658 659 659 660 660 660 661 662 664 664 664 665 665 669 669 671 672 672 673 674 674 674 674 675 675 675 675 676 676 676 677 677 679

Rhodea (Rhodeopteridium) Sphenopteris Sphenopteris sensu stricto Eusphenopteris Spiropteris Taeniopteris Tinsleya Triphyllopteris, Genselia, and Charbeckia Mesozoic Foliage Anomozamites Cladophlebis Coniopteris Ctenis Deltolepis and Cycadolepis Dictyophyllum Dictyozamites Doratophyllum Macrotaeniopteris Matonidium Mesodescolea Nilssonia Nilssoniopteris Otozamites Pachypteris, Komlopteris, and Thinnfeldia Phlebopteris Pseudoctenis Pseudocycas Pterophyllum Ptilophyllum Ptilozamites Ruflorinia Taeniozamites Ticoa Wingatea Yabeiella Zamites CHAPTER

17

Cycadophytes

680 680 682 682 683 683 685 685 685 687 687 688 689 689 689 689 690 690 690 690 690 691 693 695 696 696 697 697 698 699 699 700 700 700 700 701

703

Cycadales Leaves and Petioles Stems Paleozoic Reproductive Structures Triassic Cycads Jurassic Cycads Pollination Biology Discussion: Cycad Evolution Bennettitales Cycadeoidaceae

703 706 707 709 715 718 721 721 722 725

Contents

Stem Anatomy Reproductive Structures Development Williamsoniaceae Ovulate Structures Pollen Organs Discussion: Bennettitales CHAPTER

744 747 747 750 750 750 752 752 752 753 754 755

19

Gymnosperms with Obscure Affinities 757 Gigantopteridales Vegetative Remains Reproductive Organs Vojnovskyales Czekanowskiales Iraniales Pentoxylales Hermanophytales Gnetales Extant Genera Ephedra Gnetum Welwitschia Extant Reproductive Structures Fossil Gnetophyte Pollen Gnetophyte Megafossils Putative Gnetophytes Dirhopalostachyaceae

20

Cordaitales

758 758 762 763 765 768 768 773 775 776 776 776 776 777 777 778 781 785

787

Vegetative Features Stems Foliage

21

Conifers

743

Paleozoic Record Ginkgophyte Wood Ginkgophyte Foliage Pollen-Producing Structures Ginkgophyte Plants Ginkgoaceae Karkeniaceae Umaltolepidiaceae Yimaiaceae Schmeissneriaceae Taxa with Uncertain Affinities Conclusions

CHAPTER

Roots Reproductive Features Reproductive Organs Seeds Angaran Cordaites Phylogenetic Position and Origin of the Cordaites CHAPTER

18

Ginkgophytes

CHAPTER

725 728 730 732 734 738 739

788 788 791

xi 794 795 795 798 801 803

805

Early Conifers Voltziales Utrechtiaceae Utrechtia Ernestiodendron Ortiseia Otovicia Moyliostrobus Other Taxa Thucydiaceae Emporiaceae Majonicaceae Ullmanniaceae Bartheliaceae Other Voltzialeans Ferugliocladaceae Buriadiaceae Pollen Cones Summary Discussion: Voltzialeans Coniferales Palissyaceae Cheirolepidiaceae Summary Discussion: Cheirolepidiaceae Podocarpaceae Summary Discussion: Podocarpaceae Araucariaceae Summary Discussion: Araucariaceae Cupressaceae Cunninghamioideae Taiwanioideae Athrotaxoideae Sequoioideae Taxodioideae Cupressoideae Cupressaceous Wood Summary Discussion: Cupressaceae Sciadopityaceae Pararaucariaceae Pinaceae Pinoideae

806 807 808 809 809 809 810 811 811 814 815 816 819 820 820 823 826 826 828 830 830 831 837 838 843 843 848 849 850 851 851 852 854 857 859 859 860 861 861 863

xii

contents

Genus Pinus Pinus Wood Larix Piceoideae Abietoideae Summary Discussion: Pinaceae Cephalotaxaceae Taxaceae Summary Discussion: Cephalotaxaceae and Taxaceae Conclusions CHAPTER

22

Flowering Plants

864 866 866 867 867 868 868 869 869 870

873

Angiosperm Origins Origin of the Flower Pseudanthial Theory Euanthial Theory Microsporangial Theories Transitional–Combination Theory Habit Ecological Considerations Site of Origin Pre-Cretaceous Fossil Evidence Sanmiguelia Furcula Problematospermum Pre-Cretaceous Pollen Dispersed Pollen Early Angiosperm Evidence Pollen Pollen Evolution Evidence from Leaves Angiosperm Ancestors Caytoniales Czekanowskiales Glossopteridales Bennettitales Pentoxylales Gigantopteridales Phylogenetic Analyses and Angiosperm Origins Selected Angiosperm Families Basal Angiosperms Amborellaceae Hydatellaceae Archaefructaceae Chloranthaceae Nymphaeales

876 877 877 878 878 878 879 879 880 880 881 882 883 883 884 885 885 889 889 893 894 895 895 895 895 895 895 897 898 898 898 898 899 901

Nymphaeaceae Austrobaileyales Austrobaileyaceae Illiciaceae Schisandraceae Ceratophyllales Ceratophyllaceae Magnoliids Canellales Winteraceae Laurales Calycanthaceae Lauraceae Magnoliales Annonaceae Magnoliaceae Myristicaceae Piperales Lactoridaceae Saururaceae Monocotyledons Alismatales Alismataceae Araceae Hydrocharitaceae Zosteraceae (Seagrasses) Asparagales Agapanthaceae Hemerocallidaceae Orchidaceae Dioscoreales Dioscoreaceae Liliales Petermanniaceae Pandanales Pandanaceae Triuridaceae Commelinids Arecales Arecaceae (Palmae) Commelinales Commelinaceae Poales Cyperaceae Poaceae (Gramineae) Zingiberales Musaceae Zingiberaceae Eudicots

901 902 902 902 903 904 904 904 904 904 906 906 906 908 908 909 914 915 915 915 917 917 917 917 917 920 921 921 921 921 922 922 922 922 923 923 923 923 923 923 925 925 925 925 926 928 928 929 929

Contents

Buxaceae Trochodendraceae Proteales Nelumbonaceae Proteaceae Platanaceae Ranunculales Berberidaceae Ranunculaceae Core Eudicots Gunnerales Gunneraceae Caryophyllales Phytolaccaceae Saxifragales Cercidiphyllaceae Haloragaceae Hamamelidaceae Iteaceae Saxifragaceae Rosids Vitaceae Myrtales Lythraceae Trapaceae Myrtaceae Onagraceae Eurosids I (Fabids) Fabales Fabaceae (Leguminosae) Fagales Betulaceae Casuarinaceae Fagaceae Juglandaceae Myricaceae Nothofagaceae Malpighiales Clusiaceae Euphorbiaceae Salicaceae Malpighiaceae Oxalidales Cunoniaceae Elaeocarpaceae Rosales Moraceae Rhamnaceae Rosaceae

930 931 933 933 935 937 940 940 940 941 941 941 941 941 942 942 943 945 945 946 946 947 948 948 948 948 950 950 950 950 953 953 955 956 961 966 966 967 967 968 970 970 971 971 971 971 971 971 971

Ulmaceae Eurosids II (Malvids) Brassicales Capparaceae Malvales Tiliaceae Sapindales Anacardiaceae Meliaceae Rutaceae Sapindaceae Asterids Cornales Cornaceae Curtisiaceae Hydrangeaceae Ericales Ebenaceae Ericaceae Theaceae Euasterids I (Lamiids) Icacinaceae Garryales Eucommiaceae Gentianales Gentianaceae Rubiaceae Lamiales Avicenniaceae Byblidaceae Lentibulariaceae Oleaceae Solanales Solanaceae Euasterids II (Campanulids) Bruniaceae Quintiniaceae Apiales Araliaceae Aquifoliales Aquifoliaceae Asterales Asteraceae (Compositae) Menyanthaceae Dipsacales Caprifoliaceae Cenozoic Floras Conclusions

xiii 973 976 976 976 976 976 977 977 978 978 979 981 981 981 984 984 985 985 985 985 986 986 987 987 987 987 987 988 988 988 988 988 988 988 988 988 988 989 989 989 989 990 990 991 991 991 991 996

xiv

CHAPTER

contents

23

Interactions Between Plants and Animals 999 Early Terrestrial Ecosystem Associations Animals on Land Early Plant–Animal Associations Herbivory Defenses Against Herbivory Mechanical Protection Chemical Defenses Fossil Evidence of Herbivory Coprolites Gut Contents Marginal Feeding Defoliation Leaf Miners Wound Tissue

1001 1001 1001 1003 1004 1005 1006 1007 1007 1011 1011 1013 1013 1015

Interactions with Vertebrates Herbivory Dentition Coprolites and Stomach Contents Dispersal Plants as Habitat Other Plant–Animal Interactions Mimicry Pollination Conclusions

1016 1016 1018 1018 1018 1019 1021 1021 1022 1024

Appendix 1: Classification of Organisms Glossary References Index

1027 1031 1049 1199

This page intentionally left blank

Preface One of the major challenges faced by paleobotanists is the extraordinary interdisciplinary nature of the science. Most paleobotanists are quick to point out, however, that they were practicing collaborative and interdisciplinary science long before the concepts became fashionable in the research environment of today. We view this book as an up-to-date introduction to the discipline for advanced undergraduate and graduate students, and as a book that is more encyclopedic in organization than other paleobotany textbooks and that can be used as a reference for a number of disciplines that today encompass the biological and geological sciences, whether professional or amateur. Although this book is not technically a second edition, it does include material from The Biology and Evolution of Fossil Plants by Thomas N. Taylor and Edith L. Taylor (1993), which has long been out of print. To make the book usable to a wider range of readers, we begin each chapter with a general introduction that provides the essential characteristics of a particular group, including not only information about fossil members but also, where applicable, living representatives. In addition to a comprehensive table of contents, we have added a table to each chapter that summarizes the higher taxa in the chapter and the geologic range of each group. Chapters are subdivided to make it easier for readers to find information. For the nonbiologist, we have included a discussion of plant structure, tissue systems, and plant organs (Chapter 7) that is supplemented by illustrations and diagrams. To further assist in making the book useful, we have

expanded the glossary from Taylor and Taylor (1993) to more than 900 entries. For easy reference, a chart showing the geologic periods is included inside the front and back covers. With more than 5000 references, this book provides an introduction to the primary literature. For further literature, please see the Bibliography of Paleobotany, http://paleobotany. bio.ku.edu/BiblioOfPaleo.htm We have also included more than 2100 illustrations, many in color, and a large number unpublished. We received numerous favorable comments regarding the portraits of distinguished paleobotanists and therefore have included many more in this book. We found the following online sources to be of enormous assistance in writing this book and would like to thank those who maintain these resources: (1) Index Nominum Genericorum, http://ravenel.si.edu/botany/ing/ingForm.cfm; (2) GBIF portal, http://www.gbif.org; (3) Peter Hoen’s Glossary of Pollen and Spore Terminology 2nd edition, http://www. bio.uu.nl/~palaeo/glossary/glos-int.htm, from the University of Utrecht (4) the International Commission on Stratigraphy site, http://www.stratigraphy.org; and (5) L. Watson and M. J. Dallwitz’s The Families of Flowering Plants, http://delta-intkey. com/angio/. We have followed the 2008 International Commission on Stratigraphy (ICS) conventions on naming geologic time periods (htpp://www.stratigraphy.org) and have provided the international name in addition to local stage names throughout.

xv

This page intentionally left blank

Acknowledgments We are truly indebted to a large number of individuals and institutions who eagerly assisted us in the preparation of this book. This includes all of our colleagues both past and present, who readily contributed illustrations as well as the various professional journals, book publishers, and professional organizations that granted permission to use copyrighted material. We are also indebted to numerous colleagues who took the time to read chapters and sections of the book, and freely discussed their ideas so that we could produce a book that is as up-to-date as possible in a field covering as much information as does paleobotany. These include:

J. D. Aitken ● V. M. Albright ● A. C. Allwood ● K. L. Alvin ● H. M. Anderson ● J. M. Anderson ● H. N. Andrews, Jr. ● Anschutz Science Library staff, University of Kansas ● A. Archangelsky ● S. Archangelsky ● S. R. Ash ● K. R. Aulenback ● L. Axe ● B. J. Axsmith ● N. R. Banerjee ● H. P. Banks ● F. Baron ● D. Barr ● M. Barthel ● R. A. Baschnagel ● J. F. Basinger ● D. Bassi ● P. W. Basson ● L. H. Batenburg ● R. W. Baxter ● Bayerische Staatssammlung für Paläontologie und Geologie, Munich (BSPG) ● S. C. Beadle ● C. B. Beck ● J. Bek ● J. M. Benson ● M. L. Berbee ● S. Berger ● M. E. C. Bernardes-de-Oliveria ● C. M. Berry ● J. Bogner ● B. Bomfleur ● P. M. Bonamo ● M. N. Bose ● H. Blütmann ● S. D. Brack (Brack-Hanes) ● D. F. Brauer ● M. Brea ● W. Brenner ● D. W. Brett ● D. E. G. Briggs ● H. K. Brooks ● N. C. Brotzman ● J. T. Brown ● V. I. Burago ● J. Burgess ● N. D. Burgess ● N. J. Butterfield ● R. Butzmann ● L. Calvillo-Canadell ● D. J. Cantrill ● L. M. Carluccio ● A. V. Carozzi ● R. J. Carpenter ● S. N. Césari ● S. R. S. Cevallos-Ferriz ● W. G. Chaloner ● S. Chandra ● D. S. Chaney ● M. Chaphekar ● R. L. Chapman ● I. Chen ● S. Chitaley ● D. C. Christophel ● M. A. Cichan ● J. A. Clement-Westerhof ● J. A. Clendening ● G. T. Cole ● M. Collinson ● N. Combourieu ● M. E. Cook ● B. Cornet ● P. R. Crane ● W. L. Crepet ● A. A. Cridland ● W. N. Croft ● Y. M. Crosbie ● A. T. Cross ● N. R. Cúneo ● R. Daber ● C. P. Daghlian ● V. Daviero-Gomez ● E. Y. Dawson ● A.-L. Decombeix ● T. Delevoryas ● G. del Fueyo ● C. Delwiche ● T. Denk ● R. L. Dennis ● Denver Museum of Nature ● M. E. Dettmann ● M. DeVore ● C. Diéguez ● D. L. Dilcher ● R. M. Dillhoff ● W. A. DiMichele ● M. Dolezych ● S. Doerenkamp ● J. B. Doran ● H. Dörfelt ● N. Dotzler ● J. A. Doyle ● A. N. Drinnan ● X.-M. Du ● J. Dupéron ● M. Dupéron-Laudoueneix ● D. Edwards ● D. S. Edwards ● L. E. Edwards ● D. A. Eggert ● D. L. Eggert ● H. Eklund ● W. E. El-Saadawy ● G. Elliott ● M. S. Engel ● I. Ernstmeier ● J. Erl ● D. M. Erwin ● I. Escapa ● W. R. Evitt ● M. Fairon-Demaret ● M. Feist ● D. K. Ferguson ● P. F. Fields ● T. I. Fine ● T. Fischer ● F. Fleming ● G. L. Floyd ● V. K. Folkman ● W. H. Forbes ● C. B. Foster ● J.-P. Frahm ● F. Franzmeyer ● W. E. Friedman ● E. M. Friis ● W. L. Fry ● G. Fuchs ● J. Galtier ● M. A. Gandolfo ● Z. Gao ● R. A. Gastaldo ● C. T. Gee ● P. G. Gensel ● Geological Society of America ● Geologische Bundesanstalt, Vienna (GBA) ● E. A. George ● P. Gerrienne ● R. W. Gess ● D. E. Giannasi ● W. H. Gillespie ● C. W. Good ● I. Glasspool ● K. D. Gordon-Gray ● R. Gossmann ● K. Goth ● R. E. Gould ● F. Gradstein ● S. R. Gradstein ● A. Graham ● L. E. Graham ● N. GrambastFessard ● L. Grauvogel-Stamm ● J. Gray ● J. D. Grierson ● R. Grolle ● M. A. Haban ● D. W. Haines ● J. W. Hall ● Hancock Museum ● S.-G. Hao ● T. M. Harris ● C. M. Hartman ● H. Hass ● A. R. Hemsley ● P. S. Herendeen ● E. J. Hermsen ● G. R. Hernández-Castillo ● G. Heumann ● F. A. Hibbert ● D. S. Hibbert ● L. J. Hickey ● A. Hill ● R. S. Hill ● S. A. Hill ● N. Hiller ● L. Hillis ● J. M. Hilton ● Hirmer Verlag GmbH, Munich ● H. J. Hofmann ● J. C. Holmes ● W. B. K. Holmes ● H. J. Hoops ● R. C. Hope ● C. A. Hopping ● D. G. Horton ● C. L. Hotton ● J. Hsü ● F. M. Hueber ● A. Iglesias ● I. A. Ignatiev ● W. I. Illman ● Interlibrary Loan staff, University of Kansas ● International Commission on Stratigraphy ● L. C. Ivany ● B. F. Jacobs ● H. Jähnichen ● G. Janssen ● J. A. Janssens ● H.-B. Jansson ● S. Jardiné ● E. A. Jarzembowski ● D. M. Jarzen ● J. R. Jennings ● A. J. Jeram ● K. R. Johnson ● M. E. Johnson ● D. S. Jones ● J. H. Jones ● W. W. Jung ● M. J. Kaever ● E. Karasev ● E. E. Karrfalt ● A. E. Kasper ●

xvii

xviii

acknowledgments

K.-P. Kelber ● P. Kenrick ● E. M. Kern ● H. Kerp ● P. F. Kidwai ● B.-K. Kim ● T. Kimura ● M. Kirchner ● B. L. Kirkland ● S. Kiyokawa ● S. D. Klavins ● M. J. Knaus ● A. H. Knoll ● A. S. Konopka ● E. B. Koppelhus ● W. L. Kovach ● V. A. Krassilov ● J. Kukalová-Peck ● J. Kvacˇek ● Z. Kvacˇek ● C. C. Labandeira ● Laboratory of Palaeobotany and Palynology, Utrecht University (LPPU) ● W. S. Lacey ● W. H. Lang ● C. A. LaPasha ● I. S. Latimer, Jr. ● J.-P. Laveine ● R. L. Leary ● S. Leclercq ● G. A. Leisman ● K. U. Leistikow ● B. A. LePage ● U. Leppig ● C.-S. Li ● H.-L. Li ● J. H. Lipps ● R. J. Litwin ● Ludwig-Maximilians-Universität München ● S. T. LoDuca ● A. G. Long Collection ● C. V. Looy ● T. A. Lott ● D. R. Lowe ● B. Lugardon ● B. Lundblad ● S. D. Lys ● H. K. Maheshwari ● S. H. Mamay ● S. R. Manchester ● S. B. Manum ● G. Mapes ● F. Marsh ● D. M. Martill ● L. C. Matten ● J. D. Mauseth ● Max Kade Center for German-American Studies, University of Kansas ● H. Mayr ● A. M. McClain ● E. E. McIver ● S. McLoughlin ● C. A. McRoberts ● J. Mehl ● K. Meister ● S. V. Meyen ● H. W. Meyer ● B. Meyer-Berthaud ● J. E. Mickle ● M. A. Millay ● C. E. Miller ● C. N. Miller, Jr. ● J. M. Miller ● B. A. R. Mohr ● M. Montenari ● E. D. Morey ● J. Morgan ● J. E. Morris ● Museum für Naturkunde, Berlin (MNB) ● Národní Muzeum (National Museum), Prague ● N. S. Nagalingum ● K. K. Namboodiri ● E. M. V. Nambudiri ● Naturhistorisches Museum Schloss Bertholdsburg Schleusingen (NMS) ● Naturhistorisches Museum, Vienna (NHM) ● Naturhistoriska Riksmuseet, Stockholm (NRM) ● S. V. Naugolnykh ● D. D. Nautiyal ● M. E. Nelson ● K. J. Niklas ● E. Nisbet ● H. Nishida ● M. Nishida ● M. H. Nitecki ● K. C. Nixon ● H. Nøhr-Hansen ● R. Noll ● M. Nose ● L. L. Oestry ● T. Ohana ● Oklahoma Geological Survey ● G. Oleschinski ● J. M. Osborn ● A. D. Pan ● D. D. Pant ● B. C. Parker ● L. R. Parker ● K. Parris ● M. Parrish ● C. S. Pearsall ● K. R. Pedersen ● K. Perch-Nielsen ● C. P. Person ● J. Peters ● B. Petriella ● H. W. Pfefferkorn ● H. Pflug ● R. N. Pheifer ● T. L. Phillips ● C. J. Phipps ● K. B. Pigg ● G. Playford ● D. T. Pocknall ● S. A. J. Pocock ● G. O. Poinar ● D. Pons ● R. J. Poort ● R. W. Portell ● C. Pott ● F. W. Potter ● N. W. Radforth ● C. G. K. Ramanujam ● P. H. Raven ● C. B. Read ● J. D. Reed ● M. A. Reihman ● D. Remy ● R. Remy ● W. Remy ● G. J. Retallack ● R. Riding ● E. Rieber ● J. F. Rigby ● M. O. Rischbieter ● C. R. Robison ● D. M. Rohr ● E. J. Romero ● R. Rössler ● G. W. Rothwell ● N. P. Rowe ● A. C. Rozefelds ● P. E. Ryberg ● C. Rydin ● B. Sahni ● F. Schaarschmidt ● J. T. Schabilion ● S. E. Scheckler ● R. Schmid ● A. Schmidt ● S. Schneider ● J. W. Schopf ● J. M. Schramke ● R. E. Schultes ● S. Schultka ● R. M. Schuster ● H.-J. Schweitzer ● U. Schweitzer ● A. C. Scott ● R. A. Scott ● S. Sekido ● P. A. Selden ● A. Selmeier ● B. S. Serlin ● T. Servais ● R. L. Seymour ● G. L. Shadle ● B. D. Sharma ● W. A. Shear ● M. A. Sherwood-Pike ● C. H. Shute ● M. A. Siders ● J. Silander ● A. D. Simper ● Z. Šim˚unek ● C. A. Sincock ● R. S. Singh ● J. E. Skog ● J. J. Skvarla ● C. J. Smiley ● M. L. So ● I. Sobbe ● S. K. Srivastava ● G. D. Stanley ● P. Steemans ● W. E. Stein, Jr. ● H. Steur ● W. N. Stewart ● B. M. Stidd ● R. A. Stockey ● W. C. Stowe ● M. Streel ● C. A. E. Strömberg ● P. Strother ● S. P. Stubblefield ● T. F. Stuessy ● O. P. Suthar ● K. Sugitani ● K. R. Surange ● N. P. Swanson ● H. A. J. M. Swinkels ● R. E. Taggart ● T. Tanai ● W. R. Tanner ● D. W. Taylor ● G. H. Taylor ● J. W. Taylor ● W. A. Taylor ● G. F. Thayn ● B. A. Thomas ● J. R. Thomasson ● L. Tidwell ● W. D. Tidwell ● A. M. Torres ● J. M. Trappe ● A. Traverse ● N. Trewin ● M. L. Trivett ● G. R. Upchurch ● H. W. J. van Amerom ● M. van Campo ● R. W. J. M. Van der Ham ● D. E. Van Dijk ● J. H. A. Van Konijnenburg-Van Cittert ● G. Vasanthy ● M. Vecoli ● J. C. Vega ● R. Verwer ● Alexander von Humboldt-Stiftung ● L. Voronova ● C. Vozenin-Serra ● C. A. Wagner ● R. H. Wagner ● D.-M. Wang ● H.-S. Wang ● L. Wang ● X. L. Wang ● Y. Wang ● Z.-Q. Wang ● J. V. Ward ● S. Warner ● J. Watson ● J. A. Webb ● C. H. Wellman ● R. Werneburg ● W. Werner ● F. Westall ● E. A. Wheeler ● D. C. White ● J. F. White, Jr. ● M. E. White ● R. Wicander ● D. C. Wight ● P. Wilf ● L. R. Wilson ● R. B. Winston ● J. A. Wolfe ● V. P. Wright ● S. Xiao ● Z. L. Xu ● A. Yabe ● X. Yao ● Z. Yao ● Z.-Q. Yao ● J. R. Young ● Y. D. Zakharov ● R. J. Zakrzewski ● S.-Q. Zan ● M. S. Zavada ● Z. Zhang ● Z. Zhou ● J.-N. Zhu, if we has missed someone we are deeply apologetic. We are especially indebted to Rudolph Serbet and Andrew Schwendemann for their invaluable help and assistance in photographing specimens, adding bar scales to images, and in general overseeing the myriad tasks associated with preparing the illustrations in a volume of this size. Without their attention to detail this volume could not have been completed. Christian Pott and Hans Kerp skillfully took and graciously provided, a number of digital images from paleobotanical collections in Europe, some of which have not been published earlier. Jeannie Houts deserves special mention for her assistance in all phases of the preparation of this book. Her ability to deal simultaneously with the three of us was a challenge that she met with diligence, skill, and efficiency, and which now requires a new definition for the word “patience.”

Acknowledgments

xix

We are indebted to Frank Baron and the Max Kade Center for German–American Studies at the University of Kansas for making accommodations available when MK visited Lawrence to work on the book, and to Judith Erl, Munich, who displayed special skill and patience in ferreting out literature for the bibliography. Thanks also to J. William Schopf for his assistance with many questions on Precambrian life. We are appreciative of the extraordinary help we received from Pat Gonzalez, Developmental Editor, and Andy Richford, Senior Acquisition Editor, both of Academic Press, for their unusual patience and assistance in bringing this volume to fruition; and to Mani Prabakaran and his staff at Macmillan Publishing Solutions for their attention to detail throughout the editing process. To the many individuals who participated in this enterprise that we never directly interacted with – we sincerely thank you for your professional expertise and assistance along the way. We could not have completed the book without your help. We gratefully acknowledge the National Science Foundation and Alexander von Humboldt-Stiftung for financial assistance that has supported our research programs. This support has provided assistance in many ways that has allowed us to continue our research and scholarly activities and interests, and to actively participate in the discoveries that make the discipline of paleobotany so exciting. To our families, friends, colleagues, and institutions where we are employed—we can never truly express how very much we appreciate your support and patience during the preparation of this book. You all have made sacrifices of untold proportion and exhibited extraordinary patience as we have worked on. This book would not have been completed without your understanding and friendship. There are many individuals who have influenced our careers or have made it possible for us to participate in this writing exercise. They have been there to teach and critique, to support and challenge, and to question and inspire—and their support and enthusiasm has never wavered. George W. Burns, Bill Crepet, Fred Daniëls, Chuck Daghlian, Ted Delevoryas, David Dilcher, Don Eggert, Ray Evert, Hans Kerp, Serge Mamay, Mike Millay, Elisabeth Peveling, Richard Popham, Winfried Remy, Gar Rothwell, Ruth Stockey, and Bill Stewart deserve special mention. Finally, we thank T. Cartel for friendship and support during our careers. As every chapter begins with a quotation, we will end with this famous one from Isaac Newton: “If [we] have seen farther it is by standing on the shoulders of giants.”

About the Authors Thomas N. Taylor is a distinguished Professor in the Department of Ecology and Evolutionary Biology, and Curator of Paleobotany in the Natural History Museum and Biodiversity Research Center at the University of Kansas. He also holds a courtesy appointment in the Department of Geology. He received his Ph.D. in botany from the University of Illinois, and was a National Science Foundation Postdoctoral Fellow at Yale University. He is a member of the National Academy of Sciences. His research interests include Permian and Triassic biotas of Antarctica, early land plant–fungal interactions, the origin and evolution of reproductive systems in early land plants, symbiotic systems through time, and the biology and evolution of fossil microbes.

Edith L. Taylor has been a Professor of Ecology and Evolutionary Biology and Senior Curator of Paleobotany in the Natural History Museum and Biodiversity Research Center at the University of Kansas since 1995, and also serves as a Courtesy Professor of Geology. She received her Ph.D. in paleobotany from the Ohio State University, where she was an American Association of University Women Dissertation Fellow. She is the author of seven books or edited volumes, and more than 140 publications. She was elected a Fellow of the American Association for the Advancement of Science in 1992. Her research interests include fossil wood growth and paleoclimate, Permian and Triassic permineralized plants from Antarctica, distribution and diversity of Permian–Triassic Antarctic floras, and the structure and evolution of fossil phloem.

Michael Krings is Curator for Fossil Plants in the Bavarian State Collection for Palaeontology and Geology (BSPG) at Munich, Germany, and Professor of Plant Paleobiology at the Ludwig-Maximilians-Universität, Munich. He also holds an affiliate faculty position in the Department of Ecology and Evolutionary Biology at the University of Kansas. He received his Ph.D. in botany from the University of Münster, Germany, and was an Alexander von Humboldt Foundation Postdoctoral Fellow at the University of Kansas. His research interests include Carboniferous, Permian, and Triassic seed plants from Europe and North America, and the biology and ecology of microorganisms in late Paleozoic terrestrial ecosystems.

xxi

This page intentionally left blank

1 INTRODUCTION TO PALEOBOTANY, HOW FOSSIL PLANTS ARE FORMED WHAT IS PALEOBOTANY? .............................................................1

Cellular Preservation ...........................................................................23

THE OBJECTIVES OF PALEOBOTANY ..................................... 2

Unaltered Plant Material .....................................................................30 Summary Discussion ..........................................................................34

Reconstructing the Plants......................................................................2 Evolution of Plant Groups.....................................................................3

PALYNOLOGY ..................................................................................34

Form and Function in Fossil Plants .......................................................4

Geochronology and Biostratigraphy ...................................................36

Biostratigraphy and Correlation............................................................4

Paleoecology .......................................................................................37

Paleoecology: Plants in Their Environment ..........................................5

ABSOLUTE DATING ......................................................................38

Determining Paleoclimate from Fossil Plants .......................................6 GEOLOGIC TIMESCALE ..............................................................39

Summary ...............................................................................................7

BIOLOGICAL CORRELATION .................................................. 40

PRESERVATION: HOW PLANT FOSSILS ARE FORMED AND PRESERVED ............................................... 8

SYSTEMATICS AND CLASSIFICATION ................................ 40

Depositional Environments of Fossil Plants .........................................8

Nomenclature of Fossil Plants ............................................................41

Compressions ......................................................................................10

Classification of Organisms ................................................................42

Coal and Charcoal ...............................................................................18 BACKGROUND READING.........................................................42

Impressions .........................................................................................21 Molds and Casts ..................................................................................22

The Earth is a vast cemetery where the rocks are tombstones on which the buried dead have written their own epitaphs. Louis Agassiz … intoxicated joy and amazement at the beauty and grandeur of this world, of which man can just form a faint notion. Albert Einstein

WHAT IS PALEOBOTANY?

no flowering plants, and when the continental land masses were in different positions than they are today. Who has not been captivated by the various forms of life that are recorded in the rocks and the enormous reconstructions of dinosaurs exhibited in various museums? It is natural to wonder about such examples of prehistoric life—how these organisms

Humans are by nature curious, and we are all interested in the Earth on which we live and how various aspects have changed through geologic time. We speculate about what the Earth looked like when there were no trees, when there were

1

2

paleobotany: the biology and evolution of fossil plants

lived, what their patterns of behavior were, and even why they became extinct. Although the paleontologist is interested in the geologic history of animals, the paleobotanist is concerned with the plants that inhabited the Earth throughout geologic time (Ward, 1885) (FIG. 1.1). In a general sense, the paleobotanist is a plant historian who attempts to piece together the intricate and complicated picture of the history of the plant kingdom. Although molecular and genetic analyses of living plants have become increasingly important as tools in reconstructing the phylogeny and evolutionary history of plants, the discipline of paleobotany, in all its various forms, remains the only method by which this history can be documented and visualized. Two books that discuss paleobotany from a historical perspective and that capture the excitement of the discipline are The Fossil Hunters—In Search of Ancient Plants (Andrews, 1980) and History of Palaeobotany—Selected Essays (Bowden et al., 2005). These volumes discuss the origins of the field and the scientists who have made the science so exciting and fascinating. Fossil plants and floras from one period of geologic time are different in size and shape, level of complexity, and abundance from those of other time periods. The most logical explanation for these differences is that the types of plants changed, or evolved, through geologic time. Unless one believes that there were an almost infinite number of “special creations,” we must assume that new plant forms were derived from preexisting ones by the processes of evolution. By studying the record of fossil plants, it is possible to assess the time at which various

major groups originated, the time each reached its maximum diversity, and, in the case of certain groups, when they became extinct.

THE OBJECTIVES OF PALEOBOTANY One of the aspects of paleobotany, which makes it unusual and interesting, is that it is inherently interdisciplinary and can be approached from either a biological or a more geological perspective—or both together. Each perspective presents a variety of questions that are unique to that discipline. Today more than ever before, the questions being asked by paleobotanists necessitate that both the botanical and geological perspectives be fully understood. To a large degree, the research questions that paleobotanists ask are influenced by whether their training emphasized a biological or a more geological perspective. RECONSTRUCTING THE PLANTS

Paleobotanists who have been trained primarily as biologists are interested in research directions which include all aspects of the organisms themselves. Because the majority of fossil plants are generally preserved in rocks as disarticulated plant parts (FIG. 1.2), that is leaves (FIG. 1.3), stems, pollen, or reproductive structures, a major aim of paleobotany is to reconstruct the whole plant, that is to say, to put the pieces of the puzzle back together. Once this is accomplished,

Figure 1.2 Impression of angiosperm leaf from the Dakota Figure 1.1

Lester Ward. (Courtesy H.N. Andrews.)

Formation (Cretaceous). Bar  2 cm.

chapter 1 introduction to paleobotany, how fossil plants are formed

the research can turn to other areas, such as determining the group of living plants, if any, to which the fossil is most closely related. Some paleobotanists are interested in aspects of plant life history that can be determined from fossils. For example, how did these plants reproduce, and how and what types of propagules were disseminated? Are their reproductive strategies similar to those of closely related living plants, or have there been major modifications in the reproductive systems of certain types of plants through geologic time? If so, how did this happen and when? What can we determine about the environment in which the plants lived millions of years ago, based on features of the fossil plants? For example, fossil wood collected from the Permian and Triassic of Antarctica (FIG. 1.4) indicates that the climate was quite

3

favorable for tree growth, based on the analysis of tree rings (FIGS. 1.5, 1.6). General circulation models of Permian paleoclimate, however, have proposed that these high paleolatitudes were very cold and not favorable for plant growth. Some paleobotanists are interested in what strategies these plants, and the animals that lived among them, developed to survive in the extreme seasonality at polar latitudes. EVOLUTION OF PLANT GROUPS

Paleobotanists are also interested in the origin and subsequent evolution of major groups of plants and their interrelationships. When did plants first inhabit the Earth and what did they look like? When did the first representatives of different groups of plants first arise? Other researchers want

Figure 1.4 Permineralized wood extending from paleostream

channel in the Triassic of Antarctica.

Compressed pinna showing detail of pinnule venation (Cretaceous). Bar  2 cm.

Figure 1.3

Figure 1.5 Section of Antarctic wood (Triassic) showing sev-

eral growth rings. Bar  1.4 mm.

4

paleobotany: the biology and evolution of fossil plants

Figure 1.6 Frost ring in Triassic permineralized wood from

Antarctica. Bar  3 mm.

to know why certain types of plants developed the capacity to produce secondary tissues (such as wood), whereas others have remained small throughout their geologic history. A number of paleobotanists study not only the plants themselves, but also the interactions of the plants with other organisms in the environment, especially the symbiotic interrelationships between plants and other organisms. For example, today almost all terrestrial plants possess mutualistic associations with fungi that inhabit their roots (mycorrhizae). Paleobotanists want to know when and why these associations became established, why they are absent in some groups today, and why only certain types of fungi form these associations. Can the fossil record of plants tell us anything about various biological activities that existed millions of years ago, such as the interactions between plants and certain groups of animals that use plants for food or protection? What types of evidence can be used, and what information does this provide about the interdependence of organisms through time? Can we determine from the fossil record if plants possessed certain features that served to attract pollinators, or produced edible seeds, or whether some plants produced certain chemicals that deterred herbivory? The answer to all these questions is a resounding YES! There is a multitude of information that can be gleaned from careful examination of the plant fossil record, and the types of information that we can obtain are constantly increasing as more and more research is done on fossil plants. FORM AND FUNCTION IN FOSSIL PLANTS

From many plant fossils, it is possible to understand the relationship between form and function in ancient plants, that is, what advantages or limitations are imposed on the growth and development of a plant based on certain biomechanical

properties? For example, are all arborescent (treelike) plants constructed of cells and tissue systems of the same type? If not, in what other ways can plants grow to tower over their neighbors? Studies of this type examine the anatomical and morphological properties of various fossil plants, often using computer simulations to model growth, in an attempt to better understand broad evolutionary patterns of plant growth, as well as changes in growth form through time (Niklas, 1992; Rowe and Speck, 1998; Niklas and Spatz, 2004; Niklas et al., 2006). Biomechanical studies have been especially useful in delimiting adaptations necessary for plants to move onto the land, including upright growth, size limitations, and the nature of the conducting strand (Niklas, 1986), and, once plants became established in terrestrial environments, the influences of gravity and wind on their reproduction (Niklas, 1998), and even aerodynamic features of pollen (Schwendemann et al., 2007). Factors such as plant size and form can also be examined over a broad spectrum of plant morphologies and thus offer insights as to why certain plants and plant groups have developed particular anatomical and morphological characteristics. Examining tree growth and other factors in extant plants has demonstrated that there are a variety of variables in play. Because fossils demonstrate a number of different growth forms that are not seen in modern plants, they offer a unique resource of data to allow paleobotanists to explore a host of intriguing biological questions. Fossil plants can also be used to infer developmental processes (Sanders et al., 2007). For example, Boyce and Knoll (2002) analyze the morphospace of numerous Paleozoic leaves representing various clades and show that leaf evolution follows the identical sequence of morphologies in all groups. Such an approach provides the ability to test hypotheses using living leaf development as a proxy for the leaves seen in various groups in the fossil record. BIOSTRATIGRAPHY AND CORRELATION

Paleobotany has also played a key role in many areas of geology, especially in biostratigraphy—placing rock units in stratigraphic order based on the fossils within them. Pollen grains and spores (one aspect of palynology) have been extensively used as index fossils in biostratigraphy and in the correlation of rock units, as have various forms of algal cells and cysts. In some instances, megafossils, such as leaves and seeds, have also provided a method of correlating rock units which are widely separated geographically. Pollen and spores, as well as megafossils, are especially useful in correlating terrestrial rocks, as these are generally deposited in limited areas (former lakes, ponds, river systems, etc.), making correlation by lithology (i.e., rock characteristics) very difficult.

chapter 1 introduction to paleobotany, how fossil plants are formed

1. Lepidophloios 2. Diaphorodendron 3. Synchysidendron 4. Paralycopodites 5. Sigillaria 6. Pteridosperms 7. Tree ferns 8. Sphenopsids

3

1

2

5

5

4

A

7 6

8

B

Figure 1.7 Transect through a Westphalian D mire showing habitat partitioning among major genera (A) and distribution of other plant

groups (B). (From DiMichele and Phillips, 1994.)

PALEOECOLOGY: PLANTS IN THEIR ENVIRONMENT

Paleoecology, the study of past environments, is a rapidly changing field that involves the integration and synthesis of both botanical and geological information. In recent years there has been a concerted effort by many paleobotanists to understand the paleoenvironment of fossil land plants more completely. For example, Bateman and Scott (1990) examined the famous late Tournaisian (lowermost Carboniferous) plant-bearing deposits at Oxroad Bay, Scotland, from a number of different perspectives, including an analysis of the geologic history and sedimentology of the site, as well as the paleoenvironment and paleoecology of the plants. Their studies indicate that the Oxroad Bay flora is found at eight levels in five successive facies, and these facies show the increasing influence of nearby volcanoes in the ocean-marginal setting. Details of the depositional environments through time at this site make it possible to understand plant adaptations to a rapidly changing, lowland environment, and to better understand both the biological and evolutionary importance of the floras.

Much paleoecological work initially focused on analyses of the swamp vegetation that contributed to extensive coal deposits in the Midcontinent of North America during the Carboniferous (Phillips and Peppers, 1984). The data used in these early analyses came primarily from the study of Pennsylvanian plants in coal balls—nodular concretions of limestone that contain permineralized peat (FIG. 1.5; see section on “Preservation”), coupled with a precise stratigraphic framework for the coal ball deposits based on palynology and careful field observations and measurements. Through the pioneering efforts of Phillips, DiMichele, and coworkers, we now possess an excellent understanding of many aspects of the paleoecology of coal-swamp vegetation during the Carboniferous (Wagner and Diez, 2007). Analysis of the plants preserved at different levels in these deposits not only documents the partitioning of the habitat among the different plant groups along ecological lines (Fig. 1.7), but also records changes in the depositional environment through time (DiMichele and Phillips, 1994; Wagner and Mayoral, 2007). In one study on the peat flora just above the Mahoning coal

6

paleobotany: the biology and evolution of fossil plants

in Ohio (Conemaugh Group, Pennsylvanian), DiMichele et al. (1996), utilizing macrofossils, palynomorphs, and coal petrology, concluded that the lepidodendrid trees (lycopsids—see Chapter 9) in this succession were flooded once, recovered, and then finally drowned by another flood event. This type of information can be utilized to recognize regional (Phillips et al., 1985) and global responses of plant communities to climate change (DiMichele et al., 2001a; Wagner and Mayoral, 2007). Understanding the interplay between these swamp-inhabiting plants and a variety of environmental parameters has now made it possible to interpret large-scale ecological shifts (e.g., the role of sea level fluctuations) in the community structure of the swamps, and to examine evolutionary questions within these habitats (DiMichele et al., 1985; DiMichele et al., 1996). These studies in turn have stimulated interdisciplinary research focused on broader questions, for example the evolution of major terrestrial ecosystems through time (Behrensmeyer et al., 1992; DiMichele and Hook, 1992). Paleoecological studies are very important in revealing the diversity of fossil communities inhabiting a geographic area (horizontal variation in floras) at the same time. Wing et al. (1993) examined fossil floras preserved in an ash fall and in fluvial deposits from the Upper Cretaceous (midMaastrichtian) of Wyoming, USA. They found that ferns (49%), along with palmettolike palms (25%), dominated the landscape, and that other angiosperms (Chapter 22), mostly herbaceous dicots, were dominant only in disturbed areas close to streams (fluvial deposits). The flowering plants were more diverse—constituting 61% of the species present—but they represented only 12% of the vegetational cover in this area. This study, and other similar ones (e.g., McElwain et al., 2007), have detailed the difference between the diversity of plants in fossil floras and the dominance of particular taxa within the paleoecosystem. Of course, it is only possible to fully comprehend the fossil assemblages, or taphocoenoses, by comparison with extant plants in various depositional environments (Spicer, 1981; Burnham, 1989, 1997) and by being aware of the taphonomy of fossil plants (Spicer, 1989) (see section on “Preservation”). Understanding and interpreting the sedimentological nature of the fossil assemblage, whether based on megafossils or microfossils, is only one of several aspects required in determining the diversification of plants through geologic time (Wing and DiMichele, 1995; Lupia, 1999). Paleoecologists use many of the same statistical methods used in contemporary ecological studies, including a variety of multivariate methods (Spicer and Hill, 1979; McCune and Grace, 2002). These tools, and many others, now make

it possible to examine the evolutionary and ecological processes that governed the plant communities which we now document as the fossil record (Jackson and Erwin, 2006). For a more in-depth approach to the study and methodologies of plant paleoecology see DiMichele and Wing (1988), Gastaldo (1989), and Jones and Rowe (1999). DETERMINING PALEOCLIMATE FROM FOSSIL PLANTS

Understanding climates of the past has become more and more crucial to appreciating the changes occurring on our warming planet today, and paleobotany is very important in providing baseline data to reconstruct past climates and in calibrating paleoclimate models based on physical parameters (Steppuhn et al., 2007). This area is rapidly expanding, so we will only cover a few of the many ways in which plant fossils can be used to reconstruct paleoclimate: TREE RINGS Data from fossil tree rings (paleodendrology) (FIG. 1.3) represent an important source of paleoclimate information, in some instances with very fine resolution, for example, major atmospheric disturbances (Miller et al., 2006). Although initially used for Holocene climate information (especially dating of archeological sites), some of the techniques used to analyze recent and subfossil wood have been extended to older material (Jefferson, 1982; Creber and Chaloner, 1984a; Creber and Francis, 1999; Taylor and Ryberg, 2007). Based on the changes in radial cell diameter within the tree rings and the variation in ring width (FIG. 1.3), it is possible to extrapolate climate information, which is especially useful when coupled with information from megafossils, microfossils, and the sedimentological record of the site. This approach has been utilized successfully by Parrish and Spicer in their work on Late Cretaceous floras from the North Slope of Alaska (Parrish and Spicer, 1988; Spicer and Parrish, 1990). More recently, Taylor and Ryberg (2007) have examined tree rings in Permian and Triassic woods from Antarctica. Based on their analysis using a variety of techniques, they suggest that the small amount of latewood indicates a very rapid transition to seasonal dormancy in response to decreasing light levels at these high polar latitudes. The mechanisms these plants evolved to cope with life in a polar light regime are of continuing interest in this and other studies based on plants that were once living at very high paleolatitudes. NEAREST LIVING RELATIVE The nearest living relative (NLR) method has been in use since the beginnings of paleobotany, particularly when

chapter 1 introduction to paleobotany, how fossil plants are formed

dealing with late Mesozoic or Cenozoic floras, as these are more likely to have close living relatives. It is based on the premise that climatic tolerances of the fossils are very similar to those of their NLR. The paleobotanist compares as many fossils as possible within a flora to their most closely related extant taxa; the more species in a fossil flora that have NLRs, the more precise the paleoclimate estimate, and the more closely related a fossil taxon is to an extant one, the more precise the method. It depends, therefore, partly on the paleobotanist’s ability to identify the fossils very accurately. The further back in time, the less effective this method is, as more and more extinct species or taxa which have no living relative appear. As a result, NLR has been used to best effect for Cenozoic angiosperm floras (Wolfe, 1995). This method can provide a general estimate of paleoclimate, but is limited by the fact that some fossil taxa do not have the same climatic limitations as their modern counterparts. LEAF PHYSIOGNOMY Leaf physiognomy analysis is a powerful technique that has been widely used in paleobotany to reconstruct Cenozoic paleoclimates. It is based only on angiosperms, however, so its applicability before the Cretaceous is uncertain (but see Glasspool et al., 2004a). Physiognomy is the general appearance of a plant, and it has long been known that plant physiognomy, especially leaf physiognomy, can be related to climate (Bailey and Sinnott, 1916). Physiognomy is primarily independent of taxonomy, for example plants with thick water-storing stems and leaves tend to grow in arid regions of the world, even though they may belong to a number of different families of plants. For fossil floras, this means that leaves do not have to be identified in order to obtain a paleoclimate signal. In his now-classic papers, Jack Wolfe (FIG. 22.276) presented the applications of leaf physiognomy to paleobotany, based on large collections of many leaves from extant floras, which he then was able to compare with Cenozoic angiosperm floras (Wolfe, 1993, 1995). Webb (1959) had previously completed a detailed physiognomic classification of Australian floras, and his definitions of leaf types are generally used in physiognomic methods today. There are presently two methods of leaf physiognomic analysis that are in general use: leaf-area and leaf-margin analysis. Leaf area directly correlates with mean annual precipitation (MAP). CLAMP (Climate-Leaf Analysis Multivariate Program; Wolfe, 1995) measures 31 leaf character states of woody dicots (Chapter 22) and uses multivariate analysis to map leaf shape in two-dimensional space (Wolfe and Spicer, 1999). CLAMP can provide a number of climatic parameters related to precipitation, humidity, and temperature.

7

Leaf-margin analysis (LMA) relies on the relationship between leaf margin (toothed versus entire) and climate (Greenwood and Wing, 1995; Wilf, 1997). Specifically, the proportion of leaves in the flora with toothed margins can be correlated with mean annual temperature (MAT), as toothed leaves are more abundant in wet environments. Both CLAMP and LMA can provide quantitative reconstructions of past climates, including estimates of MAT and MAP. More recently, paleobotanists have refined physiognomic methods by using computer image analysis to analyze both leaf-shape and leaf-margin morphology (Huff et al., 2003; Royer et al., 2005). Further data on the ecophysiology of modern plants and the function of various leaf shapes (Royer and Wilf, 2006) will no doubt help to refine these methods and improve their accuracy in the fossil record. Both methods are very robust, as both rely on large databases of leaf physiognomy of living leaves from many different sites and habitats. STOMATAL INDEX The stomatal index (the ratio of the number of stomata to the total number of epidermal cells plus stomata within a given leaf area expressed as percentages; see Salisbury, 1927) has been widely used in recent years to reconstruct past ρCO2 levels, as the stomatal index is inversely proportional to atmospheric CO2 levels. Woodward (1987) was one of the first to demonstrate the value of this relationship for ancient climate prediction, based on comparisons of modern leaves with herbarium specimens from preindustrial times. The best results have been obtained from comparisons of the same genus and species in order to control for genetic differences, so younger fossils, such as Holocene plants, have provided reproducible results (Wagner et al., 2004). The technique has been extended further back in time, for example the Cenozoic (Royer et al., 2001), as well as to the Mesozoic and Paleozoic, although there are limitations to the technique, especially with older fossils (Roth-Nebelsick, 2005; Uhl and Kerp, 2005). For studies in deep time, researchers have coupled CO2 estimates from stomatal indices with other proxy records, such as isotope data (Beerling, 2005 and papers cited therein). A summary of the pros and cons of methods to reconstruct past levels of atmospheric CO2 can be found in Royer (2001) and Kerp (2002). Details regarding the stomatal index technique can be found in Beerling (1999) and Poole and Kürschner (1999). SUMMARY

Throughout this book there are numerous examples of many of the biological and geological questions being asked by

8

paleobotany: the biology and evolution of fossil plants

paleobotanists today, and how the fossil record contributes to answering these questions. The field of paleobotany continues to advance, not only by the discovery of new fossils but also by the use of new methods applied to existing fossils and the application of techniques from other fields to paleobotany. Plant parts preserved in different ways or ones that show additional features are continually being discovered. More sophisticated and improved methods to study the fossils and interpret the results also provide new data which contribute to an enhanced understanding of the plants and communities that existed through geologic time.

PRESERVATION: HOW PLANT FOSSILS ARE FORMED AND PRESERVED A relatively small fraction of the plants and other organisms that live on the Earth at any particular time will ever become fossils. Most dead plant material is decayed by aerobic (oxygen-loving) fungi and bacteria. So, the first requirement for fossilization is that dead plants must be deposited in an environment where air is excluded, that is an anaerobic environment. This usually involves deposition in a body of water (discussed below), but not always. Once deposited, the plants must be buried by sediments so that air is excluded. In addition, these sediments must have enough acidity that anaerobic decay is also reduced. Paleobotanists are often asked the question, where do you look for fossil plants? The answer is that they typically are found in places where the rocks containing them have been exposed in some way (FIG. 1.8); these rocks may be as far away as the Arctic (FIG. 1.4) or the Antarctic. Because streams and rivers cut down through the rocks, exposed strata along waterways are often excellent sites to prospect for fossil plants. Erosion by water in many other places also exposes fossil-bearing rocks. Sometimes it is possible to find plants in eroded cliffs along seashores. In addition to the natural exposure of plant-bearing strata, excavations are frequently the source of many fossils. Road cuts, for example, often reveal fresh surfaces with unweathered rocks that contain wellpreserved fossils. As might be expected, quarries and mines are rich sources of fossil plants, revealing rocks that would otherwise have been inaccessible to paleobotanists. Coal balls (FIG. 1.43), a type of permineralization, are frequently encountered in coal mines, and often the shales immediately above the coal seams in such mines contain abundant fossil plants. Quarries in which clay is being excavated for bricks,

Figure 1.8 Students collecting fossil plants from a narrow lens of fine-grained shale.

tiles, or pottery are sites that often provide fossils. In fact, almost any massive construction site, such as for a dam, a hydroelectric plant, or a building with a deep foundation, can, and has, yielded an abundance of fossil plants. DEPOSITIONAL ENVIRONMENTS OF FOSSIL PLANTS

Fossil plants are found in almost all regions of the Earth, the most notable exceptions being recent volcanic islands or in rocks that have been extensively metamorphosed (FIG. 1.9). Marine plants, such as various forms of algae (Chapter 4), are generally found in rocks deposited in marine environments (e.g., nearshore deposits, carbonate shelves, etc.). Although land plants are occasionally found in marine rocks, generally, wherever terrestrial sedimentary rocks occur, there is a good chance that fossil plants will be found in them. Sedimentary rocks are those formed by the accumulation of rock particles derived from the weathering and mechanical abrasion of existing rocks. The great majority of sedimentary rocks are formed by deposition of particles in water, but wind deposits (e.g., eolian sands, loess) can also form, and rock breakdown can occur by chemical weathering, with rock components being released into solution, later to solidify at some other place. Plant parts are typically fossilized, then, in areas where sediment is accumulating. The delta of a river is just such a depositional environment. As the course of the river constantly shifts, channels are abandoned and new ones initiated; natural levees are destroyed during flooding, and new ones built up later. A meandering river cuts into the bank on the outside of each meander, and deposits sediment on the inside of meanders, often covering plants growing along the water. When the river breaks

chapter 1 introduction to paleobotany, how fossil plants are formed

9

Figure 1.9 Exposed rocks in Antarctica. Zone of dark red rocks are volcanic and lack fossil plants, including microfossils.

through a levee, a rush of water and sediment, called a splay deposit, can rapidly cover the adjacent floodplain, inundating the plants growing there. Associated with the deltaic system are abandoned stream channels, often referred to as oxbow lakes, and vegetation growing along the banks of these abandoned channels may be undisturbed for some time. If a subsequent flood destroys the levee, knocking down trees and other plants growing on it, these plant parts can be carried to abandoned channels and other places where a high concentration of sediment would bury the plant fragments and fill in the oxbow. As might be expected, plant parts carried for great distances would tend to be fragmented and shredded, and those deposited close to the place of growth would be less distorted and better preserved. Plants that become preserved at the same locality where they were growing are termed autochthonous (e.g., many Pennsylvanian coal ball deposits), whereas those assemblages that have been transported are termed allochthonous. Preservation of whole plants or plant parts (usually stems and roots) in growth position is termed in situ. The plants that once made up a community, together with the other organisms in the ecosystem, are preserved in the Earth’s crust in a variety of ways, and different kinds of physical and chemical processes were involved during the process of preservation (FIG. 1.6). Moreover, various environmental settings and depositional processes also result in fossils that occur in a variety of forms. Taphonomy is the study of all the processes occurring between the time the organism died and its discovery today as a fossil. These include burial by sediments of some type (e.g., sand, fine mud, ash, etc.), and diagenesis, which is defined as all the chemical and physical changes to the sediment (and the fossils within

Figure 1.10 Cuticle preparation of the seed fern Blanzyopteris praedentata showing numerous trichomes extending from the surface. Bar  1 mm.

it) as it is converted into rock (Gastaldo, 1989; Gee and Gastaldo, 2005). Because of the countless ways in which plants are preserved, the paleobotanist must employ different techniques to extract the maximum amount of information from fossils. For example, when a paleobotanist finds a fossil leaf, it would first be compared to modern leaves, based simply on the overall size and shape, that is, the morphology of the leaf, to identify it. This can include describing the shape and distribution of teeth on the margin of the leaf, if present, and the shape of the base of the leaf compared to the tip, as well as the length and shape of the petiole. Next, the discoverer might examine the pattern of veins in the leaf—the venation, followed by a microscopic examination of the types and distribution of stomata (pores for gas exchange) and other structures on the surface, such as hairs (trichomes) (FIGS. 1.10, 1.11), or trichome bases if the hairs themselves are no longer attached to the leaf. Still later, the paleobotanist might study the ultrastructure of the cuticle (the waxy

10

paleobotany: the biology and evolution of fossil plants

Figure 1.11 Multicellular, uniseriate trichomes on Blanzyopteris praedentata cuticle. Bar  400 μm.

covering on leaves), or the molecules that remain part of the leaf after diagenesis (geochemistry). It is possible to determine the proportion of carbon isotopes (13C versus 12C) in many fossil plants, and to use these to reconstruct paleoenvironment or the type of photosynthetic pathway (C3 versus C4 photosynthesis) employed by the plant. And in the future? Perhaps paleobotanists will be able to extract information that reveals details about biochemical pathways, developmental mechanisms, and families of genes involved in response to parasites or herbivores that attacked the leaf surface, all from a fossil plant leaf! Although there are numerous variations on the ways in which plant fossils are preserved, there are a few basic types which we discuss later. It is important to keep in mind, however, that all preservation types can intergrade, or a fossil plant may be preserved in more than one way, for example, a compressed plant with a stem that is partially petrified. Finally, there may be certain structures that appear to be an organism, but are not of organic origin. One of the most common of these are dendrites (FIG. 1.12).

Figure 1.12 Pseudofossil (dendrites) that look like the leaves

of a plant, but are manganese oxide that has grown on the bedding plane of a limestone (Jurassic). Bar  2 cm. (Courtesy BSPG.)

COMPRESSIONS

As sediments accumulate, such as in an oxbow lake, water is squeezed out, so the sediments become much more compact, and plant fragments contained within them become flattened (Rex and Chaloner, 1983; Chaloner, 1999a). Internal structure is usually obliterated as the cells become flattened, and frequently all that is left is a delicate carbonaceous film that conforms to the original outline of the plant part. This type of fossil is called a compression (FIG. 1.13), and it is one of the most common types of plant fossils. As you might predict, if the sediment grains that bury the plant parts are large and angular, such as sand grains, the resulting compression

Figure 1.13 Compression specimen of Osmunda claytoniites

from the Triassic of Antarctica. Bar  2 cm.

shows less detail than if the sediment particles are smooth edged and very fine, such as clay particles (FIG. 1.14). There is a vast range of sediment size and structure; plant parts have even been preserved in conglomerates (rock made up

chapter 1 introduction to paleobotany, how fossil plants are formed

11

Figure 1.14 Portion of compressed fern frond (Cretaceous).

Bar  2 cm.

of variously sized pebbles), but generally, compressed plants will be better preserved in clays or shales than in sandstone deposits. Compressions are not always formed in deltaic and fluvial (river) systems; they may be formed in lagoons, along meandering rivers (not necessarily near deltas), and in ponds, swamps, or other depositional systems, as well as in wind-blown sediments. Most often a terrestrial ecosystem is involved (as opposed to a marine environment), although there are instances in which terrestrial plants are even preserved in marine limestones. Plant compressions can also be found in consolidated volcanic ash. These fossils represent plants growing near an area where there was volcanic activity that spewed clouds of ash into the air. Often there is severe atmospheric turbulence near a volcano and thunderstorms may develop as a result. The rainwater and the ash make a fine-grained mud which cascades down the hillsides, picking up and burying plant parts as it goes (Burnham and Spicer, 1986). When the mud hardens, it entombs pieces of plant material, in many cases exactly in the position in which they grew (in situ). For example, in the Cretaceous Baqueró Group in Santa Cruz Province,

Figure 1.15 Portion of a fern frond preserved in tuff from

Argentina (Cretaceous). Bar  2 cm.

southern Patagonia, the tuff deposits were laid down so rapidly that it is possible to trace the fern Gleichenites vertically from the rhizome through the sediments to the leaves (FIG. 1.15) (Archangelsky, 2003). So well preserved are some of the compressions from this site that the plant material can be sectioned and examined with the transmission electron microscope (Archangelsky and Villar de Seoane, 2004). Fossil cytoplasm has even been described in one unusual compression specimen (Hall, 1971) (FIG. 1.16).

12

paleobotany: the biology and evolution of fossil plants

Figure 1.17 Compression of pinnae in which only the axes and

pinnule venation are preserved (Pennsylvanian). Bar  1 cm.

Figure 1.16 Pollen grain Kollospora extrudens with contents

(cytoplasm?) extending from the wall. Bar  35 μm. (From Hall, 1971.)

A very unusual kind of matrix in which compressions occasionally occur is diatomite—a rock formed from the siliceous shells (frustules) of diatoms (Bacillariophyceae; see Chapter 4), which today inhabit both fresh water and marine sites. Diatomite is especially fine grained and preservation of plant remains in it is often superb. Since the diatom frustules are actually the cell walls of these microscopic algae, in this method of preservation one organism is serving as the preserving matrix for another organism. Leaves are some of the most common plant parts preserved as compressions (FIG. 1.13) and, in many instances, they occur in numerous, closely spaced layers within the rock matrix. Often a collector can uncover the leaves by splitting the rock along bedding planes with a knife (if the matrix is clay) or with a hammer and chisel (if the rock is harder), although sometimes the paleobotanist must resort to more energetic measures to uncover fossils, such as using a jackhammer or even dynamite! Many compressions are of value in showing surface details and overall morphology. Experimental evidence suggests, however, that the size and shape of fossils can vary depending on the matrix in which they are preserved (Rex, 1986). Such features as leaf shape, presence or absence of

a petiole (leaf stalk), leaf margin, trichomes, and the pattern of venation (FIGS. 1.17, 1.18) are generally readily discernible. In some cases it is possible to examine the distribution of stomata in the leaf surface. When there is an abundance of leaves presumably from one species of plant, it is possible to determine the degree of variability exhibited. In other cases this is far more difficult, especially as leaves tend to be the most plastic in their morphology of any plant part. For example, it is often difficult to find two leaves that are morphologically identical on some modern plants, such as the common mulberry tree (Morus spp.). Many conifers (Chapter 21) produce juvenile leaves with a different morphology than that of mature leaves. When found as fossils, these may have been described as two different species. For these reasons, paleobotanists must be extraordinary sleuths in uncovering features that will help distinguish variability within a single species (intraspecific variability) from differing leaf forms that reflect different species (interspecific variability) or genera. Compressions preserved as relatively dark carbonaceous films on a lighter colored rock matrix make examination and imaging of the fossil relatively easy. Sometimes, however, the matrix and the fossil have similar color values and imaging is more difficult. In these cases, details can be enhanced by using sidelighting (a light source at an oblique angle to the compression) or a polarized light source, or by submerging the fossil in some liquid, such as water, xylene, or alcohol. In many instances, the use of cross-polarization

chapter 1 introduction to paleobotany, how fossil plants are formed

13

Figure 1.19 Arctocarpus cuticle showing paracytic stomata

with cuticular ridges (Cretaceous). Bar  4 μm. (Courtesy G. R. Upchurch.)

Figure 1.18 Cuticle preparation showing venation of Barthelopteris germari pinnule. Bar  1 mm.

(polarized light sources together with a polarizing filter over the camera lens) can significantly enhance contrast (Schaarschmidt, 1973; Crabb, 2001). Another method that has been used with compressed animal remains uses backscattered electron imaging to help differentiate among superimposed anatomical features compressed into a single plane (Orr et al., 2002). CUTICLE Although most compressions show only superficial details, in some instances it is possible to learn a great deal about cellular details of the epidermis from preservation of the cuticle (FIG. 1.19). Primary aerial parts of all vascular land plants are covered with a thin film of waxy material, the cuticle, which prevents excess water loss from the surfaces of the plant. Cuticle is continuous over the surface of the plant, except over the stomatal openings; it is a non-cellular, amorphous layer that is deposited on the outside and into the walls of the epidermal cells. It closely conforms to the contours of the surface of the epidermal

cells and may also extend slightly downward between these cells in flanges which are perpendicular to the surface of the plant. Plant cuticles, as well as waxes deposited on top of the cuticle, are important for the plant in controlling transpiration—the movement of water through the plant, from the roots to its eventual evaporation from the leaves. Most water is lost from stomatal openings, but some can evaporate through the cuticle if it is thin enough or its texture allows for cuticle transpiration. Cuticle is also important in control of gas exchange with the environment, in repelling water from the leaf surface, in attenuation of photosynthetically active radiation (PAR), and in blocking ultraviolet (UV) radiation. The cuticle, especially the leaf cuticle, serves as an interface for a host of biotic interactions between plants and other organisms in their environment, including parasites and herbivores (Riederer and Müller, 2006). Because cuticle is inert and resistant to decay, it is widely preserved in the fossil record, and represents a valuable source of paleobotanical information (Mösle et al., 1997). The leaf cuticle is often preserved as an intact envelope which once surrounded the living leaf tissue (FIG. 1.20). Many fossil cuticles are fragile and must be protected prior to transport from the collecting site. One way to do this is to apply a mixture of nitrocellulose to the surface of the fossil (LePage and Basinger, 1993, 1994), thus preventing loss and breakage of specimens as the freshly excavated sediment dries. Common nail polish has also been used in a similar way. Covering the cuticle with a preservative in the field, however, may prevent subsequent geochemical study of the cuticle (Collinson, 1987).

14

paleobotany: the biology and evolution of fossil plants

Figure 1.20 Cuticle preparation of Odontopteris brardii pinnule. Bar  1.5 mm.

Figure 1.21 Laboratory set up for preparation of cuticle

Figure 1.22 Pinnule of Pseudomariopteris cordato-ovata showing adaxial epidermal anatomy (Pennsylvanian). Bar  1 mm.

mounts from Eocene clay. Note staining jars at left and pieces of cuticle before and after bleaching, right. (Courtesy D. L. Dilcher.)

It is possible to remove the cuticle from many fossil leaf specimens either mechanically, with a needle or brush, or chemically by dissolving away the rock matrix. Pieces of cuticle retrieved in this way can then be bleached and stained with common biological stains, such as safranin (Bartholomew et al., 1970; Dilcher, 1974; Kerp and Krings, 1999; Krings, 2000a) (FIG. 1.21). When mounted on a slide and examined under a microscope, the cuticle or cuticular fragments reveal considerable epidermal detail (Kerp, 1990). Cuticles of epidermal cells are apparent (Fig. 1.22), along

with the structure of the stomatal complex (the cells associated with the pores, i.e., stomata, in the leaf (Fig. 1.23)), the distribution of stomata, presence of hairs or glands, and other distinguishing features. The cuticle in plants is very much like a fingerprint, in that many species have distinctive epidermal features and patterns that can be useful in identification. Furthermore, it is often possible to demonstrate that disarticulated plant parts, such as leaves, stems, flowers, and seeds, actually belong to the same plant because the individual parts have the same complement of cuticle structures.

chapter 1 introduction to paleobotany, how fossil plants are formed

Figure 1.23 Cuticle preparation of Dicksoniites pluckenetii

stomatal apparatus. Bar  20 μm.

15

Cuticle and epidermal features can be investigated by transferring the compression fossil from the rock matrix in the form of a transparent film, which can then be examined under a microscope. The film can be made by pouring on a liquid-plastic substance (e.g., clear fingernail polish), letting it dry, and then teasing away the film, with the cuticle adhering to it, from the rock matrix. A comparable technique is to embed the surface of the fossil in liquid plastic (such as that used for preparing biological mounts) and then macerating away the rock with an appropriate acid. In some cases, the cuticle can simply be removed from the surface of the rock with a dissecting needle, without the need for transfer film or maceration. Cuticles can also be transferred onto polyester overlays, which reduces the time of preparation and preserves the fossil from which they were taken (Kouwenberg et al., 2006). In addition to preservation of the cuticle (FIG. 1.23), or some chemically altered form of it (see Gupta et al., 2006), some compressions contain cellular remains as part of the carbonaceous layer (Niklas, 1981a). Niklas et al. (1978) embedded and sectioned exquisitely preserved, compressed fossil leaves from the Succor Creek Formation (middle Miocene) of Oregon, USA, for transmission electron microscopy (TEM) and showed that the cellulosic microfibrillar organization of the cell walls could be seen. Even more astounding was the fact that organelles within the mesophyll cells of the leaves, including chloroplasts with grana stacks (FIG. 1.24) and starch deposits, nuclei (FIGS. 1.25,

Figure 1.24 Stacks of grana membranes in the chloroplast of a fossil Betula leaf (Miocene). Bar  100 nm. (From Niklas, 1981a.)

16

paleobotany: the biology and evolution of fossil plants

Figure 1.25 Lycospora spore with structures interpreted as

possible chromosomes (Pennsylvanian) Bar  12 μm. Figure 1.26 Tetrad of Flemingites spores showing cell contents interpreted as nuclei (Pennsylvanian). Bar  20 μm.

1.26, 10.20) with condensed chromatin, and plasmodesmata, were preserved in these fossils! The possibility always exists, however, that subcellular structures may be contaminants or artifacts formed as a result of the compression and dehydration of other cell components during diagenesis (Niklas, 1982). Chloroplasts have also been reported from Eocene leaves of Metasequoia collected from the Canadian High Arctic (Schoenhut et al., 2004). These authors suggested that the extraordinary preservation may have resulted from high concentrations of tanniniferous cells in the leaves, which may have inhibited microbial degradation and thus left the cell organelles intact. Some compressed Eocene angiosperm leaves from the Geiseltal in Germany and Clarkia beds in Idaho are still green when the rock is split open (FIG. 1.27), which suggests that the chlorophyll is still intact (e.g., Dilcher, 1967).

Figure 1.27 Compressed angiosperm leaves from the Clarkia

beds with chlorophyll preserved (Miocene). Bar  3 cm.

BIOFILMS AND PLANT FOSSIL PRESERVATION The reason that some very delicate structures are preserved is difficult to understand, but in recent years there has been great progress in elucidating the role that various microorganisms, especially those in biofilms, can play in the preservation process (Borkow and Babcock, 2003). Biofilms consist of an aggregation of microorganisms held together in a slimy matrix of extracellular polysaccharides, which are secreted by certain bacteria in the biofilm. We now know that biofilms are ubiquitous on the Earth, and can be found in environments ranging from streams to desert crusts, to hot

springs; the dental plaque on your teeth represents a type of biofilm. One of the first researchers to recognize the importance of biofilms in fossil preservation was Jean-Claude Gall (1990) in his studies of the beautifully preserved, soft-bodied organisms in the Early Triassic Grès à Voltzia Formation (Voltzia Sandstone) in northeastern France. These organisms, both plants and animals, are believed to have been rapidly covered by biofilms which entombed the animals in lowoxygen conditions that inhibited decay. For more information

chapter 1 introduction to paleobotany, how fossil plants are formed

17

TEM of fossil cuticles has proved useful in detailing the intricate structural organization of the stomatal complex in certain fossil plants and in understanding preservation processes (Villar de Seoane, 2003). Fossil cuticles have also been examined for their chemical constituents (Tegelaar et al., 1991; Mösle et al., 2002; Gupta et al., 2006). Some cuticles are too thin for standard preparation techniques to be effective, or have been fragmented into minute pieces during fossilization so that they cannot be removed from the rock surface. Under these circumstances, incident light, dark field, or epifluorescence microscopy have been useful in revealing certain cuticle and epidermal characters (Kerp and Krings, 1999; Thomas et al., 2004). Figure 1.28 Surface of Pseudofrenelopsis cuticle showing

distribution of stomata (Cretaceous). Bar  15 μm. (Courtesy C. P. Daghlian.)

on this fascinating subject, see the excellent compendium of papers in Krumbein et al. (2003a). Fossil deposits which show a great diversity of organisms preserved, or excellent preservation, or both, are called Lagerstätten (sing. Lagerstätte). Lagerstätten of various ages have provided paleobotanists with a wealth of information on plants of the past (see, e.g., Chapter 6). ELECTRON MICROSCOPY Scanning electron microscopy (SEM) has become a commonly used research tool in paleobotany to illustrate pollen grains and plant cuticles, because of the extensive range of magnifications available (up to 100,000 times) and the extreme depth of focus that can be achieved. In some instances, compressed leaf surfaces and various structures on them (e.g., stomata and trichomes) can be examined directly with the SEM (FIG. 1.28). In other cases, it is necessary to make latex replicas of the plant surfaces in order to interpret complex structural details. TEM has also been employed in the study of fossil plant cuticles. For TEM studies the plant cuticle is embedded in an appropriate embedding medium (e.g., Spurr epoxy resin), sectioned on an ultramicrotome, and stained in much the same way as living plant tissues are prepared for ultrastructural examination. Many fossil cuticles reveal lamellae and delicate structural features similar to those in modern cuticles (Archangelsky and Taylor, 1986; Guignard and Zhou, 2005). Varying patterns in the ultrastructure of cuticle from the same leaves have been documented and suggest that such differences may reflect cuticle from sun and shade leaves (Guignard et al., 2001). In addition,

CONFOCAL MICROSCOPY Recently, three-dimensional confocal laser scanning microscopy (CLSM) has been added to the arsenal of tools used by paleobotanists to extract information from the fossil record (Schopf et al., 2006). This technique utilizes a sequence of closely spaced images that can provide information in three dimensions. The technique is non-destructive and noninvasive, and has been successfully applied to Precambrian microscopic fossils in order to characterize not only morphology, but the nature of the preservation, including possible cell contents. Because the specimens examined must provide an autofluorescent signal, the fossils cannot have been geochemically altered. Like many techniques used in paleobotany, the procedure needs to be investigated on a particular fossil and, if the desired information cannot be obtained, the investigator needs to modify the technique or explore another means of obtaining the information needed. MACERATION AND DÉGAGEMENT Doran (1980) employed a bulk maceration technique to study Devonian compression fossils that were preserved in a silicified tuff matrix. He submerged the rock in hydrofluoric acid (HF) until the fossils were freed from the matrix. This technique provided nearly complete plants, and Doran could thus more accurately reconstruct the complete morphology of the plants. This method is useful because plant axes that extend into the matrix can be totally freed, but it only works on material where enough organic matter is preserved for the plants to remain intact through the maceration process. A more widely used technique to uncover compressed plant parts within the rock matrix is dégagement. This technique was developed primarily by Suzanne Leclercq (1960) and involves removing the rock matrix—often grain by grain— using fine steel needles (FIG. 1.29).

18

paleobotany: the biology and evolution of fossil plants

valuable for studying specimens preserved in highly metamorphosed shales where much of the specimen is covered by the matrix and where bedding planes are poor. This technique not only provides details that cannot be obtained with conventional techniques but also makes some threedimensional reconstructions possible through the use of stereoscopic X-ray analysis. Uncovering the three-dimensional nature of fossil plant parts that are compressed or otherwise encased in the rock matrix is a goal in many paleobotanical studies. More recent improvements in the capture and analysis of X-ray images suggest that this technique will be more widely used in paleobotany in the future (Dietrich et al., 2000). X-ray computed tomography (CT) scans have been widely used in medicine, and vertebrate paleontologists have adapted these methods, using high-resolution scans (HRXCT), as a non-destructive method to produce three-dimensional images of vertebrate bones (Conroy and Vannier, 1984). Only recently, however, these methods have been used on fossil plants. Devore et al. (2006) used HRXCT to image the morphology and anatomy of pyritized fruits and seeds from early Eocene London Clay Formation. Because fossils preserved the in pyrite are fragile and deteriorate over time if exposed to air, they are conserved by placing them in sealed tubes of silicon oil. Using HRXCT it is possible to examine the fossils without removing them from the vials, thus decreasing the chances of exposing the specimens to air and subsequent deterioration. This technique makes it possible to study type specimens non-destructively and to reexamine characters that were initially used to define taxa and to evaluate forms for which the taxonomic affinities remain equivocal. It also preserves the integrity of the fossils so that they may be utilized in subsequent studies, perhaps when other, newer techniques are developed (Matysová et al., 2008). COAL AND CHARCOAL

Figure 1.29 Impression-compression of stem surface of Colpodexylon deatsii (Devonian). Bar  2 cm.

OTHER TECHNIQUES Although X-ray analysis has been used for many years by paleontologists working with animal fossils, historically this technique has been little used for fossil plants (Stürmer and Schaarschmidt, 1980). X-ray analysis has been especially

Technically, coal (FIG. 1.30) comes under the definition of a compression fossil, since it represents a complex, heterogeneous mixture of macromolecular organic compounds derived from plant material that has been compressed over time (Scott, 1987). In general, the lower the rank of the coal (the degree of coalification), the more details of plant structure one can observe. The higher the rank, the more the coal has been metamorphosed and the higher the carbon content. Ranks from lowest to highest are lignite, subbituminous coal, bituminous coal, and anthracite. Lignite represents an early stage in coal formation, so the plant parts within lignite are not excessively crushed or decayed and are generally easily recognizable. In some instances, SEM has been a

chapter 1 introduction to paleobotany, how fossil plants are formed

19

Figure 1.30 Exposure of coal in Antarctica (Permian). Ham-

mer  30 cm.

useful tool in identifying plant parts preserved in lignites (Alvin and Muir, 1969). In some lignites it is possible to tease apart the plant fragments and make whole mounts of various structures, for example in the Brandon lignite, a famous early Miocene plant locality in Vermont, USA (Haggard and Tiffney, 1997). Bituminous coal is more metamorphosed and the plant parts are more flattened, but it is still possible to study plant fragments within the coal. Anthracite coal, the most highly metamorphosed type, is altered to such an extent that little of the original plant material is recognizable. Some coals can be thin sectioned for microscopic examination, and pollen grains, spores, and fragments of cuticle can be discerned. In other instances it is possible to coat the polished surface of coal with epoxy resin and etch it in a low-temperature plasma asher (Winston, 1989). Pieces of coal, or peels of the etched surfaces, can then be examined by light microscopy or SEM to determine the biological composition of the coal. This procedure makes it possible to quantify the plants in various types of coals in instances where coal ball permineralizations (see section “Cellular Preservation”) are not available (Winston, 1986). Coal can also be macerated using chemicals that break down the solid coal and release the plant fragments. It is possible to recognize cuticle, pieces of bark, bits of wood, solidified resins, and especially spores and pollen grains in this type of preparation. Examination of these components allows one to determine the kinds of plants that were growing in the ancient swamps where the coal was formed. Application of 13C nuclear magnetic resonance (NMR) and pyrolysis–gas chromatography–mass spectrometry techniques has been used to define stages in the coalification process more accurately (Hatcher et al., 1989). This same technique has also been used for Cenozoic

Figure 1.31

Marie C. Stopes.

leaf tissues and wood to identify various biomolecules (Yang et al., 2005). The components of coals can also be useful in documenting paleoecology (Poole et al., 2006). Macerals are defined as the organic constituents that comprise coal as seen in polished thin sections. The system of maceral types was originally proposed by the paleobotanist Marie C. Stopes (FIG. 1.31) in 1919, expanded in 1935, and is constantly kept up to date (ICCP, 2001). Maceral names end in -inite; for example, funginite is made up of fungal spores and various fungal bodies, secretinite is composed secretory deposits formed by medullosan seed ferns (Lyons, 2000), and sporinite consists of the sporopollenin walls of fossil pollen and spores. More detailed studies of coal composition provide valuable information about environmental parameters. For example, G. Taylor et al. (1989) suggested that the association of alginite and inertodetrinite (redeposited small particles of fusinite) in the Permian coals of Australia indicates a paleoenvironment of wet, cool summers and freezing winters. It is possible to distinguish between angiosperm and gymnosperm woods in some coals using macerals (Sýkorová et al., 2005). Fossil charcoal or fusain (carbonaceous residue that results from the incomplete combustion of organic material;

20

paleobotany: the biology and evolution of fossil plants

also called fusinite) is also an important source of paleobotanical data (Cope and Chaloner, 1985; Lupia, 1995), with charcoalified plant remains dating back to the earliest land plants (Glasspool et al., 2004b). There are several techniques used to examine fossil charcoal (Sander and Gee, 1990; Guo and Bustin, 1998; Figueiral et al., 2002) which provide information on taphonomy and paleoecology (Scott et al., 2000), including past atmospheric composition (Scott and Glasspool, 2006) and the presence of fire in paleoecosystems (Uhl et al., 2004, 2007a; Collinson et al., 2007). The discovery of beautifully preserved charcoalified flowers in Cretaceous (Tiffney, 1977; Friis and Skarby, 1981) rocks from around the world has contributed large amounts of information to our knowledge of early flowering plants (see Chapter 22). In North America, Carboniferous coals of different ages are typically characterized by the dominance of different types of swamp plants (Cross and Phillips, 1990). Among the Pennsylvanian coal beds for example, lycopsids, tree ferns, calamites, seed ferns, and cordaites constitute the major types of tropical–subtropical arborescent plants that contributed to the peat formation. Although Carboniferous peat swamps have represented the model system in most interpretations of coal-forming ecosystems, the plants lived in atypical terrestrial communities in which the pH and available nutrients were low. It is now becoming apparent that such factors as types of litter accumulation, nature of the biomass, preservational characteristics of certain tissue systems, microbial diversity, biology of the plants, paleoclimate, and paleogeography are but a few of the parameters necessary to understand and properly interpret coal-forming ecosystems through geologic time (Cross and Phillips, 1990; DiMichele et al., 2002, 2007a). An important study by Gastaldo et al. (2004) underscores the fine resolution needed to understand fossil plant community structure, using an in situ three-tiered forest above a Pennsylvanian coal in Alabama, USA, as the data set. Detailed sampling of this fossil forest indicated that the proportion of canopy, understory, and ground-cover plants was variable across the study area, and that wet–dry gradients and/or increasing habitat specialization did not control the distribution of the plants species in this swamp ecosystem. In rare instances, a coal is formed that consists entirely of cuticular fragments and amorphic organic material (DiMichele et al., 1984). The cuticle is so abundant that it can be peeled off in thin layers. This type of coal is termed a paper coal, alluding to its papery appearance, and is known from relatively few localities, some as early as the Devonian, for example the famous Orestovia paper coal from Siberia

(Ergolskaya, 1936a; Krassilov, 1981a). It is a simple matter to isolate these cuticular fragments by using a chemical base such as potassium hydroxide. The cuticle can then be washed, stained in some cases, and mounted directly on slides for examination (DiMichele et al., 1984; Kerp and Barthel, 1993). Lenses of leaf fragments may sometimes be preserved within coals; these apparently formed in small depressions containing acidic water, which inhibited the normal degradation activities of various microorganisms (Gastaldo and Stub, 1999). In other instances, coals have been found to be made up exclusively of algal remains (see Chapter 4), some as early as the Precambrian (Tyler et al., 1957). Kerogen is a type of fossilized insoluble organic matter that is widely found in sedimentary rocks, and is a common component of various paleobotanical preparations. It is the most abundant form of organic carbon on Earth—more even than coal deposits. The presence of kerogen in rocks has been used as evidence of some of the earliest life on Earth (Moreau and Sharp, 2004; see Chapter 2). Understanding the chemical composition and source of kerogen, termed “typing” the kerogen, is especially important, since these factors help to determine the petroleum-generating potential of source rocks. In the past, kerogen and coal were generally analyzed using just thin sections and light microscopy. Today, both substances are also characterized using standard geochemical methods, such as pyrolysis, gas chromatography– mass spectrometry (GC–MS), analysis of carbon isotopes for total organic carbon (TOC), and others. There are a variety of other techniques available today to investigate the nature of the organic matter that remains after fossilization and to compare the carbonaceous residue to determine the chemical-structural characteristics. Raman spectroscopy has been used to characterize carbonaceous matter in highly metamorphosed rocks for some time (Nestler et al., 2003), but has recently been applied to microscopic fossils preserved in chert of varying ages (Schopf et al., 2005). Unlike most geochemical techniques, Raman spectroscopy is a non-destructive means to analyze ancient organic matter. It provides information on the original biochemistry of the organism, and can help resolve the nature of certain ancient fossil-like organisms (see Chapter 2). Another approach that has been used to examine the chemical composition of fossil plant materials involves energydispersive X-ray microanalysis (EDXMA). With this technique the elemental composition and spatial distribution of fossils can be studied without damaging the specimen (Briggs et al., 2000). Other have used EDXMA to map the distribution of elements in fossil cells (Boyce et al., 2001), and cell walls (Boyce et al., 2002).

chapter 1 introduction to paleobotany, how fossil plants are formed

Figure 1.32 Impression specimen of Osmunda claytoniites

from Triassic of Antarctica. Bar  2 cm.

Figure 1.33 Impression of several whorls of Annularia stellata

leaves (Pennsylvanian). Bar  1 cm. IMPRESSIONS

When a paleobotanist splits a rock that contains fossil plant fragments along a bedding plane, it is sometimes possible to see the carbonaceous film of a compression along one face, and a negative imprint of the plant part, with little or no carbon adhering, on the other face (FIG. 1.32); these two faces are called part and counterpart in paleobotany. The fossil with little or no carbonaceous material is called an impression

21

1.34 Impression of Sigillaria leaf bases showing parichnos scars and position of leaf trace (Pennsylvanian). Bar  1 cm. (Courtesy BSPG.)

Figure

(FIGS. 1.33, 1.34). The imprint will show all the surface details of the compression, such as leaf shape and venation, but there is no actual plant material, that is no carbon, preserved. If you have ever seen a leaf imprint in a concrete sidewalk, you have seen an impression. The process involved in the formation of an impression is also analogous to these modern “fossils.” Such imprints are formed when leaves fall and settle into the wet concrete just after it is poured. As the concrete hardens, it conforms to the contours of the lower side of the leaf that rests on it. Eventually, the leaf disintegrates and the pieces are blown away, but a negative replica of the leaf remains on the hardened concrete. If you have ever put your initials in wet concrete, you have formed an impression. The impression of dinosaur footprints represents an excellent example of this type of fossilization process. When several footprints are of the same type or a series of trackways are discovered in close proximity, it may be possible to extrapolate the stride of the organism and, from this, infer something of the biomechanics of the animal. No cellular details can normally be seen on an impression because there is no adhering organic material, but, in some instances, especially where the matrix is exceedingly fine grained, a replica of the impression can be made with latex or similar material. The replica faithfully reproduces whatever surface details were on the original organism when it was impressed into the mud. Examination of part of the replica with the SEM may reveal details with great clarity, such as the pattern of the epidermal cells, hairs, glands, or other surface features. Some impression fossils are covered with mineral encrustations of different composition, for example iron (Spicer, 1977). These deposits may be the result of the

22

paleobotany: the biology and evolution of fossil plants

Figure 1.35 Cast of large seed fern seed (Pennsylvanian).

Bar  2 cm.

activities of microorganisms during the decay process. Regardless of their origin, however, the mineral crust may provide an excellent replica of the surface of the plant part, and this can be studied using a variety of imaging modes. MOLDS AND CASTS

Figure 1.36 Cast of arborescent lycopod (Protostigmaria egg-

In addition to two-dimensional plant parts, such as leaves, three-dimensional structures, such as stems, seeds, or fruits, can also be carried into sites where sediment is accumulating and buried. During flood events, massive trunks and tree branches can be moved some distance before they are eventually deposited. If these plant parts became crushed over time, they would be preserved as compression or impression fossils. If, however, the sediment surrounding the threedimensional plant parts hardens before the plant fragment is crushed, the sediment will form a three-dimensional negative, or mold, of the plant fragment. As the plant material disintegrates, a hollow remains in the sediment, which can subsequently be filled in with sediment, thus forming a cast inside the mold. The surface of the cast and the mold can often faithfully reproduce the surface features of a particular plant part, such as characteristic leaf bases on the surface of a stem or the ornamentation of seeds (FIG. 1.35) and fruits. The sediment that fills in the cavity of the mold and solidifies becomes a three-dimensional cast of the original plant part

(FIG. 1.36). In almost all molds and casts no actual plant material remains, but the surface contours are the same as those of the original plant part. The formation of fossil molds and casts parallels the method by which a sculptor creates a bronze statue. The sculptor does not carve directly on a block of bronze, but creates a sculpture with some other medium— wood or wax perhaps. A mold is then constructed around the original sculpture and, when the mold is complete, the original is removed in some fashion (disassembling the mold temporarily or melting the wax). When the mold is reassembled, molten bronze is poured into it, and an exact replica of the original sculpture (but one that involves none of the original material in that sculpture) is formed. Rates of sedimentation in certain areas where molds and casts were formed must have been spectacular. As an example, the sea cliffs at Joggins, Nova Scotia, reveal exposed casts of Pennsylvanian tree trunks 3–8 m tall. The trees must have been buried quite

ertina) (arrows) (Mississippian). Hammer  15 cm.

chapter 1 introduction to paleobotany, how fossil plants are formed

23

1.37 Compressed trunk cast of Eospermopteris (Devonian). (Courtesy W. E. Stein.)

Figure

Figure 1.39 Cast of several tracheids showing circular bor-

dered pits (Miocene). Bar  55 μm.

Figure 1.38 Large pith cast of Calamites gigas (Permian).

Bar  20 cm.

rapidly in place. Sediment hardened and the trees subsequently died, leaving hollows (molds) in the hardened rock that were subsequently filled with other sediments (casts). Compressions, and casts (FIG. 1.37) are important in showing the external form of plant parts in a three-dimensional fashion. Root casts of trees can provide important morphological information useful in determining the type of soil formation and soil drainage conditions when the plants were growing (Retallack, 1990). They also may reveal specialized taphonomic processes and how degradation of organic tissues may have proceeded (Driese et al., 1997). A special form of cast is the calamite pith cast or steinkern, which is a common form of preservation of larger calamite stems and branches. Pith casts are casts (FIG. 1.38) of the hollow pith or medullary region in calamites and

preserve an impression of the outside of the pith cavity, which represents the inside of the vascular tissue and cortex (see Chapter 10 for further details). An unusual example of a mold and cast is represented by fossil wood that was exposed to colloidal silica during the diagenesis; the silica permeated the cell cavities, but somehow did not impregnate the cell walls. After precipitation of the silica within the cell cavities, the cell walls (molds) disintegrated and all that is left are casts of the cavities of the wood cells (FIG. 1.39). These cell casts have the negative contours of the insides of the cell walls and show counterparts of specialized wall structures, such as bordered pits (Chapter 7). CELLULAR PRESERVATION

With few exceptions, none of the preceding types of fossil preservation provide the opportunity to examine the internal structure of a plant or plant part. In the case of permineralizations and petrifactions, however, it is possible to study the internal anatomy of ancient plants (Schopf, 1975). In these fossils, one can examine cells and tissue systems within a plant, as well as produce a series of serial sections that can be used to reconstruct the three-dimensional organization of a structure. This type of fossil is called a permineralization or a petrifaction.

24

paleobotany: the biology and evolution of fossil plants

Figure 1.40 Several silicified logs in the petrified forest of Patagonia, Argentina. S. Archangelsky, left and T. Delevoryas, right.

In both types, the process begins when a plant part becomes immersed in water containing a high concentration of dissolved minerals, the most common being silica (silicon dioxide, SiO2), which is often readily available in areas of volcanic activity. The plant part, for example a log, gets thoroughly waterlogged, with water and dissolved minerals permeating all the cells and tissue systems. The dissolved minerals may be silica compounds (silicification) (FIG. 1.40), carbonates (e.g., calcium carbonate, CaCO3), oxides, pyrites (iron sulfide, FeS2), or some other type of chemical. At this stage it is unclear what happens, but something triggers precipitation of the dissolved mineral (e.g., a change of pH) so that it hardens around and within the plant fragment. The cell walls of the plant itself may serve as nucleation sites for the growth of the mineral crystals. When the mineral is completely solidified, the plant fragment, in effect, is entombed within solid rock. The fossil can now be sectioned by various means and examined under the microscope to see internal details of the plant. Although several authors have attempted experimental silicification of wood in the laboratory (Leo and Barghoorn, 1976; Laroche et al., 1989), the preservation seen in fossils is often much better. Preservation of plants as petrifactions or permineralizations probably involves several stages of mineral growth, with different sizes of crystals involved. Some more recent work (Channing and Edwards, 2004) suggests that colloidal or gel phases of some minerals may be involved in the apparently rapid preservation of minute details. Permineralizations and petrifactions can both be studied by means of thin sections, sometimes called petrographic thin sections (FIG. 1.41) (Hass and Rowe, 1999). The piece of rock containing the fossil is cut and ground thin enough

Prepared commercial thin section by the James Lomax company of Carboniferous coal ball plants.

Figure 1.41

to transmit light through the section, essentially, the same technique that geologists use to make petrographic rock sections. Rock saws are available that can cut through most types of rock matrix; most have steel blades with diamond particles embedded in the cutting edges. Oil or some other coolant is used to keep the blade from getting too hot as it slices through the rock. Saw blades covered with or made of particles of silicon carbide or some other abrasive can be used to cut through softer material. The fossil to be studied is cut out with a saw, and the surface of the fossil is polished with an appropriate abrasive (e.g., silicon carbide of various grades) until it is smooth. The surface showing the fossil is then attached to a piece of glass with some type of adhesive. After the adhesive has solidified, the glass slide with the piece of permineralized material is placed back in a saw, now a specialized thin-sectioning saw, and the remainder of the rock is cut away to leave as thin a slice as possible. At this point, the rock is still opaque. The next step involves grinding the surface, either by hand on a lapidary wheel or plate, or on a thin-sectioning machine, so that more and more light can pass through the specimen. Eventually, the sliver

chapter 1 introduction to paleobotany, how fossil plants are formed

of rock is thin enough to be examined with a microscope. Sometimes the piece of glass to which the fossil material is attached is the actual slide used for study. In that case, a permanent adhesive, such as epoxy resin, may be used. Some prefer to transfer the ground specimen to a clean microscope slide. In those instances, a cement that can later be remelted is used, such as Lakeside thermoplastic resin. Before the thin section is transferred, it is coated with a transparent plasticlike material to keep the section intact. This thin slice is then placed on a clean slide with a natural or synthetic mounting medium and covered with a cover glass. Once the medium has hardened, the slide can be examined with a compound microscope. Some paleobotanists use no cover glass, but rather examine the rock surface directly using oil immersion microscope objectives; this method has been applied very successfully with the Early Devonian Rhynie chert (see Chapter 8). PERMINERALIZATION In a permineralization, minerals fill the cell lumina and the intercellular spaces, but do not completely replace the cell walls. The cell walls still consist of organic matter, although they may be chemically altered to various degrees. Chemically, the various layers of the cell wall may still be distinct (Boyce et al., 2002), and the permineralization may faithfully reproduce the microstructure of the wall, for example the position of cellulose microfibrils (Smoot and Taylor, 1984). Cellular contents have even been described from permineralizations! The processes involved in the formation of certain types of permineralization in silica are being studied in modern hot springs ecosystems like Yellowstone National Park, USA (Channing and Edwards, 2004), and in filamentous microbes from similar ecosystems in New Zealand (Renaut and Jones, 2003; Jones et al., 2004; Phoenix et al., 2005). These studies underscore the complexity of the preservation process. In general it involves the formation of opaline silica (opal-A) films that coat structures and colloidal silica that permeates cells and tissue systems. More recent work has shown that bacteria may be involved in or even necessary for many mineralization processes, and the field of geomicrobiology is a rapidly growing area of study. In biomineralization, the bacteria may serve as catalysts for chemical reactions and also as nucleation sites for mineralization (see Konhauser, 2007 for additional information on this topic). An analogy of the process of permineralization is the technique used to embed and section living biological material. For example, a piece of plant is killed and fixed in an appropriate chemical. It is then passed through a series of

25

alcohols to dehydrate the tissue, and finally transferred to molten paraffin or plastic. When the paraffin is cooled, the plant part is completely embedded in it—paraffin is present within the tissues as well as around them. The entombed specimens can then be serially sectioned to reveal details of the cells and tissue systems. PEEL TECHNIQUE. The peel or acetate peel technique (FIG. 1.42) is a simple and rapid method for preparing sections of permineralized plants (Joy et al., 1956; Galtier and Phillips, 1999). In order to use the peel technique, there must be a certain amount of organic matter still present in the cell walls of the fossil plant. If not, thin sections have to be prepared. The rock containing the fossil is sliced with a rock saw (FIG. 1.43) and the resulting slab is polished (FIG. 1.44), first with a coarse abrasive (100–400 grit size) on a lapidary wheel and finally with abrasive of progressively finer grain sizes (600 grit size). The polished surface is then ready to be etched. If the fossils are entombed in calcium carbonate (see coal balls below), etching is done in a dilute solution (5%) of hydrochloric acid (FIG. 1.45). The acid reacts with the carbonate, but not with the organic remains, so the mineral material (CaCO3 in this case) is slowly etched away, leaving the plant cell walls (and cellular contents, if present) projecting in relief from the surface of the slab (FIG. 1.46); the etched surface should not be touched at this stage as the cell walls are very delicate. After the surface has been rinsed and air dried, it is ready to be peeled. Acetone, which is an organic solvent, is poured on the etched surface and, before it evaporates, a thin sheet of transparent cellulose acetate (or similar plastic) is carefully rolled on the surface (FIG. 1.47). The acetone will partially dissolve the lower surface of the acetate sheet, converting it to a liquid that flows in and around cell cavities and intercellular spaces. Because acetone is quite volatile, it evaporates readily, so the lower surface of the acetate sheet quickly solidifies, embedding the cell walls within it. When the acetate is completely dry, it can be pulled from the surface of the rock, taking with it a thin section of the entombed plant (FIG. 1.48). The greatest advantage of the peel technique is that serial sections can be made quickly down through the rock by polishing, etching, and repeating the process again and again (FIG. 1.49). The peel technique (Stewart and Taylor, 1965) can be used for different types of permineralizations, but when the matrix is something other than a calcium or magnesium carbonate, a different acid or a different concentration of acid must be used. When the peel technique was first devised, preformed sheets of cellulose acetate were not available; rather, a solution of parlodion, butyl acetate, amyl alcohol, xylene,

26

paleobotany: the biology and evolution of fossil plants

Diagrammatic representations of the steps involved in the preparation of the coal ball peel technique. A. Section of coal ball slab (calcium carbonate matrix) containing plant material (crosshatched); B. coal ball slab after acid etching to partially expose plant material; C. etched coal ball slab surface with cellulose acetate sheet in place; D. cellulose acetate sheet (peel) being pulled from the surface with adhering plant material; and E. coal ball peel containing embedded plant material. (From Taylor and Taylor, 1993.)

Figure 1.42

Figure 1.44 Figure 1.43

Several pieces of coal ball after sectioning.

abrasive powder.

Polishing the coal ball slab on a glass plate using

chapter 1 introduction to paleobotany, how fossil plants are formed

Figure 1.45 Etching the coal ball slab in dilute hydrochloric

27

Figure 1.48 Removing the peel from the coal ball slab

acid.

surface.

Figure 1.46 Etched surface of coal ball slab prior to flooding

Figure 1.49 Coal ball peel, left, and coal ball slab at right from which it was removed.

the surface with acetone.

castor oil, and ether was poured on the surface and allowed to dry (Darrah, 1936) (FIG. 1.50). This resulted in peels that were not uniform in thickness and were sometimes difficult to mount on microscope slides. Another drawback was the amount of time required for the poured peels to dry on the coal ball surface, as compared with the approximately 20 min required for cellulose-acetate-sheet peels to dry. Despite these drawbacks, the poured peels may still be useful, especially when examining very delicate structures and surfaces that are irregular.

Rolling the acetate sheet into position on the coal ball slab. Bottle contains acetone.

Figure 1.47

COAL BALLS. We know more about the anatomy, morphology, and biology of Carboniferous coal-swamp plants than those from any other time period, and this is primarily due to coal balls. During the Carboniferous, North America

28

paleobotany: the biology and evolution of fossil plants

Figure 1.52

Digging coal balls from a stream bank in Illinois,

USA.

Figure 1.50 William C. Darrah.

Figure 1.53 Transporting bags of coal balls from a site in

Kentucky, USA.

Figure 1.51 Collecting coal balls at a strip mine in southern

Illinois.

and Europe were close to the equator and contained extensive tropical forests which contributed to the extensive coal deposits characteristic of these areas today. Associated with some of these coal deposits are coal balls (FIGS. 1.51–1.53), variously shaped nodules which occur in bituminous coal seams. Coal balls represent permineralized peat deposits and are composed almost entirely of plant parts preserved in calcium carbonate. Some of the first ones found in England

were nearly spherical, hence the name, coal ball, but they can be irregular in shape and range from a few centimeters across to many meters in thickness. Some of the oldest ones come from the upper Namurian (Upper Mississippian) of Germany and the Czech Republic, but they are also known from Permian coal deposits in China. They can be readily studied by means of the peel technique. The method of formation of coal balls has been examined by a number of paleobotanists (Falcon-Lang, 2008), beginning with Stopes and Watson (1908), but the process is still not fully understood. When fresh or partially decayed, the peat was infiltrated by carbonates (fibrous calcite) before there was extensive compaction of the plants within. Since some coal balls are associated with marine limestones, it has been suggested that the plants were growing in lowlying, swampy areas close to the sea, and this hypothesis fits

chapter 1 introduction to paleobotany, how fossil plants are formed

with the paleogeography of Midcontinent North America during the Carboniferous. During storms or marine transgressions (Mamay and Yochelson, 1962), the coal swamp was inundated by seawater, which provided a source of calcium carbonate for permineralization. This hypothesis explains the mixed nature of some coal balls in which both plant and marine animal remains are preserved. Scott and Rex (1985) suggested that all coal balls are not formed by the same process and put forward a non-marine model of formation in which the permineralizing fluids are derived from percolating groundwater high in carbonates. Scott et al. (1996) examined the origin of Carboniferous and Permian coal balls from Euramerica and China and concluded that several different mechanisms were involved, depending on the region and the location of the coal balls within the coal seam. Based on carbon isotopes, they found that some coal balls involved a mixture of marine and meteoric fresh water percolating through the peat and noted that most coal balls formed in freshwater basins with at least some marine influence. There can be little doubt that the formation of coal balls was a highly specialized process, as none are known after the Carboniferous–Permian. To the coal miner these calcium carbonate coal balls represent impurities in the coal that are often termed “fault,” but to the paleobotanist they provide a source of fascinating information that can be used to investigate the biology of the plants that lived in the peat swamps hundreds of millions of years ago. OTHER PERMINERALIZATIONS. Many permineralizations contain silica as the embedding mineral (FIG. 1.54). In fossil peat from Permian and Triassic rocks (FIGS. 1.55, 1.56) from the central Transantarctic Mountains of Antarctica, the silica is in the form of chalcedony (Schopf, 1971). It is possible to make acetate peels of silica permineralizations; however, they must be etched in concentrated HF. When using HF, precise safety procedures (e.g., etching in a fume hood, proper gloves and other protective clothing, and eyewear) must be employed because of the very dangerous nature of this acid. In some cases, especially in certain Devonian fossils, preservation involves permineralization via pyrite (iron sulfides) or limonite (hydrated iron oxides). Plant parts preserved by these minerals have been difficult to study because the material often breaks apart during grinding and is lost. To eliminate such problems, specimens of this type need to be first embedded in plastic prior to cutting (Stein et al., 1982). For fossils preserved in ironstone (fine-grained sedimentary rock), a useful technique involves selectively macerating the specimen in acid to free the silicified axes and then embedding the

29

Figure 1.54 In situ stump of Triassic tree in Antarctica. Yellow

pen for scale.

Figure 1.55 Block of Triassic chert (orange color) from

Antarctica.

axes in bioplastic to examine internal anatomy (Aulenback and Braman, 1991; Serbet and Rothwell, 2006). The degree of detail that can be preserved by permineralization is truly extraordinary, with such delicate structures

30

paleobotany: the biology and evolution of fossil plants

Figure 1.58 Spores of Cyathotheca tectata showing dis-

tal (left) and proximal surfaces, and distinct ornamentation (Pennsylvanian). Bar  18 μm. Figure 1.56 Block of permineralized peat from Antarctica showing root of Glossopteris, with wood wedges alternating with lacunae, Vertebraria (Permian). Bar  2 cm.

Figure 1.57 Two animals (Ebullitiocaris oviformis) (arrows)

attached to an Aglaophyton major stem (Devonian). Bar  0.5 mm. (Courtesy H. Kerp.)

as starch grains, nuclei, various types of membranes, tapetal deposits, and cells of seed-plant microgametophytes are known. Spores of lycopsids and microspores of Pentoxylon have been interpreted as containing chromosomes (BrackHanes and Vaughn, 1978; Bonde et al., 2004). The flagellum of a chytrid zoospore (FIG. 3.20) (Taylor et al., 1992) and rotifers (FIG. 1.57) from the Rhynie chert, and sperm within the pollen chamber of a Permian seed (Nishida et al., 2003) have been described from permineralized plant remains. There are numerous examples of exceptionally well-preserved plant structures throughout this book.

In some instances, the matrix of the permineralization is too crumbly to allow preparation of ground thin sections or does not lend itself to the peel technique. In such cases, it may be necessary to examine the cut and polished surface with reflected light. If a series of sections is necessary, one must make a photographic record or a series of drawings, because the specimen will be lost as it is continually ground away, leaving no actual record of each face examined. PETRIFACTION Cellular details can also be observed in a petrifaction. In this case, all of the original organic matter in the plant has been replaced by minerals. Many fossil woods, such as those from the Triassic Petrified Forest in Arizona and the Cerro Cuadrado (Jurassic) Petrified Forest in Patagonia, Argentina (FIG. 1.40), are preserved in this manner. It is necessary to make thin sections to study petrifactions, since the etching involved in producing a peel preparation would completely dissolve the specimen. Other techniques such as cathodeluminescence are providing a new source of information about silicified wood (Matysová et al., 2008). UNALTERED PLANT MATERIAL

Some plant parts are found as fossils in an unaltered form, either as body fossils or as chemical fossils. Pollen grains and spores (FIG. 1.58), diatom frustules, cuticle envelopes, various types of resins, such as amber (FIG. 1.59) and calcium carbonate remains of certain types of algae are all examples of unaltered plant fossils. In some instances even the soft parts are sufficiently preserved so that comparisons can be made at the cytoplasmic and ultrastructural level (Wolfe et al., 2006). Holocene peat is an example of relatively unaltered plant material (Williams and Yavitt, 2003). Plant parts became

chapter 1 introduction to paleobotany, how fossil plants are formed

31

Figure 1.59 Stamen (arrow) embedded in amber in the proc-

ess of shedding pollen as it was being preserved (Miocene). Bar  1 mm. (Courtesy G. O. Poinar.)

Phytolith of a dicot (Oligocene). Bar  20 μm. (Courtesy C. A. E. Strömberg.)

Figure 1.61 Phytolith (laminated trichome-type) from a dicot (Eocene). Bar  30 μm. (Courtesy C. A. E. Strömberg.)

Figure 1.60

incorporated into peat bogs, and because of the high acidity in the bogs, microbial activity is greatly reduced, so little or no decomposition occurs. The accumulated plant debris may build up to a considerable thickness, and while there is some disassociation of plant parts as well as flattening, the bits and pieces preserved can be handled like the same parts of modern plants. Another excellent example of unaltered plant material is diatomaceous earth. Although the cell contents are no longer present, the silica cell walls remain intact and are preserved in such fine detail that the exquisite sculpturing can be easily detected on the surface. Phytoliths also represent unaltered plant material secreted by the living plant in the form of calcium carbonate or opaline silica (FIGS. 1.60, 1.61). They occur in various types

Figure 1.62 Phytolith (echinate sphere-type) from a palm (Eocene). Bar  15 μm. (Courtesy C. A. E. Strömberg.)

of grasses and some trees and, depending upon the species, vary in morphology and size (FIGS. 1.62, 1.63). Although they have been used for many years to study Holocene or Pleistocene habitats, recent research has utilized these plant markers to study older paleoenvironments. They have been especially useful in the interpretation and reconstruction of grassland ecosystems (Strömberg, 2004). Extracting fossil phytoliths can be accomplished using a variety of techniques (Parr, 2002), but it is important to avoid methodological bias, as discussed in Strömberg (2007, 2008). Spores and pollen grains are represented in the fossil plant record in great abundance because the wall (sporoderm) of

32

paleobotany: the biology and evolution of fossil plants

Figure 1.63 Phytolith consisting of epidermal long cells with papillae and a row of short, siliceous vertical cross cells from a grass (Cretaceous). Bar  50 μm. (Courtesy C. A. E. Strömberg.)

the spore or pollen grain is composed of an especially resistant material called sporopollenin. As such they also represent unaltered plant material. Spores and pollen grains may be common in certain rocks, even when there is no evidence of any other plant parts. It is possible to extract these pollen grains and spores from the rock and mount them on microscope slides. TEM has provided a wealth of new information about fossil pollen and spore walls that has been used to examine the development of the pollen wall (Taylor and Alvin, 1984). More recently, Scott and Hemsley (1991) have discussed the value of laser scanning and scanning acoustic microscopy in the study of paleobotanical materials. CHEMICAL FOSSILS Chemical fossils can also represent a type of unaltered plant material (Hemsley et al., 1996). Chemical signatures, sometimes called biomarkers or geomolecules, are specific for certain groups of organisms. These biomolecules are transformed through time with lipids perhaps having the best opportunity of being preserved (Itihara et al., 1974), whereas nucleic acids degrade more rapidly (Briggs et al., 2000).

Biomarkers in the form of terpenoids have recently been reported in the fossil wood type Protopodocarpoxylon from the Jurassic of Poland (Marynowski et al., 2007). An example of chemical preservation is the presence of hydrocarbons within Ordovician oil deposits that are known to be produced only by the putative cyanobacterium Gloeocapsomorpha (Hoffman et al., 1987; Foster et al., 1989; 1990; Blokker et al., 2001). Pristanes and phytanes are also a type of biomarker. These molecules are believed to be derived principally from chlorophyll degradation, but can also be produced by non-photosynthetic organisms (Hahn, 1982). A number of new techniques have been added to the arsenal of the paleobotanist, and these promise to provide significant advances in a number of areas. They include elemental analysis, chemolysis, pyrolysis, and lipid analysis, and have been used to study a large number of organic compounds throughout the geologic column. Pyrolysis and chemolysis have been used to screen for the chemical composition of the fossil material. Such procedures and techniques have increasing application in determining the systematic affinities of a particular organism, as well as determining various taphonomic processes that may have altered the fossil. These various paleobiochemical techniques involve the extraction of organic constituents still associated with the fossil or represented as residues in the rock matrix. Classes of chemical compounds, such as sterols, aromatics, carboxylic acids, polysaccharides, various types of lignin, fatty acids, and nalkanes, are but a few of the chemical constituents that have been identified in fossil plants. Organic chemical profiles have been used effectively with extant angiosperms and, at one time, chemosystematics represented a basic and almost routine technique in plant systematics (Crawford, 1990). These techniques have sometimes been applied to the study of fossil plant systematics, but it is important to consider diagenetic changes, especially in older fossils. Niklas et al. (1985) used steroid and other cycloalkane–alkene profiles to show that a Miocene Liriodendron leaf was chemotaxonomically more similar to one particular living species despite the fact that the fossil shares morphological characters with two living species. More recently, Mösle et al. (2002) demonstrated that the biomolecular cuticle signature was more comparable between more closely related plants, such as Cordaites and Walchia, than between these seed plants and certain coeval seed ferns. There are, however, certain limitations to such paleobiochemical approaches. For example, various microbial activities may alter the organic chemicals shortly after the organism dies, or there may be modifications to the chemical constituents as a result of diagenesis. Some organic compounds may have formed abiotically, rather than

chapter 1 introduction to paleobotany, how fossil plants are formed

33

representing the original chemical constituents of the plants. Still others may have percolated through the surrounding rocks and constitute contaminants, even in rocks as hard as cherts. Thus, for paleobiochemistry to be useful to paleobiologists, analyses must follow strict qualitative and quantitative protocols that can be standardized and repeated (Van Bergen, 1999). ANCIENT DNA During the last two decades, systematists working with extant plants have switched, in part, from secondary metabolites to the use of molecular sequences, including nuclear, chloroplast, and mitochondrial DNA, as the macromolecules of choice in developing phylogenetic hypotheses for plant relationships. Golenberg and colleagues (1990) reported the extraction and amplification of an 820-base-pair DNA fragment from the chloroplast rbcL gene (ribulose-1,5bisphosphate carboxylase–oxygenase or RuBisCO) of a Magnolia latahensis leaf. The leaf was collected from the famous mid-Miocene (Langhian) Clarkia beds of northern Idaho, which are dated at 15–15.5 million years old (Ma). In another study on fossils from the same site, Kim et al. (2004) reported amplified ndhF (NADH dehydrogenase) from the same type of leaf (M. latahensis), and rbcL from a specimen of Persea, further suggesting long-term preservation of ancient plant DNA. Recently, however, these studies and others, including DNA sequences from bacteria in insects in amber, DNA in dinosaur bones, and in salt crystals have been challenged on various grounds. As a result some believe that DNA in excess of 1 million years old is probably an artifact (Pääbo et al., 2004). Others believe the study of ancient DNA holds promise at some level (Gugerli et al., 2005), but that the evidence must be compelling, and from multiple sources. Obviously, additional samples from the same site that demonstrate similar results will help verify such reports, and also a closely followed set of protocols will be especially useful in demonstrating the authenticity of ancient DNA (Gilbert et al., 2005). Although barely classified as ancient, an interesting study has recently reported 1000-year-old DNA from excavated wood samples using a strict set of procedures to insure that the material was not contaminated (Liepelt et al., 2006). MUMMIFICATION When conditions of burial are rapid, and especially in very dry or cold environments, wood or other plant parts may survive for millions of years in a relatively unaltered condition. Such mummified remains have been described from Cenozoic deposits (Basinger et al., 1988; Francis and Hill,

Figure 1.64 Mummified leaf of Cryptocarya (Lauraceae)

(Eocene). Bar  2 cm. (Courtesy D. C. Christophel.)

1996; Fukushima et al., 1996) and represent a special preservation type in which the plant tissue was rapidly dehydrated and buried (FIG. 1.64). So well preserved are the cells and tissue systems of these mummified plants that they can be studied by the same techniques as those used to examine extant tissues. Mummified wood is not mineralized, so it can be sectioned using techniques identical to those utilized by wood anatomists for extant material. AMBER Another example of unaltered plant material is amber (FIG. 1.65), a name typically applied to a wide variety of fossilized plant resins (Rice, 1987). Amber has been found in rocks from the Carboniferous to the Pleistocene, but most deposits have been reported from Cretaceous and Cenozoic strata. In the authoritative text, Plant Resins, Langenheim (2003) restricted the term amber to a lipid-soluble mixture of terpenoid or phenolic compounds, distinguishing it from gums, waxes, mucilage, oils, and latex. Amber is produced by a

34

paleobotany: the biology and evolution of fossil plants

the original plant part, but may contain some components of the original plant. Certain types of algae are more common as fossils because they precipitate or bind calcium carbonate around or in their cells. This calcium carbonate skeleton can build up to a considerable thickness and provides excellent preservational potential. When the alga dies, the calcium carbonate skeleton persists, often for millions of years. Sectioning of these calcium carbonate residues allows one to reconstruct the three-dimensional appearance of the alga by following the configurations of the hollows within the calcium carbonate sheath. These limestone-precipitating algae play a very important part in the build up of some so-called coral reefs; in these cases, the bulk of the reef is produced by the accumulation of CaCO3 precipitated by algae, rather than by the corals living on the reef (see Chapter 4).

PALYNOLOGY Figure 1.65 Winged angiosperm seeds preserved in amber

(Miocene). Bar  1 cm. (Courtesy G. O. Poinar.)

large number of plants; phytochemistry, infrared spectrophotometry, and X-ray diffraction have proved to be important analytical tools in determining the botanical origins of fossil resins (Langenheim, 2003). Because of its sticky consistency when it was produced by the plant, amber has also served as the fossilizing matrix for other organisms. In addition to pollen grains and other wind-borne microscopic plant parts, small flowers, fungi, a variety of insects (Peñalver et al., 2006), and other organisms are often preserved within pieces of amber. Even something as delicate as oil bodies in cells of liverwort leaves (Grolle and Braune, 1988), plant organelles such as chloroplasts and mitochondria (Poinar et al., 1996; Koller et al., 2005), a strand of spider silk (Zschokke, 2003), and amoebae (Schmidt et al., 2004) have been preserved in amber. Poinar (1992) provided an excellent historical account of amber, and the importance of this plant resin in examining the diversity of life preserved in this unique manner, and Grimaldi and Engel (2005) demonstrated the extraordinary preservation and diversity of insects entombed in amber in their comprehensive work, Evolution of the Insects. SUMMARY DISCUSSION

The preceding provided examples of the most common ways in which plants become fossilized, but there are other forms as well, or combinations of the preservational types just discussed. For example, some stem casts may contain a faint outline of the conducting system in the center of the cast. In this case, the cast is not simply a three-dimensional replica of

The science of palynology or, perhaps in a geologic context, paleopalynology is devoted to the study of pollen grains and spores, and also encompasses the investigation of other organic microfossils, such as chitinozoans, acritarchs (Javaux and Marshall, 2006), scolecodonts, dinoflagellates, certain types of microscopic algae, microforaminifera, rotifers, testate amoebae, chitinous fungal remains, and other forms of organic debris sometimes termed varia. Characteristic features such as grain shape (FIG. 1.58), wall sculpture, presence or absence of pores, ridges, furrows, or other types of structural features make it possible to distinguish among grains of various kinds and in some instances to assign them to certain groups of plants. The discipline of palynology is a critical component of understanding the biodiversity of the present and the past, and the important volumes by Wodehouse (1965), Erdtman (1969) (FIG. 1.66), Faegri et al., (1989) (FIG. 1.67), and Traverse (2007) (FIG. 1.68) provided an excellent historical context to the discipline. Palynology has greatly benefited from the introduction of various SEM techniques (Villar de Seoane and Archangelsky, 2008) that have made it possible to image and interpret complex external features on the grains (FIG. 1.69). There has also been an attempt to automate palynology, using texture analysis of SEM images (Langford et al., 1990; Vezey and Skvarla, 1990). This procedure greatly reduces the labor-intensive aspects of palynology and perhaps offers more rapid results, larger data sets, finer resolution of taxa, and possibly greater objectivity in identification (France et al., 2000). Fossil pollen grains and spores (FIGS. 1.70, 1.71) are now routinely sectioned and examined with the TEM as well.

chapter 1 introduction to paleobotany, how fossil plants are formed

35

Figure 1.68 Alfred Traverse. (Courtesy M. Streel.) Figure 1.66

Gunnar Erdtman.

Figure 1.67

Knut Faegri (right) and Ove Arbo Høeg.

These studies have provided a wealth of detailed information about features of the exine (the outer spore or pollen wall that is composed of sporopollenin) that have been useful for systematic studies. Some paleobotanists have combined SEM (FIG. 1.72) and TEM studies of dispersed spores to try to better understand the affinities of these propagules (Edwards et al., 1996), and to more accurately interpret features of the wall (Wellman, 2001). Information on pollen and spore ultrastructure is often determined from single sections

Figure 1.69 Cyathotheca tectata spore viewed with the scanning electron microscope (Pennsylvanian). (From Taylor, 1972.) Bar  23 μm.

of grains in which the plane of section is not easily determined, but techniques have been developed so that the same grain may be examined and recorded in transmitted light, and then scanning and TEM (Daghlian, 1982). In addition, it is often important to prepare serial sections of the same grain in order to view features, such as lamellae, that may not be consistently present throughout the entire wall (Johnson and Taylor, 2005).

36

paleobotany: the biology and evolution of fossil plants

Figure 1.70 Ultrathin section of the spore Horstisporites iri-

dodea showing organization of the sporoderm. Bar  2.5 μm. Figure 1.72 Fractured surface of Cyathotheca tectata spore.

Bar  3 μm. (From Taylor, 1972.)

daunting, and some palynologists have developed automation techniques to assist in the process. For example, image analysis of measurements can be used to quantify shape and ornamentation on SEM images (Treloar et al., 2004). In a companion study, Li et al. (2004) used a neural network to analyze texture and were able to correctly identify four extant pollen taxa. GEOCHRONOLOGY AND BIOSTRATIGRAPHY

Figure 1.71 Ultrathin section of Cyathotheca tectata spore

as viewed with a transmission electron microscope. (From Taylor, 1972). Bar  12 μm.

Fossil pollen and spores that are preserved within the pollen sac or sporangium (in situ) can provide valuable information on developmental patterns in the formation of the pollen grain and spore walls. In certain types of fossils, such as permineralizations, it is possible to extract many pollen grains of the same type from a single sporangium, or from multiple sporangia at different stages of development. Study of these grains can thus offer insights into biological processes that took place millions of years ago (Taylor and Alvin, 1984). The process of identifying numerous palynomorphs, especially those that are dispersed (i.e., not in situ) can be

Perhaps the most widespread application of palynology is in geochronology, the dating of events in the history of the Earth. Palynomorphs and certain microfossils can be used in geochronology, that is, dating rock units. Fossils can only provide a relative date for strata, that is in relation to other units. Absolute dating relies on other methods to give a specific date (see radiometric). Dating with palynomorphs is possible because many change through time or possess unique features that allow them to be distinguished from other types. For example, Upper Cretaceous and lower Paleogene rocks in the Northern Hemisphere contain a unique type of fossil angiosperm pollen termed triprojectates (Farabee, 1990). These grains are unusual in that they possess polar and equatorial projections (FIG. 1.73) and a variety of ornamentation patterns that make them especially good biostratigraphic markers. Although the botanical affinities remain problematic, it has been suggested that at least some triprojectates possess sufficient characters to include

chapter 1 introduction to paleobotany, how fossil plants are formed

Figure 1.73 Aquilapollenites grain (Cretaceous). Bar  10 μm. (From Jarzen, 1977.)

them within the modern flowering plant order Apiales (Farabee, 1991). Closely associated with the dating of rocks based on the presence of certain types of palynomorphs is the correlation of rock units in time and space based on the fossils within them, a discipline termed biostratigraphy (Gray et al., 1985). Various types of fossils can be found in different sedimentary environments, for example terrestrial and marine, and each may have its own biostratigraphic markers. For example, calcareous nannofossils are especially good markers in marine sediments from the Jurassic to the recent; plant spores are useful in terrestrial rocks from the Devonian onwards and diatoms (a type of microfossil) from the Paleogene to the recent. At one time, applied biostratigraphy was the method of choice in petroleum exploration. Although a number of other methods are used today, palynostratigraphic techniques are still used today to correlate strata based on the presence of certain types of microfossils. Today there are a large number of consulting companies that provide services in exploration and interpretation of petroleum and mineral deposits based on biostratigraphy and geochronology. The underpinnings of these commercial firms are based on the types and distribution of microfossils. PALEOECOLOGY

Palynology has also been extensively used as a method of characterizing past depositional systems (paleoenvironments) (Farley and Traverse, 1990). Here palynomorphs play an important role in defining, for example, the extent of a marine or terrestrial environment. In other instances, certain types of

37

palynomorphs may provide valuable information about water depth, temperature, salinity, and nutrient levels where the organisms once lived. In a few cases where vertebrates and invertebrates are found with palynomorphs and plant megafossils, an even greater degree of paleoecological resolution can be obtained (Westgate and Gee, 1990). Other detailed ecological studies are possible based on the frequency and types of pollen present both geographically and stratigraphically within a confined area (Graham, 1990). The use of pollen data in association with megafossil information has had a profound influence on the interpretation of paleophytogeographic patterns throughout the world (see, e.g., Graham, 2000, 2003, on plant distribution in the Caribbean). Such studies are especially valuable when they incorporate both extant and fossil data and are founded on well-defined geographic regions of the world (Graham, 1972; 1973). Other investigations have utilized paleoecological data to show that the early flowering plants were herbs or small trees living in unstable habitats during the Cretaceous (Wing and Boucher, 1998). Certain climatic parameters can also be defined by the occurrence of certain palynomorphs, since various plants respond to minor environmental fluctuations (Pocknall, 1990). Tracing the appearance and disappearance of various palynomorphs vertically in the geologic column provides a method of tracking certain types of climatic shifts. Pollen analysis is a branch of palynology in which the relative proportions of pollen and spores are mapped vertically and horizontally; these proportions are then used to reconstruct the paleoenvironment by comparison with modern proportions of the same or closely related taxa. Although primarily applied to Quaternary deposits, similar techniques have been used in older sediments. Recovery of DNA from Holocene pollen (Bennett and Parducci, 2006) has potential for more accurate identification of certain pollen types, as well as tracking populations of plants through time. Using modern pollen and spores, Traverse (1990) determined the palynomorph load in various types of bodies of water in the Trinity River of Texas. Understanding the dynamics of a modern model system such as this is important in the interpretation of past vegetation. Pollen extracted from marine sediments, together with stable isotopes and radiolarian microfossils extracted from ocean sediment cores, were used to provide data about ocean variability on millennial timescales (Pisias et al., 2001). This information can then be used to directly compare the climate responses of continental and oceanic systems, and incorporated into broader scale climate models. Palynomorphs or microfossils are preserved from every time period of geologic history and in many types of depositional environments, so they are a valuable source

38

paleobotany: the biology and evolution of fossil plants

Figure 1.74 Elaterosporites klaszii (Cretaceous). Bar  20 μm. (From Jardiné and Magloire, 1965; courtesy M. S. Zavada.)

Figure 1.75 Elaterocolpites castelaini (Cretaceous). Bar  20 μm. (From Jardiné and Magloire, 1965; courtesy M. S. Zavada.)

of information with which to characterize changes in paleoecosystems at different scales (Taggart and Cross, 1990). Although several books have been written on various aspects of palynology, the three-volume set, Palynology, Principles and Applications, edited by Jansonius and McGregor (1996) and the volume by Traverse (2007), Paleopalynology, provide very comprehensive and up-to-date surveys of the discipline. The recent volume edited by Van Geel (2006) focuses on the importance of various microfossils in the interpretation and reconstruction of Quaternary environments and the Glossary of Pollen and Spore Terminology (Punt et al., 2007) will be helpful in understanding the complex terminology used to described pollen and spores. Even with our current extensive knowledge of pollen and spores, some palynological preparations occasionally contain structures which cannot be identified (FIGS. 1.74, 1.75). Graham et al. (2000) described sinuous to coiled filaments similar to the elaters of Equisetum spores (Chapter 10), secondary wall thickenings of conducting elements, or germinating fungal spores. They showed that these filaments actually represent artifacts. Termed petrofilaments (FIG. 1.76), they form when hydrocarbons (asphaltenes) react with solvents in the mounting medium. Figure 1.76 Petrofilaments. Bar  25 μm. (Courtesy A. Graham.)

ABSOLUTE DATING One of the most frequent questions asked of paleobotanists is, “How do you date fossil plants?” Most paleobotanists are familiar with the various groups of plants that lived at different points in geologic time. Consequently, when encountering

a new assemblage of plant fossils, they usually recognize the general age immediately, but this has not always been the case. Our current understanding of the age of fossil floras is based on a long series of efforts to date the various rocks

chapter 1 introduction to paleobotany, how fossil plants are formed

in which they are found. At the present time, the best absolute dating method involves the use of naturally occurring radioactive isotopes contained in various minerals that make up a rock unit. The inherently unstable radioactive isotopes undergo a series of complex transformations (decay) that lead to stable isotopes and, in the process, release energy. The rate of decay, λ, for a radioactive isotope of a given element, sometimes called the half-life, is constant (t½  0.693/λ). Therefore, by measuring the present amount of the radioactive isotope and the present quantity of the stable product, one can calculate how much time has elapsed since the minerals in the rock formed. For example, it is known that the long-life uranium isotope 238U decays to 206Pb with a halflife of 4.5 billion years. Consequently, by measuring the relative quantities of 238U and 206Pb in a sample, it is possible to determine the length of time the decay has been going on and thus the time of formation of the rock. A widely used technique involves the analysis of a very small amount of the relative quantities of uranium and lead contained within zircon crystals (Harrison et al., 2005). These crystals, which may be 0.1 mm in size, form as molten rock begins to cool and thus lock small amounts of uranium into their crystalline structure. This technique, which utilizes a high-resolution ion microprobe, uses a powerful beam of ions to vaporize a tiny portion (two-billionths of a gram) of a zircon crystal (Davis et al., 2003). The vapor is then passed through a mass spectrometer where the different elements are separated and analyzed. Zircon crystal geochronology has been applied widely across geologic time, including dating major earth events such as the formation of the continental crust (Harrison et al., 2005). Other radioactive isotopes differ in their half-lives, for example, 87Rb (rubidium), 48.5 billion years; 40K (potassium), 1.25 billion years; and 235U (uranium), 0.704 billion years. One difficulty in employing these dating techniques is that radioactive isotopes occur more commonly in igneous and metamorphic rocks, whereas almost all fossils occur in sedimentary deposits. Today direct isotopic dating for sedimentary rocks is possible, but only when they contain minerals that have crystallized in the environment of deposition at or near the time they were deposited. One of these is glauconite, a silicate mineral that contains potassium (Smith et al., 1998). Since the potassium consists in part of 40K, the potassium–argon method can be used. Rubidium–strontium dating of some very fine-grained sedimentary rocks also has been successful, but the procedure is difficult and not routinely applicable. A technique has been developed in which actual fossils can be dated. In the upper atmosphere, cosmic rays bombard 14N (nitrogen) isotopes to form an isotope of carbon (14C) that is radioactive. This carbon unites with oxygen to

39

produce carbon dioxide (CO2). Plants take in and fix (assimilate) this carbon dioxide along with that containing the more common isotopes of carbon, 12C and 13C. Carbon dioxide is continuously assimilated during the lifetime of a plant. When the plant dies, however, it no longer exchanges carbon dioxide with the atmosphere, and thus the ratio of 14C to 13C or 12 C is fixed at that time. At that point, the 14C begins to decay to 14N with its characteristic decay rate (t½ of 14C is 5730 years). For this reason, the ratio of 14C to 12C or 13C is proportional to the age of the fossil. An age limit of 50,000 years (Balter, 2006) applies to this technique because of the short half-life of 14C. This technique obviously has somewhat limited usefulness in paleobotany because the bulk of the fossil plant record is far older. Human influence on the Earth has even altered the usefulness of the 14C dating method, because combustion of fossil fuels and nuclear testing have artificially altered the 14C content of the total carbon reservoir, and this has caused problems in maintaining reliable modern standard samples of carbon. Loss or addition of 14C to specimens and apparent fluctuations of past atmospheric 14C abundance also impose limitations on this dating method. Analytical techniques have been developed that allow direct detection of 14C atoms using high-energy accelerators. This method is especially important as it requires 1 mg of carbon (as opposed to 1000 mg in the conventional methods), and dates can be determined in a matter of hours rather than days.

GEOLOGIC TIMESCALE Frequent references will be made to the geologic timescale in succeeding chapters and a summary of geologic time is provided on the inside front and back cover of this volume. In many ways, the naming of rock units is similar to the naming of organisms, in that the geologic timescale is not fixed, but is constantly updated and refined by the International Commission on Stratigraphy (ICS), part of the International Union of Geological Sciences (IUGS). At the formation level, there is a type section of that formation, where rocks that are typical for that formation are exposed. Periods are formally defined at the base of each period by a Global Stratotype Section and Point (GSSP); the GSSP is designated on the ICS timescale as a golden spike. The global stratotype section for the base of a period is somewhere in the world where an excellent section of rocks is exposed and is agreed upon by a committee of experts. For more information, please refer to Gradstein et al. (2004) or the ICS web page, www.stratigraphy.org. There are two types of units in geologic time: rock units and time units. Rock (lithostratigraphic) units refer to the

40

paleobotany: the biology and evolution of fossil plants

physical rocks themselves and the terms, Lower, Middle, and Upper are used for these, for example this fossil was found in Lower Devonian rocks. Time (chronostratigraphic) units refer to the period of time represented by those rocks, and use time units (i.e., Early, Middle, and Late), for example, plants of the Early Devonian. Throughout this book, you will see the following abbreviations for geologic time: ka, Ma, Ga—These stand for, respectively, thousands of years, millions of years, and billions of years (gigayears) before present. These are used for dates, for example this plant lived 450 Ma (450 million years ago). kyr, myr, byr (sometimes written Gyr)—These stand for intervals of time, for example this group survived for 20 myr.

BIOLOGICAL CORRELATION Because radiometric dates are not available for all sequences of rocks in specific geographic regions, it becomes necessary to position a given rock unit accurately relative to its absolute age, a type of relative dating. One way in which a sequence of sedimentary rocks can be grouped according to age is through the use of index fossils. To be effective, an index fossil should (1) distinguishable from other fossils and easily identifiable, (2) have existed during a relatively short period of geologic time, (3) be abundant, (4) be widely distributed geographically, and (5) have lived in different environments, so that it may be preserved in different types of sedimentary rocks. Obviously, not many fossils fulfill all these requirements, and assemblages of several fossil taxa (assemblage zone) are typically more useful than a single species. Generally, the most useful organisms for correlation from one section of rocks to another are those that lived in ancient seas. Pelagic organisms (those that live in the open sea and not on the bottom) provide the best long-range correlations because of their worldwide distribution, at least within certain climatic zones. These organisms include such planktonic forms as diatoms (Chapter 4), foraminifera, silicoflagellates, and coccoliths (FIG. 4.43). These organisms are especially important because their skeletal remains are so small that a large number can be concentrated in a small sample, such as the cuttings obtained from a well boring. Other organisms, such as those that inhabited the ocean floors (benthic forms), typically have a spatially restricted distribution that enables them to be used effectively in correlations of a more local extent. In terrestrial rocks, some of the best index fossils are pollen grains and spores (Gonzalez et al., 2006; Souza, 2006).

They can be carried long distances by wind and, consequently, can be deposited in a wide variety of sedimentary environments. Palynostratigraphy has been an especially important tool in providing correlation between marine and non-marine rocks and in determining the various ecological conditions under which plants lived (Dimitrova et al., 2005). Plant megafossils have also been useful in biostratigraphy, especially when used in the form of assemblage zones. Such studies extend from some of the earliest land plants (Edwards and Richardson, 2004) to measuring Holocene vegetational changes. Noteworthy among plant megafossils used as index fossils are a variety of Carboniferous foliage types (see Chapter 16), which have proven useful in establishing stratigraphic sequences in certain geographic regions (Diaz, 1983, 1985). For example, late Paleozoic foliage types have been useful in delimiting biostratigraphic zones in North America (Read and Mamay, 1964; Gillespie and Pfefferkorn, 1979), southwestern Germany (Germer, 1971), and northern France (Laveine, 1987). In some cases, megafossils have been more reliable than palynology, as palynomorphs are often difficult to extract from the high-rank coals that comprise some of the stratotype sections. In other instances, the identification of particular taxa has been useful in precisely dating tectonic events, such as the Upper Carboniferous folding phases in northwestern Spain (Wagner, 1966), and in documenting climatic changes (Wagner, 2004). There can be little doubt that as fossil plants are better understood they will become increasingly important as stratigraphic markers in biozonation and correlation.

SYSTEMATICS AND CLASSIFICATION This book emphasizes the origin, evolution, and diversity of the major groups of plants based on the fossil record, and their relationships through geologic time; floristic changes through time are discussed to a lesser extent. To do this, we need to address the systematics of plants. The field of systematics is concerned with classifying, naming, and determining the evolutionary relationships of taxa. Taxon (pl. taxa) is a general name to indicate any level of organization (i.e., a species, a genus, a family, etc.). Within systematics, taxonomy is the process of describing and classifying organisms into natural groups and nomenclature is the process of naming taxa. In this book, we will use Linnean nomenclature, in which each plant has a two-part name, sometimes called a binomial, which consists of a genus name and a species name (specific epithet). The rules of naming plants are complex and are encoded in the International Code of Botanical Nomenclature (ICBN) (McNeill et al., 2006), which is refined every 4 years.

chapter 1 introduction to paleobotany, how fossil plants are formed

In spite of long study and continued refinements, naming plants still represents a highly subjective exercise. Generally, all classification systems are based on the same type of evidence: shared features. Shared features allow one to recognize genera, families, and other higher categories of a classification scheme (Funk and Brooks, 1990). Such features may fall into two general categories. In one group are the primitive features (plesiomorphies) that evolved relatively early in the evolution of a group of organisms, such as the vascular tissue present in most terrestrial plants. Features of this type may be regarded as evolutionary holdovers that have persisted, but tell us little about the relationships among members of the group, because every member of the group has the same feature. The other group of characters is believed to have evolved more recently. These advanced or specialized features (apomorphies) can be used to identify organisms that have a common ancestry. The cladistic or phylogenetic system of systematics has the goal to produce a hierarchical organization of taxa based on shared, derived features (synapomorphies) that reflect the evolution of particular groups of organisms (Duncan and Stuessy, 1984). Classifications that group organisms based on the overall similarity of characters, whether both primitive or derived, are termed phenetic systems. Another system that has been proposed for classifying living organisms, including plants, is the Phylocode (Cantino and de Queiroz, 2006). This system is very controversial, but is meant to reflect phylogenetic systematics more than Linnean taxonomy (Nixon et al., 2003). Monophyletic groups, that is those consisting of a single common ancestor and all descendants of that ancestor (clade), are defined solely by their position on the tree of life. Clades may have any rank, but the rank is added after nomenclature is completed. It will be interesting to see if this system gains recognition within the plant systematic community since many plant taxa are now thought to be paraphyletic (Rieseberg and Brouillet, 1994); paraphyletic groups include the ancestor and some, but not all, of its descendants). It would appear reasonable to assume that a classification scheme like the Phylocode, which includes only taxa that fit into monophyletic groups, will not be an accurate and useful tool for arranging the enormous biological diversity represented in the fossil record (Briggs and Crowther, 2001). Moreover, it is difficult to envision how fossils would be treated in the phylogenetic nomenclature of this classification system.

the plants are represented in the fossil record as disarticulated parts. This has resulted in the establishment of a special system of nomenclature for parts of fossil plants. As in other areas of botany, fossil plants are named according to the rules in the ICBN, but in paleobotany each disarticulated part is given a separate generic and specific name. In the past, paleobotanists had two types of names for parts of fossil plants. An organ genus was designated when there was enough information to assign a plant part to a family. For example, Lepidodendron, Stigmaria, and Lepidostrobus (Brack-Hanes and Thomas, 1983) (FIG. 1.77) are generic names used to designate parts (stem, roots, and cones) belonging to a particular type of Carboniferous lycopsid. The form genus was used for fossil plant parts that could not be assigned to a family, for example a piece of wood that could be assigned to the gymnosperms, but not to any particular group of gymnosperms. Originally, an organ genus was considered to represent a more natural (i.e., phylogenetic) taxon than a form genus, but confusion arose because names have been given to the same plant parts in different states of preservation or development. Today, the term morphotaxon has replaced the designations form and organ genera in paleobotany. A morphotaxon is a fossil taxon which, for nomenclatural purposes, comprises only the parts, life history stages, or preservational states represented by the corresponding nomenclatural type (Chaloner, 2004). The nomenclatural type is the plant fossil on which the name is based. Why do paleobotanists give names to different parts of the same plant? The first reason for naming parts is so that the fossils can be studied and referred to in publications

NOMENCLATURE OF FOSSIL PLANTS

Historically, paleobotanists have utilized a somewhat artificial classification system, since in almost all instances

41

Figure 1.77

Sheila Hanes.

42

paleobotany: the biology and evolution of fossil plants

and discussed with other paleobotanists. The other reason is that some identical plant parts may be attached to different plants, for example the Carboniferous lycopsid rooting organ Stigmaria, a morphogenus, has been found attached to different genera of stems. In a case like this, the name of the part is maintained, even though the entire plant has subsequently been reconstructed. The name is also maintained because fossil Stigmaria is still found unattached, and a name is necessary to describe and study the part. In addition, some fossil plant parts, despite extraordinary preservation, cannot be distinguished as belonging to only one group of plants. For example, some species of the Carboniferous foliage type Sphenopteris were borne by marattialean ferns (Chapter 11), and other species of this morphotaxon were produced by lyginopteridalean seed ferns (Chapter 16). Since most plants are constructed of many parts, referring to the entire plant once it has been reconstructed requires a complex system of nomenclature. In general, three procedures are followed when the entire plant is reconstructed: (1) the entire organism is provided with a new name, (2) the whole organism bears the generic name of the part that has priority, that is the first part given a formal name, or (3) the whole plant is referred to informally, for example the “Lepidodendron” plant. Non-paleobotanists may find the nomenclature used in paleobotany confusing and perhaps cumbersome, but the way that fossil plants are preserved necessitates its use, and it is currently the only system that provides for an orderly arrangement of names and, most importantly, for the retrieval of information on plant parts. Some have suggested that the Linnaean system of nomenclature be abandoned for certain types of fossils, especially the use of names that suggest affinities with extant taxa when the exact affinities are unknown (Spicer, 1986, for Cretaceous and Cenozoic angiosperm leaves). Hughes (1989) championed a system in which pollen and other plant parts were given artificial names, so-called paleotaxa (Chapman and Smellie, 1992 for fossil wood), but the system has never been in wide use among paleobotanists. CLASSIFICATION OF ORGANISMS

Each author has his or her own ideas concerning the way organisms should be organized, or in the case of plants, whether they represent a single kingdom or multiple kingdoms. With this in mind, the classification scheme in Appendix 1 is presented merely as a guide to the groups of

algae, fungi, bryophytes, and vascular plants that are discussed in this book. In the case of some groups, such as the hyperdiverse flowering plants, there are so many families with a meager fossil record, or no fossil record at all, that it would be impossible to include them all, so we have tried to provide a sampling of major groups and interesting examples. For the angiosperms, we have followed the system in Cronquist (1988) for the most part, with attention to the system of the Angiosperm Phylogeny Group (1998, 2003); for the algae (Chapter 4), the system in Lee (1999), for the hornworts and bryophytes (Chapter 5), the system in Frahm (2001a), and for the fungi (Chapter 3), The Mycota, Volumes VIIA and VIIB (McLaughlin et al., 2001a, 2001b). Some readers may wish to adapt the plant groups presented in the following chapters to a system with which they feel more comfortable.

BACKGROUND READING There are many approaches that one might take in the preparation of a volume dealing with fossil plants. Through the years there have been many excellent books on paleobotany that have covered the discipline from many perspectives. We have included a number of these in the bibliography so that the reader may obtain additional information on some of the plant groups presented here, or additional ones not discussed. The following volumes (and references cited therein) will provide supplemental details on many fossil plants: Hirmer, 1927; Arnold, 1947; Darrah, 1960 (FIG. 1.50); Andrews, 1961; Delevoryas, 1962; Mägdefrau, 1968; Archangelsky, 1970; Banks, 1970; Emberger, 1968; Beck, 1976a, 1988; Hughes, 1976 (FIG. 22.15); Remy and Remy, 1977; Taylor, 1981; Thomas, 1981a; Stewart, 1983 (FIG. 14.116); Gensel and Andrews, 1984; Tiffney, 1985; Spicer and Thomas, 1986; Meyen, 1987; Thomas and Spicer, 1987; Friis et al., 1987; Taylor and Taylor, 1993; Kenrick and Crane, 1997a; Stewart and Rothwell, 1993; Jones and Rowe, 1999 (for methods in paleobotany and palynology); Gensel and Edwards, 2001; Willis and McElwain, 2002; Anderson and Anderson, 2003; Kenrick and Davis, 2004; Anderson et al., 2007; and the Traité de Paléobotanique series, published under the direction of E. Boureau (FIG. 10.108) (Boureau, 1964; Boureau et al., 1967; Andrews et al., 1970; Boureau and Doubinger, 1975) (FIG. 15.14).

2 PRECAMBRIAN LIFE THE ORIGIN OF LIFE ON EARTH ......................................... 44

OXYGENATION OF THE EARTH (2.45–2.2 Ga) ................... 57

Origin of Life: Theory and Biology ................................................... 46

PROTEROZOIC LIFE .....................................................................59

EARLIEST RECORD OF LIFE ON EARTH ............................ 47

Paleoproterozoic..................................................................................59

Historical Background ....................................................................... 47

Origin of Eukaryotes .......................................................................... 61

Earliest Records of Life: Paleoarchean (3.6–3.2 Ga) ......................... 47

Mesoproterozoic ................................................................................ 64

MESOARCHEAN–NEOARCHEAN LIFE .................................54

Neoproterozoic ................................................................................... 64 CONCLUSIONS ............................................................................ 70

CONCLUSIONS: ARCHEAN LIFE ............................................ 55

We find no vestige of a beginning, no prospect of an end James Hutton (1788)

Although Hutton was speaking about the geology of the Earth in this famous quotation, it could equally apply to the presence of life on Earth. The past 10 years have seen an explosion of information on the origin of life and evidence for the earliest life on our planet. On the geologic side, this explosion has benefited from the applications of new or improved geochemical, isotopic, and microscopic techniques to Precambrian rocks. On the biologic side, the discovery of life in the deep subsurface, both in terrestrial habitats and in the deep ocean, has broadened our knowledge of where organisms are able to survive and has also influenced ideas on where life may have begun on Earth. A textbook in paleobotany would not be complete without a discussion of the earliest evidence of life on Earth. Admittedly, the first organisms were neither plant nor animal, but it was from such simple bacterial or archaeal biological systems that more complex types of plants later evolved. In the past, geologic time was divided into two major eons: the Cryptozoic (literally, hidden life), now called the Precambrian, and the Phanerozoic (visible life) (inside front and back covers). Unlike the Phanerozoic timescale, where divisions are based on rock units (see Chapter 1), the divisions of time in the Precambrian (FIG. 2.1) are based on

absolute, that is, radiometric dates. These divisions are the Archean, for rocks older than 2.5 Ga (Giga annum or billion years), and the Proterozoic for rocks dated from 2.5 Ga to the Precambrian–Cambrian boundary, which is currently 542 Ma (Gradstein et al., 2004). The last period within the Proterozoic was recently established as the Ediacaran (Knoll et al., 2006a), and it is the only period within the Precambrian based on chronostratigraphy, that is, there is a series of rocks, called the global stratotype, that serve as the reference point for the Ediacaran. The period of time between 4.5 and 4.0 Ga is termed the Hadean Eon, although this is not a formally accepted name. The age of the Earth itself is currently thought to be somewhere around 4.53 Ga; thus, the Precambrian represents almost 88% of geologic time! Isotopic analyses suggest that continental crust formed soon after accretion of the planet, possibly by 4.5–4.3 Ga (Bizzarro et al., 2003; Harrison et al., 2005), and that oceans existed by 4.4–4.2 Ga (Mojzsis et al., 2001; Wilde et al., 2001). These dates are based on analyses of the radioactive decay of the elements lutetium and hafnium (176Lu decays to 176Hf with a half-life of 370 myr) within detrital zircons (see Chapter 1) from the Jack Hills in Western Australia (Kramers, 2001). Due to the continuing

43

Age (Ma)

System period

Erathem era

Paleobotany: the biology and evolution of fossil plants

Eonothem eon

44

542 Ediacaran Neoproterozoic

~635 Cryogenian 850 Tonian

Proterozoic

1000 Stenian Mesoproterozoic

1200 Ectasian 1400 Calymmian 1600 Statherian 1800

Precambrian

Paleoproterozoic

Orosirian 2050 Rhyacian 2300 Siderian 2500

Neoarchean 2800 Archean

Mesoarchean 3200 Paleoarchean 3600 Eoarchean 4000 Hadean (informal) ~4600

Figure 2.1 International Stratigraphic Chart showing the

Precambrian. (From the International Commission on Stratigraphy; Courtesy F. Gradstein.)

dynamic processes of plate tectonics, new crust is created and old subducted, and there are few rocks still available that formed in the Paleoarchean. Currently the oldest known rocks are from the Acasta gneisses of the Great Slave Lake area in Canada (Northwest Territories), dated at 4.03–4.0 Ga (previously 3.96 Ga; Bowring et al., 1989) based on uraniumlead isotopes (U-Pb) (Bowring and Williams, 1999) (FIG. 2.2). Slightly younger rocks are known from Isua and Akilia, West Greenland, including the oldest known sedimentary rocks, dated to 3.82 Ga (Manning et al., 2006). Although no older rocks have survived, several studies suggest that continental crust was present, based on the occurrence of detrital zircons, which were weathered out of older rock. The zircons have

ages slightly older than 4.0 Ga (Iizuka et al., 2006) suggesting that weathering processes were also active at this time. Geochronology (dating) methods have improved significantly in recent years; however, the controversy over these very old rocks will remain, as they all have experienced a long and complex history of metamorphism (Kamber et al., 2001). Although it was previously thought that the Earth formed over many hundreds of millions of years, the convergence of dates for the oldest continental crust with the age of the Earth itself suggests that continental crust began to form very soon after Earth accreted, perhaps within 30 myr (Boyet and Carlson, 2005). What was the environment of Earth at that time? The composition of Earth’s early atmosphere has long been a topic of debate. Stanley Miller, in his classic experiments in the 1950s (Miller, 1953), suggested that Earth’s early atmosphere was a reducing one (high hydrogen content), composed of methane (CH4), ammonia (NH3), hydrogen (H2), and water (Miller, 1953). When a spark (e.g., from lightning) was introduced into this system in the laboratory, amino acids formed. The reducing environment hypothesis fell out of favor for many years, although it has resurfaced in a model that contains more CO2 than previous models (Tian et al., 2005) and seems to be the currently prevailing model. An important aspect of the presence of CO2 and especially methane in the early atmosphere is that they are both greenhouse gases. The early Sun is thought to have been 30% less bright than today, an idea called the Faint Young Sun hypothesis; thus, some means of global warming was needed in the Archean for Earth to be hospitable to life (Kasting, 2005). Concentrations of greenhouse gases, especially methane, which is a much stronger greenhouse gas than CO2, would have been important to increase global temperature. These gases may have formed a haze-like atmosphere (Kasting, 2005), and it has been suggested that this haze may have protected early life from harmful ultraviolet rays before the ozone layer was formed. Oxygen is believed to have been only a very minor component of the atmosphere before 2.4 Ga. Most recent research suggests that the early Earth before 3.2 Ga was hot, perhaps as hot as 60–73°C (Lowe and Tice, 2007). Whatever the initial composition of the atmosphere on Earth, it must have affected the evolution of life and, in turn, was itself changed by the presence of organisms, as will be discussed in the following sections.

THE ORIGIN OF LIFE ON EARTH As far as we know, the land surface supported no life at this time, but the accumulation of bodies of water eventually provided an environment for life to evolve and flourish. The

CHAPTER 2

EONS

ERAS and EONS

Mesozoic 0.251

Neoproterozoic

Paleozoic

Oldest traces of invertebrate animals Bitter Springs biota 0.83–0.80 Ga

Oldest sphaeromorphic acritarchs 1.85–1.4 Ga

Gunflint biota 1.9 Ga

Cyanobacterial biomarkers and steranes (eukaryotes?) 2.7–2.6 Ga

Neoarchean

Major deposits of banded iron formations

Precambrian

Decline of uraninite

0.542 Chuaria and other megascopic fossils worldwide 1.0 Peak in stromatolite diversity 1.3–1.1 Ga

1.6 Paleoproterozoic

Proterozoic

Bangiophyte red alga 1.2 Ga Oldest spiny acritarchs 1.5 Ga

2.0

Evidence of atmospheric O2 2.5

2.8

Oldest terrestrial rocks (Great Slave Lake, Canada) 4.03–4 Ga

Eoarchean Paleoarchean

Presence of oxygen, 2.45–2.2 Ga 1. Decline of BIFs 2. Disappearance of uraninite 3. Appearance of red beds 4. Change in sulfur isotopes

3.2 Strelley Pool chert—diverse stromatolites 3.430 Ga 3.6

4.0 Hadean

Oldest stromatolites (Warrawoona Group) 3.5–3.4 Ga

Archean

Oldest microbial fossils 3.5–3.4 Ga

Red beds increase in abundance 2.0–1.8 Ga Oldest red beds ~2.2 Ga

Mesoarchean

Barberton Greenstone Belt 3.5–3.2 Ga

45

AGE (Ga) 0 0.66

Mesoproterozoic

Oldest calcareous algae

Phanerozoic

Cenozoic

PRECAMBRIAN LIFE

Last major bombardment phase 4.1–3.8 Ga

Oceans present 4.4–4.2 Ga 4.5 4.6

Continental crust 4.5–4.2 Ga Age of the Earth 4.53 Ga

Figure 2.2 Summary of major events in Precambrian history. (Modified from Schopf et al., 1983; and Taylor and Taylor, 1993.)

proliferation of life is thought to have been hampered by early bombardment of Earth by asteroids, which continued until around 4.1–3.8 Ga, the Late Bombardment phase, although this idea is still controversial (Koeberl, 2006). Most theories on the origin of life on Earth suggest that life began with the synthesis of organic compounds, both in the atmosphere and on the surface of the Earth (Chang et al., 1983). These

became increasingly complex, and eventually molecules arose that had the ability to duplicate themselves (selfreplicating) and to perform other complex syntheses. The reports of organic matter in carbonaceous chondrites, a type of meteorite, confirm that organic synthesis has occurred in our solar system and beyond (Ehrenfreund et al., 2001), and numerous laboratory experiments have been performed

46

Paleobotany: the biology and evolution of fossil plants

to replicate these early syntheses. Many researchers have suggested that an RNA (ribonucleic acid) world preceded the DNA (deoxy-ribonucleic acid) world that exists today (Gesteland et al., 2006). The RNA world hypothesis is based in part on the fact that RNA is a simpler molecule than DNA. It can also replicate itself, encode and build proteins, and function to catalyze reactions (Joyce, 2002; W. R. Taylor, 2006). Moreover, RNA is the principal component of ribosomes, the intracellular bodies where proteins are synthesized, thus suggesting that RNA, rather than DNA, was the original information-containing molecule governing protein synthesis. Others suggest that a pre-RNA molecule, perhaps a peptide nucleic acid, may have been important in the early development of life (Nelson et al., 2000). This chemical synthesis of life is termed abiogenesis. ORIGIN OF LIFE: THEORY AND BIOLOGY

Darwin hypothesized in the late nineteenth century that life on Earth probably arose in a “warm little pond,” and since then this idea has been widely believed. There are several barriers to the origin of life in shallow pools, however, including the necessity of shielding early life on the surface from damaging ultraviolet rays. There is also the question of what would have happened to early life on the surface during the Late Bombardment phase, although it has been suggested that the craters formed by the bombardment would have provided an excellent environment for early life (Cockell, 2006). One of the problems with the origin of life in a primeval soup of prebiotic, organic molecules is whether a mix of molecules in liquid would be sufficiently concentrated to ensure the reactions necessary to form more complex macromolecules. It has been suggested that organics may have been attached to layers of clay or pyrite, thus providing the close proximity needed to catalyze reactions (Ferris, 2006). If the early Earth was hot, as some researchers suggest (Lowe and Tice, 2007), these surface organisms would most likely have been anoxygenic photosynthetic hyperthermophiles. In contrast to the warm little pond hypothesis, the idea that life may have arisen on the ocean floor is more recent. This hypothesis gained support when entire ecosystems were discovered in the early 1980s surrounding deep-sea hydrothermal vents (Waldrop, 1990). Subsequent work has shown that bacterial and archaeal life exists deep in the terrestrial subsurface as well (Amend and Teske, 2005; Chapter 1). In the past 20 years, the diversity of life in extreme environments (extremophiles) has been widely demonstrated, ranging from Antarctic ice to hot springs (thermophiles). Studies of microbial evolution based on rRNA propose that

a hyperthermophile represents the ancestral condition in the Archaea and possibly in the Eubacteria as well (Woese, 1987). This deep biosphere is chemosynthetic and thus not dependent upon sunlight for energy. Although some modern hyperthermophiles live in an aerobic environment, because of the low solubility of O2 at such high temperatures and the presence of reducing gases such as H2S, these organisms are basically anaerobic (Stetter, 2006). These findings suggest that Archean life could have survived periods of asteroid bombardment deep within the ocean or within rocks in earth’s crust. As early as 1988, Gunter Wächterhäuser proposed that life may have arisen around deep-sea vents. The Iron–Sulphur World hypothesis suggests that life began in association with pyrite crystals (FeS2). This idea holds that carbonate, phosphate, and sulfide ions would be attracted to iron pyrite and would rapidly cover every surface. With prebiotic molecules concentrated in this way, and in the high heat and pressure of hydrothermal vents, which are a ready source of iron and sulfur, reactions could proceed much more rapidly than they could at surface temperatures and pressures (Russell and Hall, 2006). Experimental evidence has shown that pyruvic acid (Cody et al., 2000), acetic acid, and peptides readily form in these conditions (Huber and Wächterhäuser, 1997, 1998). More recently, support for the origin of life near hydrothermal vents has come from an interdisciplinary perspective involving biophysics in combination with molecular biology. In a simulation experiment, Baaske et al. (2007) demonstrated that nucleotides concentrated around hydrothermal pores, such as the pore spaces of rocks found at vents. This concentration depended only on the size of the pores and a hydrothermal gradient. These authors suggest that prebiotic molecules could have been concentrated in this way to form the first protocells. Thus, the Iron– Sulphur World hypothesis appears to be a viable alternative to the warm little pond and provides several advantages over that idea, including survival of life during bombardment and changes in surface conditions on the early Earth (Wächterhäuser, 2000). Although originally considered only in the realm of science fiction, the possibility that the early Earth was seeded with organic matter from comets or asteroids, termed panspermia, has received serious consideration. For example, Anders (1989) calculated that only dust-sized particles would be slowed by the atmosphere sufficiently to prevent destruction of organics on impact. Chyba et al. (1990), however, assumed that early CO2 atmospheres would be denser and estimated that organics could have accumulated at the rate of 106–107 kg/yr from 4.5 to 3.9 Ga during the Late

CHAPTER 2

PRECAMBRIAN LIFE

47

Bombardment phase. The announcement of evidence of life in a Martian meteorite from the Allan Hills, Antarctica (ALH84001; McKay et al., 1996) seemed to support the idea of panspermia. The evidence has been refuted, however, based on several different types of data (Schopf, 1999; Barber and Scott, 2002). There are a number of excellent reviews on the prebiotic (i.e., pre-cellular) origin of life, as well as the atmosphere and environment of early Earth, including books on Precambrian geology and life (Coward and Ries, 1995; Lazcano and Miller, 1996; Orgel, 1994; Schopf, 2002; Knoll, 2003b; Eriksson et al., 2004; Schoonen et al., 2004; Kesler and Ohmoto, 2006; Reimold and Gibson, 2006; Schopf et al., 2007a).

EARLIEST RECORD OF LIFE ON EARTH HISTORICAL BACKGROUND

Before discussing current evidence, it is important to mention historical work on Precambrian life. Although there were some earlier reports of Precambrian life, the work of Stanley Tyler and Elso Barghoorn (FIG. 2.3) on the Gunflint Iron Formation (Gunflint chert) of the Canadian Shield provided the first detailed, irrefutable evidence for life in the Precambrian. These researchers began with a preliminary report on the organisms (Tyler and Barghoorn, 1954) and the description of coal from the same area (Tyler et al., 1957). A detailed morphological analysis of the organisms themselves appeared in 1965 (Barghoorn and Tyler, 1965). At the time, the diversity of organisms they described from this 1.9 Ga site (Paleoproterozoic) must have seemed phenomenal. Their work was quickly followed by other reports of Proterozoic fossil organisms and eventually by the discovery of Archean microorganisms. The Gunflint biota will be discussed in more detail later (see section on “Paleoproterozoic” below). EARLIEST RECORDS OF LIFE: PALEOARCHEAN (3.6–3.2 Ga)

The earliest records of life on Earth are of several types, including geochemical evidence (biomarkers or carbon or sulfur isotopes), microfossils (body fossils of microbes), stromatolites, and other microbially influenced sedimentary features, and indirect evidence from molecular phylogenies of living microbes. Each record is controversial if considered separately, but the totality of evidence from many sources continues to push back the date of earliest life. Each type of

Figure 2.3 Elso S. Barghoorn. (Courtesy Schultes and Knoll,

1987.)

evidence has its own positive and negative aspects, which we will try to address on a case-by-case basis. The majority of the research on early life comes from three localities: (1) Warrawoona Group in the Pilbara Craton, Western Australia (3.515–3.427 Ga); (2) Barberton Greenstone Belt, Kaapvaal Craton of South Africa and Swaziland (3.55–3.33 Ga); and (3) the Isua Group in southwest Greenland, including the island of Akilia (3.9–3.7 Ga). GEOCHEMISTRY The geochemical evidence for life in the Paleoarchean continues to increase. Although some of the records remain controversial, the increasing number of records from various sources, such as biomarkers, stable isotopes of carbon and sulfur, and metal isotopes, supports the hypothesis that life began on Earth soon after the formation of continental crust and oceans. The major problem with the Archean signature of life, however, is the nature of these very old rocks themselves. Many have undergone tectonic events and all are metamorphosed to some extent, which can have a significant effect on isotopic signatures. In addition, even very hard cherts are porous to some degree, so contamination after deposition must also be taken into consideration. Evidence points to the earliest life, probably consisting of anaerobic, chemosynthetic microorganisms. Free oxygen is not found in appreciable amounts until the Paleoproterozoic, so these

48

Paleobotany: the biology and evolution of fossil plants

early microorganisms are believed to have lived in a reducing environment. Shen et al. (2001) reported on the presence of microscopic sulfides in 3.47 Ga barites from the Dresser Formation (3.515–3.458 Ga), North Pole, Australia. The δ34S isotopic evidence supports the presence of sulfate-reducing microbes. Based on molecular phylogenetics of extant Archaea and Bacteria, sulfate reduction is thought to be an ancient metabolic process (Shen and Buick, 2004). Although Philippot et al. (2007) disagree that these sulfides were formed by sulfate reducers, they do acknowledge that they are biogenic, that is, produced by organisms. They conclude that the mix of sulfur isotopes points to a microorganism that was metabolizing elemental sulfur rather than sulfates. Another important source of Paleoarchean rocks is the Barberton Greenstone Belt of South Africa and Swaziland. Using a combination of petrographic observations, X-ray mapping of elements, and carbon isotope measurements, Banerjee et al. (2006) described 3.5–3.4 Ga micron-scale tubular structures as evidence of life, although aspects of this research are controversial (Kerr, 2004). The tubes have organic carbon associated with them, as well as low δ13C isotopes indicative of life. The samples come from subaerial volcanic rocks, including pillow lavas—not a typical site in which to look for early life. The tube morphology and texture, however, are almost identical to etching seen in the glassy edges of modern pillow lavas, at least some of which are microbially mediated (see also Furnes et al., 2004). Similar structures have been described from 3.35 Ga rocks from Western Australia (FIG. 2.4) (Banerjee et al., 2007). A range of organisms have been found around modern hydrothermal, deep-sea vents, including members of both the Bacteria and the Archaea, although the Archaea appear to dominate at very high temperatures. Some of these are sulfate reducers and obligate chemoautotrophs (Reysenbach and Shock, 2002), but methanogenic microbes are also found. Studies on modern basaltic glass show that the presence of microbes greatly enhances weathering and may produce different chemical products than abiotic weathering (Staudigel et al., 1998). Kerogen, the insoluble carbonaceous matter in rocks, is another source of geochemical evidence of life (Chapter 1). Kerogen has been reported from Precambrian rocks of various ages, but there is controversy over interpretation of the carbon isotopic data from some of this kerogen (see discussion in Pavlov et al., 2001). Marshall et al. (2007) examined kerogens from the 3.4 Ga Strelley Pool chert (North Pole, Australia), which is also an important source of stromatolites (see below). Using a variety of spectroscopic techniques,

Figure 2.4 Photomicrograph of interpillow hyaloclastite from Euro Basalt showing tubular structures (Paleoarchean). Bar  20 μm. (Courtesy N. R. Banerjee.)

including Fourier transform infrared, Raman, and NMR (nuclear magnetic resonance) spectroscopy, they found similarities between this Archean kerogen and younger, Mesoproterozoic kerogen (1.45 Ga), which is known to be biogenic. Based on these results, Marshall et al. (2007) suggested that the Strelley Pool kerogen is also derived from organic matter that had a biogenic origin. After analyzing hundreds of specimens from the North Pole area using petrography and isotope analyses, Ueno et al. (2004) concluded that the kerogens in their samples could have been produced by anaerobic chemoautotrophs, including methanogenic microbes. Their evidence suggests that the organic matter did not come from aerobic photoautotrophs, such as photosynthetic cyanobacteria (Chapter 3). In another study on fluid inclusions in quartz from the 3.5 Ga Dresser Formation (North Pole), Ueno et al. (2006) report the presence of microbial methane (CH4). The methane is highly depleted in δ13C isotopes, suggesting it was produced by methanogenic microbes. Controversy has surrounded the possible evidence of life in rocks from southwest Greenland ever since graphite and microfossils (see below) were found there in the late 1970s. As newer geochemical techniques have become available, the interpretation of Isua rocks has recently generated a new wave of controversy. Graphite was described from Isua rocks almost 20 years ago (Schidlowski, 1988) and more recently from the island of Akilia (Mojzsis et al., 1996), which is dated as 3.825Ga (Manning et al., 2006). The graphite occurs as inclusions within apatite (a calcium phosphate mineral) and was interpreted as being biogenically produced,

CHAPTER 2

based on the isotopically light (i.e., more negative) carbon isotopes obtained from it. Much of the controversy centers around the exact strata from which samples were obtained, as the rocks are exceptionally complex; additional sampling has yielded no graphite (Lepland et al., 2005). Others have suggested that the carbon entered the deposits at a later time than the formation of the rocks themselves (Fedo and Whitehouse, 2002). McKeegan et al. (2007) used Raman spectroscopy to examine the same samples used by Mojzsis et al. (1996) and concluded that graphite is completely contained within apatite, and that it is isotopically light carbon. This supports the claim that the carbon represents a biomarker for ancient life. For a review of the controversy and the issues involved, see Eiler (2007) and references cited therein. MICROFOSSILS (BODY FOSSILS) Although it is clear that life was present in the Paleoarchean, the fossil evidence for life at this time is difficult to interpret due to the microscopic nature of the organisms, a lack of diagnostic and preservable morphologic features to distinguish them, and the changes that have occurred in these ancient rocks since they were first formed. For these reasons, there have been a number of reports of Archean unicells, filaments, and other growth forms that have later been reinterpreted as representing abiogenic structures (i.e., not formed by living organisms). These structures have been labeled in the literature as either pseudofossils (formed abiotically) or dubiofossils (uncertain origin). ISUA GREENSTONE BELT, GREENLAND. Perhaps one of the most discussed examples of a dubiofossil was the description of Isuasphaera, a yeast-like microorganism found in 3.8–3.7 Ga Isua Greenstone Belt metamorphic rocks from Greenland (Pflug, 1978). Isuasphaera consisted of spherical–elliptical structures, some of which gave the appearance of budding yeasts. The description of these complex, eukaryotic organisms from some of the oldest rocks on Earth generated a great deal of controversy at the time, which extended to analyses of amino acids supposedly preserved in these rocks. Bridgwater et al. (1981) described a broad range of diameters for the spherical bodies in these rocks and suggested that the structures represent limonitestained, fluid-filled inclusions in the metaquartzite rather than microorganisms. Their hypothesis was widely accepted (Bridgwater et al., 1981; Schopf and Walter, 1983). More recently, Appel et al. (2003) laid the matter of Isuasphaera to rest. Appel was the one who originally collected the samples containing the spherical bodies, and he and his team were subsequently able to reexamine the exact site and show that

PRECAMBRIAN LIFE

49

the rocks in this zone (metamorphosed chert) had undergone extreme stretching deformation. Since the spherical bodies could not have been preserved through this event, these researchers concluded that they were formed as a result of pre-Quaternary weathering. WARRAWOONA GROUP, AUSTRALIA. A similar controversy has developed over the rocks and vestiges of life from the Warrawoona Group in the North Pole Dome area of Western Australia. Currently, the oldest microfossils come from this area, in the Apex chert (3.465 Ga), now part of the Salgash Subgroup of the Warrawoona Group (Van Kranendonk et al., 2002). The Warrawoona Group rocks (3.515–3.427 Ga) occur within the East Pilbara terrane and contain both stromatolitic deposits (with and without microfossils) (Lowe, 1980; Walter et al., 1980; Walter, 1983; Allwood et al., 2006, 2007) and cherts that contain preserved microfossils. The relatively high diversity of the fossil assemblage from the Apex chert (Awramik et al., 1983; Schopf and Walter, 1983; Schopf, 1993 and references cited therein) is noteworthy and provides evidence that life may have arisen earlier than 3.556 billion years ago. This silicified microbiota includes four types of filamentous bacteria, colonial unicells, organic spheroids, and radiating filaments, which are hypothesized to have been living in a shallow-water environment. The spheroids and radiating filaments are now regarded as only possible microfossils or dubiofossils. In total, 11 species of filamentous organisms are known, which range from 0.5 to 19.5 μm in diameter; although the smaller diameter fossils are probably filamentous bacteria, the larger ones are most comparable to cyanobacteria. The four types of simple filaments in this microbiota include (1) very narrow forms, Archaeotrichion, that range from 0.3 to 0.7 μm in diameter and up to 180 μm long; (2) possibly septate filaments, up to 340 μm long (0.8–1.1 μm in diameter), Eoleptonema; (3) large, tubular sheaths 3–9.5 μm wide and up to 600 μm long (Siphonophycus); and (4) large septate filaments, up to 120 μm long, which consist of mostly isodiametric cells 4–6 μm in diameter (Primaevifilum) (FIG. 2.5). Based on morphology alone, these microorganisms can be compared with a large number of living bacteria and cyanobacteria (Chapter 3), including anaerobic, autotrophic, and heterotrophic microorganisms. Perhaps the most interesting fossils in this microbiota are the relatively large (8–20 μm in diameter) unicells enclosed by what are described as lamellated sheaths. Morphologically, these fossils are comparable to extant chroococcaleans (cyanobacteria) and, as such, may represent the first evidence

50

Paleobotany: the biology and evolution of fossil plants

Figure 2.5 Primaevifilum amoenum (Warrawoona Group).

Bar  10 μm. (Courtesy J. W. Schopf.)

of oxygen-producing, photosynthetic life. There have been several accounts disputing the biotic origin of these fossils (Brasier et al., 2002), followed by a series of papers on the geochemistry of the deposits, including the presence of kerogen in the cherts (Pinti et al., 2001) and the use of various techniques, such as Raman spectroscopy, to study the carbon isotopes in the microstructures themselves and to reveal the microfossils in three dimensions (summarized in Schopf et al., 2007b). Some of the difficulties with previous isotopic results on these ancient structures and ways to address these problems are discussed in Marshall et al. (2007). More recently, analyses have shown that the kerogen in the chert is of biological origin (Derenne et al., 2008). The strongest evidence that at least some of the Apex microfossils represent life is the increasing number of reports of structurally preserved microfossils from rocks of similar or slightly younger age—some 14 rock units containing 40 described morphotypes of microfossils (Schopf et al., 2007b). For example, Rasmussen (2000) described filamentous microfossils from the slightly younger Sulfur Springs Group (3.235 Ga) from the same area of the Pilbara Craton in Australia. These organisms were interpreted as living in the rock pores in the shallow subsurface of the seafloor in a hydrothermal environment. They consist of unbranched filaments, 0.5–2 μm in diameter and up to 300 μm long. The filaments are abundant, occurring in dense groups, with many intertwined. These microfossils are interpreted as being the remains of thermophilic, chemotrophic microorganisms, similar to prokaryotes that occur in those environments today. More recently, Sugitani et al. (2007) described a diverse assemblage of microfossils from black chert in the slightly younger Gorge Creek Group (3.19–2.97 Ga, Warrawoona Group), using a combination of morphology (via thin sections) and geochemistry (isotopic analyses of C, N, H, and S). The fossils are an integral part of the rock (i.e., not contaminants) and include four morphological types including threadlike filaments, small spherical structures (FIG. 2.6), spindle-shaped microfossils, and film-like objects (FIG. 2.7). The fossils are abundant and all are 1 μm in diameter, with some filaments extending up to 100 μm in length.

Figure 2.6 Colony-like aggregation of small spheroidal

microstructures (Warrawoona Group). Bar  10 μm. (Courtesy K. Sugitani.)

Perhaps the most interesting fossils are these film-like structures, some of which have small spheres associated with them. Some sheets are folded or wrinkled and range from 50 to 500 μm in size. These may represent parts of microbial mats or perhaps fossilized biofilms (Gall, 1990), which can be especially important in the preservation of many other fossils (Chapter 1). One of the most interesting recent discoveries is a diverse assemblage of silicified microfossils from the 3.446 Ga Kitty’s Gap chert of the Warrawoona Group (Westall et al., 2006a). This deposit occurs in volcaniclastic rocks and represents a nearshore environment of channel infill, termed channel-and-flat sediments. The biota includes two sizes of coccoid cells (0.4–0.5 μm and 0.75–0.8 μm), rare, rod-shaped microfossils 1 μm long, and small, short filaments about 0.25 μm wide. Carbon isotopes from the same layers are very light (26‰ to 30‰), consistent with the presence of microbial life. The coccoids occur in colonies, and the cells within each colony are of the same size and shape; cell sizes differ slightly between colonies, as would be expected of living organisms. There is evidence of life in the form of dividing cells (FIG. 2.8) and the formation of chains of cells, and the presence of collapsed cells with wrinkled surfaces, suggesting cell death. Most important, extracellular polymeric

CHAPTER 2

PRECAMBRIAN LIFE

51

Figure 2.9 Parallel and overturned filaments in a microbial

mat (Warrawoona Group). Bar  10 μm. (From Westall et al., 2006b.) Figure 2.7 Film-like microstructures with small sphere (arrow) (Warrawoona Group). Bar  50 μm. (Courtesy K. Sugitani.)

EPS

Figure 2.8 Colony of coccoidal microfossils and extracel-

lular polymeric substances (EPS). Arrow indicates small coccoid (Warrawoona Group). Bar  2 μm. (From Westall et al., 2006a.)

substances (EPS) are ubiquitous in these deposits. EPS are produced by most bacteria (e.g., the mucilaginous sheaths of cyanobacteria) and serve to attach the organism to a substrate and to aggregate cells together into biofilms (Chapter 1). In the Kitty’s Gap biota, the EPS either surround individual or dividing coccoid cells, or extend laterally to cover and embed entire colonies of coccoid cells (FIG. 2.9) (Westall et al., 2006b). Westall et al. interpreted the filaments as

possible anoxygenic photosynthetic microbes. A biofilm with coccoid colonies occurs as a coating on volcanic grains and is interpreted as including chemolithotrophic microbes. Thus, the Kitty’s Gap chert biota is not only excellent evidence of a microbial ecosystem that lived at about the same time as the Apex microfossils, but also provides compelling evidence for the presence of ancient microbial biofilms. BARBERTON GREENSTONE BELT, SOUTH AFRICA. Another important source of data on Archean life is the Onverwacht Group (3.55–3.33 Ga) in the eastern Transvaal, South Africa. The younger Fig Tree Group (3.26–3.23 Ga) is from this same area (Lowe and Byerly, 1999; Altermann, 2001). Although a number of spherical unicells have been described from the Onverwacht cherts, many of them are now regarded as questionable microfossils (Schopf and Walter, 1983). Smaller cells, including several thought to be in the process of dividing (Knoll and Barghoorn, 1977), are more likely to represent biogenic remains. Walsh and Lowe (1985) described 3.5 Ga filamentous microfossils from the Hooggenoeg Formation (Onverwacht Group). These consist of threadlike filaments 0.2–2.6 μm in diameter. Filaments from the overlying Kromberg Formation are 0.1–0.6 μm in diameter and 10–150 μm long (Walsh and Lowe, 1985). Stromatolites from the younger Fig Tree Group (3.26– 3.23 Ga) provide evidence for the complexity of Archean life (Byerly et al., 1986). More recent studies on the Barberton rocks include the discovery of tubular structures in basaltic glass (Furnes et al., 2004; Banerjee et al., 2006) (see above), the presence of microbial mats (Westall et al., 2006b),

52

Paleobotany: the biology and evolution of fossil plants

and the research of Noffke et al. (2006) on sedimentary structures (see section “Sedimentary Evidence”). Tice and Lowe (2004) described laminated, carbonaceous material that is believed to represent the remains of microbial mats in the Buck Reef Chert (3.416 Ga). Possible coccoid and rod-shaped bacteria, replaced by minerals, have also been described from the Onverwacht Group (Westall et al., 2001), although the biogenicity of some of these structures has been questioned (Altermann, 2001). STROMATOLITES Especially compelling evidence of early life occurs in the form of stromatolites, which can be defined as layered (laminated), organosedimentary microbial deposits that form on

Figure 2.10 Modern stromatolites in Shark Bay, Western

Australia. (Courtesy J. W. Schopf.)

the bottom of a body of water (Riding, 1999). Although the term stromatolite has been used for layered structures of abiogenic origin, we will use the definition as discussed in Riding (1999) to include structures that are assumed to be biogenic. Stromatolites were widespread in the Precambrian, beginning at 3.5 Ga (Buick et al., 1981) and continuing through the Proterozoic. They are much less common in the Phanerozoic, although they occur in all periods as local reefs; stromatolites are found today only in relatively restricted environments. Modern stromatolites are formed by aggregations of microorganisms, most commonly cyanobacteria, but also green algae (Chlorophyta) and diatoms (Bacillariophyceae) that trap, bind, or precipitate calcium carbonate (CaCO3) in thin layers. Stromatolites are lithified by calcium carbonate when “living,” but most of the Precambrian fossilized stromatolites were taphonomically replaced with silica (for more information on stromatolites, see Walter, 1976; Riding, 1991; Bertrand-Sarfati and Monty, 1994). Probably the best-known site where stromatolites are actively formed today is Shark Bay (FIG. 2.10), a hypersaline lagoon along the western coast of Australia. The organisms that form these stromatolites are filamentous and coccoid members of the Cyanobacteria (Chapter 3), which live in colonies on the upper surface of the accumulating calcium carbonate. These microorganisms typically have mucilaginous or gelatinous, polysaccharide sheaths (EPS) that serve to trap particles of carbonate in the seawater. As they photosynthesize, the cyanobacteria deplete the carbon dioxide in the immediately surrounding water, which also causes the precipitation of calcium carbonate (FIG. 2.11). The CaCO3 deposition continues in thin layers, or lamellae (FIG. 2.12), as the algal colonies continue to grow on the

Figure 2.11 Modern stromatolites from Laguna Mormona, Mexico. (Courtesy J. W. Schopf.)

CHAPTER 2

upper surface of the columnar calcium carbonate. When the colony dies, the calcareous structure may persist as evidence of the organisms that formed it. Stromatolites were widespread and important during the Precambrian, as it was a microbially dominated world until the latest Proterozoic. Some Proterozoic oil deposits found associated with stromatolites are thought to be microbially produced. Some fossil stromatolites contain remains of the microorganisms that formed them, but more often only the layered structure is preserved. If the organisms themselves are preserved, it is usually the result of secondary silicification of the stromatolite. Unlike later stromatolites, the Paleoarchean ones were probably produced by anoxygenic (non-oxygen producing) microorganisms, a hypothesis that is supported by isotopic studies of Archean rocks and by modern experimental work (Bosak et al., 2007). In a recent review of Archean life, 48 sites were listed that contain Archean stromatolites believed to be biogenic (Schopf et al., 2007b); a decade ago there were only a handful. A much more detailed picture is beginning to emerge of the environment during deposition of these rocks. Van Kranendonk (2006) provided evidence that the Warrawoona Group began as a continental volcanic platform. The Pilbara Supergroup (FIG. 2.13) includes autochthonous deposition of volcanics with numerous, interbedded chert layers over 565 myr. The Strelley Pool Chert (SPC) (Kelley Subgroup, Van Kranendonk et al., 2002) contains the most widespread stromatolites (FIG. 2.14). Allwood et al. (2006) described seven different morphologies (FIG. 2.15) of stromatolites

Figure 2.12 Modern stromatolite from Laguna Mormona,

Mexico, showing lamellae of cyanobacterial colonies. (Courtesy J. W. Schopf.)

PRECAMBRIAN LIFE

53

over a distance of several kilometers in the SPC (3.430 Ga). From a detailed sedimentologic analysis, these authors suggested that the stromatolites were best developed in relatively restricted parts of a peritidal carbonate platform (Allwood et al., 2007). The morphologies range from typical, laminated dome-like stromatolites to conical forms and wavy or bumpy (egg carton) types (FIG. 2.16). A biogenic origin of these stromatolites is supported by the fact that all have morphologies similar to known microbialites (microbially induced sedimentary structures) and none can be explained by abiogenic development. The diversity of forms present and their extended distribution suggest an entire ecosystem composed of microbial communities. SEDIMENTARY EVIDENCE A number of other sedimentary and organosedimentary structures are often used as evidence of past microbial life. The general term for such structures is a microbialite, which is defined as an organosedimentary structure formed by the interaction of microbial communities with sediment. This interaction can include trapping or binding sediment, or biologically mediated calcification (Burne and Moore, 1987; Riding, 2006b). Stromatolites are thus a type of microbialite—one that has a finely laminated internal structure. Thrombolites are microbialites with a clotted internal texture. Microbially induced sedimentary structures (also called MISS) also provide evidence of past life (Noffke et al., 2001). Noffke et al. (2006) describe evidence of microbial mats in the form of wrinkle structures, desiccation cracks, and roll-up structures in 3.2 Ga sandstone rocks from South Africa. Similar structures have also been described from the Mesoarchean (2.9 Ga) of South Africa (Noffke et al., 2008) and from various localities throughout the Precambrian (Simonson and Carney, 1999; Schieber, 2004). These Archean sedimentary features are comparable to those observed in similar modern environments and form via the stabilization of sediment by overlying microbial mats. The Archean structures formed in a tidal flat setting and resemble modern mats in cross section, showing a laminated structure; carbon isotope values are also consistent with a biogenic origin (Noffke et al., 2006). Cyanobacteria are the most common organisms involved in the formation of similar modern sedimentary features, but whether these Archean mats were formed by cyanobacteria, or whether they were oxygenic or anoxygenic cyanobacteria, is not yet known. If the mats were formed by oxygenic photosynthesizers, then they would represent an early source of atmospheric oxygen on Earth (Noffke et al., 2006).

54

Paleobotany: the biology and evolution of fossil plants

Cherts with putative biosignatures 3.240 Ga (Kangaroo Caves Fm.)

V

V

Gorge Creek Group

Sulfur Springs Group

V 3.325–3.315 Ga (Wyman Fm.)

V V

V

V

V

3.346 Ga

V

V

V

Kelly Group

V

V V

V

V

3.35 Ga (Euro Basalt) Strelley Pool Chert “Kitty’s Gap” chert

V

3.458–3.426 Ga (Panorama Fm.)

V V

V

V Apex chert

Warrawoona Group

3.471–3.463 Ga (Duffer Fm.)

V

V V

V

Key

V

V

V

V

Shale 3.477 Ga (McPhee Fm.)

Sandstone/conglomerate

Dresser Fm.

V

Banded iron formation

V

V

3.496 Ga

V

Chert

V V V

Felsic volcanic rocks

V 3.508 Ga

V

V 3.515 Ga (Coucal Fm.)

V

V V V



Basaltic/komatiitic rocks

 Granitoid rocks 

V

V

V

V Basalt



U-Pb or Pb-Pb age (formation dated)

Figure 2.13 Generalized stratigraphic column of the Pilbara Supergroup showing position of various putative fossil-bearing chert formations. (From Allwood et al., 2007.)

MESOARCHEAN–NEOARCHEAN LIFE The record of life on Earth becomes more widespread and more diverse during the Mesoarchean (3.2–2.8 Ga) and

Neoarchean (2.8–2.5 Ga). Filamentous microorganisms have been described from rocks of the Fortescue Group in Western Australia (2.768 Ga) (Schopf and Walter, 1982). Although poorly preserved, these appear as two types of

CHAPTER 2

PRECAMBRIAN LIFE

55

Beukes and Lowe (1989) reported on 3 Ga stromatolites from South Africa (Pongola Supergroup) that they believe were formed in a range of shallow depositional settings, including intertidal mud flats and subtidal channel environments. Noffke et al. (2008) reported microbially influenced sedimentary structures (MISS) from ancient tidal flat facies of the 2.9 Ga Pongola Supergroup, South Africa, which are similar to the Paleoarchean ones described earlier (Noffke et al., 2006). They delimit four levels of microbial mats, which can be distinguished by their specific MISS, and which also occur in modern tidal flat settings. Although the analogous modern features are formed by benthic cyanobacteria, what types of microorganisms formed the fossil features remain unknown.

CONCLUSIONS: ARCHEAN LIFE

Figure 2.14 Stromatolite from SPC showing concentric lami-

nations (Pilbara Supergroup). Bar  20 cm. (Courtesy J. W. Schopf.)

filaments: narrow threads (1 μm wide) consisting of diskshaped cells and larger filaments (10 μm in diameter) made up of barrel-shaped cells enclosed by a multilamellated sheath. From 3.2-billion-year-old Mesoarchean rocks of the Cleaverville Group, Pilbara Craton, Kiyokawa et al. (2006) described an extensive stromatolitic layer, which extends for 1 km and contains wavy, colloform laminations similar to those seen in columnar stromatolites. Within the same carbonaceous Black Chert Member, a diverse microbiota is preserved in silica, including six morphological types: (1) spiraled structures 50–150 μm long and 10 μm wide; (2) rodshaped filaments 50–80 μm long, with cell walls 2–5 μm thick; (3) delicate, short filaments, which are compared to Primaevifilum from the Apex chert; (4) branching, dendritic filaments (FIG. 2.17) 100 μm long and 1 μm in diameter, which appear to be made up of small rod-shaped units; (5) carbonaceous, mat-like material that may represent part of a biofilm; and (6) spherical masses of carbonaceous material. The putative microbial fossils are syngenetic, that is, they were there at the time the rock formed and are preserved in three dimensions in the fine chert.

Despite the fact that there are doubters for almost all of the evidence for life in the Archean, evidence for early life has expanded tremendously in the past 20 years, from body fossils to isotopic and geochemical studies. These data are supported in most cases by detailed stratigraphic and sedimentologic evidence of the environments of early life, and by knowledge of the depositional and metamorphic history of the rocks that contain ancient carbonaceous matter. Nisbet (2000) provided an excellent summary of the variety of environments for Archean life, including deep-sea hydrothermal vents, open ocean, lacustrine environments, hydrothermal sites around active volcanoes, and anywhere that microbial mats occur today, for example, in coastal sediments (FIG. 2.18). Perhaps more than in any other area of paleobiological research, the multidisciplinary approach has been used to address questions and provide answers relating to the earliest life on Earth. The advances made in phylogenetic and evolutionary microbiology have also contributed to an increased understanding of early life. For example, based on studies of modern taxa and on geochemical studies, it now appears likely that the first organisms were not photosynthetic cyanobacteria, but perhaps chemosynthetic organisms (e.g., chemolithotrophs) that lived around deep-sea vents, or anoxygenic photosynthesizers which could live closer to the surface (Nisbet and Fowler, 1999). Both Archaea and Eubacteria are known to live in these habitats today, and there is isotopic evidence that points to the presence of sulfate-reducing bacteria and methanogenic Archaea around 3.5 Ga. The knowledge that earliest life did not necessarily have to arise in a “warm little pond” has, in some ways,

56

Paleobotany: the biology and evolution of fossil plants

Encrusting/domical laminites Large complex cones

Egg carton laminite

Cuspate swales

Iron-rich laminite

Small crested/conical laminite

Wavy laminite

Figure 2.15 Synoptic profiles of seven stromatolite facies from the SPC. (From Allwood et al., 2006.)

Figure 2.16 Example of bumpy stromatolite morphology

Figure 2.17 Dendritic carbonaceous material (Cleaverville

(Pilbara Supergroup). Bar  10 cm. (Courtesy J. W. Schopf.)

Group). Bar  100 μm. (From Kiyokawa et al., 2006.)

CHAPTER 2

Hydrothermal communities around andesite volcanoes

Lake communities

Coastal sediment S-microbial mats

Light

Mid-ocean ridge chemotrophic community

Hydrothermal systems around komatiite shields

Mg, SO4

Light

Light Organic debris S cycle and methanogens

Light

PRECAMBRIAN LIFE

Fe, Mn, S, CH4,H2

57

Ni, Co, Fe, S, Mg

Cu, Mo Zn, S

Light

P cya lankto n o b n ic act e r ia

Stromatolites Lava

Hydrothermal supply of metals and reductant in deeper water

Open ocean

Light

More oxidized More reduced

Figure 2.18 Diagrammatic representation showing numerous sites where early life may have flourished. (From Nisbet, 2000.)

revolutionized the picture of early life on Earth. Nisbet and Fowler (1999) noted that anaerobic photosynthesizers, similar to modern green gliding bacteria, could have colonized shallow-water habitats. Bacterial sulfur metabolizers would have colonized anaerobic sites, perhaps at the bottom of microbial mats, where they would live off of decaying matter from the autotrophic organisms above them. Organisms similar to modern purple bacteria (Proteobacteria), which are anoxygenic photosynthesizers, would occur in more microaerophilic areas within the same microbial mat community, perhaps similar to some of those described by Allwood et al. (2006) from 3.4 Ga rocks. Cyanobacteria would be present on the top surface of such a mat. Similar microstrata occur in modern microbial mats. As oxygenic cyanobacteria diversified and their biomass and distribution increased, the world began to change.

OXYGENATION OF THE EARTH (2.45–2.2 Ga) Around the Archean–Proterozoic boundary, Earth began a transition phase in a number of different areas. The two supercontinents that came together in the Neoarchean began to break apart in the Paleoproterozoic, and a major glaciation (the Huronian) occurred from 2.45 to 2.22 Ga. There were large eruptions of flood basalts and an increase in

atmospheric oxygen, eventually leading to the oxygen-rich environment we live in today (Melezhik et al., 2005). Some researchers have suggested that the rise in O2 triggered the glaciation, as methane in the atmosphere was oxidized to CO2, thereby reducing greenhouse gases (Kopp et al., 2005). Banded iron formations (BIFs) also reached their greatest extent during this time. As a result of all of these interrelated events, especially the rise of free oxygen, large transformations in biogeochemical cycles must also have taken place (Konhauser, 2007). Although the upper parts of the ocean gradually became enriched with oxygen, it is hypothesized that the deeper parts of the ocean remained anaerobic and were rich in hydrogen sulfide (H2S). The origin of oxygenic photosynthesizers (those that produce oxygen) represents an important benchmark in biological evolution (Schopf et al., 1983). Among the prokaryotes, cyanobacteria are the major organisms that fill this ecologic role. They were no doubt present in the Archean, but perhaps not in sufficient numbers to make a difference on a global scale. In the Paleoproterozoic, however, this condition began to change and the increase in numbers of these organisms and their distribution on Earth eventually resulted in an atmosphere enriched in oxygen. The presence of free atmospheric oxygen subsequently resulted in the ozone layer, thus making it possible for organisms to live on the land without being destroyed by ultraviolet radiation (Chapter 6). The presence of adequate oxygen was also necessary for

58

Paleobotany: the biology and evolution of fossil plants

the evolution of multicellular animals (Berner et al., 2007). The evolution of cyanobacteria thus signaled a revolution in life on Earth. These organisms were not only able to utilize the most abundant form of energy on Earth—sunlight—to split water molecules, but were able to do it more efficiently. Since cyanobacteria utilize both photosystems I and II (as do all green plants), their appearance in the fossil record also signals the evolution of an advanced biochemical system compared to the single system in other photosynthetic bacteria. It has also been hypothesized that the appearance of cyanobacteria and oxygenic photosynthesis resulted in a rapid colonization of shallow-water habitats on Earth (Nisbet and Fowler, 1999). There has been a continued controversy over the timing of the first appearance of oxygen-producing photosynthesizers (Towe, 1990), that is, cyanobacteria. Hofmann and Schopf (1983) suggested that some of the organisms found in Paleoproterozoic biotas represented aerobic photoautotrophs, cyanobacteria, based on the presence of morphologies comparable to extant cyanobacteria. Stromatolitic evidence of early photosynthesizers comes from Neoarchean (2.7 Ga) lacustrine deposits from Western Australia (Fortescue Group) (Buick, 1992). In this study, a combination of morphologic, geochemical, and sedimentologic data were used to propose that the stromatolites were probably formed by filamentous, phototropic bacteria. Based on the complex trophic system in these lakes, Buick concluded that O2-producing photosynthesizers must have been present, since a deficiency of sulfates in the system would have excluded anaerobic photosynthesis. Summons et al. (1999) reported the occurrence of 2-methylhopanoids or their derivatives, which represent a biomarker for cyanobacterial photosynthesis, in rocks as old as 2.5 Ga. They also noted that the abundance of these biomarkers at this time supports microfossil evidence that cyanobacteria arose before 2.5 Ga (Brocks et al., 2003). In the past, the occurrence of BIFs (FIG. 2.19) in Proterozoic rocks has been used as evidence of early oxygenproducing photosynthesis. BIFs are laminated units containing iron oxides, sometimes in very fine laminae, separated by silica-rich layers. Some of the layers are laterally quite extensive, for example, a layer just a few centimeters thick can be followed over 50,000 km2 in the Hamersley Basin of Western Australia. BIFs were widespread from 2.2 to 1.9 Ga and rare afterwards. It was once thought that they formed abiotically, by the oxidation of ferrous iron in the presence of free oxygen. More recent work, however, suggests that only biotic interactions could have precipitated such large quantities of iron (Konhauser et al., 2002). The source of the dissolved ferrous iron is thought to be hydrothermal activity at

Figure 2.19 Banded iron formation (Negaunee Formation). (Courtesy J. W. Schopf.)

mid-ocean ridges. Several studies have suggested that ironoxidizing bacteria (chemolithotrophic organisms) could have precipitated the amount of iron present in the Hamersley Group of Western Australia, even at cell densities lower than those seen in modern coastal waters (Konhauser et al., 2002; Kappler et al., 2005). Thus, it can no longer be assumed that the presence of BIFs represents proxy evidence for early oxygenic photosynthesizers. The distribution of other types of rocks, however, does appear to be important in dating the origin of an oxidizing atmosphere. Detrital uraninite, siderite, and pyrite occur in 3.25–2.75 Ga fluvial deposits from Australia, but are extremely rare after 2 Ga (Rasmussen and Buick, 1999). There is evidence that these minerals had been transported in water that was well aerated. All of these compounds are known to be unstable under oxidizing conditions (J. Walker et al., 1983). In addition, red beds are known to be rare or absent in the Archean. Red beds obtain their characteristic color from the presence of ferric oxides, which form as a result of subaerial oxidation, usually in fluvial rock sequences. Despite the occurrence of suitable rock sequences on the early Earth, red beds are far more abundant in the Proterozoic than in the Archean, and this change is generally attributed to an increase in atmospheric oxygen around 2.3 Ga. Another method to extrapolate paleoatmospheric O2 levels involves the fractionation of sulfur isotopes (Pavlov and Kasting, 2002). Using this method, Bekker et al. (2004) suggested that the atmosphere already included pO2 levels 105 of the present atmospheric oxygen levels (PAL) by 2.32 Ga. The timing for the disappearance of uraninite deposits, appearance of red beds, decline of BIFs, and shift in sulfur isotopes approximately coincides at 2.3–2.2 Ga,

CHAPTER 2

and this time no doubt represents evidence of global atmospheric oxygen, even though the levels were lower than today. The only viable source for this oxygen is the product of photosynthesis by cyanobacteria, since they represent the earliest oxygen-producing photosynthetic organisms in the fossil record. Oxygenic photosynthesis certainly must have evolved earlier than this time, however, in order to build up enough oxygen in the atmosphere and ocean to leave a chemical signature in the various rocks. Tomitani et al. (2006) used molecular phylogeny of extant cyanobacteria, within the context of geochemical evidence, to suggest that oxygen levels rose rapidly between 2.3 and 1.9 Ga. These results coincide with the disappearance of BIFs and suggest that oxygen levels were probably 10% PAL by 1.9 Ga and 100% PAL by the Neoproterozoic. Fossil evidence for these oxygen levels, however, is indirect. Some cyanobacteria form specialized cells called heterocysts, where nitrogen fixation occurs in an anaerobic environment. The evolution of heterocysts is thought to have occurred in response to increasing oxygen levels, but the only Precambrian evidence of heterocysts is indirect. Although heterocysts have been reported from the Gunflint Chert (discussed later), these structures could also be interpreted as the result of diagenesis. Golubic et al. (1995) proposed that the Mesoproterozoic rod-shaped microfossil Archaeoellipsoides represents the akinetes of a cyst-forming cyanobacterium and attribute this fossil to the Nostocales, a group of heterocystous cyanobacteria with living members (Chapter 3). Based on the monophyly of the extant heterocystous forms (Tomitani et al., 2006), Archaeoellipsoides is hypothesized to be the earliest evidence of heterocyst-forming cyanobacteria and used as a benchmark for paleo-oxygen levels (Tomitani et al., 2006). More recent evidence, however, suggests that atmospheric oxygen was present by 2.5 Ga, in the latest Archean. Both sulfur isotopes (Kaufman et al., 2007) and the metals molybdenum and rhenium (Anbar et al., 2007) from the Mount MacRae Shale in Western Australia show a shift in isotopic values at 2.5 Ga years. These authors interpret these shifts to show the presence of oxygen, which correlates with previous data from carbon isotopes. Kaufman et al. noted that equivalent strata in South Africa show the same shift in sulfur isotopes, suggesting that widespread oxygenation of the ocean was present around 2.5 Ga, prior to oxygenation of the atmosphere at 2.45 Ga. The time period for evolution of oxygenic photosynthesis is further constrained by the absence of evidence for oxygen in a 2.7 Ga paleosol (Yang et al., 2002). Brocks et al. (1999) reported on lipids characteristic of cyanobacterial photosynthesis in Neoarchean rocks (2.7–2.6 Ga) from the lowermost Hamersley Group

PRECAMBRIAN LIFE

59

(Pilbara Craton). Additional data from carbon isotopes also support widespread oxygenic photosynthesis in the ocean before the land (Eigenbrode and Freeman, 2006).

PROTEROZOIC LIFE If the Archean was a prokaryotic world, the Proterozoic was a time of transition to a eukaryotic and eventually a multicellular world. Microfossils of eukaryotes appeared around the Paleoproterozoic–Mesoproterozoic boundary, and by the late Mesoproterozoic—early Neoproterozoic, the major algal clades were present. Eukaryotes (based on acritarch diversity) appear to have gradually increased in diversity until the midNeoproterozoic (700 Ma), the time of the first (Sturtian) Neoproterozoic glaciation. After the second glaciation (630 Ma), their diversity increased rapidly until the latest Neoproterozoic, when they underwent an extinction event. PALEOPROTEROZOIC

There is abundant evidence of diverse life forms in the Proterozoic. In their review of Paleoproterozoic microfossils, Hofmann and Schopf (1983) listed 122 taxa (including 23 dubiofossils and pseudofossils) in 40 different biotas (2.5–1.6 Ga); this number does not include microorganisms that were not named. They classify these genera into five morphological categories: (1) coccoid unicells; (2) septate, unbranched, filaments; (3) tubular, unbranched forms; (4) branched filaments; and (5) bizarre or unusual forms, that is, those with unusual morphologies and uncertain affinities. Coccoid forms dominate most Paleoproterozoic assemblages, both in terms of diversity and abundance. In contrast to later Proterozoic and early Paleozoic forms, they generally exhibit a simple and unornamented morphology and are relatively small (25 μm; most are 2–7 μm). Included in this classification would be coccoid forms that occur in colonies which are usually surrounded by a sheath-like structure. Septate filamentous forms are also usually small and simple, ranging from 1 to 2.5 μm in diameter and generally not surrounded by a sheath. Tubular microfossil forms are less commonly found and many appear to represent the remains of microbial sheaths, similar to those in modern filamentous cyanobacteria. Branched filaments and unusual forms are relatively rare in most biotas. Based on their size and simple organization, Hofmann and Schopf (1983) considered all of these forms as representing prokaryotes. Thus, it would appear that Paleoproterozoic biotas were composed for the most part, if not entirely, of primitive prokaryotic organisms.

60

Paleobotany: the biology and evolution of fossil plants

Figure 2.20 Stromatolites from the Gunflint Formation.

(Courtesy J. W. Schopf.)

Figure 2.22 Filaments and spheres from the Gunflint Form-

ation. Bar  10 μm. (Courtesy J. W. Schopf.)

Figure 2.21 Thin section from the Gunflint Formation showing stromatolitic laminae. (Courtesy J. W. Schopf.) Figure 2.23 Eosphaera tyleri (Gunflint Formation). Bar 

One of the most extensively studied Paleoproterozoic biotas comes from the Gunflint Formation of southern Ontario (1.9 Ga) and, as noted earlier, it includes the first well-preserved Precambrian organisms described (Tyler and Barghoorn, 1954; Barghoorn and Tyler, 1965). See also Cloud (1965), which helped to give credibility to this work at a time when most believed that the Precambrian was devoid of life. The organisms from this formation are structurally preserved in stromatolitic (FIGS. 2.20, 2.21) and non-stromatolitic cherts and can be divided into four basic morphologic types: coccoid forms, (Huroniospora) (FIG. 2.22); septate, filamentous forms (Gunflintia); tubular,

10 μm. (Courtesy J. W. Schopf.)

unbranched forms (Animikiea); and unusual or bizarre forms (Archaeorestis, Kakabekia, Eoastrion, Eosphaera) (FIG. 2.23). Huroniospora is a simple, spherical—ellipsoidal form originally described by Barghoorn, that has since been found at several localities with budlike outgrowths attached to the cells (Barghoorn and Tyler, 1965). Gunflintia is a narrow filament (1–4 μm in diameter) composed of a single row of cylindrical cells, whereas Animikiea is a broader (6–12 μm) tube that sometimes contains a septate filament.

CHAPTER 2

PRECAMBRIAN LIFE

61

Figure 2.25 Kakabekia umbellata (Gunflint Formation). Bar 

5 μm. (Courtesy J. W. Schopf.)

Figure 2.24 Archaeorestis sp. (Gunflint Formation). Bar 

10 μm. (Courtesy J. W. Schopf.)

Some of the most interesting forms in the Gunflint microbiota are those that have uncertain taxonomic affinities. Eoastrion is the name given to a star-shaped group of radiating, filamentous structures considered by most to represent some type of bacterium. Archaeorestis is a non-septate, irregularly branched organism with filaments that range from 2 to 10 μm in diameter and up to 200 μm long (FIG. 2.24). Its affinities are uncertain, but it has been interpreted as a budding bacterium (Awramik and Barghoorn, 1977) and a possible Kakabekia by Hofmann in Hofmann and Schopf (1983). Kakabekia umbellata exhibits a tripartite organization, consisting of a bulb-like base, a so-called stipe, and an umbrellalike crown (FIG. 2.25); some specimens extend up to 30 μm in length. The affinities of these fossils continue to remain obscure. Some of the structures in the Gunflint chert may represent taxa that are not related to any known forms, and it is difficult to assign these fossils to any group of organisms

with certainty. Although there is noticeable morphological similarity between some of the fossil organisms and certain modern bacteria, it would be almost impossible to categorize many of these fossil organisms without a knowledge of their biochemistry and molecular biology. Information on cell morphology alone is not enough to make classification possible. Overall, the types of organisms present in Paleoproterozoic biotas are similar in their morphology and occurrence to both older (Neoarchean) and younger (Mesoproterozoic) types. They differ from those of the Archean by their greater diversity and by the presence of both benthic and planktonic forms. What can be generalized is that there is a trend toward increasing diversity and increasing cell and filament size throughout the Proterozoic. ORIGIN OF EUKARYOTES

Almost 40 years after Schopf (1968) originally reported presumed green algae from the Bitter Springs Formation in the Amadeus Basin (Australia), it is difficult to appreciate the controversy this discovery caused at the time. Originally thought to be 1–0.9 Ga in age, the Bitter Springs is now considered to be 830–800 Ma and is discussed in more detail later (see Neoproterozoic). The deposit includes fossilized microbial mats with microfossils preserved in chert. The microfossils Caryosphaeroides and Glenobotrydion

62

Paleobotany: the biology and evolution of fossil plants

are spherical cells with opaque material inside the cell walls (Schopf, 1968; Schopf and Oehler, 1976), and it was suggested that these microfossils could be assigned to the Chlorophyta (green algae; Chapter 4). Cells of C. pristina average 13 μm in diameter and those of Glenobotrydion, 9 μm; by comparison, the filamentous and spherical bacterial cells in this same deposit are generally 10 μm in diameter. Both of these organisms contained opaque material, which was interpreted as the remains of nuclei in Caryosphaeroides and the remains of a pyrenoid-like body in Glenobotrydion (Oehler, 1977). Knoll and Barghoorn (1975), however, examined stale cultures of living cyanobacteria (Chroococcus) and concluded that all of the fossil intracellular objects could be explained as artifacts resulting from the coagulation of cytoplasm in a prokaryotic cell. Despite the fact that interpretation of these structures remains controversial, there is no doubt that the presence of intracellular material in Precambrian cells, whether cytoplasm or nuclei, represents a remarkable case of fossil preservation. Eotetrahedrion is another important (and controversial) fossil from the Bitter Springs biota; it consists of a tetrahedral tetrad of spherical cells surrounded by a sheath-like structure (FIG. 2.26) (Schopf and Blacic, 1971). Cells average 9.4 μm in diameter and many include a triradiate mark on the surface. What is especially interesting is that this type of mark is commonly seen on spores of land plants after meiosis has occurred (e.g., see Chapters 5, 6). Such spores are produced during sexual reproduction by the process of reduction division, in which a single cell (spore mother cell) with a diploid complement of chromosomes (referred to as 2n) divides to produce four identical products (spores), each with one half the chromosome complement of the mother cell (n). Many of these spores are produced in a tetrahedral tetrad (4 spores), and the triradiate or trilete mark on the spores denotes the contact face where each spore in the tetrad is in contact with the other three. The presence of a triradiate mark on spores is generally assumed to represent

evidence of meiosis. Thus, it was initially suggested that the marks on Eotetrahedrion indicated the occurrence of meiosis and, therefore, of sexual reproduction during the Proterozoic. Some green algae and cyanobacteria, however, are known to produce mitotically derived tetrahedral tetrads of cells or spores, so this arrangement of cells cannot necessarily be considered definitive evidence of meiosis or sexual reproduction in the fossil record (see Chapter 6). Work completed since Schopf’s (1968), and Schopf and Blacic’s (1971) research on the Bitter Springs chert has demonstrated that eukaryotes had definitely evolved by 830–800 Ma; in fact, they were present far earlier. Cloud (1976) suggested that the earliest eukaryotic cells occur in the 1.3 Ga Beck Spring Dolomite of California, based on the occurrence of filamentous forms resembling siphonaceous green algae and unicells that range from 40 to 62 μm in diameter (Cloud et al., 1969). Similar to the cells described by Schopf from the Bitter Springs microbiota, nearly all of these forms include a dark body within the cell. Because of the diversity of forms present in this biota, Cloud (1983) suggested that the eukaryotes may have arisen anywhere from 2 to 1.3 Ga. Further support for this assumption is the report by Schopf (1977) of an increase in cell size of both filaments and unicells starting at 1.4 Ga. More recently, several authors have suggested specific criteria for identifying early eukaryotes, some of which are similar to standards used to identify the earliest life. Cells presumed to be eukaryotes must have a morphology comparable to known organisms, and this morphology should have a definite range of variability. Knoll et al. (2006b) suggested three distinct criteria for recognizing early eukaryotes, but in most cases these will only apply to acritarchs. Acritarchs (Chapter 4) are generally unicellular microfossils, many of unknown affinity, although some have been attributed to planktonic, cyst-forming algae; the vast majority are assumed to represent eukaryotes (Martin, 1993). Many acritarchs have a complex ornamentation consisting of elongate processes on

Figure 2.26 Different focal planes of Eotetrahedrion sp. (Bitter Springs Formation). Bar  10 μm. (Courtesy J. W. Schopf.)

CHAPTER 2

the surface of the cell wall (FIG. 2.27). Knoll et al. (2006b) proposed that to be definitively classified as eukaryotes, microfossils must (1) be large, (2) have a preservable wall, and (3) have processes (elaborations of the external wall). Prokaryotic cells, with a few notable exceptions, are generally 10 μm in diameter, whereas eukaryotic cells can range from 10 μm to hundreds of microns in diameter. Javaux et al. (2001, 2003) described well-preserved microfossils of Tappania and other acritarchs from coastal facies of the early Mesoproterozoic (1.5 Ga) Roper Formation of north central Australia. Various acritarchs are found in a range of rocks representing marginal marine to basinal settings; the diversity and abundance of eukaryotic fossils decreases from onshore to offshore. Tappania extends up to 160 μm in diameter and is characterized by hollow, cylindrical processes with expanded tips (FIG. 3.9). Z. Zhang (1997) described sphaeromorphic acritarchs from the Changzhougou Formation in northern China (1.85 Ga). Although somewhat poorly preserved, the assemblage contained a large number of smooth-walled acritarchs whose diameters clustered around 60 μm and extended up to 238 μm. From the Meso-Neoproterozoic Ruyang Group (1150–950 Ma), L. Yin (1997) described two acanthomorphic acritarchs, Tappania and Shuiyousphaeridium. The latter is spherical, from 110 to 250 μm in diameter and bears elongate processes on the exterior. Tappania is vase shaped and small, 45–60 μm long. Meng et al. (2005) suggested that Shuiyousphaeridium represents an early dinoflagellate, and support this conclusion

Figure 2.27 Acritarch (Trachyhystrichospahaea) from Siberia

Russia (Lakhanda Formation). Bar  50 μm. (Courtesy J. W. Schopf.)

PRECAMBRIAN LIFE

63

with evidence of the biomarker, dinosterane, in the same rocks as the fossils in the Mesoproterozoic Beidajian Formation (see also Chapter 4). Other evidence, however, suggests that eukaryotes may have arisen much earlier. Brocks et al. (1999) found steranes in the same Neoarchean rocks (2.7–2.6 Ga) from which they extracted cyanobacterial biomarkers as evidence of oxygenic photosynthesis. Steranes are only produced by eukaryotes, but Summons et al. (2006) have shown that previous reports of steranes in cyanobacteria can be attributed to eukaryote contamination of cultures. In addition, sterane biosynthesis requires oxygen, so this research suggests that both eukaryotes and oxygenic photosynthesis were present 2.6 Ga (Brocks et al., 1999). To date, no generally accepted microfossil evidence has been found to corroborate these geochemical hypotheses. Several authors have suggested that the spiraled compression fossil, Grypania, may represent the earliest eukaryote. Han and Runnegar (1992) described ribbon-shaped, spiraled carbonaceous films from the Negaunee Iron Formation (2.1 Ga) as Grypania, and noted that the thallus could reach 0.5 m in length, although it was only 0.5 mm in diameter. Samuelsson and Butterfield (2001) suggested that the Grypania specimens from the Negaunee Iron Formation are unlike those from younger rocks (which are clearly eukaryotic) and should not be considered the same organism. A number of similar carbonaceous films have been described from Proterozoic rocks, and many of these have been shown to be pseudofossils (Lamb et al., 2007). Possible evidence for the origin of eukaryotic organisms around 1 Ga comes from reports of similar megascopic fossils from various localities around the world (Walter et al., 1976; Hofmann, 1985b). It has been suggested that some of these megascopic fossils represent eukaryotic algae. Chuaria is one of the most widespread genera of this type and has been described from the Little Dal Group in northwest Canada (1080–780 Ma) as black, circular compressions ranging in size from microscopic up to 4.6 mm in diameter (Hofmann and Aitken, 1979). Concentric wrinkles on the surface are present on many specimens. Chuaria occurs worldwide and was originally thought to be restricted to rocks 1.1–0.6 Ga, where it was used as an index fossil. In the Little Dal Group, Chuaria occurs with ribbonlike, often curved, sometimes tubular compressions several centimeters long that are given the name Tawuia. This biota also includes filamentous bacteria, cyanobacterial sheaths, and possible acritarchs. Chuaria was originally interpreted as a planktonic eukaryotic alga and Tawuia as a probable alga, although metazoan affinities could not be completely discounted for

64

Paleobotany: the biology and evolution of fossil plants

the latter. Better-preserved specimens of both taxa from northern China (Sun, 1987) suggest that at least some of these compressions may represent prokaryotic aggregations. These fossils were studied by means of cellulose acetate peels and bioplastic transfers of the rock surface, techniques which revealed that both taxa represent colonies of filamentous cyanobacteria; Sun compares them to the living genus Nostoc. This interpretation helps to explain the variable size and morphology of these unusual remains. Other authors, however, regard Chuaria and Tawuia as developmental stages of the same organism, which is thought to be some type of benthic, tubular macroalgae (Xiao and Dong, 2006). Similar fossils from Paleoproterozoic rocks were described as seaweed-like algae (W. Zhu and Chen, 1995; Yan and Liu, 1997), but these specimens are now considered to have uncertain affinities or represent aggregations of prokaryotes (Butterfield, 2000; Knoll et al., 2006b). Grazhdankin and Gerdes (2007) examined discoidal compressions made up of concentric rings. These had previously been identified as metazoans, but these authors compare them to ring structures in modern microbial mats and conclude that they are formed by either bacteria, fungi, or protists. The earliest fossil evidence for eukaryotic life occurs in the Paleoproterozoic in the form of acritarchs. Although there is geochemical evidence in the latest Archean, there is yet no widely accepted fossil evidence to support this finding. There is excellent evidence, however, that eukaryotes diversified into multicellular forms in the Mesoproterozoic. MESOPROTEROZOIC

During the Mesoproterozoic (1.6–1 Ga), the supercontinent Rodinia was formed (around 1.1 Ga), atmospheric oxygen levels continued to rise, and stromatolites formed large and widespread reefs. Unicellular, eukaryotic organisms probably arose either in the late Paleoproterozoic or early Mesoproterozoic, and continued to diversify throughout this time period. Coccoid microfossils show an increase in size during the Mesoproterozoic and biotas exhibit an increasingly diverse assemblage of cyanobacterial remains, both filamentous and coccoid. As noted earlier, cyst-forming, planktonic, eukaryotic algae (acritarchs) are believed to have emerged in the Mesoproterozoic, 1.4 Ga ago (Knoll, 1985a), with the earliest evidence for multicellularity not long thereafter. Oehler (1977) described a deep-water biota from the Mesoproterozoic of Australia. The assemblage was dominated by filamentous bacteria, unlike those known from shallow-water deposits or stromatolitic assemblages. Oehler suggested that these organisms lived below the photic zone under anoxic conditions.

EARLIEST MULTICELLULAR LIFE Many textbooks, journal papers, and online sites identify the origin of multicellular life as the famous 610 Ma Ediacaran biota—an assemblage of strange, soft-bodied metazoans that are preserved in only a limited number of sites around the world. These reports tend to forget the fact that irrefutable evidence for multicellular algae existed long before the Ediacaran. In 1990, Butterfield et al. described a permineralized red alga similar to living members of the Bangiophyceae (Rhodophyta) from the Hunting Formation of arctic Canada. This formation consists of shallow-water carbonates and includes stromatolites, as well as cyanobacterial fossils; the rocks that contain the algae are dated at 1200 Ma (late Mesoproterozoic) (Butterfield, 2000). The fossils of Bangiomorpha pubescens (FIGS. 4.43, 4.44) consist of filaments made up of stacked, discoidal cells enclosed within a sheath; many have a holdfast at their base. The filaments, which are up to 2 mm long and range from 15 to 45 μm in diameter, show radial cell division within the filament, a feature characteristic of bangiophytes. As Butterfield (2000) noted in the formal description of B. pubescens, the fossil is so similar to the modern red alga Bangia as to be “indistinguishable.” Based on the different morphology of filaments and circular objects within them, Butterfield suggests that B. pubescens fossils also show the earliest fossil evidence of sexual reproduction. The shallow-water carbonates that include these red algae contain a total of four fossil assemblages, three of which are dominated by mat-forming, photosynthetic prokaryotes (Butterfield, 2001). Each assemblage appears to occupy a specific niche within a larger ecosystem. In the Bangiomorpha assemblage, however, matforming prokaryotes are excluded, perhaps because of the vertical growth of the algae. Butterfield (2001) suggested that this biota documents the oldest case of competitive exclusion of prokaryotes by multicellular eukaryotes. He hypothesized that such exclusion may explain the relatively rapid, even explosive, diversification of eukaryotes in the Neoproterozoic. NEOPROTEROZOIC

A greater diversity of eukaryotic algae begins to appear around the Mesoproterozoic—Neoproterozoic boundary (1 Ga), and the divergence of the major clades of eukaryotes is also believed to have occurred about this time (Knoll et al., 2006b). There is an increasing diversity of ornamented acritarchs (Butterfield, 2000). Although the diversity of taxa and number of clades increase, it is not until the middle Neoproterozoic that the primary radiation of eukaryotes occurred (Porter, 2004), so that many early Neoproterozoic biotas still contain diverse and abundant

PRECAMBRIAN LIFE

65

Figure 2.28 Cephalophytarion grande (Bitter Formation). Bar  10 μm. (Courtesy J. W. Schopf.)

Springs

CHAPTER 2

cyanobacteria, as well as other bacteria. Porter (2004) suggested that eukaryotic diversification occurred in response to selective pressures, which included the appearance of microbial predators, for example, testate amoebae (Porter and Knoll, 2000), and changes in seawater chemistry. Certainly, there were extensive environmental changes during the Neoproterozoic. The supercontinent Rodinia, which formed in the Mesoproterozoic, broke up in the early–middle Neoproterozoic, and two major glaciations, the Sturtian (750–700 Ma) and the Marinoan (635–624 Ma; the so-called snowball Earth glaciation), occurred in the Neoproterozoic (Bodiseltisch et al., 2005). Canfield (1998) proposed that the deep ocean did not become oxygenated until 1–0.54 Ga, based on sulfur isotopes. However, Butterfield (2004) suggested that there has been taxonomic inflation of eukaryotic taxa in the Proterozoic and that eukaryotes evolved at a much slower rate in the Proterozoic than they did in the Phanerozoic. A number of Neoproterozoic fossils can definitely be attributed to eukaryotic clades. Butterfield (2004) described fossil filaments of Jacutianema and attributed them to the Vaucheriales, an order within the Xanthophyceae (Chapter 4), or yellow-green algae. The fossils come from the middle Neoproterozoic, Svanbergfjellet Formation of Spitsbergen and are attributed to the Vaucheriales based on unique constrictions within the filaments. This discovery gives additional credence to an earlier report of a vaucheriacean, Palaeovaucheria, from the 1 Ga Lakhanda Formation of Siberia (Hermann, 1981). Butterfield and Rainbird (1998) described a diverse acritarch biota from the 1077–723 Ma Wynniatt Formation, arctic Canada. Three of the taxa resemble dinoflagellate cysts and may represent an early microfossil record of this group.

2.29 Oscillatoria amena (Extant). Bar  10 μm. (Courtesy J. W. Schopf.)

Figure

Figure 2.30 Filiconstrictosus cf. extant Oscillatoria amena (Bitter Springs Formation). Bar  10 μm. (Courtesy J. W. Schopf.)

Figure 2.31 Oscillatoriopsis sp. (Bitter Springs Formation).

BITTER SPRINGS BIOTA One of the better-known Proterozoic biotas comes from the Bitter Springs Formation of central Australia (Schopf, 1968; Schopf and Blacic, 1971; Oehler, 1976). This diverse biota (830–800 Ma; A. Hill, 2005) includes an abundance of filaments (FIGS. 2.28, 2.30, 2.31) and spherical unicells. Some filaments have been given names such as Palaeolyngbya, Oscillatoriopsis (FIG. 2.31), and Palaeoanacystis, and bear a striking morphologic resemblance to various extant cyanobacteria (FIG. 2.29). Palaeolyngbya is a filament with evenly spaced cross walls that are surrounded by a sheath up to 1 μm thick. Cells toward the center of the filament are usually rectangular, 2.6–3.1 μm  8.3–10.9 μm in diameter. The terminal cell of the filament is characteristically rounded. Morphologically, Palaeolyngbya is closely comparable to modern Lyngbya.

Bar  10 μm. (Courtesy J. W. Schopf.)

Some filaments show a narrowing of cells at the extremities of the filaments so that their tips appear pointed. Most of the filaments are 5 μm in diameter, although some reach up to 15 μm. Present-day cyanobacteria may be just as slender, but a proportionately larger number are more robust. The diversity of the Bitter Springs biota is illustrated by the fact that five families of extant cyanobacteria representing two major orders (Chroococcales and Nostocales) can be identified (Schopf and Blacic, 1971). Some of the cyanobacteria in the Bitter Springs Formation were compared with modern, cyst-forming filamentous forms (e.g., in the Nostocales) and these similarities were used to suggest that heterocysts had evolved by this time (Schopf

66

Paleobotany: the biology and evolution of fossil plants

Figure 2.32 Eomycetopsis robusta (Bitter Springs Formation).

Bar  10 μm. (Courtesy J. W. Schopf.)

and Blacic, 1971). As noted earlier, heterocysts are the site of nitrogen fixation in extant cyanobacteria. Since N2 fixation is inhibited by oxygen, the cyanobacterium maintains an anaerobic environment within each heterocyst; the evolution of heterocysts is presumed to indicate high oxygen levels, possibly comparable to modern levels (Tomitani et al., 2006). As originally described, the Bitter Springs biota included numerous species of cyanobacteria, both filamentous and unicellular, several eubacteria, fungus-like filaments, possible dinoflagellates (Dinophyta), and spheroidal green algae. The filaments that were originally described as fungi, Eomycetopsis (FIG. 2.32), have since been reinterpreted as cyanobacterial sheaths. STROMATOLITES Stromatolites occur beginning in the Paleoarchean, but are particularly widespread and diverse in the Proterozoic. One interesting stromatolite that appears to have a living counterpart growing in hot springs environments is the genus Conophyton. This stromatolite differs from others in that it has acute conical laminations and a distinct axial zone. Columns of Conophyton range from 5 to 10 mm in diameter. In one form from the Chichkan Formation of southern Kazakstan (650 Ma), both cyanobacteria (FIGS. 2.33, 2.34) and eukaryotic algae were identified in the silicified lamellae (Schopf and Sovietov, 1976). It has long been suggested that stromatolites reached their peak in diversity and abundance around 1250 Ma, after which they began to experience a decline (Walter and Heys, 1985). They never again reach the levels of diversity

Figure 2.33 Paleopleurocapsa reniforma (Chichkan Forma-

tion). Bar  10 μm. (Courtesy J. W. Schopf.)

Figure 2.34 Myxococcoides inornata (Chichkan Formation). Bar  10 μm. (Courtesy J. W. Schopf.)

or abundance that they exhibited during the Proterozoic, although there are reefs found in every period of the Phanerozoic which are dominated by stromatolites (Pratt, 1982). Garrett (1970) and Awramik (1971) suggested that this decline could be correlated with the rise of metazoans that grazed on or burrowed into the stromatolites. This theory would seem to be supported by the restricted occurrence of living stromatolites, either in areas of high salinity (e.g., Shark Bay, Australia) or in zones with high sediment influx (e.g., Highborne Cay, Bahamas; Andres and Reid, 2006). There is no fossil evidence, however, for metazoan grazing or damage on Proterozoic stromatolites. Awramik and Sprinkle (1999) examined changes in stromatolite taxonomic diversity for 1187 forms and found that stromatolites reached their peak 1350–1000 Ma,

CHAPTER 2

PRECAMBRIAN LIFE

67

Proterozoic peak Late Proterozoic decline

Cambrian– Early Ordovician resurgence % 100

Archaean– Proterozoic increase 200

50

Reefal microbial carbonates

Stromatolite taxa per 50 MYR interval

400

Ordovician– present-day episodic decline

0

0 3000

2500

2000

1500

1000

500

0

Age in millions of years

Figure 2.35 Distribution of stromatolite taxa through time. (From Riding, 2006a, b.)

after which they declined. Riding (2006a) combined stromatolite diversity data (FIG. 2.35) (Awramik and Sprinkle, 1999) with data on the abundance of reefal microbial carbonates (Kiessling, 2002) and metazoan diversity (Sepkoski, 1997). He found that stromatolites increased in abundance until 2250 Ma, remained at that level until 1450 Ma, and then markedly increased in abundance until 1350 Ma. After 1100 Ma, they began an irregular decline until 550 Ma, the latest Ediacaran. Stromatolites and other microbial carbonates showed an increase in the Cambrian and Early Ordovician (Riding, 2006a, b), and there is little to no correlation with marine metazoan diversity during the Phanerozoic, suggesting that metazoan grazing was not the primary reason for Proterozoic stromatolite decline. If the rise of grazing metazoans is not the primary cause of stromatolite decline, what is? Competitive exclusion by eukaryotic algae has also been implicated. Butterfield (2001) suggested competitive exclusion to explain the monotypic Bangiomorpha community in the Mesoproterozoic Hunting Formation (see above). Andres and Reid (2006) studied modern stromatolites in Highborne Cay, the Bahamas, and found that stromatolites persist in areas where sediment influx is high, because metazoans and macroalgae are excluded. If sedimentation patterns change, both metazoan borers and macroalgae colonize the substrate, outcompeting and eventually excluding the microbial stromatolites. Climate change in the Proterozoic, including two major glaciations, is also hypothesized to have contributed to stromatolite

decline. Riding’s (2006a) analysis showed that the decline itself is more complex than previously thought and cannot be explained by metazoan grazing alone. Further fossil studies are clearly needed, especially on substrates and modes of growth of stromatolites, if we are to fully understand this important Proterozoic event. OTHER MICROFOSSILS Like stromatolites, Neoproterozoic non-stromatolitic microfossils are diverse and are found in varied habitats, ranging from supratidal to open-shelf environments (Knoll, 1985a). Cyanobacteria are particularly diverse; for example, 11 of the 13 genera in the family Chroococcaceae (FIG. 2.36) are found as Proterozoic fossils. Most cyanobacteria occur in stromatolites, shallow marine muds, or tidal pools. The Entophysalidaceae are represented by mat-building organisms in intertidal zones. Pleurocapsales are an important part of tidal flat communities. Filamentous cyanobacteria are well represented as mat builders in a variety of environments. Knoll noted that Proterozoic cyanobacteria are very similar in their morphology and life-history patterns to those found in similar environments today, and that these patterns were established early. Throughout the Meso- and Neoproterozoic (2.0–1.4 Ga), intertidal zone, mat-building assemblages are dominated by Eoentophysalis-type microfossils (FIGS. 2.37, 2.38). Knoll concluded that, by the middle of the Neoproterozoic, cyanobacteria were essentially modern in their morphology and, by inference, in their physiology. In fact, most of the major evolutionary

68

Paleobotany: the biology and evolution of fossil plants

Figure 2.36 Sphaerophycus medium (Draken Formation).

Bar  50 μm. (Courtesy A. H. Knoll.)

Figure 2.37 Eoentophysalis sp. (Sukhaya Tunguska Form-

ation). Bar  10 μm. (Courtesy J. W. Schopf.)

Figure 2.38 Eoentophysalis belcherensis (Kasegalik Formation). Bar  10 μm. (Courtesy J. W. Schopf.)

changes in the cyanobacteria appear to have been established by the Neoarchean–Paleoproterozoic (Knoll, 1985b). The same cannot be said for the eukaryotes, however. Eukaryotic diversity, as measured by morphotaxa of acritarchs, especially large complex acritarchs (FIGS. 2.39, 4.75) underwent a number of changes in the Neoproterozoic. These planktonic microfossils show a gradually increasing morphologic diversity beginning in the Mesoproterozoic and extending into the mid-Neoproterozoic (1200–700 Ma) (Knoll, 1994; Porter, 2004) (FIG. 2.39), followed by a decline around the time of the Sturtian glaciation (750– 700 Ma). Toward the end of the Marinoan glaciation (625 Ma), acritarchs exhibit a rapid increase in diversity (Zang and Walter, 1989) followed by a major extinction in the latest Neoproterozoic (Vidal and Knoll, 1982; Porter, 2004). Acritarchs again underwent a diversification event in the Early Cambrian concomitant with the radiation of shelly animals. A very diverse assemblage of acritarchs in the Pertatataka Formation of central Australia (650–600 Ma) includes about 25 genera of complex acritarchs, much larger than those found in younger rocks (Zang and Walter, 1989). The lower part of the formation contains simple spheroidal acritarchs, whereas the upper part contains as many as 40 taxa of acritarchs, including many spheroidal forms, but also large, ornamented types. The most abundant taxon in the upper assemblage is Cymatiosphaeroides, which is 180 μm in diameter and characterized by numerous thin spines on the main vesicle. Zang and Walter compared these forms to younger microfossils, but noted that this assemblage is unique. Up to half of the assemblage (10–15 genera) occur only in this formation; a third represent taxa previously known from other Proterozoic assemblages, and only a few taxa are closely related to Paleozoic forms (Knoll and Butterfield, 1989). Perhaps most interesting is that these forms with Paleozoic affinities are consistently two to three times larger than their younger relatives. Zang and Walter suggested that this type of assemblage had not previously been recognized in the Ediacaran because it represents an environmental setting (offshore marine) that is poorly represented in rocks of this age. Extinction of these forms occurred in the latest Neoproterozoic. Whether or not eukaryotic diversity and extinction are related to the Neoproterozoic glaciations is still uncertain. The original idea of a “snowball” Earth included global glaciation and massive extinctions, but the microfossil record does not support this hypothesis (Corsetti et al., 2006). Corsetti et al. (2003) reported a diverse microbial biota containing two types of stromatolites, prokaryotes, and both

CHAPTER 2

Late Paleoproterozoic Early Mesoproterozoic (ca 1800–1300 Myr)

MidLate NeoprotMesoproterozoic Early erozoic Neoproterozoic (ca 850– (ca 1300–850 Myr) 720 Myr)

PRECAMBRIAN LIFE

69

Ediacaran Early (632–ca 560 Myr)

Late (560– 542 Myr)

Macroscopic Microscopic multicellular

70

Vase-shaped microfossils 60



 Acritarchs, symmetrical processes

Total taxa per assemblage

Acritarchs, asymmetrical processes 50

Ornamented acritarchs Unornamented acritarchs

40 Global glaciation 30

20

10

0

Figure 2.39 Composition and taxonomic richness of non-metazoan eukaryotic morphospecies for selected Proterozoic to Early

Cambrian fossil assemblages. (Modified from Knoll et al., 2006.)

heterotrophic and autotrophic eukaryotes preserved in chert and carbonate from the Kingston Peak Formation in the western US. The biota occurs in Sturtian glacial deposits but is comparable to the biota in underlying preglacial rocks. This is a particularly interesting deposit, in that the diversity within the biota does not suggest merely a few “disaster” taxa but rather a complex community with several trophic levels. Z. Zhou et al. (2007) examined the stratigraphic distribution of the acritarch biotas that are typical of both the Pertatataka Formation in Australia and the Doushantuo Formation in South China. Specifically, using chemostratigraphy and biostratigraphy, they detailed first appearances of this assemblage with regard to Neoproterozoic glaciations. Their research suggests that this biota appeared immediately after glaciation in the East Yangtze Gorges area; its appearance at Weng’an, however, cannot be correlated with a glaciation event. Interestingly, acritarch biozonation in Australia places the first appearance of acanthomorphic species immediately

after the Acraman impact (580 Ma), and this has led some to suggest that this event may have triggered subsequent eukaryotic diversification (Grey et al., 2003; Willman et al., 2006). It appears that more data are needed from both paleobotany and geochronology before the evolution of these acritarch biotas can be confidently correlated with glaciations or other environmental perturbations. The fossil record from the Neoproterozoic is the most complete in the Precambrian, and correlations between particular biotas and their paleoenvironments can be discerned with much higher resolution (Knoll et al., 1989). Planktonic microfossils show an inshore—offshore pattern, as they do today, with biotas from inshore deposits typically exhibiting low diversity and dominance by one or two taxa. Open-shelf assemblages, however, are much more diverse and the individuals are more complex (Vidal and Knoll, 1983). Three different assemblages of microorganisms were described from the Ryssö Formation of Svalbard, Spitsbergen (Knoll and Calder, 1983). One assemblage consists of typical

70

Paleobotany: the biology and evolution of fossil plants

stromatolitic microorganisms. Another represents an open coastal environment and is dominated by planktonic forms such as acritarchs, and the third contains a large number of vase-shaped protists. The occurrence of these three microbiotas within a single formation provides a unique opportunity to evaluate organisms from three very different habitats and to develop a more complete picture of life from 800 to 700 Ma. Although much of our information on Precambrian organisms has come from structurally preserved fossils in chert deposits, more and more data are available on compressed organisms from other paleoenvironments than those represented by chert biotas. Knoll and Swett (1985) described a series of biotas from Spitsbergen which include unicellular and filamentous prokaryotes preserved in shales. They are able to compare these organisms to those found in the Bitter Springs chert and to slightly younger, silicified biotas from Svalbard. Butterfield et al. (1988) described a well-preserved fossil biota recovered from subtidal marine shales of the Neoproterozoic of Spitsbergen (800–700 Ma). This biota includes sphaeromorphic acritarchs, eukaryotic multicellular algae comparable to extant Ulvophyceae, cyanobacterial sheaths, rod-shaped and filamentous forms similar to heterotrophic bacteria, and what appear to be germinating zoospores of filamentous protists. Other forms that were present in this biota have been interpreted as allochthonous (i.e., not preserved in situ). These include Chuaria, Tawuia, and a number of morphologically complex acritarchs. Some of the more interesting forms include large vesicles (150– 250 μm in diameter) with long, terminally flared processes. These structures have been interpreted as encystment structures. This interpretation is supported by their occasional occurrence within even larger forms (300 μm in diameter). DOUSHANTUO FORMATION Our knowledge of Neoproterozoic life and the evolution of multicellular life has been greatly enriched by fossils from the Doushantuo Formation (590–555 Ma) of south central China. The formation consists of a series of finely laminated carbonaceous and siliceous shales that were deposited in a quiet, subtidal environment, probably in a restricted basin (Xiao et al., 2002). Fossil animals, algae, and cyanobacteria occur in cherts, phosphorites, and black shales (Xiao and Knoll, 1999). In a reexamination of the compressed Miaohe biota, which includes compressed carbonaceous macrofossils, from the uppermost black shales in the Doushantuo,

Xiao et al. (2002) proposed that there are only 20 taxa present, although more than 100 have been described. Of these, eight taxa are definitely eukaryotic algae, nine represent possible algae, and two taxa are cyanobacteria. The biota includes coenocytic green algae and thalli comparable to members of the Rhodophyta (red algae) and Phaeophyceae (brown algae; Chapter 4). This biota occurs immediately after the Marinoan glaciation and just prior to the appearance of the metazoan Ediacaran fauna, but there appear to be no metazoans in this compressed biota. Algae, including acritarchs, are also preserved as three-dimensional phosphatized remains in the Doushantuo. Xiao et al. (2002, 2004) reported on exceptionally well-preserved, multicellular and pseudoparenchymatous red algae, as well as cell division in algal cells. These, along with other Neoproterozoic algae, are described in more detail in Chapter 4.

CONCLUSIONS In the past, paleobiologists viewed the Precambrian as a long interval of time in which very little biological evolution occurred. Now however, now we recognize that this interval of geologic time was one of the most exciting in terms of biological change and innovation. The complete Precambrian record provides some of the best evidence for evolution through geologic time, beginning with the simplest, unicellular prokaryotic microorganisms and progressing through the earliest, unicellular eukaryotes into more complex colonial types, and finally, multicellular forms. Several authors have suggested that all the important biochemical and cellular evolution took place in the Precambrian and the various Phanerozoic life simply represents the morphological and physiological elaboration of forms that evolved in the Precambrian. The volume and importance of the work done in this area in the past 15–20 years is extraordinary. Perhaps nowhere else in the geologic record has such a diversity of evidence been applied to answer evolutionary questions, including biomarkers, geochronology, isotope chemistry, molecular phylogenetics of extant organisms, and microfossils. At this point in time, we are beginning to gain a clearer picture of the most ancient life—not just the types of organisms present, but also the environments in which they lived, their nutritional modes, and their interactions with each other and the abiotic world. As research on Precambrian life continues, one thing stands out, as it does in our modern world—the ubiquity and versatility of life on Earth.

3 FUNGI, BACTERIA, AND LICHENS Fungi ................................................................................................. 71

Fungal Spores....................................................................................111

Earliest Fossil Fungi............................................................................73

Fungal-like Organisms ......................................................................112

Systematics of Fungi ...........................................................................77

Eubacteria and archaea ................................................ 112

Fungal Life-History Strategies............................................................98 Archaea .............................................................................................113

Fungi–Animal Interactions ...............................................................105

Eubacteria .........................................................................................113

Geologic Activities of Fungi .............................................................107

Lichens ...........................................................................................117

Epiphyllous Fungi .............................................................................108

Rain, and then the cool pursed lips of the wind draw them out of the ground Mary Oliver, Mushrooms

Fungi

Higher taxa in this chapter:

Kingdom Fungi Chytridiomycota Zygomycota Zygomycetes Trichomycetes Glomeromycota Ascomycota Pezizomycotina Basidiomycota Agaricomycotina Fungal-like organisms Peronosporomycetes (Oomycota) Kingdom Archaea Kingdom Bacteria Cyanobacteria Lichens

Fungi are primarily terrestrial, achlorophyllous, eukaryotic organisms that were at one time grouped with plants. Today, however, fungi are regarded as a monophyletic group more closely related to animals than plants and they occupy their own Kingdom (FIG. 3.1). Plesiomorphies shared with animals include the presence of chitin, food stored as glycogen, and, in the mitochondrial RNA, the bases uracil–guanine– adenine (UGA) code for the amino acid tryptophan (plants use UGG to code for tryptophan). The multicellular fungal body consists of filaments, called hyphae (sing. hypha), which together make up the mycelium or vegetative body of the fungus. From this simple organization, however, fungi can form many types of complex sexual and asexual reproductive structures, and they have developed adaptations that allow them to live in every habitat on Earth. Approximately 100,000 species of fungi have been described, but estimates

71

Paleobotany: the biology and evolution of fossil plants

Laboulbeniomycetes Lecanoromycetes Eurotiomycetes Sordariomycetes (Xylariales, Meliolales)

PEZIZOMYCOTINA

Leotiomycetes Dothideomycetes Pezizomycetes

Ascomycota

72

SACCHAROMYCOTINA (yeasts) DIKARYA TAPHRINOMYCOTINA

AGARICOMYCOTNA USTILAGINOMYCOTINA PUCCINIOMYCOTINA

Basidiomycota

Glomeromycota MUCOROMYCOTINA Zygomycota ENTOMOPHTHOROMYCOTINA HARPELLOMYCOTINA BLASTOCLADIOMYCOTINA (treated with chytrids here) Chytridiomycota

Figure 3.1 Relationships of major fungal groups. (Modified from J. Taylor and Berbee, 2006.)

are that more than 1.5 million species are yet to be discovered (Hawksworth et al., 1995). Other organisms that were historically included with the fungi, but are now placed in the Kingdom Chromista, include the Oomycetes (water molds), Hyphochytridiomycetes, and Labyrinthulomycetes. Of these three clades, only the water molds will be discussed here. Slime molds were also once considered to be Fungi, but are now included in the Kingdom Protista. The fossil record of slime molds (FIGS. 3.2–3.4), once called the myxomycetes, is meager, restricted only to several found in Eocene amber (Dörfelt and Schmidt, 2006), so they will not be covered in this book. Together with a few other groups of heterotrophic organisms (e.g., bacteria), the fungi are the principal decomposers in the biosphere, releasing CO2 into the atmosphere and nitrogenous compounds into the soil. They also function as bioweathering agents and transformers of minerals and rocks (Burford et al., 2003). Fungi are involved in numerous types of associations with other organisms, ranging from those that produce diseases in plants and animals to a variety of beneficial, symbiotic relationships (e.g., mycorrhizae in the roots of most vascular plants) (Newsharn et al., 1995). Despite the fact that fungi, as a

group, have a long and interesting geologic history (Tiffney and Barghoorn, 1974), only relatively recently have they been studied in any detail. As a result, their importance in the evolution of past biotas and our knowledge of the evolutionary history of the major fungal groups through time have been minimized. There are several reasons why our understanding of fossil fungi has lagged behind our knowledge of many other fossil organisms. One of these is the long-held belief that fungi are too fragile to be adequately preserved in the fossil record (FIG. 3.5). Additionally, the study of fossil fungi may have been avoided due to difficulties in recognizing and interpreting them (FIGS. 3.6, 3.7). This is especially true in situations where fungi have been found within permineralized vascular plants (FIG. 3.8), which typically represent the principal focus of the research (LePage et al., 1994). In addition, when collecting fossil plants, paleobotanists may introduce their own inherent bias by focusing on collecting the best specimens of a taxon (e.g., those that provide the most informative characters of the organism). As a result, specimens that might show the effects of fungal activities, such as degraded tissue or necrotic areas, are simply not brought back to the laboratory for study. Some of this bias has been overcome by the modern use of

CHAPTER 3 fungi, bacteria, and lichens

73

Figure 3.3 Arcyria sulcata (slime molds), capillitium com-

posed of capilitial threads (see FIG. 3.2) (Eocene). Bar  300 μm. (Courtesy A. Schmidt.)

Figure 3.2 Arcyria sulcata (slime molds), sporocarp composed

of a stalk and cupuliform base of the pteridium (cup) (Eocene). Bar  100 μm. (Courtesy A. Schmidt.)

quantitative field techniques, in which all available material within a certain area is examined or collected. Finally, the environment of deposition and fossilization may affect preservation. For example, Carboniferous swamp plants often occur in peat deposits, and the chemistry in the swamp may have discouraged extensive fungal activity. Nevertheless, despite all of these limitations, the literature on fossil fungi and their biotic and abiotic interactions is rapidly increasing. As might be expected, some of the details, especially those that characterize levels of fungal interaction in ecosystems (Taylor, 1990; Taylor et al., 2004) are more difficult to determine from fossils. Earliest Fossil Fungi

Although fungi were probably present during the Precambrian, the earliest fossil record of them has been difficult to interpret (Brunel et al., 1984). There are spore-like bodies and filaments in Precambrian rocks that may represent the remains of fungi, but the affinities of some of them remain equivocal. For example, although initially described as

Protophysarum balticum showing three sporocarps emerging from a plant fragment (Eocene). Bar  100 μm. (Courtesy A. Schmidt.)

Figure 3.4

septate fungal filaments, Eomycetopsis robusta (Schopf, 1968), (FIG. 2.32) from the Neoproterozoic Bitter Springs Formation (830–800 Ma) of Australia is now regarded as a cyanobacterial sheath. Tappania (FIG. 3.9) is a Neoproterozoic acritarch (1.43 Ga) from marine carbonates and shales that has been suggested to be fungal (Butterfield, 2005). Extending from the spherical main body of this organism is what are interpreted as septate hyphae, some

74

Paleobotany: the biology and evolution of fossil plants

Figure 3.5 Zoosporangium with discharge opening found in

phloem cells of a Carboniferous fern. Bar  20 μm.

Figure 3.7 Chlamydsospore with chytrids on the surface (Devo-

nian). Bar  300 μm.

Figure 3.6 Tracheid from Sphenophyllum containing fungi (Pennsylvanian). Bar  500 μm.

of which anastomose. Other organisms have been described from late Mesoproterozoic rocks as possibly fungal in origin (Hermann, 1979). Convincing evidence that any of these are some type of reproductive organ from a marine fungus, however, has not yet been forthcoming. Several structures regarded as fungal have been reported from the Lakhanda Series of Siberia dated at 1 Ga (Hermann and Podkovyrov, 2006). Although the specimens are hypothesized to be possible zygomycetes, their preservation on organic sapropelic films makes assignment difficult. An even earlier putative fungus from Kola Peninsula (northwestern Russia) is dated at 2 Ga (Belova and Akhmedov, 2006). Petsamomyces varies in morphology and bears what are termed hyphal-like

Figure 3.8 Cortical cell filled with fungal hyphae (Pennsylvanian).

Bar  10 μm.

appendages. Like many of these organic-walled microfossils, the affinities of these structures remain problematic. Despite these uncertainties, the presence of fungi in the Proterozoic has been indirectly inferred on the basis of divergence times using molecular clock assumptions, but see Taylor and Berbee (2006) on problems associated with many of

CHAPTER 3 fungi, bacteria, and lichens

75

Figure 3.9 Tappania (Neoproterozoic). Bar  100 μm. (From Butterfield, 2005.)

Branching hypha (Silurian). Bar  25 μm. (From Sherwood-Pike and Gray, 1985.)

Figure 3.10

these estimates. Based on these calibrations it is suggested that fungi may have diverged from metazoans about 1.5 Ga (Hedges et al., 2004; Taylor and Berbee, 2006). If these estimates are even close to being accurate, then it would be expected that more fossil fungal remains will be described from these ancient sediments in the years ahead. Some of the oldest fossil remains that are convincingly fungal in origin occur in Early Silurian (Llandoverian) rocks of Virginia, USA (Pratt et al., 1978). The bulk maceration of these terrestrial rocks produced small (6 μm wide), septate, and branched filaments (FIG. 3.10). Some filaments had specialized cells morphologically identical to those of fungi that produce endogenously formed chains of conidia (FIG. 3.11). To date, the single most important source of ancient fungi is the Early Devonian (Pragian—earlist Emsian) Rhynie chert from Aberdeenshire, Scotland. This Lagerstätte represents an entire ecosystem that is petrified in silica—plants, animals, and microbes are all present. Numerous spores (FIG. 3.12), hyphal filaments (FIG. 3.13), and sporocarps have been described from the silicified matrix and from tissues of the land plants Asteroxylon mackiei, Rhynia gwynne-vaughanii, Aglaophyton, Nothia, and Horneophyton. Rhynie chert fungi, such as Palaeomyces (FIG. 3.14) (Kidston and Lang, 1921a),

Septate conidium (Silurian). Bar  15 μm. (From Sherwood-Pike and Gray, 1985.)

Figure 3.11

76

Paleobotany: the biology and evolution of fossil plants

Figure 3.12 Section of Aglaophyton major axis with two clusters of fungal spores (Devonian). Bar  300 μm. Figure 3.14 Palaeomyces sp. spores within large resting spore

(Devonian). Bar  75 μm.

Figure 3.13 Hyphae and chlamydospores in Rhynie chert plant tissue (Devonian). Bar  120 μm.

Palaeoblastocladia (Remy et al., 1994), and Glomites (Taylor et al., 1995), are but a few of the morphotaxa known from the Rhynie chert ecosystem. These are described later in the sections “Glomeromycota” and “Chytridiomycota” (Palaeoblastocladia). Other fungi in the Rhynie chert cannot yet be assigned to major groups. They consist of non-septate hyphae that branch at irregular intervals, as well as hyphae that are distinctly septate, with the central region of the septum slightly thickened (Kidston and Lang, 1921a). At various points along the hyphae, ovoid to pear-shaped vesicles occur which are believed to have developed into large (250 μm), thickwalled sporangia. In other thin-section preparations, especially those of the chert matrix, larger sporangia occur, but these have not been found attached to hyphae. Those with stratified walls were named Palaeomyces gordonii var. major and are now thought to be members of the Glomeromycota (discussed later). Other sporangia contained a variety of thick-walled structures termed resting spores or resting sporangia. There is no doubt that several different natural forms are represented by these Rhynie chert fungi, but they cannot yet be assigned to a particular clade with certainty, as important parts of their life cycles have not yet been discovered. Recent work indicates that there is not only a considerable diversity of fungi within the Rhynie chert but also that these specimens offer the opportunity to examine the life history

CHAPTER 3 fungi, bacteria, and lichens

of some of these early terrestrial microorganisms as well as their biological interactions with other components of the ecosystem. The biological relationships of the Rhynie chert fungi, especially those preserved within or on land plants in the chert, have been variously interpreted historically. Kidston and Lang (1921a) suggested that the nutritional mode of most of these fungi was saprophytic, but that some may have represented symbionts. Other workers (Boullard and Lemoigne, 1971) hypothesized that some of the Rhynie chert fungi were parasitic. Further work has shown that some of the Rhynie plants exhibit host response features and these can provide clues to the nutritional mode of certain fossil fungi. The presence of fungi in two different taxa from this site, Rhynia gwynne-vaughanii and Aglaophyton major (formerly Rhynia major), was at one time used to support the suggestion that these two different plants represented independent phases of the same life cycle. As will be discussed in Chapter 8, we now

Multicelled fungal spore (Cretaceous). Bar  10 μm. (Courtesy J. M. Osborn.)

Figure 3.15

77

know that R. gwynne-vaughanii and A. major are not only different organisms, but probably belong to different clades. Fungal spores represent one of the most common examples of fungi in the fossil record, and are found in a variety of facies from the Paleozoic (Pirozynski, 1976a; Ediger and Alisan, 1989) to the recent (see section “Fungal Spores”). The identification of fungal palynomorphs (FIG. 3.15) is difficult, but there are now several glossaries of descriptive terms relating to spore and thallus morphology and structure (Elsik et al., 1983; Kalgutkar and Jansonius, 2000). Systematics of Fungi

In the last edition of this book (Taylor and Taylor, 1993), fungi and fungal-like organisms were described within a stratigraphic framework, and when possible, comments were offered on where they might fit within a modern classification of the Kingdom Fungi. With more information now available about fungal diversity through time, and many more specimens known in greater detail, many of the fossil taxa can now be discussed within the context of modern fungal groups. Phylogeny of the fungi was once based on morphology and, in some cases, characteristics in laboratory cultures. Today, the fungal tree (FIG. 3.1) of life is constantly refined using various molecular sequences. Most analyses include five Phyla within the fungi, three of which, the Glomeromycota, Ascomycota, and Basidiomycota, are considered monophyletic. The other fungi, which are considered to be the earliest-diverging fungi based on molecular phylogenies, include the paraphyletic groups, Chytridiomycota and Zygomycota. See Blackwell et al. (2006) for a summary of progress on the fungal tree of life, as well as the special issue of Mycologia, A Phylogeny for Kingdom Fungi (98(6), December 2006). CHYTRIDIOMYCOTA Members of this phylum are the only fungi that produce motile spores (zoospores) at some stage in their life cycle. Today chytrids are found in both soil and freshwater, with many functioning as saprotrophs and some as parasites (FIG. 3.16). Phylogenetic studies based on molecular markers view the chytrids as an early diverging, probably paraphyletic group within the Kingdom Fungi and the sister group of the remaining non-flagellated fungi (Zygomycota, Glomeromycota, Ascomycota, Basidiomycota) (James et al., 2006a). Increased resolution within the classification of living members of the Chytridiomycota is occurring at several levels (Letcher et al., 2005; James et al., 2006b). For

78

Paleobotany: the biology and evolution of fossil plants

Figure 3.16 Chytrid (arrow) with discharge papillae attached

to pollen grain (Extant). Bar  25 μm. (Courtesy D. Barr.)

Figure 3.19 Chlamydospore with mycoparasite developing

between wall layers (arrow) (Devonian). Bar  20 μm.

Figure 3.17 Chytrid zoosporangia embedded in outer wall of

Palaeonitella cranii cell. Bar  20 μm.

Figure 3.20 Several chytrid zoospores. Arrow indicates flagellum (Devonian). Bar  10 μm.

Figure 3.21 Several chytrid zoosporangia on the surface of a Figure 3.18 Detail of chlamydospore wall showing mycopara-

site developing between wall layers (Devonian). Bar  12 μm.

land plant spore. Arrow shows zoosporangium exit site (Devonian). Bar  10 μm.

CHAPTER 3 fungi, bacteria, and lichens

example, at the molecular level, the relationship between members of the Chytridiomycota and Glomeromycota is strengthened by the presence of certain tubulin genes in both groups (Corradi et al., 2004). Although there have been some putative chytrids described from Cambrian rocks (see Butterfield, 2005 for a review), the best-preserved forms are from the Lower Devonian Rhynie chert (FIG. 3.17) (Illman, 1984; Taylor et al., 1992). These fossils possess thalli and discharge tubes (FIGS. 3.18 and 3.19) similar to those in certain modern chytrids, and the preservation is so detailed that it is possible to demonstrate the presence of a single flagellum (FIG. 3.20) on a fossil zoospore (Taylor et al., 1992). These fossil chytrids possess thalli that are epi- and endobiontic (living on or within other organisms), and are associated with land plants, spores (FIG. 3.21), and algae (FIG. 3.22), as well as occurring isolated in the matrix. One especially interesting form is Palaeoblastocladia (Remy et al., 1994), a fungus that shares many features with members of the extant Blastocladiales

(FIG. 3.23). Palaeoblastocladia milleri consists of thalli that were produced beneath the cuticle of stems of Aglaophyton and extended out from the surface about 0.5 mm (FIG. 3.24). Some thalli had terminal zoosporangia (FIG. 3.25), whereas others produced pairs of terminal gametangia (FIG. 3.26). This complement of characters suggests that P. milleri possessed an isomorphic alternation of generations with sexual reproduction, a combination of features that is very rare in modern fungi. It is suggested that P. milleri was a saprotrophic, but other Rhynie chert chytrids are believed to have been parasites. Despite the small size of chytrids, their ubiquity in the Rhynie chert provides the opportunity to study their life history biology (FIG. 3.27), since various developmental stages are preserved. Geologically younger fossil chytrids are also known (FIG. 3.28). Some of the first late Paleozoic chytrids to

Sporothallus

Figure 3.23

Figure 3.22 Chytrid zoosporangium. Arrow indicates position

of discharge tube (Devonian). Bar  20 μm.

79

Gametothallus

Life history of Palaeoblastocladia milleri.

Figure 3.24 Two tufts (arrows) of Palaeoblastocladia milleri extending from epidermis of Aglaophyton axis (Devonian). Bar  400 μm.

80

Paleobotany: the biology and evolution of fossil plants

Figure 3.26 Two pairs of Palaeoblastocladia milleri gametangia (Devonian). Bar  30 μm.

Figure 3.25 Zoosporangium of Palaeoblastocladia milleri

showing zoospores with dark central body (Devonian). Bar  15 μm.

Mature holocarpic thallus Zoospore discharge Immature thallus

Zoospore Protoplast Penetration tube

Figure 3.27

Life history of a Devonian chytrid.

CHAPTER 3 fungi, bacteria, and lichens

81

be accurately identified were Grilletia spherospermii and Oochytrium lepidodendri (FIG. 3.29) which occur in Carboniferous gymnosperm seeds and anatomically preserved tissues of the lycopsid Lepidodendron sp. (Renault and Bertrand, 1885a (FIG. 3.32); Renault, 1895, 1896a). These authors not only related these fossils to modern chytrid genera but also suggested that they represented parasitic fungi. Other microfungal remains associated with permineralized Lepidodendron tissues are more difficult to interpret (FIGS. 3.30, 3.31). Fossil chytrids (FIG. 3.33),

Figure 3.28 Thick-walled lycopsid spore containing fungal

spores (chytrids?) (Pennsylvanian). Bar  225 μm.

Arthroon rochei, thick-walled spore in periderm of Lepidodendron (Mississippian). Bar  100 μm. (Courtesy N. Dotzler.)

Figure 3.30

Figure 3.29 Oochytrium lepidodendri in Lepidodendron tracheids (Mississippian). Bar  20 μm. (Courtesy N. Dotzler.)

Figure 3.31 Palaeomyces gracilis. (Mississippian). Bar  35 μm. (Courtesy N. Dotzler.)

82

Paleobotany: the biology and evolution of fossil plants

Figure 3.34 Several chytrids attached to substrate (Extant). Bar  50 μm. (Courtesy D. Barr.)

Figure 3.32 Paul Bertrand.

Figure 3.33 Numerous chytrids on surface of fungal spores

(Devonian). Bar  20 μm.

like their modern relatives, are often found on spores and pollen grains (FIG. 3.34) in bodies of water. Millay and Taylor (1978a) described epibiotic and endobiotic chytrid thalli (FIGS. 3.35, 3.36) in association with Pennsylvanian cordaitean pollen grains (Millay and Taylor, 1978a). Some of the chytrid zoosporangia recovered from the Eocene Green River Formation, which is known for its excellent

Saccate pollen grain with two chytrid zoosporangia (arrows) attached (Pennsylvanian). Bar  25 μm. Figure 3.35

preservation, are so nearly identical to those of modern forms that the fossils are assigned to living genera (Bradley, 1964, 1967). ZYGOMYCOTA Zygomycetous fungi are distinguished by the production of thick-walled zygospores (non-flagellated) that form in a

CHAPTER 3 fungi, bacteria, and lichens

83

Figure 3.36 Epibiontic chytrid zoosporangium extending from

corpus wall of pollen grain (Pennsylvanian). Bar  10 μm.

Figure 3.38 Dubiocarpon (Pennsylvanian). Bar  175 μm.

Figure 3.37 Suggested reconstruction of Endochaetophora antarctica thought to represent an ascomycete sporocarp (Triassic).

special sporangium, the zygosporangium, following gametangial fusion. Hyphae are generally aseptate and asexual reproduction occurs by the formation of internally produced spores. The group is highly diverse and includes saprotrophs (e.g., black bread mold) and certain pathogens, including some that infect other fungi, called mycoparasites. Some forms live as obligate symbionts within the gut of various arthropods (Lichtwardt et al., 1999; 2001). The Zygomycota are currently interpreted as paraphyletic and are believed

to have diverged from the chytrids before the colonization of the land. Two classes are currently recognized (White et al., 2006): Zygomycetes and Trichomycetes. The only fossil record for the Trichomycetes will be discussed under the section “Fungi–Animal Interactions”. Various spore-like bodies that are now interpreted as zygomycetous fungi occur in permineralized peat (FIG. 3.37), especially in Carboniferous coal balls. The discovery of these unique structures, termed sporocarps, is no doubt directly related to the long history of the study of vascular plants in coal balls. The fungal affinities of these fossils are based on the structure of the wall, which consists of aseptate, interlaced hyphae. Some of the most common forms are spherical and 1 mm in diameter. The most common morphogenera include Sporocarpon, Dubiocarpon (FIG. 3.38), Mycocarpon, Coleocarpon, and Traquairia (FIG. 3.39) (Stubblefield et al., 1983). Mycocarpon, a common Middle Pennsylvanian form, consists of a central spore-like structure 550 μm in diameter surrounded by a wall of interlaced, hypha-like cells four layers thick. Inside is an amorphous, cuticle-like membrane. In some species, for example M. bimuratus, the central cavity is filled with small spores. In Sporocarpon, the sporocarp is smaller (200 μm) and constructed of a pseudoparenchymatous tissue,

84

Paleobotany: the biology and evolution of fossil plants

Figure 3.39 Traquairia williamsonii sporocarp (Carboniferous).

Bar  100 μm.

which extends outward into numerous narrow, conical processes, each 6–7 cells high and 1–3 cells wide. Another form, Dubiocarpon, is distinguished by radially oriented, elongate cells that extend out from the sporocarp wall as spines, some with bifid tips. The most ornate form is Traquairia, a fungus that was initially described from the Lower Coal Measures of Great Britain by Carruthers (1872a). Since that time, numerous specimens have been reported from many localities in Pennsylvanian rocks (Stubblefield and Taylor, 1983). Individual specimens are roughly spherical and up to 1 mm in diameter. The wall is complex, with the outer portion constructed of branching hyphae, some of which are organized into hollow spines. In Roannaisia, from the Visean of central France, the hyphae branch and the lumen of the sporocarp contains a single, multilayered spore (Taylor et al., 1994b). As noted earlier, some of these sporocarps contain one to several spherical structures in the central lumen (FIG. 3.40). Initially these fungi were interpreted as ascomycetes, and the sporocarps as closed ascocarps (cleistothecia). Larger spherical structures inside would represent asci (sporangia), and smaller ones, ascospores (Stubblefield and Taylor, 1988). These sporocarps are now interpreted as zygomycetes, however, and the large, inner spore-like body represents a zygospore, similar to those produced in mycelial sporocarps of modern Mucorales (White and Taylor, 1989a). The smaller internal spores reported in some sporocarps (formerly interpreted as ascospores) are now regarded as mycoparasites, most likely some type of chytrid.

Figure 3.40 Sporocarp of Dubiocarpon containing several larger spores (Pennsylvanian). Bar  225 μm.

An interesting fossil believed to be a zygomycete is Protoascon missouriensis, a fungus found within a seed-like structure in a Pennsylvanian coal ball from Missouri, USA (Batra et al., 1964). The description is based on multiple specimens of a bulb-like structure 150 μm diameter (FIG. 3.41). At one end is a whorl of 12 elongate, aseptate appendages (FIG. 3.42). The appendages are curved to form a loose, basket-like structure which surrounds a highly ornamented sporangium containing a single, thick-walled spore. As the name suggests, this fossil was initially thought to be an ascomycete, but later reinterpreted as a chytrid (Baxter, 1975), and most recently as a zygomycete (Taylor et al., 2005a). Based on reexamination of the type material and additional specimens, Taylor et al. (2005a) concluded that the aseptate appendages (or suspensors) partially enclose a sporangium, which is either an azygosporangium (asexual) or a zygosporangium, containing a single, thick-walled spore (FIG. 3.43). GLOMEROMYCOTA The Glomeromycota are a clade that was instituted based on rDNA phylogenies of living members (Schüssler et al., 2001; Redecker and Raab, 2006). The phylum, which includes the arbuscular mycorrhizal (AM) fungi, was formerly included

CHAPTER 3 fungi, bacteria, and lichens

85

AZ

AZ

S

Figure 3.41 Protoascon missouriensis showing suspensor (S)

with appendages and (a)zygosporangium (AZ) (Pennsylvanian). Bar  50 μm. Figure 3.43 Protoascon missouriensis thick-walled (a)zygosporangium (AZ) (Pennsylvanian). Bar  50 μm.

Figure 3.42 Suspensor appendages of the mucoralean fungus Protoascon missouriensis (Pennsylvanian). Bar  50 μm.

within the Zygomycota, and is now considered to be the sister group of the clade formed by the Ascomycota  Basidiomycota, based on molecular phylogenies (Blackwell et al., 2006; Redecker and Raab, 2006). Extant Glomeromycota are comprised of obligate symbionts that may form arbuscules in plant roots; they produce large (40–800 μm),

multilayered spores which are attached to non-septate hyphae. More than 90% of extant land plants have a symbiotic (mutualistic) relationship with mycorrhizal fungi in their roots. There are two basic types of extant mycorrhizae: ecto- and endomycorrhizae. Endomycorrhizae are formed by members of the glomeromycetes and are the most common form today. The fungal hyphae grow within the host root, and although they penetrate the host cell walls, they do not penetrate the plasma membranes. Most produce arbuscules, highly branched hyphal structures that provide for exchange between the fungal symbiont and its host. Some also produce storage organs called vesicles (the vesicular-arbuscular mycorrhizae). Ectomycorrhizae are formed by members of the Basidiomycota and a few ascomycetes; they are less common today and occur primarily in woody plants of the temperate zone, including many conifers (Chapter 21). In this case, the fungal hyphae form a net around the outside of the plant root, which penetrates between the cells of the root itself. Although there are many hypotheses for the establishment of this fungal–land plant association (e.g., resistance against drought, defense against root herbivory, etc.), the mutualistic association provides for increased mineral nutrient uptake by the plant in exchange for a source of carbon for the fungus.

86

Paleobotany: the biology and evolution of fossil plants

Figure 3.45 Intercalary glomalean chlamydospore (Devonian).

Bar  70 μm.

Figure 3.44

Bernard Renault.

The fossil record of Glomeromycota is believed to be ancient, extending well back into the Paleozoic. Spores and hyphae of a glomeromycotan type have been reported from rocks as old as the Cambrian (Pirozynski and Daplé, 1989), and from 460–455 Ma Upper Ordovician rocks (Redecker et al., 2000). Palaeoglomus grayi has aseptate (coenocytic) hyphae and spores that resemble living Glomus spores (Redecker et al., 2002). In these reports, however, there is no association with land plant remains, and thus the symbiotic association of these fungi remains unresolved. In addition, Butterfield (2005) noted that the very simplicity of these organisms does not provide enough diagnostic characters to separate them from other protists or parasites. Palaeomyces is a name first used by Renault (1896a) (FIG. 3.44) and later by Kidston and Lang (1921a) and others to describe large, isolated spores associated with Paleozoic plant remains; some of these are now included in the morphotaxon Glomites (FIG. 3.45). Although originally described from the Rhynie chert, they are also known from other sites. As well as being a source for beautifully preserved fossil fungi, the Rhynie chert ecosystem provides some of the best fossil evidence of fungi and land plants interacting in a symbiotic association. Glomites rhyniensis

Figure 3.46 Detail of mycorrhizal arbuscule zone (dark band) in Aglaophyton (Devonian). Bar  325 μm.

consists of aseptate hyphae and globose, multilayered spores, which occur in the tissues of several macroplants from this Devonian site (Taylor et al., 1995). Of special significance is the discovery of highly branched, intercellular arbuscules of this species within the cells of the land plant Aglaophyton major. In extant plants with AM fungi, arbuscules are

CHAPTER 3 fungi, bacteria, and lichens

87

Figure 3.48 Section of Aglaophyton major axis showing cortical cells and hyphae (arrows) of Glomites rhyniensis in intercellular spaces. Bar  32 μm.

Figure 3.47 Aglaophyton major showing arbuscule trunk hyphae (arrows) and opaque structures representing arbuscules (Devonian). Bar  80 μm. (Courtesy H. Kerp.)

confined to cells of the root, but in A. major the arbuscules occur in a narrow zone of cells inside the cortex, termed the mycorrhizal arbuscule zone (FIGS. 3.46, 3.47), which extends throughout the rhizome and proximal portions of the aerial axes (FIG. 3.48). More recently another land plant from the Rhynie chert, Rhynia gwynne-vaughanii, was described with Glomites fungi in the cortical tissues (Karatygin et al., 2006). In this taxon, G. sporocarpoides, arbuscules were not identified, but there were large, glomoid sporocarps containing numerous spores (20–24 μm in diameter) in the tissues. Rhizomatous and upright axes of Nothia aphylla, another land plant from the Rhynie chert, host a glomeromycotan fungus that closely resembles G. rhyniensis. Glomites rhyniensis is an intercellular endophyte, however, that becomes intracellular only within a well-defined region of the cortex where it forms arbuscules. The fungus in N. aphylla is initially intracellular, but later becomes intercellular in the cortex where it forms vesicles (FIG. 3.49) and thick-walled spores (Krings et al., 2007b). If this fungus is

Figure 3.49 Vesicles and hyphae in Nothia aphylla axes.

Bar  100 μm.

functioning as an endomycorrhiza, N. aphylla displays an alternative mode of colonization by endomycorrhizal fungi, which may be related to the peculiar internal anatomy of the rhizomatous axis. In this part of the axis of N. aphylla, the cells are arranged in radial rows with virtually no intercellular spaces, so perhaps there is no intercellular infection pathway into the cortex (Krings et al., 2007a, b) this part of the axis of (see Chapter 8 for additional data on N. aphylla).

88

Paleobotany: the biology and evolution of fossil plants

Figure 3.51 Cross section of Psilophyton dawsonii axis contain-

ing chlamydospores (Devonian). Bar  750 μm. (From Stubblefield and Banks, 1983.)

Figure 3.50 Arbuscules (arrows) in gametophyte of Lyonophyton rhyniensis (Devonian). Bar  35 μm.

We now know that several of the gametophytes of the Rhynie chert plants were also endomycorrhizal (FIG. 3.50). Although the Rhynie chert does not represent the oldest terrestrial ecosystem, the extraordinary preservation at this site does indicate that the association of plants and certain types of fungi is ancient and, in fact, lends support to the hypothesis that fungi may have been critical in the early colonization of land by plants (see Chapter 6; Pirozynski and Malloch, 1975). In addition to the fungal remains associated in the Rhynie chert plants, large spores like those associated with mycorrhizal symbiosis in the Rhynie chert plants have been found in many other Paleozoic plants. Within the tissues of Psilophyton dawsonii (Late Devonian) are spherical spores up to 175 μm in diameter (FIG. 3.51), which appear similar to the chlamydospores (thick-walled, asexual resting spores) of certain modern fungi (Stubblefield and Banks, 1983). Such chlamydospores (FIG. 3.52) are also relatively common in the tissues of a large number of Carboniferous plants preserved in coal balls (Wagner and Taylor, 1982). In some instances they occur within the tissues of underground

Figure 3.52 Lycopod megaspore containing chlamydospores (arrows) (Pennsylvanian). Bar  150 μm. (From Stubblefield and Taylor, 1988.)

organs such as Stigmaria, the underground organ of the arborescent lycopsids; in other cases they are solitary within the matrix. Spores of this type range from 100–400 μm in diameter and possess a multilayered wall. Many are preserved attached to the hyphae that produced them, and, in some specimens, the attachment area is structurally identical to those seen in modern endomycorrhizal fungi like Glomus. Dotzler et al. (2006) described thick-walled spores from the Rhynie chert that are unusual because they possess a germination shield (FIGS. 3.53, 3.54), a structural feature associated with spore germination. The shield is a lobed structure that occurs between the inner and outer walls of the spore. What is especially intriguing is that typical germination shields only occur in a few glomeromycotan genera. To date the function of the germination shield remains unknown. Although molecular studies of the Glomeromycota have not resolved whether the germination shield is a derived or

CHAPTER 3 fungi, bacteria, and lichens

89

Figure 3.55 Arbuscule in root of Antarcticycas (Triassic). Bar 

25 μm.

Figure 3.53 Scutellosporites devonicus showing germination shield in cross section (arrow) (Devonian). Bar  50 μm.

Figure 3.54 Tongue-shaped germination shield of a glomero-

mycotan fungus (Devonian). Bar  100 μm (Courtesy H. Kerp and H. Hass.)

pleisomorphic character, its discovery in the Rhynie chert is important as a direct marker for calibrating molecular clock hypotheses on endomycorrhizal fungi. In spite of the morphological similarities between many of the fossil spores and the thick-walled spores of the modern endomycorrhizal fungi, the presence of spores does

not conclusively demonstrate that all of the plants were mycorrhizal. Thick-walled spores are also produced by several groups of non-mycorrhizal fungi and thought to represent resting propagules. It is the distribution of the fungal hyphae within the tissues of the host plant and the presence of arbuscules, the specialized highly branched structure of the fungus that forms within cells of the host, that are the most conclusive pieces of evidence for endomycorrhizal associations. Although arbuscules have been reported from as early as the Paleozoic in permineralized plants, with the exception of those in the Rhynie chert, all have subsequently been reinterpreted as produced by nonmycorrhizal fungi or as some form of coalesced cell contents. One example of a Mesozoic host plant with true arbuscules comes from silicified Middle Triassic (Anisian) peat collected in Antarctica (Stubblefield et al., 1987a). Within the roots of the cycad Antarcticycas schopfii are large spores, hyphae (FIG. 3.55), and highly branched structures that are most certainly arbuscules. Like the arbuscules of modern AM fungi, the Triassic specimens possess hyphae that arise from a single point on the cell wall and repeatedly branch to fill the host cell lumen. Additional evidence supporting the existence of AM fungi in this Triassic ecosystem is the presence of the thallus of another mycorrhizal fungus, Sclerocystis. This fossil consists of an aggregation of nonseptate hyphae that give rise to 30 terminally produced thick-walled spores (FIG. 3.56) (Stubblefield et al., 1987b). The fungus appears most similar to the living species S. rubiformis. Stockey et al. (2001) reported coiled hyphae and arbuscules in the roots of a middle Eocene taxodiaceous conifer, Metasequoia milleri. Hopefully, these various examples of arbuscules will prompt the reexamination of underground organs of other fossil plants, not only to substantiate the existence of obligate biotrophs but also to offer

90

Paleobotany: the biology and evolution of fossil plants

Figure 3.56 Chlamydospores of Sclerocystis sporocarp (Triassic). Bar  15 μm.

information that may be useful in tracing the evolution of the important nutrient-exchange structure, the arbuscule.

ASCOMYCOTA The ascomycetes, or sac fungi, are the most diverse clade of living fungi, with more than 40,000 species. In recent phylogenies, the group includes three subclasses: the Taphrinomycotina, an early-diverging clade, the Saccharomycotina (the yeasts), and the Pezizomycotina, which includes the vast majority of the multicellular ascomycetes. The clade includes both singlecelled (yeasts) and filamentous forms, and the hyphae of the latter are septate. Sexual reproduction is characterized by an ascus (pl. asci), a sac-like sporangium in which nuclear fusion (karyogamy), followed by meiosis, take place, leading typically to the production of 4, 8, or 16 ascospores (FIG. 3.57). Asexual reproduction occurs by fission, fragmentation, or the formation of asexual spores. Ascomycetes are worldwide in distribution and can be found in a wide variety of habitats, today and in the past (White and Taylor, 1988). Several important pathogens today are ascomycetes (e.g., chestnut blight, powdery mildews, ergot), as well as the edible truffle. Some ascomycetes form symbiotic associations with certain algae and/or cyanobacteria to form lichens. The Ascomycota are regarded as sister to the Basidiomycota based on morphological and molecular phylogenies (Blackwell et al., 2006). Morphological features that ally them with the basidiomycetes include the segmentation of the hyphae in the form of cross walls (septa, sing. septum),

pairs of unfused nuclei after mating and before nuclear fusion, and structures in both groups that coordinate simultaneous mitosis of the two nuclei. Molecular clock analyses (Taylor and Berbee, 2006) and recent fossil evidence indicate that the Ascomycota are far older than once believed. The earliest fossils attributed to the Ascomycota are specialized cells (phyllides) macerated from middle Silurian rocks of Gotland, Sweden (Sherwood-Pike and Gray, 1985). These remains consist of tubular filaments with perforate septa; on some filaments the surface is ornamented by short branches (FIG. 3.10). Multiseptate spores with 1–9 cross walls occur in the macerates as well (FIG. 3.11). Evidence of ascomycetous fungi has also been reported on the cuticle of the Devonian enigmatic thalloid plant Orestovia devonica from the Early Devonian of Siberia (Krassilov, 1981a), including haustoria, hooked hyphae, and asci. One of the best-preserved examples of an early ascomycete is Paleopyrenomycites from the Rhynie chert (Taylor et al., 2005b). Specimens of P. devonicus occur as perithecial ascocarps just beneath the epidermis of the vascular plant Asteroxylon mackiei (FIG. 3.58). Perithecia are spherical with a short ostiolate neck (FIG. 3.59), and asci and paraphyses (sterile, hair-like filaments) line the inner surface of the ascocarp. Each ascus produces up to 16 ascospores (FIG. 3.60) with 1–5 cells each. In addition to the presence of the sexual (telemorphic) stage, P. devonicus is also known from tufts of conidiophores that are interpreted as the asexual (anamorphic) component of the life history. The presence of both phases makes P. devonicus one of the most completely known fossil fungi. Taylor et al. (2005b) noted that the features of this fungus compared with those in several extant groups (e.g., Xylariales in the Sordariomycetes); a cladistic analysis was not able to clarify relationships. The extraordinary preservation of this Early Devonian ascomycete should help to define character evolution in the ascomycetes, as well as serving as a benchmark in tracing the evolution of major lineages within the Ascomycota (Taylor and Berbee, 2006). Mycokidstonia sphaerialoides is another Rhynie chert fungus interpreted as an ascomycete (Pons and Locquin, 1981). They describe ascomata (fruiting bodies that give rise to asci and ascospores), which are spherical and 175 μm in diameter, but these structures share features with chytrid zoosporangia, so the attribution to the Ascomycota is not certain. Palaeosclerotium pusillum represents a sporocarp found among plant debris in Carboniferous coal balls and is believed to have some affinities with the Ascomycota (Rothwell, 1972). Each sporocarp consists of an ovoid cleistothecium 1.2 mm in diameter with a two-parted wall. The wall consists of a central zone of branched, septate hyphae,

CHAPTER 3 fungi, bacteria, and lichens

Asexual reproduction by spores (conidia)

91

Ascogonium Antheridium

Ascus Germinating ascospores (n)

Trichogyne

Ascospores (n)

Plasmogamy

Mitotic division Ascospore formation Developing ascogenous hyphae (nn)

Second meiotic division

Dikaryotic ascogenous hyphae (nn)

First meiotic division Young ascus (2n) Karyogamy (nn)

Ascocarp (Apothecium)

Sterile hyphae (n)

Figure 3.57 Life history of an ascomycete. (From Taylor and Taylor, 1993.)

Figure 3.58 Section of Asteroxylon mackiei enations show-

ing position of several Paleopyrenomycites devonicus perithecia (Devonian). Bar  10 μm.

Figure 3.59 Section through ostiole (arrow) of perithecium of Paleopyrenomycites devonicus and numerous ascospores in cavity (Devonian). Bar  60 μm.

92

Paleobotany: the biology and evolution of fossil plants

3.60 Several asci containing ascospores Paleopyrenomycites devonicus (Devonian). Bar  15 μm.

Figure

from

surrounded by a zone of pseudoparenchymatous tissue (FIG. 3.61). What is especially interesting about P. pusillum is that it appears to exhibit a combination of characteristics of several extant fungal groups. Within the cleistothecium are what have been interpreted as asci (FIG. 3.62) containing a variable number of spores (4–8). Hyphae possess dolipore-like septa and clamp connections, features typical of basidiomycetous fungi, but also found in other groups (McLaughlin, 1976). The taxonomic affinities and nutritional role of this interesting fossil have been variously interpreted (Dennis, 1976). Some regard it as a fungus intermediate between Ascomycota and Basidiomycota (McLaughlin, 1976), whereas others see it as an ascomycetous fruiting structure closely related to the Eurotiales, which has been parasitized by the mycelium of a basidiomycete (Singer, 1977). The Eurotiales are an order of ascomycetes that includes the sexual stages of Penicillium and Aspergillus, with the latter found in Baltic amber from Russia (FIGS. 3.63–3.65). In another interpretation, Palaeosclerotium is used as an example of a fungus that links the Basidiomycota symbiotically with possible nematophytes (Chapter 6), an extinct group of early land organisms (Pirozynski and Weresub, 1979). Other fossil evidence of Paleozoic members of the sac fungi include a variety of multicelled fungal spores with thick septa, which have been interpreted as ascomycetes (Kalgutkar and Jansonius, 2000). Another group of ascomycetes that are represented in the fossil record are the sooty molds belonging to the Capnodiaceae family (Dothideomycetes), a group that today produces colonies of dark hyphae on living plants often associated with aphid infestations. The aphids parasitize the plants, which, in turn, exude a sap that becomes the substrate for fungal growth. Fragments of Bitterfeld (24–22 Ma,

Figure 3.61 Palaeosclerotium pusillum showing outer, pseu-

doparenchymatous zone and central region of branched and septate hyphae (Pennsylvanian). Bar  500 μm.

Figure 3.62 Palaeosclerotium pusillum with central region containing spore-like structures (Pennsylvanian). Bar  500 μm.

Oligocene–Miocene) and Baltic (55–35 Ma, Eocene) amber contain brown hyphae with tapering distal ends that are identical to the extant sooty mold Metacapnodium (Rikkinen et al., 2003). Today some regard the sooty molds as monophyletic based on molecular studies (Reynolds, 1998; Schoch et al., 2006).

CHAPTER 3 fungi, bacteria, and lichens

93

Figure 3.63 Springtail overgrown by Aspergillus collembolorum (Eocene). Bar  0.5 mm. (Courtesy A. Schmidt.)

Figure 3.65 Aspergillus collembolorum, conidial head with radial chains of conidia (see FIG. 3.64) (Eocene). Bar  25 μm. (Courtesy A. Schmidt.)

Figure 3.64 Aspergillus collembolorum, sporulating conidio-

phores (see FIG. 3.63) (Eocene). Bar  50 μm. (Courtesy A. Schmidt.)

Stigmatomyces succini is a parasitic ascomycete that has been found attached to the thorax of a stalk-eyed fly preserved in Baltic amber (Rossi et al., 2005). It represents the earliest fossil account for the ascomycete order Laboulbeniales, an enigmatic group which is now considered to represent ascomycetes (Weir and Blackwell, 2001). BASIDIOMYCOTA The Basidiomycota, a monophyletic sister group to the Ascomycota, includes 30,000 extant species divided into three major lineages (subphyla): the rusts (Puccinomycotina), smuts (Ustilagomycotina), and mushrooms (Agaricomycotina). They are known from both terrestrial and aquatic habitats around the world and include important plant pathogens (e.g., wheat rust, corn smut), as

well as the edible mushrooms. The most diagnostic feature of the basidiomycetes is the basidium (pl. basidia), a generally club-shaped cell where nuclear fusion (karyogamy) takes place and the structure upon which the sexual spores (basidiospores) are produced. Some basidia are borne on complex, multicellular fruiting bodies, for example the mushrooms (FIG. 3.66). Other basidiomycete features include hyphal outgrowths termed clamp connections, and the presence of a dikaryon phase in the life cycle, a condition in which each cell in the thallus contains two nuclei. Some basidiomycetes are involved in ectomycorrhizal associations, whereas others are symbiotically associated with various insects, for example with leaf-cutter ants and termites. One example of a fungus–termite interaction is the Miocene–Pliocene termite trace fossil Microfavichnus; this ichnogenus is thought to represent fungus combs of fungus-growing termites (Duringer et al., 2007). The combs consist of alveolar masses in which the walls have a pelletal structure. The Basidiomycota today play an important role in the carbon cycle by decaying organic matter, including wood;

94

Paleobotany: the biology and evolution of fossil plants

Diploid nucleus undergoes meiosis

Pair of nuclei fuse

2n Fertilization

Meiosis nn

n Basidiospores

Basidium

Sterigmata

Basidium Cap Portion of gill

Gill Developing basiodiocarp

Button

Monokaryotic mycelium

Annulus

Spore

Stalk

Spore

Dikaryotic mycelium

Monokaryotic mycelium

Figure 3.66 Life history of basidiomycete fungus. (From Taylor and Taylor, 1993.)

presumed basidiomycetous fungi have been found as early as the Middle Devonian, not long after land plants first began producing secondary growth. There are a number of Carboniferous fossils that superficially resemble modern basidiomycetous fruiting bodies, or basidiocarps (Lindley and Hutton, 1831–1837; Lesquereux, 1877; Herzer, 1893; Hollick, 1910) (FIG. 3.67). Almost all of these reports, however, were later questioned and the specimens reinterpreted as non-fungal (Pirozynski, 1976b). The fact that there are so few reports of basidiomycetes associated with fossil wood is perplexing, since today they are the major decomposers of cellulose and lignin, the major components of plant cell walls. One well-documented basidiomycete from the Pennsylvanian is Palaeoancistrus martinii, a fungus found in the tracheids of the fern Zygopteris (Dennis, 1970). There are several features of this fungus that suggest affinities with living saprotrophic members of the Basidiomycota. One of these is the presence of smooth, narrow, septate hyphae

(4.8 μm in diameter) that follow a straight course within the tracheids. Associated with the hyphae are both terminal and intercalary chlamydospores. Some hyphae possess incomplete clamp connections in which the hook of the clamp does not form a complete union with the hypha, whereas in others the clamp connections are well developed. Another basidiomycete has been described in the secondary xylem of several Paleozoic and Mesozoic woods from Gondwana (FIG. 3.68), but in this case it is the symptoms caused by the fungus, in association with the fungus itself that provides the identification. The activities of this fungus result in the formation of numerous longitudinally oriented, spindle-shaped pockets of decay in the secondary xylem (FIG. 3.69), called pocket rot (Stubblefield and Taylor, 1986). In other regions of the secondary xylem, for example in the root Vertebraria and stems assigned to the morphogenus Araucarioxylon, septate hyphae with simple and medallion clamp connections are present (FIG. 3.70). The elongate,

CHAPTER 3 fungi, bacteria, and lichens

95

Figure 3.69 Fractured surface of extant wood showing elon-

gate spindles (white) of white pocket rot. Bar  1 cm.

Figure 3.67 Arthur Hollick.

3.70 Medallion clamp connection (arrow) in Araucarioxylon wood infested by pocket rot (Triassic). Bar  20 μm. (From Stubblefield and Taylor, 1986.) Figure

Figure 3.68 Cross section of Araucarioxylon wood showing symptom (white areas) of white pocket rot (Triassic). Bar  2 cm.

spindle-shaped areas in the fossil are identical to the symptoms caused by modern white pocket rot fungi (Blanchette, 1980). Other structural features in the fossil woods indicate the sequential delignification of the secondary cell wall (FIG. 3.71), for example the loss of the middle lamella between the wood cells, and these are also characteristic of modern white rot fungi. Palaeofibulus is another fossil fungus with clamp connections (FIG. 3.72) and thick-walled

spores, known from Middle Triassic permineralized peat of Antarctica (Osborn et al., 1989). Investigations of woodrotting fossil fungi offer the potential to indirectly examine the biochemical evolution of fungi, based on patterns and features associated with delignification and removal of cellulose from the host cell wall. Some of the most common members of the Basidiomycota are the mushrooms (Hibbett, 2006). To date the oldest gilled mushrooms (Agaricales or agarics) come from amber. The oldest of these is Archaeomarasmius leggeti (FIG. 3.73) entombed in a piece of Late Cretaceous (Turonian; 94– 90 Ma) amber from the Raritan Formation of New Jersey, USA (Hibbett et al., 1997). The pileus, or cap of the mushroom, ranges from 3–6 mm in diameter and contains elliptical

96

Paleobotany: the biology and evolution of fossil plants

Separation of wood cell wall layers due to fungal degradation (Triassic). Bar  55 μm.

Figure 3.71

Figure 3.73 Archaeomarasmius leggeti (Cretaceous). (Courtesy

D. Hibbert.)

Figure 3.72 Palaeofibulus showing hyphae, spores, and clamp

connection (arrow) (Triassic). Bar  35 μm.

basidiospores up to 8 μm long. Another gilled form is Protomycena electra from the younger Dominican amber (Miocene; 20–15 Ma) (Hibbett et al., 1997); this mushroom is similar to the extant leaf-litter and wood-rotting genus Mycena. Coprinites dominicana is another gilled mushroom found in Dominican amber. It has a cap 3.5 mm in diameter with 28 gills extending from the lower surface (Poinar and Singer, 1990). The most recent homobasidiomycete reported from Dominican amber is Aureofungus yaniguaensis (Hibbett et al., 2003). It is similar to the other fossil agarics in amber, and suggests that among certain homobasidiomycete lineages there has been relatively little morphological change since the Cenozoic, at least in those features that can be compared within the amber matrix. Basidiomycetes that lack gills but possess large basidiocarps with poroid hymenophores (spore producing layers in the fruiting bodies) have been described from the

fossil record as bracket, shelf, or polypore fungi (surveyed in Fleischmann et al., 2007). Members of this group are generally saprotrophs involved in the degradation of cellulose and lignin, but some are also parasites of woody plants. Quatsinoporites cranhamii is a fragment of an Early Cretaceous (Barremian) basidiocarp described from permineralized marine calcareous concretions of British Columbia, Canada (Smith et al., 2004). The hymenophore consists of numerous parallel tubes, each up to 540 μm in diameter. Appianoporites is also a poroid (bracket) fungus constructed of smaller tubes, from younger Eocene rocks from the eastern side of Vancouver Island, British Columbia. Both specimens possess septate hyphae; the basidiocarps are interpreted as persistent and placed in the Hymenochaetales. Another fossil polyporous fungus is Ganodermites libycus (FIG. 3.74) from the lower–middle Miocene (Neogene) of North Africa (Fleischmann et al., 2007). The basidia are clavate and produce elliptical basidiospores, each up to 6.5 μm long with two wall layers. The basidiocarp shows evidence of incremental growth and is placed in the extant family Ganodermataceae, a predominantly tropical group of fungi that are characterized by basidiospores with a double (so-called ganodermatoid) wall. The presence of tunnels containing fecal pellets in the basidiocarp indicates that this bracket fungus was visited by fungivores.

CHAPTER 3 fungi, bacteria, and lichens

97

H

Figure 3.74 Ganodermites libycus, longitudinal thick section (polished surface) through the basidiocarp showing hymenophoral strata (H) (Miocene). Bar  2 cm. (Courtesy BSPG.)

Basidiomycetes also form a variety of symbiotic associations with other organisms, including many families of temperate forest trees, for example Fagaceae and Pinaceae. Distinctive among these are ectomycorrhizae, estimated to occur on 10% of plants. Ectomycorrhizae are characterized by intercellular hyphae that often form a loose aggregation, called a Hartig net, around the root tip. In addition, the fungal association can also cause a change in root morphology, making the roots shorter, wider, and more branched. The oldest evidence of ectomycorrhizae to date comes from the middle Eocene (50 Ma) Princeton chert of the Allenby Formation, British Columbia, an extraordinary site in which many whole plants are permineralized by silica (LePage et al., 1997). Ectomycorrhizal roots of Pinus from this site contain a dense aggregation of small septate hyphae that extend into the cortex of the roots and represent the Hartig net. This discovery represents another example of the long standing symbiotic relationship in a particular group of seed plants. It is unclear why the majority of living ectomycorrhizae coevolved with woody rather than herbaceous plant hosts. Perhaps the structure and organization of the root, soil type, microbial community, or some other combination of biotic and abiotic factors favored ectomycorrhizal fungi over endomycorrhizal forms in certain hosts. A striking example of the diversity of fossil basidiomycetes is seen in a fossil specimen assignable to the Gasteromycetes, the group that contains the puffballs and some of the earthstars. Geastrum tepexensis (FIG. 3.75) is a compressed late Eocene basidiocarp (called a peridium in this group) approximately 2.5 cm in diameter from the Coatzingo Formation in Puebla, Mexico (Magallón-Puebla and Cevallos-Ferriz, 1993). The endoperidium (inner layer of the peridium) is circular and characterized by a small ostiole; ornamented spores 7 μm in diameter were also

Figure 3.75 Geastrum tepexensis showing central endoperidium surrounded by triangular-shaped segments of the exoperidium (Cenozoic). Bar  7.5 mm. (Courtesy S. R. S. Cevallos-Ferriz.)

found in the structure. Although the morphological features of the fossil make it difficult to place within a modern genus, this discovery does expand the geographic distribution of the Lycoperdales into the tropics during the Cenozoic. Extending the range further back into the Cretaceous is Geastroidea lobata, a compressed earthstar from Mongolia (Krassilov and Makulbekov, 2003). OTHER FUNGAL REMAINS Another group of fossil fungal remains includes the asexual stages (anamorphs) of Ascomycota and Basidiomycota; these have historically been called deuteromycetes, fungi imperfecti, conidial, or mitosporic fungi. Today most of these organisms are saprotrophs or weak parasites of terrestrial plants and a few aquatics. Since information on their sexual reproduction is incomplete, they are therefore placed in artificial groups, which are based on conidial characters, that is, asexual reproduction. Many of the small spores recovered from palynological samples represent conidiospores of these fungi. The Coelomycetes are an artificial group known from temperate and tropical regions today. One form from the Cretaceous of Japan is Archephoma cycadeoidellae (Watanabe et al., 1999). It consists of mature pycnidia up to 250 μm wide that contain numerous smooth, aseptate conidia. In another form the conidia are elliptical and divided by a dark septum.

98

Paleobotany: the biology and evolution of fossil plants

Other coelomycetes have been reported as epiphytes on dicot and grass leaves preserved in amber (Poinar, 2003). Although paleomycology represents a very old subdiscipline of paleobotany, for reasons noted earlier almost all of the studies to date have focused on the description and identification of the fungi. The study of fossil fungi does, however, provide the opportunity to examine several levels of biotic and abiotic interactions that extend beyond the description and classification of the organisms (Stubblefield and Taylor, 1988). Studies of fossil fungi may examine symptoms of the host as well as the fungi themselves, and can reveal different levels of interaction, as well as providing indirect evidence of coevolutionary relationships between fungi and land plants. In some instances these may be beneficial levels of biological cooperation (e.g., mycorrhizae, lichens), whereas in others they indicate parasitic and pathogenic associations. Following are several examples of such interactions that have been determined based on the fossil record of fungi. Fungal Life-History Strategies

SAPROTROPHISM One of the most obvious and essential activities of fungi today is the degradation of plant and animal tissue. Without this activity, life on Earth would cease. As a result of the carbon cycle, atmospheric CO2 is fixed into organic molecules in plants via photosynthesis. Then after the plant dies, it is ultimately degraded by fungi (and bacteria); this process in turn releases the CO2 back into the atmosphere. One of the oldest examples of saprophytism comes from Late Devonian specimens of the progymnosperm wood morphotaxon Callixylon (Stubblefield et al., 1985a). Inside many of the secondary xylem tracheids are numerous hyphae (FIG. 3.76), some branching and exhibiting intercalary swellings (FIG. 3.77). The surface of some filaments is smooth, whereas others have numerous rounded knobs. In addition to these specimens in the wood, some of the cells of the vascular rays contain spherical structures, which were interpreted as fungal resting spores or ergastic deposits. In some areas of the wood, the tracheid walls are characterized by erosion troughs, which represent areas that have been delignified by the enzymatic activities of the fungus. The patterns produced on the tracheid walls in Callixylon are similar to those formed by living basidiomycetous fungi responsible for white rot (Otjen and Blanchette, 1986). In this particular example, the fossil fungi cannot be identified with certainty, but features of the degradation process provide information about the level of fungus– plant interaction. The production of lignified, secondary tissues (wood) by vascular plants had evolved in some groups by the

Figure 3.76 Callixylon tracheid containing fungal hyphae

(Devonian). Bar  35 μm.

Givetian (Middle Devonian). Today, fungi continue to be the primary decomposers of lignin in the ecosystem. The presence of wood-rotting fungi in Callixylon, one of the oldest known trees (Meyer-Berthaud et al., 1999), provides compelling evidence that this important saprophytic association between fungi and vascular plants arose around the same time that wood development began. Fossil evidence of this association also provides a proxy record of the type of decay process and the basidiomycetes that were responsible for it. For example, white rot (degradation of both cellulose and lignin) and brown rot (degradation of only the cellulosic part of the wall) can be distinguished by observing tracheid cell walls in fossil wood. As noted earlier, wood-rotting fungi in several Paleozoic and Mesozoic woods from Gondwana produced symptoms of white pocket rot similar to those seen in extant wood. One of the interesting aspects of the fungi responsible for white pocket rot is their apparently long geologic history. The woody plants they infected in the Paleozoic and Mesozoic are long extinct, but the fungi appear to have produced the same features in

CHAPTER 3 fungi, bacteria, and lichens

99

Figure 3.78 Seedling of Picea baltica with Gonatobotryum piceae (arrow) at the base of the unfolding cotyledons (Eocene). Bar  1 mm. (Courtesy A. Schmidt.)

Figure 3.77 Septate hypha with terminal chlamydospore in

Callixylon wood (Devonian). Bar  12 μm.

wood for several hundred million years. Does this mean that the relationships between certain types of fungi and woody plants coevolved as the fungi continued to adapt to new plant hosts? Perhaps the enzymatic system responsible for selective wood degradation evolved several times or in multiple groups of fungi, or perhaps the microenvironment of wood is so similar in all gymnosperms that fungi, once adapted to this habitat, showed little change. It may be possible to answer questions such as these in the near future by using molecular or isotopic techniques applied to modern and fossil wood-rotting fungi. PARASITISM Simply stated, parasites live at the expense of other organisms (FIGS. 3.78, 3.79). Living fungi demonstrate a variety of interactions with their hosts, including those that obtain nourishment without causing death (biotrophs), those that absorb nutrients from dead tissues (necrotrophs), those that cause disease (pathogens), those in which the partners provide mutual benefits (mutualists), and a variety of intermediate levels of interaction. Of all the potential levels of

Figure 3.79 Gonatobotryum piceae, young conidiophores

showing apical conidiogenesis (Eocene). Bar  50 μm. (Courtesy A. Schmidt.)

interaction between fungi and other organisms, parasitism is perhaps the most difficult to demonstrate and distinguish from saprotrophism in the fossil record. Without clear evidence of some host response to the infection, this type of interaction appears similar to saprophytism and even mutualism when examined in fossils. In some instances, the plant host may show several different responses to fungal invasion, further complicating our understanding of specific interactions (Oliver and Ipcho, 2004). One of the best examples of host response in living plants involves the production of swellings or appositions on the inner surface of cell walls (FIG. 3.80). In extant plants, such structures form in response to the invasion of fungi and are regarded as a mechanism by the host to isolate the fungus from uninfected cells nearby. Examples of cell wall appositions in tissues heavily infected with fungal hyphae are known from Late Pennsylvanian coal ball material of Illinois (Stubblefield et al., 1984a). In the Rhynie chert plant, Nothia aphylla, three different fungal endophytes were found in the plant axes (FIGS. 3.81, 3.82)

100

Paleobotany: the biology and evolution of fossil plants

Figure 3.80 Section of Araucarioxylon wood showing wall appositions (arrows) (Triassic). Bar  50 μm.

Figure 3.81 Cortical cell in rhizoidal ridge of Nothia aphylla containing numerous fungal spores (Devonian). Bar  30 μm.

(Krings et al., 2007a). Parts of the rhizome have hypodermal cells with thickened walls (FIG. 3.83) (Krings et al., 2007a). Other parts show areas that are devoid of cells (FIGS. 3.84, 3.85), or include degraded cells, suggesting that the host may have responded to infection through programmed cell death. Both responses are seen in living plants and function to prevent further spreading of the parasite. The production of resinous material (FIG. 3.86) is another response to fungal invasion of living plant tissue. This type of host response has been noted in wood of Callixylon newberryi from the Upper Devonian and may be more widespread in fossil woods, but as yet underreported. Pathogenic fungi are believed to be responsible for the symptoms found in Late Triassic tree trunks in the Petrified Forest of Arizona (Creber and Ash, 1990) (FIG. 3.87). In cross section the fossil wood shows numerous tubes associated with

Figure 3.82 Nothia aphylla rhizoid with swollen area contain-

ing endophytic fungi (Devonian). Bar  60 μm.

Figure 3.83 Partial section of Nothia aphylla rhizoidal ridge and host response to fungal attack in the form of zigzag line (arrows) of secondarily thickened cell walls (Devonian). Bar  500 μm.

Figure 3.84 Nothia aphylla rhizoidal ridge axis showing rhizoids and space where tissue is lacking (Devonian). Bar  60 μm.

CHAPTER 3 fungi, bacteria, and lichens

101

Figure 3.85 Section of Nothia aphylla rhizoidal ridge showing

void in response to fungal attack (Devonian). Bar  100 μm.

Figure 3.87 Geoffrey Creber (left) and Sidney R. Ash.

Figure 3.86 Ray parenchyma cells of Callixylon newberryi containing globules (Devonian). Bar  25 μm. (From Stubblefield et al., 1985a.)

areas of disrupted cells. Similar symptoms are known in living trees and are associated with the extant fungi Heterobasidion and Armillaria. The large number of trees with similar symptoms at the same stratigraphic level suggests that perhaps the forest was attacked on a large scale, much like the action of modern Dutch Elm disease (Creber and Ash, 1990).

Another interesting host response which has been observed in fossils is hypertrophy, an abnormal increase in cell size. A charophyte alga (Chapter 4) from the Rhynie chert, Palaeonitella cranii, shows greatly enlarged cells in response to infection by the fossil chytrid Krispiromyces discoides (FIGS. 3.88, 3.89) (Taylor et al., 1992). The consistent presence of this chytrid embedded in the walls of the hypertrophied algal cells indicates that the alga was alive when infection took place, and that the abnormal increase in cell size represents a host response (FIG. 3.90). Interestingly, identical examples of hypertrophy have been reported in the extant charophyte Chara (Karling, 1928). Other forms of chytrid parasites are reported in and on various other fungi in the Rhynie chert (FIG. 3.91) (Hass et al., 1994). For example, chlamydospores associated with the AM fungus Glomites rhyniensis contain papillae extending from the inner surface of the spore wall (FIG. 3.92) which are identical to those produced in living glomeromycotan (FIG. 3.93) spores attacked by chytrids (Boyetchko and Tewari, 1991). Other spores have minute holes like those produced by certain actinomycetes (Lee and Koske, 1994). These examples of mycoparasitism (FIG. 3.94) underscore just a few of the microbial dynamics that occurred in and around the freshwater pools of the Rhynie ecosystem in the Early Devonian.

102

Paleobotany: the biology and evolution of fossil plants

Figure 3.88 Section of hypertrophoid cell of Palaeonitella with chytrid (Krispiromyces discoides) (arrow) penetrating cell wall (Devonian). Bar  80 μm.

Figure 3.90 Two hypertrophoid cells of Palaeonitella cranii. Arrow indicates normal cell size of axis (Devonian). Bar  275 μm.

3.89 Krispiromyces (Devonian). Bar  25 μm.

Figure

discoides

(see

FIG.

3.88)

Fungal parasitism of fossil plants can also be inferred based on particular characteristics of the fungus. For example, in certain epiphyllous fungi growing on middle Eocene (Paleogene) angiosperm leaves from the Claiborne Formation of Tennessee, haustoria and haustorial pores were found in many of the leaves (Dilcher, 1965). Although the possibility exists that these fungi obtained their nourishment

CHAPTER 3 fungi, bacteria, and lichens

103

Figure 3.91 Spore tightly packed with coenocytic hyphae.

Arrow indicates possible discharge papilla of fungus (Devonian). Bar  10 μm. (From Hass et al., 1994.) Figure 3.93 Chlamydospore of extant glomeromycotan fungus

showing host response in the form of wall papillae. Bar  100 μm.

Figure 3.92 Chytrid (arrow) on surface of chlamydospore.

Note host response in form of papilla inside spore wall (Devonian). Bar  20 μm.

from some other source (e.g., host leaf excretions or exudates from animals), the presence of haustoria like those of modern parasitic fungi, and the reaction of the leaf to this pattern of penetration, suggests that these fungi parasitized the angiosperms on which they grew. MUTUALISM Mutualistic interactions are those symbioses in which both partners benefit from the relationship. Most of those that have been reported in the fossil record involve mycorrhizal fungi

Figure 3.94 Parasitic or saprotrophic fungi inside glomeromy-

cotan spore (Devonian). Bar  80 μm.

and vascular plants; lichens are another example of a mutualistic association (discussed later). For endomycorrhizal fungi, the existence of such interrelationships in the fossil record has been based on the presence of non-septate hyphae

104

Paleobotany: the biology and evolution of fossil plants

and various forms of vesicles and chlamydospores within the underground rhizomes or aboveground prostrate axes of permineralized vascular plants. Since their initial description in tissues of the Rhynie chert plants, the occurrence of variously shaped spores and hyphae in these plants has been used as the basis for establishing the early occurrence of endomycorrhizal associations. Based on the presence of such structures, Pirozynski and Malloch (1975) hypothesized that such fungal–plant interactions were necessary for the establishment of plants on the land (Chapters 6, 8). According to these authors, the fungi in this mutualistic association would be provided with a carbon source, whereas the land plants would benefit from increased nutrients and more efficient water uptake from the substrate. Some of the chlamydospores and hyphae that are so common in the Rhynie chert plants may also represent the remains of saprophytic fungi (Taylor and White, 1989). Others, however, confirm the existence of arbuscular mycorrhizae based on well-defined arbuscules in a restricted zone of the cortical tissues (Taylor et al., 2005c) in both sporophytes and gametophytes (FIG. 3.95). Although a morphological structure

cannot be used to conclusively demonstrate a physiological function in a fossil plant, the direct correspondence between arbuscule formation in extant and fossil plants, including location within the host plant and morphology of the arbuscules (FIG. 3.96), is striking. Other forms of evidence, such as molecular sequence data calibrated to molecular clock assumptions, also support the existence of endomycorrhizal symbioses by the Early Devonian (Simon et al., 1993). It is hypothesized that one selective advantage of mycorrhizae is the ability to increase the plant’s uptake of phosphorus via the extended hyphal network. This certainly may have been an important attribute in what must have been a nutrientpoor substrate during the Early Devonian of the Rhynie site. Unfortunately the earliest land plants (Cooksonia-type organisms) are preserved as impressions or compressions, and so their mycorrhizal status is unknown. Molecular biology does suggest, however, that the Glomeromycota extend well back into the Paleozoic (Berbee and Taylor, 2001; Taylor and Berbee, 2006), and even into the Precambrian based on some accounts (Heckman et al., 2001). By conservative estimates this would mean that by the time of the Early Devonian Rhynie ecosystem, fungal relationships with plants were well established, and this is borne out by morphological and structural features of both the host and the fungus. Generally the underground organs of fossil plants, unless they are petrified or permineralized, are not studied in great detail. There is sufficient permineralized plant material available today, including roots, however, from many different geologic horizons and representing most of the major groups of plants, that a systematic study looking for endomycorrhizae

Figure 3.95 Hypha (arrow) extending through gametangiophore stalk of Lyonophyton rhyniensis (Devonian). Bar  35 μm.

Figure 3.96 Detail of arbuscules in Aglaophyton major corti-

cal cells (Devonian). Bar  30 μm.

CHAPTER 3 fungi, bacteria, and lichens

could be successfully undertaken. Although it is estimated that 90% of all living plants enter into some type of mycorrhizal symbiosis, there remains little information about the spatial and temporal distribution of mycorrhizae in geologic time. For example, we know a great deal about Carboniferous coal swamp plants. Were they all mycorrhizal, and if not, why not? What is the distribution of mycorrhizae in major groups of plants which are anatomically preserved, that is either as petrifactions or as permineralizations? These and many other questions can have important implications in deciphering the evolution of certain types of paleoecosystems, as well as tracking fungal-plant interactions through time.

105

1982, 1984). From Cenozoic amber collected in Mexico, fungi have even been described inside fossil nematodes (FIG. 3.98) (Poinar, 1984; Jansson and Poinar, 1986). Schmidt et al. (2007; 2008) have reported carnivorous fungi preserved in Cretaceous amber that possess hyphal rings

Fungi–Animal Interactions

Despite the obvious role fungi play in modern ecosystems as decomposers, parasites, and mutualists, the interactions between fungi and animals are not extensively documented in the fossil record. A common fungus–animal interaction known from the Paleozoic to the recent is evidence of borings made by endolithic fungi (also algae and bacteria) in calcium carbonate skeletons of marine invertebrates (Gatrall and Golubic, 1970; Grahn, 1981). This represents a trace fossil (or ichnofossil), in that no organic material of the fungus remains, only trace evidence in the form of damage to shells, and so on. Another early example of fungus–animal interactions involves middle Silurian fungal remains from Sweden (Sherwood-Pike and Gray, 1985). These rocks contain spindle-shaped aggregates of hyphae associated with amorphous material up to 260 μm long. The fungal remains are interpreted as frass (fecal pellets) produced by a microarthropod. Another explanation is that the arthropod was using the fecal remains as their primary food source. This example demonstrates the difficulty in determining nutritional modes and interactions in paleoecosystems. In some cases, the morphological similarity between a fossil and extant fungus can be used to infer the nutritional mode and degree of interaction. Geotrichites glaesarius represents a conidial fungus, that is, a fungus that forms external, asexual spores of a particular type (Stubblefield et al., 1985b). It was found on the surface of a partially decomposed abdomen of a spider (FIG. 3.97) preserved in late Oligocene–early Miocene amber from the Dominican Republic (Stubblefield et al., 1985b). The fact that the fungus had not invaded the body cavity of the spider and the pattern of conidial formation suggest that this interaction was saprotrophic. Two other reports of entomophthoralean (Zygomycota) fungi from the Dominican amber include a parasitic infection of a winged termite and a fungus, found on a fossil ant, which resembles modern members of an insect pathogen (Poinar and Thomas,

Figure 3.97 Chains of conidia attached to a spider leg preserved in Dominican amber (Miocene). Bar  50 μm.

3.98 Nematode containing fungus (Oligocene). Bar  15 μm. (From Jansson and Poinar, 1986.)

Figure

106

Paleobotany: the biology and evolution of fossil plants

that serve to trap nematodes (FIGS. 3.99–3.101). Associated with the fungi are small nematodes that probably represented the prey. Modern carnivorous fungi are found in the Zygomycota, Ascomycota, and Basidiomycota, and the presence of these forms in amber indicates that complex devices to trap motile organisms had evolved by the early Cretaceous. Although these reports are important in documenting cases of specific interactions in the fossil record, there are still too few reports currently available from older rocks to make any substantive comment regarding the evolution of these complex interactions. The associations between fungi and animals are perhaps nowhere more unusual than those known from the

Trichomycetes. Today, trichomycetes inhabit the lower digestive tracts of various types of insects and other arthropods, and based on molecular sequences, it is suggested that many groups are polyphyletic (White, 2006). These endosymbiotic microfungi live in freshwater and include more than 130 species; however, this probably represents a fraction of the total number of living forms. In this obligate mutualistic association, the fungi are not capable of existing outside the host gut. The only fossil trichomycete known to date is a specimen from the Triassic of Antarctica (White and Taylor, 1989b). It consists of a small fragment of presumed arthropod cuticle to which are attached numerous, elongate thalli (FIG. 3.102), each anchored by a holdfast cell. At the distal end of each

Figure 3.101 Reconstruction of the carnivorous (nematode

trapping) fungus (Cretaceous). (From Schmidt et al., 2007.) Figure 3.99 Carnivorous fungus, trapping ring (Cretaceous).

Bar  10 μm. (Courtesy A. Schmidt.)

Figure 3.100 Carnivorous fungus, yeast-like growth forming along a hypha (Cretaceous). Bar  10 μm. (Courtesy A. Schmidt.)

Figure 3.102 Palisade organization of thalli on the inner sur-

face of putative arthropod cuticle (Triassic). Bar  100 μm.

CHAPTER 3 fungi, bacteria, and lichens

thallus is a small plug and numerous spores. Although the diagnostic features of the cuticle necessary to identify the host are not present, the occurrence of numerous fecal pellets associated with plant tissues from the same site substantiates the existence of arthropods in this Triassic ecosystem (Kellogg and E. Taylor, 2004). Fungal remains have also been documented in a variety of coprolites (FIG. 3.103) and have been especially useful in Quaternary palynology (Davis, 2006 and references therein). In some examples the producers of the coprolites are believed to have been saprobes dwelling on all sorts of organic particles, whereas in other instances a high percentage of the coprolites contain fungal remains, suggesting that the producers were true fungivores (Pratt et al., 1978). Fungi have been reported in various types of dinosaur coprolites, and the presence of certain types of leaf-borne fungi has been used to infer a foliage diet for these animals (Sharma et al., 2005). An interesting report of partially decayed wood found in coprolites of herbivorous dinosaurs (Chin, 2007) suggested that the animals may have eaten wood that had been partially rotted by fungi when there were few other food sources available. The wood would only be useful as a nutrient source after fungi had begun to break down the lignin and cellulose within it. The report and description of fungi associated in other biotic interactions (e.g., additional types of coprolites and a variety of different types of borings) in Miocene wood is an important contribution that will help to underscore the geological history of fungal-animal interactions (Sutherland, 2003). An unusual interaction between fungi and animals has been documented from the Jurassic (Martill, 1989). On the surface of fossil fish teeth are a series of meandering

Figure 3.103 Coprolite composed of hyphae (Devonian).

Bar  50 μm.

107

borings that are restricted to the surface enamel. These have been given the name Mycelites enameloides. Similar borings have been reported on teeth, scales, and bone as early as the Devonian (Gouget and Locquin, 1979). The report of fungal–animal interactions and associations in the fossil record is a very important area in paleomycology. Although many of the associations may be the activities of fungal saprotrophs, others may represent early stages of symbiotic interrelationships that are widespread in modern ecosystems, and which represent one of the cornerstones of biodiversity today (Zook, 2002). Geologic Activities of Fungi

Throughout the geologic past, fungi have played an important role in their interactions with the depositional environment, and together with cyanobacteria, are today receiving increasing attention because of their role in shaping the geosphere. The discipline of geomicrobiology involves the study of organisms, their interactions, and the materials that they colonize (Konhauser, 2007). Certain fungi are known to degrade and liquefy coal (Sterflinger, 2000), and they are the only organisms that can completely break down lignin. Fungi in mycorrhizal and lichen symbioses are also involved in weathering rock, a form of biological weathering (Landeweert et al., 2001; Hoffland et al., 2004). Fungi in very cold or desert environments appear to be important as rock colonizers (Selbmann et al., 2005). Fungi are involved in the alteration of several minerals, such as carbonates and silicates; for example, fungi can be involved in both the formation and destruction of carbonate deposits (Sterflinger, 2000). The roles of fungi in geological activities are only beginning to be fully appreciated as the focus in the past has been mostly on bacteria (Gadd, 2007; Perri and Tucker, 2007). One type of fungal activity seen in the fossil record includes borings in rock. Since there are several groups of organisms, however, that are borers, including cyanobacteria, algae, lichens, sponges, bryozoans, gastropods, and several other invertebrate groups, distinguishing the organism responsible for the boring is often very difficult (Elias and Lee, 1993). In one study on endolithic fungi from Jurassic rocks, scanning electron microscopy was used after resins casts were made of the molds (Gatrall and Golubic, 1970). Based on comparison with extant forms, it was determined that these borings were the result of fungi, although many earlier contributions had regarded such trace fossils as the result of algal borers. Fungal and algal borers may also be involved in a process termed micritization, in which carbonate particles are reduced in size and combined with chemically precipitated

108

Paleobotany: the biology and evolution of fossil plants

carbonate to produce a finely crystalline form of carbonate, also called calcite mud. Biogenic micritization, which is mediated by organisms, has been noted in rocks dating back to the Paleozoic; in a few cases the filaments of the endolithic borers are preserved in the rocks, but in most cases their presence can only be inferred (Harris et al., 1997). Endolithic filaments are known from the Upper Silurian of Morocco (Barbieri et al., 2004). Although fungi are involved in some of these structures, many are also mediated by various types of microorganisms (cyanobacteria and algae), which produce micrites as byproducts of their metabolism. In some instances these structures demonstrate a faint banding pattern that suggests they may have been produced during different seasons. Fungi have also been important in the geologic record through their activities in biomineralization—the synthesis of minerals from simple compounds by organisms. For example, the occurrence of needle-fiber calcite in Carboniferous paleosols (fossil soils) has been regarded as evidence of ectomycorrhizal associations in the ecosystem (Wright, 1986). Certain extant vascular plant roots accumulate secondary soil carbonate through biomineralization. Some of these roots have larger diameters because of calcification of cortical parenchyma cells. It is suggested that carbonate biomineralization and acid extrusion in these root cells may represent a mechanism used by plants to acquire soluble nutrients from the rhizosphere, or a method to live in soils with excessive calcium (Kosir, 2006). As paleobotanical data are further integrated with sedimentology, it may be possible to document such an adaptive feature in the roots of certain fossil plants.

been overlooked. If this is not the case, perhaps the plants possessed chemical substances that deterred epiphyllous fungi, or fungi simply may not have evolved an epiphyllous habitat by the Paleozoic. Stomiopeltites is perhaps the oldest (Early Cretaceous), well-documented member of the Dothideales (Pezizomycotina, Ascomycota), a group that includes numerous epiphyllous taxa (Alvin and Muir, 1970). It consists of dome-shaped plectenchymatous thyrothecia (a type of ascocarp with a palisadelike layer of hyphae), each bearing a small ostiole. Hyphae are small (1.7–3 μm in diameter) and form the layered wall of the thyrothecia. Pycnidia (asexual reproductive structures) are common and appear to have been borne directly over a stoma on the leaf surface. No spores were described in these specimens. This fungus is unique, not only because of its age but also because it is one of the few fossil epiphyllous fungi known to occur on conifer leaves (Frenelopsis). Stomiopeltites (FIG. 3.104) is also known from the Upper Cretaceous (Cenomanian) of France, also on Frenelopsis leaves (Pons and Boureau, 1977) (FIG. 3.105), and from the Miocene Clarkia locality of Idaho, USA, on leaves of three angiosperm genera (Phipps and Rember, 2004). Numerous structurally preserved ascomycetes have been described from Cretaceous conifer leaves of the Brachyphyllum-type (Van der Ham and Dortangs, 2005). The fossil fungi (FIGS. 3.106, 3.107) are thought to be

Epiphyllous Fungi

Mesozoic and younger rocks contain numerous examples of ascomycetous fungi (Taylor, 1994), many of them preserved on compressed leaves. Epiphyllous fungi are commonly encountered on many forms of vegetation as early as the Cretaceous, and probably far earlier, although few good examples exist. Despite the tremendous diversity of foliage types during the Carboniferous, for example, there are very few reports of fungi associated with the leaves. One example is reported in cuticle preparations of rachides and pinnules of Callipteridium, a pteridosperm from the Stephanian of France (Krings, 2001). The fungus consists of septate hyphae, but no reproductive structures were found. In preservation of this type, it is especially difficult to determine if the fungus was associated with the plant when it was alive, or simply represents a saprotroph that colonized decaying plant remains. As is true for many facets of paleomycology, epiphyllous fungi in the Paleozoic may exist, but have simply

Figure 3.104 Epiphyllous fungus Stomiopeltites amorphos (Miocene). Bar  50 μm. (Courtesy of C. J. Phipps.)

CHAPTER 3 fungi, bacteria, and lichens

109

Figure 3.107 Pteropus brachyphylli hypostroma (Cretaceous).

Bar  40 μm. (From Van der Ham, 2005.)

Figure 3.105

Denise Pons.

Figure 3.106 Pteropus brachyphylli fruiting body (Cretaceous).

Bar  100 μm. (From Van der Ham et al., 2005.)

closely related to the extant Phaeocryptopus, which infects the leaves of a number of extant conifers, thus suggesting specific parasitic relationships with certain conifers since the Late Cretaceous. Trichopeltinites is a microthyriaceous (Dothideales) fungus that can be traced back to the latest Cretaceous (Maastrichtian). Specimens from rocks of the Cretaceous–Paleogene boundary in Alberta, Canada, suggest that this fungus can be correlated with the accumulation of organic matter (coal) under wet conditions, rather than with a particular paleolatitude (Sweet and Kalgutkar, 1989). Two interesting ascomycetes were found within the permineralized leaf of an extinct palm from the Eocene Princeton chert (Currah et al., 1998). Palaeoserenomyces allenbyensis is compared to modern Serenomyces, which forms leaf spots (so-called tar spots) on palms. Within the locules of the stromata of Palaeoserenomyces are globose ascomata of a mycoparasite, Cryptodidymosphaerites princetonensis. This mycoparasite appears similar to the extant form, Didymosphaeria, which today is a mycoparasite on ascomycetes that form stromata (a type of pseudoparenchymatous fruiting body). The Vizellaceae (Dothideales) is a small family of epiphyllous ascomycetous fungi that can be traced into the Eocene. It is one of only a few families of fungi that can be identified in the fossil record solely on the basis of non-reproductive (vegetative) features, as it has distinctive septate hyphae with regularly alternating dark and hyaline segments (Phipps, 2007). One of the most comprehensive studies of fossil epiphyllous fungi is the contribution by Dilcher (1965), who described a large number of fungi found in association with

110

Paleobotany: the biology and evolution of fossil plants

middle Eocene angiosperm leaves, including several members of the Vizellaceae. In several instances it was possible to reconstruct the entire life history of the fungus, including both the sexual (telemorphic) and asexual (anamorphic) phases, thereby making assignment to modern groups even more reliable. One of the more common forms in this biota is Vizella memorabilis, which is characterized by hyphae with alternating short and long cells (Selkirk, 1972). In the Eocene clay pit (Claiborne Formation) where this fungus was collected, it is most commonly found on the upper surface of angiosperm leaves of the genus Sapindus. During the growth of the fungus the hyphae dichotomize to form colonies that may be up to 450 μm in diameter. Irregular proliferations of cells form a short, lateral branch that produces two types of fruiting bodies. The larger fruiting bodies (ascocarps) produce two-celled spores, whereas the smaller ones (pycnidia) produce single-celled spores. Up to 60 ascospores are produced in each ascocarp, which are 10  14 μm. Pycnidiospores are oval, unicellular, and oriented in chains. Upon germination, the pycnidiospores form small germinal tubes that penetrate the leaf epidermis. Fossil fungi of this general type, but which lack spores, are often assigned to the genus Entopeltacites (FIG. 3.108) (Phipps, 2007). Epiphyllous fungi of the Vizella type are morphologically identical with the modern genus Manginula, a form that was included within the Microthyriales. In addition to determining stages in the life history of several of these Eocene epiphyllous fungi, Dilcher was able to document various levels of host specificity. For example, V. memorabilis was found on a number of different types of leaves, whereas others, such as Meliola anfracta, appeared consistently only on certain leaf genera. In other instances, fungi appeared to be specific for only one surface of the leaf. Determining host specificity for particular types of fossil fungi is a difficult and long-term endeavor that will require careful documentation of both the fungus and host plant from many localities and stratigraphic levels (Venkatachala and Kar, 1968). Nevertheless, it represents an important line of research that can provide new data on the evolution and variety of fungal interactions in the fossil record. Identification of the spores produced by certain types of Eocene fungi is also important, as it provides a method of correlating sediments where fruiting bodies are not preserved (Sheffy and Dilcher, 1971). A rich assortment of fossil fungi in a flora, such as the ones described from the Eocene of Tennessee, represents an important potential source of information on the paleoclimate of a region (SherwoodPike, 1988). Modern microthyriaceous fungi are principally tropical in distribution. The most important limiting factor

Figure 3.108 Epiphyllous fungal hyphae of Entopeltacites

remberi (Miocene). Bar  50 μm. (Courtesy C. J. Phipps.)

in their geographic distribution, however, appears to be precipitation, not temperature, and these fungi can be correlated with regions where mean annual precipitation reaches 100 cm/yr (Elsik, 1978). On the basis of leaf types present at the Tennessee site, it has been suggested that the region was subtropical during the middle Eocene. A comparison of the fossil fungi with the ranges of comparable modern forms (the Nearest Living Relative method applied to fungi) supports the theory that this area was a low-lying, coastal region characterized by a moist, subtropical climate. It may be possible in other fossil floras to utilize various forms of fungi to provide more accurate information on paleoclimatic regimes (Lange, 1978), especially in those instances where modern fungi can be used as analogs for microhabitat determination. Another fungus that has been used to infer paleoclimatic information is Meliolinites dilcheri (Daghlian, 1978a). This form, from the early Eocene of Texas (Rockdale Formation), consists of colonies 2 mm in diameter (FIG. 3.109). The hyphae (FIG. 3.109) produced numerous short, lateral, twocelled branches consisting of a short stalk and a capitate head. The fossil fungus is included within the Meliolaceae, the modern members of which are parasitic and common in warm, humid, forested tropical areas today. Phragmothyrites (FIGS. 3.110, 3.111) is another fossil epiphyllous fungus; it consists of ascomata with radiating, pseudoparenchymatous hyphae (Phipps and Rember, 2004). Although there are numerous reports of fossil epiphyllous fungi on Cenozoic leaves, the geologic antiquity of this plant-fungal interaction remains relatively unexplored. The increasing attention to documenting fossil fungi over wide geographic areas, however, is a positive step (Herbst and Lutz, 2001). It has been proposed that leaf-associated

CHAPTER 3 fungi, bacteria, and lichens

111

Figure 3.109 Colony of the epiphyllous fungus Meliolinites dilcheri with radiating hyphae and hyphopodia on cuticle of Eocene angiosperm leaf. Bar  50 μm. (Courtesy C. P. Daghlian.)

Figure 3.111 Detail of epiphyllous fungus Phragmothyrites

concentricus (Miocene). Bar  50 μm. (Courtesy C. J. Phipps.)

Another interesting approach that has been used to identify the presence of epiphyllous fungi in fossils is based on the analysis (gas chromatography–mass spectrometry) of lipid fractions extracted from uninfected and infected conifer twigs of Cretaceous age (Nguyen et al., 2000). This study suggests that it is possible to determine lignin degradation products in the fossils produced by the fungi even when the fungi themselves are not present. Fungal Spores Figure 3.110 Several Phragmothyrites concentricus thalli

(Miocene). Bar  25 μm. (Courtesy C. J. Phipps.)

fungi did not arise until the Cretaceous, perhaps in association with the early radiation of flowering plants (Pirozynski, 1976b). The presence of laminar leaves as early as the Middle Devonian, however, indicates that there was a phylloplane microhabitat available for fungi much earlier. In addition, the discoveries of ascomycete-like fungi as early as the Silurian (Sherwood-Pike and Gray, 1985) provide evidence that this group may have evolved far earlier than we previously thought. Perhaps the early relatives of leafinhabiting fungi occupied other types of habitats. It is also possible that fungi became associated with leaf surfaces far earlier than the Cretaceous, but they have not been found. There is some evidence that the methods used to separate fossil leaf cuticles from the rock and prepare them for study may be destroying epiphyllous fungi.

Although fungal spores are known throughout geologic time, it is only beginning in the Late Jurassic that they constitute a significant and important fraction of the palynomorphs recovered from most rocks (Elsik, 1976). Spores of obvious fungal origin, however, are known as early as the Late Silurian. There is an extensive terminology applied to fungal palynomorphs based on living fungi (Elsik et al., 1983). Spores are often named based on the concept of morphogenera (Elsik, 1989), in which various morphologic features (e.g., size, shape, number of septations or pores, and type of ornament) form the basis of the generic concept (e.g., Fusiformisporites). In addition to spores, fungal fragments recovered in macerations may include various forms of fructifications, hyphae, and mycelia. Funginite is the name applied to sedimentary organic matter (macerals) composed chiefly of fungal material (see Chapter 1). The study of spores and other fungal remains represents an important, yet still largely untapped, paleomycological

112

Paleobotany: the biology and evolution of fossil plants

and paleobotanical data source. Although fungal spores have been utilized in some stratigraphic applications in Cenozoic rocks, especially around the Cretaceous–Paleogene boundary (Vajda and McLoughlin, 2004), perhaps their greatest untapped potential lies in the areas of biogeography, paleoecology, and paleoclimatology (Sherwood-Pike, 1988). Jarzen and Elsik (1986) examined fungal palynomorphs from a modern, subtropical environment of open savanna and riparian forest in eastern Zambia. The results of that study provide a basis for quantifying the various fungal remains preserved in modern fluvial systems, with applications to similar environments in the fossil record. It also suggests another method for inferring paleoecological parameters when other microfossils and megafossils are lacking.

Figure 3.112 Albugo-like oogonia in Pennsylvanian seed tissues. Bar  100 μm.

Fungal-Like Organisms

There are a number of organisms that historically were classified with the fungi, but as a result of subsequent research, and especially molecular systematic approaches to modern forms, have been removed from the Kingdom Fungi. Nevertheless, each of these organisms has a phylogenetic history, and some are preserved as fossils. PERONOSPOROMYCETES (OOMYCOTA) The oomycetes, or water molds, are organisms that are now included within the Kingdom Straminipila; some authors refer to this group as the Peronosporomycetes. Most forms are filamentous, possess cell walls of cellulose rather than chitin, and lack cross walls, except where reproductive cells are produced. Asexual reproduction occurs by biflagellated zoospores. Sexual reproduction includes the production of sperm in antheridia which fuse with one to several eggs produced in an oogonium. Although there are several pre-Cenozoic reports of peronosporomycetes in association with fossil plants, many of these have been discounted because of the absence of a complete suite of diagnostic features (Johnson et al., 2002), whereas others regard at least some of these fossils as authentic (Blackwell and Powell, 2000). One interesting fossil that perhaps provides the best evidence of peronosporomycetes is Hassiella monospora (T. Taylor et al., 2006). This Rhynie chert organism consists of aseptate hyphae that randomly branch to form terminal oogonia, each approximately 30 μm in diameter. At the base of the oogonium is a funnelshaped structure interpreted as the antheridium. Intracellular peronosporomycetes (Combresomyces cornifer) in lycopod periderm have also been described and illustrated from the upper Visean cherts of central France (Krings et al., 2007e; Dotzler et al., 2008). Another possible peronosporomycete is

found in a seed-like structure from the Carboniferous (Stidd and Cosentino, 1975). Each spherical oogonium (100 μm in diameter) contains a single oosphere (FIG. 3.112); however, no well-defined antheridia are preserved. What makes this specimen so interesting and increases the chances that it represents a peronosporomycete is that the disruption of the tissues in the seed are identical to those found when the extant peronosporomycete Albugo infects flowering plants.

Eubacteria and archaea Despite the microscopic size of the cells (only a few micrometers long), in terms of numbers of individuals and metabolic diversity, bacteria are the dominant organisms of the biosphere and, geologically, the oldest organisms on Earth (Chapter 2). They live in every possible habitat—soil, water, and even the deep subsurface of the continents and oceans; they can even be found in radioactive waste. They are critical in nutrient recycling (Nardi et al., 2002), including the fixation of nitrogen from the atmosphere into a form that is usable by plants. It has been estimated that there are from 40 million to 2 billion bacteria in a single gram of soil (Whitman et al., 1998). Bacteria are classified today based on genetic sequence data. In older classification systems, five kingdoms of organisms were recognized: Protista, Monera (all prokaryotes), Fungi, Plantae, and Animalia. Although modern phylogenetic work has long since shown that this system is out of date for the prokaryotes, it is still reproduced in some textbooks. The most widely used classification of organisms today includes three domains (a level above the level of

CHAPTER 3 fungi, bacteria, and lichens

Kingdom): Eubacteria, Archaea, and Eucarya (Woese and Fox, 1977; Woese et al., 1990). Within the Domain Eubacteria are the Cyanobacteria (including the Chloroxybacteria), and other bacteria (e.g., purple, green sulfur, Gram-positive, spirochetes), with each group occupying the level of kingdom. All possess prokaryotic cells, which differ from the eukaryotic cells of plants, animals, and fungi. Prokaryotic cells are small (0.5–10 μm), and lack a nucleus or other membrane-bound organelles. Neither meiosis nor mitosis occurs; prokaryotes reproduce asexually by means of binary fission. All the nutritional modes used by eukaryotic cells are also found among the prokaryotes, plus several that are unique to the group. Despite the fact that bacteria are very small (0.5–5 μm) and delicate, there is an excellent fossil record of the group dating back to the Archean, with structurally preserved specimens reported as old as 3.7 Ga. All morphologic forms (coccoid, bacilloid or rod-shaped, spiral, and filamentous) have been recognized in a variety of configurations that correspond to those of existing types. They occur as structurally preserved and compressed forms within most types of common mineral matrices and have also been found in fossil vertebrates, invertebrates, plant tissues, and coprolites (fossil fecal material). Archaea

The application of molecular tools has been useful in establishing relationships and the divergence of major groups within the prokaryotes. Using 16S rRNA phylogenetic trees, Woese and Fox (1977) identified a group of prokaryotes called the Archaea. Although they had many features in common with other prokaryotes, they differed in certain genetic processes, and are generally considered to be more closely related to the Eucarya than to the Eubacteria. Today, many archeans live in some of the most severe environments on Earth and are termed extremophiles. They can be found in hot springs, deep within the ocean, in the digestive tracts of animals, and a variety of other harsh environments. Archeans include halophiles (organisms that live in high concentrations of salinity), thermophiles (live in hot and acidic environments), and methanogens (utilize hydrogen [H2] to reduce CO2 to methane). Many believe that the Archaea evolved from single-celled organisms 4 billion years ago. Many of the environments they inhabit today are believed to be the types of conditions present in the first billion years of Earth’s history, and it has been suggested that life may have arisen in these extreme habitats (Leach et al., 2006) (Chapter 2). As a result of ribosomal sequences of modern halobacteria, some suggest that these are the most primitive of the

113

archean group. Although no body fossils of Archaea have been confirmed to date, there is evidence of biomarkers from archeans in the fossil record. Perhaps the largest obstacle is distinguishing fossil from living Archaea, since members of these group are widespread in the deep biosphere. Eubacteria

It is currently not possible to determine which of the bacterial groups is the oldest geologically. Of the earliest microfossils known to date, all are morphologically simple and at least some were probably phototactic. Early microbiotas contained both anaerobes and photoautotrophs (Knoll, 1985a). In 1966, Barghoorn and Schopf described small rod-shaped cells from the Onverwacht Group of South Africa (3.5 Ga ago) under the name Eobacterium. Although originally thought to represent rod-shaped bacteria based on transmission electron microscopy (TEM), these structures are now regarded as non-fossil or modern bacterial contaminants (Schopf and Walter, 1983). Research with modern bacteria growing in hot springs environments is providing information about how bacteria are preserved, which in turn will aid paleobotanists in their interpretation of fossil microbes (Phoenix et al., 2005), as many ancient hot springs environments resulted in structurally preserved bacterial fossils. Bacteria from the Early Devonian Rhynie chert hot springs environment have been described as both spherical clusters and as unicells. Some of these structures appear similar to bacteria-like organisms (BLOs) that have been described inside the spores of some modern mycorrhizal fungi (Bonfante, 2003; Taylor et al., 2004; Duponnois and Kisa, 2006). Bacteria have also been reported in permineralized plants and coprolites of Carboniferous and Permian age (Renault, 1901) as circularto-elongate structures 4 μm in diameter (FIG. 3.113). These were named Micrococcus and Bacillus, but some of them represent inorganic particles of various types (Taylor and Krings, 2005). Fossil iron bacteria have been found in pyrite from Middle Pennsylvanian rocks and examined by means of replication techniques for TEM (Schopf et al., 1965). The morphological similarity between the fossil and extant types resulted in some of these fossils being assigned to modern taxa. Bacteria have also been suggested as the causal agent for various borings in vesicles of Ordovician chitinozoans (Grahn, 1981). In modern ecosystems, a number of forms of saprophytism involve bacterial decay in association with the activities of fungi (Daniel et al., 1987). Earlier in this chapter, it was noted that saprotrophic fungi have been identified as early as the Devonian, and perhaps earlier. To date the bacteria that are typically associated with certain types of

114

Paleobotany: the biology and evolution of fossil plants

Figure 3.114 Annella capitata (white spheres) on surface of pollen grain. Arrow indicates a cavity caused by a bacterium (Pennsylvanian). Bar  5 μm. (From Srivastava, 1976a.)

Bacterial colony. (Mississippian). Bar  10 μm. (Courtesy N. Dotzler.)

Figure 3.113

wood rot have not been identified either directly or indirectly in these fossil woods. Fossil woods, however, offer potential for future research on both the bacteria and their interrelationships with fungi during the decay process. Actinomycetes are a type of filamentous, Gram-positive bacteria. Actinomycetes have been reported from Eocene amber (Waggoner, 1994), a Carboniferous seed plant cell (Smoot and Taylor, 1983a), and Eocene dicot wood (Wilkinson, 2003). Although some of these reports suggest that the fossils are actinomycetes based on morphological features, some may represent modern contaminants or coagulated cytoplasm. One method that has been used to indicate that these organisms are fossils is the presence of the filaments growing through the crystals in the wood cells (Wilkinson, 2003). Leptotrichites is a sheathed bacterium from Cretaceous amber that resembles the modern genus Leptothrix (Schmidt and Schäfer, 2005). The fossil forms are interpreted as living in ponds in forest woodlands and became entrapped in amber flowing from the trees. Various forms of bacteria, including forms suggestive of sulfate-reducing bacteria and several types associated with decomposition, have been reported from the Eocene Green River Formation (Mason, 2005). Another source of information about fossil bacteria includes indirect evidence based on the presence of characteristic patterns or scars on sporomorphs. Some of the

microbes were extremely specific, attacking only certain layers of the pollen grain wall (Elsik, 1970). Such occurrences have been documented as early as the Carboniferous, where coccoid and bacilloid microbes were found on the surface of miospores (FIG. 3.114). Instances of fossil bacterial degradation patterns are probably far more common than reported, since palynologists typically search out only well-preserved sporomorphs showing a complement of diagnostic features. This inherent bias toward well-preserved specimens probably eliminates the documentation of more examples of microbial degradation. However, with organisms that are as small as prokaryotes, there is always the possibility of ascribing biotic properties to abiotic artifacts. A case in point involves the description of unmineralized fossil (Cretaceous) bacteria associated with scraps of organic material in lake muds. Upon reexamination, these bacteria turned out to be fluorite artifacts formed during the maceration process (Bradley, 1968). There are other potential sources of confusion regarding the authenticity of fossil bacteria, for example tapetal cell components, such as orbicules produced during microsporogenesis in certain plants, have been mistaken for coccoid bacteria. Due to the ubiquity of bacteria in the modern world, there is always the problem of specimen contamination. For example, it has been reported that fossil bacteria could be isolated and grown from Permian salt deposits (250 Ma) (Dombrowski, 1963; Vreeland et al., 2000; Satterfield et al., 2005). Others suggest that the organisms represent modern contaminants formed during the recrystallization of the

CHAPTER 3 fungi, bacteria, and lichens

salt. It has been noted by several authors that DNA decays quickly, and that no metabolic processes of any organisms could survive for so long. Several analyses have suggested that these bacteria are too similar genetically to modern taxa, so evolution would have to be extremely slow in these organisms (Nickle et al., 2002) and that bacterial spores, which can survive for long periods of time, generally have no DNA repair enzymes within them, so DNA this ancient would have been very fragmented (Graur and Pupko, 2001). The problem of modern bacterial contamination and the formation of bacteria-like artifacts during certain types of paleobotanical techniques is discussed by Edwards et al. (2006a). These studies elegantly underscore that, in dealing with fossil organisms, the pervasive distribution of microbes now and in the past can impact research results at several levels of inquiry. The complexity of the fossilization process and subsequent diagenesis, the procedures used in sample extraction and preparation, and the establishment of biogenicity all must be considered in the analysis of microbial fossils. For example, rod-shaped bacteria, together with certain types of organic compounds (polycyclic aromatic hydrocarbons), were described from a Martian meteorite collected in Antarctica (McKay et al., 1996). Subsequent research has suggested that the “microbes” are in fact the result of certain preparation techniques, and/or contaminants from melt water. What does the future hold for the study of extraterrestrial life based on fossil evidence? One approach will require the cataloging of various types of biosignatures in all forms of sediments, including how modern microbes can change the potential signature (Cady et al., 2003). CYANOBACTERIA The cyanobacteria are the most common and widespread group of photosynthetic bacteria today, and are the primary producers and initial source of free atmospheric oxygen. Their fossil record is among the oldest for any group of organisms, and can be traced back to the Archean (Golubic and Seong-Joo, 1999; Chapter 2). Combining fossil evidence with molecular data of living cyanobacteria has made it possible to hypothesize that cell differentiation (e.g., heterocysts and akinetes) occurred in this group as early as 2.450 Ga (Tomitani et al., 2006). Heterocysts are thick-walled cells where anaerobic nitrogen fixation occurs and akinetes are resting spores. What is especially interesting is that, when modern and fossil cyanobacterial communities are compared, the data suggest that there has been relatively little morphological and probably biochemical change during more than 2 billion years of Earth history (Knoll, 1985a; Sergeev et al., 2002).

115

Among the other bacteria, the Cyanobacteria appear to be most closely related to Gram-positive bacteria, based on molecular signatures. Cyanobacteria obtain their energy through photosynthesis and occur as unicellular, colonial and filamentous forms. They all contain chlorophyll a and accessory pigments in the form of phycobilins. Most cyanobacteria possess a mucilaginous sheath, which may be variously pigmented. Many of these ancient organisms have the ability to fix atmospheric nitrogen and, combined with their photosynthetic abilities, are thus the most nutritionally autonomous organisms on the Earth. Some extensive petroleum deposits in the world were produced by the decay and accumulation of various cyanobacteria such as Gloeocapsomorpha, a form responsible for some of the Middle Ordovician oil shales of Estonia (Foster et al., 1990) and the central United States (Iowa) (Jacobson et al., 1988). Although present throughout geologic time, cyanobacteria are a dominant component of many Precambrian biotas discussed in Chapter 2; a few additional examples of geologically younger cyanobacteria will be presented here. One indirect method of determining the activities of bacteria in the geologic record involves the identification of certain bacteriogenic isotopic sulfides (δ34S). The geochemical analysis of certain Archean rocks indicates that sulfate has been continuously present since at least 3.5 Ga (Schidlowski, 1989). The identification of the heavy sulfur isotope has also been proposed as a useful tool in determining past environmental conditions and as a stratigraphic marker at the Neoproterozoic–Cambrian boundary (Schröder et al., 2004). However, bacteriogenic sulfide patterns are also regarded as difficult to identify earlier than about 2.8 Ga. As a result, it is currently impossible to determine whether some of the geologically early sulfides were the result of sulfatereducing organisms or were inorganically formed, that is they are abiotic in origin. Activities of cyanobacteria in the fossil record can also be examined by the presence of various biomarkers in the sediments. Molecular fossils in the form of biological lipids have been recovered from 2.7 Ga shales in Australia (Brocks et al., 1999), and in slightly younger rocks from Canada. Akinetes have been reported from 2.1 Ga cherts from Gabon, Africa (Amard and Bertrand-Sarfati, 1997), and slightly younger ones from Siberian cherts that are morphologically identical to those produced by the living cyanobacterium Anabaena. Well-preserved cyanobacteria have also been described from the Lower Devonian Rhynie chert, both coccoid, for example Rhyniococcus (Edwards and Lyon, 1983), and filamentous forms, including the morphogenera Archaeothrix, Kidstoniella, Langiella, Rhyniella (Croft and George, 1958),

116

Paleobotany: the biology and evolution of fossil plants

Figure 3.115 Elongated tuft-like colonies of Croftalania venusta (Devonian). Bar  500 μm.

Figure 3.117 Spiny spheres (probably peronosporomycete

oogonia) that are associated with the microbial mats formed of Croftalania venusta. Bar  25 μm.

Figure 3.116 Microbial mat formed by Croftalania venusta

(Devonian). Bar  500 μm.

and Croftalania (Krings et al., 2007c). Langiella scourfieldi is a heterotrichous form consisting of a horseshoe-shaped basal portion that produced a number of uniseriate branches. At the base of each filament is a heterocyst which is typically smaller than the other cells of the sheath; at the distal end of each filament is a hair-like extension. Kidstoniella fritschii, from the same locality, contains highly branched filaments. Croftalania venusta is a sessile filamentous form that grows on sediment and submerged plant parts (FIG. 3.115). It is associated with the formation of microbial mats (FIG. 3.116), but may also occur in structured colonies where the individual filaments are aligned into flat, irregular stands or united radially into hemispherical aggregates; it may also form fan-shaped tufts. Associated with some of the mats

are spiny spheres that may represent peronosporomycete oogonia or some other microbial reproductive structure (FIG. 3.117). The Rhynie chert also provides the earliest fossil evidence for endophytic cyanobacteria in land plants in the form of Archaeothrix-type filaments that colonize prostrate axes of the land plant Aglaophyton major (Taylor and Krings, 2005). The cyanobacteria enter the axes through stomata and initially colonize the substomatal chambers, some also occur in voids of degrading sporangia of Aglaophyton. From the substomatal chambers they spread through the outer cortex and individual filaments or groups of filaments penetrate cortical cells to form coils (FIG. 3.118). Both transmission and scanning electron microscopy have provided valuable information about the morphology and ultrastructure of several fossil cyanobacteria, as well as suggesting a basis for determining stages in the growth of some fossil bacteria. Sphaerocongregus is the generic name applied to three morphologically distinct cell types collected from uppermost Proterozoic siliceous shales from southwestern Alberta, Canada (Moorman, 1974). One cell type, assigned to S. variabilis, includes coccoid cells 3–5 μm in diameter; another cell type is larger (5–6 μm) and often arranged in chains. The third morphologic category consists of globose masses 5–20 μm in diameter that are composed of small coccoid subunits. Surrounding these larger cell masses

CHAPTER 3 fungi, bacteria, and lichens

Figure 3.118 Coils of cyanobacteria inside Aglaophyton cell (Devonian). Bar  50 μm.

is a delicate envelope. Comparisons with living cyanobacteria suggest that the various morphologic types assigned to S. variabilis represent different stages in the life cycle of a single organism. Accordingly, the larger masses represent endosporangia that contain endospores, and some of the other cells represent stages in the vegetative plant body.

Lichens Lichens are unique, double organisms that consist of two unrelated components, an alga and/or cyanobacterium (photobiont) and a fungus (mycobiont). The organisms that make up the lichen live in a close symbiotic relationship in which the photobiont gains mechanical protection, increased water availability, reduced desiccation, and an improved ability to obtain nutrients from the mycelium of the fungus. The fungus, in turn, gains organic nutrients synthesized by the photobiont(s) that is, a source of carbohydrates for growth.

117

If an alga and a cyanobacterium are both involved, then the alga also gains a source of nitrogen. Recent research with modern lichens suggests that the relationship between the partners, although stable, may be highly variable, ranging from mutualism to parasitism. Although the thallus organization of most lichens suggests that they would be easily preserved, there are relatively few substantiated reports of fossil lichens of any antiquity. Hallbauer and van Warmelo (1974) described a putative fossil lichen from the Precambrian of South Africa under the name Thuchomyces lichenoides. It consisted of a horizontal thallus with a cortex of erect columns 5 μm high. The cortex included several zones of branched, septate hyphae suggestive of certain types of lichens. A more recent interpretation is that T. lichenoides was a parasitic, filamentous microorganism (Hallbauer et al., 1977). The occurrence of these structures in rocks that were no doubt strongly heated during diagenesis (a gold-bearing, uranium–lead-oxide conglomerate) makes their assignment as lichens perhaps less convincing. Interestingly, similar filamentous structures have been obtained in the laboratory abiotically using comparable preparation techniques (Cloud, 1976). One structure that may represent some type of lichen symbiosis is a fossil termed a biodictyon (Krumbein et al., 2003b). A biodictyon is part of a biofilm that penetrates the surface on which it grows and forms a three-dimensional net-like structure. Certain types of biofilms have been reported from as early as the Precambrian (Barghoorn and Tyler, 1965). Net-like structures have been described from phosphorites from the famous Neoproterozoic (551–635 Ma) Doushantuo Formation Lagerstätte in south China (Yuan et al., 2005). What makes these fossils so lichen-like is the presence of groups of coccoid cells (cyanobacteria or algae) within the spaces of what appears to be a net of fungal mycelia (FIG. 3.119). The specimens were deposited in a shallow subtidal environment and the site has yielded abundant algal fossils. Although the association of the coccoid cells and the hyphae of the net cannot unequivocally demonstrate a lichen symbiosis, the ordered arrangement of the cells of the two symbionts in the fossil indicates a regular and close physiological relationship between the two organisms. Molecular clock estimates from living lichens have also suggested that lichen symbioses may have existed during the Precambrian. The best known Paleozoic lichen comes from the Rhynie chert. Winfrenatia reticulata, named for Winfried and Renate Remy, is constructed of superimposed layers of aseptate hyphae that form numerous shallow depressions (FIG. 3.120) on the upper surface of the thallus (Taylor et al., 1997). Extending from the sides of the depressions are hyphae organized into a three-dimensional net (FIG. 3.121),

118

Paleobotany: the biology and evolution of fossil plants

Figure 3.119 Hyphal net-like structure containing what

is interpreted as cyanobacteria (Precambrian). Bar  20 μm. (Courtesy S. Xiao.)

Figure 3.121 Hyphal net of Winfrenatia reticulata enclosing cyanobacterial unicells (Devonian). Bar  60 μm.

Figure 3.120 Section of thallus of Winfrenatia reticulata showing hyphal pockets (arrows) with cyanobacteria inside (Devonian). Bar  800 μm.

with each net space filled by a coccoid, cyanobacterial cell surrounded by a thick sheath (FIG. 3.122). Each depression of the thallus shows various stages of cyanobiont cell division (FIG. 3.123); in some depressions the cells of the cyanobiont are moribund, suggesting that in these regions the cyanobacteria have become fully parasitized by the fungus. The affinities of the fungus are not known, and the cyanobacterium could be any one of a number of coccoid forms. It has been suggested that this fossil represents a colony of cyanobacteria parasitized by a fungus (Poinar et al., 2000). Although the thallus organization of W. reticulata is unlike that of modern lichens, the consistent association of coccoid cells and aseptate hyphae in this Devonian fossil satisfies the organismal and functional components of a lichen symbiosis, and the morphology of W. reticulata is similar

Figure 3.122 Cyanobacterial cell from Winfrenatia reticulata

showing thickened sheath (Devonian). Bar  10 μm.

CHAPTER 3 fungi, bacteria, and lichens

Figure 3.123 Eight-celled stage of Winfrenatia reticulata

cyanobacterium (Devonian). Bar  10 μm.

to the Proterozoic fossils of Yuan et al. (2005). In addition, lichen symbioses represent a type of controlled parasitism, and lichens are believed to have evolved from true parasitism to the more mutualistic relationships seen in many modern taxa. Flabellitha is a leaf-like film interpreted as a lichen from the Devonian of Kazakhstan. The fossil consists of septate hyphae and slightly sunken apothecia with asci; each ascus contains two bicellular spores (Jurina and Krassilov, 2002). Pelicothallos, found on the angiosperm leaf Chrysobalanus sp. from the Eocene of Tennessee, was originally described as an epiphyllous fungus (Dilcher, 1965) and later reinterpreted as an alga (Reynolds and Dilcher, 1984). The thallus bears setae and several dark fruiting bodies. On the basis of the spores and spore-bearing structures, Sherwood-Pike (1985) suggested that the fossil is similar to Cephaleuros (Trentepohliaceae), a genus of green algae that includes forms parasitic on land plants (Joubert and Rijkenberg, 1971), as well as those that function as lichen phycobionts (e.g., in the lichen Strigula). Indirect evidence of a cyanobacterial or algal-fungus symbiosis to form a lichen has been reported from Eocene amber collected from the Baltic region (Garty et al., 1982). This fossil is a multibranched thallus 2 cm wide, with the individual branches 1 mm in diameter. On the surface of the thallus are several structures thought to represent aeration pores, apothecia, and spores. The fossil is encased in amber, and elemental analysis suggests the presence of sulfur, calcium, iron, silicon, and aluminum. The presence of certain elements in the fossil is suggested as demonstrating

119

that the lichen accumulated iron and sulfur from the Eocene atmosphere prior to fossilization and thus can be regarded as a bioindicator of air pollution at that time. Other lichens have been described from Dominican amber (Rikkinen and Poinar, 2008), including a foliose thallus that morphologically appears similar to modern forms in the Parmeliaceae (Poinar et al., 2000). Additional indirect evidence of fossil lichens comes from the suggestion of Klappa (1979) that some laminar calcretes were formed by so-called lichen stromatolites. It has been known for some time that laminar calcretes (caliche), a type of calcareous soil deposit, contain algal filaments, fungal hyphae, and layers of organic-rich and organic-poor material, but the formation of these structures was not well understood. Lichens are known to be primary colonizers of rock surfaces, and the changes that they cause in their substrate represent the beginnings of soil formation. Klappa proposes that laminar calcretes are formed by a cycle of lichen colonization, followed by hardening of these biologically formed surfaces, followed by further lichen colonization, and so on. Over time, a layered structure, the so-called lichen stromatolite is formed. It is surprising that more lichens have not been described from the fossil record, since many of them, such as the foliose forms, have tissues capable of preservation. The recognition of well-preserved photobionts and mycobionts represents a major problem in the identification of these organisms since both are relatively fragile and may be difficult to recognize in permineralizations. In addition to a possible inability to recognize fossil lichens, another problem may simply be that the total number of lichens throughout geologic time was relatively small. Although many modern species reproduce asexually, which rapidly increases their distribution, the ancestral forms of lichens may have relied exclusively on the sexual mechanism of the fungus, a relatively slow way to distribute new individuals into the environment (Bowler and Rundel, 1975). In addition, many extant lichens grow in dry, exposed habitats, such as on bare rock surfaces, where the chance of fossilization is greatly diminished. Molecular sequence data from living ascomycetes involved in lichen symbioses suggest that not only have lichens evolved multiple times, but that lichens are far older than we once believed (Lutzoni et al., 2001). As more is learned about fossil fungi, algae, and cyanobacteria, obstacles to recognizing lichens in the fossil record and their geologic history may be substantially decreased. This is turn provides an opportunity to discuss not only the timing of lichenization through time but also the evolution of various structures unique to this mutualistic association.

This page intentionally left blank

4 Algae CHLOROPHYTA (GREEN ALGAE) ......................................... 123

Dictyochophyceae (Silicoflagellates) ............................................... 142

Prasinophyceae................................................................................. 124

Xanthophyceae (Yellow-Green Algae) ............................................. 142

Chlorophyceae ................................................................................. 126

Phaeophyceae (Brown Algae) .......................................................... 143

Ulvophyceae..................................................................................... 128

PRYMNESIOPHYTA (HAPTOPHYTES) ............................... 144

Charophyceae ................................................................................... 133 RHODOPHYTA (RED ALGAE) ................................................ 145 EUGLENOPHYTA ......................................................................... 138

Solenoporaceans .............................................................................. 146

DINOPHYTA (DINOFLAGELLATES) ..................................... 139

Other Calcified Red Algae ............................................................... 149

HETEROKONTOPHYTA ............................................................ 141

Uncalcified Red Algae ..................................................................... 150

Bacillariophyceae (Diatoms)............................................................ 141

ACRITARCHA (ACRITARCHS)................................................158

We have lingered in the chambers of the sea By sea-girls wreathed with seaweed red and brown Till human voices wake us, and we drown. T.S. Eliot, The Love Song of J. Alfred Prufrock

The algae are a large informal grouping of heterogeneous, polyphyletic or paraphyletic groups of primarily aquatic organisms ranging from tiny, flagellated unicells only a few microns in diameter to multicellular organisms up to 80 m long, such as the giant kelps (Graham and Wilcox, 2000). Unlike vascular plants, the algal body (thallus) lacks organ differentiation, although some forms have developed structures functionally similar to roots, shoot axes, and leaves. Most algae are photoautotrophic; some forms, however, are mixotrophic and derive energy both from photosynthesis and uptake of organic carbon by osmotrophy, myzotrophy, or phagotrophy. A few forms have reduced or lost their photosynthetic capacities and are entirely heterotrophic. Extant

algae are classified on the basis of a complement of features that do not normally lend themselves to fossil preservation. These include the type of pigments present, storage products, type of flagellation if present, and degree of cellularization. Moreover, molecular, biochemical, and ultrastructural characteristics are increasingly important in algal systematics and phylogeny (Brodie and Lewis, 2007). Although algae thrive in a spectrum of habitats, including Antarctic ice, rock and tree surfaces, animal fur, human and animal skin, and desert sand, most forms are aquatic. Algae are critical in modern aquatic ecosystems, not only in producing oxygen for other aquatic life, but also in serving as primary producers of organic matter at the base of the food

121

122

paleobotany: the biology and evolution of fossil plants

web (Round, 1981). Stands of larger algae (e.g., kelp forests) are used by animals as shelters and nursing grounds; other algae are pivotal in the physiology of some aquatic animals and involved in various vital processes (e.g., symbioses with corals [cnidarians]). Still other algae are structural contributors to the formation of reefs. With this recognition of the ecological significance of extant algae, a sound record of the evolutionary history of these organisms, including the roles they played in biological and ecological processes in the past, is critical in understanding the complexity and evolution of both ancient and modern aquatic ecosystems. Molecular clock hypotheses suggest that the first algae occurred on Earth during the late Paleoproterozoic or early Mesoproterozoic, some 1.5 Ga or even earlier (Yoon et al., 2004). This date corresponds with the long fossil record of these organisms, which can be traced into the Precambrian. Despite the often fragile nature of the plant body of many algae, there are numerous algal remains throughout the fossil record. Some algae have contributed to the formation of petroleum and thus are chemical fossils, whereas others are represented by thousands of feet of accumulated siliceous diatom shells. Algae that precipitate or deposit calcium carbonate, CaCO3, are often common in the fossil record. The actual algal thallus may not be preserved, but in many instances, the calcareous structures can form extensive deposits (Coniglio and James, 1985). Calcareous algae first appear in the late Neoproterozoic and become widespread during the Cambrian (Riding and Voronova, 1985). Among extant algae, 10% of benthic multicellular forms are calcified, with about 90 genera known for the red (Rhodophyta), 10 for the green (Chlorophyta), and 2 for the brown algae (Phaeophyceae) (Lüning, 1990; Kraft et al., 2004). Various calcareous algae play a major role in the formation of modern coral reefs (FIG. 4.1) (Wood, 1998), especially some coralline red algae (see section “Rhodophyta”) and the green alga, Halimeda (see section “Ulvophyceae”). Although sections of the fossils rarely indicate any of the cellular details of the original organism, the existence of calcified algae is determined by the presence of calcium carbonate that accumulated around the thallus. Cellular details of fossil algae can be preserved in certain depositional environments, for example, various Proterozoic organisms, including various algae, have been preserved in cherts. In contrast to many groups of microalgae and calcareous macroalgae, the fossil record of uncalcified macroscopic algae is meager, and hence the evolutionary history of these plants remains poorly understood. The failure to more fully

Figure 4.1

Coral (Extant). Bar  2 cm.

document uncalcified macroalgae throughout geological history is probably because algal thalli, with no real strengthening tissue, are not easily preserved (Tappan, 1980). Many marine forms may live in the intertidal zone, where they are easily destroyed after death. In addition, they do not produce characteristic skeletons that can be used for comparison with modern forms. As a result, hypotheses on the affinities of uncalcified fossil macroalgae, especially those preserved as impressions and compressions, are often formulated solely on basic morphological comparisons with modern forms. These scenarios thus remain speculative and place serious constraints on the interpretation of the roles these plants played in ancient ecosystems. Numerous impression and compression fossils were described, especially in the nineteenth and early twentieth centuries, that superficially resembled algae; these fossils, however, lacked certain distinctive features, so it was difficult to establish their biological affinities. To classify such forms, a system of artificial morphotaxa based on a few characteristic macromorphological traits was instituted. To a certain degree this taxonomy is still used today. For example, within the older literature there are numerous references to the morphogenus Fucoides (Harlan, 1830; Taylor, 1834; Lesquereux, 1869). This taxon was instituted by Brongniart (1822), for fossils which he interpreted as

chapter 4

similar to the extant brown alga Fucus and similar forms in the order Fucales (Phaeophyceae). The algal nature of most of these structures, some of which are as old as the Cambrian, however, has been challenged and many of them today are thought to be ichnofossils (Jensen and Bergstrom, 1995), graptolites (Bulman, 1963), or different sedimentary structures. The morphogenus Algites was established by Seward (1894) for fossils that he thought were probably algal, but which, because they lacked reproductive or internal characteristics, could not be assigned to any modern group. Over the years, Algites has become a “wastebasket taxon” for all sorts of enigmatic fossils that are similar to algae in basic organization. One species, A. enteromorphoides from the Devonian of Missouri, is composed of narrow, branching filaments and has been suggested to be a member of the Chlorophyta (Basson and Wood, 1970). Thallites is another morphogenus used for fossil thalloid plants that may have been either algae or bryophytes. Webb and Holmes (1982) suggested that these fossils be referred to as intermediate thalloid fossils. Thallites dichopleurus (Middle Pennsylvanian) is a dorsiventral thallus with a well-developed midrib (DiMichele and Phillips, 1976). The undulate lamina is slightly less than a centimeter wide, and the entire thallus dichotomizes several times. Latex peels made from the surface of the specimen and examined in the scanning electron microscope show epidermal cells of the midrib and thallus to be of different sizes; no epidermal appendages or openings are present. Other species of Thallites from Mesozoic rocks have been delimited on the basis of size, character of the margin, and nature of dichotomies. Since uncalcified macroalgae are not easily preserved, one of the most important discoveries in paleoalgology is numerous, uncalcified macroalgal floras that are well preserved as carbonaceous fossils in Precambrian and Cambrian marine rocks from North America, Europe, and especially China. Over the last few decades, paleobotanists working on the material from China have been particularly productive in deciphering these organisms (Cao and Zhao, 1978; Du and Tian, 1985, 1986; Duan et al., 1985; Y. Zhang, 1989; Zhang and Yuan, 1992; M. Chen et al., 1994; Steiner, 1994; R. Yang et al., 1999). Today many forms can be assigned to one of the major groups of algae with some confidence, based on structural similarities to extant forms, although the affinities of others still remain inconclusive. We have included several examples of fossil microalgae, as well as calcified and uncalcified macroalgae within the major groups:

algae

123

Higher taxa in this chapter (based largely on the classification system of Lee, 1999):

Phylum Chlorophyta (see text for detail) (green algae) Prasinophyceae Chlorophyceae Ulvophyceae Charophyceae Phylum Euglenophyta (Middle Ordovician–recent) Phylum Dinophyta (dinoflagellates) (?Silurian–recent) Calciodinellaceae Phylum Heterokontophyta (heterokonts) Bacillariophyceae Dictyochophyceae Dictyochales Xanthophyceae (yellow-green algae) Vaucheriales Phaeophyceae (brown algae) Ectocarpales Ectocarpaceae Laminariales Fucales Cystoseiraceae Phylum Prymnesiophyta (haptophytes) Coccolithophorales Discoasters Phylum Rhodophyta (red algae) Solenoporaceans (artificial group) Rhodophyceae Bangiales Corallinales Graticulaceae, Sporolithaceae, Corallinaceae Nemaliales Ceramiales Delesseriaceae Gigartinales Acritarcha (artificial group)

CHLOROPHYTA (GREEN ALGAE) The Chlorophyta (FIG. 4.2), or green algae, are the most diverse group of algae in the world today in terms of number of species (at least 7000 species), organization of the plant body (unicellular to multicellular), and habitat (from the surface of snow to a variety of symbiotic relationships) (Graham and Wilcox, 2000). The green algae are generally assumed to include the ancestral group that has given rise

124

paleobotany: the biology and evolution of fossil plants

 Gametophyte (n)

 Spore (n)

 Gametangia (n)

Germinating  spore

Germinating  spore  Spore (n)  Gametangia (n)

Spores (n) Meiosis  Gametophyte (n)

 Gamete (n)  Gamete (n)

Sporangia (2n) Sporophyte (2n) Germinating zygote

Zygote (2n)

Syngamy

Fusing gametes

Figure 4.2 Life history of the green alga Ulva showing an isomorphic alternation of generations. (From Taylor and Taylor, 1993.)

to the embryophytes (land plants). The green algae and the embryophytes are today interpreted as a monophyletic group, the Viridiplantae, consisting of two lineages, Chlorophyta and Streptophyta, the latter of which includes all embryophytes and the green algal class Charophyceae. Although the fossil record of the green algae is extensive, the fossil record has provided almost no information about the evolutionary steps involved in this transition (see Chapter 6). Instead, the phylogenetic relationships between these two groups have been hypothesized based on the study of biochemical, cytological, and ultrastructural features of living taxa (Lewis and McCourt, 2004) and these have been confirmed by molecular phylogenetic studies (Karol et al., 2001; Turmel et al., 2007). There are several late Neoproterozoic microfossils that morphologically resemble members of the Chlorophyta. One of these is Caryosphaeroides, a spheroidal unicell from the Bitter Springs Formation (800–830 Ma) of Australia (Schopf, 1968). The cells range from 6 to 15 μm in diameter and lack an outer sheath; a few specimens appear to have been contained within an amorphous organic matrix. Morphologically, these microfossils appear similar to the living green algae Chlorococcum and Chlorella (Chlorococcales). Neoproterozoic (700–800 Ma) shales from northeastern Spitsbergen have yielded a rich, well-preserved microfossil assemblage, including several

types morphologically similar to members of the Chlorophyta (Butterfield et al., 1988). One of these is a complex branched structure 10–50 μm in diameter and 1 mm long. It has been suggested that these structures are similar to the rhizoids of certain members of the Chaetophorales, a group of green algae that possess a prostrate basal system from which arise erect, branching filaments. Also included in this shale assemblage are branching, filamentous thalli constructed of cylindrical cells that range from 50 to 800 μm in diameter. Some fragments of this alga were a centimeter long. Certain extant members of the cladophoralean green algae (Cladophorales) are morphologically similar to these fossils. PRASINOPHYCEAE

The Prasinophyceae is a large, poly- or paraphyletic group of single-celled, flagellate green algae hypothesized to have diverged early in chlorophyte phylogeny (Lewis and McCourt, 2004). The group consists of four (Nakayama et al., 1998) or six to seven (Teyssèdre, 2006) distinct clades, one of which is the Pyramimonadales. The life history of several species of Pyramimonadales includes a unique non-motile stage characterized by the formation of organic-walled cyst-like structures termed phycomata (sing. phycoma). The alga remains metabolically active within the phycoma and undergoes

chapter 4

algae

125

Higher taxa of green algae in this chapter (based largely on the classification system of Lee, 1999):

Class Prasinophyceae Pyramimonadales Cymatiosphaeraceae, Tasmanaceae Class Chlorophyceae Chlorococcales Hydrodictyaceae, Scenedesmaceae, Chlorococcaceae Tetrasporales Botryococcaceae, Palmellaceae Chaetophorales Volvocales Lageniastraceae Class Ulvophyceae Cladophorales Dasycladales (Cambrian–recent) Families: See Table 4.1 Receptaculitida (artificial group, may not be algae) Cyclocrinales (Cambrian–Devonian) Caulerpales Codiaceae, Caulerpaceae Ulvales Class Charophyceae Charales Eocharaceae, Calvatoraceae Moellerinales Moellerinaceae (Silurian–Permian) Sycidiales Sycidiaceae, Trochiliscaceae, Chovanellaceae, Pinnoputamenaceae Zygnematales Coleochaetales (see Chapter 6)

vegetative reproduction; as a result, the phycoma increases in size over time. Prasinophycean phycomata have decayresistant walls and thus have preservation potential as fossils (Quintavalle and Playford, 2006) (FIG. 4.3); the oldest persuasive specimens come from the Precambrian and are more than 1.2 Ga old (reviewed in Teyssèdre, 2006). The majority of fossil phycomata have been described from strata deposited in marine or brackish water, but there are also several reports of these structures from freshwater deposits (Dotzler et al., 2007). Phycomata assigned to the genus Cymatiosphaera (200 species currently recognized) are globular (spherical– ellipsoidal) vesicles, up to 100 μm in diameter, with the external surface divided into a polygonal reticulum by

Figure 4.3 Geoffrey Playford.

membranous expansions or muri, which are perpendicular to the surface (FIG. 4.4). Cymatiosphaera fossils are structurally similar to the reticulate forms within the extant prasinophycean genus Pterosperma (Colbath and Grenfell, 1995). Phycomata with a surface reticulum (Cymatiosphaeraceae) are present at least by the Ordovician (Tappan, 1980). They diversified extensively in the Devonian and later Paleozoic but are rare in Mesozoic and Cenozoic rocks. Because prasinophytes are most abundant in the absence of other phytoplankton, they became less important when dinoflagellates and other algal groups diversified during these periods. One hypothesis suggests that blooms of prasinophytes and acritarchs observed across the Triassic–Jurassic boundary are in response to rapid CO2 increases which acidified the upper ocean and thus reduced the ability of other organisms to calcify (van de Schootbrugge et al., 2007). Tasmanites (Tasmanaceae) is another type of algal microfossil which occurs in many marine facies from the Cambrian to the Miocene (Martín-Closas, 2003). Specimens are commonly found in the maceral type (Chapter 1) called tasmanite, as well as in certain oil shales. At least one species of this family, Pleurozonaria maedleri, however, has been recorded from Late Pennsylvanian–Early Permian non-marine deposits from France (Doubinger, 1967). The fossils are

126

paleobotany: the biology and evolution of fossil plants

Figure 4.4 Phycoma of the prasinophyte Cymatiosphaera

(Devonian). Bar  20 μm. (Courtesy N. Dotzler.)

preserved as compressed cysts, usually ranging from 100 to 600 μm in diameter. Haptotypic features are absent, but the surface is covered by numerous, regularly spaced punctae or small pits. Ultrathin sections of the wall show concentric banding and two types of pores which traverse the wall in a radiating pattern (Jux, 1968). The affinities of Tasmanites have remained in doubt since the original description by Newton in 1875. Most researchers now believe, however, that these microfossils represent the phycomata of planktonic prasinophytes. Guy-Ohlson (1988), using a variety of techniques, was able to demonstrate the life history of Tasmanites based on specimens of Toarcian (Early Jurassic) age from southern Sweden. Her investigation substantiates the affinities of Tasmanites with the modern prasinophycean algae Pachysphaera, Halosphaera, and Pterosphaera. CHLOROPHYCEAE

VOLVOCALES The Chlorophyceae encompass the widest range of morphologies in the green algae. The colonial Volvocaceae (Volvocales) and their unicellular relative Chlamydomonas reinhardtii (Chlamydomonaceae) have frequently been used as a model in studies addressing the evolutionary pathways leading from unicellularity to multicellularity, including a division of labor within the algal thallus (Kirk, 1998, 1999). Molecular evidence suggests a minimum age of 400–500 Ma for a few Chlamydomonas species (Van den Hoek et al., 1988).

Nevertheless, the fossil record of Chlamydomonaceae is virtually nonexistent, and that of Volvocaceae is meager, perhaps because colonies (coenobia) disintegrate almost immediately upon death (Tappan, 1980). Fossil unicellular algae suggestive of Chlamydomonas are preserved in Cenomanian (Late Cretaceous) amber from southern Germany (Schönborn et al., 1999). The algal cells are thick-walled, oval in lateral view, up to 10 μm long and display a single, cup-shaped chloroplast, which is characteristic of extant Chlamydomonas; flagella are not recognizable. A rare microfossil that has been interpreted as a volvocacean alga is Eovolvox silesiensis from the Devonian of Poland (Kaz´mierczak, 1975, 1981). This fossil consists of hollow spherules with a surface layer composed of closely spaced, ovoid, pyriform, or spindle-shaped isomorphic cells. It is similar in basic structure to Symphysosphaera radialis from the Lower Cambrian of China (Yin, 1992). Affinities with the Volvocaceae have also been suggested for a few other Paleozoic and Mesozoic microfossils (reviewed in Kaz´mierczak, 1981). Few of these records, however, can be regarded as unequivocal (Kirk, 1998). The most biologically interesting fossil with possible affinities to the Volvocales is the endophyte Lageniastrum macrosporae (Lageniastraceae) from the Lower Carboniferous (Viséan) of France (Renault, 1896a; Krings et al., 2005a). This alga occurs inside lycopsid megaspores in the form of dome-shaped, three-dimensional colonies composed of up to 500 lens- to pear-shaped cells arranged in a single layer and bounded by a transparent membrane. Lageniastrum macrosporae colonies display a striking similarity in organization to certain extant species of Volvox, including the presence of radiating protoplasmic strands that interconnect adjacent cells in the colony (FIG. 4.5). TETRASPORALES Several groups of chlorophycean green microalgae have been important geologically. Some are responsible for the formation of certain coals and also may have contributed to the formation of petroleum in oil shales (Wolf and Cox, 1981). These deposits contain irregularly shaped, yellow bodies that were formed by the hydrocarbon-producing alga Botryococcus braunii. This taxon, the only member in the family Botryococcaceae, is a living planktonic colonial green alga in the order Tetrasporales that is known from both temperate and tropical climates throughout the world. The fossil colonies consist of pear-shaped cells arranged in radial rows and surrounded by a mucilage layer which has been described as a cuticularized layer. In the fossils, the yellow bodies are thought to represent paraffins and fatty acids secreted by the cells and bound together by the mucilage-like

chapter 4

algae

127

Figure 4.5 Lageniastrum macrosporae colony showing proto-

plasmic strands interconnecting cells. Bar  50 μm.

substance of the sheath. Botryococcus colonies are known from every geologic period beginning in the Ordovician (Martín-Closas, 2003) and extending to the present day, and they can be the dominant plankton of certain freshwater ecosystems (Batten and Grenfell, 1996), for example, in the Rotliegend (Late Pennsylvanian–Early Permian) of the Saar-Nahe Basin in Germany (Clausing, 1999). Fossil Botryococcus has been shown to represent valuable proxy indicator for certain paleoenvironments (Guy-Olsen, 1998). Modern Botryococcus colonies live in standing bodies of both fresh and brackish water (Colbath and Grenfell, 1995). The alga has recently attracted attention in biotechnology since it represents an unusually rich, renewable source of hydrocarbons and other chemicals (Banerjee et al., 2002). CHLOROCOCCALES Chlorophycean green algae belonging to the order Chlorococcales have been reported from the Messel Oil Shale (middle Eocene) of Germany (Goth et al., 1988). These kerogen-rich sediments have a very high percentage of organic matter made up of unicellular algae that are morphologically identical to the modern species Tetraedron minimum (Chlorococcaceae). The rectangular unicells range from 5 to 20 μm long. Geochemical analyses of these cells indicate the presence of an insoluble, non-hydrolyzable, highly aliphatic biopolymer that is believed to represent a significant precursor in the formation of n-alkanes in crude oils. Laminar concentrations of fossil T. minimum have been described from lower Miocene lacustrine sediments near Hausen in the Rhön Mountains, central Germany (Goth and Schiller, 1994). The genus Pediastrum (Hydrodictyaceae) consists of diskshaped coenobia or colonies composed of a variable number

Figure 4.6 Coenobia of Plaesiodictyon decussatus (Triassic). (From Brenner and Foster, 1994.)

of cells (Gray, 1960). Cells are arranged in a concentric pattern, with each cell of the outer ring containing one to three spines. The geological range of Pediastrum remains uncertain. Stanevich et al. (2007) suggested that the Neoproterozoic acritarch Dictyotidium minor from the Chencha Formation of eastern Siberia is structurally similar to the extant P. boryanum, and thus may represent an early relative of that genus. Fossils that appear similar to Pediastrum have also been found in Silurian (Deflandrastrum; Combaz, 1962) and Triassic rocks (Plaesiodictyon (FIG. 4.6); Brenner and Foster, 1994; Wood and Benson, 2000). The first unequivocal representatives of Pediastrum come from the Early Cretaceous of Britain and North America (Batten, 1996); others have been reported from the Upper Cretaceous–Neogene of southern South America (reviewed in Zamaloa and Tell, 2005). The genus is also known from the Miocene of Oregon and the Eocene of southern Sumatra. The Miocene Pediastrum fossils from southern South America are 8- to 32-celled coenobia

128

paleobotany: the biology and evolution of fossil plants

Figure 4.7 Coelosphaeridium cyclocrinophilum (Ordovician). Bar  1 cm. (Courtesy BSPG.)

with the marginal cells each bearing two spines (Tell and Zamaloa, 2004). Today the genus lives exclusively in freshwater, and its presence in Cretaceous marine rocks suggests a difference in the physiologic tolerance of the taxon, or that the genus is not as good a proxy record for freshwater environments as has been suggested (Evitt, 1963a). Another extant chlorococcalean green alga that is also known from the fossil record is Scenedesmus. It consists of cylindrical cells with rounded or pointed ends that are joined laterally into 4- to 16-celled coenobia. Two fossil species (S. hanleyi and S. tschudyi) have been reported from the Upper Cretaceous and lower Paleocene of Colorado and Mexico (Fleming, 1989). Coenobia of S. tschudyi consist of four or eight cells, with the terminal cells possessing elongate extensions, whereas coenobia of S. hanleyi have four oval cells. Both Pediastrum and Scenedesmus are green algae almost exclusively restricted to freshwater habitats and important non-marine paleoecological indicators of the presence of lacustrine environments (but see above). ULVOPHYCEAE

DASYCLADALES Among the most commonly encountered members of fossil green macroalgae are those forms assignable or structurally similar to the Dasycladales (FIG. 4.7). There are 180 genera currently recognized within this order, of which only 11 still exist today. Berger and Kaever (1992) classified the 180 genera into five families, that is, the Seletonellaceae (Cambrian– Cretaceous), the Diploporaceae (Devonian–Triassic), the

Figure 4.8 Suggested reconstruction of Triploporella remesii. (From Pia in Hirmer, 1927.)

Triploporellaceae (Ordovician–Eocene) (FIG. 4.8), the Dasycladaceae (Jurassic–recent), and the Acetabulariaceae (Carboniferous–recent) (FIG. 4.9). In addition, a sixth family, the Beresellaceae (Late Devonian–Permian), has traditionally been placed in the Dasycladales (Deloffre, 1988), but Berger and Kaever (1992) regarded this family as a heterogeneous group of organisms that does not belong to the order Dasycladales and may not even represent algae (Table 4.1). Adams and Al-Zahrani (2000), however, retained the Beresellaceae in the order Dasycladales and reported on Kamaena khuraisensis from Saudi Arabia, a form that extends the fossil record of the family into the Late Jurassic (Kimmeridgian). Morphologically, fossil dasycladalean algae are radially symmetrical, with a central axis that produces whorls of lateral appendages, some of which are branched. Reproduction includes the formation of operculate cysts containing isogametes. Most members of the Dasycladales secrete lime around the thallus, and this greatly increases their potential for preservation. Uncalcified dasycladaleans are comparatively rare as fossils; an example of one is Chaetocladus (LoDuca, 1997; Kenrick and Vinther, 2006). The morphologically simple, bottle-brush-shaped thalli of Chaetocladus (Ordovician– Devonian) range from 2 to 6 cm high and 0.5 to 1.5 cm

chapter 4

algae

129

Figure 4.9 Acetabularia acetabulum gametangium-bearing

cup (Extant). Bar  1 mm. (From Berger and Kaever, 1992.) Table 4.1 Geologic range of dasycladalean families.

Family

Geologic range

Seletonellaceae Diploporaceae Triploporellaceae Dasycladaceae Acetabulariaceae Beresellaceae?

Cambrian–Cretaceous Devonian–Triassic Ordovician–Eocene Jurassic–recent Carboniferous–recent Late Devonian–Permian (?Late Jurassic)

Source: From Berger and Kaever (1992).

wide and are composed of an unbranched, parallel-sided central axis surrounded by various projections. The slender, unbranched projections terminate sharply, without rounding or tapering, and are arranged as discrete whorls around the axis. Other Paleozoic genera interpreted as uncalcified dasycladaleans include the Silurian genus Heterocladus (FIG. 4.10) (LoDuca et al., 2003), Medusaegraptus, which ranges from the Ordovician to the Silurian (LoDuca, 1990), and the Devonian Uncatoella (Kenrick and Li, 1998). Among the oldest fossil calcified dasycladalean algae may be Amgaella from the Middle Cambrian of the Amga River (Russia) and Mejerella and Seletonella from the Upper Cambrian of Kazakhstan; the precise affinities of these organisms, however, remain uncertain (Riding, 2001). Among the earliest records of bona fide calcified dasycladaleans are fossils assigned to Dasyporella, Moniliporella,

Figure 4.10 Thallus of Heterocladus waukeshaensis (Silurian).

Bar  2.0 cm. (From LoDuca et al., 2003.)

and Plexa from the Ordovician of China (Riding and Fan, 2001), and Rhabdoporella of probable Late Ordovician age (Riding, 2001). Studies by Elliott (1978) suggest that the fossil dasyclads occupied environments similar to their modern counterparts and that their adaptations to specific microenvironments (e.g., salinity, bottom sediment, attachments, and water energies) are essentially similar to those of the extant forms. In general, calcareous algae are scarce in the Early Ordovician, becoming diversified from the Middle Ordovician onward. Since they are associated with warm-climate carbonate sedimentation, none are known from Gondwana during the Ordovician, except in Australia, which occupied a low paleolatitude at that time (Poncet and Roux, 1990). In the Paleozoic and Mesozoic dasyclads, the primary whorled branches are irregularly positioned. These, in turn, bear secondary and tertiary laterals. Most were cylindrical plants, although some were club-shaped, spherical, or shaped like a string of beads (Bassoullet et al., 1977). The thallus was attached to the substrate by simple rhizoids. In the fossils, the encrusting calcium carbonate is sometimes so thick that

130

paleobotany: the biology and evolution of fossil plants

several orders of branches are covered. A slightly different morphology is apparent in Primicorallina (Seletonellaceae; Ordovician). Specimens of this dasyclad have a remarkably narrow central axis with three orders of laterals in lax arrangement (Berger and Kaever, 1992). Another form that extended from the Ordovician into the Lower Carboniferous is Rhabdoporella (Seletonellaceae), a type that produced only primary laterals, each of which were slightly clavate or terminally expanded. This genus, along with Dasyporella and Vermiporella, has been suggested as representing a primitive form or early evolutionary stage in the dasyclad algae (Herak et al., 1977). More highly evolved members are interpreted as containing a larger number of whorls and more orders of branches. Pianella is a small (3 mm long) dasyclad with funnel-shaped branches arranged in alternating whorls. This taxon is common in the Early Cretaceous of the Middle East, Italy, and the former Yugoslavia. Mizzia is a Permian dasyclad in the family Triploporellaceae that is found shelfward of the Capitan reef complex of southeastern New Mexico (Kirkland and Chapman, 1990). The modern analog of Mizzia is Cymopolia, which extends from the Cretaceous to the recent. Lithologic evidence suggests that Mizzia grew in water that was shallow, warm, and only slightly agitated, subsequently becoming hypersaline. The study by Kirkland and Chapman (1990), involving the distribution of Mizzia in time and space, uses this genus as an indicator of paleoenvironment. More recently, long-term patterns of dasycladalean biodiversity over the past 350 myr (Carboniferous–Pliocene) were compared to global fluctuations in temperature and sea level. It appears that, in general, diversity peaks occurred when warm, shallow seas were most extensively developed on the Earth (FIG. 4.11) (Aguirre and Riding, 2005). RECEPTACULITIDA AND CYCLOCRINALES Two other groups of Paleozoic calcium carbonate–depositing organisms that are known from fossil examples are the receptaculitids and cyclocrinaleans, which are often considered to be related to one another because they display comparable morphological architectures. Both groups have been classified among various higher-level taxa, including corals, bryozoans, and particularly sponges and algae. The latest interpretation is that receptaculitids are neither sponges nor algae (Riding, 2004), while cyclocrinitids (FIG. 4.7) represent a problematic group of algae, possibly a sister group to the Dasycladales (Nitecki et al., 1999). Members of the order Receptaculitida occur in the fossil record from the Ordovician to the Carboniferous (FIGS. 4.13; 4.14), and perhaps into the Permian (Nitecki, 1972; Nitecki et al., 2004). Although

they are typically found associated with coral reefs, they are not regarded as reef builders. One of the most common representatives of the receptaculitids is the Ordovician Fisherites reticulatus (Finney et al., 1993), sometimes called the sunflower coral (FIG. 4.12). This organism is spherical or cup shaped, 30 cm in diameter, and composed of a central axis from which radiate numerous, spindle-shaped lateral branches (meroms or meromes). Each merome bears a terminal rhomboidal element (FIG. 4.12), which together form a surface pattern of facets that resembles the arrangement of ripe achenes in a sunflower capitulum (inflorescence). The order Cyclocrinales displays a narrower stratigraphic range, extending from the upper Middle Cambrian to the Lower Devonian (Nitecki et al., 2004). Cyclocrinitids existed in communities at depths of 100 m of water. Cyclocrinites is a common Silurian form with a club-shaped central stem and radiating primary branches (FIG. 4.15). At the end of each primary branch shaft are bowl-shaped cortical cells with flattened tops that in turn merge to form a reticulum of hexagonal plates on the outer surface perforated by regularly spaced holes. CAULERPALES Halimeda (Caulerpaceae) is a calcified green alga that is widespread today and important as a structural component of many Cenozoic–recent reefs; the genus is estimated to be responsible for 25–30% of the CaCO3 in Neogene fossil reefs (Stanley and Hardie, 1998). The thallus of Halimeda is constructed entirely of branching filaments that are matted together to form a plant body, which in some species may reach over a meter in length. Since no cross walls are produced in any of the filaments, the plant may be regarded as one giant, multinucleated cell. Despite its dominance today, the fossil record of Halimeda extends back only to the Cretaceous and possibly to the Triassic (Hillis, 2001; Dragastan and Soliman, 2002), although taxa with halimediform morphology are known from the Permian (Poncet, 1989). Several authors (Hardie, 1996; Stanley and Hardie, 1998; Knoll, 2003a) have correlated the rise and fall of certain calcareous organisms, including some green algae, with changes in seawater chemistry through time, especially the ratio of magnesium to calcium, Mg/Ca. Those organisms which deposit massive calcium carbonate skeletons (FIG. 4.16) and have weak physiological control over mineralization are most strongly affected by changes in seawater chemistry. This theory would explain the lack of Halimeda-type algae prior to the formation of “aragonitic” seas in the Late Jurassic–Cretaceous (Ries, 2005). In fact, the dasycladalean algae are the group which contributed most to carbonate rock formation in the Triassic.

chapter 4

0

50 60

Ol. Mioc.

Cenozoic

40

Pa. Eoc.

30

131

P io

10 20

algae

70

120 130

Present sea level

110

Present sea level

90 100

Cretaceous

80

150 Age (Ma)

160 170 180

Mesozoic Jurassic

140

190 200 220 230

Triassic

210

240

270 280 290 300 310 320 330 340

Paleozoic Carboniferous

260

Permian

250

Low

High

Haq et al. (1987)

Low

High

Hallam (1992)

350 0

20

40 60 80 n° dasyclad species

100

Figure 4.11 Comparative plots of Carboniferous–Pliocene dasycladalean species richness measured against sea level estimates. Brackets at right indicate periods of maximum diversity. (Redrawn from Aguirre and Riding, 2005.)

The Paleozoic calcified marine green algae also include species assigned to the Codiaceae that lived in dense colonies in warm, relatively shallow waters (Torres, 1995, 1999; Torres et al., 2003). The vegetative thallus of Ivanovia tebagaensis, a Permian form, is coenocytic or siphonous, with a cyathiform (cup-shaped) membrane-like thallus, 1.5–2 cm wide and up to 3 cm tall. The membrane forming the

thallus had inner and outer cortices composed of utricles surrounding a central medulla (FIG. 4.17). The alga apparently reproduced asexually by forming outgrowths (buds) on a parental thallus; the mode of sexual reproduction in this species is unknown. Ivanovia is also known from Triassic rocks of the Yukon Stikine Terrane in Canada (Reid, 1986; Torres, 2003), and thus this genus was among the relatively few

132

paleobotany: the biology and evolution of fossil plants

Figure 4.12 Fisherites reticulatus (Ordovician). Bar  5 cm. (Courtesy BSPG.) Figure 4.14 Ischadites murchinsonii (Ordovician). Bar  1 cm. (Courtesy BSPG.)

Figure 4.13 Ischadites sp. showing surface pattern. Bar  2 cm.

(Receptaculitida)

Figure 4.15 Diagrammatic reconstruction of Cyclocrinites dactyloides (Silurian). (From Taylor and Taylor, 1993.)

chapter 4

algae

133

algal survivors of the massive Permian–Triassic extinction. The species from the Yukon, I. triassica, is interesting because it displays sexual reproductive structures in the form of stalked spherical outgrowths from the thallus that are interpreted as oogonia, and dome-shaped protuberances thought to be male gametangia (Torres, 2003). Another uncalcified green alga, which morphologically resembles the existing genus Caulerpa, has been reported to be preserved in Miocene diatomite from California (Parker and Dawson, 1965). This fossil, Caulerpites denticulata, consists of a diffusely branched thallus with broadened, ultimate segments.

Figure 4.16 Calcareous alga Lithothamnium fornicatum (Extant). Bar  2 cm.

TAXA INCERTAE SEDIS Early fossils interpreted as uncalcified green macroalgae include Yuknessia and Margaretia from the Middle Cambrian Burgess shale of Canada and the Marjum and Wheeler Formations in Utah (Walcott, 1919; Conway and Robison, 1988). Yuknessia simplex formed delicate, centimeter-sized thalli composed of unbranched or (repeatedly) dichotomizing erect cylindrical branches which arose from what appears to be a central holdfast. Thalli of M. dorus were threadlike, rarely branched structures up to 2 cm in diameter and probably up to 1 m long. Courvoisiella (FIG. 4.18) is a Devonian uncalcified green macroalga characterized by a comb-shaped thallus constructed of non-septate, branching tubes (Niklas, 1976a). Along some tubes are spherical structures about 200 μm in diameter that have been interpreted as some type of gametangium. Chemical analyses of the fossils confirm the presence of cellulose. These findings, along with morphological data, were used to suggest affinities with the green algae. Impression fossils of relatively large ( 15 cm long), flat, sometimes lobed thalli have been described from the Mississippian Bear Gulch Limestone in Montana (Grogan and Lund, 2002), and the Namurian of Hagen-Vorhalle in Germany (Krings, 2005). Although it is difficult to determine the affinities of these fossils because they lack characteristic features, such as reproductive structures, they have been suggested as superficially resembling the thalli of the modern green alga Ulva lactuca (Ulvales) and the red alga Porphyra umbilicalis (Bangiales). Bubnoffphycos rhombeum, described by Daber (1960) from the Permian of Germany, may also represent an algal thallus comparable to these extant forms. CHAROPHYCEAE

Figure 4.17 Cross section of Ivanovia tebagaensis show-

ing relationship between tissues A. and well-preserved membrane B. Bar  0.5 mm. (Courtesy A. Torres.)

The Charophyceae are generally considered to be the ancestral group within the Chlorophyta that gave rise to the land plants (McCourt et al., 2004). Initially, this relationship

134

paleobotany: the biology and evolution of fossil plants

Figure 4.19 Carbonaceous tuff consisting of Chara sp. plants (Extant). Bar  2 cm. (Courtesy BSPG.)

Figure 4.18 Portion of Courvoisiella ctenomorpha showing projecting tubules (Devonian). (From Taylor and Taylor, 1993.)

was based on the complex morphology of the charophytes, including differentiation into organs and the presence of enclosing structures around the egg cells (oogonia); molecular phylogenies have confirmed this relationship (Sanders et al., 2003; Turmel et al., 2007). Included in this group are four important orders, Klebsormidiales, Zygnematales, Coleochaetales, and Charales (Lee, 1999). We will discuss the Charales in greater detail and briefly address the Zygnematales. For a comprehensive survey of morphology, paleoecology, stratigraphic distribution, phylogeny, and classification of the charophytes, see Feist et al. (2005b). CHARALES The Charales, commonly known as stoneworts or brittleworts (FIG. 4.19), include six living genera that have been assigned to two tribes within the family Characeae: Chareae, including Chara (FIG. 4.20), Lamprothamnium, Nitellopsis, and Lychnothamnus, and Nitelleae, with Nitella and Tolypella (McCourt et al., 1996). They inhabit freshwater and brackish environments worldwide, where some may exceed 30 cm in length. The thallus is characterized

Figure 4.20 Stand of Chara vulgaris showing oogonia (yel-

low spots) (Extant). (Courtesy M. Feist.)

by distinct nodes and internodes, with whorls of laterals borne at the nodes (FIG. 4.21). Charales reproduce both asexually and sexually; male and female reproductive organs are produced on short branches and the female

chapter 4

algae

135

Figure 4.22 Polar view of gyrogonite showing spiral cell (Triassic). Bar  500 μm.

Figure 4.21 Portion of Chara axis with two oogonia. (From

Taylor and Taylor, 1993.)

organs of extant forms (oogonia) consist of five spirally arranged tubes or elements (FIG. 4.22) that surround the egg. A fossilized oogonium is termed a gyrogonite (FIG. 4.23) and is the principal fossil evidence of the order (FIGS. 4.24–4.26). The oldest charophyte gyrogonites are known from the Upper Silurian (Feist et al., 2005a). Gyrogonites assigned to the genus Trochiliscus (Devonian, Trochiliscaceae) are characterized by more than five enveloping elements that are dextrally coiled (twisted to the right). The gyrogonites are 0.5 mm in diameter and slightly longer than wide; the base is rounded and the distal end is elongated. Inside a gyrogonite of T. podolicus is a thin (1 μm), continuous membrane that has been interpreted as a remnant of the original oospore. In some sections, patches of disorganized cells are found associated with the membrane. The trochiliscs were initially regarded as marine organisms; now they are believed to have inhabited fresh or brackish water habitats. In contrast with the trochiliscs, the younger Devonian and Carboniferous gyrogonites demonstrate greater structural variability, with at least six elements surrounding the egg cavity, and an open pore at the apex (Peck, 1957).

Figure 4.23 Gyrogonite showing spiral cells in side view (Triassic). Bar  500 μm.

Near the close of the Devonian, the pattern of twisting in gyrogonites changed from dextral to sinistral (to the left). Beginning with Eochara (Middle Devonian, Eocharaceae), believed to be the ancestral form leading to modern taxa, there was a progressive reduction in the number of sinistrally

136

paleobotany: the biology and evolution of fossil plants

Figure 4.24 Lateral view of Perimheste horrida (Jurassic).

Bar  500 μm. (Courtesy M. Feist.)

Figure 4.26 Lateral view of Atopochara triquetra (Cretaceous). Bar  500 μm. (Courtesy M. Feist.)

Figure 4.25 Lateral view of Flabellochara grovesi (Cretaceous).

Bar  500 μm. (Courtesy M. Feist.)

coiled elements, with five (the common number in living species) established by the Pennsylvanian (Peck and Eyer, 1963). The evolution of modern forms has led to the elimination of the apical pore as a result of the tighter association of the spiral cells at the apex (Grambast, 1974) (FIG. 4.27). The reproductive structures in the Paleozoic charophycean families Sycidiaceae (Silurian–Carboniferous) (FIG. 4.28), Trochiliscaceae (Devonian), Chovanellaceae (Devonian– Carboniferous), and Pinnoputamenaceae (Devonian) include a utricle, a calcified supplementary vegetative cover that is believed to protect the zygote against desiccation and which surrounds the gyrogonite. These fossil charophyte families

Figure 4.27 Louis Grambast. (Courtesy J. Galtier.)

have been placed within a single order, the Sycidiales (Feist et al., 2005a) (FIG. 4.29). In other Paleozoic families, including the Eocharaceae (Middle Devonian–?Triassic), fructifications are more similar to those seen in modern taxa, which do not produce utricles; these families are included in the

chapter 4

sb sc

t pb pc v

Figure 4.28 Suggested reconstruction of Sycidium xizangense

utricle showing vesicle (v), thallus (t), primary (pb), secondary (sb) branches, and primary (pc) and secondary canals (sc) (Devonian). (Modified from Feist et al., 2005a.)

et al., 2005a.)

137

Charales. The only Paleozoic family containing utricleproducing and utricle-free species is the Moellerinaceae (Silurian–Permian), which has been interpreted as occupying a central position in the early phylogeny of the group. In addition to the large number of fossil gyrogonites that have been described and used as index fossils, the vegetative parts of many fossil charophytes are also known. Two genera have been reported from the Upper Devonian of South Africa (Gess and Hiller, 1995). In Hexachara (FIG. 4.30A) each node produces a whorl of six laterals, and oogonia are produced on each lateral, whereas in Octochara (FIG. 4.30B) a whorl of eight laterals are borne at each node; each lateral is branched and produces an oogonium. Palaeonitella cranii (FIG. 4.31) is a relatively small, anatomically preserved charophyte initially described by Kidston and Lang (1921a) from the Early Devonian Rhynie chert of Aberdeenshire, Scotland. The thallus consists of branched, septate filaments that possess a nodal organization. Associated with some of the filaments are long tubular cells, separated from one another by an enlarged node of

A

Figure 4.29 Suggested charophyte phylogeny. (From Feist

algae

B

Figure 4.30 Octochara crassa A. and Hexachara setacea B. (Modified from Gess and Hiller, 1995.)

138

paleobotany: the biology and evolution of fossil plants

The Clavatoraceae is a large group of exclusively Mesozoic Charales that has frequently been used in biostratigraphy of continental facies (Martín-Closas, 1996). They are known from the Oxfordian (Upper Jurassic) through Cretaceous of all continents except Australia and Antarctica, based on both fructifications and vegetative parts, often in organic connection. Clavatoracean fructifications are composed of an oogonium surrounded by a calcified utricle. Utricles are important as characters in species identification as they underscore morphological variability. The Clavatoraceae shows development of utricles similar to those in Paleozoic sycidialean families, and this is interpreted as being a result of similar external constraints, rather than expressing true phylogenetic relationships (Feist et al., 2005a). Fossils of the clavatoracean genus Clavator (Jurassic– Cretaceous) consist of strongly calcified stems with narrow internodes and six lateral branches in each whorl. Oogonia are located in a single vertical row on the adaxial side of a branch, one per node. The vegetative parts of another genus in the Clavatoraceae, Echinochara from the Morrison Formation (Jurassic) of North America, are known in some detail. The plant body consists of 12 dextrally spiraled, cortical tubes constructed of elongate cells arranged in a linear series. At the distal end of each cortical cell are five long spines. Oogonia were produced in whorls of six. cranii showing three nodes (Devonian). Bar  200 μm. (Courtesy W. Remy and H. Hass.) Figure

4.31 Palaeonitella

small cells. These tubular cells are similar to rhizoids in living Charales. Uncalcified oogonia found associated with P. cranii axes (but not in organic connection) are composed of six sinistrally spiraled cells with an equal number of coronula cells arranged in a single layer around an apical pore (Kelman et al., 2004). The shape of the oogonium is reminiscent of the extant Chareae, whereas the morphology of the thallus is similar to that of the Nitelleae. Another species, P. tarafiyensis, comes from the Upper Permian of Saudi Arabia (Hill and El-Khayal, 1983), and P. vermicularis has been reported from the Lower Cretaceous of Spain, along with three other charalean fossils (Martín-Closas and Diéguez, 1998). One of these, Charaxis spicatus, closely resembles members in the extant genus Chara, whereas the other two have been assigned to Clavatoraxis, a morphogenus created for sterile, verticillated clavatoracean (family Clavatoraceae) vegetative remains that cannot be attributed to any species of gyrogonite.

ZYGNEMATALES Thin sections of chert from the Middle Devonian of New York revealed both marine and freshwater algae (Baschnagel, 1966), including a representative of the Zygnematales (or conjugate algae). Paleoclosterium leptum is formed of solitary, elongate, lunate cells 46 μm long and 5 μm wide that appear morphologically similar to species in the extant genus Closterium. Another fossil member of the conjugates is Palaeozygnema spiralis, which occurs in Cretaceous amber from southern Germany (Dörfelt and Schäfer, 2000). This fossil has unbranched chains of cells, each 20 μm long by 14 μm wide, in which chloroplasts and zygotes are exquisitely preserved. Palaeozygnema (FIGS. 4.32, 4.33) is similar to the modern genus Zygnema, although the process of gametogenesis is apparently different in the fossil. Zygnematacean zygospores reported in palynological samples have been useful in assessing changing depositional environments (Zavattieri and Prámparo, 2006).

EUGLENOPHYTA The Euglenophyta is a diverse group of naked, motile, unicellular organisms characterized by a pellicle composed of

chapter 4

algae

139

Figure 4.33 Filament of Palaeozygnema spiralis with intact

cells showing disk-like axial plastids (arrows) (Cretaceous). Bar  10 μm. (Courtesy A. Schmidt and H. Dörfelt.)

Figure 4.32 Palaeozygnema spiralis, hypnozygote with heli-

cal ornamentation (arrow) (Cretaceous). Bar  10 μm. (Courtesy A. Schmidt and H. Dörfelt.)

helically arranged, interlocking proteinaceous strips (Walne and Kivic, 1989). The number of flagella is two, but in some species, several may be present. Within the group nutrition ranges from autotrophic to heterotrophic. The group is cosmopolitan and may be found in almost all habitats. Modern euglenoids include 43 genera and 600–800 species. They have historically been classified either as plants (Bold and Wynne, 1985) or as protozoans (Leedale, 1985). Molecular data (small subunit rRNA gene sequences) suggest that the euglenoids diverged far earlier than the radiation responsible for green plants (Sogin et al., 1986). Until recently the fossil record of the group included just a few reports of specimens from Cenozoic rocks believed to represent members of the extant genera Phacus or Lepocinclis (Bradley, 1929) and Trachelomonas (Deflandre and LeNoble, 1948). Moyeria is a Middle Ordovician–Silurian acritarch that is believed to represent a fossil euglenoid (Colbath and Grenfell, 1995). The specimens are spindle shaped, range up to 40 μm long, and display bihelical symmetry. The cells are

characterized by a series of longitudinal spiral strips that mirror the pellicle morphology of Euglena (Gray and Boucot, 1989). The fossils are abundant as early as the Middle Ordovician in non-marine, nearshore environments (Gray, 1988a). The most persuasive fossil euglenoids discovered to date are preserved in amber. Schönborn et al. (1999) reported on a diverse microcoenosis in Mesozoic (Late Cretaceous; Schmidt et al., 2001) amber from southern Germany that contains several types of protozoans, including two euglenoids. Recently, a similarly diverse but slightly older (Early Cretaceous) microcoenosis containing the colorless euglenoid Astasia was discovered in amber from Álva in northern Spain (Ascaso et al., 2005).

DINOPHYTA (DINOFLAGELLATES) In some older treatments, the organisms included within this group are classified within the Pyrrhophyta or fire algae (Chapman and Chapman, 1973). Although most dinoflagellates occur in marine waters, freshwater forms are also known. Most extant dinoflagellates are free-living components of the oceanic plankton, but saprotrophic, parasitic, symbiotic, and holozoic (heterotrophic) forms are present

140

paleobotany: the biology and evolution of fossil plants

as well. They range from 5 μm to 2 mm (e.g., Noctiluca) in diameter and are characterized by two flagella of unequal length (heteromorphic) and different beating patterns (heterodynamic) (Hausmann et al., 2003). One flagellum extends outward, but the other wraps around the equator of the cell (FIG. 4.37) and contributes to their unique mode of locomotion, by spiraling through the water. The adaptation of these organisms to a wide variety of environments is reflected by a tremendous diversity in form and nutrition, as well as an extensive fossil record (Hackett et al., 2004). Some dinoflagellates produce extensive blooms, called red tides. The toxic by-products of these organisms cause death to fish and shellfish, as well as higher organisms in the food web. Extensive fossil blooms of certain dinoflagellates were no doubt responsible for the formation of some oil deposits in the world, based on a comparison of 4-methylsteroidal hydrocarbons in petroleum deposits with 4-methylsterols in modern dinoflagellates (Robinson et al., 1984). Freshwater dinoflagellates have also been suggested as the source for some oils (Goodwin et al., 1988). Fossil evidence for dinoflagellates is based on organicwalled cysts (FIGS. 4.34–4.36), sometimes termed dinocysts, which are variable in size (25–250 μm in diameter) and shape. Since only about 15% of extant dinoflagellate species produce fossilizable cysts, it is highly probable that the fossil record represents only a small segment of the actual diversity of these organisms through time (Head, 1996). The wall of most dinoflagellate cysts consists of a substance similar to sporopollenin, which accounts for their excellent preservation as fossils. Other dinoflagellate cysts have calcified walls (see below). Cysts are subdivided by a transverse furrow (paracingulum) into an anterior epicyst and a posterior hypocyst (Evitt, 1985; Evitt et al., 1977). On the surface are numerous, polygonal paraplates separated from one another by parasutures; the plates and sutures are given numbers for descriptive purposes. Some cysts are relatively smooth, whereas others possess various forms of ornament, ranging from spinelike processes (FIG. 4.38) to elaborate horns. The earliest probable fossil dinoflagellate is Arpylorus antiquus from the Upper Silurian of Tunisia (Sarjeant, 1978); the oldest undisputed fossil member of this group occurs in the Early Triassic (Fensome et al., 1999). Certain acanthomorphic acritarchs from the Mesoproterozoic Beidajian Formation in North China, however, have been reported to display morphological and ultrastructural features similar to those seen in living dinoflagellates (Meng et al., 2005; Yin et al., 2005), and dinosteroids were isolated from the rock matrix containing the fossils. This suggests that these fossils may represent the oldest dinoflagellates. The Precambrian

Figure 4.34 Hapsidopalla exornata (Devonian). Bar  20 μm. (From Playford, 1977.)

Figure 4.35 Dinoflagellate Invertocysta tabulata (Miocene). Bar  40 μm. (Courtesy L. E. Edwards.)

origin of dinoflagellates is substantiated by geochemical analyses that show a nearly continuous dinosterane record in Precambrian to Cenozoic organic-rich marine rocks (Moldowan and Talyzina, 1998; Moldowan et al., 2001).

chapter 4

algae

141

20 μm

4.38 Florentinia, a dinoflagellate Bar  20 μm. (Courtesy M. S. Zavada.)

Figure Figure 4.36 Dinoflagellate Wilsonidium tabulatus (Eocene). Bar  25 μm. (Courtesy L. E. Edwards.) Apex Plates

Epitheca

Sutures

(Cretaceous).

in younger Mesozoic shelf and slope sedimentary environments throughout the world (Keupp, 1992), and have been used as valuable proxy indicators for (paleo-)environmental conditions (Zonneveld et al., 1999). Important features in the classification of these fossils include ultrastructure of the calcitic cyst walls, orientation of the crystallographic c-axes of the wall crystals (Young et al., 1997; Kohring et al., 2005), and position and configuration of the excystment aperture or archeopyle (Streng et al., 2004).

Cingulum

HETEROKONTOPHYTA Hypotheca

Flagella

Sulcus

Figure 4.37 Structure and morphology of a dinoflagellate cell. (From Evitt, 1985.)

The hydrocarbon dinosterane is concentrated in and nearly exclusive to the dinoflagellates, so this chemical evidence strongly indicates that dinoflagellate lineages must have existed as early as the Precambrian. The small-sized (10 or 75 μm in diameter) cysts of calcareous dinoflagellates (Calciodinellaceae sensu lato; Gottschling et al., 2005), which are sometimes also termed calcispheres, initially occur as fossils in the Late Triassic, becoming highly diverse in the Cretaceous and throughout the Cenozoic (Kohring et al., 2005). They occur in rock-forming abundance

This large and diverse group of phototrophic and heterotrophic organisms is characterized by motile cells that typically have two unequal flagella, that is, a forwardly directed tinsel flagellum and posterior directed whiplash flagellum. The tinsel flagellum is covered with lateral bristles (mastigonemes), whereas the posterior flagellum is smooth and usually shorter or sometimes reduced to a basal body (Lee, 1999). All are golden or brown in color and several classes are generally recognized, including the Synurophyceae, Pelagophyceae, Raphidiophyceae (chloromonads), Eustigmatophyceae, Chrysophyceae (golden algae), Bacillariophyceae (diatoms), Dictyochophyceae (silicoflagellates), Xanthophyceae (yellow-green algae), and Phaeophyceae (brown algae). We will cover only the last four groups here. BACILLARIOPHYCEAE (DIATOMS)

The diatoms today include more than 100,000 extant species (Round et al., 1990). Most diatoms are unicellular organisms (FIG. 4.39) (a few are filamentous) that are ubiquitous in water,

142

paleobotany: the biology and evolution of fossil plants

4.39 Diatom Cyclotella meneghiniana (Extant). Bar  10 μm. (From Hoops and Floyd, 1979.)

Figure

occurring in environments that include freshwater, brackish, terrestrial, subaerial, and marine habitats (Round et al., 1990; Hausmann et al., 2003). They are an important part of the marine phytoplankton and, as primary producers, are estimated to fix at least one-quarter of the inorganic carbon fixed in the ocean (Granum et al., 2005). Many are planktonic, but they also occur as epiphytes on, or endophytes in, other organisms. Diatoms are constructed of a cell enclosed by a rigid, cell wall or shell (FIG. 4.39) composed of opaline silica (silicon dioxide, SiO2) coated by an organic material. This shell or frustule is constructed much like a Petri dish or a box in that one valve fits inside the other valve, and it is the reason that the fossil record of diatoms is so excellent. When a diatom cell dies, the frustule separates into two parts which settle to the bottom. Deposits of fossil diatoms, termed diatomaceous earth, may be more than a thousand feet thick in some areas of the world, for example, in Lompoc, California, where diatomaceous earth is mined and has several commercial uses, as an abrasive or in filtration. Frustule shape and elaborate ornamentation are two of the important characters used to classify fossil diatoms. Two of the most common shapes are centric (radial) and pennate. The fossil record of the diatoms is extensive and can be traced into the Mesozoic (Sims et al., 2006). The earliest unambiguous marine diatoms are known from the Lower Jurassic (Rothpletz, 1896; Barron, 1987), and the earliest diverse diatom assemblages occur in the Lower Cretaceous. Molecular clock hypotheses, however, suggest an earlier origin of the group (Kooistra et al., 2003). Medlin et al. (1997) suggested that the origin of diatoms may be related

to the end-Permian (250 Ma) extinction event, after which there were presumably many marine niches available. The oldest fossil freshwater diatoms to date are latest Cretaceous (Maastrichtian) and have been described from several different localities, including the Tarahumara Formation in northern Mexico (Chacón-Baca et al., 2002) and the Deccan Intertrappean beds and Lameta Formation of India (Ambwani et al., 2003). Diatoms, including exceptionally well-preserved organelles (e.g., internal membranes, lamellate plastid fragments, extracellular mucilage bodies), have been reported from Eocene terrestrial sediments (A. Wolfe et al., 2006). Because of their elaborate ornamentation and excellent preservation, diatoms are important in biostratigraphy. They have also been widely used in paleoecology as indicators of past environments. Although the use of molecular data has proved useful in determining the relationships among various groups of extant diatoms, evaluating morphological characters and determining synapomorphies in groups of fossil diatoms have not rapidly moved forward (Williams, 2007). DICTYOCHOPHYCEAE (SILICOFLAGELLATES)

The silicoflagellates comprise a small group of autotrophic marine, planktonic organisms that range from 20 to 50 μm in length and, as the name implies, possess a siliceous skeleton composed of opaline rods fused together to form a network (Preisig, 1994; Desikachary and Prema, 1996). Some classify them with the diatoms. They are first encountered in Early Cretaceous sediments (McCartney et al., 1990), and their peak abundance is reduced in the Cenozoic (Hausmann et al., 2003) so that they are represented by a single family today. Silicoflagellates have not been used extensively as biostratigraphic markers because of the relatively slow rate of evolution within group. Their value as fossils lies in their apparent sensitivity to temperature, and they are useful for biostratigraphy at higher latitudes and in deeper water, where calcareous microfossils are less common, as well as paleoenvironmental indicators. They are potentially useful in the study of paleoclimatology and perhaps in determining productivity levels in ancient ecosystems. XANTHOPHYCEAE (YELLOW-GREEN ALGAE)

The yellow-green algae are a group of heterokonts composed primarily of freshwater forms and a few marine representatives. Many species are single-celled organisms, whereas others are colonial, living as naked cells in a gelatinous envelope, or produce long filaments of cells. The group also includes a number of coenocytic forms such as the water felt Vaucheria (Vaucheriales). Molecular data have shown that the

chapter 4

algae

143

Xanthophyceae are most closely related to the Phaeophyceae (Potter et al., 1997). The fossil record of the Xanthophyceae is scanty. One fossil representative is Palaeovaucheria from the one-billion-year-old Lakhanda Formation in Siberia. This fossil alga displays morphological traits characteristic of vaucherian xanthophytes, including branching at right angles, two sizes of filaments on the same individual, and terminal pores and septae at filament ends (Javaux et al., 2003). Other vaucherian algae have been reported from middle Neoproterozoic shales in Spitsbergen (Butterfield, 2004, 2007). PHAEOPHYCEAE (BROWN ALGAE)

The brown algae consist of 1500 extant species in more than 250 genera (Norton et al., 1996). Phaeophycean thalli range in size from microscopic uniseriate filaments, for example, in the Ectocarpaceae, to complex plant bodies with significant organ, tissue, and cellular specialization (Graham and Wilcox, 2000). In some species of the giant kelp Macrocystis (Laminariales), the thallus is more than 70 m long. With two exceptions (the genera Newhousia and Padina; Kraft et al., 2004), all brown algae are uncalcified, which has contributed to the absence of a well-defined fossil record for the group. Brown algae have large amounts of the carotenoid fucoxanthin which gives them their characteristic color, and stored food is in the form of laminarin. Modern representatives typically inhabit colder waters in the northern hemisphere; in the tropics they occur in large numbers in the Sargasso Sea region of the Atlantic Ocean. One major problem in identifying fossil brown algae is their morphologic similarity to some members of the Rhodophyta. Several of the impression–compression thalloid taxa mentioned in this chapter (Rhodophyta, Chlorophyta) may belong to the brown algae, as Leary (1986) noted in his description of three uncalcified algal impressions from Mississippian strata. One of these, Phascolophyllaphycus, has elongate blades with rounded apices and tapered bases attached to the stipe in a helical arrangement. Several pneumatocysts, air-filled sacs that are characteristic of the brown algae, occur at the base of the blades; each is 1 mm in diameter. Early megafossils interpreted as phaeophycean algae come from the Lower Cambrian of southwestern China (Xu, 2001a). Thalli of Punctariopsis latifolia are up to 20 mm high and consist of single or clumped unbranched, leaflike blades, each of which is attached to the substrate by a short, narrow stipe and a globose or irregularly shaped holdfast (FIG. 4.40). In P. simplex, the blades occur singly and have a globose base. Vendotaenia antiqua was first reported from the Neoproterozoic (late Vendian) of the former Soviet Union (Gnilovskaja, 1971) but is also known from the Lower

Figure 4.40 Reconstruction of Punctariopsis latifolia. (From

Xu, 2001.)

Cambrian of China (Xu, 2001a). The thallus is composed of ribbonlike, unbranched or occasionally branched blades up to 85 mm long, which bear sporangia. Several algal morphotypes have been described from rocks of Late Ordovician age in Canada (Fry, 1983). The preservation of the carbonaceous specimens and the sedimentology of the deposits suggest that these algae probably grew in relatively shallow water. Winnipegia has an axis bearing wedge-shaped appendages, each 3.5 cm long and lacking surface features. The thousands of specimens encountered in the rocks suggest that the plants were not transported a great distance and were probably buried close to the site in which they grew. Thalassocystis is a compressed middle Silurian alga that is thought to be either a member of the red or brown algae (Taggart and Parker, 1976). The thallus is branched, with each branch terminating in an inflated bladder about 2.5 cm long. These algae were deposited in a shallow-water marine environment, but the actual habitat of the plants is not known. Another thalloid fossil with possible affinities in the Phaeophyceae is Yeaia africana from the Late Devonian Witpoort Formation in South Africa (Hiller and Gess, 1996), a form that resembles Y. flexuosa from the Upper Silurian Baragwanathia flora (Chapter 8) of central Victoria (Douglas, 1983). The thallus of Y. africana consists of straplike, repeatedly dichotomizing blades that are ornamented with tiny spots. Brown algae resembling members of the Laminariales are known from Miocene rocks of the Monterey Formation in California (Parker and Dawson, 1965). One of the most interesting forms is Julescraneia, a large compound thallus

144

paleobotany: the biology and evolution of fossil plants

Figure 4.42 Emiliania huxleyi (Quaternary). Bar  1 μm. (Courtesy J. R. Young.)

Figure 4.41 Cystoseirites altoaustriacus thallus fragments.

partschii from Romania (Givulescu, 1975) and the Ukraine (Molhanov, 2004).

Bar  1 cm. (Miocene). (Courtesy BSPG.)

PRYMNESIOPHYTA (HAPTOPHYTES) with lateral branches arising from a central pneumatocyst 16 cm in diameter. Lateral branches extend up to 3.6 cm in diameter and are believed to have terminated in a blade much like that in the extant brown algae Pelagophycus and Nereocystis. Also present in the diatomaceous sediments are a large number of brown algae that can be directly compared to members of the extant family Cystoseiraceae (Fucales). One of these, Paleohalidrys, is a flat, linear thallus 40 cm in length and often demonstrating pinnate branching of the laterals. In many modern members of this family, the blades can become disassociated seasonally, and it is believed that this phenomenon is responsible for the large number of specimens found in the Miocene deposit. Other fossils interpreted as belonging to the Cystoseiraceae have been reported from the Cenozoic of central Europe (reviewed in Kovar, 1982). Among these is Cystoseirites altoaustriacus, a repeatedly branched thallus composed of slender, cylindrical branch segments and spherical or egg-shaped air vesicles (aerocysts), each 2–4 mm in diameter (FIG. 4.41). Another Cenozoic form closely resembling extant members of the Cystoseiraceae is Cystoseirites (or Cystoseira)

The Prymnesiophyta, also known as Haptophyta, is a group of autotrophic, planktonic uninucleate flagellates characterized by the presence of a haptonema (a filamentous, microtubulesupported appendage) that lies between two smooth, approximately equal flagella (Lee, 1999; Andersen, 2004). The group includes at least 500 extant and many more fossil species. Geologically important members of the Prymnesiophyta are certain calcareous nannofossils termed coccolithophores or coccolithophorids (FIG. 4.42). Surrounding the living cell of these organisms are small (20 μm in diameter), calcified scales termed coccoliths which demonstrate a complex morphology and structure. Coccolithophores and coccoliths are valuable biostratigraphic markers, as well as indicators of paleoclimate (Wise, 1988). The living counterparts of these unicellular organisms are included within the Coccolithophorales (Jordan and Chamberlain, 1997), a group that is principally marine, and currently makes up 45% of the total phytoplankton in middle latitudes. Coccolithophores have a significant impact on their environment since they are the major primary producers that convert dissolved CO2 in the ocean to calcium carbonate (CaCO3) (Rost and Riebesell,

chapter 4

2004; Baumann et al., 2005) and therefore influence global biogeochemical cycles. Moreover, they have the ability to increase the albedo of the Earth by reflecting light from their coccolith-covered surfaces and by producing dimethylsulfide, a gas that contributes to the formation of aerosols which enhance cloud formation in the atmosphere (Brand, 1994). Coccolithophores live in the upper 200 m of the water column and, therefore, fossil assemblages can be used to delineate former oceanographic and climatic conditions of the surface waters. Coccoliths almost always occur in the fossil record as isolated entities; rarely are the scales found attached to the nannoplankton that produced them. Because of the small size of the scales, electron microscopy has become an important research tool to study coccoliths. Based on the structure and morphology of the scales, several artificial families have been established (Bown and Young, 1997; Young and Bown, 1997a, b). Coccoliths are commonly divided into two major groups: heterococcoliths are constructed of crystal elements that differ in size and shape, whereas the crystal elements making up holococcoliths are essentially identical in size and shape (Siesser and Winter, 1994). In addition, a third category of similar status, nannoliths, occur as fossils and are most commonly defined as calcareous nannofossils of uncertain affinity, but probably related to the coccolithophores (Young and Bown, 1997b; Young et al., 1997). Nannoliths first occur in the Carnian (Late Triassic); coccoliths appear somewhat later, during the Norian, and are particularly abundant in the younger Mesozoic and Cenozoic. Coccolithophores appear to have been at their zenith during the Late Cretaceous (PerchNielsen, 1985; Bown et al., 2004), where their coccoliths often form thick deposits; the famous White Cliffs of Dover, England, consist largely of coccoliths. Their amazing abundance during the Cretaceous has been attributed to the chemistry of seawater at the time—a low Mg/Ca ratio and high Ca concentration (Stanley, 2006). In contrast, modern seawater has low Ca concentration and a high Mg/Ca ratio, which apparently limits coccolithophore population growth today. Several ideas have been advanced regarding the function of the scales in coccolithophores (Young, 1994). One suggests that the scales function to shield the cell from excessive light, although perhaps the more popular corollary argues that the convexo-concave surface of the scales actually focuses light into the cell. Their small size, along with their abundance in younger Mesozoic and Cenozoic rocks and typically restricted stratigraphic range, has made coccoliths important index fossils and biostratigraphic markers. An analysis of their distribution through time indicates that a major extinction event occurred during the latest Cretaceous,

algae

145

followed by a recovery and radiation during the early Paleocene. This recovery, however, did not result in species richness similar to that seen in the Late Cretaceous (Bown et al., 2004). The discoasters are a distinct type of calcareous nannofossil believed to be related to the coccolithophores; in some treatments, they are included within the nannoliths (Bown et al., 2004). The tiny skeletons of discoasters appear as stars or rosettes, between 10 and 35 μm in diameter. They lack modern analogs but were a conspicuous component of the nannoplankton during most of the Cenozoic (Aubry, 1984). Discoasters became extinct at the end of the Pliocene (Chepstow-Lusty and Chapman, 1995; Kitaeva et al., 1997). They differ from coccoliths in being composed of tubular forms of calcite, whereas the coccoliths are formed of rhombohedral and hexagonal calcite crystals.

RHODOPHYTA (RED ALGAE) Rhodophyta, or red algae, are distinguished from other algal groups by the presence of chlorophylls a and d in combination with certain accessory pigments (phycobiliproteins), non-aggregated photosynthetic lamellae in the chloroplasts, specialized food reserves, unique sexual reproduction, and the absence of flagellation in all phases of the life cycle. The 4000–6000 extant species are primarily marine, mostly inhabiting warm tropical waters (Graham and Wilcox, 2000; Sounders and Hommersand, 2004). Phylogenies based on molecular data suggest the group is monophyletic (Le Gall and Saunders, 2007). The red algae are almost all multicellular and structurally more complex than other algae, with specialized pit connections between cells and a complicated mode of reproduction. Various red algae are commonly preserved as fossils, because they possess calcified skeletons that form as a result of calcium carbonate precipitation within the cell walls. In this feature, they differ from other lime-precipitating algae that deposit calcium carbonate only on the thallus. In the reds, the calcite is typically deposited in a grid-like pattern. Red algae are widespread in the fossil record, extending back to the late Mesoproterozoic. To date, the oldest red alga is a member of the Bangiales, Bangiomorpha pubescens (FIGS. 4.43, 4.44), from the Hunting Formation (1.2 Ga) of Somerset Island, arctic Canada (Butterfield et al., 1990; see Chapter 2). Not only is B. pubescens the oldest taxonomically resolved eukaryote on record, but it also exhibits the oldest example of eukaryotic sex and complex multicellularity (Butterfield, 2000, 2001).

146

paleobotany: the biology and evolution of fossil plants

Figure 4.44 Bangiomorpha pubescens (Mesoproterozoic).

Bar  50 μm. (Courtesy N. J. Butterfield.)

these reproductive propagules are impossible to distinguish from vegetative cells. Figure 4.43 Bangiomorpha pubescens (Mesoproterozoic). Bar  50 μm. (Courtesy N. J. Butterfield.)

Most Paleozoic calcareous red algae grew in open-marine carbonate shelf environments, although, as a group, they tolerated a variety of environments. These wide environmental variations suggest that individual taxa may provide important clues relative to ancient environments. Far more difficult, however, is the problem of relating Paleozoic red algae to living groups, because the taxonomy of the fossil forms often remains uncertain. This is due in part to the fact that reproductive structures have rarely been described. Their absence has suggested to some that the reproductive organs were externally produced and not calcified. Others suggest, however, that spores were produced within cells and that

SOLENOPORACEANS

One of the families traditionally placed within the Rhodophyta is the Solenoporaceae. Although several genera were initially assigned to the animal kingdom (e.g., as tabulate corals, bryozoans, or sponges; see Cózar and Vachard, 2006), solenoporaceans were later generally interpreted as calcified red algae (Pia, 1927) (FIG. 4.45). Today, however, the Solenoporaceae is known to represent a heterogeneous group that includes a variety of animals, red algae, and cyanobacteria, and, as a result, it is no longer possible to support the concept of the Solenoporaceae (Ordovician– Miocene) as a coherent family (Riding, 2001). Solenoporaceans were nodular or encrusting marine organisms composed of closely packed, radially or vertically divergent rows of tubes. Their diameters were almost

chapter 4

always larger than those of living coralline red algae (see section “Other Calcified Red Algae”), which are believed to be related structurally. One common representative is Solenopora (FIGS. 4.46–4.48), established for irregularly lobed, calcium carbonate masses composed of radiating, juxtaposed tubes with shared walls. The type species, S. spongioides from the Upper Ordovician of Estonia, was initially interpreted as a chaetetid sponge, but later transferred to the red algae. Reexamination of the original illustrations and new material from the type locality, however,

Figure 4.45

Julius Pia.

algae

147

indicates that the initial interpretation was accurate and thus removes Solenopora from the algae (Riding, 2004). Another form initially placed in Solenopora is S. gotlandica from the Silurian of Sweden and Wales. This species, which represents a true red alga, was transferred to the genus Graticula (Brooke and Riding, 1998, 2000). Skeletons of G. gotlandica (FIGS. 4.49, 4.50) may be free nodular or massive encrusting and are composed of laterally joined columns, branches, or pillars, which are mushroom- or umbrella-like, and occasionally up to 10 cm high (Nose et al., 2006). The columns consist of erect-to-radiating, juxtaposed filaments, which are rounded to polygonal or irregular in cross section, and which share adjacent walls; cross partitions (cross walls) in adjacent filaments are sometimes aligned. The arrangement of cross partitions, along with the presence of sporangia in sporangial compartments arranged in irregular sori, separates Graticula from other forms traditionally placed in the Solenoporaceae. Based on these features, Brooke and Riding (1998, 2000) established the new family Graticulaceae, which they assign to the Corallinales. The Paleozoic Graticulaceae are structurally similar to the Sporolithaceae (Corallinales); the type species G. gotlandica closely resembles members in the earliest recorded modern Corallinales, whose fossil record only extends back into the Early Cretaceous (Braga and Bassi, 2007; Tomás et al., 2007). The solenoporacean genus Parachaetetes is found as early as the Late Carboniferous, and Cenozoic forms are especially common in shallow-water reef facies (Wray, 1977). It occurs as bluntly lobed growths up to several centimeters in diameter. In vertical thin sections, typical forms exhibit a tightly packed mass of elongate cells arranged in curved radial lines. In cross section, these cells are circular and variable in length. At least some species within the genus Parachaetetes

Figure 4.46 Solenopora sp. (Jurassic). Bar  2 cm. (Courtesy M. Nose.)

148

paleobotany: the biology and evolution of fossil plants

4.49 Red algal–calcimicrobe boundstone with Graticula gotlandica (Silurian). Bar  5 cm (Courtesy M. Nose.) Figure

Figure 4.47 Solenopora condensata (Jurassic). Bar  2 cm

(Courtesy M. Nose.)

Figure 4.48 Solenopora sp. (Jurassic). Bar  2 cm (Courtesy

LMU.)

seem to represent true red algae (Aguirre and Barattolo, 2001), whereas others, again, are interpreted as chaetetid sponges today (Riding, 2004). The cells of Solenomeris are polygonal in outline, with those of adjacent rows forming a zigzag configuration. Formerly considered to be a red alga, today Solenomeris belongs with the incrusting foraminifera (Bassi, 2003). Marinella lugeoni (Late Jurassic–Oligocene) was originally described as a cyanobacterium but has more recently been included in the Codiaceae (Chlorophyta) and the Solenoporaceae. This organism forms either encrusting thalli, several centimeters in diameter, or erect, digitiform, and branching thalli with branches up to 9 mm high that are

Figure 4.50 Graticula gotandica framestone from Gotland with laterally linked red algal pillars (Silurian). Bar  2 cm (Courtesy M. Nose.)

attached to the substrate by a narrow base. The internal tissue is composed of radially oriented and densely packed filaments. Specimens of M. lugeoni from the Upper Jurassic of Portugal imply a close relationship with the corallinaceans, but structural similarities to the solenoporaceans have also been noted (Leinfelder and Werner, 1993). What were initially thought to be oval, aggregate reproductive structures in the

chapter 4

Miocene genus Neosolenopora (Mastrorilli, 1955) have been reinterpreted as an unusual bryozoan (Tillier, 1975). Several other genera traditionally referred to the Solenoporaceae, such as Metasolenopora, Petrophyton, and Solenoporella may belong, or be related to the Corallinaceae (Riding, 2004). The identification of various organisms traditionally included in the Solenoporaceae has been useful in understanding how skeletal organisms and microbes interacted to produce certain reef limestones (Adachi et al., 2007). OTHER CALCIFIED RED ALGAE

Fossil red algae, like many of those today, were important in the formation of reef communities as early as the Silurian. The Graticulaceae (see above) were important structural components of Silurian (Nose et al., 2006) and Devonian (Adachi et al., 2007) reefs. Another reef-building red alga from the Devonian is Archaeoamphiroa, which occurs along a 48-km outcrop within the Alexandra Reef Complex (Frasnian) in the Northwest Territories, Canada (Magathan, 1985). This fossil closely resembles the extant coralline red alga Amphiroa (Corallinales) and consists of alternating long and short rows of cells. The lithofacies suggest that this alga grew in a shallow, surf-swept reef in association with other filamentous algae. Masloviporidium is a cosmopolitan calcareous red alga from the Carboniferous, ranging from Russia to central Texas (Groves and Mamet, 1985). The thallus is sheetlike and consists of rows of wedge-shaped cells surrounded by calcified partitions. Each cell is connected to cells above and below by pores in the partitions. Nothing is known about the mode of growth or reproduction. Another Carboniferous (Pennsylvanian) red alga from marine rocks is Litostroma (Mamay, 1959a). This small alga is formed of irregularly shaped, thalloid platelets that are one cell thick and up to 6 mm in diameter. Filaments arise from the surface of the thallus, and irregularly shaped perforations are present along the margin. Many of the cells contain shrunken cell contents or nuclei. In addition to being assigned to the red algae, Litostroma has also been suggested as a member of either the green or brown algae. CORALLINALES One of the geologically most important orders of marine calcified red algae is the Corallinales. The crown group Corallinales includes two living families, the Corallinaceae and the Sporolithaceae, both of which have been documented from the Mesozoic and Cenozoic (Aguirre et al., 2000). Both families are characterized by macroscopic crustose or erect and branched thalli differentiated into two distinct histologic zones. The hypothallus, or medulla, which forms the basal

algae

149

part of the crustose plants and central part of the erect forms, is constructed of loosely arranged, relatively large cells. The perithallus (cortex), or second zone, which constitutes the major portion of the thallus, is located above the hypothallus in crustose forms and outside of the hypothallus in branched forms. It is characterized by smaller cells (Lee, 1999). In addition, a distinct surface layer, the epithallus, may be present and this is composed of one to a few layers of small, thin-walled cells (Xiao et al., 2004). Sporolithaceae and Corallinaceae are indistinguishable in vegetative anatomy but differ in sporangial structure: the former family is characterized by sporangial chambers grouped in sori, whereas the latter produces sporangial conceptacles with one (uniporate; Lithophylloideae and Mastophoroideae) or several (multiporate; Melobesioideae) small pores for the dispersal of spores (Aguirre et al., 2000). A common member of the Sporolithaceae is Lithothamnion (FIG. 4.16), which can be traced from Cretaceous tropical seas to present-day temperate and polar areas (Johnson,

Figure 4.51

J. Harlan Johnson.

150

paleobotany: the biology and evolution of fossil plants

1962) (FIG. 4.51). Earlier (Late Jurassic) records of the genus remain questionable with regard to both age attribution and taxonomic circumscription (Aguirre et al., 2000). Taxa demonstrate a variety of growth forms, ranging from tiny crusts to large aggregations 30 cm in diameter. Based on a study of Miocene non-geniculate coralline red algae from Crete (Greece), Kroeger (2007) suggested that various species of Lithothamnion, along with other forms, such as Sporolithon, can be used as valuable proxy indicators for paleoenvironmental studies in Cenozoic sediments, because these algae appear to form specific and consistent associations depending on water depth and water temperature. Early Cretaceous coralline algae from India have also been used as a proxy indicators for water depth and water temperature (Misra et al., 2006). Another widely cited genus is Archaeolithothamnion (Johnson, 1963), which has been regarded as a junior synonym of Sporolithon (Moussavian and Kuss, 1990; Tomás et al., 2007). One Early Cretaceous (Barremian–Albian) species originally assigned to Archaeolithothamnion is Parakymalithon phylloideum (Moussavian, 1987). The thallus of this alga is composed of a relatively thick hypothallus and only a few distinct horizontal rows of cells in the perithallus. Asexual reproduction occurs in the form of ovalshaped or fusiform sporangia arranged in a nemathecium-like receptacle that lies cushioned above the perithallial tissue. Parakymalithon is believed to be phylogenetically intermediate between the Sporolithaceae and Corallinaceae. Recently, however, the separate status of Parakymalithon as a genus has been questioned (Tomás et al., 2007). Karpathia (FIG. 4.52) is an example of a Cenozoic (Paleocene) form assigned to the Corallinaceae (Bassi et al., 2005). The hypothallus of this encrusting alga is composed of straight, loosely arranged filaments of irregularly shaped, large cells, which resulted from cell fusion. The perithallus is formed of densely arranged filaments or a series of smaller, thick-walled cells. Sporangial conceptacles are uniporate. UNCALCIFIED RED ALGAE

The most exquisitely preserved, early uncalcified red algae, some of which may be distantly related to the Corallinales, were discovered in the Neoproterozoic (600 Ma) Doushantuo Formation at Weng’an, South China (Xiao et al., 2004, and references therein). The fossils are phosphatized and thus possess cellular preservation. They are particularly interesting because they help bridge the stratigraphic and evolutionary gap between the earliest rhodophyte fossils and the calcified red algae described from younger strata. The most distinctive feature of the phosphatized Doushantuo algae described by Xiao et al. (2004) is their filamentous construction or cell fountain

Figure 4.52 Section of Karpathia sphaerocellulosa thallus (Paleocene). Bar  300 μm. (From Bassi et al., 2005.)

architecture. Thalli of Wengania globosa are broadly spherical, nodular, or irregular, 70–750 μm in diameter in thin section, and display a simple pseudoparenchymatous construction (FIG. 4.53), in which cortex and medulla are not differentiated. The thalli consist of cuboidal cells that form regular files, radiating outward and branching toward the thallus margin. In W. exquisita, cell files are less regularly arranged, whereas in the third species, W. minuta, regular cell files are absent (Xiao, 2004). Similar simple thalli are formed by Thallophycoides pholeatus and Gremiphyca corymbiata (FIG. 4.54). These thalli may be lobed, however, with the most profound lobation occurring in G. corymbiata. Thalli of Thallophyca ramosa and T. corrugata (FIG. 4.55) display a complex pseudoparenchymatous construction, in which there is a clear differentiation into an inner medulla and outer cortex (FIG. 4.56), with cortical cells either smaller than the medullary cells or arranged differently, and clustered cell islands interpreted as reproductive cells (Y. Zhang et al., 1998). The pseudoparenchyma is arrayed as upward diverging splays of filaments or cell fountains. Diverging filaments form fan-shaped lobes separated by deep invaginations. Particularly striking are cylindrical, cell-lined invaginations that resemble conceptacles. Thalli of Paramecia incognata are millimeter-sized nodules characterized by cortex– medulla differentiation, thallus compartmentalization (FIG. 4.57), and absence of well-developed invaginations (FIG. 4.58). Xiao et al. (2004) stated that the complex pseudoparenchymatous thalli of Thallophyca and Paramecia display features resembling those seen in some Paleozoic members of the Corallinales, and thus these forms may be interpreted as stem group corallinaceans. They are clearly different from the crown group corallines, however, in that they are not calcified in life (Xiao and Knoll, 1999).

chapter 4

4.53 Simple pseudoparenchymatous thallus of Wengania globosa (Neoproterozoic). Bar  50 μm. (From Xiao et al., 2004; courtesy S. Xiao.)

Figure

algae

151

Figure 4.55 Complex pseudoparenchymatous thallus of Thallophyca corrugata (Neoproterozoic). Bar  100 μm. (From Xiao et al., 2004; courtesy S. Xiao.)

Figure 4.56 Detail of complex pseudoparenchymatous thallus Figure 4.54 Gremiphyca corymbiata showing simple pseu-

doparenchymatous, lobed thallus (Neoproterozoic) Bar  100 μm. (From Xiao et al., 2004; courtesy S. Xiao.)

Another genus of anatomically preserved multicellular algae from the Doushantuo Formation is Sarcinophycus (Xiao and Knoll, 1999; Xiao, 2004). It differs from the previously detailed Doushantuo algae in that the thalli lack the cell fountain architecture and possess marginal protuberances

of Thallophyca corrugata (Neoproterozoic). Bar  100 μm. (From Xiao et al., 2004; courtesy S. Xiao.)

(FIGS. 4.59, 4.60). The systematic affinities of Sarcinophycus remain uncertain. Paratetraphycus giganteus (Z. Zhang, 1985) is yet another multicellular alga-like organism from the Doushantuo Formation. Although initially assigned to the cyanobacteria (Chroococcaceae-like coccoids; Z. Zhang, 1985), certain features have been noted that are reminiscent of extant bangiomorphic red algae (Y. Zhang et al., 1998;

152

paleobotany: the biology and evolution of fossil plants

Figure 4.58 Complex parenchymatous thallus of Paramecia incognata (Neoproterozoic). Bar  100 μm. (From Xiao et al., 2004; courtesy S. Xiao.)

Figure 4.57 Detail of thallus of Paramecia incognata showing

compartmentalized clusters (arrow) of larger, probably reproductive cells (Neoproterozoic). Bar  100 μm. (From Xiao et al., 2004; courtesy S. Xiao.)

Saunders and Hommersand, 2004). Paratetraphycus giganteus has been reported from the Meso-Neoproterozoic (middle Riphean) in the southern Urals (Sergeev and Lee, 2006). Some of the earliest compression fossils indicative of diverse, marine uncalcified macroalgal floras come from the Neoproterozoic of South China. These morphotypes, such as Doushantuophyton (centimeter-sized thallus fragments composed of erect, repeatedly forking branches) (FIG. 4.61), Konglingiphyton (centimeter-sized dendritic thalli), and Miaohephyton (millimeter-sized dichotomously branched thallus fragments characterized by horizontal constrictions), suggest affinities with the Rhodophyta. Other genera, such as Enteromorphites (centimeter-sized, branched thalli

attached to the substrate by a holdfast) (FIG. 4.62), Gesinella (centimeter-sized, lanceolate or strap-shaped thalli with a basal holdfast), and Yemaomianiphyton (centimeter-sized tuft-like thalli composed of a strong holdfast and forked or unforked erect branches), have been interpreted as either red, green, or brown algae (Steiner, 1994). It has also been interpreted as fragmentary specimens of Enteromorphites (Xiao et al., 2002). A detailed study on the holdfast structures formed by some of the Neoproterozoic macroalgae from China and their paleoenvironmental implications has recently been published by Wang and Wang (2006). The assignment of Miaohephyton to the red algae has been questioned by Xiao et al. (1998), who suggested that this form may represent a brown alga with possible affinities in the order Fucales. Enteromorphites siniansis is an organism composed of several forked hollow tubes 17 mm long that arise from a basal holdfast. It was initially interpreted as being structurally similar to the extant green alga Enteromorpha (Ulvales) (Zhu and Chen, 1984), but Steiner

chapter 4

algae

153

Figure 4.60 Detail of Sarcinophycus radiatus cell packets

(Neoproterozoic) Bar  25 μm. (From Xiao and Knoll, 1999; courtesy S. Xiao.) Figure 4.59 Sarcinophycus radiatus showing radiating packets of cells (Neoproterozoic) Bar  100 μm. (From Xiao and Knoll, 1999; courtesy S. Xiao.)

Figure 4.61 Doushantuophyton lineare. (Redrawn from Steiner, 1994.)

154

paleobotany: the biology and evolution of fossil plants

Figure 4.62 Enteromorphites intestinalis thallus. (From Xu,

Figure 4.63 Reconstruction of Paradelesseria sanguinea.

2001.)

(From Xu, 2004.)

(1994) noted that this structure may also resemble various red and brown algae. Enteromorphites intestinalis from the Early Cambrian Chengjiang Biota in southeastern China is a considerably larger (7 cm high) form, composed of hollow unbranched tubes that were attached to the substrate by a basal rhizoidal cell or tubular proliferations (Xu, 2001b). Paradelesseria sanguinea is a Delesseria- or Phycodryslike red alga from the Chengjiang Biota (Xu, 2004). The thallus is composed of leaf-like, lanceolate, or oblanceolate and petiolate blades 90 mm long that are attached to a subcylindrical, nodose stipe (FIG. 4.63). Two other enigmatic organisms from the Chengjiang Biota that have been variously interpreted as algae are Longfengshania cordata, a small thallus (2 cm high) composed of a bladelike distal portion and a hollow, stalk-like proximal portion (FIG. 4.64), and Plantulaformis sinensis (1 cm high), a cotyledon-like thallus composed of a two-parted distal blade and hollow, stalk-like proximal portion (FIG. 4.65) (Xu, 2002). Hofmann (1985a) suggested that Longfengshania may represent algae with possible affinities in the red or brown algae. Z. Zhang (1988), however, interpreted Longfengshania as an early

bryophyte. Another diverse early algal flora has been reported from the Middle Cambrian Kaili Biota in Guizhou Province, China (Yang et al., 2001). This flora consists of more than 20 genera, 5 of which have tentatively been assigned to the red algae, Palaeocodium, Paraamphiroa, Wahpia, Dalyia, and Bosworthia. Thalli of Paraamphiroa siniansis are 1.5 cm high and composed of a main branch that distally produces a cluster of second-order, bi- or trifurcating branches, each consisting of uncalcified joints and calcified, cylindrical segments. This alga is believed to represent the earliest fossil evidence for a calcified red alga (Yang and Zhao, 2000). Dalyia has also been reported from the Middle Cambrian Conasauga Formation in northwestern Georgia (USA) where it co-occurs with several putative green algae (Schwimmer and Montante, 2007), and from the Middle Cambrian Burgess Shale in British Columbia, Canada (Walcott, 1919), along with a second red alga, Waputikia ramosa (FIG. 4.66). The latter differs markedly from other Precambrian and Cambrian algae as it displays a more complex morphology in the form of a central axis interrupted by large branches,

chapter 4

algae

155

Figure 4.64 Reconstruction of Longfengshania cordata. (From

Xu, 2002.) Figure 4.65 Reconstruction of Plantulaformis sinensis. (From

Xu, 2002.)

Figure 4.66 Suggested reconstruction of Waputikia ramosa. (From Briggs et al., 1994.)

which in turn bear a profusion of smaller branches and terminal filaments (Briggs et al., 1994). Several extant red algae possess morphologies that are similar to the Late Ordovician alga Manitobia patula. In this form, the laminar thallus divides in a single plane, with each segment further divided into three, nearly uniform segments. The margin of the thallus is entire, and the apex of each segment typically truncated (Fry, 1983). Three macroscopic Devonian algae occur in marine rocks in New York (Fry and Banks, 1955). Drydenia (FIG. 4.67) consists of elliptical laminae (8.5 cm long) that are basally attached to a narrow stipe terminating in a branching holdfast. In Hungerfordia (FIG. 4.68), also recorded from the Late Devonian of South Africa (Hiller and Gess, 1996), the lamina is highly dichotomous with the distal segments lobed. Specimens of Enfieldia are circular (5.0 cm in diameter), with the outer margin lobed, and characterized by distinct reticulations. Both Drydenia and Hungerfordia have been compared with existing red and brown algae; Enfieldia is more difficult to position systematically, perhaps representing a thalloid liverwort.

156

paleobotany: the biology and evolution of fossil plants

Figure 4.68 Dichotomous thallus of Hungerfordia dichotoma

(Devonian). Bar  2.5 cm. (From Fry and Banks, 1955.)

Figure 4.67 Drydenia foliata (Devonian). Bar  4 cm. (From

Fry and Banks, 1955.)

The genus Perissothallus (FIG. 4.69) comprises several types of uncalcified macroalgae from Late Pennsylvanian and Early Permian freshwater environments of North America and Europe. The thallus consists of repeatedly dichotomizing, erect cylindrical branches that radiate from a small holdfast (Krings et al., 2007d). Vegetative reproduction occurs in the form of secondary thalli produced on prostrate branches. Striking similarities in basic structure exist between Perissothallus and members of the extant marine red algal genus Scinaia (Nemaliales), but the fossils superficially resemble species of the extant genera Codium (green algae) and Dictyota (brown algae). Another putative late Paleozoic red alga consists of a central axis with nodes of lateral appendages (FIG. 4.70). Bassonia hakelensis is a Cenomanian (Late Cretaceous) marine fossil from the Haqel fish beds in Lebanon that

Figure 4.69 Perissothallus showing branches radiating from holdfast (Pennsylvanian). Bar  1 cm.

appears as a compressed, irregularly branched thallus, 20.0 cm long (Basson, 1972; Krings and Mayr, 2004). The thallus is more or less monopodially organized and attached to the substrate by a circular holdfast (FIG. 4.71). At irregular intervals, the main long shoot produces second-order long shoots. Numerous determinate short shoots bearing spine-like outgrowths (?trichoblasts) extend from the long shoots and give the whole thallus a spiny appearance. The basic morphology of the spine-bearing short shoots of B. hakelensis closely resembles the extant red alga Pithyopsis tasmanica (Ceramiales) that occurs along the coasts of southern

chapter 4

Figure 4.70 Grateloupia sp. (Permian). Bar  2 cm.

Australia and Tasmania (Krings and Mayr, 2004). Another red alga from the Haqel fish beds is Delesserites lebanensis (Basson, 1981). This compressed, Delesseria-like thallus is 9.5 cm long and includes several blades that radiate from a common stipe (FIG. 4.72). Nothing is known about the holdfast or reproductive organs of this alga. The early Oligocene Haeringiella multifidiformis (FIG. 4.73) is characterized by a bladelike thallus attached to the substrate by a disk-like holdfast (Krings and Butzmann, 2005). The thallus is formed of a central axis that produces irregularly shaped lateral branches. Branches in the proximal portion of the thallus are closely spaced or clustered

algae

157

Figure 4.71 Bassonia hakelensis, young thallus with circular

basal holdfast (Cretaceous). Bar  3 cm. (Courtesy BSPG.)

and relatively undifferentiated, whereas those in the distal portion are irregularly forked and give off secondorder branches. Branch tips may be fringed. The most similar modern red alga is perhaps the gametophytic thallus of Sphaerococcus coronopifolius (Gigartinales). Various morphotypes believed to represent uncalcified red algae have been reported from middle and upper Miocene rocks from the Monterey Formation in California (Parker and Dawson, 1965). In general, they consist of planated axes with equally spaced laterals.

158

paleobotany: the biology and evolution of fossil plants

ACRITARCHA (ACRITARCHS)

Figure 4.72 Thallus of Delesserites lebanensis (Cretaceous). Bar  2.5 cm. (From Basson 1981.)

At one time, spiny vesicular microfossils were often termed hystrichosphaerids (a designation that alludes to their morphology) regardless of their biological affinities. In 1961, Evitt conclusively demonstrated that many hystrichosphaerids were the cysts of dinoflagellates. Consequently, the remaining “hystrichosphaerids” whose affinities remain uncertain or unknown are placed in the artificial group Acritarcha (Evitt, 1963 b,c). As a result the acritarchs represent a highly heterogeneous group of organic-walled vesicular microfossils (FIGS. 4.74–4.75) interpreted as (cysts of) protists of different biological affinities (Mendelson, 1993; Colbath and Grenfell, 1995; Strother, 1996; Montenari and Leppig, 2003), including the cysts of some naked dinoflagellates (Tappan, 1980). Other acritarchs, however, have been interpreted as multicellular (green) algae (Butterfield, 2004; Stanevich et al., 2007), algal (zygnematacean) spores

Figure 4.73 Haeringiella multifidiformis. Bar 1 cm. (From Krings and Butzmann, 2005.)

Figure 4.74 Dicrodiacrodium sp. (Ordovician). Bar  10 μm. (Courtesy M. Vecoli and T. Servais.)

chapter 4

Figure 4.75 Evittia sp. (Silurian). Bar  10 μm. (Courtesy M. Vecoli and T. Servais.)

(Grenfell, 1995), and fungi-like organisms (Butterfield, 2005); other possible origins outside of the phytoplankton have been reviewed by Colbath and Grenfell (1995). Microfossils assignable to the Acritarcha are among the first eukaryotes preserved in the fossil record (Huntley et al., 2006); they first occur in the Paleoproterozoic (Vidal and Moczydłowska-Vidal, 1997; Huntley et al., 2006) and extend to the Holocene (Mendelson, 1987). Acritarchs dominate the microfossil record in Proterozoic and Cambrian rocks. Some of the earliest accounts come from the Paleoproterozoic of China (Z. Zhang, 1997; S. Sun and Zhu, 2000). Other early acritarchs include the forms described from the Mesoproterozoic Roper (Javaux et al., 2001) and Bangemall (Buick and Knoll, 1999) Groups in Australia, and the Billyakh Group in northeastern Siberia (Sergeev et al., 1995). Paleo- and Mesoproterozoic acritarch assemblages are characterized by rather simple forms, whereas more complex forms typify Neoproterozoic assemblages (Sergeev et al., 1995). Slightly younger Proterozoic acritarchs have been described from shales of the Meso-Neoproterozoic Ruyang Group in China (L. Yin, 1998) and the Neoproterozoic Doushantuo Formation in South China (C. Zhou et al., 2001; Xiao, 2004). Acritarchs are variable in size (on average 5–200 μm) and shape (FIG. 4.78); the vesicle (body) of an acritarch ranges from oval to triangular in outline and may possess various forms of projections. They are classified in an artificial system of morphotaxa based on a complement of morphological characters, like those of spores (FIG. 4.76), including

algae

Figure 4.76 Emphanisporites annulatus Bar  10 μm. (Courtesy M. Vecoli and T. Servais.)

159

(Devonian).

Figure 4.77 Stelliferidium sp. (Ordovician). Bar  10 μm.

(Courtesy M. Vecoli and T. Servais.)

size and shape of the vesicle, number and form of projections, number of symmetry levels, and form of the exit rupture (Williams et al., 2000; Montenari and Leppig, 2003). Regardless of their natural affinities, the widespread occurrence and complex morphology of acritarchs, as well as their rapid rate of evolution, have made them extremely valuable in long-range correlation and biostratigraphic zonation

160

paleobotany: the biology and evolution of fossil plants

Type genera

Possible biological relationships

Acanthomorphitae (akantha  spine)

Microhystridium Multiplicisphaeridium Vulcanisphaera

?

Polygonomorphitae (Poly  gonia  polygonous)

Diexallophasis Goniosphaeridium Veryhachium

?

Sphaeromorphitae (sphaira  sphere)

Leiosphaeridia Synsphaeridium Trachysphaeridium

Netromorphitae (netron  spindle)

Deunffia Domasia Leiofusa

Herkomorphitae (herkos  fence)

Cymatiogalea Cymatiosphaera Dictyotidium

Diacromorphitae (di  akron  two tops)

Acanthodiacrodium Dasydiacrodium

Pteromorphitae (pteros  wing)

Duvernaysphaera Fimbriaglomerella Pterospermella

Prasinophyceae

Prismatomorphitae (primsa  prismatic)

Marrocanium Octoedryxium Polyedryxium

Prasinophyceae

Oömorpitae (oön  egg-shaped)

Aranidium Nothooidium Ooidium

Group

Morphology

Prasinophyceae

?

Prasinophyceae

?

?

Figure 4.78 Basic acritarch morphology. (Modified from Montenari and Leppig, 2003.)

(Smelror, 1987; Montenari and Leppig, 2003) and as indicators of climatic change (Mohr, 1990). In addition, acritarchs have been useful as paleoecological indicators of specific marine environments. For example, Staplin (1961) used several acritarch associations to predict distances from reefs in the Upper Devonian of Canada.

5 Hornworts and Bryophytes Early Fossil Evidence ........................................................163

Marchantiophytina (Liverworts or Hepatophytes) .......................... 167

Anthocerotophyta (Hornworts) .......................... 165

Bryophytina (Mosses) ..................................................................... 174

Bryophyta (Bryophytes) ...................................................166

These tiny bryophytes reveal their beauty slowly and up close, as do good friends. John Caddy, Morning Earth Poems

Prior to the development of the first efficient microscopes in the late eighteenth century, most people, including scientists, regarded hornworts, liverworts, and mosses as tiny flowering plants. The distinctness of these organisms first became widely acknowledged in the decades following the documentation of the bryophyte life cycle (FIG. 5.1) by Wilhelm Hofmeister (FIG. 5.2) in 1851. Beginning in the twentieth century, scientists became increasingly interested in the origin and evolution (including the fossil record) of the hornworts and bryophytes, and their relationships to other groups of fossil and modern plants. Bryophytes and hornworts are unique among extant embryophytes in that they have a gametophyte-dominant life cycle; the sporophyte usually is relatively short lived and permanently dependent upon the gametophyte. Physiologically, they are poikilohydric, meaning that they cannot control water loss; when the environment dries out, bryophytes also desiccate. Some are reported with endomycorrhizae (e.g., Y. Zhang and Guo, 2007). When moisture is available, they rehydrate; in other words, desiccation tolerance is physiological and not structural, as it is in most vascular plants. The plant body in these groups is a thallus—a simple plant body which is not differentiated into stems, roots, and leaves. Although they have no vascular tissue, some bryophytes have conducting cells in the form of hydroids and leptoids.

Hornworts and liverworts have been interpreted as occupying a position intermediate between the green algae and vascular plants (Smith, 1938). Others suggested that the bryophytes in general represented examples of evolutionary failures; perhaps they originated from early vascular plants, such as the rhyniophytes (Chapter 8), and then, as the group continued to evolve, vascular tissue was lost. Today, however, these views have changed on the basis of a variety of ultrastructural, biochemical, and molecular data (Duckett and Renzaglia, 1988) suggesting that the principal bryophyte groups had separate origins, and that hornworts, liverworts, and mosses represent the earliest divergent lineages of extant land plants, although the specific order of their divergence still remains unresolved (Friedman et al., 2004; Shaw and Renzaglia, 2004). Phylogenetic analysis strongly supports the liverworts as sister to all other land plants, and provides moderate to strong support for hornworts as the sister group to the vascular plants (Nickrent et al., 2000; Qui et al., 2006). The earliest recognizable bryophytes in the macrofossil record include liverworts that appear to have their closest affinities with the Metzgeriales (Krassilov and Schuster, 1984; Frahm, 2001a). True mosses are first encountered in the Mississippian, and appear to be well established by Permian time, including some modern orders (Neuburg, 1960a (FIG. 5.3); Jovet-Ast, 1967). Mosses easily referable to

161

162

Paleobotany: the biology and evolution of fossil plants

Sporangium with spore mother cells Sporophyte

Operculum Peristome

Gametophyte

Zygote

Calyptra Spore mother cells undergo meiosis, producing

Seta

2n Meiosis

Fertilization n

Spores Female gametophyte

Sperm

Bud

Protonema

Archegonium Egg

Antheridium

Male gametophyte “Leaf” Rhizoids

Figure 5.1 Life history of typical moss. (From Taylor and Taylor, 1993.)

modern genera do not appear until the Cretaceous. Hornworts are not encountered until the Cretaceous, and their occurrence is based almost exclusively on dispersed spores (N. Miller, 1980). The majority of Cenozoic bryophytes are assignable to modern genera, which may indicate that most of the modern families arose during the Cretaceous. Frey (1990) suggested that most of the leafy liverwort taxa as well as some mosses may have evolved in southern Gondwana and migrated north into tropical and Laurasian regions during the Cretaceous and Cenozoic.

Higher taxa in this chapter (based on the classification system of Frahm, 2001a):

Phylum Anthocerotophyta (hornworts); Dendrocerotaceae Phylum Bryophyta (bryophytes) Marchantiophytina (liverworts or hepatophytes) Treubiopsida Marchantiopsida (thalloid liverworts) Marchantiales (Continued)

CHAPTER 5

Sphaerocarpales Calobryales Ricciales Marchantiaceae, Ricciaceae Jungermanniopsida (leafy liverworts) Jungermanniales Lophoziaceae, Scapaniaceae, Frullaniaceae, Porellaceae, Lejeuneaceae Bryophytina (Mosses sensu lato) Sphagnopsida (Sphagnum mosses) *Protosphagnales Takakiopsida Bryopsida (Mosses sensu stricto; true mosses) Bryales Dicranales Dicranaceae Pottiales Leucodontales Hypnales Polytricales Polytricaceae, Dawsoniaceae

Hornworts and Bryophytes

Figure 5.2 Wilhelm Hofmeister.

Early Fossil Evidence The origin and early evolution of the bryophytes is a complex and perplexing problem (Goffinet, 2000). Phylogenetic evidence suggests that bryophytes in general, and liverwort-like plants in particular, should have been important components of early terrestrial floras (Bateman et al., 1998; Renzaglia et al., 2007). The fossil record of bryophytes, however, especially of early bryophytes, is meager and those known from fossils appear comparable in many ways to extant taxa. Even the earliest bryophyte-like fossils have the basic thallus organization also seen in many living forms. Based on this evidence, it is possible that the bryophytes evolved far earlier than the fossil record suggests, and fossils of the earliest bryophytes have not been found to date. However, it is possible that paleobotanists simply may not recognize the earliest bryophytes, because morphologically they do not resemble modern forms. One interesting hypothesis suggests that several of the enigmatic Cambrian to Devonian fossils traditionally included in the nematophytes (see Chapter 6) may represent remains of ancient liverworts which shared certain features with modern marchantioids (Graham et al., 2004). Some of these fossils consist of sheets of (pseudo-)cells (so-called cuticle) and tubes of some resistant material. Cosmochlaina (Silurian–Devonian) was originally considered to be of

Figure 5.3 Maria Neuburg. (Courtesy H. N. Andrews.)

163

164

Paleobotany: the biology and evolution of fossil plants

Figure 5.4 Cuticle of Cosmochlaina verrucosa (Devonian).

Bar  20 μm. (From Edwards, 1986.)

uncertain affinity, but has been interpreted as the lower epidermal surface of a marchantioid liverwort (FIG. 5.4). At least some of the tubular aggregations assigned to nematophytes (and possibly to Nematothallus) have been reinterpreted as masses of resistant liverwort rhizoids, and are sometimes attached to fragments of lower epidermal tissue. It has also been suggested that some of the dispersed Early Silurian cell sheets and tubes may represent fragmentary remains of bryophyte sporangia (Graham and Gray, 2001). One of the oldest reports of bryophyte-like megafossils is Parafunaria sinensis, a compression from the Early–Middle Cambrian Kaili Formation of China (Yang et al., 2004). The specimen is up to 2 cm long and comprises 4–5 leaves (each 5–15 mm long by 5 mm wide) that are densely borne around a short stem. The basal portion of the stem consists of a footlike structure, whereas distally it bears what is interpreted as a short seta with capsule. The authors suggest that the arrangement of leaves is similar to that in the extant moss Funaria hygrometrica (Bryopsida). Due to the age of these fossils, additional supporting evidence will be important in confirming their assignment to the Bryopsida. The study of dispersed spores (sporae dispersae) is an important source of data on the composition of the earliest land floras (e.g., Wellman and Gray, 2000; Steemans and Wellman, 2004). Although in many instances it is difficult to distinguish dispersed bryophyte spores from those of vascular plants, there is increasing support for the suggestion that some of the spores in Ordovician and Silurian rocks resemble those of modern liverworts. The spores most often suggested as bryophytic occur in permanent (sometimes envelope-enclosed) tetrads in these assemblages; it is hypothesized that such tetrads came from plants at a bryophytic, most likely a liverwort, grade of organization (Gray, 1985). Additional support for this hypothesis comes in the form of tiny spore-containing plant fragments from Ordovician rocks of Oman (Wellman et al.,

Figure 5.5 Reconstruction of Sporogonites exuberans. (From Taylor and Taylor, 1993.)

2003). These fossils indicate that the spore producers, although diminutive in size, were true land plants, which produced sporangia containing large numbers of spores—a minimum of 7450 tetrads in some specimens. Ultrastructural features of the spore wall also suggest affinities with the liverworts. In addition to microfossil remains, there are several types of meso- and macrofossils that have been considered important in understanding the early history of bryophytes (Edwards, 2000), including isolated sporangia with in situ spores and axes with conducting elements similar to those in extant bryophytes. The Early Devonian compression fossil Sporogonites (Halle, 1916a, 1936) morphologically resembles a bryophyte. This plant was originally found in Norway and consists of stalks 5 cm long which terminate in elongate capsules (FIG. 5.5). Several longitudinal furrows ornament the base of the sporangium and extend onto the stalk. The sporangium is multilayered, and there is some suggestion that it may contain a central columella-like projection. Inside the sporangia are trilete spores that range up to 30 μm in diameter. Many specimens of S. exuberans are preserved with the sporangial stalks in a more or less parallel orientation (Andrews, 1960), suggesting that they were produced from a common thallus; some stalks appear to be attached at the base to an irregularly shaped, carbonaceous film 15 cm long (FIG. 5.5). The fact that not all stalks bear

CHAPTER 5

Hornworts and Bryophytes

165

sporangia as well as the subtending axis are characteristically twisted (FIG. 5.6). Sporangial dehiscence is unknown, but suggested to have occurred along a preformed line that split the sporangium into two vertical valves. The morphology and spiraled architecture of the axis and sporangium closely resemble those in the living moss, Takakia ceratophylla (Takakiopsida) (Renzaglia et al., 1997). Morphological correspondence with members of the modern liverwort genus Pellia (Jungermanniopsida), however, have also been noted. Gerrienne (1997) pointed out that Tortilicaulis shares a number of important features with trimerophytes, and suggested that this taxon might therefore be ancestral to the Trimerophytina (Chapter 8). Although the affinities of many of these early fossils will continue to be controversial, fossil organisms that possess bryophytic characters (dominant gametophyte and parasitic, inconspicuous sporophyte) will play an increasingly important role, as paleobotanists continue to decipher steps leading to the colonization of the terrestrial realm.

Anthocerotophyta (Hornworts) Figure 5.6 Tortilicaulis transwalliensis sporangium with twisted stalk (Devonian). (From Taylor and Taylor, 1993.)

sporangia suggests that they were perhaps easily detached or abscised once the spores were mature. Although the compressed nature of the specimens of Sporogonites reveals little about the internal organization of the tissues, it appears that vascular elements are lacking, thus supporting its classification with the bryophytes. It is suggested that Sporogonites may represent a compressed gametophytic thallus bearing upright sporophytes of a primitive moss or perhaps an early hornwort (Poli et al., 2003). Both interpretations are supported by the suggestion of a columella-like projection within the sporangium, because this structure is exclusively known in hornworts and the putative primitive lineages of extant mosses (Goffinet, 2000). Crandall-Stotler (1984) suggested that the basal thallus may represent a persistent protonema, or even a small leafy gametophore, a hypothesis consistent with affinities to the mosses. Tortilicaulis is a Late Silurian–Early Devonian fossil that shares morphological features with Sporogonites and certain extant bryophytes (Edwards, 1979, 1996; Edwards et al., 1994). Specimens consist of unbranched or isotomously branched axis fragments that terminate in solitary or branched, elongate sporangia (FIG. 5.6) containing trilete spores. The

The hornworts include 300 living species (e.g., Duff et al., 2007) and differ from mosses and liverworts in the nature of their spore dispersal, which is accomplished by a longitudinal splitting of the capsule into several valves along with the action of pseudoelaters. The capsules (sporangia) characteristically have a central column of sterile tissue, called a columella. Hornworts produce symmetrical spermatozoids, and on the ventral surface of the gametophytic thallus are specialized, apically derived mucilage clefts surrounded by two cells resembling guard cells. Some authors have considered these homologous to the stomata which occur on the sporophytes (Renzaglia and Vaughn, 2000). In most genera, cells typically have a single, cup-shaped chloroplast (with pyrenoids) that is reminiscent of the chloroplasts seen in some green algae (Frahm, 2001a). There are no reliable reports of fossil hornworts prior to the Cretaceous, although they have been variously implicated as the oldest extant lineage of land plants (e.g., Renzaglia and Vaughn, 2000). The earliest macrofossil that bears some resemblance to a modern hornwort is perhaps Dendroceros victoriensis from the Lower Cretaceous Koonwarra Fossil Bed in Australia (Drinnan and Chambers, 1986). Morphologically it appears to be a sporophyte arising from a thalloid gametophyte. Another fossil that has been assigned to the hornworts is Notothylacites filiformis from the

166

Paleobotany: the biology and evolution of fossil plants

uppermost Cretaceous of Bohemia (Nˇemejc and Pacltova, 1974). Frahm (2005) noted, however, that this fossil displays a midrib, which does not occur in modern Notothylas, and thus may represent a Riccia-like thallus instead. Additional evidence of Cretaceous hornworts occurs in the form of spores. Three types of spores have been described from Maastrichtian (latest Cretaceous) rocks and compared to the extant genus Anthoceros (Phaeoceros) (Jarzen, 1979). All are trilete, circular, and possess a distinct cingulum; ornamentation is variable among all the three types. Other spores referred to the hornworts come from the Neogene of Hungary (Nagy, 1968) and Cenozoic from elsewhere in Europe and other sites (e.g., Lacey, 1967); these have been assigned to Phaeocerosporites, Rudolphisporis, and Saxosporis, morphogenera defined to accommodate spores that appear similar to those of extant hornworts. Several macrofossils assignable to the Anthocerotophyta have been recorded for the Cenozoic. The most complete representative (Frahm, 2005) is preserved in Dominican amber of early–middle Miocene age (Iturralde-Vinent and MacPhee, 1996). The specimen consists of a thallus segment (3 mm  3.5 mm long) bearing several cylindrical, hornlike sporophytes, which are similar to those in the Dendrocerotaceae (e.g., Dendroceros or Megaceros). Another fossil with possible affinities to the hornworts is Shuklanites deccanii from the uppermost Cretaceous Deccan Intertrappean beds exposed at Mohgaon Kalan village, Madhya Pradesh, India (Singhai, 1973). This specimen represents an isolated, 1.5 mm long, pear-shaped sporogonium (the sporophyte generation in bryophytes and hornworts), which consists of a short, bulbous foot and an elongate capsule containing thin-walled, trilete spores and abundant filamentous structures interpreted as pseudoelaters. A central columella, however, which is characteristic of most extant hornwort capsules, is not present. Two isolated sporogonia reminiscent of those in the extant hornwort Notothylas were also described from the same locality (Gupta, 1956; Chitaley and Yawale, 1980). Another interesting bryophyte from this site is Krempogonium mohgaoensis, an enigmatic fossil (5.5 mm long) composed of a twisted stalk and an oval capsule (FIG. 5.7), containing 10 vertically oriented spore sacs that are separated from each another by parenchymatous septa (Nambudiri et al., 2003). The spore sacs contain numerous small (25–30 μm in diameter) spores and pseudoelaters. The distal half of the capsule is covered by a smooth calyptra-like structure. Krempogonium mohgaoensis is interpreted as a bryophytic sporophyte that contains a mosaic of features seen in extant hornwort, liverwort, and moss sporophytes.

Figure 5.7 Reconstruction of Krempogonium mohgaoensis sporangium (Cretaceous). (From Nambudiri et al., 2003.)

Bryophyta (Bryophytes) Living bryophytes are represented by 900 genera and nearly 24,000 species. They are not a conspicuous portion of the Earth’s flora, although they may dominate the vegetation in certain special environments, for example Sphagnum in certain types of bogs. Most bryophytes are small plants, many 2 cm long. The largest forms rarely exceed 60 cm in length (e.g., species in the genus Dawsonia). In general, bryophytes are most abundant in relatively moist areas. They range throughout the world and can even be found in coastal areas of Antarctica. Bryophytes differ from true vascular plants by the absence of vascular tissue and by the presence of a nutritionally independent gametophyte generation in their life cycle (FIG. 5.1). Although bryophytes do not contain true vascular tissue, some have specialized conducting elements (Hébant, 1977) (FIG. 5.8). In certain mosses, the stem of the gametophyte and the seta (stalk) of the sporophyte contain elongate, non-lignified water-conducting cells called hydroids (analogous to xylem in vascular plants). Surrounding the hydroids are assimilate-conducting cells

CHAPTER 5

Hornworts and Bryophytes

167

Figure 5.9 Tetrapterites visensis (Mississippian). (From Taylor and Taylor, 1993.)

Figure 5.8 Charles Hébant. (Courtesy J. Galtier.)

termed leptoids, which are comparable to sieve elements in the phloem of higher plants. The fossil record of the Bryophyta is based on both spores and macrofossil remains. Numerous spores have been described as bryophytic, but most of these are sporae dispersae, which limits their taxonomic usefulness and may obscure their biological affinities. One of these is Tetrapterites visensis, an unusual dispersed structure isolated from Mississippian rocks (Hibbert, 1967). It consists of a tetrahedral, non-cellular membrane with winglike ridges; each ridge is attached to a single, trilete spore (FIG. 5.9). Sullivan and Hibbert (1964) compared Tetrapterites to the persistent tetrads of some liverworts, including Sphaerocarpus. The Paleozoic macrofossil record of bryophytes is poor. The earliest accepted specimens come from the Carboniferous (e.g., Walton, 1925, 1928; Oostendorp, 1987), and a few examples have been found in rocks as old as the Devonian (discussed below). Just why there is no extensive record of bryophytes in the Carboniferous is debatable, since the coal-swamp forests of the Pennsylvanian should have provided a wealth of suitable habitats for bryophytes. In addition, the preservation of other fossils from these paleoenvironments is often excellent (e.g., in coal balls). Some believe that the preservational potential of bryophytes is so poor that they are simply not preserved in sufficient numbers to be accurately recorded. Experiments conducted

by Hemsley (2001), however, show that the preservational potential of bryophytic plant material is similar to that of vascular plants, which suggests that bryophytes were indeed rare elements of the Carboniferous coal-swamp ecosystems. Until approximately 25 years ago, the fossil record of the Bryophyta consisted principally of vegetative remains of the gametophyte; with few exceptions, the sporophyte generation remained unknown, and almost no fossil bryophytes were known with identifiable sex organs. This situation has changed, however, because a number of well-preserved fossils of moss and liverwort sporophytes have been described from Mesozoic rocks (e.g., Konopka et al., 1997, 1998) and Cenozoic amber (e.g., Grolle, 1998; Frahm, 1999b, 2001b; Grolle and Schmidt, 2001) (Table 5.1; FIGS. 5.10–5.16). In several of the liverworts in amber, even the androecium, perianth, and gynoecium are preserved (Grolle, 1990, 1998). Although bryophyte remains in amber were noted as early as the nineteenth and early twentieth centuries (e.g., Göppert, 1853; Caspary, 1887; Dixon, 1922), only recently have these fossils received wider scholarly attention. Today, amber fossils represent the single most important source of evidence for the evolutionary history and biodiversity of bryophytes in the Cenozoic. Marchantiophytina (liverworts or hepatophytes)

Recent divergence-time estimates of the origin of the liverworts obtained using penalized likelihood suggest a Late Ordovician divergence of the liverworts (Heinrichs et al., 2007), based on a maximum age from Wellman et al. (2003) for the oldest fossils generally accepted as land plants, and Kenrick and Crane (1997a) for the oldest split of vascular plants. The earliest liverwort in the fossil record is

168

Paleobotany: the biology and evolution of fossil plants

Table 5.1 Bryophyte genera known from amber.

Source/Age

Group

Taxa

Source

Baltic (Eocene) and Bitterfeld (Eocene or Oligocene–lowermost Miocene)

Jungermanniales (leafy liverworts)

Bazzania, Calypogeia, Cheilolejeunea, Cylindrocolea, Frullania (FIG. 5.10), Jungermannia, Lophozia, Mastigolejeunea, Metacalypogeia, Nipponolejeunea, Notoscyphus, Plagiochila, Porella, Ptilidium (FIG. 5.11), Radula, Scapania, Spruceanthus

Surveyed in Grolle and Meister (2004a)

Dominican Republic (early–middle Miocene)

Jungermanniales

Archilejeunea, Bazzania, Blepharolejeunea, Bryopteris, Caratolejeunea, Cyclolejeunea, Cyrtolejeunea, Drepanolejeunea, Leucolejeunea, Lopholejeunea, Marchesinia, Mastigolejeunea (FIG. 5.12), Neurolejeunea, Radula, Stictolejeunea

Grolle (1984a, 1987, 1990, 1993), Gradstein (1993)

Baltic, Bitterfeld

Mosses (Bryophytina)

Aptychella, Atrichum, Barbella, Barbula, Bartramia, Bescherellea, Boulaya, Brachythecium, Brotherella, Calomnion, Campylium, Campylopodiella, Campylopus, Ctenidium, Dichodontium, Dicranum, Dicranites, Eurohypnum, Fabronia, Grimmia, Haplocladium, Hymenostomum, Hypnodontopsis, Hypnum, Mastopoma, Merilliobryum, Muscites, Phascum, Polytrichum, Rhizogonium, Rhytidiadelphus, Sematophyllites, Symphyodon, Trachycystis, Trichostomum, Tristichella

Frahm (1996a, b, 1999a, b, 2000, 2001b, 2004a, b, 2006b)

Dominican Republic

Mosses (Bryophytina)

Acroporiites, Adelothecium, Calymperes (FIG. 5.13), Calyptothecium (FIG. 5.14), Caribaeohypnum, Clastobryum, Entodon, Homalia, Hypnum, Leucobryum, Mittenothamnium, Mniomallia, Octoblepharum (FIG. 5.15), Orthostichella, Orthostichopsis, Plagiomnium, Porotrichum, Syrrhopodon (FIG. 5.16), Thuidium

Frahm and Newton (2005), Frahm (2006a)

Dominican Republic

Anthocerotophyta (hornworts)

Dendroceros or Megaceros

Frahm (2005)

Citations for ages: Dominican (Iturralde-Vinent and MacPhee, 1996); Baltic (Weitschat and Wichard, 1998; Knuth et al., 2002, see discussion in Schmidt and Dörfelt, 2007); Bitterfeld (see discussion in Dunlop and Giribet, 2003 and references cited therein).

Figure 5.10 Frullania schumannii (Eocene). Bar  4 mm. (Courtesy J.-P. Frahm.)

Figure 5.11 Ptilidium sp. (From Grolle and Meister, 2004b.)

CHAPTER 5

Hornworts and Bryophytes

169

Figure 5.14 Calyptothecium duplicatum (Miocene). Bar  6 mm. (Courtesy J.-P. Frahm.)

Figure 5.12 Mastigolejeunea bidentula (Eocene). Bar  1 mm. (Courtesy S. R. Gradstein.)

Figure 5.13 Calymperes palisoltii (Miocene). Bar  5 mm. (Courtesy J.-P. Frahm.)

Figure 5.15 Octoblepharum cylindricum (Miocene). Bar  3 mm. (Courtesy J.-P. Frahm.)

170

Paleobotany: the biology and evolution of fossil plants

Figure 5.16 Syrrhopodon incompletus (Miocene). Bar  3 mm. (Courtesy J.-P. Frahm.)

Metzgeriothallus sharonae from Givetian (upper Middle Devonian) shales and siltstones from New York (Van Aller Hernick et al., 2008). The fossils are preserved as carbonaceous films, and display dorsiventral thalli up to 32 mm long and 1.5 mm wide, that consist of a median costa and entire-margined wings. What appear to be unicellular, ribbon-like rhizoids extend from beneath the costa. Associated with the gametophytic thalli is an elongate sporophyte capsule with four valves. Another slightly younger liverwort is Pallaviciniites (Hepaticites) devonicus from the Upper Devonian (Frasnian) of New York (Hueber, 1961). The specimens consist of compressions preserved in a fine-grained shale together with numerous other plant remains. The liverworts were removed by bulk maceration (see Chapter 1) of the shale in concentrated hydrofluoric acid (HF); the carbonaceous, thalloid specimens were then floated onto microscope slides for examination. Pallaviciniites

devonicus is a simple, two-parted, flattened thallus with a central midrib and marginal lamellae or wings. The thallus is dichotomously branched, and along the margin of the wings are closely spaced teeth. The rhizomatous portion of the plant shows outlines of elongate parenchymatous cells, some bearing non-septate rhizoids. No reproductive structures are known. Various species of Pallaviciniites have been described from the Carboniferous to the Pleistocene, and they have been compared with such living genera as Pallavicinia, Metzgeria, Treubia, and Fossombronia (Schuster, 1966). Other late Paleozoic liverwort thalli come from the Carboniferous and have been assigned to morphogenera such as Blasiites, Metzgeriothallus, and Treubiites (He-Nygrén et al., 2006). Treubiites kidstonii from Scotland was initially believed to be similar to the extant Treubia, but later was shown to closely resemble extant Blasia because of its ventral scales (Krassilov and Schuster, 1984). Naiadita is a Triassic liverwort that was preserved in large numbers and many different stages of development. As a result, a great deal of information is known about the total biology of this bryophyte. The most comprehensive and detailed treatment of Naiadita is that of Harris (1938), based on specimens collected in Worcestershire and Warwickshire, England (Late Triassic). The plant is small (rarely exceeding 3 cm) and consists of an unbranched stem with helically arranged, lanceolate leaves (FIG. 5.17). Individual leaves are rounded at the apex (FIG. 5.18) and generally 1–5 mm long. Near the base of the stem are numerous, unbranched, non-septate rhizoids. Located along the stem are gemmae cups, which are specialized vegetative reproductive structures. They produced oval (500 μm in diameter) gemmae, which represent one of the most common components of the Naiadita fossiliferous beds. Gemmae are small pieces of thallus tissue, which can grow into a new plant. Some specimens possess stem-borne archegonia, which are 300 μm long and surrounded by a “perianth” of leaflike lobes. Although antheridia are not known, numerous stages in the development of the embryo and sporophyte are preserved. The fossil sporophyte of N. lanceolata consists of a short foot, slender stalk, and bulbous sporangium. The capsules are about 1.2 mm in diameter and contain spores in tetrahedral tetrads. The spores (100 μm in diameter) are lens shaped with an equatorial flange. On the proximal surface are numerous small, pointed spines; the distal surface bears larger, irregular projections. Dispersed spores with the same complement of morphologic features from rocks of equivalent age, referred to the genus Naiaditaspora, were examined at the fine-structural level (Hemsley, 1989a). The exine is organized into five distinct zones in which the inner

CHAPTER 5

Figure 5.17 Reconstruction of Naiadita lanceolata (Triassic). (From Taylor and Taylor, 1993.)

Figure 5.18 Leaf of Naiadita lanceolata (Triassic). (From Taylor and Taylor, 1993.)

regions contain numerous lamellae; the outermost region is granular. Based on a comparison with extant liverwort spores, the spores of Naiadita are most similar to members of the Marchantiales and Sphaerocarpales. The type of

Hornworts and Bryophytes

171

spores, presence of unicellular rhizoids, and organization of archegonia and capsules suggested to Harris (1938) that Naiadita represented a liverwort similar to extant Riella (Sphaerocarpales). Naiadita also shares some vegetative features with certain modern liverworts included in the Calobryales (Schuster, 1966). Late Paleozoic, Mesozoic, and Cenozoic impression and compression fossils of liverworts or liverwort-like thalli have been assigned to various morphogenera. These include: Thallites, for thalloid fossils that may represent liverworts or algae (Chapter 4), Hepaticites, for thalli that can confidently be assigned to the liverworts, but cannot be classified further, and Jungermannites, Metzgeriites, or Marchantites, for thalli that can be classified to the ordinal level within the hepatophytes (Cantrill, 1997b). Liverwort thalli in general are relatively rare as fossils, but there are several reports from the Mesozoic, in which bedding planes, which sometimes extend for several square meters, are covered with densely spaced thalli (e.g., Banerji, 1989; Pole and Raine, 1994). In many of the liverwort-rich beds, the thalli are preserved in situ, and therefore are interpreted as colonization horizons of freshly deposited sediment (Cantrill, 1997b). Beautifully preserved, compression fossils assignable to the Marchantiopsida occur in dense mats on bedding planes in the Aptian (Lower Cretaceous) of Spain (Diéguez et al., 2007b). These fossils consist of small, rosette-forming dichotomously branched thalli (FIG. 5.19). Marchantites cyatheoides and M. tennantii are impression fossils of thalloid liverworts referred to the Marchantiales from the Upper Triassic Molteno Formation in South Africa (Anderson, 1976). Marchantites tennantii has dichotomizing thalli (FIG. 5.20), in which individual branches range from 2.5 to 4 mm wide, each with a prominent midrib 1 mm wide. The lateral regions of branches have a surface pattern of polygonal areas between 0.75 and 1.5 mm wide that are arranged in rows arching away from the midrib. Similar patterns of regularly arranged polygonal fields are common in modern Marchantiaceae, where they represent the surface expressions of the subsurface air chambers. Four species of Marchantites were described from the Lower Cretaceous of Alexander Island, Antarctic Peninsula (Cantrill, 1997b), where they functioned as colonizers of fresh sediment near rivers and as an important part of the understory in both fern thickets and conifer forests. The taxa are distinguished based on thallus form, size, and the presence of features such as arcuate ribbing and air pores. Thalli of M. pinnatus are pinnate with short lateral branches ( 10 mm long), which have a prominent midrib and numerous rhizoids arising from the midrib region on the ventral

172

Paleobotany: the biology and evolution of fossil plants

Figure 5.19 Rosette-like dichotomizing thalli of marchantioid liverwort (Cretaceous). Bar  5 mm. (Courtesy C. Diéguez.)

Figure 5.20 Suggested reconstruction tennantii. (From Anderson, 1976.)

of

Marchantites

side. Marchantites rosulatus is a rosette-like thallus with individual thallus branches 3–5 mm wide. Thalli of M. taenioides are ribbon-like, up to 60 mm long, and only sparsely branched. On the dorsal surface are numerous circularelliptical pores; on the ventral side rhizoids are borne along the midrib. The fourth species, M. arcuatus, has thalli that display open branching. The affinities of these taxa within the Marchantiales are based on thallus morphology and the presence of structures resembling air pores on the thalli. Other liverwort thalli from Alexander Island were assigned

to the morphogenera Hepaticites and Thallites. Cantrill (1997b) noted that the liverworts of these late Albian floras were both diverse and abundant, and appeared to occupy a number of different ecological niches in this high latitude site. Based on an analysis of the δ13C of the Cretaceous liverwort thalli from Alexander Island, and subsequent comparisons of the results with modern analogs, Fletcher et al. (2005) showed that fossilized bryophytes can be used to gather information on paleoatmospheric CO2 concentrations, and thus offer new methods and insights into paleoclimatic reconstructions. Marchantiolites is an Early Cretaceous liverwort from central Montana (Brown and Robison, 1976). One specimen is 4 cm long and contains a prominent midrib. On the ventral surface are numerous rhizoids, whereas the dorsal surface displays air pores surrounded by specialized subsidiary cells. A slightly different air-pore frequency and morphology is present in the Rhaeto–Liassic (Late Triassic–Early Jurassic) species M. porosus from Scania, Sweden (Lundblad, 1954). As the name suggests, Marchantiolites has been included in the Marchantiales based principally on the general organization of the air pores. Lundblad (1954), however, noted that living species of Marchantia exhibit compound pores, whereas the fossil forms contain mostly simple ones. In general, the record of thalloid fossils interpreted as members of the Ricciaceae ranges from the Pennsylvanian (Walton, 1949a) to the Quaternary (Jovet-Ast, 1967); however, the affinities of most pre-Jurassic forms are still uncertain (Oostendorp, 1987). Fossil Ricciaceae are usually placed in the morphogenus Ricciopsis, but some have also been assigned to the modern genus Riccia (e.g., Sheikh and Kapgate, 1982). Ricciopsis florinii (Ricciaceae) is a rosette-shaped thallus from the Late Triassic–Early Jurassic of Sweden (Lundblad, 1954); it has four main branches, each of which dichotomizes twice. Although most records of fossilized ricciacean thalli come from Europe and Asia, to date only a single species has been described from North America, Ricciopsis speirsae from the Paleocene of Alberta, Canada (Hoffman and Stockey, 1997). This form differs from all other living and fossil Ricciaceae by displaying occasional constrictions and dilations of the repeatedly dichotomizing thallus. These fossils occur in what is interpreted as an oxbow lake deposit, along with lemnaceous angiosperms. Extensive occurrences of fossil Ricciaceae (e.g., Ricciopsis algoaensis) are known from the Lower Cretaceous (Berriasian–Valanginian) of South Africa (Anderson and Anderson, 1985), and dispersed ricciacean spores of Paleocene age have been reported as rare

CHAPTER 5

elements in the Sonda coal deposits in Pakistan (Leghari et al., 2001). Another interesting fossil liverwort is Diettertia, a Cretaceous form from Montana initially described as a moss (Brown and Robison, 1974). Based on additional material, Diettertia is now regarded as a bilaterally symmetrical, leafy liverwort with affinities to the Jungermanniales (Schuster and Janssens, 1989). The gametophyte consists of unistratose, bifid leaves inserted in two ranks on stems approximately 0.5 mm in diameter. Rhizoids are long and slender, non-septate, and up to 25 μm in diameter. Many features of Diettertia suggest that it represents a highly specialized member of the Jungermanniales. As a result of this fossil and other evidence, Schuster and Janssens (1989) suggested that the order probably evolved much earlier, perhaps in the late Paleozoic. Amber represents a valuable source of information about the Cenozoic biodiversity of the liverworts. The richest ambers containing liverwort remains come from the Baltic (Eocene), Bitterfeld in Saxony, Germany (Sachsen-Anhalt), dated as Eocene or Oligocene to lowermost Miocene, and the Dominican Republic (early–middle Miocene; IturraldeVinent and MacPhee, 1996). Amber is often very difficult to date. The Bitterfeld amber occurs in lower Miocene sediments (Barthel and Hetzer, 1982), but some authors have suggested that it is equivalent to the Baltic amber, based on the similarity of some faunal elements (see discussion in Dunlop and Giribet, 2003). Several dozen beautifully preserved specimens, representing more than 50 species, have been documented in recent years (e.g., Grolle, 1983, 1984a, b, 1987, 1993; Grolle and Braune, 1988; Gradstein, 1993; Grolle and Heinrichs, 2003; Grolle and Meister, 2004a, b). The liverworts in Baltic and Bitterfeld amber are almost exclusively leafy liverworts (Jungermanniales, Table 5.1) assignable to 17 extant genera (reviewed in Grolle and Meister, 2004b). Included are several specimens that provide insights into the reproductive biology of Cenozoic liverworts. For example, vegetative reproduction in the form of angular gemmae produced in globules at the tips of leaf lobes (FIG. 5.21) has been reported for Lophozia kutscheri (Lophoziaceae) from Bitterfeld amber (Grolle and Meister, 2004a). In a specimen of Scapania hoffeinsiana (Scapaniaceae) also from the Bitterfeld amber, both a cyathiform perianth (surrounded by involucral and subinvolucral leaves) and capsule (split to the base into four narrow valves) on a seta are preserved (Grolle and Schmidt, 2001). Frullania baltica (Frullaniaceae) from Baltic amber shows a capitate androecium that is positioned on a very short branch (Grolle, 1998). Leaves arise in six leaf circles below a clavate, beaked perianth, which is

Hornworts and Bryophytes

A

173

B

Reconstruction of Lophozia kutscheri showing dorsal surface (left) and gemmae at tips of leaves. (Modified from Grolle and Meister, 2004a.)

Figure 5.21

Figure 5.22 Porella subgrandiloba branch; ventral (left) and dorsal view (Eocene). (From Grolle and So, 2004.)

positioned terminally on the main axis and surrounded by involucral and subinvolucral leaves. At the tip of one perianth beak, the globose capsule of a developing sporophyte is visible. Archegonial branches (FIG. 5.22) with dentate bracts and bracteoles of Porella subgrandiloba (Porellaceae) are also known from Baltic amber (Grolle and So, 2004). Antheridial and archegonial reproductive structures and/or “perianth parts” have been recorded for Leucolejeunea antiqua (Grolle, 1990), Drepanolejeunea eogena (Grolle, 1993), and Mastigolejeunea auriculata (Gradstein, 1993) from Dominican amber. Although the former two taxa are only known as fossils, the latter, along with the amber fossils of Marchesinia brachiata and Stictolejeunea squamata (Lejeuneaceae) from the Dominican Republic, represent

174

Paleobotany: the biology and evolution of fossil plants

fossils of species that still exist today, and thus indicate that these species are archaic and already existed at least by the early Neogene, 20 Ma (Gradstein, 1993). Bryophytina (mosses)

The fossil record of the mosses is less complete than that of the liverworts. Krassilov and Schuster (1984) suggested that perhaps the earliest mosses evolved rapidly into droughttolerant forms that occupied sites where fossilization was unlikely. Nevertheless, there is some record of mosses as early as the Carboniferous (e.g., Walton, 1928). One of these is Muscites plumatus, an impression of a small leafy shoot in rocks of Mississippian age (Thomas, 1972). The axis is covered with helically arranged leaves with clasping bases. Each leaf is about 7.5 mm long and terminates in an elongate tip. Sex organs, sporophyte capsule, or rhizoids were not present. Additional species of Muscites have been described from the Pennsylvanian of France (Renault and Zeiller, 1885, 1888) and the Triassic of Africa (Townrow, 1959; Anderson, 1976). Although there are only a few records of mosses from the Carboniferous, numerous species have been described from Permian rocks, including an extensive moss flora from Siberia (Neuburg, 1960a). From this flora of well-preserved vegetative shoots, Neuburg described six genera that she assigned to the Bryales and three to a new order, the Protosphagnales. Protosphagnum (FIG. 5.23) has leaves similar to those of the extant genus Sphagnum, except for the presence of a midrib. Ignatov (1990) has also described a diverse flora of well-preserved mosses from the Upper Permian of the Russian Platform. The dozen specimens represent remains of the gametophyte generation and include forms that are referable to the extant orders Dicranales, Pottiales, Funariales, Leucodontales, and Hypnales. To date the best documented permineralized late Paleozoic moss comes from the Permian of Antarctica (Smoot and Taylor, 1986a). Merceria augustica consists of delicate axes 1 mm in diameter to which are attached helically arranged leaves and numerous rhizoids (FIG. 5.24). Each leaf is unistratose and contains a thickened midrib (FIG. 5.25). Although no evidence of reproductive organs or sporophytes was found associated with the specimens, the shape of the leaf cells, structure of the leaf margin, and anatomy of the axes suggest affinities within the Bryales. Megafossils of Triassic and Jurassic true mosses are relatively rare and consist almost entirely of compression specimens, although these plants are often represented in palynofloras (e.g., Zavattieri and Volkheimer, 2003).

Figure 5.23 Reconstruction of Protosphagnum nervatum (Permian). (From Taylor and Taylor, 1993.)

Figure 5.24 Cross section of Merceria augustica showing

leaves near apex (Permian). Bar  110 μm.

Sphagnophyllites triassicus consists of isolated Sphagnum-like leaves from the Triassic of India (Pant and Basu, 1978). Another Mesozoic moss is Tricostium, a small Late Jurassic– Early Cretaceous specimen from Siberia, which consists of a stem densely covered with imbricate leaves, each approximately 1.2 mm long (Krassilov, 1973a); on some leaves are elongate pores. In another form from the same area, Yorekiella, the awl-shaped leaves are two or three ranked and lack midribs.

CHAPTER 5

Hornworts and Bryophytes

175

Figure 5.25 Detail of Merceria augustica leaf (Permian).

Bar  110 μm.

Figure 5.27 Sporophyte of Campylopodium allonense showing a lateral view of the capsule and partially attached calyptra (Cretaceous). Bar  100 μm. (From Konopka et al., 1998; courtesy P. S. Herendeen.)

Figure 5.26 Diagrammatic view of capsule of Eopolytrichum

antiquum. (From Konopka et al., 1997.)

Some of the most exquisitely preserved younger Mesozoic mosses (sporophytes and gametophytes) come from the Late Cretaceous (Santonian) of North America. Gametophytes of Eopolytrichum antiquum (Polytrichaceae) (Konopka et al., 1997) have broadly lamellate leaves densely arranged around the stem. Some specimens have male “inflorescences” in the form of conspicuous rosettes of overlapping perigonal bracts, antheridia, and clavate paraphyses. Female gametophytes have not been found. Sporophytes (FIG. 5.26) consist of oblong capsules, which are terete in cross section and somewhat flattened dorsiventrally. The capsule wall consists of bulging-mammillose exothecial cells with abruptly thinned outer periclinal walls. The operculum is rounded, and an annulus is missing. The peristome consists of a short

Figure 5.28 Lateral view of peristome of Campylopodium allonense showing teeth with divided tips (Cretaceous). Bar  100 μm. (From Konopka et al., 1998; courtesy P. S. Herendeen.)

peristomal membrane, which originates just within the rim of the capsule. Peristomal teeth are apparently lacking, but cells in the circumference of the peristomal membrane suggest that, if this species had developed teeth, there would have been 32 peristome teeth. Stomata are restricted to the apophysis and resemble those seen in extant Polytrichaceae. Campylopodium allonense (Dicranaceae) (Konopka et al., 1998) has oblong, curved, and nodding capsules (FIG. 5.27) that display a distinct basal stomatiferous swelling. The operculum is obliquely rostrate (beaked) with a smooth, cucullate (hood-shaped), or cone-shaped calyptra. Beneath the operculum is a peristome composed of a single cycle of 16 bifid teeth (FIG. 5.28). The spores are spherical, 10–12 μm

176

Paleobotany: the biology and evolution of fossil plants

Figure 5.29 Hypnum lycopodioides (Oligocene). Bar  1 cm. (Courtesy C. Gee and G. Oleschinski.)

in diameter, and display a delicate rugose sculpture. Associated sterile gametophytes consist of leafy stem portions; fertile gametophytes have not been found. The Cenozoic record of mosses is considerably more extensive than that of the Paleozoic and Mesozoic, and includes impression–compression specimens (FIG. 5.29– 5.31) and amber fossils (N. Miller, 1980, 1984) (Table 5.1). Aulacomnium heterostichoides from deep-water varved clays (Eocene) of a freshwater lake in British Columbia (Janssens et al., 1979) is an example of a well-preserved compression of a moss gametophyte. The plant is irregularly branched (FIG. 5.30) and contains helically arranged, elliptical leaves with multicellular teeth along the upper half of the margin. The cells of the upper laminal surface are regularly isodiametric and unipapillose, whereas those of the basal surface are rectangular with slightly thicker walls. Based on an analysis of characters in existing populations of Aulacomnium, the fossil is believed to be most closely related to A. heterostichum, a living species found in eastern North America and eastern Asia. Numerous Cenozoic mosses are preserved in amber from the Baltic, Bitterfeld (Germany), and the Dominican Republic (e.g., Frahm, 1993, 1994, 1996a, b, c, 1999a, b, 2000, 2001b, 2004a, b, 2006a, b; Frahm and Reese, 1998; Frahm and Newton, 2005) (Table 5.1). The mosses present in the amber can either be assigned to modern species or represent fossil species belonging to modern genera; others cannot be referred to any modern species, and are therefore assigned to morphogenera such as the compression genera Dicranites (FIG. 5.31), Hypnites, and Muscites. At least 30 genera of mosses have been recorded from Baltic and Saxon amber. Several of the moss species preserved in Baltic amber (e.g., Trachycystis flagellaris) still exist today, but their occurrence is restricted to eastern and/ or southeastern Asia. This suggests that these forms became

Figure 5.30 Several leafy branches of Aulacomnium heterosti-

choides (Eocene). Bar  2 mm. (From Taylor and Taylor, 1993.)

Figure 5.31 Dicranites rottensis (Oligocene). Bar  3 mm.

(Courtesy C. Gee and G. Oleschinski.)

extinct in Europe as a result of the Quaternary climatic changes, but persisted in Asia (Frahm, 2001a). Other mosses from Baltic amber (e.g., Haplocladium angustifolium) represent species that today still occur in Europe, but are restricted to the Southern Alps as Cenozoic relicts. The inventory of mosses preserved in Dominican amber includes representatives of almost 20 genera (Frahm and Newton, 2005; Frahm,

CHAPTER 5

2006a). Most of the species are also known today from the Neotropics, which suggests that the foundation of the neotropical moss flora was in existence as early as the Paleogene (45–25 Ma) (Frahm, 2001a). Most mosses in amber occur as sterile gametophytes or gametophyte fragments; fertile gametophytes and sporophytes are comparatively rare. One fossil example of a gametophyte with attached sporophyte is Dicranites grollei from Eocene Baltic amber (Frahm, 1999b). This fossil consists of the distal portion of a stem with 14 linear leaves, each up to 1.8 mm long. The stem apex bears a twisted seta that terminates in a round, distally constricted capsule, which displays 6 (probably originally 16) short, lancet-like, undivided peristome teeth. Another example that includes gametophyte and sporophyte in organic connection is Hypnodontopsis conferta (Baltic amber) (Frahm, 2001b, 2004a). One of the specimens is a complete plant, consisting of a stem bearing numerous linear leaves and a complete sporophyte (FIG. 5.32). The seta is 1.5–2 mm long, twisted, and cygneous (shaped like the neck of a swan). It terminates in a short, oval capsule with 16 longitudinal ribs. The capsule is open and displays the peristome, where the 16 peristomal teeth are fused into eight pairs. Complete plants displaying gametophyte and sporophyte in organic connection have also been reported from Saxonian amber (e.g., Campylopodiella cf. himalayana; Frahm, 1996b).

Hornworts and Bryophytes

177

Figure 5.32 Hypnodontopsis conferta (Eocene). Bar  1.5 mm. (Courtesy J.-P. Frahm.)

This page intentionally left blank

6 The move to the land Enigmatic organisms .......................................................180

The transition to land ................................................194

Nematophytes....................................................................................180

Anchorage and Water Uptake ......................................................... 194

Spongiophytaceae .............................................................................185

Structural Support and Water Transport.......................................... 195

Other Enigmatic Organisms ..............................................................186

Protection Against Desiccation and Radiation ................................ 195

Isolated fragments: clues to the transition

Gas Exchange.................................................................................. 195

to land? ....................................................................................... 189

Reproduction on Land..................................................................... 196 Life History Biology ....................................................................... 196

Cuticle and Cuticle-like Material ......................................................189

Animals ............................................................................................198

Spores and Spore Tetrads ..................................................................189

A Fungal Partner ..............................................................................198

Tubes .................................................................................................192

Conclusion.............................................................................. 199

Land plant ancestors ...................................................... 193

All theory, dear friend, is grey, but the golden tree of life springs ever green. Johann Wolfgang von Goethe, Faust: The Tragedy, Part I Although land surfaces must have been available for colonization soon after life evolved in the Precambrian, the earliest record of terrestrial animals begins around 425 Ma (Ward et al., 2006) and that of plants somewhat earlier, with microfossil evidence from the Ordovician (discussed below). This delay in colonization of terrestrial habitats has been related to oxygen levels in the paleoatmosphere (Berner et al., 2007) and specifically to the lack of a sufficient ozone shield to protect terrestrial organisms from ultraviolet (UV) radiation, which acts as a strong mutagen in organisms. Because UV radiation can penetrate only a few centimeters into water, algae and other forms of marine life were protected from its effects. Ozone (O3) is formed in the stratosphere by a reaction between free oxygen (O2) and UV radiation, so oxygen levels needed to reach a certain level for an ozone layer to develop. Once formed, the ozone layer would mitigate the lethal effects of UV radiation and allow for colonization of the land. Since plants are rooted in place, they would have had to develop mechanisms to shield their cells from UV

radiation very early in evolution. Today plants have a number of adaptations that help to filter UV radiation, including pigmentation, secondary compounds, and cuticle. It is not known precisely when photosynthetic organisms began to exploit terrestrial habitats. Some have suggested that cyanobacteria existed in shallow pools on the earth’s surface relatively early in geologic time (Westall et al., 2006b; see Chapter 2). We do know, however, that beginning in the Ordovician there is evidence of organisms with characters suggesting that they lived in a desiccating environment. For many of these organisms, it is not known whether they existed for all or part of their life history on the land, but many demonstrate features that are today found only in land plants. Although the primary aim of paleobotany is to identify plants in the fossil record and to understand their phylogeny, biological activities, and paleoenvironment as much as possible, in some instances it is impossible even to place organisms in a major hierarchical category. This is true for a number of these late Paleozoic (mainly Silurian and Devonian)

179

180

PALEOBOTANY: the biology and evolution of fossil plants

fossils that exhibit some but not all features of land plants, (Edwards et al., 1998a) even though some are well preserved and have been intensively studied. Some of these organisms have been compared to vascular land plants, whereas others have a morphology or internal organization different from any other living or fossil group (Edwards, 1982). We discuss several of these unusual organisms in the section that follows.

Enigmatic organisms

internal organization constructed of longitudinally oriented tubes (hyphae, according to Hueber, 2001) of three different types (FIGS. 6.2, 6.3). Adopting the terminology used for hyphal types in trimitic extant basidiomycetes, that is those have three types of hyphae, Hueber (2001) distinguished (1) skeletal hyphae, which are thick-walled, large, long, straight or flexuous, aseptate, and unbranched; (2) generative hyphae, which are large and thin-walled, septate with an open or occluded pore, and profusely branched (FIG. 6.2); and (3) binding hyphae, which are small, thin-walled with a pore in the septum, and profusely branched. Arrangement of the

Nematophytes

The nematophytes or Nematophyta (Strother, 1993) is an informal grouping of enigmatic Silurian–Devonian organisms that range from tiny millimeter-sized structures to severalmeter-long “logs.” Their plant bodies are constructed entirely of variously sized and shaped tubes. Although nematophytes have been studied intensively for more than 150 years, little is known about their systematic affinities, biology, and ecology. PROTOTAXITES This impressive organism is the largest representative of the Nematophyta known to date. It occurs in Silurian and Devonian rocks in the form of compressed or silicified axes (sometimes also termed pseudostems), some of which are up to 1.25 m in diameter (FIG. 6.1) and more than 8 m long (Arnold, 1952b; Jonker, 1979; Bahafzallah et al., 1981; Schweitzer, 1983, 2000; Chitaley, 1992; Hueber, 2001). Charcoalified pieces of Prototaxites have been reported from the Lower Devonian of the Welsh borderland and interpreted as products of wildfire activity (Glasspool et al., 2006). The outer surface of Prototaxites is smooth or mildly ribbed. Thin sections show a pseudoparenchymatous or plectenchymatous

Figure 6.1 Cross section of Prototaxites showing eccentric incremental growth (Devonian). Bar  3 cm.

B

A

C

Figure 6.2 Longitudinal section of Prototaxites southworthii

showing three types of tubes: A. binding, B. generative, and C. skeletal (Devonian). Bar  50 μm.

Figure 6.3 Section of Prototaxites southworthii showing large skeletal and narrow binding tubes/hyphe. (Devonian). Bar  50 μm.

CHAPTER 6

“tissue” made up of these hyphae, along with the presence of well-defined borders of growth increments (FIG. 6.1) marked by increased tissue density, suggests some type of periodicity in growth. In addition to the tubes, numerous lacunae are scattered throughout the pseudoparenchyma. These spaces often contain disintegrated tissue and occasionally tubes, and are interpreted as formed as a result of some type of parasite. When this gigantic nematophyte was initially discovered, the specimens were believed to represent some type of fossilized conifer wood, and the organism was named Prototaxites (Dawson, 1859). After closer examination, however, the fossils were reinterpreted as algal like, transferred to another genus (Nematophycus), and formally classified with the Codiaceae (green algae). Prototaxites has also been compared to the brown alga Lessonia (Carruthers, 1872b; Kräusel, 1936). When tubes of P. southworthii were examined ultrastructurally (Schmid, 1976), however, distinct cross walls (septa) were identified, each with a centrally located elliptical aperture or pore (FIG. 6.4). These pores are superficially similar to the pores and pit connections found in certain red algae, but corresponding structures also occur in fungi (e.g., basidiomycetes) in the form of doliporous septa. In some of the tubes or hyphae, small outgrowths occur close to the septa that remotely resemble basidiomycetous clamp connections (Hueber, 2001). Both mycologists and phycologists have debated the affinities of Prototaxites for many years. Early chemosystematic data suggested an affinity with the algae, whereas the presence of cutin and suberin in the samples implies that the organisms may have been terrestrial. Evidence from other biomarkers has been used to hypothesize that

Figure 6.4 Ultrathin section of Prototaxites tube showing sep-

tal pore (arrow) (Devonian). Bar  2 μm. (Courtesy R. Schmid.)

The move to the land

181

Prototaxites represents a failed experiment during terrestrialization (Abbott et al., 1998), as Lang (1937) originally suggested, or that it belongs to one of several algal groups which were in the process of adapting to a terrestrial habitat during the Devonian, but failed to survive to the present time (Niklas and Smocovitis, 1983). Another interpretation of Prototaxites has been offered by Schweitzer (1983, 1990) based on a reinterpretation of Mosellophyton hefteri (FIG. 6.5) a large, irregularly branched axis from the Lower Devonian of the Mosel Valley in Germany (Wehrmann et al., 2005), which was originally interpreted as a tracheophyte (Schaarschmidt, 1974). Schweitzer (1983) combined M. hefteri with Prototaxites, and provided a reconstruction (FIG. 6.5) that depicts a kelp-like alga consisting of a monopodial trunk (Prototaxites) with a distal, multibranched

Figure 6.5 Suggested reconstruction of Prototaxites based on isolated “stems” and Mosellophyton hefteri. (From Schweitzer, 1983; Courtesy U. Schweitzer and R. Gossmann.)

182

PALEOBOTANY: the biology and evolution of fossil plants

Figure 6.7

Francis M. Hueber.

Figure 6.8 Hypothetical reconstruction showing the habit and

habitat of Prototaxites. (By M. Parrish; from Hueber, 2001.) Figure 6.6 Basal holdfast of Prototaxites hefteri (Devonian).

Bar  2.5 cm. (From Schweitzer, 2000; courtesy U. Schweitzer and R. Gossmann.)

crown (Mosellophyton). The alga is believed to have inhabited shallow, tidally influenced coastal marine waters where it was attached by a root-like holdfast. More recently, a branched Prototaxites specimen (FIG. 6.6) was discovered

in the Waxweiler quarry in Germany, and interpreted as a portion of the basal holdfast of this organism (Schweitzer, 2000). It should be noted, however, that M. helteri and Prototaxites have never been found in organic connection, and it remains questionable as to whether they actually belong to a single organism.

CHAPTER 6

The most recent interpretation of Prototaxites is the comprehensive paper on P. loganii by Hueber (2001) (FIG. 6.7), who hypothesizes that the large “logs” represent a giant (FIG. 6.8), terrestrial, saprotrophic organism that may belong with the basidiomycetes (Chapter 3). Thus far, however, there are no easily identified structures that would support the assignment to the basidiomycetes or any other fungal group. Carbon isotope analyses of Prototaxites and land plants that lived in the same environment indicate that Prototaxites has a much wider variation in its C12/C13 ratio than would be expected in any plant (i.e., autotroph), indicating that Prototaxites was a heterotrophic organism (Boyce et al., 2007). Although a heterotroph the size of Prototaxites remains a possibility, the structure of the typical Late Silurian–Early Devonian ecosystem, which consisted of a sparse vegetation of exceedingly small plants and probably algal and cyanobacterial growths, brings into question whether there would be a sufficient source of carbon in the ecosystem to support a heterotrophic organism more than 8 m tall. Still another suggestion is that Prototaxites is an example of an ancient mutualistic association of two (or more) different organisms that combined both heterotrophy and some level of a lichen-like nutritional mode (Selosse, 2002). Although there is no direct evidence of an associated photobiont within the permineralized specimens, the hypothesis has some merit. Perhaps the photobiont consisted of cyanobacteria that were associated with the outer surface of the organism and that have either not been preserved or simply unidentified to date. Regardless of whether Prototaxites was an alga, a fungus, a lichen-like association of several different organisms, or some other type of life form not remotely related to any modern organism, it must have presented an imposing structure. As increased attention is directed at the evolution of early terrestrial ecosystems, it will be interesting to see where the affinities of this unique Paleozoic organism eventually reside. NEMATOTHALLUS Perhaps the earliest detailed account of an enigmatic organism initially believed to represent an early land plant was Lang’s (1937) description of Nematothallus from the Lower Devonian of the Welsh borderland. Specimens of Nematothallus are flat, thallus- or leaf-like, and only a few centimeters in diameter. Like Prototaxites, they are composed of a system of interlacing tubes, often of two distinct orders of size (some with internal thickenings), and usually covered by a cuticle with a pseudocellular pattern (Strother, 1988, 1993). Isolated cuticle sheets of this type are known from the Ordovician to the Early Devonian, and have flanges on the inner surface that would be interpreted as marking

The move to the land

183

the outlines of epidermal cells if found on a vascular plant. Scattered among the tubes and within the cuticle are spores of various sizes (Strother, 1993). It is not entirely clear, however, whether the spores, tubes, and cuticle-like layers actually represent parts of the same organism. Perforations up to 100 μm in diameter are present on some of the cuticles and these possess various types of thickenings near the surface. Like the pores found in other enigmatic organisms and modern liverworts, those of Nematothallus may have functioned in some form of gas exchange. Edwards and Rose (1984) noted, however, that similar pores have not been found in all Nematothallus cuticles. Other possible functions for the cuticle pores in Nematothallus, as well as those found in some other Paleozoic enigmatic organisms of this type, include sites of gamete liberation or a response to wounding. Corsin (1945), Jonker (1979), and others have suggested that Nematothallus represents the remains of leaf-like structures that were produced on Prototaxites-like axes. Although this is an interesting hypothesis, no specimens in organic connection are currently known that might support any of these hypotheses. Strother (1993), after examining a large assemblage of Nematothallus-like organisms, suggested that the construction of these organisms may be indicative of a subaerial habitat. NEMATOPLEXUS Nematoplexus rhyniensis is a silicified, thalloid fossil known only from incomplete, partially decayed, and generally fragmentary remains from the Rhynie chert (Lower Devonian). The overall morphology and size of the organism remain unknown. It consists of a meshwork of interlaced aseptate and septate tubes of two or three different sizes (Lyon, 1962). Most of the larger, aseptate tubes have annular or helical thickenings; some of the narrow, smooth-walled tubes display distinct septa with a centrally located elliptical aperture or pore (FIG. 6.9). These pores are similar, if not identical, to the pores found in Prototaxites. Nematoplexus tubes apparently do not branch, except for certain areas termed branch knots in which aggregations of tubes form dense, multibranched clusters (FIG. 6.10). The biological significance of the branch knots remains unknown. It has been suggested that N. rhyniensis may represent the permineralized equivalent of Nematothallus. A second, even less well-documented and understood nematophyte from the Rhynie chert has been given the name Nematophyton taiti (Kidston and Lang, 1921a). NEMATASKETUM DIVERSIFORME This Devonian organism is structurally similar to Prototaxites (Burgess and Edwards, 1988) and is based on coalified, three-dimensionally preserved specimens. It also

184

PALEOBOTANY: the biology and evolution of fossil plants

Figure 6.9 Nematoplexus rhyniensis wide tube with helical

thickenings and narrow tube with septal pore (arrow) (Devonian). Bar  100 μm. (Courtesy H. Hass.)

Figure 6.11 Fractured surface of Nematasketum diversiforme showing large tubes and smaller interwoven filaments (Devonian). Bar  100 μm. (From Burgess and Edwards, 1988.)

Figure 6.10 Nematoplexus rhyniensis with branched knot

(Devonian). Bar  45 μm. (Courtesy H. Hass.)

consists of tubes of two sizes (FIG. 6.11), including some with internal thickenings. In attempting to decipher the affinities of both Nematasketum and Prototaxites, Burgess and Edwards (1988) proposed either fungal or algal affinities,

but offer equally plausible reasons to refute inclusion in either group. Instead, they place both genera in the Nematophytales, an order erected by Lang (1937). Burgess and Edwards (1991) later developed an artificial classification system, based on one used for dispersed palynomorphs, to aid in description and comparison of some of these organisms constructed of various types of tubes.

CHAPTER 6

The move to the land

185

Figure 6.12 Diagrammatic view of Pachytheca showing two-

layered wall. (From T. Taylor, 1988a.)

PACHYTHECA This genus is known from the Late Silurian to the Middle Devonian (Barber, 1889, 1891; Stockmans and Willière, 1938; Lang, 1945; Schmidt, 1958), and is often cited in discussions of possible land plant origins. Pachytheca specimens are small (0.1–0.6 cm in diameter), spherical, and organized into two distinct zones (Niklas, 1976c; Gerrienne, 1991). The inner zone (medulla) consists of densely spaced, intertwined tubes (FIG. 6.12), whereas the outer layer (cortex) is constructed of tubes that are distinctly radially aligned. Between cortex and medulla is a narrow area in which the tubes seem to change from random to radial alignment. In some specimens, a narrow canal is visible that appears to connect the medulla with the exterior. Pachytheca has repeatedly been interpreted as a dispersal unit (cystocarp) of Prototaxites (Schmidt, 1958; Jonker, 1979; Schweitzer, 1983); however, fossils in organic connection have not been discovered to date. Niklas (1976c) suggested that Pachytheca represents the juvenile form of Parka decipiens (discussed below), and thus the two taxa are merely different ontogenetic stages of a single organism. In an interesting set of experiments, Graham et al. (2004) observed extant specimens of the liverworts Marchantia and Conocephalum at various stages of controlled tissue degradation. Based on the results of these experiments, they suggest that some of the enigmatic fossils commonly referred to as nematophytes may represent remains of ancient liverworts in various stages of decay and that some of these plants shared certain features with modern marchantioids. Although these authors do not suggest that extant liverworts existed in the Devonian, they discuss the possibility that some of the cuticle sheets and tubes of these extinct forms may represent characters that have persisted during the evolution of the group (see Chapter 5).

Figure 6.13 Airpores (arrows) in Marchantia thallus. (Extant)

Bar  100 μm.

Spongiophytaceae

The Spongiophytaceae an artificial taxon used to include a variety of enigmatic early organisms, some of which may represent or be related to vascular land plants (Kräusel, 1954; Kräusel and Venkatachala, 1966). Some organisms included in this family show features of vascular plants, but not all; as a result, they are sometimes interpreted as depicting transitional stages in the evolution of true vascular land plants. SPONGIOPHYTON Spongiophyton is a thalloid fossil that has been reported from numerous Devonian localities throughout the world (Kräusel, 1954; Kräusel and Venkatachala, 1966; Boureau and Pons, 1973; Chaloner et al., 1974; Zdebska, 1978; Gensel et al., 1991; Guerra-Sommer, 1993; Griffing et al., 2000), and has variously been interpreted as a colonial animal, alga, vascular plant, and bryophyte. Specimens are small (25 mm wide by 2.5 cm long), branched axes with rounded tips (FIG. 6.14). It has been suggested that Spongiophyton grew in a desiccating environment, based on the presence on one surface of the thallus of circular pores ranging 200–300 μm in diameter; these have been compared to the pores in certain modern liverworts like Marchantia (FIG. 6.13). The pores are believed to have functioned in gas exchange, and perhaps represent a transitional stage in the evolution of stomata (Chaloner et al., 1974). Additional support for life in a desiccating environment is the presence of a cuticle (or cuticlelike surface layer) covering the Spongiophyton thallus. On the thallus surface which contains the pores, the cuticle is approximately three times thicker than on the non-poral

186

PALEOBOTANY: the biology and evolution of fossil plants

surface (FIG. 6.14). Nothing is known about the internal tissues nor the reproductive organs of this organism, although it has been suggested that perhaps there was a hyphal network like that found in certain lichens (Stein et al., 1993; Taylor, W. A. et al., 2004). The lichen hypothesis was initially supported by carbon isotope ratios from Spongiophyton (Jahren et al., 2003). Fletcher et al. (2004), however, found that the range of δ13C values in Spongiophyton is not significantly different from those of lichens, liverworts, or mosses, so carbon isotopes are not useful in identifying this fossil. As a result, the affinities of Spongiophyton remain conjectural. Continued research is needed to clarify whether this organism possesses a complement of features that are intermediate between algae and land plants. ORESTOVIA Another Early–Middle Devonian organism included in the Spongiophytaceae by Kräusel and Venkatachala (1966) is Orestovia. This interesting, compressed organism has been isolated from paper coals of western Siberia and consists of naked, unbranched cutinized axes up to 20 cm long and 2 cm wide that taper distally (Ergolskaya, 1936; Ishchenko and Ishchenko, 1980; Krassilov, 1981a). Most specimens are preserved as hollow cuticular envelopes, often displaying an epidermis-like (?pseudo-)cellular pattern. Extending the length of each axis is a delicate strand of elongate, tracheidlike tubes with annular to reticulate thickenings on the internal wall. Many tubes contain a central core that is thought to represent resin. On the outer surface of the axes are irregular swellings thought to have had a secretory function. Small, sunken pores are randomly distributed on the axes and these are surrounded by several rings of (?pseudo-)cells; these structures have been interpreted as stomatal complexes with sunken guard cells and subsidiary cells arranged in several

rings. In a few specimens, spores occur in the cortex of the axes and range from 150 to 190 μm in diameter. Fungi have been reported on cuticles of Orestovia, but the nature of these remains equivocal. Further analysis of the structure of the cuticle and stomata prompted Gensel and Johnson (1994) to suggest that Orestovia is a tracheophyte of uncertain affinity. Snigirevskaya and Nadler (1994) reconstructed Orestovia as a semiaquatic plant resembling the extant marsileaceous fern Pilularia globulifera in overall appearance, with creeping rhizomes and naked orthotropous axes that display circinate vernation when immature. Aculeophyton is a third genus assigned to the Spongiophytaceae by Kräusel and Venkatachala (1966). It comes from the Lower Devonian of Siberia and differs from Orestovia in that the surface of the axes is covered by massive, conical papillae. Other Enigmatic Organisms

PROTOSALVINIA Although a great deal of information is known about Protosalvinia (initially named Foerstia), the biological affinities of this Devonian organism remain problematic (Niklas and Phillips, 1976; Niklas et al., 1976; Schopf, 1978b; Gray and Boucot, 1979). Like Spongiophyton, it has been interpreted as a fern, alga, bryophyte, or some form of semiaquatic organism. It is typically found compressed (FIG. 6.15), but may assume a variety of morphologic shapes, ranging from nearly circular to clavate. Some of the largest specimens approach 1 cm in diameter. At least three taxa, P. arnoldii, P. ravenna, and P. furcata (FIG. 6.16), have been suggested as representing different growth forms of a single biological species. According to this developmental chronology, the thallus becomes more clavate and develops depressions

Figure 6.14 Diagrammatic reconstruction of Spongiophyton

nanum thallus showing pores and cuticular flanges. (From T. Taylor, 1988a.)

Figure 6.15 Cleared tip of Protosalvinia thallus (Devonian).

Bar  1 mm.

CHAPTER 6

The move to the land

187

Figure 6.16 Protosalvinia furcata apical notch (Devonian).

Bar  450 μm. (From Niklas et al., 1976.)

Figure 6.18 Protosalvinia braziliensis showing position of tetrads (Devonian). Bar  475 μm. (From Niklas et al., 1976.)

Figure 6.17 Protosalvinia ravenna showing conceptacles at

apical end (Devonian). Bar  275 μm. (From Niklas et al., 1976.)

containing spores as development progresses (Phillips et al., 1972). On the surface of the thallus are cell patterns that suggest that thalli dichotomized, and specimens with an apical notch are known. At the tips of the thallus are slight depressions (FIG. 6.17), termed hypodermal conceptacles, which contain tetrads (FIGS. 6.18, 6.19) of large (200 μm), thick-walled spores (Niklas and Phillips, 1976). Each conceptacle is slightly 0.5 mm in diameter and constructed of two distinct cell layers. Spores are large and thick-walled (FIG. 6.20). In the Curiri Formation (Famennian) of the Amazon Basin, Melo

Figure 6.19 Tetrad of spores in depression in thallus of Protosalvinia ravenna (Devonian). Bar  125 μm. (From W. Taylor, and T. Taylor, 1987.)

and Loboziak (2003) correlated spores of the sporae dispersae genus Retusotriletes with Protosalvinia, where they characterized a distinct biostratigraphic unit throughout the basin. It has been debated whether the spores represent meiotic

188

PALEOBOTANY: the biology and evolution of fossil plants

Figure 6.21 Suggested reconstruction of Parka decipiens

showing lower (left) and upper surfaces. (From T. Taylor, 1988a.)

Figure 6.20 Protosalvinia spore (Devonian). Bar  200 μm.

or mitotic products (FIG. 6.20), and thus how they may have functioned in the life history of the organism. The spore wall ultrastructure suggests that they are the result of meiosis, even though they do not appear to have been constructed of sporopollenin (W. Taylor and T. Taylor, 987). The presence of tetrads of spores does not unequivocally establish Protosalvinia as a land plant, since some red and brown algae also produce spores and eggs that are morphologically similar to the spores recovered from Protosalvinia, but the similarity does not extend to the ultrastructural level. The discovery of lignin-like compounds in the fossils suggests land plant affinities for Protosalvinia (Romankiw et al., 1988), but it has also been interpreted as related to the brown algae based on biochemical evidence (Niklas, 1976b). Others consider Protosalvinia inhabited to be an alga that shows some land-adapted features, which may result from convergent evolution at a time when land plants were first becoming established. Gutschick and Sandberg (1991) suggested that Protosalvinia inhabited brackish coastal swamps. Despite all of the tools available to paleobotanists to examine the nature of the organic matter in Protosalvinia, the biomolecular signature is neither distinctly marine nor terrestrial (Mastalerz et al., 1998). Some Protosalvinia specimens have been suggested to represent parts of the cephalopod Sidetes (Hannibal, 1994). Although the biological affinities of Protosalvinia may remain unclear, it has served as a significant index fossil, that is the Protosalvinia Zone, within Upper Devonian black shale sequences of the eastern United States (Schopf and Schwietering, 1970; Murphy, 1973; Schwietering and Neal, 1978). It has also been useful for stratigraphic correlation

among Devonian rocks of the Michigan, Illinois, and Appalachian Basins (Matthews, 1983). More recent studies, however, have shown that the age of the Protosalvinia Zone locally varies from middle to late Famennian, which reduces its value as an index fossil and correlation aid. As noted earlier, a Protosalvinia Zone has also been documented in the Amazon Basin of northern Brazil (Grahn, 1992; Loboziak et al., 1997; Melo and Loboziak, 2003). PARKA Another Late Silurian–Early Devonian thalloid organism of uncertain affinity is Parka decipiens (Fleming, 1831; Reid et al., 1897; Don and Hickling, 1917; Neuber, 1979). This fossil attracted the attention of biologists because of its morphological similarity to some members in the Coleochaetales, an extant order of green algae that many believe is significant in deciphering the ancestry and origin of land plants (discussed below). Specimens of P. decipiens are 7 cm in diameter with an oval outline and slightly undulating margin (FIG. 6.21). Impressions of cells on the surface of some compressed specimens suggest that the underlying tissue was pseudoparenchymatous, and growth simulation models suggest that the thallus grew by means of both apical and anticlinal intercalary growth (Niklas, 1976c). On the surface of the thallus are disk-shaped structures which have been interpreted as some type of sporangium (FIG. 6.22); many contain small (25–45 μm) compressed bodies thought to represent spores. None possess haptotypic features such as a trilete mark. At the ultrastructural level, the wall consists of lamellae of various thicknesses (Hemsley, 1989b) similar to those in the spores of certain extant liverworts (Hemsley, 1990). Parka has also been suggested as a potential ancestor to Coleochaete in the transition to a terrestrial habitat (Niklas, 1976c). Although this hypothesis is intriguing, major obstacles remain, including the time gap of more than

CHAPTER 6

Figure 6.22 Diagrammatic reconstruction of Parka decipiens thallus with structures interpreted as sporangia. (From T. Taylor, 1988a.)

400 million years between the two taxa, together with uncertainties regarding the life history of the fossil. Moreover, antheridia have not been found in P. decipiens, and if these two taxa are similar, the presumed spores in Parka would actually represent zygotes.

Isolated fragments: clues to the transition to land? Isolated cuticle fragments, spores, and tubes are found in Lower to Middle Ordovician rocks, and extend through the Silurian into the Lower Devonian (Edwards et al., 1979; Wellman, 2001; Edwards, 1986). Although even less is known about these fragments than the enigmatic fossils described earlier, they constitute an important source of paleobiological data because they occur at a point in geologic time that coincides with the origin of a terrestrial biota (T. Taylor, 1982a, 1988a; Kenrick, 2003; Steemans and Wellman, 2004). Cuticle and Cuticle-Like Material

Isolated sheets of resistant material in the fossil record are routinely referred to as cuticle or cuticle-like sheets; early records of this type of microfossil include specimens from the Early Ordovician of Tunisia (Combaz, 1967) and early Middle Ordovician of Saudi Arabia (Le Hérissé et al., 2007). The affinities of the organisms that produced the cuticle were probably very diverse and included not only plants, but animals, fungi, and perhaps lichen-like associations. Although most cuticle sheets lack distinctive ornamentation, similar to those of Nematothallus (Edwards and Rose, 1984), a few have pores or projections. Most of these cuticle sheets have been obtained from bulk maceration of rock in acid (usually

The move to the land

189

hydrofluoric acid) and subsequent oxidation of the sample. Using such techniques, Edwards (1986) recovered numerous cuticles from Lower Devonian shales in Great Britain. On the inner surface of some specimens are inwardly directed flanges that may have marked the position of interior tissues. On the outer surface are projections of low relief. Because the biological affinities of these cuticles remain unknown, an artificial classification was created so that future workers would be able to correlate the various types of cuticles geographically and stratigraphically. These data may ultimately make it possible to trace the organisms in time and space once the questions of biological affinities are resolved. As paleobiologists learn more about the spatial and temporal distribution of cuticle fragments, and their macromolecular composition, the possibility of identifying which organisms produced the cuticle sheets also increases, and perhaps what evolutionary adaptations they represent. Spores and Spore Tetrads

In recent years, isolated spores and spore tetrads of Ordovician and Silurian age have received considerable attention as potential proxy records for early land plants (Wellman and Gray, 1988b, 2000; Edwards and Wellman, 2001; Steemans and Wellman, 2004) (FIG. 6.23). Although spore-like microfossils have been reported from as early as the Middle Cambrian (Strother, 2000; Strother and Beck, 2000), to date the oldest spore assemblages believed to have been produced by some land-inhabiting plant occur in lower Middle Ordovician (Llanvirn) rocks (Vavrdová, 1984, 1990; Gray, 1985; Strother et al., 1996; Steemans, 1999; Le Hérissé et al., 2007). Nearly identical spore assemblages occur from the Llanvirn to the late Llandovery (Early Silurian) (Richardson, 1988; Steemans et al., 1996; Rubinstein and Vaccari, 2004; Richardson and Ausich, 2007), suggesting a period of relative evolutionary stasis 40 myr in duration (Wellman and Gray, 2000; Steemans, 2001). Commonly referred to as cryptospores (Richardson, 1996), these microfossils occur as monads, dyads, and tetrads (FIGS. 6.24, 6.25), some surrounded by a smooth, loosely attached membrane (FIG. 6.26) (Strother, 2000). Strother and Beck (2000) categorized cryptospores (FIG. 6.27) as a class of organic-walled microfossils of probable terrestrial origin, whereas Steemans (2000) restricted the term to spore-like microfossils that are unambiguously attributable to the Embryophyta (Le Hérissé et al., 2007). Individual spores show no clearly defined exit site and are termed inaperturate. Other spores that first appear in Silurian rocks are solitary and possess a distinct trilete laesura (FIG. 6.28) (Steemans, 2000). In younger rocks there is a

190

PALEOBOTANY: the biology and evolution of fossil plants

40 30

Number of miospore species

20

5a

Llanvirn

5b

5c

Caradoc

5d

6a 6b 6c Ashgill

DARRIWILIAN

M. ORDOVICIAN 472

LLANDOVERY

UPPER ORDOVICIAN 460.5

Telychian

Arenig

4c

Aeronian

3a 3b 4a 4b

Rhuddanian

10

443

naked laevigate monads naked monads with pseudo-trilete mark (i.e., Imperfectotriletes) naked fused dyads (i.e., Pseudodyadospora) naked unfused dyads (i.e., Dyadospora) naked fused tetrads (i.e., Tetrahedraletes) naked unfused tetrads (i.e., Stegambiquadrella) smooth envelope enclosed dyads (i.e., Segestrespora) envelope enclosed ornamented tetrads ornamented envelope enclosed dyads (i.e., Segestrespora) smooth envelope enclosed tetrads (i.e., Velatitetras) ornamented envelope enclosed tetrads (i.e., Velatitetras) envelope enclosed monads (i.e., Sphaerasaccus) Trilete spores (Ambitisporites)

Stratigraphic ranges of numerous miospore morphologies showing evolution of biodiversity from the Ordovician into the Silurian. (From Steemans and Wellman, 2004.)

Figure 6.23

Figure 6.24 Tetrahedraletes medinensis tetrad (Silurian). Bar  15 μm. (Courtesy P. Strother.)

6.25 Quasiplanar cryptospore tetrad (Cambrian). Bar  5 μm. (Courtesy P. Strother.)

Figure

CHAPTER 6

Figure 6.26 Quasiplanar cryptospore tetrad with thin enclosing wall (Cambrian). Bar  10 μm. (Courtesy P. Strother.)

The move to the land

191

Ambitisporites dilutus (Silurian). Bar  25 μm. (Courtesy P. Strother.) Figure 6.28

Figure 6.29 Bifurcating tip of Grisellatheca salopensis

(Devonian). Bar  500 μm. (From Edwards et al., 1999.)

6.27 Dicryptosporites radiatus monad (Silurian). Bar  20 μm. (Courtesy P. Strother.)

Figure

pronounced increase in the diversity of spore types, as well as in the complexity of ornamentation (Gray et al., 1974; Wood, 1978; Beck and Strother, 2001), suggesting a major radiation of land plants during this period. Based on comparisons of Late Silurian plant microfossils from China with assemblages reported from other parts of the world, Wang et al. (2005) hypothesized that Late Silurian floras were cosmopolitan and exhibited little paleogeographic differentiation.

Although a few fragments of sporangia (FIG. 6.29) have been found containing these early spores (Wellman et al., 1998, 2003), their biological affinity and significance has continued to be challenged. Some solitary spores are believed to have been produced by algae, whereas the dyads and tetrads are thought to have represented a liverwort grade of land plant, or perhaps even vascular plants (Steemans and Wellman, 2004). The presence of land plants with a bryophytic level of organization as early as the Ordovician is supported by both paleobotanical and molecular evidence (Bateman et al., 1998; Qiu et al., 2006; Renzaglia et al., 2007). The ultrastructure (FIGS. 6.30, 6.31) of some of these spores is also expanding the database of wall organization patterns (Taylor, 2002, 2003) and, as a result, is opening new lines of evidence directed at determining the affinities of

192

PALEOBOTANY: the biology and evolution of fossil plants

Figure 6.30 Ultrathin section of scabrate spore walls (Cambrian). Bar  500 nm. (Courtesy W. Taylor.)

(FIG. 6.24), some of which possess an outer membrane. Smooth-walled, solitary, trilete spores mark the second zone, which extends from the Lower to Middle Silurian. The third zone (Middle and Upper Silurian) contains spores with various types of external ornament, suggesting increasing levels of diversity. Tetrads of spores are also found throughout this last zone, but without enclosing membranes. Despite the uncertainties regarding the affinities of various cryptospores, their presence, together with other types of fragmentary debris as early as the Middle Cambrian, indicates the presence of some type of subaerial photosynthetic cover during this time period (Strother et al., 2004). These authors also noted that at least some of the cryptospores are no doubt homologous with younger Ordoviocian and Silurian forms, but that the record to date suggests that during the Cambrian there was a diverse mesoflora of eukaryotic photoautotrophs that were derived from chlorophytes and/or charophytes, together with various thalloid organisms constructed of filaments. Tubes

Figure 6.31 Ultrathin section of cryptospore wall showing lamellae (Silurian). Bar  500 nm. (Courtesy W. Taylor.)

major plant groups. These studies will also serve to distinguish phylogenetically important features from those that are developmental in scope (Wellman, 2004). Despite the fact that most early spore types cannot be traced with certainty to major groups of plants, they are still a significant source of information with regard to biostratigraphy, paleophytogeography, and paleoecology (Steemans et al., 2007). An excellent synthesis is the study by Gray (1985) who identified three microfossil zones based on dispersed spores. The first zone extends from the Middle Ordovician to the Lower Silurian, and is dominated by spores in tetrads

Isolated tubes represent the third type of fragmentary fossil found as early as the Late Ordovician (Vavrdová, 1988). One study reports two types of tubes with annular–helical thickenings from the Mesoproterozoic (1–1.6 Ga) of northern China (Yin et al., 2004). Most tubes can be separated into two groups. In one group are narrow tubes, 8–20 μm in diameter and about 50 μm long, which are generally smooth, aseptate, and unbranched. The second group has tubes that are longer (200 μm long) with annular–helical thickenings on the inner surface. Some have a papilliform or bulb-shaped tip. Burgess and Edwards (1991) developed an artificial classification system for isolated tubes and filaments found in Ordovician–Lower Devonian rocks, and suggested that many belong to the nematophytes (Lang, 1937). The attempt to define the stratigraphic and morphological extent of these tubes may ultimately prove an important component in understanding their biology. Some tubes are like those found in certain green algae, especially members of the Dasycladales; others suggest the tubes were produced by a variety of different organisms. Although the biological affinities of the cuticle-like sheets, isolated spores, and tubes continue to remain speculative, they all suggest features that are analogous to those of the land plants found in slightly younger rocks (Edwards et al., 1998a). The presence of cuticles on aerial plant parts and spores enclosed in a wall made of sporopollenin are primary antidesiccation features found in vascular plants and a few bryophytes today. Although the Ordovician–Silurian tubes

CHAPTER 6

The move to the land

193

do not precisely fit the definition of a tracheid or any other known type of putative water-conducting cell, they nevertheless show structural and functional equivalents in organisms that were adapting to a desiccating environment, a major feature in the transition to the land (discussed below).

Land plant ancestors For many years the green algae (Chlorophyta) were the group thought to be most likely to have given rise to the land plants. Today most regard the green algae and embryophytes together as a monophyletic group, the Viridiplantae, which consists of two monophyletic lineages, the Chlorophyta and the Streptophyta. Included in the Streptophyta are all embryophytes, that is bryophytes and vascular plants, and a distinct group of green algae traditionally known as the Charophyceae, which includes the Charales, Coleochaete, and the Zygnematales, among other taxa (Simon et al., 2006) (Chapter 4). Although the fossil record of the green algae in general is extensive (Chapter 4), fossils have thus far provided relatively little information about the steps involved in the transition to a land habitat. In contrast, studies of living green algae have produced an array of molecular and morphological data that underscore the phylogenetic relationships between these two groups. Based on multiple characters, the Charophyceae are regarded as the green algal lineage most closely related to land plants (Huss and Kranz, 1997; Nishiyama, 2007). This evidence was initially biochemical and ultrastructural (Mattox and Stewart, 1984), but more recently has included molecular sequence data (Kranz et al., 1995; Karol et al., 2001; Lewis and McCourt, 2004; McCourt et al., 2004; Qiu, 2008). The question remains, however, as to which of the orders within the Charophyceae—Charales or Coleochaetales—is more closely aligned to the land plants (Chapman and Waters, 2002), as several molecular analyses provide conflicting information (Karol et al., 2001; Delwiche et al., 2002; Lewis and McCourt, 2004). Based on comparative genomic analyses, Turmel et al. (2007) suggested that the Charales are sister to a clade which consists of the Coleochaetales, Zygnematales, and land plants. They note that the question of which particular group of charophycean algae is most closely related to the land plants is a complex one for which we still do not have a clear answer. Much has been written about Coleochaete (FIG. 6.32) as a model organism in the study of land plant ancestors (Graham, 1996). The morphology of this genus has been used to suggest a way in which a filamentous thallus might have given rise to a parenchymatous land plant body

Figure 6.32 Diagrammatic view of Coleochaete with setae.

(From T. Taylor, 1988a.)

Figure 6.33 Coleochaete pulvinata showing several zygotes in various stages of cortication (Extant). Bar  120 μm. (Courtesy C. F. Delwiche.)

(Graham, 1982, 1984), and the apically biflagellate male gamete is also consistent with the sperm of land plants. The non-motile female gamete in Coleochaete and oogamous reproduction in embryophytes (Blackwell, 2003) have also been compared. Moreover, Coleochaete is the only living green alga in which some species possess zygotes (FIG. 6.33) that are retained on the maternal plant and corticated by a layer of sterile gametophytic cells (Graham, 1984;

194

PALEOBOTANY: the biology and evolution of fossil plants

McCourt et al., 2004). In at least one species, the zygote receives nourishment from placental transfer cells with wall ingrowths that increase the area in contact with the zygote (Graham and Wilcox, 1983). This pattern is similar to the archegonial venter cells of lower embryophytes. The larger number of meiotic products (8–32) in Coleochaete has been suggested as an adaptation in competing for available substrate space (Hopkins and McBride, 1976). Sporopollenin in the inner wall of the zygote of Coleochaete and lignin in the thallus, perhaps functioning in an antimicrobial manner, add additional characters that have been used to suggest a relationship between this charophycean alga and early land plants (Delwiche et al., 1989).

The transition to land The transition from an aquatic habitat to life on land was a major evolutionary event in the history of photosynthetic organisms, which involved a number of important physiological changes and structural modifications to the plant body (T. Taylor, 1982a, 1988a; Graham et al., 2000). When plants first moved onto the land, the Earth’s surface was probably already inhabited by various cyanobacteria, algae, fungi, and perhaps lichens (Taylor et al., 1997). As in arid regions today, microbial mats and biological crusts consisting of communities of microorganisms were probably important in soil formation, and there is evidence of fossil soils (paleosols) as early as the Proterozoic (Hasiotis, 2002; Prave, 2002). Molecular data obtained from the scaly green flagellate Mesostigma viride (Streptophyta) suggest that several major physiological changes, for example, in the regulation of photosynthesis and photorespiration, took place early during the evolution of the streptophytes, that is before the transition to land (Simon et al., 2006). The transition to a desiccating terrestrial habitat probably occurred sometime in the Ordovician or earliest Silurian. By the Late Silurian–Early Devonian, there is evidence of a radiation of land plants and a number of unique adaptations are found in several groups of organisms (Bateman et al., 1998). In some ways, the move of plants onto the land is similar to the evolution of the soft-bodied, late Neoproterozoic marine faunas—the Ediacaran faunas. In both cases, there were apparently a number of open niches and the plants that filled these niches exhibited a number of unusual morphologies, many of which did not survive beyond the Devonian. Although perhaps overly simplistic, it is now apparent that the ability to exist on land is the result of numerous complex interactions that involved the interplay between structural and physiological adaptations in the plants themselves,

symbiotic interactions at several levels, and physical and chemical changes in the environment. In the following sections, we will consider the major adaptations that are necessary for life on the land, including anchorage and water uptake, support for the upright plant body, water movement through the plant, desiccation prevention, some mechanism for gas exchange, reproduction in a terrestrial environment, and life history strategies. All of these features are interrelated and some structures perform multiple roles. Anchorage and Water Uptake

Long before there was any real appreciation of the diversity of fossil plants, the French botanist Octave Lignier (1908) advanced a theory about the morphological changes necessary during the move onto the land and the evolution of roots. His hypothesis used an algal ancestor with a threedimensional, dichotomously branched system that was periodically desiccated during fluctuations of available water. According to Lignier’s scenario, one segment of the branching system became covered with substrate and over time assumed the function of an anchoring and absorbing organ, much like a root. Thus, the “root” of this early land plant would be homologous with an aerial branch system, differing only in function. Although Lignier’s scenario began with an alga that we now know is not closely related to land plants, his hypothesis of a morphological model is strengthened by the occurrence of some early land plants with no organ differentiation between aerial stem and prostrate rhizome (e.g., Aglaophyton major, see Chapter 8). These permineralized axes possess the same complement of cells and tissue systems in both the aboveground aerial axes and the rhizomes. Additionally, the rhizomes produced tufts of rhizoids only on those portions in direct contact with the substrate and stomata can also be found on the rhizome, indicating that these parts of the plant may have been photosynthetic. Lignier’s model addresses a fundamental need for land plants—some way to anchor themselves to the substrate. In addition, unless a plant has a flat, thalloid plant body, the underground portions must also serve to anchor upright axes. The early land plants and the modern vascular cryptogams have no true roots, that is, with the specialized anatomy and morphology of roots (Chapter 7). Their anchoring organ is generally a rhizome, a horizontal stem which is either below ground level or on the surface of the substrate. These early plants produced rhizoids, small, usually unicellular hair-like structures, on the rhizomes, which absorbed water and minerals from the substrate. Once an early land plant was anchored in the substrate, some method of moving water and nutrients from the

CHAPTER 6

substrate to the rest of the plant was needed. The ancestral aquatic alga would have been suspended in water and water could easily enter the organism by diffusion and osmosis; the terrestrial realm, however, was a hostile and desiccating environment. One of the most important structural adaptations of the plant body was the evolution of mechanisms and structures to both obtain and conserve water. A plant growing on land has some water in the surrounding air, for example as rain, fog, or dew, but potentially more water in the substrate in which it is anchored. Thus, early land plants had to have a system to absorb water from the air and/or the substrate, as well as effective mechanisms to prevent water loss (discussed later) in order to survive periods of drought. Extant land plants overcome these obstacles in two different ways. Bryophytes are poikilohydric, that is they have no specialized mechanism to prevent desiccation, but many can tolerate desiccation and rehydrate later. Vascular plants, however, have evolved homoiohydry, that is the capacity to remain hydrated internally. This adaptation, however, except in a few rare cases (e.g., Selaginella lepidophylla, the resurrection plant), is coupled with vegetative intolerance of desiccation (Proctor and Tuba, 2002; Proctor et al., 2007). Water uptake by bryophytes occurs in the form of simple diffusion and osmosis, either internally or externally. Some extant liverworts, for example certain members of the Calobryales and Metzgeriales, contain endohydric conduits (those on the inside of the thallus), and some mosses, for example in the Bryales and Polytrichales, possess hydroids and leptoids that are functionally equivalent to the xylem and phloem of vascular plants, although the hydroids are structurally very different from tracheids (Hébant, 1977; Ligrone et al., 2000). In contrast, vascular land plants use a variety of specialized subterranean and aerial absorbing structures, which allow them to live in almost any environment on Earth. Structural Support and Water Transport

As noted earlier, a successful transition to land required structural modification for upright support of the plant. It is impossible to separate a discussion of support in early, upright land plants from water transport, since in most living vascular plants, the vascular tissues (xylem and phloem) are involved in both support and conduction (Chapter 7). In many early land plants, however, it appears that the central strand initially functioned primarily in conduction and that the plant stood erect as a result of turgor pressure in the parenchymatous cells of the axis (Speck and Vogellehner, 1988; Niklas, 1990). As land plants continued to evolve and grew larger, vascular tissue also took over the role of support, as it does in modern vascular plants. Tracheary elements in

The move to the land

195

the xylem and fibers in the phloem (Chapter 7) have secondary cell walls impregnated with lignin, a polyphenolic polymer which provides structural support and flexural stiffness to the plant organ. Many of the early land plants, however, exhibit a central strand of conducting elements, but these do not have the secondary wall thickenings that are characteristic of xylem tracheids (Kenrick and Crane, 1991). Instead, the central strand is made up of a series of tubes, some with internal or external bands that superficially resemble tracheid thickenings (Chapter 8). These bands, however, are made of primary cell wall material and do not appear to be lignified, but based on their location and structure they must have served in conducting water throughout the plant. The discovery of these interesting cell types in the central strand of plants that were once thought to contain vascular tissue necessitates a continued reexamination of all early land plant cell types, especially those involved in conduction. Protection against Desiccation and Radiation

In addition to structural support and conduction, evolving land plants also required a method to retain water in a desiccating environment. It is believed that these organisms existed for at least a portion of their life history in a terrestrial, desiccating environment where uncontrolled transpiration, and thus water loss, presented a major physiological problem. Many of the enigmatic organisms discussed earlier in this chapter possess a non-cellular outer envelope that has been referred to as a cuticle or cuticle-like layer, and may have been effective as a boundary layer against excessive transpiration. The presence of sporopollenin in the spore wall represents a similar adaptation to prevent desiccation of reproductive propagules. At the same time, the cuticle or cuticle-like layer may have been effective in the attenuation of UV radiation, including the especially dangerous UV-B (Raven, 2000). As noted earlier, this would have been a particularly important function of the cuticle in terrestrial habitats. It is interesting to note that many of the enigmatic Silurian–Devonian organisms, such as Orestovia and some early embryophytes, are characterized by a massive cuticle, which exceeds in thickness that of most plants found in geologically younger rocks. Gas Exchange

Once early terrestrial plants had developed a cuticle, they would then need some means for gas exchange (Raven, 2002), as the cuticle is only very weakly permeable to gases. Although the process of photosynthesis functions in both aquatic algae and terrestrial plants, the source of carbon dioxide for

196

PALEOBOTANY: the biology and evolution of fossil plants

each system is quite different. In algae, carbon dioxide dissolved in the water is available to chloroplast-containing cells through osmosis, whereas in most land plants carbon dioxide enters the plant through specialized openings termed stomata (Chapter 7). The regulation of these openings provides for a physiological balance within the plant (Hetherington and Woodward, 2003), a system that is regulated by other means in algae. Stomata have been identified on the naked aerial axes of many early land plants (Edwards et al., 1998a; Habgood et al., 2002); they have also been reported as occurring on the axes of the free-living gametophyte generation of several Rhynie chert plants (Kerp et al., 2004) (Chapter 8). Less specialized pores in the cuticle (FIG. 6.14) occur on other presumably terrestrial organisms, such as Spongiophyton and Orestovia, suggesting that these plants also may have carried out some level of gas exchange (Chaloner et al., 1974). Like the pores in extant liverworts, however, these unspecialized pores do not appear to include a mechanism to regulate their opening and closing, as do the stomata of vascular plants. Reproduction on Land

A primitive land plant requires several adaptations in order to reproduce on land. One of these is a method to move gametes from one gametophyte to another in order to effect fertilization, and the other is some method of spore dispersal in which the dispersal units are protected from the desiccating environment. In addition, it is important to move the sporophytic reproductive units up off the substrate, both to prevent infection by microbes and presumably to disperse spores further, in order to colonize new substrates. Ancestral algae produce motile gametes and spores in an aqueous environment. In the most primitive land plants, gamete transfer is still dependent on water, as it was in the algae, but spores of even the earliest land plants are already protected by a wall of sporopollenin, which persists through all the subsequent land plant lineages. Thus, land plant sporophytes produce spores or seeds that resist desiccation and can be transported great distances by abiotic or biotic vectors. The simplest method to raise dispersal units off the substrate is a sporangium positioned terminally on upright axes, and numerous examples of the earliest bona fide land plants show terminal sporangia on tiny unbranched or branched naked axes (Edwards and Wellman, 2001). As will be seen in later chapters, the vascular plant sporophyte in higher plants has become so dominant that the gametophyte phase is completely dependent on the sporophyte for its survival. In the more specialized vascular plants, that is the seed plants, water as a medium for fertilization is no longer necessary. Whether we are speaking

of a cycad, club moss, lily, or beech tree, each has basically the same life cycle pattern: a dominant, diploid sporophyte capable of producing spores as a result of meiosis, with each spore germinating to produce a haploid gametophyte (or gametophytes) that produce two types of gametes—egg and sperm. The life cycle is complete when the two gametes fuse to initiate the diploid phase of the new sporophyte (Graham, 1985). Life History Biology

There remain a number of significant gaps in our understanding of the transitional steps required to move from a charophycean algal ancestor to a land plant. One of these involves significant differences in the life histories of these two groups of organisms (Nishiyama, 2007). In the haplobiontic life cycle of a charophycean alga, zygotic meiosis leads to the formation of haploid zoospores, each of which develops into a mature haploid organism. In this life history the only cell that is diploid is the zygote. This is in marked contrast to the life history displayed by vascular plants, in which the multicellular sporophyte, that is the diploid organism, is the dominant phase (Graham et al., 2000), or the life cycle of bryophytes, in which a multicellular sporophyte is produced which is dependent on the gametophyte (Chapter 5). Historically there have been two different theories on the evolution of the alternation of generations in land plants (reviewed in Blackwell, 2003; Haig, 2008). HOMOLOGOUS THEORY According to the homologous or transformation theory, both the gametophyte and sporophyte phases of land plant ancestors were morphologically identical (FIG. 6.34), that is they possessed an isomorphic alternation of generations. These two phases differed only in that the gametophyte was haploid and produced gametes, whereas the sporophyte was diploid and produced spores. The presence of extant green algae with isomorphic alternation of generations, and some bryophytes with photosynthetic sporophytes, has been used to support the homologous theory. The proponents of this idea suggest that, during the course of land plant evolution, in particular the evolution of vascular plants, the sporophyte phase eventually evolved as the dominant, nutritionally independent phase. ANTITHETIC THEORY The second idea, termed the antithetic or interpolation theory, begins with the premise that the gametophyte phase was primitive and that the sporophyte phase was later added to the life cycle as a result of a delay in zygotic meiosis (FIG. 6.35). This theory, which was championed by the eminent

Sperm Egg Zygote (2n) Gametangia

Spore producing unit

Isomorphic alteration of generations

Multicellular gametophyte (n)

Heterotrichous habit Sporangium

Multicellular sporophyte (2n)

Meiosis

Meiosis Heteromorphic alternation of generations (land plants)

Spores

Spores

Figure 6.34 Hypothesized stages in the origin of the alternation of generations according to the homologous theory. Beginning with a green alga with an isomorphic alternation of generations, the sporophyte becomes structurally and physiologically more complex, whereas the gamete producing phase becomes reduced. (From Taylor and Taylor, 1993.)

Heterotrichous habit

Sperm Delay meiosis

Egg Zygote (2n) Gametangia

Multicellular gametophyte (n)

Haplontic life history

Meiosis

(New) Multicellular sporophyte (2n)

Sporangium

Meiosis

Spores Heteromorphic alternation of generations (land plants)

Spores

Figure 6.35 Hypothesized stages in the origin of the alternation of generations according to the antithetic theory. Beginning with a

haplontic green alga with zygotic meiosis, a delay in meiosis results in the interpolation of a new multicellular sporophyte that becomes heterotrichous. Meiosis occurs and the spores give rise to a multicellular gametophyte. (From Taylor and Taylor, 1993.)

198

PALEOBOTANY: the biology and evolution of fossil plants

Figure 6.36

Frederick Orpen Bower. (Courtesy A. C. Scott.)

pteridologist F. O. Bower (1908) (FIG. 6.36), views the sporophyte as gradually evolving away from a parasitic dependence on the gametophyte to become a physiologically independent, photosynthesizing organism. The fossil record provides no direct evidence as to which of these ideas is correct as it is not possible to know the sequence of events that took place after the spores of early land plants germinated. Both ideas do, however, allow for subsequent patterns of plant evolution regardless of which theory may be correct. Those supporting the antithetic theory suggest that the diploid phase was initially better adapted to life on land than the haploid phase (Keddy, 1981). As a result the diploid phase continued to evolve adaptations that contributed to increasing organ differentiation, whereas the haploid phase remained small. Proponents of the homologous theory argue that, although both phases were initially identical, environmental influences provided selective pressures, so that each phase continued to evolve different levels of specialization and reproductive strategy. Why is the diploid phase better adapted to the terrestrial environment? One answer is that water is necessary for fertilization in the most primitive land plants. Sperm cells of primitive extant plants are motile (perhaps a retention of an algal characteristic) and require the presence of water to reach the egg cell. Water may be in the form of a film of dew, a swampy environment, or splashing raindrops. Whatever the source, water is essential to fertilization. Once the new

sporophyte (diploid plant) has developed, spores are produced, released, and carried away by air currents. Even though only a small proportion of the diploid plant body is used for spore production, the absolute number of spores produced can be quite large. Spores may be distributed randomly over wide areas, with each capable of producing a haploid gametophyte. In general, the larger the number of spores produced, the greater the probability that some will settle in places suitable for germination and the eventual production of gametes. It should be pointed out, however, that in some extant plants not all spores produced are viable, and this no doubt happened in the fossil record as well. Our understanding of early land plants is based primarily on features of the sporophyte and dispersed spores. The discovery of gametophytes from the Lower Devonian Rhynie chert has provided important details about the gametophyte phase of some early land plants and these will be discussed in Chapter 8 (Remy, 1982). In these plants the gametophytes are all free-living, autotrophic organisms that differ morphologically from the sporophyte. It is not known, for example, whether any of the early land plants possessed a life history in which the gametophyte phase existed in an aquatic environment, whereas the sporophyte occupied a terrestrial habitat. Perhaps a separation of the two life history phases in the early evolution of land plants is one reason why the sporophyte has become the dominant phase in vascular plants today (but see Bennici, 2005). Animals

The earliest evidence of terrestrial animals comes from Cambrian–Ordovician trace fossils (MacNaughton et al., 2002), but the oldest body fossil to date is from the Early Silurian (Llandovery; Wilson and Anderson, 2004). Age estimates based on molecular clock assumptions, however, suggest a much earlier occurrence date (Pisani et al., 2004). As is the case for plant fossils, there appears to have been a radiation of terrestrial animals in the Late Silurian–Early Devonian when the fossil record of terrestrial animals becomes more abundant and diverse. It is interesting that the arthropods from the Siluro–Devonian (e.g., millipedes, centipedes, arachnids), including those known from the famous Gilboa, New York site (Shear et al., 1984; see Chapter 23), are all believed to have been predators (Shear and Selden, 2001). This underscores the probability that a complex terrestrial ecosystem existed prior to the Late Silurian (Jeram et al., 1990). A Fungal Partner

Another important component of the successful colonization and exploitation of the terrestrial realm by plants may

CHAPTER 6

have involved mutualistic associations with certain fungi (Pirozynski and Malloch, 1975). These symbioses are ubiquitous today (Chapter 3) and may have provided early land plants with an increased ability to obtain nutrients as a result of the extensive hyphal network of the fungus. In exchange for an increased ability to scavenge nutrients, the fungus gained access to a stable source of carbon. The fact that several Early Devonian land plants display well-established endomycorrhizae in both the sporophyte and gametophyte phases (T. Taylor et al., 1995, 2005c) adds credibility to this scenario. To further test this hypothesis we need to either find additional structurally preserved land plants with mycorrhizal fungi in their tissues, or develop techniques to detect the presence of specific mycorrhizal fungal biomarkers in compression fossils of early land plants.

The move to the land

199

Conclusion The scenario presented earlier hypothesizes that land plants evolved from a green algal ancestor(s) that became adapted to the desiccating environment on land through various physiological, structural, and functional modifications. Some of the transitional phases necessary for the successful colonization of the terrestrial realm are currently impossible to evaluate from the fossil record. Others, however, can be discussed and assessed based on the available fossil evidence. Some of these examples are discussed in Chapter 8.

This page intentionally left blank

7 INTRODUCTION TO VASCULAR PLANT MORPHOLOGY AND ANATOMY PLANT ORGANOGRAPHY .......................................................202

Epidermis ..........................................................................................208

CELL TYPES .................................................................................... 203

ANATOMY OF STEMS AND ROOTS ..................................... 210

Parenchyma .......................................................................................203

Arrangement of Primary Tissues ......................................................210

Collenchyma .....................................................................................203

Primary Xylem Maturation Patterns .................................................212

Sclerenchyma ....................................................................................203

Secondary Development ...................................................................212

Sieve Elements ..................................................................................206

Stele Types ........................................................................................216 LEAF MORPHOLOGY AND ANATOMY ............................... 221

PLANT TISSUES AND PRIMARY GROWTH..................... 207

Leaf Anatomy ....................................................................................221

Xylem Tissue.....................................................................................207

Leaf Evolution...................................................................................222

Phloem Tissue ...................................................................................207

FURTHER READING ...................................................................222

Meristems..........................................................................................208

After all, I guess it doesn’t matter whether you look down [through a microscope] or up [through a telescope] — as long as you look. John Steinbeck, Sweet Thursday It is the presence of green vegetation on the surface of the Earth that makes it a pleasant and interesting place to live. Frequently, we take this green mantle for granted, forgetting that for most of earth history the landscape was barren. Cyanobacteria, algae, and algal-like organisms must have lived in terrestrial habitats before true land plants evolved, but surely they did not have the same effect on the appearance of the Earth as true land plants do. We now know that the early land flora included both vascular and thalloid forms (Chapter 6), as does the terrestrial flora today. This chapter will be restricted, however, to a discussion of land plants with vascular tissues. In some systems of classification, vascular plants are placed in a formal division, the Tracheophyta, and molecular data support monophyly for the vascular plants (Nickrent et al., 2000). In this book, the groups of vascular plants are elevated to phylum level (see Chapter 1), in part

reflecting the fossil record of several of these groups, especially the early evolution of vascular plants (Chapter 8). One of the perplexing problems in the history of plant life has been the long interval between the appearance of green, photosynthetic organisms and the evolution of vascular land plants. There is compelling evidence that autotrophic organisms existed at least 2.5 billion years ago (see Chapter 2). There is fossil evidence that plants with conducting systems which functioned like modern vascular plants existed in the Late Silurian, a little more than 400 million years ago. There is also fossil evidence that land organisms were around earlier, perhaps during the Ordovician, based on the spore and microfossil record. As you have seen in Chapter 6, some of these, no doubt, represented relatively short-lived attempts to colonize the land surface. We now know that some of the plants that have traditionally been regarded as

201

202

PaleOBotany: the biology and evolution of fossil plants

Figure 7.2 Longitudinal section of Coleus stem showing apex and axillary buds (arrows) (Extant). Bar  600 μm.

Figure 7.1 Cross section of Aglaophyton major stem showing

conducting strand (Devonian). Bar  500 μm. (Courtesy H. Kerp.)

the earliest vascular land plants, for example, Cooksonia and Aglaophyton (FIG. 7.1), have conducting elements that are different from those of true vascular plants and which may have had a different evolutionary origin (see Chapter 8). Since vascular plants contain rigid, resistant conducting tissues and a waterproofing cuticle, they are much more frequently found fossilized than algae, fungi, and bryophytes. This chapter is a brief review of vascular plant morphology and anatomy for those who have not studied these subjects previously, including a brief outline of the principal cell types, tissue systems, and structures found in vascular plants. It is a general introduction to the subject and does not take into account the many exceptions to the definitions and principles listed in the following sections. For more detailed information on morphology or anatomy, we have listed some additional sources at the end of this chapter.

PLANT ORGANOGRAPHY The vascular plant sporophyte consists of a shoot axis stem, roots, and laterals. The primary shoot (stem and leaves) and the root arise in the embryo and are responsible for elongation

growth, that is, vertical growth, in the plant. Laterals (leaves and branches) are borne at nodes on the stem separated by internodes (FIG. 7.2). Roots have no nodes or internodes. Although all extant plants bear laterals of some sort, some of the earliest land plants (Chapter 8) were leafless, photosynthetic axes. Branches are produced either at the apex, by division of the apex, or in seed plants, they can also be produced from axillary buds. See Chapter 8 for the different types of apical branching that occur in early vascular plants. Axillary buds develop at nodes in the angle (the axil) where the leaf meets the stem (FIG. 7.2). Most vascular cryptogams (vascular plants that do not reproduce with seeds) do not have axillary branching, although there are exceptions to this rule in the fossil record. Many plants also produce adventitious organs, the most common being adventitious roots. Adventitious organs are those that develop somewhere on the plant where they do not normally arise. Since a primary root usually arises from the base of the primary embryonic axis, roots are commonly termed adventitious if they arise anywhere else on the plant, for example, from the nodes of the stem and leaf (Barlow, 1994). The basic organs of plants thus consist of stems, roots, and leaves, although the stem and leaf may be treated together as the shoot. Reproductive structures, such as cones and flowers, represent combinations and modifications of these basic organ types, for example, cones or strobili consist of a central axis (stem) with helical or whorled laterals. The laterals are most commonly sporophylls—modified leaves that bear

203

Chapter 7 INTRODUCTION TO VASCULAR PLANT MORPHOLOGY AND ANATOMY

sporangia. Floral parts in the angiosperms are also believed to represent evolutionarily modified leaves (see Chapter 22), and the flower itself consists of a shoot bearing floral parts. In some cases, the fossil record provides evidence of the evolution of these various modifications of plant organs, but in other cases, the evidence is less clear. Most plant anatomy texts concentrate on the anatomy of the flowering plants. Details that differ from angiosperms in various groups will be covered in those particular chapters, for example, ferns (Chapter 11), and seed plants (Chapters 14–22). In addition, details on seed anatomy and evolution are included in Chapter 13, so they will not be repeated here.

C

CELL TYPES PARENCHYMA

Plant organs are made up of cells. The most basic cell type, which makes up the ground tissue in plants, is the parenchyma cell (FIG. 7.3). Although all tissue types contain parenchyma, certain tissues are predominantly parenchyma, including the cortex and pith in stems and roots, and the mesophyll in leaves. Parenchyma cells are alive at maturity, have primary walls that are relatively thin, and can vary in their shape, from elaborately branched to almost isodiametric. Because they contain the full complement of cellular organelles, parenchyma cells have the potential to become meristematic and are totipotent, that is, they contain all the genetic material to develop an entire plant. They are a general-purpose cell and function in photosynthesis, so they may contain chloroplasts, and in storage of water, photosynthates (reserve foods), and many other compounds. Because of their thin walls (FIG. 7.3) and usually high water content, parenchymatous tissue is not generally an important structural component of plants, except in some of the earliest land plants (see Chapter 8). Parenchyma cells are less commonly preserved in fossils than some other cell types, especially sclerenchyma. COLLENCHYMA

Collenchyma cells are similar to parenchyma in which they are alive at maturity and can be isodiametric in shape. They have irregularly thickened primary cell walls (FIG. 7.3) and function as support in plants either during elongation growth or in plants without much secondary growth. They are most often found just beneath the epidermis in some stems and often contain chloroplasts. SCLERENCHYMA

Sclerenchyma cells are characterized by relatively thick, lignified secondary cell walls. All plant cells initially have only

Figure 7.3 Cross section of Apium sp. petiole showing paren-

chyma (arrow), epidermal, and collenchyma cells (C). Bar 100 μm.

Figure 7.4 Astrosclereid (arrow) in Castalia sp. leaf (Extant).

Bar  150 μm.

a primary wall made predominantly of cellulose. As sclerenchyma develops, a secondary wall with a high proportion of lignin is deposited inside the primary wall. The protoplast usually dies during development so that the typical sclerenchyma cell is dead at maturity. The lignified wall gives sclerenchyma cells their rigidity, and they function primarily in mechanical support and water conduction. They also make up most rigid parts of the plant (e.g., seed coats and some fruit walls) and are often positioned so that they provide mechanical protection for softer plant parts. There are three basic types of sclerenchyma cells: sclereids (FIG. 7.4), fibers, and tracheary elements, although there are intergradations

204

PaleOBotany: the biology and evolution of fossil plants

B

P

C

X Figure 7.5 Cluster of brachysclereids (arrows) in Pyrus sp.

fruit (Extant). Bar  250 μm.

among these types. Sclereids and fibers function solely in support, whereas tracheary elements function both in support and water conduction. Sclereids are variously shaped, from isodiametric to elongate and branched. They are characterized by a very thick wall with simple pits, that is, there is no special ornamentation associated with the pits. Generally, sclereids are shorter than fibers. Fibers have very thick secondary walls like sclereids (FIG. 7.5) but are elongated, spindle-shaped cells with long, narrow cell lumens; the lignified wall generally contains simple pits with slit-like apertures. They form supportive structures in tissues after the elongation growth has ceased and occur in many plant parts; many fibers, such as flax and jute, are of economic importance. Fibers commonly grow by intrusive growth, that is, the cells elongate and grow between other cells until they reach their mature length. They are common in both the primary and secondary xylem and phloem, especially in woody angiosperms, and can make up a considerable proportion of some of these tissues. Conifers have no fibers in their xylem but may have some in the secondary phloem. In some angiosperms, fibers are commonly found as a “cap” (FIG. 7.6) on the outer surface of the vascular bundle or as a band near the periphery of the stem. Fibers are classified according to their position in the plant into xylary (fibers in the xylem) and extraxylary (fibers elsewhere) fibers, for example, in the phloem or cortex. TRACHEARY ELEMENTS The term tracheary elements includes the two basic types of water-conducting cells in the xylem of vascular plants: tracheids and vessel elements. Tracheids differ from vessel elements; vessel elements have perforated end walls, whereas

Figure 7.6 Cross section of Helianthus sp. stem vascular bundle showing prominent bundle cap (B), phloem (P), cambium (C), and xylem (X) (Extant). Bar  250 μm.

tracheids have primary wall material present on their end walls; both have lignified secondary walls and both can occur in primary and secondary xylem. Both cell types also may have various types of secondary wall patterns on their side walls. Vessel elements are more efficient at water conduction, as there is no barrier to water movement from cell to cell vertically, whereas water must diffuse through primary wall at the end of each tracheid. Although tracheids are often narrower and more elongated than vessel elements, this is not always the case. TRACHEIDS. The tracheid is the basic cell in the xylem, that is, all plants have tracheids, but not the more highly evolved vessel elements. Tracheids are generally spindle shaped, very elongate, and have tapered ends. Tracheids have a dual function of support and water conduction, whereas vessel elements, except perhaps for some primitive types, function in conduction only. Both cell types are readily preserved in fossils and are easily recognized by their secondary wall thickenings. During development, the secondary wall is deposited in various patterns on top of (inside) the continuous primary wall, including rings (annular FIG. 7.14A), helical bands (FIG. 7.14B), ladderlike transverse bars (scalariform) (FIG. 7.7), or continuous except for pits. Pitted tracheids and vessels may have simple pits (no border) or pits that are surrounded by a thickened rim of wall

Chapter 7 INTRODUCTION TO VASCULAR PLANT MORPHOLOGY AND ANATOMY

205

Figure 7.7 Tracheid of Pteridium sp. showing scalariform

secondary wall thickenings (Extant). Bar  40 μm.

material—bordered pits (FIG. 7.8). The border is a dome-shaped structure, made of secondary wall, that surrounds and arches over the opening in the secondary wall, that is, the pit. There is an opening in the center of the “dome,” the aperture (FIG. 7.9). It is important to remember that the pit itself is not a “hole” in a tracheid or vessel wall, it is merely an area where there is only primary wall and middle lamella (the area between two cells) present. This area is called the pit membrane. Neighboring tracheary elements often develop bordered pits, a pit pair, at the same location on adjacent cell walls. In conifers and a few angiosperms, pit pairs can be very elaborate. In the center of the pit membrane is a thickened area, the torus, which is slightly larger than the pit aperture. Around the torus some of the primary wall and middle lamella are partially dissolved, so this area, the margo, is thinner and very porous, consisting only of strands of cellulose microfibrils. The permeable margo allows for more efficient water conduction in these pit pairs. The margo is somewhat flexible and under water stress, it can move to one side of the pit pair and seal it off. When small bubbles of air form within the water column in the xylem, they can restrict water flow (embolism), or in some cases, the water column can suddenly collapse (cavitation). Severe cavitation can cause the collapse of the tracheary element. Both drought and the freeze-thaw cycle can precipitate cavitation in the xylem.

Figure 7.8 Pinus sp. secondary xylem tracheids with circular

bordered pits (Extant). Bar  50 μm.

Figure 7.9 Circular bordered pits on secondary xylem tracheids of Sequoiadendron giganteum. Note uniseriate and biseriate pattern (Extant). Bar  100 μm.

206

PaleOBotany: the biology and evolution of fossil plants

can vary from simple perforation plates, with a single large hole, to those with scalariform or reticulate openings. The gnetophytes (Chapter 19) have a unique type—foraminate perforation plates. The most specialized vessel elements are relatively short with horizontal end walls and simple perforation plates, whereas some woods contain vessel members that are more elongate, with oblique end walls (Bailey and Tupper, 1918). Tracheids and vessels occur in both primary and secondary xylem tissue. Like tracheids, vessels also exhibit various secondary wall thickenings on their side walls which range from annular to pitted. Although vessel elements have sometimes been thought to be a synapomorphy of the flowering plants, they apparently arose independently in a number of plant groups (Bailey, 1944). The anatomical evidence for independent origin has been well illustrated in the detailed studies of Schneider and Carlquist in homosporous (Schneider and Carlquist, 1998) and heterosporous (Schneider and Carlquist, 2000c) ferns, and in certain lycopsids (Schneider and Carlquist, 2000a,b). Vessel elements are also known in several gymnospermous groups, including the gnetophytes (Carlquist, 1996; see Chapter 19) and the enigmatic fossil group, the gigantopterids (Li et al., 1996; see Chapter 19). Figure 7.10 Cross section of Cucurbita sp. vascular bun-

dle showing larger diameter vessel elements (center), many with tyloses in them (Extant). Bar  650 μm.

Tracheids and vessel elements with annular or helical thickenings are extensible, so they are most often found in the earliest matured primary xylem (protoxylem, see below), since they can stretch somewhat as the axis continues to elongate. Scalariform thickenings are variable; depending on the amount of wall material deposited, it may grade between helical and pitted. In a tracheary element with pits, most of the primary wall is covered by secondary wall, so these elements are not extensible and occur in primary xylem that has matured after the axis has ceased elongation growth; this part of the primary xylem is called metaxylem. Secondary xylem is made up predominantly of pitted tracheids, although some plant groups also have scalariform secondary xylem tracheids. VESSEL ELEMENTS Vessel elements represent a more specialized type of tracheary element (FIG. 7.10). They are usually shorter than tracheids and have perforated end walls called perforation plates. Individual vessel elements are connected end to end in vertical rows to form vessels; each vessel is a continuous tube with little or no obstruction to water flow, depending on the size and type of perforation plates in the end walls. These

Sieve Elements

Sieve elements are thin-walled cells that are alive at maturity, although the protoplast is greatly changed, and they generally lack nuclei. Sieve elements are elongated and function as the basic photosynthate-conducting cell type in the phloem of vascular plants. The walls of sieve elements contain sieve areas, circular-to-elliptical parts of the wall that are thinner. Each sieve area (FIG. 7.11) includes a number of sieve pores, which allow for transport from one sieve element to the next (FIG. 7.12). Sieve pores are not actual holes in the wall as perforation plates are. Rather, they are protoplasmic connections between two living cells and are lined with a plasma membrane. In addition to the basic sieve element, there are two more specialized types of sieve elements: sieve cells, which occur in conifers, and sieve tube elements, which are a synapomorphy for the angiosperms. Sieve cells are generally long, narrow, and tapered at the ends, whereas sieve tube members are shorter and wider with more horizontal end walls. Sieve tube elements, like vessel elements, are connected end to end in vertical rows to form sieve tubes. Sieve plates occur on the end walls of sieve tube elements; these are groups of sieve areas, usually with larger pores than those on the lateral walls of the cell. Since sieve elements are under great hydrostatic pressure while functioning, they often collapse after death. Thus, preservation of phloem tissue in fossils is relatively rare.

Chapter 7 INTRODUCTION TO VASCULAR PLANT MORPHOLOGY AND ANATOMY

207

P

Figure 7.13 Transverse section of Lepidodendron stem showing extensive periderm (P) (Pennsylvanian). Bar  1 mm.

Figure 7.11 Sieve areas (arrows) on phloem cell of fos-

sil Cycadeoidea sp. stem (Cretaceous). Bar  10 μm. (Courtesy P. Ryberg.)

contain one cell type and complex tissues, more than one type. Under the system of Sachs (1875), there are three tissue systems: the dermal, vascular, and ground (or fundamental) tissues. Dermal tissue is the outer covering of the plant, consisting of epidermal cells and cuticle in young plants, and periderm (FIG. 7.13), in plants that produce extensive secondary growth. Vascular tissue is the conducting tissue in the plant and consists of xylem (water conduction) and phloem (food conduction) tissue. Ground tissue is basically everything else in the plant. It can include simple tissues like parenchyma or complex tissues that include parenchyma and other cell types, for example, sclerenchyma. XYLEM TISSUE

Figure 7.12 Sieve areas (arrows) on phloem cells of Sequoia sp. (Extant). Bar  100 μm.

PLANT TISSUES AND PRIMARY GROWTH A plant tissue is a group of cells having a similar origin, structure, or major function. Plant tissues contain a characteristic complement of one or more types of cells; simple tissues

Of all the components in a typical plant axis, xylem elements are found most frequently in the fossil record. Xylem is a complex tissue made up of several types of cells. As noted earlier, all xylem tissue contains tracheids and some plants have both tracheids and vessels in their xylem. All xylem also contains living cells, xylem parenchyma, and sometimes fibers (xylary fibers). PHLOEM TISSUE

In addition to conducting cells, phloem tissue also contains unspecialized parenchyma, the phloem parenchyma, and often strengthening cells such as phloem fibers. When sieve tube members are present, they are always accompanied by companion cells (FIG. 7.15), specialized parenchyma cells that develop from the same procambial initial (mother cell) as the associated sieve tube member.

208

PaleOBotany: the biology and evolution of fossil plants

sieve tube member. The corresponding cell in gymnosperm phloem is called an albuminous cell, but it does not develop from the same initial cell as does the sieve cell. MERISTEMS

Sieve elements (S) and companion cells (C) in Cucurbita sp. Note conspicuous sieve plate (Extant). Bar  50 μm.

Although all parenchyma cells in a plant have the potential for growth and production of new cells, unless a plant has been wounded, cell production normally occurs in meristematic tissue, which consists of parenchyma cells that remain capable of dividing and producing daughter cells throughout the life of the plant. Because of these meristems, vegetative growth in plants is indeterminate, which means that the plant body is not fixed in its development but is potentially capable of continuous growth. There are two basic types of meristems, apical meristems and lateral ones, and one more specialized type, an intercalary meristem. Apical meristems are responsible for growth in length (height), or primary growth, and lateral meristems or cambia (sing. cambium) for growth in diameter (width) (see section “Secondary Development”). As the name implies, apical meristems occur at the apex of every stem and twig and the apex of every root in a plant. The cells that make up the apical meristem undergo repeated divisions to produce daughter cells; a short distance back from the growing tip, the daughter cells begin to differentiate into xylem, phloem, sclerenchyma, etc. In stems, the apical meristem produces cells and tissues that form the stem, leaves, and axillary buds, or lateral branches. In roots, the apical meristem produces only the cells that make up the root; laterals are produced further back from the meristem itself. Vascular tissue first appears as procambial strands, which will develop into mature xylem and phloem. The first xylem to mature is termed protoxylem and, as noted earlier, usually consists of extensible tracheary elements with helical or annular secondary wall thickenings (FIG. 7.14). Although protoxylem cells are usually smaller in diameter than metaxylem cells, the nature of the wall thickenings is a more exacting way to distinguish these cells in fossil plants. Further back from the meristem, after the stem has for the most part ceased to elongate, the metaxylem matures. It generally consists of non-extensible tracheary elements, such as pitted or scalariform (FIG. 7.7) tracheids. Both proto- and metaxylem are primary xylem, that is, they are produced by the apical meristem. Similar terms are used for phloem cells, that is, protophloem and metaphloem.

At maturity companion cells exhibit dense cytoplasm and have numerous protoplasmic connections with their associated sieve tube member. Physiologically, companion cells are thought to control movement of solutes into and out of the

The epidermis is the layer of cells on the outside of all primary parts of plants (FIGS. 7.3, 7.16); it functions in protection against water loss and infection, gas exchange with the

Figure 7.14 Longitudinal section of Zea mays xylem showing

annular (A) and helical thickenings (B) (Extant). Bar  60 μm.

S

S C e S

C

C

Figure 7.15

EPIDERMIS

Chapter 7 INTRODUCTION TO VASCULAR PLANT MORPHOLOGY AND ANATOMY

209

S

C

Figure 7.16 Section of Podocarpus urbanii stem showing thick

cuticle (arrows), hypodermal sclereids (S), and stomatal chamber (C) (Extant). Bar  50 μm.

Figure 7.18 Abaxial epidermal cells of Kalanchoe sp. with stomata (Extant). Bar  350 μm.

Figure 7.17 Section of Yucca sp. leaf showing thick cuticle

(arrow) (Extant). Bar  50 μm.

atmosphere, and sometimes photosynthesis. If a plant does not exhibit secondary growth, the epidermis may be retained throughout the life of the plant. In plants with secondary growth, the periderm will take over these functions as development continues (see section “Secondary Growth”). The epidermis is most often one cell layer thick, although there are exceptions. Epidermal cells are primarily parenchymatous, as they are alive at maturity, and most are compactly arranged with very few or no intercellular spaces. CUTICLE Covering the epidermis is the cuticle, a non-cellular layer which covers all the aerial parts of plants (FIG. 7.16); epicuticular waxes occur on the outer surface of the cuticle. The cuticle impregnates the outer cell wall of the epidermal cells, may form flanges between epidermal cells, and is also deposited on the surface of the cell (FIG. 7.17). It consists of varying proportions of two lipid polymers, cutin and cutan. Cutan is resistant to decay and was thought to account for the widespread preservation of cuticle in the fossil record (Chapter 1), but Gupta et al. (2006) have recently questioned this assumption after finding little cutan in some fossil leaves.

P

Figure 7.19 Cross section of Cycas revoluta leaf showing papilla (P) and sunken guard cells (arrows) (Extant). Bar  50 μm.

They suggested that a combination of cutin, waxes, and lipids from the interior of the plant may combine during diagenesis to form the so-called cuticle found on many fossils. STOMATA The epidermis also contains stomata (sing. stoma), the openings in aerial plant parts that allow for exchange of O2, CO2, and water vapor during photosynthesis. The stomatal complex (FIG. 7.18) consists of the stoma itself (FIG. 7.19), which is the pore or opening in the epidermis, two guard cells, and subsidiary cells, if present. Subsidiary cells are epidermal cells that surround the guard cells and differ in function and often also in morphology from other epidermal cells. In the ontogeny of the stomatal complex, the guard cells and subsidiary cells may develop from the same initial, a pattern called mesogenous, or syndetocheilic. If the guard cells and subsidiary cells develop from two different initials, this pattern is called perigenous or haplocheilic. These two types,

210

PaleOBotany: the biology and evolution of fossil plants

glandular and secrete substances that are important in deterring herbivory. They are known throughout the fossil record and range from multicellular glandular types (Krings et al., 2003a) to simple one-celled forms. Trichome bases may be preserved on fossil leaf cuticles, even when the trichome itself is no longer present, and have been useful in paleoecological studies. The type and arrangement of trichomes can also provide systematic information.

ANATOMY OF STEMS AND ROOTS Figure 7.20 Glandular trichome on Coleus sp. epidermis

(Extant). Bar  50 μm.

along with other epidermal characteristics, have been used to classify fossil foliage in the cycadophytes (Chapter 16). The basic types of stomatal arrangements include: (1) anomocytic (irregular), in which subsidiary cells are lacking, that is, the cells around the guard cells are not distinguishable from other epidermal cells; (2) paracytic (parallel), in which elongated subsidiary cells flank the guard cells; (3) anisocytic (unequal), which has three subsidiary cells, with one substantially smaller than the other two; (4) actinocytic (ringlike), with a circle of subsidiary cells around the guard cells; and (5) diacytic (cross-celled), in which there are two subsidiary cells, and the wall that these two cells share is at right angles to the guard cells. There are many variations on these types and gradations between them; different types can even occur side by side on the same organ (Kothari and Shah, 1975). For further information on these and additional types, see Van Cotthem (1970), Leaf Architecture Working Group (1999), Beck (2005), Carpenter (2005), and Evert (2006). Stomata may be on level with the surface of the leaf or may be sunken, which is common in plants that live in arid environments. Guard cells are usually bean shaped and covered with cuticle. They allow for the controlled opening and closing of the stomatal pore through changes in turgor pressure within the cells. Guard cells are known to respond to various exogenous environmental stimuli, including light (especially UV-B radiation) and CO2 concentration. Certain plant hormones, especially abscisic acid, mediate the internal reactions and ionic changes that control stomatal opening and closing through guard cell water potentials. TRICHOMES Certain epidermal cells produce outgrowths (FIG. 7.20) which develop into trichomes (hairs). Trichomes are common, especially on leaves, and have a variety of functions, including additional protection against water loss. Some of them are

ARRANGEMENT OF PRIMARY TISSUES

The arrangement of tissues in stems and roots is similar, although there are differences between monocots and dicots, as well as between seed plants and non-seed plants. These will be discussed later under section “Stele Types.” In gymnosperm and dicot stems, the center of the axis is occupied by parenchymatous pith tissue, which functions as a storage tissue and sometimes a water source. Surrounding the pith is a ring of vascular bundles, each containing xylem and phloem. The most common type of bundle is collateral, with xylem toward the pith and phloem immediately outside the xylem. Outside of the vascular tissue is the cortex, with the epidermis external to it. Both the pith and cortex are primarily parenchymatous, although sclereids, fibers, secretory cells or canals, and other specialized cells may occur in either tissue. In the stems of monocotyledons, the vascular bundles are not arranged in a ring but are scattered through the stem. In this case, it is impossible to distinguish pith and cortex, so the tissue in which the bundles are embedded is simply called ground tissue. The vascular tissue in stems and roots, together with associated parenchyma, including the pith if present, is termed the stele. In the roots of most plants, there is no pith, and the xylem occupies the center of the axis; monocot roots have a central pith with vascular tissue in a ring. The xylem is ribbed and the phloem occupies the area between the ribs. In cross section, these ribs resemble two, three, or four arms extending from the center of the root (FIG. 7.21). A fundamental difference between stem and root anatomy is the position of the phloem with respect to the xylem. In stems, the phloem is immediately outside the xylem, but in roots, its position alternates with that of the xylem, so that phloem and xylem are situated on different radii in a root, but the same radius in a stem. This concept is important in classifying fossil axes as roots or stems when there are no laterals attached. In roots and in stems of certain primitive plants, the peri-cycle is the tissue immediately outside the vascular tissue; it may be one or more layers thick. The pericycle is

Chapter 7 INTRODUCTION TO VASCULAR PLANT MORPHOLOGY AND ANATOMY

211

Figure 7.21 Cross section of Ranunculus sp. root show-

ing tetrarch xylem strand with phloem between arms (arrows). Note starch grains (stained pink) in cortical cells (Extant). Bar  200 μm.

P

Figure 7.22 Longitudinal section of Lupinus sp. root showing

origin of lateral root from pericycle (P) (Extant). Bar  500 μm.

part of the stele, as it arises from the same group of cells in the apex that produce vascular tissue. The pericycle is an important tissue in roots, as this is the site where branch or secondary roots arise (FIG. 7.22). Cells in the pericycle become meristematic and differentiate into a lateral root, which, as it grows, pushes its way out through the cortex (FIG. 7.23) and epidermis to emerge on the outside of the

Figure 7.23 Cross section of branch root arising from pericycle of Salix sp. root (Extant). Bar  250 μm.

root. As noted earlier, the laterals produced by stems, that is, leaves and axillary buds, are produced by the apical meristem early in primary growth, but this is not the case in roots. Because leaves arise from the outer tissues of the stem, they are said to have an exogenous origin. Since lateral roots arise deep in the root tissue in the pericycle, they have an endogenous origin. In addition to stelar differences, the origin of laterals is another fundamental difference between roots and stems and is often an important anatomical character used to identify isolated axes found in the fossil record. Immediately outside the pericycle is the endodermis (FIG. 7.24), a single layer of cells lacking intercellular spaces that can be recognized by characteristic thickenings, Casparian strips, in its transverse and radial walls. The endodermis represents the innermost layer of the cortex and functions as a physiological barrier that regulates movement of solutes into and out of the vascular tissue in the stele. Casparian strips surround the sides, top, and bottom of each cell, but do not cover the inner and outer tangential walls; they are rich in suberin, making them hydrophobic. These strips effectively seal off the tiny canals and spaces within the walls, or apoplast, of the endodermis, and thus force diffusion of solutes through the plasma membrane and cell lumen of the endodermal cells, that is, materials must move through the symplast or the living parts of the cells, thus providing some physiological control over lateral movement into and out of the stele. In a young stem or root, there is a parenchymatous cortex surrounding the stele; in some stems this tissue may contain

212

PaleOBotany: the biology and evolution of fossil plants

Figure 7.24 Transverse section of Pinus leaf showing endodermis and thickened casparian strips (arrows) (Extant). Bar 75  m.

is situated toward the outside of the stem, with metaxylem toward the center, the maturation, or development, of the xylem is described as being centripetal (from the outside in) or exarch. The exarch condition is found in almost all roots and in many primitive plants, such as the lycopsids. In axes with a pith, when the protoxylem is next to the pith, that is, closest to the center of the axis, and the metaxylem develops outside of it, the maturation pattern is endarch (centrifugal development of xylem). This is the typical maturation pattern found in the primary xylem of most seed plants. In axes with a solid core of xylem, called a protostele (see section “Stele Types”), if the protoxylem occupies the center of the protostele, this type of maturation is called centrarch. Centrarch xylem maturation is relatively common in several groups of Devonian plants, for example, the Rhyniophyta and Trimerophytophyta (Chapter 8). In axes that have a pith, if the protoxylem occupies the center of the xylem and the metaxylem develops on both the outside and the inside of the protoxylem, this pattern is called mesarch (development is both centrifugal and centripetal). Mesarchy occurs in many ferns. These primary xylem configurations can be seen in anatomically preserved fossils and have been used when attempting to determine relationships among groups. While it is not possible to observe the actual sequence of development in fossils, it is often possible to infer the maturation pattern based on the relative cell size and secondary wall patterns of protoxylem and metaxylem elements. SECONDARY DEVELOPMENT

Figure 7.25 Cross section of Elodea sp. stem showing aeren-

chymatous cortex (Extant). Bar  650 μm.

large intercellular spaces (FIG. 7.25). The cortex is principally a storage tissue, although some of the outer cells in young stems may be photosynthetic. The periphery of the cortex may contain collenchyma in young axes, but sclerenchyma is more common in older stems and roots; epidermis covers the outside of the plant. If a plant does not have secondary growth, then maturation of the primary tissues completes growth. See section “Secondary Development.” PRIMARY XYLEM MATURATION PATTERNS

As mentioned earlier, protoxylem consists of those tracheary elements (usually smaller in diameter) that are the first to mature. Xylem maturation patterns describe the location of the protoxylem in relation to the metaxylem; these terms are used only for primary xylem. When the protoxylem

Secondary growth is responsible for increase in diameter in roots and stems; secondary tissues are produced by lateral meristems or cambia. There are two basic types of cambia, a vascular cambium, which produces vascular tissue, that is, secondary xylem and phloem, and a cork cambium or phellogen, which produces tissues to replace the epidermis (see the following sections). VASCULAR CAMBIUM The vascular cambium arises between the primary xylem and phloem of a young stem or root. Parenchymatous cells become meristematic and begin to produce secondary xylem or wood toward the inside of the cambium and secondary phloem toward the outside of the cambium. The cambium itself remains meristematic, except in some unusual cases, for example, in the Carboniferous arborescent lycopsids (Chapter 9) and may range from a single layer to several layers of meristematic cells (FIG. 7.26). If the primary xylem is a solid core, as in some fossils,

Chapter 7 INTRODUCTION TO VASCULAR PLANT MORPHOLOGY AND ANATOMY

C

Figure 7.26 Cross section of Pinus sp. stem showing radial files of vascular cambium initials (C) (Extant). Bar  100 μm.

the cambium begins development as a complete cylinder (a ring, as seen in cross section) between the primary xylem and phloem. If the primary vascular tissue occurs in bundles, as is the case in woody dicots and gymnosperms, the cambium begins development within the bundle—the fascicular cambium. Then, parenchyma cells between the bundles become meristematic—the interfascicular cambium—and connect the fascicular cambia together so that the cambium eventually forms a complete ring around the axis, between the primary xylem and phloem. Cambial cells or initials divide primarily by periclinal divisions (parallel to the surface of the axis) on their inner and outer faces, producing files of cells along the radii of the axis. The presence of these orderly files is one way to distinguish secondary growth in fossil axes. Cambial initials must also divide anticlinally (perpendicular to the surface) to produce more cambial cells as the circumference of the axis continues to increase due to the production of secondary tissue. There are two types of initial cells in the vascular cambium. Fusiform initials are elongate cells that produce the conducting cells in both the secondary xylem and secondary phloem and the other cells in the axial system. Ray initials are shorter, generally rectangular cells, which give rise to cells in the ray system (see section “Secondary Xylem”). Generally, many more secondary xylem cells are produced than secondary phloem; indeed, in most living trees the bulk of the trunk represents secondary xylem or wood. The vascular cambium in roots arises in the same place as in stems, that is, between the primary xylem and phloem, but since the primary xylem in many roots is lobed or furrowed, the cambium initially also has this shape. As the root continues to develop, however, more secondary xylem is produced in the furrows so that the cambium eventually has a cylindrical shape, just as it does in stems. See section “Secondary Xylem” and “Phloem” (later) for the cell types produced by the vascular cambium.

213

CORK CAMBIUM (PHELLOGEN) As the vascular cambium continues to produce cells, the stem or root increases in diameter and the peripheral portion of the cortex and epidermis, which are not meristematic, would eventually be split apart. In older axes, therefore, periderm tissue performs the function of the primary epidermis, that is, to protect the plant from infection and desiccation. The periderm includes the phellogen or cork cambium, cork cells (phellem), and sometimes phelloderm. Like the vascular cambium, the cork cambium produces cells to the inside (phelloderm) and the outside (cork). Also like the vascular cambium, the production of cells is not equal on the two faces, but, in this case, more cells are usually produced on the outside (cork) than on the inside, with the exception of some members of the Lepidodendrales (Chapter 9), which produce more phelloderm than cork. Phelloderm cells are parenchymatous, but cork cells are non-living at maturity and their walls are impregnated with suberin; they thus prevent water loss and also provide a barrier to infection by fungi and bacteria. The cork cambium can arise close to the outside of the stem, that is, subepidermally, or deeper within the cortex or in the secondary phloem. It can even arise in the epidermis itself. The process of development is the same as for the vascular cambium which parenchyma cells become meristematic and produce files of cells by periclinal divisions of the cork cambial initial cells. The cork cambium also undergoes anticlinal divisions to increase in circumference. The cork cambium may initially arise in certain areas of the axis but eventually becomes continuous around the stem or root. As the stem continues to increase in diameter, the older (i.e., outermost) phellem ruptures and may be sloughed off the outside of the stem. Newer cork cambia then differentiate inward of the original cork cambium, initially within the primary cortex but later within the secondary phloem. It is the arrangement of these subsequent cork cambia and the amount of cork they produce that gives the outer bark, or rhytidome, of particular species its characteristic appearance. Smooth bark (e.g., in some species of Betula) forms where there is little cork produced, whereas rough, fissured bark (e.g., in Quercus) results from extensive cork production. The fossil aquatic angiosperm Decodon allenbyensis, from the Eocene of British Columbia, has a very complex rhytidome, and the same structure does not occur in living species of this genus (Little and Stockey, 2006). Roots of D. allenbyensis produce a lacunate phellem, with alternating elongate and isodiametric cells. Bark is a non-technical term that includes all the tissues outside the vascular cambium. If the axis is young, the bark may include, from the cambium outward, secondary phloem,

214

PaleOBotany: the biology and evolution of fossil plants

Figure 7.27 Cross section of Pinus sp. stem showing 3 years of

growth and resin canals in wood (arrows) (Extant). Bar  250 μm.

primary phloem, primary cortex, phelloderm (if present), cork cambium, and phellem (cork). In older trees, the bark may consist only of secondary phloem, cork cambium, and phellem. In roots, the cork cambium may also arise near the surface of the axis but most commonly arises in the pericycle. SECONDARY XYLEM Secondary xylem (wood) is a much more complex tissue than primary xylem and consists of a number of different cell types arranged in specific ways. Wood includes an axial system, which moves water and minerals up the stem, and a ray system, which runs horizontally through the stem, that is, in a radial direction. The axial system contains the vascular tissue, tracheary elements (tracheids and/or vessels), and axial parenchyma (vertical strands of parenchyma). In certain angiosperms (hardwoods), the axial system may also contain support cells such as fiber-tracheids or libriform fibers, a type of xylary fiber. Gymnosperms do not have fibers in their wood (although fiber-tracheids may be present) and are called softwoods by foresters. Some conifer wood contains resin ducts (FIG. 7.27) or canals in both the axial and ray system, that is, they are oriented both vertically and horizontally. Resin ducts form by the separation of parenchyma cells during development; at maturity, they are hollow tubes which are lined with an epithelial layer, whose cells produce the resin. Resin ducts also form in many conifers as a response to wounding or infection by various pathogens. The ray system extends at right angles to the tracheary elements and is involved in conducting water and nutrients in a radial direction in the mature axis, as well as storage

Figure 7.28 Tangential (A) and radial (B) section of Pinus strobus wood showing vascular rays (arrows) (Extant). Bar  360 μm.

in the older secondary xylem. It consists of vascular rays, which are principally composed of parenchyma cells (homocellular rays). Some conifers have ray tracheids in their rays (heterocellular rays); these are shaped like parenchyma cells but have pitted walls and are non-living at maturity. Vascular rays in conifers are usually uniseriate or biseriate, that is, from one to two cells wide (FIG. 7.28), and can range from one to usually 20 cells high (Evert, 2006). In gymnosperms (e.g., Ephedra) and angiosperms rays range from uni- to multiseriate in width and from one to many cells high, up to several centimeters (FIG. 7.29) (Evert, 2006). Both ray and axial parenchyma cells in the wood may form tyloses where they border a tracheary element. The wall of the parenchyma cell extends through the pit cavity and balloons out into the lumen of the neighboring vessel or tracheid. Tyloses (FIGS. 7.10, 12.40) most often appear in xylem that is no longer functional and are thought to function as a means of sealing off tracheary elements, or perhaps as a host response to infection. Because of the complexity of secondary xylem, three different planes of sections are needed to fully characterize the anatomy of the wood and, in many cases, to classify isolated pieces of wood (e.g., fossils) to genus. A cross, or transverse, section is made at right angles to the axis of a stem or root; this is the section exposed when a tree is cut down. If the tree exhibits growth rings (tree rings), they will be visible in a cross section. A radial section is a longitudinal section which is cut along the radius of the axis. A tangential section is also a longitudinal section but is cut perpendicular to a radial section on a tangent to the surface of the stem.

Chapter 7 INTRODUCTION TO VASCULAR PLANT MORPHOLOGY AND ANATOMY

215

E

L

E

Figure 7.29 Tangential section of Ephedra trifurca wood showing multiseriate rays (Extant). Bar  650 μm.

A cross section shows the tracheids, fibers, vessels and other cells of the axial system in cross section, and the cells appear as squares, rectangles, or polygons. The vascular ray cells in this section are in longitudinal section, since they are elongated along the radius of the axis. In a cross section, you can measure the width of tracheary elements, but not their length; conversely, it is possible to measure the length of ray cells, but one cannot determine their height. In a radial section, tracheids and fibers appear in longitudinal section as very elongate cells with tapered end walls. Secondary wall patterns can best be seen in a radial section, as pits are more common on the radial faces of the cells. Vessels, if present, are revealed as elongated series of cells, one above the other, and perforation plates in their end walls appear in side view. Vascular rays also appear in side view and the cells that make up the ray look like the face of a brick wall that is many bricks (or cells) long and many bricks (cells) high. From this view, we get no idea of the number of cells that make up the thickness of the ray (or, to continue the analogy, the thickness of the brick wall). It is also very difficult to determine the height of rays in a radial section, as the cut would need to be exactly through the middle of a ray to reveal the full height. In a tangential section, the axial system, that is, tracheids, fibers, vessels, and xylem parenchyma, all look more or less as they do in a radial section, except that bordered pits are not always seen. Vascular rays can be seen in cross section

Cross section of Pinus sp. early wood (E) and late wood (L) transition (Extant). Bar  150 μm.

Figure 7.30

in this view or, in other words, at right angles to the ray’s length. It is now possible to see the height and width (thickness) of the ray and the ray cells but not the length of the ray. To continue the brick wall analogy, in a tangential section you see the ray head on, like looking at the end of a brick wall, but you cannot determine its length. Growth rings or tree rings occur in many woody plants (FIG. 7.30). Those that grow in temperate zones usually, but not always, produce a single ring each year and this can be counted to determine the age of the tree. Many tropical trees also produce rings, however, and they often correspond to wet and dry seasons. In some areas of the temperate zone, trees can produce multiple rings per year due to seasonal precipitation. Tree rings are made up of earlywood and latewood, sometimes called spring wood and summer wood. In the spring, the apical meristems and young growing leaves of the plant produce the plant hormone, auxin, which is integral to the functioning of the vascular cambium. In the spring when the stems and roots are still elongating, auxin levels are higher, and the cambium produces earlywood—larger diameter tracheids with relatively thin walls. As elongation slows down and eventually stops, less auxin moves down the axis (or up in a root), and cambial production switches to producing latewood, which consists of smaller diameter

216

PaleOBotany: the biology and evolution of fossil plants

tracheids with thicker walls. When the tree goes dormant in the fall, either due to seasonal deciduousness (leaf drop) or to cold temperatures if it is evergreen, the cambial cells cease to divide. The following spring, the cambium begins producing earlywood again. The contrast between the latewood of the previous growing season and the earlywood of the present season is the ring boundary, which can be very sharp and visible to the naked eye. SECONDARY PHLOEM Secondary phloem, the tissue produced to the outside of the vascular cambium, is also a complex tissue that includes an axial and a ray system. Like the xylem, the axial system in secondary phloem includes conducting cells, either sieve cells in conifers or sieve tube members in the angiosperms, which conduct solutes from the sites of photosynthesis to other parts of the plant. Phloem parenchyma occurs in the axial system, as well as companion cells (angiosperms) and albuminous cells (conifers). Fibers are very common in the secondary phloem of both conifers and angiosperms (FIG. 7.31), and the pattern of fiber production by the cambium can sometimes be used to identify secondary phloem and bark tissue taxonomically. Although some conifers can produce regular, repeating bands of sieve cells, fibers, and parenchyma, they do not seem to produce these on an annual cycle, so it is not possible to determine the age of bark as it is to date wood by counting the tree rings. Usually only a narrow band of phloem close to the cambium is actively involved in conduction—the functional phloem or inner bark. As the older phloem becomes nonfunctional, there are many histological changes in the tissue, including the collapse of sieve elements, the development of sclereids from parenchyma cells, and/or the deposition of ergastic substances in parenchyma cells. These changes have also been identified in fossil phloem (Smoot, 1984c). It is in the nonfunctional phloem that subsequent cork cambia may arise in older axes. Vascular rays in the secondary phloem are continuous from the secondary xylem into the secondary phloem and consist only of parenchymatous ray cells. In some plants, the secondary phloem increases tangentially as the stem increases in diameter. This increase can occur by a tangential elongation of either axial or ray parenchyma cells. Some parenchyma cells, especially ray cells, may become meristematic and divide radially to produce additional cells. This process is called dilatation growth and can substantially increase the width of phloem rays. Secondary phloem rays are also important in ethylene signaling during plant responses to wounding and pathogens (Hudgins and Franceschi, 2004).

V P

X

Figure 7.31 Cross section of Tilia sp. stem showing secondary xylem (X), phloem (P), and dilating vascular rays (V) (Extant). Bar  650 μm.

STELE TYPES

As noted earlier, the stele is defined as all tissues inside of, but not including, a distinct physiological barrier or boundary layer such as the endodermis (including the conducting tissue), after a concept called the stelar theory, which was initially developed by Van Tieghem and Douliot (1886a, b). The stelar theory is not used today by most botanists working with living plants as it is sometimes difficult to recognize the outer boundary of the stele as originally defined, and because stelar configuration can vary at different developmental stages of the plant or at different levels within a single axis. Nevertheless, the concept has been useful in comparative and phylogenetic studies of fossil vascular plants. In a general sense, stele types become progressively more complex in the fossil record and certain plant groups are characterized by particular types, so a knowledge of stele types is useful in paleobotany. For a treatment of stelar terminology and classification, see Schmid (1982) or Brebner (1902). PRIMITIVE VASCULAR PLANTS (VASCULAR CRYPTOGAMS) The simplest type of stele is a protostele, which consists of a solid core of xylem (no pith) in the center of the axis. Stems of many primitive plants and most roots are protostelic. There are three basic types of protostele: haplostele (FIG. 7.32), actinostele, and plectostele (FIG. 7.33). In a haplostele, the xylem is circular in cross section or cylindrical in three dimensions; phloem is immediately outside the xylem. An actinostele exhibits armlike projections of the xylem in cross section or ridges in three dimensions, with the phloem in

Chapter 7 INTRODUCTION TO VASCULAR PLANT MORPHOLOGY AND ANATOMY

Figure 7.32 Cross section of Gleichenia sp. rhizome showing

haplostele (Extant). Bar  650 μm.

Figure 7.33 Cross section of Lycopodium serratum stem

showing actinostele with exarch xylem development (Extant). Bar  200 μm.

the furrows. Many roots are simple actinosteles, as they are usually diarch (two arms), triarch, or tetrarch. A plectostele exhibits many lobes in cross section, and it may appear as if the xylem is in separate plates (FIG. 7.34); phloem occurs between the plates. A three-dimensional view of a plectostele, however, indicates that the plates are interconnected. Both plectosteles and actinosteles occur in the extant Lycopodiales. A siphonostele occurs in vascular cryptogams that have a pith, with the xylem and phloem forming a continuous cylinder around the pith (FIG. 7.35). A solenostele, or amphiphloic siphonostele (FIG. 7.36), has phloem on both the outside and inside of the xylem (FIG. 7.37); an

217

Cross section of Lycopodium sp. stem showing plectostele (Extant). Bar  300 μm.

Figure 7.34

Figure 7.35 Cross section of Helmenthostachys siphonostele showing leaf trace (arrow) (Extant). Bar  780 μm.

ectophloic siphonostele has phloem only on the outside of the xylem. Both ectophloic and amphiphloic siphonosteles are found in the ferns. When leaves are produced, a leaf trace is given off from the stem stele and it supplies the leaf with xylem and phloem. In a series of cross sections through a fern stem in the region of leaf attachment, parenchyma cells appear in the siphono-stele at the point of leaf trace emission and there is continuity between the pith and cortex at that level. At higher levels, the leaf trace extends upward and outward into the base of the leaf. This interruption of the stelar cylinder is called a leaf gap, even though it is not actually a space as the name implies, but rather parenchymatous tissue. At higher

218

PaleOBotany: the biology and evolution of fossil plants

Figure 7.38 Cross section of Pteridium aquilinum rhizome showing dictyostele (Extant). Bar  2 mm.

Figure 7.36 Cross section of Marsilea quadrifolia rhizome showing amphiphloic siphonostele (Extant). Bar  300 μm.

Figure 7.39 Cross section of Osmunda sp. rhizome showing dissected stele and leaf traces (arrows) (Extant). Bar  2 mm.

Figure 7.37 Cross section of Psaronius vascular tissue showing preservation of phloem (arrows) (Pennsylvanian). Bar  1mm.

levels, the xylem and phloem at the edges of the gap appear closer together and eventually the interruption is no longer present. In some taxa, many leaf gaps are present in a single cross section and the stele appears to be dissected into

segments. Such a dissected siphonostele is called a dictyostele (FIG. 7.38). Dictyosteles are typical of many extant and fossil ferns (FIG. 7.39). Since the simplest and oldest stelar type is a protostele, it is hypothesized that plants with siphonosteles evolved from protostelic ancestors, and there are examples in the fossil record that support this hypothesis. Two principal theories have historically been advanced to explain the origin of the siphonostele from the protostele (Ogura, 1972). The intra-stelar origin theory (FIG. 7.40) suggests that, during the course of stelar evolution in some plant groups, cells in the center of the protostele did not mature into tracheids (FIG. 7.41B). The resulting medullated protostele would represent an intermediate

Chapter 7 INTRODUCTION TO VASCULAR PLANT MORPHOLOGY AND ANATOMY

(A)

(B)

(C)

219

(D)

Figure 7.40 Diagrammatic stages in the evolution of the siphonostele according to the intrastelar theory: A. haplostele; B. medullated protostele; C. siphonostele with the beginning of a leaf trace; D. siphonostele with C-shaped leaf trace. (From Taylor and Talyor, 1993.)

(A)

(B)

(C)

(D)

Figure 7.41 Diagrammatic stages in the evolution of the siphonostele according to the extrastelar theory: A. haplostele; B. departure of leaf trace causing an interruption in the vascular cylinder; C. vascular cylinder closed at arrow; D. siphonostele with C-shaped vascular trace. (From Taylor and Taylor, 1993.)

stage between a protostele and a siphonostele, in which the central region contains both tracheids and parenchyma cells. According to this theory, the siphono-stele has evolved by the failure of certain procambial cells to develop into tracheids. Fossil evidence to support this theory can be found within the lycopsids, where protostelic forms occur early in the history of the group, followed by plants with medullated protosteles, and finally by those with siphonosteles (Chapter 9). The second hypothesis on the origin of the siphonostele, often termed the extrastelar theory (FIG. 7.41), views the siphonostele as evolving by the continued expansion of cortical parenchyma toward the stem center during the production of leaf traces from the surface of a protostele. In this scenario, cortical parenchyma became “trapped” as the xylem became continuous after trace departure (FIG. 7.41C). The production of a large number of leaf traces from a protostele would

eventually result in a stele in which the center contains parenchymatous pith. Some of the Paleozoic ferns best illustrate this pattern of stelar evolution, for example, the early protostelic botryopterid ferns and Grammatopteris, considered to be a progenitor of the osmundaceous ferns (Chapter 11). SEED PLANTS The primary vascular tissue in seed plants is arranged in a fundamentally different way from non-seed plants. Xylem and phloem occur in distinct strands called sympodial bundles or sympodial strands (vascular bundles), which are embedded in parenchymatous ground tissue. This stelar type is characteristic of seed plants and is called a eustele. The vascular strands are arranged either in a ring around the central pith, as in gymnosperms and dicotyledonous angiosperms, or scattered throughout the ground tissue (atactostele)

220

PaleOBotany: the biology and evolution of fossil plants

(FIG. 7.42), as in monocots. In a single cross section, a eustele may look like a dictyostele, in that the cylinder of vascular tissue appears dissected. The sympodial strands in seed plants, however, are discrete and continue throughout the stem (FIG. 7.43C, D). The ground tissue is also continuous from the pith to the cortex, that is, around the sympodial bundles. When a leaf trace is produced from a eustele, a stelar

Figure 7.42 Cross section of Zea stem atactostele showing

scattered vascular bundles (Extant). Bar  2 mm.

(A)

(B)

bundle divides in to two, with one part of the bundle separating tangentially to supply the leaf trace and the other part remaining as a sympodial strand (FIG. 7.43C, D) (Namboodiri and Beck, 1968a, b). Vascular bundles in seed plants are most commonly collateral, with primary xylem on the inside and primary phloem on the outside. Bicollateral bundles also occur; these have phloem both internal and external to the xylem. It was once believed that seed plants had their origin within the ferns (Jeffrey, 1917) and that the eustele evolved by continued dissection of a siphonostele, in part because of the similarity of stelar anatomy in cross section. We now know, based on many lines of evidence, including a better knowledge of early land plant lineages, that the ferns and seed plants had separate evolutionary histories, and the siphonostele and eustele did not have a shared evolutionary history. Progymnosperms and early members of the seed ferns provide evidence of the evolution of the eustele. Namboodiri and Beck (1968c) proposed that the eustele evolved through the continued longitudinal dissection of a protostele (FIG. 7.43B), that is, the intercalation of parenchyma into the vascular tissue. When leaf traces were produced, no gap was formed in the stele. Several Devonian taxa, including the progymnosperm Aneurophyton (Chapter 12), had ribbed protosteles and could serve as model starting points in this progression (FIG. 7.43A). Continued medullation of a ribbed protostele resulted in a three-stranded vascular system (FIG. 7.43B), similar to that seen in several species of Stenomyelon, and finally in a central pith with bundles around the periphery, a stele type seen in Archaeopteris. Finally, a

(C)

(D)

Figure 7.43 Suggested stages in the evolution of the gymnosperm eustele. A. Lobed protostele, for example, Stenomyelon primaevum (Chapter 14). B. Longitudinal dissection of protostele to form pith, for example, S. tuedianum. C. Continued dissection giving rise to discrete sympodial bundles. Trace formation via tangential division of sympodia, e.g., Calamopitys. D. Trace formation via radial division of sympodial bundles, resulting in the formation of a primary vascular system like that in most gymnosperms. (Redrawn from Namboodiri and Beck, 1986; in Taylor and Taylor, 1993.)

Chapter 7 INTRODUCTION TO VASCULAR PLANT MORPHOLOGY AND ANATOMY

change in the production of leaf traces (FIG. 7.43D), so that sympodial bundles divided tangentially to produce traces, is illustrated by several Carboniferous seed ferns, such as Lyginopteris (Chapter 14). Some modern conifers (Chapter 21) have sympodial strands that undulate through the ground tissue.

LEAF MORPHOLOGY AND ANATOMY Leaves demonstrate the greatest morphological variability of any plant organ. They are also known for their plasticity, that is, difference in leaf form within a single species. Except in the earliest vascular plants, modern Equisetum, Psilotum, and Ephedra, and a few specialized angiosperms such as stemsucculent cacti, which have highly reduced leaves or are leafless, leaves function as the primary photosynthetic units in plants. Cotyledons, or seed leaves, represent the first leaves produced during embryonic development and they function as storage organs, providing food for the developing plant until the first true leaves appear and begin to photosynthesize. In many plants that live in arid environments, leaves may be fleshy and function in water storage. Modified leaves, which may be non-photosynthetic, are important as protective structures, for example, bud scales, and as parts of reproductive structures, such as flower petals and sepals or floral bracts, where they may be brightly colored and serve as signals for pollinators. Although leaf form is highly variable, most consist of a flattened blade borne on a narrow, elongate axis—the petiole. At the point of attachment to the stem (the axil of the leaf), many leaves have an abscission layer, which functions in separating the leaf from the stem and preventing water loss at the same time. Morphologically, leaves are classified as simple or compound; compound leaves are made up of leaflets. If the leaflets are attached at a single point, the leaf is palmately compound, for example, Sagenopteris, a leaf type found in the Mesozoic Caytoniales (Chapter 15). If the leaflets are attached along the petiole, the leaf is pinnately compound. This leaf morphology is very common in ferns, where multiple levels of leaflets may occur (Chapter 11). In a pinnately compound leaf, the term petiole on stipe used below the level of leaflet attachment. The region where the leaflets are attached is called the rachis (pl. rachides). The morphology of fossil angiosperm leaves, including overall shape and features of the margin, are widely used by paleobotanists to reconstruct paleoclimates using leaf physiognomy (see Chapter 1). Leaf venation is highly variable among vascular plants and has evolved differently in different groups of plants at various points in geologic time. The non-seed plants tend

221

to have venation patterns in which the vascular bundles dichotomize to fill the leaf, and veins often end at the leaf margin (open venation). Seed plants have more complex venation. If they exhibit dichotomous venation, the veins may also anastomose (join together) and then dichotomize again, forming enclosed meshes. This type of venation is seen in the Permian pteridosperm, Glossopteris (Chapter 14) and several other fossil seed plants. Veins often do not end at the leaf margin but bend back to fuse with other veins (closed venation). Some fossil gymnosperms exhibit relatively complex venation with multiple orders of veins (primary, secondary, tertiary, etc.), but the flowering plants have evolved the most diverse and complex pattern of venation. They include multiple, distinct orders of venation, and the ultimate veinlets enclose small patches of the leaf, termed areoles. Doyle and Hickey (1976) were able to trace the progressive changes in venation patterns in some of the earliest angiosperm leaf fossils and demonstrated a progressive increase in the organization of veins and the number of orders of veins (see Chapter 22) in successively younger rocks. Venation patterns in fossil leaves are an important systematic character, especially when coupled with leaf morphology and epidermal anatomy. LEAF ANATOMY

Leaves are composed of three principal tissue systems: epidermis, mesophyll, and vascular tissue, but the organization and extent of each of these systems are almost as variable as leaf morphology. Because a leaf is typically a dorsiventral structure, the epidermis of the abaxial surface is often different from that on the adaxial side. When stomata occur on both the adaxial (upper) and abaxial (lower) surface, the leaf is amphistomatic, on only the upper surface, epistomatic, and on only the lower surface, hypostomatic. The epidermis of leaves often bears trichomes that help to decrease water loss or glandular trichomes, which may help to discourage herbivores. The adaxial epidermis often has a thicker cuticle and fewer stomata than the abaxial side. On some leaves a hypodermis of thick-walled cells may be present immediately internal to the epidermis; this layer is believed to provide mechanical support for the leaf. The mesophyll tissue makes up the major part of the leaf and often consists of two types of cells. Palisade parenchyma consists of thin-walled, columnar cells with numerous chloroplasts (FIG. 7.44). These chlorenchyma cells are typically arranged in rows and are the principal photosynthesizing cells in most leaves. Beneath the palisade parenchyma (PP) is the spongy mesophyll (SM), which is characterized by thin-walled cells that are widely separated by lacunae or

222

PaleOBotany: the biology and evolution of fossil plants

PP

SM

may extend for several orders of branching (primary, secondary, etc.), depending upon the particular leaf type or plant group. The leaves of gymnosperms (FIG. 7.45) possess a slightly different tissue organization than that found in a typical angiosperm leaf. They often possess greater amounts of sclerenchyma and, in many instances, the vascular bundles are surrounded by a transfusion tissue composed of parenchyma and short tracheids. This tissue is believed to assist in conduction of materials between the vascular bundles and the mesophyll. Some leaves produced secondary vascular tissue, and stomata may be scattered on both surfaces. Many conifers, however, produce stomata in rows, termed stomatiferous bands. LEAF EVOLUTION

Figure 7.44 Section of Nymphaea leaf showing (palisade

parenchyma) and SM (spongy mesophyll). Arrow indicates a sclereid with simple pits on the wall (palisode parenchyma) (Extant). Bar  200 μm.

True leaves appear to have evolved at least twice in plant evolution. Microphylls are believed to have evolved from enations (Chapter 8) and represent a synapomorphy for the Lycophyta. They are generally, but not always, small, have a single vascular bundle, and a leaf trace that leaves no gap when it departs the stele. The evolution of microphylls is discussed in more detail in Chapter 9. Megaphylls evolved from branching systems generally have more complex vasculature and produce a leaf gap when the leaf traces departs the stele. Contrary to the name, they are not all large in size. All groups of vascular plants except the lycopsids possess megaphylls, and the evolution of this leaf type will be discussed in Chapter 11.

FURTHER READING

Figure 7.45 Cross section of Pinus sp. leaf with two vascular

bundles and resin canals (arrows) in cortex (Extant). Bar  350 μm.

intercellular spaces (FIG. 7.44). This tissue is ideally suited for the circulation of carbon dioxide and also provides a degree of flexibility to the leaf. Spongy mesophyll cells also contain chloroplasts, but they are generally not as densely packed as in palisade cells. The third component of a leaf is the vascular system or vein system. A transverse section of a leaf reveals many vascular bundles, each surrounded by a layer of cells termed a bundle sheath. Vascular bundles in a leaf are usually collateral, but bicollateral and concentric ones occur in some groups. When collateral, the xylem is located toward the abaxial surface with the phloem below it; secondary tissues may be present as well. The veins

Obviously, all facets of the morphology and anatomy of vascular plants cannot be covered in this brief summary. A comprehensive general botany textbook is a good source for more information about the structure and diversity of vascular plants (Mauseth, 2003; Raven et al., 2005; Graham et al., 2006; Stern, 2006). For additional details on plant morphology, see the classic three-volume textbook by Karl Goebel (1928–1933), or Bierhorst (1971) or Gifford and Foster (1989). There are a number of excellent plant anatomy texts, including Esau (1965, 1977), Fahn, (1990), Mauseth (1988), Dickison (2000), Beck (2005), Evert (2006), and Cutler et al. (2007), as well as those that concentrate on angiosperm anatomy, for example, Rudall (2007), or the two series of volumes, Anatomy of the Dicotyledons (Metcalfe and Chalk, 1979, 1983; Metcalfe, 1987) and Anatomy of the Monocotyledons (currently nine volumes), both series edited by anatomists at Kew Gardens.

8 EARLY LAND PLANTS WITH CONDUCTING TISSUE

CONDUCTING ELEMENTS IN EARLY LAND PLANTS .. 224

Discussion: Rhyniophyte Evolution ................................................. 251

HISTORY OF DISCOVERY ...................................................... 225

ZOSTEROPHYLLOPHYTES...................................................... 252

RHYNIOPHYTES......................................................................... 227

Zosterophyll Evolution..................................................................... 259

Rhynie Chert Plants ..........................................................................228

TRIMEROPHYTES ........................................................................259

Gametophyte Generation ................................................................. 241

Trimerophyte Evolution ................................................................... 262

Other Rhyniophytes ......................................................................... 246

EARLY LAND PLANT EVOLUTION ..................................... 263

I would be met and meet you so, In a green airy space, not locked in. Denise Levertov, About Marriage Many paleobotanists regard the upper part of the Silurian as the point in geologic time when the first plants with organized conducting tissue appear. Others have suggested that vascular plants occur in strata as old as the Cambrian (Kryshtofovich, 1953). Some of these pre-Devonian fossils have subsequently been demonstrated to be the remains of nonvascular plants or even animals (Theron et al., 1990). In other instances, reinterpretation of the age of the rocks containing the fossils has negated reports of early vascular plants. The early lycopsid, Baragwanathia, is an unusual case in this regard. (Garratt, 1978) It was initially described from Upper Silurian compressions from Australia (Lang and Cookson, 1935) and the age was based on the occurrence of the graptolite Monograptus. Baragwanathia (FIG. 8.1) will be discussed in more detail in Chapter 9, but it is most certainly a vascular plant, with annular-helical tracheids forming the conducting strand. The age of this plant has been debated in the literature ever since as to whether the rocks are truly Upper Silurian or Lower Devonian. Subsequent studies have confirmed the Late Silurian (Ludlow) age (Rickards, 2000). The age has

been hotly debated (Thomas, 1984) because Baragwanathia represents a relatively complex vascular plant at a point in geologic time when all other vascular plants were comparatively simple. This suggests that either vascular plants evolved far earlier or that tracheids perhaps evolved more than once during the terrestrialization of the earth. As we learn more about early land plants, including those with well-defined conducting tissues, it is becoming clear that a number of these early plants did not possess vascular tissue like that in extant vascular plants. Although identifying plants with vascular tissue early in the geologic record is important, what is perhaps equally significant is understanding the nature of these cells.

223

Higher taxa in this chapter:

Rhyniophytes (Silurian–Devonian) Zosterophyllophytes (Silurian–Devonian) Trimerophytes (Silurian–Devonian)

224

Paleobotany: the biology and evolution of fossil plants

Figure 8.1 Baragwanathia longifolia (Silurian). Bar  2 cm.

Figure 8.2 Oblique view of Gosslingia breconensis conducting

elements (Devonian). Bar  10 μm. (From Edwards and Kenrick, 1988a.)

CONDUCTING ELEMENTS IN EARLY LAND PLANTS Kidston and Lang (1920a), in their detailed descriptions of plants from the Rhynie chert (see below), noted that they were unable to resolve secondary wall thickenings in the central strand of Rhynia major (now Aglaophyton) (FIG. 7.1). When David S. Edwards (1986) demonstrated that these conducting elements were structurally different from tracheids, he provided the first evidence that Aglaophyton was not a true vascular plant. Additional detailed studies of water-conducting elements in other early land plants (Kenrick and Crane, 1991) indicate a level of structural complexity in these elements, and demonstrate that they are quite different from the tracheids of vascular plants. Kenrick and Crane (1991) delimit three types of water-conducting cells in Early Devonian plants: (1) unornamented, elongate cells, found in the Rhynie chert taxa Aglaophyton major, Nothia aphylla, and the gametophytes Lyonophyton rhyniensis and Kidstonophyton discoides (Edwards, 2003). These are comparable in structure to the hydroids of mosses; (2) S-type, and (3) G-type conducting cells. Kenrick et al. (1991a) reported S-type water-conducting cells in the central strand of Sennicaulis, an Early Devonian rhyniophyte from Wales (Edwards, 1981). Each cell in the central strand has simple helical thickenings, but at the ultrastructural level the wall is unlike that of vascular plant tracheids. In S. hippocrepiformis, the cell wall consists of two principal layers. The outer layer makes up the helical thickenings, which possess a spongy internal organization. This is bounded on the lumen side by a thin microporate

layer. In addition to Sennicaulis, S-type conducting cells are characteristic of various rhyniophytes, including Huvenia kleui, Stockmansella langii, and S. remyi (Fairon-Demaret, 1985, 1986; Kenrick et al., 1991b; Schultka and Hass, 1997). G-type elements were first described in the Early Devonian zosterophyll Gosslingia breconensis (FIGS. 8.2, 8.3) (Kenrick and Edwards, 1988a), but also occur in other zosterophyllophytes, Hsüa (Li, 1992), Barinophyton (Brauer, 1980), and various early lycophytes (e.g., Asteroxylon, Drepanophycus; Chapter 9) (Berry and Fairon-Demaret, 2001). G-type cells have a two-layered wall. The inner, decay-resistant layer makes up a series of annular thickenings, but also forms a continuous wall area between the thickenings, where it contains small pits. The outer layer is non-resistant and may be mineralized. Following Kenrick and Crane’s (1991) delimitation of these original three types, three more forms of water-conducting cells have been recognized (Edwards, 2003) (FIG. 8.4): the P-, C-, and I-types. Species of Psilophyton are representative of the P-type. These cells exhibit scalariform bars, but the bars are narrowly attached to the cell wall, giving the appearance of scalariform bordered pitting. They have a decayresistant inner layer with decay-resistant material covering the pit apertures, forming strands or holes within the scalariform bars (Hartman and Banks, 1980). Structurally similar cells occur in later Paleozoic lycophytes, including the late Middle Devonian herbaceous form Minarodendron (Li, 1990). The C-type is known exclusively in Cooksonia pertonii (Edwards et al., 1992). These cells resemble conventional annular or

CHAPTER 8 EARLY LAND PLANTS WITH CONDUCTING TISSUE

225

with the hypothesis that turgor pressure in parenchyma cells was the primary means of maintaining rigidity in these early land plants (Speck and Vogellehner, 1988). Although some authors call these cells “tracheids,” we have chosen to use the term conducting cells, since their homologies with either true tracheids or bryophyte hydroids remain uncertain. These cell types do, however, provide a functional perspective that can be used to test hypotheses about water conduction in early terrestrial ecosystems (Edwards et al., 1998b; Cook and Friedman, 1998; Friedman and Cook, 2000) (FIG. 8.5).

HISTORY OF DISCOVERY

Figure 8.3 Diagrammatic section of G-type conducting ele-

ment of Gosslingia breconensis. (Courtesy P. Kenrick.)

spiral tracheids, except that the imperforate lateral walls are thicker than the primary wall of protoxylem elements (Edwards, 2003). Indeterminate stomatiferous axes illustrate the I-type. In this bilayered wall, the outer layer is imperforate and fused with that of adjacent cells, whereas the inner layer has rounded perforations. In section, the perforations are slightly wider toward the middle lamella and so superficially resemble bordered pits (Chapter 7) (Edwards and Axe, 2000; Edwards et al., 2003). Such internal wall thickenings may represent structural features of the cells which help to reduce embolism in early homoiohydric plants. This hypothesis is supported by the presence of G-type cells in the main axes of Sawdonia ornata from Röragen, Norway, but not in the laterals (Edwards et al., 2006b). The spongy organization of the wall may represent the template where lignin was deposited once this biochemical pathway evolved. It is not known whether the central strand cells in many early plants contained sufficient amounts of lignin to be useful in support. The small size of the central strand compared to the diameter of the stem, however, is more consistent

Our understanding of early vascular plants has an interesting history that, to a large degree, has greatly influenced many areas of paleobotany. In 1859, the Canadian geologist and paleobotanist Sir John William Dawson (FIG. 8.6) published a report on a Devonian vascular plant collected from the Gaspé region of Nova Scotia. His reconstruction showed a horizontal rhizome bearing upright, leafless, dichotomizing axes, to which were attached pairs of sporangia. Dawson named this interesting plant Psilophyton princeps (FIG. 8.7). Dawson’s scientific colleagues virtually ignored this important discovery, however, perhaps because the plant he reconstructed looked so unusual and certainly because of its age. Several years later (Dawson, 1871) he described additional specimens, but, again, these were not seriously considered by the scientific community of the day. In the years that followed, other discoveries were made on plants with obvious vascular tissue, and gradually Dawson’s initial report of Devonian vascular plants gained acceptance (Dawson, 1888). One of the most spectacular discoveries in paleobotany finally proved beyond any doubt that vascular plants existed by the Early Devonian. Beginning in 1917, Robert Kidston (FIG. 8.8) and William Lang published a series of papers detailing some exquisitely preserved vascular plants collected near the village of Rhynie, in Aberdeenshire, Scotland. This fossil-bearing rock (FIG. 8.9) consists of a fine-grained chert that is now regarded as coming from the upper part of the Lower Devonian, and dated at approximately 400 Ma (Rice et al., 1995). Recent palynological studies suggest a Pragian–?earliest Emsian age for the deposits (Wellman et al., 2006; and Wellman, 2007). Most of the fossils from the Rhynie locality showed that these plants did in fact consist of dichotomizing and, in general, leafless aerial stems (FIG. 8.10) arising from a horizontal aboveground or subterranean rhizomatous system. At the ends of some axes were terminal sporangia.

226

Paleobotany: the biology and evolution of fossil plants

Carboniferous

1

363

2

3

4

5

6

7

8

9

10

11

12

13

14

15

Upper

Cenozoic

Famennian 367 Frasnian 377 Middle

Givetian 381 Eifelian 386 Emsian 390

Lower

Mesozoic

Devonian

65

Pragian (Siegenian) 396 Lochkovian (Gedinnian) 409

250

Pridoli 411

Silurian

Palaeozoic

Ludlow

421 Wenlock 430 Llandovery 439

Precambrian

Ordovician

Ashgill 570

Caradoc Llandeilo/ Llanvirn Arenig

Cambrian

500

Known range Possible range Probable range but derivation different Probable range based on megafossils

Tremadog

Figure 8.4 Stratigraphic ranges of microfossils, megafossils and various anatomical features. Numbers across the top refer to (1) Dyads and tetrads of possible bryophytic origin. (2) Single spores with trilete marks. (3) Cooksonia megafossils. (4) Bifurcating axes of putative vascular plants. (5) Nematophyte cuticle. (6) Higher plan cuticle. (7) Stomata on axial fossils and lycophytes. (8) Banded tubes. (9) C-type conducting elements. (10) G-type conducting elements. (11) Zosterophylls. (12) Baragwanthia, Drepanophycus, lycopsids. (13) S-type conducting elements. (14) P-type conducting elements. (15) Trimerophytes. (From Edwards, 2003.)

Since the discovery and publications by Kidston and Lang (1917–1921), additional simple land plants have been described from Devonian and pre-Devonian rocks. For many years most of these early plants were placed in a single order, Psilophytales. These fossils, along with two living

plants, Psilotum and Tmesipteris, made up a separate subdivision of vascular plants, Psilopsida (or Psilophyta). Today this classification is rarely used. Some include Psilotum and Tmesipteris in their own division (Gifford and Foster, 1989), whereas others regard the two extant genera as having their

CHAPTER 8 EARLY LAND PLANTS WITH CONDUCTING TISSUE

Resistant layer

Resistant layer

Spongy layer

Degradationprone layer

Primary cell wall

S-type conducting cells (Rhyniopsida)

Primary cell wall

G-type conducting cells (early Lycophytina)

227

Resistant layer

Resistant layer Degradationprone layer Primary cell wall

Primary cell wall

P-type conducting cells (early Euphyllophytina)

Seed plant tracheids (recent Euphyllophytina)

Figure 8.5 Longitudinal section of cell wall thickenings in fossil S-, G-, and P-type conducting elements and tracheids of extant seed plants showing primary cell wall components. (Modified from Cook and Friedman, 1998.)

Figure 8.6

John W. Dawson. (Courtesy H. N. Andrews.)

closest affinities with certain ferns (Bierhorst, 1968; 1971), a placement that is supported by molecular studies (Manhart, 1995). As additional Silurian–Devonian plants were discovered and carefully evaluated (Andrews et al., 1977), it became apparent that there were suites of characters that might be used to define larger taxonomic groups among the fossils (Høeg, 1954). Harlan Banks (1975) (FIG. 8.11) was the first to propose abandonment of the Psilophytales, which had become a repository for all types of unrelated early plants. In its

place he established three subdivisions—the Rhyniophytina, Zosterophyllophytina, and Trimerophytina. In the earliest cladistic analysis of early land plants (Kenrick and Crane, 1997a, b) the rhyniophytes and trimerophytes were not considered monophyletic, whereas the zosterophyllophytes were similar to the Zosterophyllophytina of Banks, with the inclusion of several other taxa. These fossils are now included in the polysporangiates, a clade of all land plants that bear multiple sporangia in the sporophyte phase, which includes both vascular plants and nonvascular plants (e.g., Aglaophyton). The Eutracheophytes contain all extant vascular plants and most vascular plant fossils, and are further subdivided into the Euphyllophytina and Lycophytina. The characters that are interpreted as plesiomorphic or derived (apomorphic) will be continually debated in the case of certain fossil plants. It is especially difficult to characterize basal groups, such as the earliest land plants, as by definition they will contain multiple plesiomorphies that are shared by all later-evolving land plants. With those limitations in mind, this chapter will discuss the Late Silurian–Early Devonian record of “vascular” land plants by reference to Banks’ original three groups.

RHYNIOPHYTES Some of the plants included in this group were previously included in the Rhyniophyta (Rhyniophytina of Banks, 1975). They can be characterized by dichotomously branched, naked aerial axes with terminal sporangia. The aerial axes arise from horizontal, dichotomizing rhizomes that bear rhizoids; no true roots are known. Sporangial shape

228

Paleobotany: the biology and evolution of fossil plants

Figure 8.8 Robert Kidston. (Courtesy University of Aberdeen.)

Figure 8.7 Suggested reconstruction of Psilophyton princeps.

Figure 8.9 Trench exposing Rhynie chert bed. Hagen Hass,

(From Taylor and Taylor, 1993.)

far right.

varies from ellipsoidal to branched, and some rhyniophyte sporangia appear to have an opening at the tip; others have been described with an abscission layer at the base of the sporangium. When axes are structurally preserved, they contain a small, terete, centrarch conducting strand of S-type conducting elements. Spores are all of the same morphological type, and hence the plants are considered homosporous. Cooksonia (FIG. 8.12), considered by many to represent the first vascular plant, is typical of this group (see section “Other Rhyniophytes”). We will begin, however, with the

plants from the famous Rhynie chert Lagerstätte, as they represent the best-known of the early land plants and have provided so much information about early land plants in their ecosystems, even though a number of them are now placed within the zosterophylls (i.e., Ventarura, Trichopherophyton, early lycophytes (Asteroxylon). RHYNIE CHERT PLANTS

There can be little doubt that the Rhynie chert organisms have had a profound influence on many of our hypotheses

CHAPTER 8 EARLY LAND PLANTS WITH CONDUCTING TISSUE

229

Figure 8.10 Section of Rhynie chert showing closely spaced

upright axes and matrix (Devonian). Bar  600 μm.

Figure 8.12 Suggested reconstruction of Cooksonia caledonica.

(From Taylor and Taylor, 1993.)

lignieri, and Asteroxylon mackiei. Since the original description, additional land plants have been described from the chert lenses, including Nothia aphylla (Lyon, 1964; ElSaadawy and Lacey, 1979) and Trichopherophyton teuchansii (Lyon and Edwards, 1991), and from coeval deposits at the nearby Windyfield chert site, that is, Ventarura lyonii (Powell et al., 2000).

Figure 8.11

Harlan P. Banks. (Courtesy H. N. Andrews.)

about the early evolution of land plants. In recent years, however, some of these ideas have been challenged in light of new discoveries from the chert beds and the reexamination and reinterpretation of some of the plants. In their pioneering series of papers (1917–1921), Kidston and Lang described four plant taxa: Rhynia major, R. gwynne-vaughanii, Hornea

AGLAOPHYTON MAJOR The best known plant from the Rhynie chert is Aglaophyton major (FIG. 8.13), a macroplant originally described as Rhynia major, but transferred to a new genus by Edwards based on a reexamination of the original Kidston and Lang slides and the discovery of new specimens (D.S. Edwards, 1986). In a very real sense, A. major is the Arabidopsis of the Devonian, not because its genome is well known, but because the plant and its life history are known in such detail (FIG. 8.14). Aglaophyton major is now reconstructed as a plant 18.0 cm tall that consists of a system of naked, stomatiferous (FIG. 8.15), more or less cylindrical, and sinuous prostrate axes, which are loosely lying on the substrate surface and function as rhizomes. The prostrate axes dichotomize repeatedly, periodically turning upwards and passing

230

Paleobotany: the biology and evolution of fossil plants

Figure 8.13 Model of Aglaophyton major. (Courtesy N. Trewin.)

8.15 Stoma of Aglaophyton major (Devonian). Bar  25 μm. (Courtesy H. Kerp.)

Figure

Figure 8.14 Suggested life history of Aglaophyton major/

Lyonophyton rhyniensis showing stages in the development of the dimorphic gametophytes. Mature sporophyte (lower left, not to scale) bears sporangia with spores of two types. Blue spores (color is arbitrary) develop into antheridiophores; orange spores develop into archegoniophores (From Taylor et al., 2005c.)

into upright, fertile axes 6.0 mm in diameter that bear terminal sporangia (FIG. 8.16) with trilete spores (Wellman et al., 2006). There is some suggestion that the apex twisted during the development of the sporangia or that torsion of the sporangium was somehow involved in spore release. This sporangial feature is also found in the Early Devonian compression genus Tortilicaulis and has been compared with the morphology of certain bryophyte sporangia (Chapter 5).

The anatomy of the prostrate and upright axes is simple (FIG. 8.17); most of the axis consists of a parenchymatous cortex (FIG. 8.18), which is subdivided into inner and outer zones. The outer cortex is composed of densely packed elongated cells with narrow intercellular spaces. The cells of the inner cortex are more loosely spaced and exhibit a well-developed intercellular system (FIG. 8.19). The outer cortex is surrounded by hypodermal tissues and epidermis. Between the outer and inner cortex, there is a well-defined region of cortical tissue in which an endomycorrhizal fungus, Glomites rhyniensis, forms intracellular arbuscules (Chapter 3). The most significant component of Edwards’ (1986) work, however, and the reason that a new genus was created, involves the nature of the central conducting strand. As mentioned earlier, although Kidston and Lang (1920a) originally described A. major as a vascular plant, they were never able to discern secondary wall thickenings on the conducting cells. They ascribed the absence of thickenings to poor preservation, but David Edwards showed that the conducting strand of A. major was actually something quite different. He found that the central strand consisted of three regions. The outer zone, topographically in the position of the phloem, was made up of

CHAPTER 8 EARLY LAND PLANTS WITH CONDUCTING TISSUE

231

Figure 8.18 Transverse section of Aglaophyton major axis showing central strand and dark cell zone (arrow) containing arbuscules (Devonian). Bar  650 μm.

Figure 8.16 Longitudinal section of Aglaophyton major sporangium filled with spores (Devonian). Bar  4 mm. (Courtesy H. Kerp.)

Figure 8.19 Transverse section of Aglaophyton major axis showing thin-walled cells of inner cortical zone surrounding central strand (Devonian). Bar  110 μm.

Figure 8.17 Cross section of chert block in FIG. 8.10 showing numerous axes (Aglaophyton major) in transverse section (Devonian). Bar  1 cm.

thin-walled, elongated cells with oblique, S-shaped end walls. The remainder of the central strand consisted of two zones of opaque cells that were initially thought to represent tracheids. The cells in the outer zone are circular in transverse section and 50 μm in diameter. In longitudinal section these cells exhibit a reticulate patterning on their walls, but this appears to be the result of cell wall degradation or silica crystallization. Cells in the center of the strand are uniformly thin walled and more angular in cross section. In longitudinal section, the conducting elements in A. major have S-shaped walls with partly helical

232

Paleobotany: the biology and evolution of fossil plants

Figure 8.20 Tetrads of spores (Devonian). Bar  65 μm.

Figure 8.21 Germinating spore of Aglaophyton major (Devonian). Bar  50 μm. (Courtesy H. Kerp.)

Figure 8.22 Germinating spore (Devonian). Bar  70 μm. (Courtesy H. Kerp.)

thickenings, consisting of a spongy outer layer and inner, delicate decay-resistant component. Edwards concluded that the cells of the conducting strand of Aglaophyton are more similar to the leptoids and hydroids found in certain bryophytes than to the sieve cells and tracheids of vascular plants. It is clear from this important work that this plant combines features of both bryophytes and vascular plants, and thus it is placed within an informal group of early land plants, the cooksonioids (rhyniophytoids), which includes dichotomizing axes for which there is no information about the presence or absence of a conducting strand. One of the most influential reports involving the Rhynie chert plants was the discovery of free-living gametophytes in the chert lenses (Remy and Remy, 1980a, b). Since the initial description of Lyonophyton rhyniensis, which is the gametophyte of A. major, additional details have been added, (Fig. 8.14) including a sequence of stages leading from spore germination (FIGS. 8.20–8.28) to the development of

CHAPTER 8 EARLY LAND PLANTS WITH CONDUCTING TISSUE

233

Figure 8.23 Germinating Aglaophyton major spore showing initial division (arrow) (Devonian). Bar  30 μm. (Courtesy H. Kerp.)

Figure 8.25 Spore (arrow) germinating and giving rise to

multicellular gametophyte Lyonophyton rhyniensis (Devonian). Bar  20 μm. (Courtesy H. Kerp.)

Figure 8.24 Germinating spore of A. major showing stages of

cell division (Devonian). Bar  50 μm. (Courtesy H. Kerp.)

mature, unisexual antheridiophores—gametophytes that bear antheridia (FIG. 8.29) and archegoniophores—gametophytes that bear archegonia (FIG. 8.30) (Remy and Hass, 1996). Clusters of spores or spore balls are typically found in the

chert matrix, suggesting that the spores may have been shed en masse (FIG. 8.31). The gametangiophores (gametophytes) are fragmentary and there is still some question as to whether they arise from a common thallus (i.e., they are bisexual), or whether each arises from a single spore that only produces antheridiophores or archegoniophores, as has been hypothesized (Taylor et al., 2005c). The same type of gametangia-bearing structures occur in other Rhynie chert sporophyte–gametophyte associations, including Rhynie gwynne-vaughanii–Remyophyton delicatum (Kerp et al., 2004), Nothia aphylla–Kidstonophyton discoides (Remy and

234

Paleobotany: the biology and evolution of fossil plants

Figure 8.28 Section of Lyonophyton rhyniensis protocorm (Devonian). Bar  400 μm. (Courtesy H. Kerp.)

Figure 8.26 Multicelled gametophyte extending from ruptured spore wall (arrow) (Devonian). Bar  30 μm. (Courtesy H. Kerp.)

Figure 8.27 Multicellular gametophyte of Lyonophyton rhyniensis. Arrow indicates possible pyramid-shaped apical cell (Devonian). Bar  40 μm. (Courtesy H. Kerp.)

Hass, 1991b), and Horneophyton lignieri–Langiophyton mackiei (FIG. 8.32) (Remy and Hass, 1991c), perhaps adding support to the hypothesis that the Aglaophyton–Lyonophyton gametangiophores (FIG. 8.33) were unisexual. All of the Rhynie chert gametophytes contain conducting cells (FIG. 8.34) and are mycorrhizal (see below).

Figure 8.29 Antheridium with escaping sperm (Devonian). Bar  30 μm. (Courtesy H. Kerp.)

All of the Rhynie chert plants studied in detail suggest that various forms of vegetative reproduction (FIG. 8.35) were widespread. The presence of reduced branches, bulbils (FIG. 8.36), and bulges with disorganized vascular tissue on

CHAPTER 8 EARLY LAND PLANTS WITH CONDUCTING TISSUE

235

Figure 8.30 Transverse section of Lyonophyton rhyniensis

archegonium neck showing eight neck Bar  30 μm. (From Remy and Hass, 1991c.)

cells

(Devonian).

Figure 8.32 Suggested reconstruction of distal end of Langiophyton mackiei archegoniophore (Devonian). (From Taylor and Taylor, 1993.)

Figure

8.31 Spore

ball in Rhynie chert (Devonian).

Bar  1 mm.

axes and various types of fragmentation suggest that these plants were clonal and possessed multiple reproductive strategies. The environment in which they lived has been reconstructed as a freshwater ecosystem with volcanic influence including hot springs (Trewin et al., 2003). Thus, it seems likely that the various forms of asexual reproduction may have allowed rapid colonization in a variable environment. RHYNIA GWYNNE-VAUGHANII This species is reconstructed as a small plant, 18.0 cm tall, with upright, dichotomizing axes that arise from a rhizome

Figure 8.33 Longitudinal section through the antheridio-

phore of Lyonophyton rhyniensis showing two antheridia (arrows) (Devonian). Bar  4 mm. (Courtesy H. Kerp.)

236

Paleobotany: the biology and evolution of fossil plants

Figure 8.34 Longitudinal view of conducting elements in Lyonophyton rhyniensis (Devonian). Bar  12 μm. (Courtesy H. Kerp.)

Figure 8.36 Bulbil (arrow) extending from vegetative axis of

Aglaophyton (Devonian). Bar  1 mm. (Courtesy H. Kerp.)

Figure 8.35 Diagrammatic population of Aglaophyton major plants showing single individual (yellow box), association of plants by endomycorrhizae (brown filament), and fragmentation (red lines).

bearing delicate, threadlike rhizoids. The stems are 2.0– 3.0 mm in diameter with a narrow central strand. Surrounding the central conducting strand are some poorly preserved, thin-walled cells that have been suggested to be phloem (FIG. 8.37); some of these cells have been described as containing sievelike thin areas on their walls (Satterthwait and Schopf, 1972). The remainder of the axis consists of a twoparted, parenchymatous cortex. The inner cortex consists

Figure 8.37 Detail of Rhynia gwynne-vaughanii axis showing

histologic differences in cells (Devonian). Bar  125 μm.

of cells with large intercellular spaces and the narrow outer cortex of more tightly packed cells. A thin cuticle is present on the epidermis, as well as stomata, each with two simple, kidney-shaped guard cells. Sections of the rhizome show the

CHAPTER 8 EARLY LAND PLANTS WITH CONDUCTING TISSUE

237

Figure 8.38 In situ stand of Remyophyton delicatum axes

(Devonian). Bar  6 mm. (Courtesy H. Kerp.)

same complement of tissues as the aerial axes. On the tips of some aerial axes are ellipsoidal, thick-walled sporangia up to 3.0 mm long. The sporangial wall is constructed of a zone of outer palisade cells with thickened lateral walls termed a cohesion tissue. Rhynia gwynne-vaughanii was homosporous and produced spores in tetrahedral tetrads; each spore is 40 μm in diameter and ornamented by closely spaced spines. The gametophyte phase is called Remyophyton delicatum (Kerp et al., 2004) and consists of a dense cluster of 200 unbranched axes, 0.2–0.7 mm in diameter (FIG. 8.38), with the basal region represented by globular prothalli (protocorms) bearing rhizoids. Gametophytes are unisexual—either archegoniophores or antheridiophores. Archegoniophores are larger (10–15 mm long) and bear massive archegonia with long protruding necks. Antheridiophores are smaller (4–8 mm long), with stalked antheridia arising from the flattened upper surface. Conducting tissue is made up of S-type elements. A globular mass slightly larger than the gametophyte axis has been described extending from the archegonium in one specimen and this has been interpreted as a putative sporophyte (Kerp et al., 2004). In his reexamination of some of the Rhynie chert plants, David S. Edwards (1980) provided additional information about R. gwynne-vaughanii. His studies confirmed the existence of hemispherical projections and numerous short adventitious branches along the stem. Some of the latter structures are interpreted as underdeveloped lateral branches (FIG. 8.39), whereas others may represent asexual reproductive structures, which presumably broke off the parent plant. The hemispherical projections may produce rhizoids when they are in contact with the substrate. It has also been suggested that these projections may represent possible secretory structures or hydathodes. By documenting the presence

Figure 8.39 Rhynia gwynne-vaughanii showing possible early

stage of adventitious branch (arrow) (Devonian). Bar  650 μm.

of the adventitious branches, Edwards was able to show that R. gwynne-vaughanii was more monopodial in growth architecture than Aglaophyton. In addition, he suggested that sporangia abscised after the release of spores, and in some instances, new branches formed distal to the abscission zone (Edwards, 1980). HORNEOPHYTON LIGNIERI This plant was originally described by Kidston and Lang (1920a) as Hornea, a name that was previously occupied by a flowering plant. Some features of Horneophyton lignieri are of special interest because they differ from those of other Rhynie chert plants. The plant consists of naked and dichotomously branched aerial axes (FIG. 8.40) up to 20 cm high and 2.0 mm in diameter, which are attached to a lobed, cormlike structure that bears numerous rhizoids on the lower surface. Although the structural features of the aerial axes are similar to those of Rhynia, with a central conducting strand, the basal corm lacks any evidence of vascularization. Recent studies indicate that in the basipetal regions of the aerial stem the phloem cells become indistinct, with conducting elements losing their characteristic features in more basal sections. At the transition between stem and corm, opaque parenchyma cells replace the conducting elements and these are distinguishable in the corm proper. Sporangia of H. lignieri are borne terminally at the tips of some of the branches. Each sporangium is branched, consisting of two to four lobes of varying length. Sporangial lobes tend to be ellipsoidal–cylindrical in shape, with the distal end truncated. It has been suggested that sporangial

238

Paleobotany: the biology and evolution of fossil plants

The sporangium in Horneophyton is unique among plants, both living and fossil, in that it consists of a branched fertile unit of four lobes that resulted from dichotomies of the stem apex (Eggert, 1974). Consequently, each fertile lobe was produced by its own apex, and these apices must have remained meristematic for a short time, as spores within the sporangium show evidence of acropetal maturation. The lobed Horneophyton sporangium also has been interpreted as transitional, leading to a synangium in which some differentiation and partitioning of sporogenous tissue took place.

Figure 8.40 Suggested reconstruction of Horneophyton lig-

nieri (Devonian). (From Taylor and Taylor, 1993.)

dehiscence took place through an apical pore. Extending into the sporangial cavity is a central column of sterile tissue around which a continuous zone of sporogenous tissue, or spores, developed. Both isobilateral and tetrahedral tetrads of spores occur in Horneophyton sporangia (Sharma and Bohra, 1985). The spores (see Wellman et al., 2004 for details) are radial, trilete, and irregularly ornamented by cavities that may represent some type of exine degradation. They range from 39–49 μm in diameter and are most similar to the sporae dispersae taxon Emphanisporites decoratus (Wellman et al., 2004). Some have been described as containing multicellular gametophytes (Bhutta, 1973). The plant is believed to have been homosporous. The gametophytes are described under the generic name Langiophyton mackiei (Fig. 8.32) (see below).

ASTEROXYLON MACKIEI The most complex element in the Rhynie chert flora is Asteroxylon mackiei. It is a particularly interesting plant that is sometimes included in the Drepanophycales (Chapter 9), although it lacks true leaves. Unlike the other Rhynie taxa, A. mackiei is characterized by numerous small flaps of tissue (leaflike appendages or enations) that cover the aerial stems, as well as a more complex central strand and conducting system. Asteroxylon mackiei was homosporous like the other Rhynie chert plants, but the sporangia were located laterally along the stem, not apically, as in the other taxa. In the original description of this plant, Kidston and Lang (1920b) described narrow and naked distal branches with terminal sporangia that, although not actually attached, were thought to be the fertile portion of the plant. The presumed terminal position of sporangia in A. mackiei was considered to be evidence that all the Rhynie plants were closely related. Subsequent studies of chert blocks containing A. mackiei axes, however, indicate that the sporangia were borne laterally on the stems, near the axils of leaflike appendages, instead of in a terminal position (Lyon, 1964). This important discovery greatly altered the taxonomic position of the genus. Asteroxylon mackiei was probably up to 50.0 cm tall and consisted of upright, monopodial axes supported by a horizontal, subterranean rhizome (Fig. 8.41), which may be up to 4.2 mm in diameter (Kidston and Lang, 1920b). From the principal aerial branches arose secondary branches that were regularly dichotomous. In contrast to other Rhynie plants, the aerial stems of Asteroxylon were densely covered by numerous leaflike flaps of tissue, each up to 5.0 mm long, and the taxon is known from compressions (Fig. 8.41) as well. These structures have been called leaves, scalelike leaves, leaflike scales or appendages, or enations. They are not true leaves (Chapter 7), as they contain no vascular tissue and are not produced in a regular pattern (phyllotaxy) from nodes, as true leaves are. The enations in Asteroxylon have been hypothesized to represent the precursors to microphylls, the leaf type which is unique to the Lycophyta (see Chapter 9 for a discussion). It is not certain that stomata present on the stems also existed on the enations.

CHAPTER 8 EARLY LAND PLANTS WITH CONDUCTING TISSUE

239

Figure 8.43 Transverse section of Asteroxylon mackiei showing actinostele and highly lacunate cortex characteristic of mature axes. Arrows indicate several traces in the cortex (Devonian). Bar  4 mm.

Figure 8.41 Suggested reconstruction of Asteroxylon mackiei

(Devonian).

Figure 8.42 Cross section of Asteroxylon mackiei stele

(Devonian). Bar  1 mm.

In the rhizomatous portion of Asteroxylon the conducting strand has a central core of thick-walled elements, while the strand is stellate or star shaped in cross section in the aerial axes (Fig. 8.42), the feature from which the generic name is derived. The stele of Asteroxylon has been interpreted as an Asteroxylon-type protostele (Ogura, 1972) or an actinostele. Primary xylem is slightly mesarch, with the protoxylem elements situated near the edges of the xylem ridges. Xylem elements have annular and helical secondary thickenings. Thin-walled cells in the furrows between the xylem ridges are thought to represent phloem. In a transverse section of an aerial axis, numerous small strands of vascular tissue can be seen in the cortex (FIG. 8.43). These traces originate at the outer edges of the xylem ridges and extend through the cortex, ending abruptly near the periphery of the stem at the bases of the enations; they do not pass into the enations themselves. Reniform sporangia, with dehiscence along the distal edge, are produced on short, vascularized pedicels scattered among the enations, apparently in no particular relation to the enations themselves. Sporangia are up to 7.0 mm long; spores are 40–60 μm in diameter and ornamented by closely spaced spines on the distal surface. NOTHIA APHYLLA Kidston and Lang (1920b) initially described naked axes bearing pear-shaped sporangia as the sporangia of Asteroxylon. Today these sporangia are recognized as belonging to N. aphylla (Lyon, 1964; El-Saadawy and Lacey, 1979). The

240

Paleobotany: the biology and evolution of fossil plants

Figure 8.44 Section of Nothia aphylla rhizome (Devonian). Bar  1.0 mm (Courtesy H. Kerp.) Figure 8.45 Suggested reconstruction of Nothia aphylla

N. aphylla plant consists of an aerial system of upright, dichotomously branching axes that arise from subterranean rhizomes 2.5 mm in diameter (FIG. 8.44); the entire plant was 20 cm tall (Kerp et al., 2001). The aerial axes are covered by slightly elongated emergences (FIG. 8.45), each of which bears a single stoma. While the aerial axis is anatomically similar to that of other Rhynie chert plants, the anatomy of the prostrate rhizome is unique. In cross section, the rhizome is bilaterally symmetrical with a median ridge (rhizoidal ridge) that extends the length of the axis on the lower surface; extending from the ridge are unicellular rhizoids up to 1.5 mm long (Fig. 8.44). The rhizome anatomy includes a central stele with a strand of fibrous conducting cells surrounded by a narrow zone of phloem-like tissue. Outside the phloem is a parenchymatous cortex, hypodermal tissues, and epidermis. The internal anatomy of the ventral rhizoidal ridge includes a rhizoid-bearing epidermis, several hypodermal layers of relatively large and radially arranged cells, files of thin-walled parenchymatous cells that connect to the stelar body, and individual, extrastelar conducting elements (xylematic elements of Kerp et al., 2001).

(Devonian). (From Daviero-Gomez et al., 2004.)

Sporangial position in N. aphylla is highly variable, with individual sporangia borne on adaxially recurved stalks whose attachment may range from helical to almost whorled. Sporangia are reniform (3.0 mm wide by 2.0 mm long) with dehiscence through a long, apical transverse slit; spores are radial (65 μm) and trilete. The gametophytes are assigned to Kidstonophyton discoides and are unisexual (Remy and Hass, 1991b). Like many, and perhaps all of the Rhynie chert plants, N. aphylla is highly clonal, with numerous plantlets borne on the rhizomes (Daviero-Gomez et al., 2005). These appear as lateral buds on the rhizome, and in sections of the chert in which the plants are preserved, it is possible to see several levels of rhizomes in section view. In certain anatomical features, N. aphylla may be compared with members of the old Rhyniophyta, whereas the laterally borne, stalked, reniform sporangia are features that suggest affinities with the Zosterophyllophyta.

CHAPTER 8 EARLY LAND PLANTS WITH CONDUCTING TISSUE

TRICHOPHEROPHYTON TEUCHANSII This plant from the Rhynie chert is very rare. The aerial axes are up to 2.5 mm in diameter, branch dichotomously and pseudomonopodially, and exhibit circinate vernation. They are covered with small, unicellular spinelike projections. The central strand has exarch maturation with annular–helical secondary wall thickenings on the conducting elements (Lyon and Edwards, 1991) (FIG. 8.46). Sporangia are reniform and attached laterally to the axes by a small stalk, characters which identify T. teuchansii as a zosterophyll. The sporangia also bore unicellular spines and produced smooth, trilete spores. VENTARURA LYONII To date, this taxon is known only from the Windyfield chert site, which is just a short distance from the original Rhynie chert locality, and of the same age (Fayer and Trewin, 2004). The naked aerial axes of V. lyonii are larger, up to 7.2 mm in diameter, branch dichotomously, and contain a conducting strand which consists of a cluster of thin-walled cells with uneven bands like those seen in G-type cells (Powell et al., 2000); tips of axes are circinately coiled. The middle cortex contains a zone of thick-walled cells interpreted as sclerenchymatous. Like Trichopherophyton, sporangia are reniform and attached laterally to the axes by small stalks. They appear in a row along the axis, suggesting that perhaps they were produced in a loose

Figure 8.46

Geoffrey Lyon. (Courtesy N. Trewin.)

241

spore-producing structure, such as a strobilus. The wall of the sporangium is multilayered and trilete spores 75 um in diameter were produced by this homosporous plant. The shape and position of the sporangia and the coiled apices suggest that V. lyonii is a zosterophyll. Fayers and Trewin (2004) describe Ventarura growing in association with Asteroxylon and Nothia and suggest that this plant community colonized organic-rich, sandy substrates. To date nothing is known about the gametophyte generation of either Trichopherophyton or Ventarura. GAMETOPHYTE GENERATION

In this chapter, in fact throughout this book, the primary focus is on the sporophyte generation of vascular plants. This is not by choice, but rather reflects the general paucity of information about the gametophyte phase in most fossil plants. Although there is some information about the gametophytes in certain seed plants (Millay and Eggert, 1974; Chapter 14), information about fossil vascular plant gametophytes is generally rare. Until recently this has been especially true in discussions of early land plants and has resulted in a variety of theories to explain both the origin of the dominant sporophyte phase and the absence of gametophytes in the fossil record (Chapter 6). Some of these ideas have influenced the interpretation of the Rhynie chert sporophytes. One of the early theories suggested that Aglaophyton major represented the sporophyte and Rhynia gwynne-vaughanii, the free-living gametophyte phase of a single plant (Pant, 1962). The idea gained support from the homologous theory of the alternation of generations (Chapter 6), since there were no specimens of R. gwynne-vaughanii known to possess sporangia at that time. This theory has been abandoned since R. gwynne-vaughanii was subsequently shown to possess terminal sporangia (Edwards, 1980) and A. major is now known to be structurally somewhat different from a vascular plant. Another theory suggested that the small protrusions near the base of R. gwynne-vaughanii axes represented archegonia (Lemoigne, 1968). According to this hypothesis, the rhizomatous portion of the plant represented the gametophyte phase and the aerial upright axes bearing sporangia constituted the sporophyte. Such a situation might be viewed as evidence for the antithetic theory and parallels the life history in modern bryophytes, where the sporophyte phase is permanently parasitic on the gametophyte. Although this hypothesis cannot be totally discounted, to date there is little structural evidence in the form of a transitional zone or foot-like structure, which should be present at the base of all R. gwynne-vaughanii plants. One fossil that was thought to support the hypothesis of a bryophyte-like gametophyte is Horneophyton lignieri.

242

Paleobotany: the biology and evolution of fossil plants

Several features of this genus are similar to those found in bryophytes, especially the columellate sporangia and cormlike base. Some suggested that the relatively simple modern hornwort Anthoceros could represent an early morphological type that would lead to the more complex sporophyte of H. lignieri. It is important to understand that these ideas were advanced at a time when many hypothesized that the bryophytes represented an intermediate stage in the evolution of vascular plants from the green algae. Today the hornworts and bryophytes are mostly considered to be paraphyletic (Chapter 5). There is continuing disagreement, however, as to whether the liverworts or the hornworts represent the basal group among the bryophytes (Nickrent et al., 2000; Friedman et al., 2004; Shaw and Renzaglia, 2004). We are still a long way from understanding the evolution of the sporophyte of vascular plants, as well as the nature of the plant body in a number of fossils that exhibit a bryophytic level of organization. Some exceptionally well-preserved fossils that do contribute to a better understanding of the sporophyte–gametophyte relationship in early land plants have been described by the Remys (FIG. 8.47) from the Rhynie chert (Remy and Remy, 1980a, b; Remy et al., 1993). One of these is Lyonophyton rhyniensis (FIG. 8.48), a small axis that terminates in a shallow, bowl-shaped structure (5.0 mm in diameter), described as a gametangiophore (gametophyte) (Remy and Hass, 1991a). Antheridia are distributed over the entire upper surface. Each antheridium is characterized by a central mass of sterile tissue (columella), and many contain exceptionally well-preserved, coiled spermatozoids (FIG. 8.49). Stomata (FIG. 8.50) occur on the lower surface of the gametangiophore bowl. In the center of the stalk are dark, elongated conducting cells which do not extend into the bowl itself.

Figure 8.47 Renate and Winfried Remy. (Courtesy D. Remy.)

Figure 8.48 Longitudinal section of Lyonophyton rhynien-

sis antheridiophore showing central conducting strand (arrow) and antheridia in cup-shaped distal end (Devonian). Bar  2 mm. (Courtesy H. Kerp.)

The similarity of epidermal cells and conducting elements has been used to demonstrate that Lyonophyton rhyniensis is the gametophyte of Aglaophyton major. Kidstonophyton discoides (FIG. 8.51) is another gametophyte discovered in the Rhynie chert that resembles Lyonophyton (Remy and Hass, 1991b). It consists of a stalk terminating in a shallow, cup-shaped structure (antheridiophore) (FIG. 8.52) that contains numerous antheridia interspersed among ridges of sterile tissue. Spermatozoids have also been described. The sporophyte in the Kidstonophyton life cycle is Nothia aphylla. Associated with these two taxa in the Rhynie chert is another gametophyte that Remy and Hass (1991c) named Langiophyton mackiei (Fig. 8.32). This plant consists of upright axes 6.0 mm long; each terminates in a flattened, peltate structure from which arise numerous (about 30) archegonia. The upper surface of the peltate disk is irregular, and each archegonium is characterized by an elongated neck and deep-seated venter. Antheridiophores are bowl-shaped

CHAPTER 8 EARLY LAND PLANTS WITH CONDUCTING TISSUE

Figure 8.49 Section of antheridium showing coiled sperm (Devonian). Bar  30 μm. (Courtesy H. Kerp.)

and contain up to 50 antheridia. Langiophyton mackiei is the gametophyte of Horneophyton lignieri. The most recent gametophyte described from the Rhynie chert is Remyophyton delicatum (FIGS. 8.53, 8.54) (Kerp et al., 2004). Unlike the other Rhynie chert gametophytes, those of R. delicatum are attached to rhizoid-bearing protocorms (Fig. 8.38). The upright axes possess S-type conducting elements like those of the sporophyte R. gwynne-vaughanii. Gametangiophores are unisexual with the larger ones bearing archegonia. All of the gametophytes known to date from the Rhynie chert are morphologically quite similar to one another, each consisting of an elongated stalk bearing a flattened or saucer-shaped head that contains antheridia or archegonia (FIG. 8.55). These gametophytes appear similar in organization to those found in some members of the hepatic order Marchantiales (Chapter 5), which produce upright antheridiophores and archegoniophores from a prostrate thallus. In the Rhynie chert gametophytes, at least one taxon, Remyophyton delicatum, appears to arise from a protocorm-like structure, not a thallus. All possess stomata and conducting elements, and some have been demonstrated

243

Figure 8.50 Paradermal section of Lyonophyton rhynien-

sis gametangiophore showing stoma (Devonian). Bar  80 μm. (Courtesy H. Kerp.)

8.51 Diagrammatic section of distal end of Kidstonophyton discoides antheridiophore (Devonian). (From Taylor and Taylor, 1993.)

Figure

244

Paleobotany: the biology and evolution of fossil plants

to be endomycorrhizal (FIG. 8.56) (Taylor et al., 2005c). As more gametophytes are discovered it will be interesting to see if any of the antheridiophores and archegoniophores are found attached to some type of thallus-like structure.

Figure 8.52 Longitudinal section of distal end of Kidstonophyton discoides antheridiophore showing several antheridia (arrows). Compare with FIG. 8.51. (Bar  1 mm).

Morphologically, Lyonophyton, Langiophyton, Kidstonophyton, and Remyophyton show some similarity to the Early Devonian compression genus Sciadophyton (FIG. 8.57), which also consists of radiating axes terminating in cuplike structures (FIG. 8.58) (Remy et al., 1980a, b; Schweitzer, 1980a). A few elongated cells in the center of the axis are interpreted as conducting elements, but because the specimens are known only as compressions, these features remain equivocal. Although initially regarded as a sporophyte, the presence of oval bodies on the inner surface of the cup, similar in size to those in the Rhynie chert gametophytes, are suggestive of gametangia (Kenrick and Crane, 1997a). The presence of Sciadophyton in the same rocks with several morphological forms considered to be developmental stages prompted Schweitzer (1981) to regard Sciadophyton as the gametophyte of Stockmansella langii. Kenrick et al. (1991b), however, based on mineral casts of the conducting elements, interpret Sciadophyton as the gametophyte of either Stockmansella or Huvenia. As suggested by Remy et al. (1992), it is highly probable that the

Figure 8.54 Distal end of Remyophyton delicatum gametangiophore showing archegonium (arrow) (Devonian). Bar  300 μm. (Courtesy H. Kerp.)

Figure 8.53 Several Remyophyton delicatum gametangiophores with antheridia (arrows) (Devonian). Bar  2 mm. (Courtesy H. Kerp.)

Figure 8.55 Diagrammatic representation of a population of Lyonophyton gametangiophores associated with mycorrhizae.

CHAPTER 8 EARLY LAND PLANTS WITH CONDUCTING TISSUE

245

morphotaxon Sciadophyton represents a gametophyte stage of several early land plants. The anatomical and morphological similarity of the gametophytes described from the Rhynie chert to date would appear to strengthen Remy’s hypothesis.

Figure 8.56 Arbuscules (arrows) in gametangiophore of

Lyonophyton rhyniensis (Devonian) Bar  36 μm (Courtesy of H. Kerp.)

Figure 8.57 Suggested reconstruction of Sciadophyton sp.

(Devonian). (From Taylor and Taylor, 1993.)

Figure 8.58 Sciadophyton steinmannii, a gametophyte believed to belong to the sporophyte Zosterophyllum rhenanum (Devonian). Bar  2 cm. (Courtesy BSPG.)

246

Paleobotany: the biology and evolution of fossil plants

Calyculiphyton is an Emsian (Lower Devonian) compression fossil that may also represent a gametophyte (Remy et al., 1991). In this type, elongate axes terminate in cupshaped structures that morphologically look much like those of Sciadophyton. The anatomically preserved gametophytes from the Rhynie chert provide the first real fossil evidence that pertains to the evolution of alternation of generations. The nature of the conducting strand and relationship to the sporophyte phase of Rhynie chert plants substantiate the existence of free-living, terrestrial gametophytes as early as the Early Devonian. The conducting strand of Aglaophyton major and several other early land plants provided the impetus to reevaluate the conducting elements of many early land plants and to use this feature to establish the biological affinities of the sporophyte and gametophyte phase. To date there is little evidence of post-fertilization stages or the new sporophyte (embryo) in any early land plants. Kerp et al. (2004) described and illustrated some tissue extending from an archegonial neck of Remyophyton delicatum that might represent the gametophyte–sporophyte junction and development of the new sporophyte generation of R. gwynne-vaughanii. OTHER RHYNIOPHYTES

As mentioned earlier, Cooksonia, a plant that has been historically regarded as the oldest vascular plant, is included in this group. Compressed specimens have been described from localities all over the world, including North America, Great Britain, North Africa, Europe, Siberia, and South America (e.g., Fig. 8.59). Some Cooksonia specimens discovered in Wales are known from deposits as old as the Ludlovian; other specimens suggest that the taxon extended into the Early Devonian (Emsian). Sporangia thought to belong to Cooksonia have been described from Wenlock (mid-Silurian) rocks (Edwards and Feehan, 1980). Cooksonia hemisphaerica, from the Upper Silurian of Wales consists of dichotomous branches up to 6.5 cm long with axes 1.5 mm wide. Stomata occur along the aerial axes (Edwards, 1979). Sporangia are terminal and vary from hemispherical to spherical; in C. pertoni and C. downtonensis, sporangia are wider and longer. In C. caledonica, the shape of the sporangium is highly variable. None of the specimens shows a distinct dehiscence mechanism and all plants were apparently homosporous. One might suggest that all of the Cooksonia specimens described to date merely represent the distal branches of a much larger plant. The uniform sizes of the many specimens described thus far, however, favor the interpretation of Cooksonia as a small plant. More recently, a larger specimen interpreted as Cooksonia was described

from the early Lochkovian (Upper Devonian) of Brazil which includes five orders of dichotomous branching (Gerrienne et al., 2006). Several hypotheses are offered as to whether the specimen represents a sporophyte, a gametophyte, or a sporophyte arising from a thalloid gametophyte (prothallium). Although Cooksonia is considered by many to represent the oldest vascular plant, the evidence for this assumption is still equivocal. When the genus was originally described by Lang (1937), none of the axes with terminal sporangia in the original specimens displayed conducting elements; the presence of a central strand composed of specialized (perhaps conducting) cells was only observed in isolated axes lacking sporangia. Axes containing terminal sporangia of the Cooksonia type have been found to possess thick-walled cells of a sterome (Edwards et al., 1986), but no vascular elements (Edwards et al., 1992). In addition to axes lacking clearly defined vascular tissue, specimens of Cooksonia have also been reported with a variety of sporangial morphologies (Edwards et al., 2004), variously ornamented spores, axes with and without stomata, and various forms of branching. Thus, it appears that the more we learn about Cooksonia, the more difficult it is to interpret precisely what the genus Cooksonia represents. Edwards and Edwards (1986) place Cooksonia in a group they refer to as rhyniophytoids, whereas Taylor (1988a) terms this group cooksonioids, defined as small plants with terminal sporangia borne on narrow axes that lack true tracheids. Other authors simply include the Cooksonia-like axes in the Eutracheophytes (Kenrick and Crane, 1997a). All of these interpretations underscore that the cooksonioids as presently

Figure 8.59 Several Cooksonia specimens from Brazil (Devonian). Bar  5 mm. (Courtesy P. Gerrienne.)

CHAPTER 8 EARLY LAND PLANTS WITH CONDUCTING TISSUE

understood represent a highly artificial group of plants that existed during the Late Silurian–Early Devonian, and that may include forms that are ancestral to either bryophytes or vascular plants, or possibly both. Uskiella (FIG. 8.60) is used for both permineralized and compressed cooksonioids from the Lower Devonian of southern Wales. They have naked, simple isotomous branching and ellipsoidal sporangia (Shute and Edwards, 1989). The sporangial wall has several cell layers thick, with a longitudinal row of thin-walled cells along which the sporangium splits (Fanning et al., 1992). The spores have been described as alete and range from 28–42 μm in diameter; they possess a two-layered sporoderm.

Figure 8.60 Suggested reconstruction of Uskiella spargens (Devonian). (From Taylor and Taylor, 1993.)

247

Another genus that morphologically resembles Cooksonia and Uskiella is Dutoitea (Rayner, 1988). Several species have been described from Lower Devonian (Lochkovian?) compressions of Cape Province, South Africa. Some axes show a thin median line that may represent a conducting strand, perhaps even that of a bryophytic level of organization. In at least one specimen, multicellular spines are present on the axes. Nothing is known about the spores. In the cladistic analysis of Kenrick and Crane (1997a), Hsüa (FIG. 8.61), from the Middle Devonian of Yunnan, China, is included among the zosterophyllophytes. This simple plant consists of main axes 1.0 cm wide that divide to produce lateral branches, some of which terminate in reniform sporangia (C.-S. Li, 1982). Stomata occur along the stems in H. robusta together with tubercles. In the center of each axis in H. deflexa (FIG. 8.62) is a terete, centrarch protostele with G-type conducting cells (D.-M. Wang et al., 2003a). Spores range from 18–36 μm and are trilete. In H. deflexa from South China, spines are present along the axes (D.-M. Wang et al., 2003b). Although features of Hsüa suggest affinities with the cooksonioids, the pattern of branching and features of the sporangia are strikingly similar to that found in some zosterophylls. Another plant at the cooksonioid level of organization is the Early Silurian genus Steganotheca (Edwards, 1970a). Specimens preserved as compressions are 5.0 cm tall and contain several orders of branching, each bearing a terminal sporangium. Although no conducting elements have been identified from the specimens, each axis contains a centrally located striation that may represent some form of conducting strand or

Figure 8.61 Suggested reconstruction of Hsüa robusta (Devonian). (From Kenrick and Crane, 1997a.)

248

Paleobotany: the biology and evolution of fossil plants

Figure 8.62 Suggested reconstruction of Hsüa deflexa (Devonian). (From D.-M. Wang et al., 2003b.)

sterome. Sporangia are elongated (2.5 mm) and the surface is striated. Nothing is known about the sporangial contents. Hedeia corymbosa consists of dichotomizing axes, each terminated by an elongated sporangium (Cookson, 1935). The specimens, which are Early Devonian in age, are not known in sufficient detail, so features of the sporangium and spores have not been described. Specimens of Yarravia also consist of the distal ends of dichotomizing branches, and may represent a preservational state of Hedeia (Hueber, 1983). In Hedeia the fertile axes show some similarity to that found in the trimerophytes. Another cooksonioid is Salopella (Edwards and Richardson, 1974). It consists of compressed, naked, dichotomously branched axes up to 2.0 mm wide which bear terminal sporangia (Larsen et al., 1987). The trilete spores are all of the same morphological type and described as azonate. At the present time the spores are sufficiently different to allow Salopella to be distinguished from other rhyniophyte taxa (Edwards and Fanning, 1985). Eogaspesiea gracilis is the name given to tufted, dichotomously branched axes up to 10.0 cm long (Daber, 1960a). This Early Devonian taxon includes a tangled mass of axes believed to have been attached to a rhizome. At the end of some of the axes are elongate sporangia, each up to 2.5 mm long and containing thin-walled, perhaps alete spores. Not all of the early land plants had cylindrical axes. Some taxa, such as Taeniocrada, which ranges from the Lower to the Upper Devonian, included plants with flattened axes and dichotomous branching (FIG. 8.63). Sporangia are

Figure 8.63 Suggested reconstruction of Taeniocrada deche-

niana based on Kräusel and Weyland, 1930. (Courtesy H. Kerp.)

typically terminal on specialized fertile, repeatedly forked branches and range from 3.0–7.0 mm long; a few sporangia appear in a lateral position (FIG. 8.64). The occurrence of this fossil in dense mats and the apparent lack of stomata on the stem surfaces prompted some to suggest that it may have been aquatic or semi-aquatic. Taeniocrada stilesvillensis (Upper Devonian of New York) consists of axes that branch dichotomously or pseudomonopodially, with hairlike projections arising from ridges along the axes. This taxon may also be distinguished by exarch maturation of the probable G-type conducting elements (Taylor, 1986). The genus Stockmansella has been instituted by Fairon-Demaret (1985, 1986) for forms previously assigned to Taeniocrada

CHAPTER 8 EARLY LAND PLANTS WITH CONDUCTING TISSUE

249

Figure 8.64 Fertile portion of Taeniocrada sp. (Devonian). Bar  4 mm. (Courtesy H. Kerp.)

that bear single sporangia in a lateral position on the axes. Stockmansella remyi from the Eifelian of northwestern Germany consists of a system of prostrate, aboveground axes up to 10 cm long and 3.1 mm wide which repeatedly bifurcate and are characterized by a central xylem strand consisting of S-type cells (Schultka and Hass, 1997). These basal or rhizomatic axes produce narrow laterals at right angles; these may turn upward and become smooth, erect axes. The rhizomatic axes also produce arrested laterals positioned in the axils of bifurcations and irregularly distributed sporangia arising from nonvascularized pads (sporangiophores) that are attached laterally to the main axes. Sporangia are elongate or ovoid, up to 2.2 mm long and 1.0–2.2 mm wide, and dehisce by one to several longitudinal fissures. Prostrate axes, whether main axes or laterals, produce scattered rhizoid-bearing bulges on all sides. Taeniocrada dubia is a Devonian plant that was historically included in the rhyniophytes, but Hueber (1982) has suggested that it contains a central strand composed of tubes

8.65 Suggested reconstruction of Huia gracilis (Devonian). (From Wang and Hao, 2001.)

Figure

of varying diameters. On the inner surface of each tube is a series of helical thickenings that represent a component of the primary wall rather than being secondarily deposited. The wall in these tubes has three parts, consisting of an inner microporate layer, a middle spongy zone that constitutes the majority of the cell wall, including the helical thickenings, and an outer fibrillar layer. Huia gracilis is a permineralized Early Devonian (Yunnan Province, China) plant with oval to reniform sporangia (FIG. 8.65) (Wang and Hao, 2001). The presence of G-type conducting elements in a centrarch strand, K- and

250

Paleobotany: the biology and evolution of fossil plants

Figure 8.66 Sporangia of Huvenia kleui showing twisted organization of the wall. Sporangium at right shows attachment region (Devonian). (From Kenrick and Crane, 1997a.)

H-type branching, and sporangia borne on elongate stalks are features found among more than a single group of early land plants. Whether H. gracilis represents a plant which is transitional between the rhyniophytes and zosterophyllophytes or is intermediate in a line leading to the trimerophytes is yet to be determined. The Early Devonian (Pragian) plant Huvenia also has what are interpreted as flattened axes (Hass and Remy, 1991; Schultka, 1991). The axes often bifurcate and bear small protrusions interpreted as rhizophores. The conducting stand contains S-type elements. Twisted fusiform sporangia (FIG. 8.66) are attached to stout sporangiophores and borne on the primary axes near branches. In a specimen of Huvenia sp. from Gaspé, Hotton et al. (2001) describe sporangia attached to small branches. Small disks associated with the specimen are interpreted as vegetative reproductive units. It is unknown whether the twisted configuration of the sporangium reflects a postmortem change, the actual morphology, or structural variation perhaps related to periodic drying. Spores of Huvenia from Gaspé are compared to the sporae dispersae taxa Retusotriletes or Calamospora (Hotton et al., 2001). One plant that combines features of two early vascular plant groups is Renalia (FIG. 8.67) (Gensel, 1976). The characteristics of Renalia further underscore the inherent problems in classifying many of the Devonian and pre-Devonian land plants. Specimens of R. hueberi occur as compressions in the Battery Point Formation from the famous Gaspé region of Québec (Lower Devonian). It is estimated that the plant was up

Figure 8.67 Suggested reconstruction of Renalia hueberi (Devonian). (From Taylor and Taylor, 1993.)

to 30.0 cm tall, and consisted of a pseudomonopodial main axis to which were attached lateral branches dichotomizing several times and terminating in reniform sporangia. The spores measure 46–70 μm in diameter and are trilete. Although a few helical–scalariform conducting elements have been recovered in macerates, virtually nothing is known about the conducting system. Terminal sporangia and pseudomonopodial branching are features that suggest affinities with the rhyniophytes, but the large, reniform-shaped sporangia (with dehiscence along the distal margin) are characteristics common to members of the zosterophyllophytes. Pinnatiramosus quianensis was originally described from the Lower Silurian (Llandovery) of China as the oldest vascular

CHAPTER 8 EARLY LAND PLANTS WITH CONDUCTING TISSUE

251

the endogenous origin of the laterals (Edwards et al., 2007). Detailed analysis of the matrix suggests that although the specimens are fossil, they probably represent roots of a geologically younger (?Permian) extinct plant that grew into the Silurian rocks below and subsequently became fossilized. DISCUSSION: RHYNIOPHYTE EVOLUTION

Figure 8.68 Pinnatiramosus Bar  1 cm. (Courtesy D. Edwards.)

quianensis

(Silurian).

Figure 8.69 Detail of Pinnatiramosus quianensis show-

ing interdigitating laterals (Silurian). Bar  5 mm. (Courtesy D. Edwards.)

plant (Geng, 1986; Cai et al., 1996). The plant includes naked, compressed branching systems, some up to 40 cm long, that produced closely spaced, interdigitating laterals (FIGS. 8.68, 8.69). Cells macerated from the main axis show tracheid lumen casts with circular-bordered pits. A more recent reexamination of this fossil within its depositional environment concludes that it represents a rooting system, rather than aerial axes, based on the two dimensionality of the branching systems and

The rhyniophytes have traditionally occupied the position of the oldest and simplest vascular plants. Many authors, including Banks (1975), have suggested that the rhyniophytes gave rise to the trimerophytes. As new information becomes available, however, it is clear that some of the so-called rhyniophytes are not true vascular plants, but share features of both bryophytes and vascular plants; these plants have been accommodated in an artificial group, the cooksonioids, until we know more about some of these early forms. From the foregoing section, it should be apparent that although these Late Silurian–Devonian plants share many features, they include so many plesiomorphic characters that it is difficult to include them in higher taxonomic categories. In addition, the information on conducting elements from exceptionally well-preserved specimens like Aglaophyton major suggests that some of our concepts about early land plants with conducting tissues may be in need of modification. Now that gametophytes are known in at least some of the rhyniophytes, their structure and morphology may help to define broader whole-plant concepts leading to a more robust classification of many of these plants. Even in specimens lacking anatomy, some morphological characters suggest a bryophytic level of organization; it is important to keep in mind, however, that tissue systems, like various organs (e.g., leaves and roots) were evolving as plants adapted to a new, terrestrial environment. The fossil record suggests that conducting elements may have evolved in more than one group and at different times during the early colonization of the earth (Taylor, 1986; Kenrick et al., 1991a). Not all of these elements were true tracheids, but they apparently functioned like tracheids in that they not only had to provide support (mechanical stability), but also function in translocation, an idea suggested many years ago by F. O. Bower and later by H. P. Banks. Some of these plants may represent the ancestral stock of certain bryophyte lineages and others may be true vascular plants, but the majority may simply represent failed attempts in the colonization of the land. As additional specimens are discovered and new information evaluated, there are certain to be additional modifications and refinements of our understanding of these plants. Rhyniophytes are certainly some of the simplest, upright land plants, but are they the oldest? As you will see later, zosterophyll and even lycopsid megafossils are now known from the Upper Silurian in some diversity. At the

252

Paleobotany: the biology and evolution of fossil plants

present time, the oldest megafossil evidence demonstrates parallel evolution of rhyniophytes and zosterophylls; unfortunately, the spore record currently provides few characters that would enable us to distinguish these early plants from one another, as all contain simple trilete spores. In addition, few are known with spores in situ, which would provide for correlation of sporae dispersae with parent plants and demonstrate plant diversity prior to the Late Silurian.

(A)

(B)

(C)

(D)

(E)

(F)

ZOSTEROPHYLLOPHYTES The zosterophyllophytes (or zosterophylls) are the second major group of vascular plants established by Banks (1975), as the Zosterophyllophytina. They range from Late Silurian to Late Devonian and represent some of the most interesting early vascular plants, in part because there is a considerable amount of anatomical and morphological detail known about them. They demonstrate diversity as early at the Ludlovian (Late Silurian) (Kotyk et al., 2002) and were widespread geographically by the Early Devonian, including Gondwana (Zhu and Kenrick, 1999). As a group they share many features with the Lycophyta and may have given rise to the lycopsids; others interpret them as paraphyletic (Crane, 1990), or as the sister group to the lycopsids (Gensel, 1992). Most zosterophylls exhibit dichotomous branching, although in some genera there is a tendency toward a pseudomonopodial habit; the branching is generally planar (Kenrick and Crane, 1997a) and ultimate branches exhibit circinate vernation (development by means of uncoiling). When specimens are found permineralized, the exarch protostele is more robust than in the rhyniophytes and often elliptical in transverse section. The synapomorphy that distinguishes zosterophyllophytes from other early vascular plants is the presence of sporangia that are borne laterally along the stem (FIG. 8.70); they may either be sessile or attached by short branches. In many taxa, sporangia are aggregated often into terminal clusters or cone-like structures. Sporangial shape varies from globose to reniform, with dehiscence typically occurring along the distal edge and separating the sporangium into two valves. Most zosterophyll sporangia also exhibit a thickened zone bordering the dehiscence line. All zosterophyllophytes are homosporous, although the size range of the spores can be rather extensive in some Devonian forms. In other plants that possess zosterophyllophyte features (e.g., Barinophyton and Protobarinophyton), each sporangium contains both large and small spores (Taylor and Brauer, 1983; Chapter 9). One plant that has an interesting taxonomic history and is now included among the zosterophyllophytes is Sawdonia. As mentioned earlier, Dawson described a plant from the

(H)

(G)

(I)

(J)

Figure 8.70 Morphology and arrangement of zosterophyll sporangia. A,H. Gosslingia breconensis; B,I. Zosterophyllum myretonianum; C, J. Zosterophyllum fertile; D. Zosterophyllum spectabile; E. Sawdonia acanthotheca; F. Konioria andrychoviensis; G. Psilophyton princeps (Devonian). (From Kenrick and Edwards, 1988b.)

Devonian of the Gaspé Peninsula in 1859 as Psilophyton princeps. In his initial reconstruction, Dawson illustrated a dichotomizing plant that terminated in straight or circinately curved branches. He later modified his interpretation of P. princeps and described the reproductive units as sporangia borne at the ends of slender branches (Dawson, 1870). This was followed a year later by an emended diagnosis based on additional specimens, some naked and some with helically arranged spines (Dawson, 1871). At this time, the spiny fossils were referred to the taxon P. princeps var. ornatum. Subsequent workers were able to identify globose sporangia in lateral positions on the axes of some specimens of P. princeps var. ornatum. The literature at the time now contained a contradiction relative to Psilophyton. The P. princeps plant originally described by Dawson consisted of spiny axes with terminal sporangia, and P. princeps variety ornatum, which was regarded as the type of the species, represented a very different plant with lateral sporangia. Today we know that Dawson combined two

CHAPTER 8 EARLY LAND PLANTS WITH CONDUCTING TISSUE

253

Deheubarthia splendens axis with axillary tubercle branch (Devonian). (From Kenrick and Crane, 1997a.) Figure 8.72

Figure 8.71 Suggested reconstruction of Deheubarthia splend-

ens (Devonian). (From Kenrick and Crane. 1997a.)

separate plants in his early reconstruction. Psilophyton princeps is used for plants with naked or spiny axes and terminal sporangia. Those specimens with spiny axes and lateral, globose sporangia (Fig. 8.70E) are referred to as Sawdonia ornata (Hueber, 1971). A detailed account of the historical development of the problem and its solution can be found in Banks et al. (1975). It may be appropriate that Dawson’s original description and reconstruction of Psilophyton, which was ignored for so many years, involved numerous researchers and took more than a century to fully understand! Sawdonia ornata is thought to have been 30.0 cm tall and constructed of pseudomonopodial axes that arose from a rhizome. Lateral axes branch dichotomously and are characterized by circinate tips. Branches are covered by numerous tapered spines, which typically show dark tips (Edwards et al., 1989). The sporangia seem to be confined to the distal ends of branches, where they form loosely aggregated spikes. Individual sporangia are reniform and borne on short

stalks in two vertical rows. Dehiscence occurred along the convex margin dividing the sporangium into two equal valves. Sawdonia ornata apparently was a homosporous plant with sporangia containing round–subtriangular trilete spores, each 64 μm in diameter. In structurally preserved specimens, the stele has a solid core of conducting elements with annular secondary wall thickenings that exhibit interconnecting bars suggestive of G-type cells (Rayner, 1983). Epidermal cells are unusual in S. ornata. Some bear papillae, whereas others consist of a central cell surrounded by elongate radiating cells that may represent some type of hair base (Wang and Hao, 1996). Similar cells have been noted in Oricilla, another Early Devonian zosterophyllophyte (Gensel, 1982a). Their function is not known, but secretion, aeration, and storage functions have been suggested (Gensel and Andrews, 1984). Stomata are also present on the epidermis of the stem of Sawdonia, but not on the spines. Deheubarthia (FIG. 8.71) has been proposed for certain Early Devonian specimens previously placed in Sawdonia (Edwards et al., 1989). Specimens of D. splendens consist of spiny axes 30.0 cm tall organized in a planar, pseudomonopodial branching system. Spines without dark tips and the occurrence of subaxillary branches (FIG. 8.72)

254

Paleobotany: the biology and evolution of fossil plants

distinguish Deheubarthia from Sawdonia. Conducting cells are of the G-type and some epidermal cells possess papillae. Another plant that closely resembles Sawdonia is Discalis longistipa (FIG. 8.73), a zosterophyll from the Lower Devonian of Yunnan, China (Hao, 1989). This plant had H- and K-shaped branching with fertile axes bearing large (3.7 mm in diameter) sporangia organized in loose spikes. Multicellular spines occur on all axes, as well as on the sporangia. Based on the presence of radial, trilete spores 30–50 μm in diameter, D. longistipa is thought to have been homosporous. Kidston and Lang (1923) described a plant from the Devonian Old Red Sandstone Series that they characterized as having a tufted growth habit with predominantly dichotomous branches. Hicklingia attained a height of 17.0 cm and is known from compressed specimens. Because the

sporangia appeared to have been terminal, Kidston and Lang regarded it as similar to Rhynia, and subsequent authors placed the genus among the rhyniophytes. A reexamination of the original specimen by Edwards (1976) indicates that the sporangia were, in fact, borne laterally on the axes, suggesting assignment to the Zosterophyllophyta. Spores are up to 50 μm in diameter and trilete. The Early Devonian genus Zosterophyllum includes a number of species. Zosterophyllum was a leafless and smooth, dichotomously branched plant that produced lateral sporangia on short, delicate stalks. In permineralized specimens of Z. llanoveranum from the Lower Devonian of Britain, the axes are 1.5 mm in diameter and contain an elliptical strand of scalariform conducting elements (Edwards, 1969). Surrounding the stele is a cortex of three zones distinguished by differences in cell size and cell-wall thickness. Sporangia

Figure 8.73 Suggested reconstruction of Discalis longistipa. Inset shows portion of fertile axis with spines. (From Taylor and Taylor, 1993.)

CHAPTER 8 EARLY LAND PLANTS WITH CONDUCTING TISSUE

255

occur in either one or two rows at the distal ends of branches; they vary from circular to reniform (FIG. 8.74), and each is borne on a small stalk that departs from the axis at an abrupt angle. Vascular tissue has not been identified in the sporangial stalk. On either side of the distal line of dehiscence is a band of elongated, thick-walled cells, grading proximally into smaller, thinner-walled cells proximally. Spores are ovoid and average 45 μm in diameter. In Z. ramosum the fertile axes branch many times, each terminating in a spike of 8–15 sporangia; trilete spores are triangular with long laesurae (Hao and Wang, 2000). Zosterophyll features are present in Macivera gracilis from the Upper Silurian (Ludlovian) of Bathurst Island, Canadian Arctic (Kotyk et al., 2002). In this leafless plant the sporangia are sessile, borne in small clusters at the distal ends of axes, and are not in rows. Hueber (1972) suggested that Zosterophyllum subgenus Platyzosterophyllum be used to accommodate species with sporangia in two rows (thus forming dorsiventral spikes) and the subgenus Zosterophyllum be reserved for species with helically arranged sporangia. In several species, such as Z. myretonianum (FIG. 8.70B,I, 8.75) and Z. divaricatum

(FIG. 8.76), the basal region of the plant is characterized by K- and H-type branching patterns. These branching patterns are apparently the result of successive, close order dichotomies. Branching of this type has been described in a number of zosterophylls and is also known to have occurred in some species of the drepanophycalean genus Drepanophycus (Chapter 9). In Z. divaricatum the fertile axes have circinately coiled apices, and the reniform sporangia contain smooth, trilete spores 50–90 μm in diameter (Gensel, 1982b). In Z. deciduum, a compression form from the Lower Devonian (Emsian) of southern Belgium (Gerrienne, 1988), sporangia appear to have been shed at maturity. Recurved branch tips are also present on the Early Devonian (Gedinnian of Germany) Anisophyton (Remy et al., 1986). Spines with truncated apices cover the axes and the sporangia are typically borne on only one side of the axis. Another Early Devonian plant that has features of the zosterophylls is Guangnania cuneata (FIG. 8.77) (D.-M. Wang and Hao, 2002; D.-M. Wang et al., 2002) from Yunnan Province in southwestern China. Elongate sporangia are borne on upright stalks, are slightly curved, and dehisce into two, unequal valves.

Figure 8.74 Structure and morphology of Zosterophyllum

Figure 8.75 Suggested reconstruction of Zosterophyllum myretonianum (Devonian). (From Taylor and Taylor, 1993.)

llanoveranum sporangia (Devonian). (From Edwards, 1969.)

256

Paleobotany: the biology and evolution of fossil plants

Figure 8.76 Suggested reconstruction of Zosterophyllum divaricatum (Devonian). (From Kenrick and Crane, 1997a.)

Lateral stalked sporangia are also found in Danziella artesiana from the Lower Devonian of France (Edwards, 2006). In Serrulacaulis furcatus (FIG. 8.78) (Upper Devonian), axes are ornamented by two rows of opposite emergences (Hueber and Banks, 1979). The triangular shape of the projections gives the axes a sawtooth appearance. These are now interpreted as triangular, prism-shaped structures that are arranged in two rows in a steplike manner on opposite sides of the axes (Berry and Edwards, 1994). Some axes bear stalked sporangia between the rows of projections on one side of the axis. Sporangia are reniform and borne on short stalks; dehiscence splits the sporangia into two unequal halves. Spores are trilete and 60 μm in diameter. Conducting cells are of the G-type (Berry and Edwards, 1994). Gosslingia breconensis is a zosterophyll that was 50.0 cm tall and is known from the Lower Devonian

Figure 8.77 Suggested reconstruction of Guangnania cuneata

fertile axis. (From D.-M. Wang and Hao, 2002.)

of Wales (Edwards, 1970b). It consists of dichotomizing axes which are up to 4.0 mm wide distally. The base is presumed to have consisted of a rhizome bearing rhizoids, although organic attachment has not been demonstrated. Like Sawdonia, Gosslingia exhibits distal tips that are circinately coiled. The aerial stems are leafless, although some specimens show small protuberances that extend a few hundred micrometers from the surface and larger tubercles

CHAPTER 8 EARLY LAND PLANTS WITH CONDUCTING TISSUE

257

Figure 8.78 Suggested reconstruction of Serrulacaulis furca-

tus (Devonian). (From Taylor and Taylor, 1993.) Figure 8.79 Suggested reconstruction of Thrinkophyton for-

mosum (Devonian). (From Kenrick and Crane, 1997a.)

that are often identified only by their scars. These latter structures, which have been termed axillary tubercles, contain conducting elements in the form of a terete trace and extend from one side of the stem just below the point of a dichotomy. Pyritized axes reveal that the conducting strand is elliptical in outline, with annular protoxylem elements located near the periphery. Conducting elements are of the G-type. The cortex is constructed of thick-walled cells and what appear to be stomata are scattered along the stem. In iron sulfide permineralizations of G. breconensis, all the conducting elements are believed to be metaxylem tracheids, the protoxylem being either lost during fossilization or at some stage in the ontogeny of the stele (Kenrick and Edwards, 1988a). The presence of pyrite in various forms is discussed in relation to the loss of cellulose and lignin during fossilization and the formation of artifacts in the cells. This type of detailed study represents an important starting point in the accurate characterization of the vascular elements and supporting cells in all early land

plants. The sporangia in Gosslingia occur in definite aggregations in the distal regions of the plant. They are variable in shape, ranging from globose to reniform (Fig. 8.70A,H), and are attached to the stems by slender stalks. Little detailed information is known about the histology of the sporangial wall, although some spores have been recovered. These range from 36–50 μm in diameter and are ornamented by small spines. Tarella is an interesting zosterophyll in which the stalked, reniform sporangia split into two equal valves much like those of Gosslingia (Edwards and Kenrick, 1986). The sporangia occur in opposite rows and are not present near the circinate stem tips. Projections, some with hooked apices, are present on both the fertile and sterile axes. Spores are trilete and 40 μm in diameter. Another zosterophyll from the Lower Devonian of Wales is Thrinkophyton formosum (FIG. 8.79) (Kenrick and Edwards, 1988b). It has pseudomonopodial and isotomous branching and is 9.0 cm tall. The stem tips are circinate with projections borne just below a branch dichotomy. Sporangia

258

Paleobotany: the biology and evolution of fossil plants

are arranged in one or two rows arising from short stalks, and are reniform. The central strand is described as consisting of xylem elements with annular to helical thickenings; xylem maturation is exarch. Konioria is an Early Devonian plant from Poland that has branch tips curved to form hooklike ends (Zdebska, 1982). On the surface of the axes are irregularly positioned spines of various lengths, the longest ones (4.0 mm) bearing delicate teeth. The stems also contain winglike outgrowths that are extended longitudinally. In the center of the axis is an exarch protostele. Sporangia were produced on stalks that extend at right angles from the stem just below the last dichotomy. They are typically 2.5 mm wide and covered by minute spines. To date nothing is known about the spores. Crenaticaulis is a zosterophyll that is known in some detail and thus provides important information about the conducting system and sporangia of this group. Crenaticaulis verruculosus was described from both compressed and structurally preserved specimens collected from Lower Devonian rocks in the Gaspé region (Banks and Davis, 1969). The largest specimen consists of a 22.0-cm-long axis that shows both pseudomonopodial and dichotomous branching. The dichotomies occur at short intervals, and the distal stem tips are slightly coiled. An unusual feature of C. verruculosus is the presence of two rows of multicellular, toothlike protuberances that are nearly oppositely arranged along the surface of the stems. These teeth are triangular and present on the circinately coiled stem apices as well as on the stalks that bear the sporangia. Epidermal cells of the stem are of two types, elongate and papillate. On some specimens, subaxillary tubercles are present; on others their position is indicated by an elliptical scar. The exarch strand is elliptical in cross section and composed of G-type conducting elements. The sporangia of C. verruculosus are clustered and occur in opposite–subopposite groups on the distal branches. They are pedicellate and nearly spherical in outline. Sporangial dehiscence has been termed distal—beginning on one side just above the attachment to the stalk and arching over the adaxial face to the opposite surface. Dehisced sporangia consist of a large abaxial and a small adaxial segment; nothing is known about the spores. Rebuchia is an Early Devonian plant from the Beartooth Formation of Wyoming (Hueber, 1972), with naked, dichotomously branched axes that gradually taper into blunt points. Sporangia are confined to distal branches and occur as spikes of up to 20 sporangia (FIG. 8.80). Individual sporangia are arranged oppositely to suboppositely and borne on short, curved stalks, so that all the sporangia point are essentially in the same direction. Dehisced sporangia have equal valves,

Figure 8.80 Suggested reconstruction of Rebuchia ovata

(Devonian). (From Taylor and Taylor, 1993.)

indicating a basipetal form of dehiscence. Rebuchia ovata was probably homosporous; all the spores recovered are unornamented, 68–75 μm in diameter, and of the Retusotriletes type. Another plant that has aggregations of sporangia in what might be termed strobili is Bracteophyton variatum from the Lower Devonian of China (Wang and Hao, 2004). Individual sporangia are adaxial and associated with pairs of bracts, or with a single bract that bifurcates at the tip. Sporangial dehiscence is distal. Axes are isotomous, naked and up to 4.7 mm in diameter. Nothing is known about the anatomy. Kaulangiophyton is an Early Devonian plant consisting of a horizontal branching system that bears numerous, irregularly spaced, short spines (Gensel et al., 1969). The spines are slightly curved and up to 2.0 mm long, with decurrent bases. Sporangia are borne on short stalks along the stem and appear to be interspersed among the spines, giving this

CHAPTER 8 EARLY LAND PLANTS WITH CONDUCTING TISSUE

plant the appearance of modern species of Lycopodium. No spores have been recovered, nor is anything known about the vascular system. In many features, K. akantha is similar to Drepanophycus (Chapter 9). ZOSTEROPHYLL EVOLUTION

There is convincing evidence that supports the lycopsids as most closely related to the zosterophylls. Both groups are characterized by exarch protosteles and laterally produced reniform sporangia. In Crenaticaulis and Gosslingia, axillary tubercles are thought to represent rhizophore-like branches, similar to those of extant species of Selaginella. The absence of leaves in the zosterophylls, although an obvious difference between the two groups, can be explained in terms of the transitional status of the group. Although they do not have true leaves, the zosterophylls are known to bear various types of laterals. These range from unicellular to multicellular spines to multicellular teeth; the latter may have functioned to increase the photosynthetic surface of the stems. Generally, these appendages were randomly scattered over the stem surfaces, although in Crenaticaulis the large, toothlike outgrowths are arranged in two rows along the stems. In the putative early lycopsid Asteroxylon, traces extend through the cortex but do not enter the base of the enations. This tendency toward a definite arrangement and vascularization may constitute the initial stages in the subsequent evolution of the microphyll (Chapter 9). The zosterophylls also demonstrate several stages in the evolution and organization of sporangia. These include forms such as Kaulangiophyton, in which sporangia are apparently helically arranged over the stem surface, to intermediate forms such as Gosslingia, in which the helically arranged sporangia are aggregated into definite spikes. A further modification might result in the arrangement seen in Rebuchia, where sporangia are organized into definite spikes, with the individual sporangia dorsiventral in arrangement (present on only one side of the axis). Niklas and Banks (1990) suggested that the zosterophylls can be separated into two groups based on the presence or absence of terminally located fertile units, as well as on the symmetry of the sporangial aggregation. These two groups can also be distinguished based on the presence of small flaps of tissue along the stems (enations), circinate tips on the axes, and, when preserved, the nature of the conducting strand. Based on their analysis, Niklas and Banks suggest that the lycopsids arose from a zosterophyllophyte-like group of plants, perhaps with a level of organization similar to Asteroxylon or Drepanophycus (Chapter 9). Sporangial dehiscence within the zosterophyllophytes typically divides the sporangium into valves of equal or nearly equal size. As additional zosterophyllophyte taxa are

259

described, a new perspective is emerging as to the diversity and spatial distribution of these interesting Paleozoic plants (Hao and Gensel, 2001; Kotyk et al., 2002; Wang and Hao, 2002). New taxa are providing an increasing data set of sporangial characters relating to dehiscence (e.g., in Crenaticaulis), where dehiscence results in two unequal valves. In addition, the discovery of new taxa is expanding our understanding of the diversity of sporangial aggregation, position and length of stalks, and pattern and orientation on the fertile axes. These characters may become more important as additional information on features of the fertile parts of these interesting plants become more fully known, and will no doubt result in an increased level of resolution relating to the phylogenetic position of these plants.

TRIMEROPHYTES The third major group that was culled from the original Psilophytales by Banks is the Trimerophytophyta. Trimerophytes1 were generally more complex than either the rhyniophytes (from which they are thought to have descended) or the zosterophyllophytes. Trimerophytes demonstrate monopodial branching of the main axes, with lateral axes showing either dichotomous or trifurcate branching. As with rhyniophytes, the sporangia are terminal, although typically they are fusiform to elongate and aggregated or clustered on fertile branches. Internally, the members of this group also exhibit greater complexity, in the form of a relatively large (compared to stem diameter), centrarch stele. Tracheid wall patterns vary from scalariform-bordered to circular-bordered pitting. Trimerophytes are considered to have given rise to all the other vascular plants, with the exception of the lycopsids. As indicated in the preceding section, Psilophyton specimens with lateral sporangia were transferred to the zosterophyllophyte genus Sawdonia. The remaining Psilophyton specimens, that is, those with terminal sporangia, are retained in the Trimerophytophyta. Psilophyton dawsonii is one of the most completely known members of the group. Both compressed and structurally preserved specimens have been described from several Lower Devonian localities (Banks et al., 1975). A reconstruction of P. dawsonii shows a highly branched plant in which the fertile lateral branches are borne alternately and distichously

1

Based on the scientific name of this group, the common name should probably be the trimerophytophytes. However, for the sake of simplicity, we have chosen to use trimerophytes.

260

Paleobotany: the biology and evolution of fossil plants

(in two opposite, vertical rows) along the main axes. Vegetative branches are naked and dichotomize at right angles, terminating in slender, blunt tips. Fertile branches typically branch up to six times before terminating in clusters of 32 sporangia. Individual sporangia are 5.0 mm long, with dehiscence occurring along the lateral surface. The trilete spores vary from 40 to 75 μm in diameter with a smooth spore wall, although in some specimens, a thin layer often appears separated from the spore body. Fine structural details of the spores have been provided by Edwards et al. (1996). If found as dispersed grains, such spores would most closely approximate the sporae dispersae taxa Retusotriletes or Phyllothecotriletes. The conducting tissue of P. dawsonii consists of a centrarch protostele that accounts for approximately one-quarter of the stem diameter. In more basal regions of the plant, the strand has several enlarged protoxylem zones and numerous, radially aligned conducting elements of the P-type (Kenrick and Crane, 1997a). They have bordered pits with secondary wall material that extends across the pit apertures (Hartman and Banks, 1980). Traces supplying the fertile branches are initially terete, becoming more rectangular in outline at higher levels. Surrounding the central strand is a multilayered cortex consisting of collenchymalike cells and, toward the periphery, substomatal chambers. In P. princeps, the stem surfaces are ornamented by a number of cup-tipped spines. The larger size of the sporangia and the nature of the vegetative branching further distinguish this species from P. dawsonii. The conducting strand of P. princeps consists of a solid strand of elements that is either mesarch or centrarch. Psilophyton crenulatum (FIG. 8.81) a species from the Lower–Middle Devonian of New Brunswick, Canada, has spiny branches that dichotomize several times and terminate in slender, recurved tips. On the surface of the axes are multicellular spines up to 6.0 mm long; some have forked tips, whereas others are trifurcate. The vascular strand is described as centrarch and constructed of elements with scalariform–circular bordered pits. The fertile units (clusters of sporangia) in P. crenulatum (FIG. 8.81) (Doran, 1980) are alternate and distichous, or may be helically arranged. In the distal regions they are covered by semicircular crenulations. The fusiform sporangia are up to 5.0 mm long and twisted; spores range from 48 to 102 μm in diameter and conform to the genus Apiculiretusispora. Short recurved lateral branches and elongate pairs of sporangia that are of unequal length are characteristic of Aarabia, an early Emsian (Early Devonian) plant from Morocco (Meyer-Berthaud and Gerrienne, 2001). Psilophyton forbesii was one of the largest species of Psilophyton, estimated as 60.0 cm tall (FIG. 8.82) (Gensel, 1979). The growth habit was either monopodial or

Figure 8.81 Psilophyton crenulatum. Bar  2 mm. (Courtesy

J. B. Doran.)

pseudomonopodial and the naked stems were marked by longitudinal striations. Ellipsoidal sporangia were produced in pairs on fertile lateral branches and spores ranged 53–96 μm in diameter. In P. dapsile (FIG. 8.83) (Kasper et al., 1974), known from the Middle Devonian of Maine, the axes measure 2.0 mm in width and the erect plant is thought to have been about 30.0 cm tall. The stems are smooth and branched dichotomously. Sporangia are small (2.0 mm long) and borne in dense clusters at the ends of closely spaced, distal dichotomies. One species of Psilophyton, P. hedei, which was described from Silurian rocks of Sweden, is now thought to represent some type of invertebrate similar to a graptolite or a pterobranch. In this instance, the branched axis may represent a colony of numerous individuals that share a system of internal tubes and the characteristic stem spines probably represent free zooidal tubes (Lundblad, 1972).

CHAPTER 8 EARLY LAND PLANTS WITH CONDUCTING TISSUE

261

Figure 8.83 Suggested reconstruction of Psilophyton dapsile

(Devonian). (From Taylor and Taylor, 1993.)

Figure 8.82 Suggested reconstruction of Psilophyton forbesii (Devonian). (From Taylor and Taylor, 1993.)

Pertica quadrifaria (FIG. 8.84) is another Devonian plant that can be assigned to the trimerophytes on the basis of its branching pattern and the nature of its fertile region (Kasper and Andrews, 1972). The genus is known from compression remains and has been described from at least two localities of Early Devonian age. The size of some specimens suggests that the plant may have exceeded 1.0 m in height. Numerous, dichotomously branched laterals are borne on the main axis in a tetrastichous pattern (i.e., in four rows). The distal ends repeatedly dichotomize at right angles to each other. Like all trimerophytes, the lateral branches were either completely sterile or completely fertile. The fertile ones consist of closely spaced dichotomies that terminate in masses of sporangia borne on short stalks. Sporangia are elliptical in outline and lack any histologic evidence of a dehiscence mechanism. Poorly preserved spores that are about 64 μm in diameter suggest that P. quadrifaria was homosporous. Pertica varia was much larger than P. quadrifaria, reaching a height of nearly 3.0 m (Granoff et al., 1976). The primary

branches of P. varia are arranged in subopposite pairs (with the successive laterals decussate. Paired sporangia are erect, but the number of sporangia per cluster is smaller than in P. quadrifaria. Spores of P. varia are subcircular–subtriangular and 90 μm in diameter. In P. dalhousii the lateral branches are spirally arranged and divided pseudomonopodially (Doran et al., 1978). Trimerophyton robustius, the type species of the group, was initially described by Dawson (1859) as Psilophyton robustius from some fragmentary specimens collected from the shore of Gaspé Bay. The generic name Trimerophyton was introduced many years later on the basis of a single specimen, also from the Gaspé (Hopping, 1956). The main stem of T. robustius is 1.0 cm wide and consists of numerous helically arranged, trifurcate, lateral branches. The primary and secondary branching patterns of the laterals are trichotomous, with further subdivisions of the dichotomous type. All axes are smooth, with the exception of some that are longitudinally striated. Ultimate branches bear erect sporangia in clusters of three (FIG. 8.85). It has been suggested that Trimerophyton may represent the distal parts of Pertica (Gensel and Andrews, 1984).

262

Paleobotany: the biology and evolution of fossil plants

Figure 8.85 Suggested reconstruction if Trimerophyton robus-

tius (Devonian). (From Taylor and Taylor, 1993.)

Figure 8.84 Branch of Pertica quadrifaria (Devonian).

Bar  2.0 cm.

Yunia is a Siegenian (Pragian, Lower Devonian) plant with spiny axes that shares some features with Pertica (Hao and Beck, 1991). The axes dichotomize in a cruciate arrangement and each contains a conducting strand with one or two protoxylem strands, depending on the level of the axis. Conducting elements have annual scalariform sculpture patterns. Closely associated sporangia are elongate and bear trilete spores with a relatively smooth sculptural pattern. Dawsonites is a Devonian morphogenus that has been used by some workers for sporangia that are not organically connected to an axis or aggregated into large clusters (Halle, 1916b). The sporangia are fusiform and typically measure 3.0–5.0 mm long. The genus is now restricted to terminal sporangia that are believed to have been produced by Psilophyton. As spores are recovered from additional fertile specimens so that comparisons can be made, many of the Dawsonites species no doubt will be placed in synonymy. Another morphogenus that has been used for naked, dichotomously branched axes of Siluro–Devonian age is Hostinella. Structurally preserved specimens consist of

small fragments up to 3.0 mm wide. The conducting strand is protostelic, centrarch, and composed of tracheids with scalariform to bordered pits. Some specimens of Hostinella are thought to represent an isolated branching fragment of Margophyton (Psilophyton) goldschmidtii, now believed to be a Lower Devonian member of the zosterophylls (Zakharova, 1981). A Late Silurian (Prˇídolí) specimen from New York State includes preserved apices and a rootlike structure (Edwards et al., 2004). TRIMEROPHYTE EVOLUTION

The trimerophytes demonstrate more complex morphology and anatomy than rhyniophytes, their presumed ancestors, although both groups are coeval. In the trimerophytes, plant architecture is monopodial or pseudomonopodial. Laterals are produced in a variety of patterns, including helical (Psilophyton sterile branches), alternately and distichous (Psilophyton fertile branches), tristichous (Trimerophyton), and tetrastichous (Pertica). In Pertica the ultimate branchlets consist of slender, three-dimensional dichotomizing structures. It has been suggested that the planation of these lateral branches would provide the morphologic equivalent of a megaphyllous leaf and that Pertica may be used as a transitional morphotype in the evolution of a frond or a leaf. In another group of Devonian plants, the Aneurophytales

CHAPTER 8 EARLY LAND PLANTS WITH CONDUCTING TISSUE

(progymnosperms), some taxa possess planated laterals, whereas in others the branching systems are more three dimensional (Chapter 12). Trimerophytes also demonstrate various patterns of sporangial attachment. In rhyniophytes, sporangia are terminal at the ends of dichotomizing axes. In some species of Psilophyton the number of sporangia is small, while in others (e.g., P. dapsile) numerous small sporangia are clustered together. One possible transformational series might involve Pertica, with its massive clusters of densely packed sporangia, leading to some Carboniferous ferns, such as some species of Botryopteris. Another line might lead to the progymnosperms through such a plant as Tetraxylopteris. One Middle Devonian plant that might bridge the evolutionary gap between trimerophytes and progymnosperms is Oocampsa (Andrews et al., 1975). Specimens of O. catheta consist of closely spaced, helically arranged branches up to 7.0 cm long that were produced from a primary axis (FIG. 8.86). The lateral branches divide pseudomonopodially and dichotomously and terminate in elongate, erect sporangia. Sporangia dehisce longitudinally and contain

263

large (96–120 μm), trilete miospores. The spores are interesting in that there appears to be some space between the wall layers, suggesting a pseudosaccate morphologic type. Oocampsa, with erect sporangia borne on helically arranged primary and secondary branches, may be transitional between certain trimerophytes, for example, Trimerophyton and Pertica, and Tetraxylopteris, a progymnosperm with pinnate arrangement of ultimate segments. Such a series is congruent with the stratigraphic occurrences of the taxa listed. There are relatively few trimerophytes that are structurally preserved. In Psilophyton dawsonii and P. princeps, there is a simple conducting strand, but one that is more massive than any known for the rhyniophytes. Conducting elements in the rhyniophytes are of the S-type, while Psilophyton contains P-type elements in which secondary wall material is deposited between the scalariform bars. Although the trimerophytes appear to be less diverse than other early land plants, there are some apparent evolutionary trends within the group, including stages in the evolution of a particular type of leaf, modification of conducting element pitting toward the circular-bordered type, and some suggestion of an early stage in the evolution of spores with an increased surface area as a result of the separation of wall layers. As additional specimens are discovered and described from Devonian or possibly even Mississippian rocks, it is obvious that this group will play an increasingly important role in our understanding of levels of specialization in early land plants and their role in the diversification of laterappearing groups.

EARLY LAND PLANT EVOLUTION

Figure 8.86 Suggested reconstruction of Oocampsa catheta (Devonian). (From Taylor and Taylor, 1993.)

During the evolution of land plants, the sporophyte generation has become increasingly more complex, both physiologically and morphologically. The gametophytes from the Rhynie chert demonstrate that during the Early Devonian gametophytes and sporophytes were more similar to each other than these generations are today. Some authors working with extant plant development suggest that the Rhynie chert gametophytes provide evidence of developmental genes being reassembled from the gametophyte phase and incorporated into the developmental pathways of the sporophyte (Floyd and Bowman, 2007). The comparison of developmental mechanisms based on various gene families in bryophytes, lycopsids, and seed plant lineages has already revealed some commonalities (e.g., among the MADS-box genes), as well as differences related to various plant structures. These morphological expressions can then be compared to structures

264

Paleobotany: the biology and evolution of fossil plants

that are present in fossils. This is another way that evidence from the fossil record can be used in understanding the morphological evolution that has led to the complexities in multicellular plants that dominate the Earth today. While cladistic analyses suggest that the land plants are monophyletic based on the presence of multiple sporangia (the polysporangiophytes), the evidence from the fossil record may be interpreted in several ways. The conducting elements in many of these early plants have been interpreted as representing a protracheophyte grade of evolution. The diversity of these elements, however, brings up the question of whether or not conducting elements evolved more than once as a response to the selective pressures of a desiccating environment. Researchers on extant algae are still debating the relationships of the charophycean algae and which group is most closely related to land plants. Many systematists that work only with extant plants believe the fossil record is too

incomplete to provide any information on the origin of land plants, but new discoveries continue to provide more data. Based on the diversity of conducting elements, the fossil record could be interpreted to suggest that land plants (hornworts, bryophytes, and vascular plants) became established in terrestrial environments several times and perhaps originated from a variety of charophycean green algae. The fossil evidence for the lycopsid lineage shows that these plants were clearly differentiated from the other vascular plants (and remarkably modern!) as early as the Late Silurian (Chapter 9). Information on the earliest land plants and the earliest vascular plants continues to be amassed. Regardless of the degree of interrelatedness of these early land plants, how and when the necessary steps took place in the transition from an aquatic to terrestrial habitat will continue to be challenging questions for paleobotanists and neontologists alike.

9 Lycophyta Evolution of the microphyll ...................................267

Lycopodiales.............................................................................310

Drepanophycales ................................................................. 268

Selaginellales.......................................................................... 312

Protolepidodendrales ................................................... 271

Pleuromeiales ..........................................................................316

Lepidodendrales ...................................................................279

Isoetales ...................................................................................... 320

Vegetative Features............................................................................282

Putative lycopsids ...............................................................325

Reproductive Biology .......................................................................294 Conclusions ........................................................................... 326

Sigillariaceae .....................................................................................303 Other Lepidodendrid Genera ............................................................307

And, again, it is an error to imagine that evolution signifies a constant tendency to increased perfection. Thomas Henry Huxley The lycopsids have an extensive geologic history, extending back into the Late Silurian (Kotyk et al., 2002). They were widespread during the Late Mississippian and most of the Pennsylvanian, representing the dominant group of plants in most of the vast Euramerican paleoequatorial coal-swamp ecosystems. The widespread coal-mining operations that have uncovered Carboniferous rocks have been responsible, in large part, for the abundant, well-preserved fossil specimens of this group. The paleoecology of these plants has been extensively studied, and many constitute the focal point of ancient landscape reconstructions in museums around the world. The comprehensive geologic history of the lycopsids and numerous, exquisitely preserved specimens have provided paleobotanists with the opportunity not only to trace the evolution of the group but also to investigate some basic facets of their biology. The Lycophyta are monophyletic and basal within the vascular plants (Bateman, 1996b); together with the Zosterophyllophyta (Chapter 8), they comprise a clade (Lycophytina of Kenrick and Crane, 1997a). Both lycopsids and zosterophylls occupied the same habitats (floodplain and channel margins) during the Devonian (Gensel, 1992). The common names of this group have become somewhat confusing in recent years. Among

paleobotanists, lycopods has been the traditional name for all of the Lycophyta (without the zosterophylls), and you will find this name in the older literature; some neontologists, however, use this name to refer only to members of the Lycopodiales. More recently, lycopsids has been widely used. The term lycophytes has been used to refer not only to the Lycophyta but also to the clade composed of the lycopsids and the zosterophylls together. Kenrick and Crane (1997a) classified the lycophytes as the subphylum (subdivision) Lycophytina, containing two classes, Lycopsida and Zosterophyllopsida. We will use the name lycopsids to refer to the plants discussed in this chapter. According to Gensel and Berry (2001), the lycophytes represent a broader concept, which includes traditional lycopsids as well as the zosterophylls and transitional taxa such as Asteroxylon and Drepanophycus. Synapomorphies of the lycopsids include helically arranged microphylls and sporangia borne in the axil or on the adaxial (upper) surface of sporophylls. Most have exarch primary xylem maturation and, like the zosterophylls, lycopsid sporangia dehisce into two valves. In addition, metaxylem tracheids contain Williamson striations (discussed below). Currently available evidence suggests that the lycopsids originated from the zosterophylls, although these groups were coeval in the Siluro–Devonian (Gensel and Berry, 2001;

265

266

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Kotyk et al., 2002). From these beginnings, the lycopsids radiated extensively during the Carboniferous, in terms of both diversity and distribution, and then began to diminish in numbers of taxa and abundance toward the end of the Carboniferous, as the climate began to change and the extensive, equatorial peat swamps diminished (DiMichele et al., 2001a). Today the clade is represented by only seven genera in most treatments, of which three genera, Huperzia, Lycopodium (club mosses), and Selaginella (spike mosses), contain most of the 1600 species (gbif.org). The group is characterized by either dichotomous branching or a combination of dichotomous and monopodial branching. Stems are densely covered with true leaves termed microphylls. Microphylls are generally small (except in some of the extinct arborescent lycopsids), helically arranged, and vascularized by a single bundle that does not produce an interruption (leaf gap) in the stele of the stem when it separates. The roots in the lycopsids, as in other vascular cryptogams, are primarily adventitious. One of the most diagnostic features of the group is the position of the sporangium, which is borne in the axil or on the upper surface of a modified leaf or sporophyll. In some members sporangia are borne on short stalks. Sporophylls may be interspersed among photosynthetic microphylls, or they may be non-photosynthetic and aggregated into loosely constructed strobili or more consolidated cones. In most species the sporophyte produces only one type of spore and these plants are therefore regarded as homosporous. In some lycopsids, however, two types of spores are produced, and they not only look different but also function differently in the life history of the plant. Small spores (microspores, developed in microsporangia on microsporophylls) produce the male gametophyte, whereas the mega- or macrospores (developed in megasporangia on megasporophylls) germinate to produce the female gametophyte. Living heterosporous members of the Lycophyta produce both types of spores in the same strobilus (bisporangiate); in the heterosporous fossil representatives, microspores and megaspores were produced either in the same or in different cones (monosporangiate). All extant lycopsids are herbaceous and do not produce secondary vascular tissue, although many fossil forms are known to have been arborescent. Despite the large size of some of the arborescent members (40 m tall), the amount of secondary vascular tissue is small compared with the total stem diameter. Maturation of the primary xylem is exarch in most lycopsids with scalariform wall thickenings, the most common type of secondary wall pattern on the tracheids. Vascular organization ranges from protostelic to siphonostelic.

Although leaf gaps are not produced, interruption in the vascular cylinder occurs when branch traces are produced. The classification of lycopsids, similar to many groups of fossil plants, is a difficult task, in part because of the large number of fossil taxa for which there is a limited amount of information, especially reproductive characters. Thomas and Brack-Hanes (1984) have suggested the formation of what they term satellite taxa to accommodate various plant organs that cannot be accurately placed in well-defined families, after the initial concept of satellite genera was proposed by Meyen (1978). Their classification is similar to the one used in this volume, with the exception that we prefer the order Lepidodendrales rather than the Lepidocarpales, and we continue to include Miadesmia within the Selaginellales. In some phylogenies, heterospory and the presence of a ligule are used to group certain taxa; in others herbaceous versus arborescent habit has been used to define hierarchy. DiMichele and Bateman (1996) included all the rhizomorphic lycopsids (Lepidodendrales, Isoetales) in a single order, the Isoetales. The rhizomorphic lycopsids are those with a stigmarian-type rooting system, which can be either laterally extensive or small and lobed (discussed below). As research continues with the lycopsids, it is increasingly clear that delineation into major clades is not easily resolved, and that transformation series leading to the origin of some modern forms that at one time seemed well defined, are today more difficult to resolve (Gensel and Berry, 2001). The following traditional classification of the lycopsids into seven orders is intended to provide a framework for discussion of this group of plants. As is true of most groups with fossil members, there are several enigmatic forms that do not fit precisely into this classification. Higher taxa in this chapter:

Lycophyta Lycopsida Drepanophycales (Devonian) Protolepidodendrales (Devonian–Mississippian) Lepidodendrales (Devonian–Permian) Lepidodendraceae, Diaphorodendraceae, Sigillariaceae Lycopodiales (Pennsylvanian–recent) Lycopodiaceae Selaginellales (Pennsylvanian–recent) Selaginellaceae Pleuromeiales (Triassic–Cretaceous) Isoetales (Upper Devonian–recent) Isoetaceae, Chaloneriaceae

CHAPTER 9

Evolution of the microphyll One of the synapomorphies of the lycophytes is the presence of microphylls. Microphylls are true leaves, and as such, are borne in a definite pattern (phyllotaxy) on the stem, but they have an evolutionary history separate from the leaves of other vascular plants (megaphylls, see discussion in Chapter 11). Extant microphylls are small, although not all fossil microphylls are small. They usually contain a single vascular bundle (vein) and there is no leaf gap formed in the stele during the production of leaf traces (Chapter 7). Microphylls are believed to have evolved from enations, which as noted in the previous chapter are small, unvascularized flaps of tissue that do not have a phyllotaxy. There are several fossils that could be used to illustrate the stages in the evolution of microphylls. Beginning with a naked, vascularized axis (FIG. 9.1A), the first laterals produced were small, scattered spines, as seen in many zosterophylls, for example Sawdonia (FIG. 9.1B). Asteroxylon illustrates an intermediate stage (FIG. 9.1C), in which the laterals (enations) are more leaf-like, but vascular tissue extends only to the base of each enation. Finally, the laterals

LYCOPHYTA

267

(microphylls) become vascularized and are borne in a regular pattern on the axis, for example as seen in Leclercqia (FIG. 9.1D). There are several fossil plants that represent intermediates between stages B, C, and D, and some of these are included in the Drepanophycales. Kenrick and Crane (1997a) suggested several hypotheses on the origin of the lycophyte sporophyll, including reduction from a lateral branch (FIG. 9.2A), fusion of a sporangium to an enation– microphyll (FIG. 9.2B), and sterilization of a second sporangium (FIG. 9.2C). A study of stelar and microphyll vasculature in two extant Lycopodium species (Gola et al., 2007) confirms the unique nature of microphylls, especially when compared to megaphylls, the leaf type found in other vascular plants (see section “Evolution of Megaphyll” in Chapter 11). This study reveals that the production of microphylls and the architecture of the stele are not closely connected developmentally, as they are in seed plants. Based on these data, the authors concluded that the origin of the microphyll and the development of vasculature to supply the leaf in the lycopsids may have evolved independently, a theory which is supported by many of the plants discussed in the next section.

(A)

(B)

(A)

(B)

(C)

(D) (C)

Figure 9.1 Hypothesized stages in the evolution of the microphyll.

A. Vascularized axis (shaded area) lacking epidermal appendages (e.g., Rhynia). B. Axis with epidermal appendages in the form of spines or enations (e.g., Sawdonia). C. Axis with vascular traces extending to the base of the enation (e.g., Asteroxylon). D. Axis with vascularized microphylls (e.g., Leclercqia). (From Taylor and Taylor, 1993.)

Suggested transformational stages in the evolution of the microphyll with sporangium in lycopsids. A. Reduction from a lateral branch system. B. Origin from a sterile stem appendage (enation theory). C. Origin from a sterilized sporangium. (Modified from Kenrick and Crane, 1997a.)

Figure 9.2

268

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Drepanophycales This group includes the oldest known lycophytes (Banks, 1960). In some treatments these Devonian plants are regarded as prelycopsids and classified based on their stratigraphic position and lack of some typical lycopsid characters (Gensel and Andrews, 1984), whereas in others they are considered as transitional between the zosterophylls and the true lycopsids. They represent a diverse collection of presumably herbaceous plants, although anatomical details and complete specimens are not known for all taxa. The upright axes have exarch primary xylem maturation and bear helically arranged appendages that sometimes appear to be in a near-whorled or pseudowhorled pattern. These appendages are either non-vascularized or partially vascularized, that is, the vein runs into the proximal portion of the appendage, but does not extend through the entire structure, so in this way they are similar to microphylls, rather than being true leaves. The appendages do not dichotomize at their tips as do microphylls in the Protolepidodendrales, but this may not be a sufficiently well-defined character, as it can vary with preservation. Fertile specimens suggest that sporangia are not aggregated into cones. In contrast to the true lycopsids, the sporangia of most Drepanophycales are not borne adaxially on sporophylls, but rather arise directly from the axis shortly above an appendage, that is, in an axillary position. The lack of definitive information on this feature, however, has been used to suggest that many of the fossils are better placed in the Lycopodiales (Bateman, 1996b). All members are homosporous. One of the oldest lycopsids is Baragwanathia (FIG. 8.1), known from the Siluro-Devonian rocks of Australia and the lower Middle Devonian of Canada (Lang and Cookson, 1935; Hueber, 1983; Garratt et al., 1984; Rickards, 2000). Baragwanathia longifolia is similar to a modern Lycopodium or Huperzia in that it had dichotomous branches that bore small, closely spaced, helically arranged appendages. The plants are much more robust than those of Lycopodium, however, with stems reaching 6.5 cm in diameter and appendages up to 4 cm long. Associated with some of the appendages are reniform sporangia that produced trilete spores 50 μm in diameter. The compressed nature of the fossil material makes it difficult to determine whether the sporangia are attached to the upper surfaces of the appendages or borne on short stalks in the axils of appendages. The stele is stellate in cross-sectional outline, and the exarch xylem has annular tracheids. Baragwanathia abitibiensis is based on fossils from the Emsian (Lower Devonian) of Canada and consists of axes up

to 3.2 cm in diameter with helically arranged, microphyll-like appendages that bend downward (Hueber, 1983). Anomocytic stomata are randomly scattered on the appendages, but are far more common on the stems. The exact configuration of the xylem cylinder could not be determined, but the stele is constructed of tracheids with both annular and helical wall thickenings that closely conform to the G-type wall thickening. Sporangia were not found on any of the specimens. One interesting feature of B. abitibiensis is the truncated end of the mature appendages. It is believed that this unusual tip morphology is the result of postmortem changes. A similar cause may be used to explain the truncated or cup-tipped spines that have been described in some zosterophylls and trimerophytes. Drepanophycus spinaeformis (FIG. 9.3) was once regarded as a good index fossil of Lower Devonian strata; however, specimens have now been discovered in Middle and Upper Devonian rocks as well. Although traditionally placed with the lycopsids, Schweitzer (1980b) included the genus in the

Figure 9.3 Suggested reconstruction of Drepanophycus spinaeformis (Devonian). (Courtesy D. A. Eggert.)

CHAPTER 9

zosterophyll complex. Aerial axes were probably produced from horizontal rhizomes (Banks and Grierson, 1968) (FIG. 9.4). Branching in D. spinaeformis is in an H or K configuration (as in zosterophylls). This pattern forms when a branch departs from a rhizome at a right angle and then dichotomizes at right angles to produce two stems that parallel the primary axis (Li, 1995). Some specimens from the Emsian of Scotland possess axes interpreted as endogenously formed roots (Rayner, 1984). Drepanophycus spinaeformis axis fragments are up to 27 cm long and 4.2 cm wide. On some specimens from the Canadian Arctic buds occur in the position of branches (Gensel et al., 2001). The surface of the stem in D. spinaeformis is covered with raised mounds (FIG. 9.5) (leaf cushions), representing bases of microphyll-like appendages that have broken off. Other specimens have been preserved so that the inner surface of the outer part of the stem is exposed. In these, the appendages are surrounded by the matrix under the specimen and, instead of mounds, one sees

269

depressed horizontal or circular areas that denote the position of the appendages still in place. Microphyll-like appendages up to 2 cm long were borne in a shallow helix. Stomata have been described as occurring randomly among elongate, polygonal epidermal cells. The stomatal apparatus of D. spinaeformis is the paracytic type, consisting of two guard cells and two reniform subsidiary cells surrounded by a ring of epidermal cells that vary in number. Permineralized axes of D. spinaeformis reveal a lobed, exarch protostele 2 mm in diameter. Individual tracheids are 70 μm in diameter and 1 mm long. Secondary wall thickenings are annular–helical and possess a perforated reticulum in the position of the middle lamella (Hartman, 1981), now interpreted as G-type thickenings. Within the genus, sporangia are borne in either an axillary position or adaxially on appendages. Little is known about the spores, although the plants are regarded as being homosporous. Another species, D. gaspianus, had more robust axes characterized by rhombic leaf bases that bore microphylllike appendages with broad bases and recurved tips. Appendages were produced in a flat helix that contained 18–22 rows. Specimens from the Lower Devonian of New York

Figure 9.5 Figure 9.4 James D. Grierson. (Courtesy M. A. Millay.)

LYCOPHYTA

2 cm.

Drepanophycus spinaeformis (Devonian). Bar 

270

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

include a lobed, stellate protostele with annular tracheids and a perforate reticulum similar to that in D. spinaeformis (Grierson and Hueber, 1968). Drepanophycus gujingensis is a species from the Emsian of Yunnan Province in China (C.-S. Li and Edwards, 1995). It has sporangia attached to the stems by short stalks and adventitious roots attached to both the fertile and sterile axes. The stalked, adaxial or axillary sporangia are considered to represent an intermediate feature between plants like Asteroxylon (discussed below) and later lycopsids with sporangia borne directly on sporophylls. Another early lycopsid that is morphologically similar to Drepanophycus is Halleophyton (Early Devonian of China). It has rhomboidal to hexagonal swollen leaf bases and sporangia that split into two equal valves (C.-S. Li and Edwards, 1997). Haskinsia is a herbaceous lycopsid that was once regarded as a species of Drepanophycus (Grierson and Banks, 1983; Xu and Berry, 2008). The helically arranged appendages are falcate and 3 mm long (FIG. 9.6). Metaxylem tracheids are characterized by various patterns of bordered pits. In H. hastata appendages are arranged in a pseudowhorl and are 5 mm long (Berry and Edwards, 1996a, b); sporangia are globose (Yang et al., 2008). These authors placed the genus with the Protolepidodendrales based on the presence of deltoid-shaped sporophylls. Another presumably herbaceous Devonian lycopsid is Haplostigma. In H. baldisii the

microphyll-like appendages are simple and possess subhexagonal bases (Gutiérrez and Archangelsky, 1997). Asteroxylon mackiei (FIG. 8.41) (Chapter 8) is sometimes included in the Drepanophycales, or within a clade that includes Baragwanathia and Drepanophycus as a sister group to all other lycopsids (Kenrick and Crane, 1997a). As noted in Chapter 8, Kidston and Lang originally described terminal sporangia that were thought to belong to this taxon, but subsequent studies showed that the sporangia were borne laterally on the stems near the axils of microphyll-like appendages and not in a terminal position (Lyon, 1964). Whereas A. mackiei is known exclusively from the Lower Devonian Rhynie chert, a second species, A. elberfeldense (FIG. 9.7), is based on impressions with partial anatomical preservation from the Middle Devonian of Germany, Scotland, and

6 cm (A)

(B)

Figure 9.7 Asteroxylon elberfeldense, proximal and distal and Figure 9.6 A. Haskinsia sagittata and B. H. hastata (Devo-

nian). (From Berry and Edwards, 1996a.)

narrow, naked axes fragments of Stolbergia spiralis (Devonian). Bar  6 cm. (Courtesy BSPG.)

CHAPTER 9

Norway. The best-preserved specimens come from Kirberg near Elberfeld, western Germany (Kräusel and Weyland, 1926; Weyland et al., 1969). Axes of A. elberfeldense (originally named Thursophyton milleri) are dichotomous to sympodial, up to 1 m long and 0.5 cm in diameter. They bear

LYCOPHYTA

271

helically arranged and densely spaced, short microphyll-like appendages. Often found in association with the spiny A. elberfeldense axes are narrower, dichotomously branched, naked axes of the Hostinella hostimensis type (sometimes misspelled as Hostimella), which were originally believed to represent the terminal portion of A. elberfeldense (Kräusel and Weyland, 1926). Fairon (1967), however, demonstrated that these axes do not belong to A. elberfeldense, but rather represent a different plant, for which she introduced the name Stolbergia spiralis. Although superficially similar to A. mackiei, the systematic affinities of A. elberfeldense remain unclear. Anatomically preserved specimens from the Middle Devonian of the Aachen region in Germany suggest that A. elberfeldense, as well as S. spiralis, belong to the lycophytes (Fairon, 1967). Hestia eremosa is a putatively herbaceous and phylogenetically primitive lycopsid with uncertain affinities that has been described based on isolated stems from a Mississippian sequence of tuffs and lacustrine deposits at Oxroad Bay, East Lothian, Scotland (Bateman et al., 2007). Stems are characterized by the combination of a stellate stele, scalariform xylem pits, and perforate sheets of wall material partially infilling the pits, a complement of features largely consistent with that seen in Huperzia, the most plesiomorphic extant genus of Lycopodiaceae. The limited number of characters preserved in the fossils, however, cannot preclude placement of H. eremosa within the Drepanophycales. An excellent example of a lycopsid that possesses characters of several groups is Smeadia (FIG. 9.8) from the Upper Devonian Cleveland Shale of Ohio (Chitaley and Li, 2004). This plant was herbaceous with a siphonostelic stem (FIG. 9.9) and helically arranged leaves. At the distal end was an erect strobilus (FIG. 9.8) that contained trilete spores 40–80 μm in diameter.

Protolepidodendrales

Figure 9.8 Distal end of Smeadia clevelandensis showing ter-

minal cone (Devonian). Bar  5 mm. (From Chitaley and Li, 2004.)

Members of the Protolepidodendrales extend from the Devonian into the Mississippian (Lower Carboniferous) and fit the definition of a true lycopsid. They were either herbaceous or subwoody plants or small trees, and possessed small, helically arranged microphylls that branched at their tips and were vascularized by a single vein (Berry, 1996). Although one genus has been found to bear ligules (discussed below), they have not been found in any other taxa. The generic name Protolepidodendron has been used for a variety of Middle Devonian, dichotomously branched lycopsids (FIG. 9.10) (Grierson and Banks, 1963). Their stems are up to

272

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

9.9 Transverse section of Smeadia clevelandensis showing pith, the ring of vascular tissue and cortex of stem (Devonian). Bar 1 mm. (From Chitaley and Li, 2004.)

Figure

Figure 9.11 Suggested reconstruction of fertile leaf of

Minarodendron cathaysiense (Devonian). (From Taylor and Taylor, 1993.)

Figure 9.10 Suggested reconstruction of Protolepidodendron (Devonian). (Courtesy D. A. Eggert.)

2 cm in diameter and covered with helically arranged microphylls that were typically bifurcated at the tip. Schweitzer and Cai (1987) described P. cathaysiense from the Middle Devonian of southern China, which they believe is identical to Leclercqia (discussed below), based on the presence of highly bifurcated microphylls. C.-S. Li (1990) combined P. cathaysiense with P. scharyanum and erected a new genus, Minarodendron. Specimens of M. cathaysiense (Givetian) are 3–4 mm wide and generally exhibit a longitudinal series of elongate cushions on which the leaves are borne (FIG. 9.11). The apex of each microphyll is trifurcate, with two tips directed up and a single, median one pointing down. The stems contain an exarch or mesarch strand of primary xylem that is toothed or triangular in cross section. Tracheids range from annular to bordered pitted, and the sporangia are globose–reniform and borne on the adaxial surfaces of unmodified sporophylls. Some Early Devonian specimens that were initially described as species of Protolepidodendron are now called Estinnophyton (Fairon-Demaret, 1978, 1979) (FIG. 9.12). Like other protolepidodendrids, E. gracile was a small, herbaceous plant with axes up to 4 mm wide. Leaves were helically arranged and up to 7 mm long. Each fertile leaf bore two pairs of sporangia, each attached by a small stalk, at a short distance from the tip (FIG. 9.13), although in E. yunnanense (Lower Devonian of China) each fertile leaf contains two

CHAPTER 9

LYCOPHYTA

273

Figure 9.14 Suggested reconstruction of Colpodexylon deatsii

(Devonian). (Courtesy D. A. Eggert.)

Figure 9.12

Muriel Fairon-Demaret.

Figure 9.13 Diagrammatic reconstruction of Estinnophyton

yunnanense. (From Hao et al., 2004.)

single stalked sporangia (Hao et al., 2004). This species also contains poorly preserved trilete spores and tracheids with annular to helical wall thickenings. Estinnophyton is included here with Protolepidodendrales principally on the bifurcate nature of the leaves. However, the paired sporangia suggest affinities with the trimerophytes, whereas their recurved organization is similar to that of some Devonian sphenophytes (Hao et al., 2004).

One of the better-preserved members of the Protolepidodendrales known from Middle and Upper Devonian rocks in New York is Colpodexylon deatsii (Banks, 1944). The dichotomously branched stems are up to 2.5 cm wide and reveal elliptical leaf bases arranged in a low helix or appearing as a pseudowhorl (FIG. 9.14). The characteristic feature of this fossil is the presence of trifurcate leaves, which reached 3 cm in length (FIG. 9.15). The primary xylem strand is lobed in cross section, with exarch– mesarch maturation and annular tracheids. Sporangia are borne on the upper surfaces of unmodified leaves that are scattered along the stem surface. In C. trifurcatum (FIG. 9.16), a Middle Devonian taxon, the trifurcating leaves are 2 mm wide. The primary difference between the two species is the larger leaf base in C. trifurcatum, which may extend up to 5 mm long. In C. camptophyllum (FIG. 9.16) from the Devonian of Venezuela, the tips of the leaves are described as being shorter (Berry and Edwards, 1995). Clwydia (formerly Archaeosigillaria) is a small, dichotomously branched, herbaceous lycopsid that extends from the Devonian into the Carboniferous (Lacey, 1962). The leaf bases range from fusiform on smaller axes (FIG. 9.17) to hexagonal on larger stems. Despite the helical arrangement of the small leaves, they appear decussate and organized into vertical ranks. Some specimens superficially resemble Lycopodites, differing only in the possession of decussate, needlelike leaves (FIG. 9.18). In other species, such as C. vanuxemii, the leaves are deltoid in outline and possess a toothed margin; extending from the apex of the leaf is an elongate hair (Fairon-Demaret and Banks, 1978). The leaves of Clwydia probably did not abscise. The vascular system consists of a lobed protostele with exarch primary xylem and scalariform metaxylem tracheids.

274

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

(E)

(A)

(B)

(C)

(F)

(D)

Figure 9.16 Morphologies of various Colpodexylon leaves:

A. C. trifurcatum; B. C. deatsii; C. C. camptophyllum; D. C. cachiriense; E. C. variabile; and F. C. coloradense (From Berry and Edwards, 1995.)

Figure 9.15 Leaf of Colpodexylon trifurcatum (Devonian). (From Berry and Edwards, 1995.)

Phytokneme is a petrified lycopsid axis about 3 cm in diameter discovered in an Upper Devonian phosphatic nodule in the Chattanooga Shale of Kentucky (Andrews et al., 1971). The specimen is exquisitely preserved, with all cells and tissue systems intact (FIG. 9.19). In P. rhodona, the middle cortical zone, which is typically poorly preserved in fossil

lycopsids, contains a network of radially aligned, ray-like strands; these appear similar to cells that characterize the axis of the Carboniferous cone Lepidostrobus kentuckiensis. Roy and Matten (1989) suggested that Phytokneme may have affinities within the Lepidodendrales. A number of lycopsids have been described from the Devonian–Mississippian New Albany Shale of Kentucky and Indiana (Roy and Matten, 1989). One of these is Fodiodendron defractus, characterized by an exarch protostele with scalariform-reticulate metaxylem tracheids. The presence of two

CHAPTER 9

LYCOPHYTA

275

Figure 9.17 Leafy axis of Clwydia (Archaeosigillaria) (Devonian). Bar  1 cm.

strands of included phloem in the leaf traces distinguishes Fodiodendron from other permineralized lycopsids. Probably the most completely known member of the Protolepidodendrales is Leclercqia (FIG. 9.20), a slender, herbaceous plant known from the Early–Middle Devonian of Australia (Fairon-Demaret, 1974; Meyer-Berthaud et al., 2003), North America (Banks et al., 1972; Gensel and Kasper, 2005; Gensel and Albright, 2006), South America (Berry, 1994), and Europe (Fairon-Demaret, 1981).

Figure 9.18 Axis of Clwydia (Archaeosigillaria) showing

prominent leaves (Devonian). Bar  2 cm.

Specimens of L. complexa are up to 46 cm in length and vary from 3.5 to 7 mm in diameter. The dichotomously or pseudomonopodially branched axes are covered by microphylls that attained lengths of 6.5 mm (FIG. 9.21). The leaves are

276

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 9.19 Partial cross section of Phytokneme rhodona

(Devonian). Bar  3 mm. (From Andrews et al., 1971.)

unusual in that they exhibit a pair of lateral divisions about half way up the lamina (Bonamo et al., 1988). Each of these divides into two acuminate tips, whereas the central portion of the leaf gradually tapers and recurves abaxially (FIG. 9.22). The closely spaced leaves, each with five slender tips, must have given the plant an unusual appearance. Stomata of the anomocytic type (Gensel and Albright, 2006) are present on the microphylls and stems, and a few have been observed on the wall of a sporangium. Ligules (FIG. 9.23) have also been reported on Leclercqia microphylls (Grierson and Bonamo, 1979). The function of these small flaps of tissue which occur only in lycopsids has been of historical interest dating back to Hofmeister (1851). Some of the many suggestions as to the function of ligules include secretion and accumulation of water, mucilage, enzymes, and/or nutrients, or superficial conduction of water (Pant et al., 2000). The presence or absence of ligules has been used to define some groups of lycopsids.

Figure 9.20 Leclercqia complexa axis (Devonian). Bar  2 cm.

Some Leclercqia specimens are known in which the vascular cylinder has been preserved as a pyrite petrifaction. In cross section, the stele is circular with up to 18 external protoxylem points and exarch primary xylem maturation.

CHAPTER 9

LYCOPHYTA

277

(A)

Figure 9.21 Suggested reconstruction of Leclercqia complexa

(Devonian). (From Kenrick and Crane, 1997a.)

Metaxylem tracheids are scalariform or have oval pits on their walls. Outside the stele is a narrow, parenchymatous cortex with scattered cells with thickened walls, which may be sclerenchyma. Each leaf is vascularized by a single trace that originates from one of the protoxylem ridges on the stele. Sporangia are attached to the adaxial surfaces of sporophylls by a small pad of tissue just proximal to the lateral segments of the leaf. The distribution of sporophylls is similar to that in many species of Hupezia, in which fertile and sterile leaves are almost indistinguishable from one another and interspersed along the stems. Immature spores of Leclercqia are preserved in tetrads. At maturity, the spores are trilete, 60–85 μm in diameter, and ornamented with numerous, closely spaced spines with expanded bases (FIG. 9.24); in situ spores of L. complexa from eastern New York State have been assigned to the dispersed spore genus Aneurospora (Streel, 1972), whereas others have been compared to Actinosporites lindlarensis (Richardson et al., 1993; Gensel and Albright,

(B)

Figure 9.22 A. Vegetative and B. fertile leaf of Leclercqia complexa (Devonian). (From Kenrick and Crane, 1997a.)

Figure 9.23 Leaf of Leclercqia complexa with distal branched tips. Arrow indicates position of ligule (Devonian). Bar  625 μm. (From Grierson and Bonamo, 1979.)

278

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 9.25 Leaf of Cervicornus wenshanensis with eight segments (Devonian). Bar  0.5 mm. (From C.-S. Li and Hueber, 2000.)

Figure 9.24 Tetrad of Leclercqia complexa spores with one

spore removed from the tetrad (Devonian). Bar  10 μm. (From Gensel and Albright, 2006.)

2006). Details of the spores, including those in tetrads, have been reported by Gensel and Albright (2006). The large number of fertile specimens with morphologically identical spores has been regarded as evidence that Leclercqia was homosporous. Cervicornus is a small herbaceous lycopsid with helically arranged leaves divided into eight segments (FIG. 9.25) that superficially appear similar to those of Leclercqia (C.-S. Li and Hueber, 2000). Hubeiia dicrofollia is a herbaceous member of the Protolepidodendrales from the Upper Devonian Xiejingsi Formation of Hubei Province in China (Xue et al., 2005). Protoxylem occurs in ridges on the outer surface of the stele and consists of annular tracheids; metaxylem tracheids are scalariform with Williamson striations, delicate, vertical strands of secondary wall material that extend between the scalariform bars, also called fimbrils. These are characteristic of the tracheids in the Lepidodendrales (discussed below). Primary phloem forms a narrow band surrounding the stele. The cortex is relatively thick. Mesarch leaf traces depart directly from the protoxylem. Leaf bases are circular or slightly elliptical in shape and arranged in low helices or in pseudowhorls with the leaf bases alternating. Leaves are

subdivided into four segments as a result of two successive dichotomies. Another interesting lycopsid from China is Wuxia bistrobilata from the Upper Devonian (Berry at al., 2003). Compressed specimens of branches are up to 1.4 cm wide and possess sterile leaves inserted in whorls of six. Megasporangiate structures occur at the dichotomies of the axes and consist of elongate megasporophylls, each with a prominent midrib. Spines occur at irregular intervals along the megasporophyll. Megaspores are up to 4 mm in diameter. Morphologically W. bistrobilata shares most features with Minarodendron cathaysiense from the Middle Devonian (C.-S. Li, 1990). Another Chinese lycopsid of Devonian age with megasporophylls, each bearing 4–6 megasporangia, is Chamaedendron multisporangiatum (Schweitzer and Li, 1996). Microsporangia are stalked and contain spores of the Longhuashanispora type. Chamaedendron is reconstructed as a narrow tree-like plant lacking secondary xylem. Longostachys latisporophyllus from the Middle Devonian of China is reconstructed based on numerous specimens from the same site. It bears elongate megasporophylls with the distal lamina recurved upward and trichome-like appendages arising from the margin and is believed to represent another small tree-like lycopsid (Cai and Chen, 1996). The stele changes from a protostele, surrounded by secondary xylem near the base, to a medullated stele that lacks secondary tissues at higher levels. It is suggested that L. latisporophyllus is intermediate between the herbaceous Protolepidodendrales and the arborescent lepidodendrids. Monilistrobus yixingensis (FIG. 9.26) is a species from the Late Devonian Wutung Formation (Famennian) of Jiangsu, China, that is similar in overall structure to Longostachys and Chamaedendron, but can be

CHAPTER 9

Figure 9.26 Monilistrobus yixingensis showing partial reconstruction of the fertile axis (Devonian). (From Y. Wang and Berry, 2003.)

distinguished from these two taxa (and from all other fossil and most living lycopsids) based on the occurrence of sporangia on modified, proximally widened sporophylls which are compactly arranged into distinct, cone-like fertile zones separated by vegetative regions with more lax microphylls (Y. Wang and Berry, 2003). Another plant believed to represent a transitional stage between the Protolepidodendrales and arborescent lycopsids of the Late Devonian and Carboniferous is Zhenglia radiata from the Lower Devonian of southeastern Yunnan, China (Hao et al., 2006). This plant is characterized by undivided microphylls, sporophylls arranged helically to form a compact area resembling a cone, and ovoid-elongate sporangia positioned adaxially on the widened proximal portion of the sporophyll. Leaf scar arrangement is similar to that seen in Lepidodendrales. Another arborescent lycopsid assigned to the Protolepidodendrales is Protolepidodendropsis pulchra from the Middle and Upper Devonian of Spitsbergen (Høeg, 1942; Schweitzer, 1965, 1999). The largest compressions discovered to date indicate that the stems of this plant, which were originally described as Bergeria mimerensis by Høeg (1942), reach a thickness of 10 cm and a

LYCOPHYTA

279

height of 1.5–2 m. Stem surfaces are covered with helically arranged, broad rhombic leaf cushions that vary considerably in size, depending on the position in the stem, and reach 1.5  1.2 cm in the proximal portions of the stem. In the middle of each of the leaf cushions is a prominent leaf scar; parichnos scars and ligule pit are lacking. Distally the stems dichotomize to form two branches that, in turn, dichotomize up to five times to form a loose crown. Branch fossils do not display leaf cushions, but rather show only leaf scars, which are spindle shaped, that is pointed at both ends, with a thin ridge from the lower scar tip to the upper tip of the scar below (Schweitzer, 1999). Leaves are small and simple (Høeg, 1942). Protolepidodendropsis frickei, a second species in that genus, has been described from the Upper Devonian of Bögendorf-Liebichau near Waldenburg (today Walbryzch) in Silesia (Poland) by Gothan and Zimmermann (1937). In contrast to P. pulchra, stems of P. frickei remain unknown to date, and Banks (1960) suggested that this plant was herbaceous. As we learn more about Devonian and Mississippian lycopsids, it is becoming increasingly clear that not all of them fit into present classification schemes. Such forms as Linietta and Lycopogenia from the Famennian–Tournaisian of North America (Roy and Matten, 1989), and Trabicaulis and Landeyrodendron from the Montagne Noire (Tournaisian) of France (Meyer-Berthaud, 1984a) underscore the necessity of continuing to reevaluate early lycopsids (Meyer-Berthaud, 1984a).

Lepidodendrales The Lepidodendrales includes the arborescent lycopsids that were the most conspicuous elements of the Carboniferous landscape. Members of this group are responsible, to a large extent, for the extensive quantities of plant material that resulted in the formation of Carboniferous coal seams around the world. It is estimated that up to 70% of the biomass in the extensive Westphalian coal-swamp forests of Euramerica was produced by members of the Lepidodendrales (DiMichele et al., 1985). Toward the end of the Westphalian, however, their numbers were in decline and, in the Stephanian epicontinental swamp forest ecosystems, these plants are responsible for merely 5% of biomass production (see Kerp, 2000 and references therein). At the end of the Carboniferous, most arborescent lycopsids become extinct in Europe and North America, and are replaced by tree ferns that, for the first time in geologic history, formed a relatively closed forest canopy; in China, however, the arborescent lycopsids persist

280

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

into the Middle Permian. It is suggested that the disappearance of the Euramerican Lepidodendrales toward the end of the Carboniferous was due to climate change (DiMichele and Phillips, 1996a). Others hypothesize that the Variscan tectonic activity considerably reduced the size of the coalswamp ecosystems, which led to more dynamic environments and unstable conditions that were no longer suitable for the lepidodendrids (Kerp, 1996, 2000). Some authors have combined these individual assumptions and suggested that climatic changes were caused by dramatic changes in the floral composition that, in turn, were triggered by tectonic activity (Cleal and Thomas, 2005). Because of the large number of specimens collected, and the variety of ways in which they were preserved (FIG. 9.27), members of the Lepidodendrales are easily the best understood of all fossil lycopsids. The occurrence of structurally preserved members of the order, as well as the extensive stratigraphic distribution of the group, have provided paleobotanists with material to address a variety of geologic and biological aspects of these coal-swamp giants. The great diversity of fossil material has provided exceptionally detailed knowledge not only of the reproductive biology and developmental stages of the vegetative parts of these plants, but also of the role of these plants within the paleoecosystem and the structure of the coal-swamp ecosystem through time. Members of the Lepidodendrales are ligulate and characterized by the presence of secondary xylem, extensive periderm development, a three-zoned cortex, spirally arranged rootlike appendages (stigmarian rootlets) with a monarch vascular strand, and a single functional megaspore per megasporangium, which germinates within the sporangium (DiMichele and Bateman, 1996). Historically specimens preserved in many different modes, for example impressions, compressions, and structurally preserved fossils, were included in Lepidodendron, a genus defined principally on features of leaf cushion external morphology. To alleviate ambiguity and provide better resolution of the systematics and diversity of the arborescent lycopsids, DiMichele (1985) established Diaphorodendron for some structurally preserved specimens previously placed in Lepidodendron, including D. vasculare, D. scleroticum, and D. phillipsii. Later, Diaphorodendron was divided into two genera, Diaphorodendron and Synchysidendron, and placed in its own family, the Diaphorodendraceae (DiMichele and Bateman, 1992). Synchysidendron includes two species, S. dicentricum (formerly Lepidodendron) and S. resinosum (DiMichele and Bateman, 1992). According to DiMichele and Bateman (1996), synapomorphies of the family include a medullated

Figure 9.27 Base of a stem of Tomiodendron peruvianum in growth position (Mississippian). (Courtesy H. W. Pfefferkorn.)

protostele (siphonostele) and a megasporangium that is dorsiventrally flattened with proximal dehiscence; the megaspore has a massa. Synapomorphies of the Lepidodendraceae (Lepidodendron, Lepidophloios, and Hizemodendron) include a bilaterally flattened megasporangium with distal dehiscence and infrafoliar parichnos that extends below the leaf scar (discussed later), among other characters (DiMichele and Bateman, 1996). Today, the generic names Diaphorodendron and Lepidodendron are not only used for stem segments exhibiting cellular preservation but also to encompass the concept of entire plants (FIG. 9.28), including leaves, underground parts, reproductive organs, and a variety of other parts (e.g., decortication stages, spores, isolated sporophylls). Each of these plant organs is assigned both generic and specific names, as relatively few of them have been found organically attached to each other. The generic name Lepidodendron is retained for impression–compression specimens that possess a particular type of leaf cushion morphology in addition to

CHAPTER 9

LYCOPHYTA

281

Figure 9.28 Suggested reconstruction of an arborescent lycopsid (e.g., Lepidodendron). (Courtesy D. A. Eggert.)

certain anatomical features. Specimens of Lepidodendron have been discovered that indicate some trees attained heights in excess of 40 m and were at least 2 m in diameter at the base (Thomas and Watson, 1976). In the midcontinent of North America, the genus appears to have reached its zenith at the end of the Middle Pennsylvanian (Phillips et al., 1974). The massive, erect trunks of some Lepidodendron species branched profusely to produce large crowns of leafy twigs (FIGS. 9.29, 9.30). Some leaves were at least 1 m long (most were much shorter) and, when they dropped from the branches, conspicuous leaf bases were left on the stem surface.

Figure 9.29 Leafy lycopsid twig (Pennsylvanian). Bar  2 cm.

282

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

L

VB P

9.30 Compression of Lepidodendron lycopodites (Pennsylvanian). Bar  2.5 cm.

Figure

Strobili were borne at the tips of distal branches or in a zone at the top of the main trunk. The underground portions of the Lepidodendrales (sometimes called a stigmarian rhizomorph) consisted of dichotomizing axes that bore helically arranged, lateral appendages that presumably functioned as roots.

Figure 9.31 Lepidodendron leaf base. L: ligule; P: parichnos;

VB: vascular-bundle scar. (From Taylor and Taylor, 1993.)

Vegetative Features

STEM SURFACE AND LEAF BASES Some of the most commonly encountered fossils assignable to the lepidodendrids are compressions of stem surfaces marked by persistent, somewhat asymmetric, more or less rhomboidal leaf cushions (FIG. 9.31). The leaf cushion actually represents the expanded leaf base left behind after the leaf dropped off (FIG. 9.31), since abscission of the leaf did not occur flush with the stem surface. The top and bottom of the cushions, which are also called leaf bolsters, generally form acute angles; the sides are more rounded. The actual scar left by the abscised leaf is slightly above the midpoint of the cushion and is generally elliptical or rhombic in outline (FIG. 9.32). On the surface of the leaf scar are three small, pit-like impressions. The central one represents the single vascular-bundle trace that extended into the leaf. The other two

scars represent the position of channels of loosely arranged parenchyma tissue, termed parichnos (FIG. 9.31); this tissue originates in the cortex and extends through two grooves on the abaxial surface of the leaf. On Lepidodendron stem surfaces, two additional parichnos channels can be identified at a short distance beneath the leaf scar; these do not occur in the Diaphorodendraceae, where parichnos is confined to the foliar scar only. Parichnos is a system of aerating tissues within the stem. A vertical line extends from the leaf scar proper to the lower limit of the leaf base. In many specimens, lateral wrinkles cut across this line. Initially, it was thought that these wrinkles were of systematic value, but it is now understood that they are the result of the growth of secondary tissues in the stem. Just above the leaf scar is a mark that

CHAPTER 9

LYCOPHYTA

283

Figure 9.32 Several Lepidodendron leaf cushions preserved in

a Mazon Creek nodule (Pennsylvanian). Bar  3.5 mm.

represents the former position of a ligule. Additional markings occur on the leaf base in the form of vertical lines that result from lateral expansion of the stem. Rarely is preservation so good that all these features can be observed in a single leaf base. In Synchysidendron, the leaf cushion includes a groove that is formed by folding of the cushion tissue immediately below the leaf scar proper. Some compression specimens of arborescent lycopsid stems (FIG. 9.33) also provide information about the epidermis of these plants. A waxy cuticle covers the stem surface, including the leaf cushions, but is thought to be absent on the leaf scar itself (Thomas, 1966). The epidermis is simple and lacks such specialized cells as hairs and glands. Stomata are common and sunken in shallow pits (Thomas, 1974). Another tree-sized lycopsid, Lepidophloios (Q. Wang, 2007) (originally spelled Lepidofloyos; Sternberg, 1825), occurred in the Carboniferous coal swamps along with Lepidodendron (DiMichele, 1979a). Lepidophloios (Lepidodendraceae) was probably slightly smaller in stature (FIG. 9.34), but in general its features are quite similar to Lepidodendron. One notable difference between the two is the arrangement and organization of leaf bases (FIG. 9.35). In Lepidophloios, the leaves are arranged in a shallow helix, as in other lepidodendrids, but the leaf bases are flattened and wider than they are tall (FIG. 9.35). They are directed downward on the stem and overlap the bases below, much like shingles on a roof. When attached, the leaves bend upward abruptly from the leaf bases; when a leaf abscised, it left a scar on the bottom third of the base

Figure 9.33 Lepidodendron leaf bases (Pennsylvanian). Bar 

1 cm. (Courtesy BSPG.)

(FIG. 9.36). Parichnos and vascular-bundle scars on the leaf scar are like those of Lepidodendron, but parichnos scars are not present on the base itself. Lepidophloios was ligulate, with the ligule attached just above the position of the leaf scar. Many Lepidophloios stems exhibit large, circular to elliptical scars on the stem surface (FIG. 9.37), some up to several centimeters in diameter. The origin of these scars has been debated for many years. Some workers regard them as former sites of vegetative branches that abscised during the normal growth of the plant, whereas others suggest that they represent former positions of specialized branches that bore clusters of strobili. Historically, stems with helically arranged scars of this type have been given the generic name Halonia, whereas those with oppositely arranged scars are

284

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

called Ulodendron. Jonker (1976) suggested that Ulodendron scars on the axes of Lepidodendron, Lepidophloios, and a related genus, Bothrodendron, represent the former positions of branches that abscised in a manner similar to that of some existing gymnosperms and angiosperms. Thomas (1967) regarded Ulodendron as a natural genus, basing this hypothesis on the persistent leaves, shallow ligule pits, and rhomboidal leaf bases, whereas DiMichele (1980) suggested it was congeneric with Paralycopodites.

L

VB

P

Figure 9.36 Diagrammatic representation of Lepidophloios leaf base showing ligule (L), parichnos (P), and vascular-bundle scar (VB) (Pennsylvanian). (From Taylor and Taylor, 1993.)

Figure 9.34 Suggested reconstruction of distal region of

crown branches of Lepidophloios hallii (Pennsylvanian). (From DiMichele, 1979a.)

Figure 9.35 Paradermal section of Lepidophloios leaf bases

Figure 9.37 Three prominent oval branch scars along a

showing four branch scars (Pennsylvanian). Bar  1 cm.

Lepidodendron axis (Pennsylvanian). Bar  4 cm.

CHAPTER 9

STEM ANATOMY Species of Diaphorodendron have been delimited on the basis of internal stem organization. The configuration of the stele has been used as a taxonomic character as well as to suggest evolutionary changes that may have taken place within the arborescent forms. Knowledge of changes in stelar configuration at various levels of the stem has clearly demonstrated that many of the so-called species of Lepidodendron and Diaphorodendron as well as some of the species of Lepidophloios (FIG. 9.38) may represent different developmental stages of a single species. These studies have provided us with a basic understanding of just how these giant Carboniferous plants actually grew (FIG. 9.39). Before discussing the ontogenetic changes that occur during the growth of arborescent lycopsids, it is important to detail the internal structure of a typical lepidodendralean stem. Stems of D. scleroticum, D. phillipsii, and D. vasculare are relatively abundant in coal-ball deposits from the eastern interior basin of North America (DiMichele, 1979b) and the description that follows will be a composite picture of the internal organization of a typical stem. DiMichele suggests that the principal feature

LYCOPHYTA

285

of lepidodendrids, that is, leaf cushions that are longer than they are wide, masks the extensive diversity present within the genus, but there are several anatomical differences as well. Stems may be protostelic (Diaphorodendron), have a mixed pith, or be siphonostelic (Diaphorodendron and Lepidodendron) (FIG. 9.40). Protostelic stems such as L. pettycurense from the Mississippian of Scotland (see Kidston, 1907) consist of a central strand of primary xylem surrounded by a narrow ring of protoxylem that in turn is surrounded by secondary xylem. In stems with a mixed pith, parenchyma cells are interspersed with tracheids in the center, but the tracheids are shaped more like short, squat parenchyma cells rather than

B

Figure 9.39 Arborescent lycopsid (Lepidodendron) showing

branch trace (B) (Pennsylvanian). Bar  1 cm.

C

L

Figure 9.38 Several steles of the Arran Tree (Lepidophloios

wuenschianus) (Pennsylvanian). Bar  5 cm.

Figure 9.40 Transverse section of Diaphorodendron stem showing small vascular cylinder (arrow), conspicuous cortex (C), and prominent leaf bases (L) (Pennsylvanian). Bar  5 mm.

286

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

long, slender cells. This condition is often cited as evidence that the pith in lycopsids originated when immature parenchymatous cells in the center of stems failed to differentiate into tracheids. Surrounding the mixed pith is a narrow band of metaxylem tracheids with scalariform wall thickenings; protoxylem occurs at the periphery of the metaxylem. In cross section, the outer margin of the primary xylem appears fluted due to the large number of protoxylem ridges. As in most members of this division, maturation of the primary xylem is exarch. Surrounding the primary xylem in the arborescent lycopsids may be a zone of secondary xylem, which reaches a maximum thickness of several centimeters. Unlike woody trees today, secondary xylem in the arborescent lycopsids accounts for only a small proportion of the diameter of the stem; rather, the extensive development of periderm is primarily responsible for their massive trunks. Both primary and secondary xylem tracheids are scalariform and have Williamson striations (fimbrils) extending between the scalariform bars. These fimbrils are characteristic of the wood in the arborescent lycopsids, but similar fimbrils also occur in modern club mosses and spike mosses (Wilder, 1970; Schneider and Carlquist, 2000a, b), and are regarded as a shared character for the Lycopsida as a whole (DiMichele and Bateman, 1996). Numerous vascular rays radiate through the secondary xylem; these are generally a single cell wide and up to several cells high. Due to the presence of scalariform thickenings on the walls of the ray cells, it has been suggested that xylem rays in this group evolved from tracheids. Immediately outside the secondary xylem is a zone of thinwalled cells that represent the vascular cambium. Unlike the bifacial cambium of typical seed plants (Chapter 7), however, the vascular cambium in the lepidodendrids is unifacial, producing only secondary xylem on the inner face (Eggert and Kanemoto, 1977). The unusual manner in which this secondary vascular tissue was produced was determined by examination of stems with and without secondary vascular tissue, and by comparing the distribution of cell types within the stems. In Diaphorodendron and Lepidodendron, the phloem zone is separated from the secondary xylem by a band of thin-walled cells termed the parenchyma sheath. Immediately outside the parenchyma sheath are sieve elements with large, elliptical sieve areas on their walls, interspersed with strands of phloem parenchyma. In some specimens a zone of radially seriate, thin-walled cells is in contact with the secondary xylem cylinder. Although some believe that this tissue resulted from cambial activity, others suggest it was formed of living primary-sheath cells that were capable of becoming meristematic. Current evidence seems to suggest that no secondary phloem was produced within the arborescent lycopsids.

CORTICAL TISSUES The cortex of the lepidodendrids is usually subdivided into three general zones—the inner, middle, and outer cortex (FIG. 9.41), which have been defined on the basis of cell types. The inner cortex is the narrowest of the cortical zones and is constructed of small, isodiametric parenchyma cells. Aggregations of cells with dark contents, presumed to be secretory cells, in addition to lacunae and various types of sclerotic cells, also occur in this zone. The middle cortex is more extensive and constructed of larger parenchyma cells. In young stems this zone is characterized by lacunae that extend radially; in older stems the middle cortex is generally not preserved, except for a few parenchyma cells along the inner and outer edges. Cells of the outer cortex show no definite arrangement. They have slightly thicker walls and superficially resemble collenchyma cells. In some species, this zone may be distinguished by longitudinally oriented, anastomosing bands of fibers. Secondary cortical tissue, or periderm, is produced in the outer cortex and, judging from the extensive blocks of this tissue in coal balls, it is this tissue that contributed to most of the trunk diameter in these plants. In Diaphorodendron, the periderm is bizonate, with a differentiation into phelloderm (inner zone) and phellem (outer zone). The inner zone consists of alternating thick- and thin-walled cells, whereas the outer zone is homogeneous and contains cells with dark contents, described as appearing resinous. Lepidodendron has a massive periderm (FIG. 7.13), which may be homogeneous or bizonate, and there is no clear differentiation into phellem and phelloderm (DiMichele, 1985); cells may be tangentially expanded in the outer part of the cortex. Periderm production originates

Figure 9.41 Transverse section of Diaphorodendron stem with

branch trace (arrow) (Pennsylvanian). Bar  1 cm.

CHAPTER 9

from meristematic parenchyma cells (phellogen) in the outer cortex just beneath the leaf bases. The phellogen in the arborescent lycopsids produced a relatively small amount of phellem toward the outside of the stem, and a much larger amount of phelloderm toward the inside, the opposite of most extant seed plants (see Chapter 7). Cells of the periderm are radially aligned and often show a storied arrangement, that is, in longitudinal section, the end walls of the cells are lined up horizontally. In some lepidodendrids, the periderm is more complex and consists of three kinds of cells: (1) thickwalled, axially elongated fibers; (2) radial rows of chambered cells, conspicuous because they are divided by transverse walls; and (3) secretory cells aligned in sinuous bands that give the appearance of growth rings in cross sections. It has been suggested that periderm in the arborescent lycopsids formed in several ways (Eggert, 1961). Some axes show successive tangential bands of meristematic tissue at varying depths in the cortex. In this pattern of periderm production, outer cortical layers become meristematic and produce radially oriented files of cells for a period of time. As each group of meristematic cells ceases dividing, successively deeper cortical cells become meristematic and repeat the process of periderm production. The presence of short files of periderm cells in both stems and underground axes supports such a pattern of development. In some cases, thickwalled cells are produced in the outer cortex. The outer surface of arborescent lycopsid stems is not the only surface to have been preserved as fossils. Both the loose construction of the cortex and the production of large amounts of relatively thin-walled periderm contributed to the sloughing off of stem layers and tissues, either as a feature of normal development or as a result of mechanical separation during the fossilization process. As a result, a variety of presumed external stem and trunk features are represented in fossils that are unlike the leaf bases and have resulted from levels of decortication of the axis. Various generic names have been applied to these decorticated states, but the names do not conform to the concepts of morphogenera and are therefore of little value in systematic studies. Nevertheless, these various fossil forms do provide information that can be used to reconstruct developmental changes in the component tissue systems of an axis. One of the more common decortication stages is Knorria, a name used for stems in which almost all the tissues external to the xylem (and sometimes even the stele) have been lost; Knorria represents a mold–cast type of preservation and is characterized by irregular, longitudinally oriented ridges. These ridges represent sediment infillings of the parichnos strands that accompany the vascular bundles through

LYCOPHYTA

287

the stem (Hirmer, 1927). Knorria casts are known not only from Lepidodendron stems (FIG. 9.42), but have also been reported from Bothrodendron and Jurinodendron (FIG. 9.43) (formerly Cyclostigma) stems. STEM DEVELOPMENT One of the outstanding contributions to modern paleobotany is Eggert’s detailed investigation of development in arborescent lycopsids (Eggert, 1961) (FIG. 9.44). Studies of this type have not only contributed to our understanding of growth processes in fossil groups, but have also made it possible for paleobotanists to distinguish developmental differences represented by fossils from features that are useful in lycopsid taxonomy (Delevoryas, 1964a). In his analysis, Eggert utilized a large number of permineralized axes with varying diameters; the axes included different amounts of primary and secondary tissues. Using this approach, he was able to reconstruct the pattern of growth in these plants and demonstrate the successful application of this technique in the analysis of other groups of plants. In these arborescent forms, the upper portion of the main axis contains a large siphonostele that is characterized by

Figure 9.42 Lepidodendron stem or branch in Knorria preser-

vation (Pennsylvanian). Bar  1 cm. (Courtesy BSPG.)

288

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

a wide pith surrounded by a thick zone of primary xylem (FIG. 9.45). As very young specimens exhibit only a small protostele, there must have been a change in stelar configuration from the protostelic sporeling stage to the siphonostele present in the main trunk. This initial expansion of the primary body during early growth is termed epidogenesis. When the plant is relatively immature, the cortex is extensive and the outer surface of the trunk is covered with numerous rows of raised leaf bases (FIG. 9.45). As the tree continues to grow, secondary xylem and periderm are added to the stem

Figure 9.44 Donald A. Eggert.

kiltorkense in Knorria-type preservation (Mississippian). Bar  5 cm. Figure

9.43 Jurinodendron

Figure 9.45 Longitudinal and cross sections showing the distribution of primary (solid) and secondary (radiating lines) xylem in an arborescent lycopsid. (After Eggert, 1961.)

CHAPTER 9

as a result of the vascular cambium and phellogen. The increase in stem diameter results in the sloughing off of the outer cortical tissues (FIG. 9.45), including the leaf bases, so that in older parts of the plant (e.g., at the base), the outer surface of the trunk is protected by periderm. Many of the older reconstructions of Lepidodendron in museums and drawings often err in showing leaf bases extending all the way to the ground on old trunks. At higher levels in the tree, the branches have smaller steles and fewer rows of smaller leaves on the surface. Sections of stems at these levels indicate that less secondary xylem and periderm are produced. A reduction in stele size and tissue production continues until the most distal branches, which contain a tiny protostele with only a few small tracheids, no secondary xylem or periderm, and just a few rows of leaves. This stage in development, in which the plant literally grows itself out, has been termed apoxogenesis. In other words, the small, distal twigs of these arborescent lycopsids do not have the potential of developing into larger branches with time. This type of growth pattern is called determinate and contrasts with indeterminate growth, which is typical of vegetative development in most living woody plants. Paleobotanists must continually devise new methods of investigating the biology of the organisms they study. Eggert’s elegant analysis of growth in the arborescent lycopsids is one such approach. In another, the focus of the study is the nature of the unifacial vascular cambium in two Carboniferous lycopsid morphogenera, Stigmaria and Paralycopodites (Cichan, 1985a). Cichan (FIG. 9.46) prepared serial tangential sections of the secondary xylem in order to determine the pattern of production of cambial derivatives and the method of circumferential increase in the cambium. Cichan’s studies indicate that cambial activity in these plants was also determinate. Circumferential increase took place by the enlargement of fusiform initials, rather than by anticlinal divisions of existing initials, as it does in seed plants. This type of growth would result in a cambium that was limited in its capacity for radial expansion. As secondary growth ceased in the plant, fusiform initials ceased to be meristematic and matured into a cylinder of parenchyma. LEAVES The leaves of arborescent lycopsids are linear and some were up to 1 m long (FIG. 9.47). Chaloner and MeyerBerthaud (1983) demonstrated that stems with the largest diameters have the longest leaves, a feature they correlate with the determinate growth of the plants. Many of the different species established for detached leaves were probably

LYCOPHYTA

289

produced by the same kind of plant and only differed in size, shape, and anatomy because of their position on the plant. The generic name Lepidophyllum was initially used for both structurally preserved and compressed lepidodendrid leaves, but because this name had been used earlier for a flowering plant, the name Lepidophylloides was proposed in its place (Snigirevskaya, 1958). A single vascular bundle, flanked by two shallow grooves on the abaxial surface, extends the entire length of the lamina in Lepidophylloides (FIG. 9.48). Stomata occur on the abaxial surface aligned in rows that parallel the grooves and sunken in shallow pits. A well-developed hypodermal zone of fibers surrounds the mesophyll parenchyma and vascular bundle of the leaf; no palisade parenchyma has been reported. In L. sclereticum from Permian coal balls, the vascular bundle is convex in transverse section and surrounded by transfusion tracheids (S. J. Wang et al., 2002). UNDERGROUND ORGANS Underground axes of the Lepidodendrales are given the name Stigmaria. These dichotomizing structures represent one of the most common lycopsid fossils and constitute the

Figure 9.46 Michael A. Cichan.

290

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 9.48 Cross section of Lepidodendron leaf. Note

two abaxial furrows (arrows) where stomata are located (Pennsylvanian). Bar  1.5 mm. (Courtesy BSPG.)

Figure 9.47 Lepidophylloides in Mazon Creek nodule (Pennsylvanian). Bar  2 cm.

principal organ found in the clay layer or underclay immediately beneath most Carboniferous coal deposits. The underclay represents the soil layer (paleosol) in which these plants were rooted in the coal swamps (Mosseichik et al., 2003). Extensive specimens of Stigmaria have been uncovered in growth position, some with rootlike structures, commonly called stigmarian appendages, still attached (FIG. 9.49). Although there are several species of Stigmaria, our knowledge of the anatomy of these underground systems is based principally on the species Stigmaria ficoides (FIG. 9.50) (Williamson, 1887b). The stigmarian system arises from the base of the trunk as four primary axes, each of which extends out horizontally, so that the rooting system is relatively shallow. Helically arranged lateral appendages were attached to each axis. These appendages abscised during the growth of the plant, leaving characteristic circular scars (FIG. 9.51) on the main axis, and these can be seen on a variety of casts, compressions, and impressions of Stigmaria. The lateral appendages are sometimes called stigmarian rootlets (FIG. 9.52), although their helical arrangement (i.e., phyllotaxy) and abscission are characteristic of leaves rather than lateral roots (see Chapter 7). The primary axes in Stigmaria dichotomize repeatedly to form an extensive subterranean system that may have radiated up to 15 m from the trunk. Primary axes of Stigmaria have a parenchymatous pith that may include scattered tracheids at more distal levels. Primary xylem is endarch and arranged in a series of dissected bands, which are, in turn, surrounded by a vascular cambium. What is interpreted as secondary xylem is distinctive because the wide vascular rays give the wood a segmented appearance (FIG. 9.53). Secondary xylem tracheids are aligned in radial files and possess scalariform wall

CHAPTER 9

LYCOPHYTA

291

Figure 9.49 Stigmaria ficoides compression showing laterally attached rootlets (Pennsylvanian). Bar  2 cm.

Figure 9.50 Cross section of stigmarian axis showing cylinder of secondary xylem and several rootlets (arrows) attached to the outer surface of the cortex (Pennsylvanian). Bar  2 cm.

Figure 9.51 Cast of Stigmaria sp. showing pattern of rootlet scars (Pennsylvanian). Bar  1 cm.

Figure 9.52 Cross section of stigmarian rootlet showing monarch collateral bundle (Pennsylvanian). Bar  2 mm.

thickenings with fimbrils identical to those of the aerial parts. Occasional imperfections in the secondary xylem appear to be the result of a temporary cessation of cambial activity, which caused an abrupt change in the diameter and distribution of the tracheids in the wood. It has been suggested that this erratic cambial activity was a result of some abrupt environmental change. No secondary phloem has been identified in Stigmaria; the vascular cambium was apparently unifacial and translocation effected by the primary phloem. Rothwell

292

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 9.53 Cross section of the secondary xylem cylinder of Stigmaria ficoides (Pennsylvanian). Bar  4 mm.

Both primary and secondary cortical tissues of Stigmaria ficoides are complex, consisting of numerous cell and tissue systems, some of which are difficult to trace developmentally. In general, however, the production of secondary cortical tissues in the underground parts resulted in a narrow zone of periderm that is histologically similar to that of the aerial stems. A cast of the apex of Stigmaria is known from the Middle Pennsylvanian of Iowa (FIG. 9.54) (Rothwell, 1984). The specimen is 8.5 cm long and contains helically arranged rootlet scars that surround a raised rim that is believed to correspond to the shape of the meristem. Morphologically, the specimen is nearly identical to the juvenile stage of the underground rooting structure of Nathorstiana, a Mesozoic lycopsid (Karrfalt, 1984). The lateral appendages, sometimes termed rootlets, produced by the stigmarian axes are up to 40 cm long and 0.5–1 cm in diameter, usually unbranched (some may dichotomize once), and they gradually taper distally. Each rootlet has a small monarch vascular strand surrounded by a compact inner cortex. External to this is a hollow, middle cortical zone and a thin outer cortex (FIG. 9.52). At some levels, a connective extends from the outer cortex to the inner cortex (Weiss, 1902) (FIG. 9.55). Stigmaria stellata, from the Upper Mississippian Chester Series of Illinois, exhibits radiating ridges on the surface around the lateral appendage scars in both casts and impressions (Jennings, 1973). Structurally preserved axes suggest a close relationship to S. ficoides, but S. stellata differs in several anatomical features, including the absence of a connective in the lateral appendages.

Figure 9.54 Apex of Stigmaria ficoides (Pennsylvanian).

Bar  1.5 cm. (From Rothwell, 1984.)

and Pryor (1991) combined observations on mold–cast specimens with permineralized axes and hypothesized that the radially aligned tracheids in the steles of many stigmarian axes were produced by a primary thickening meristem rather than a vascular cambium.

Figure 9.55 Frederick E. Weiss.

CHAPTER 9

LYCOPHYTA

293

Figure 9.56 Detail of Lepidodendropsis sp. leaf bases (Mississippian). Bar  5 cm.

Figure 9.57 Axis of Lepidodendropsis sp. showing leaf bases

Not all stigmarian underground parts are extensive, dichotomously branched systems. Some, such as the Early Mississippian genus Protostigmaria (Jennings, 1975a), consist of cormlike axes bearing helically arranged laterals similar to those of Stigmaria. In P. eggertiana from the Price Formation of Virginia (Mississippian), both underground parts and some aerial axes are preserved in growth position (Jennings et al., 1983). Axes are up to 32 cm in diameter and preserved as impression–compression and mold–cast specimens (FIG. 1.36). The root-bearing portion is divided into lobes by a system of furrows, with additional furrows added as the plants grew; the largest number of lobes recorded in P. eggertiana is 13. Roots were arranged on each lobe in a helical pattern; each of the scars is circular and up to 8 mm in diameter. Features of the stem surface (FIG. 9.56) and its association in the same rocks strongly suggest that Protostigmaria was the underground portion of Lepidodendropsis (FIG. 9.57), a Mississippian arborescent lycopsid. One feature of Protostigmaria that deserves additional comment is the ability of the plant to maintain an upright position despite the relatively small anchoring surface of the lobed base. Except for the larger size and larger number of lobes, Protostigmaria is morphologically identical with the root-bearing structure of the extant lycopsid Isoetes, and this

(Mississippian). Bar  2 cm.

further strengthens the homologous nature of the root-bearing organ of lycopsids. Experimental studies suggest that during the development of Isoetes (Karrfalt, 1977), non-contractile roots in the furrows of the cormlike base move laterally as tissues are added, resulting in the plant being pulled farther down into the substrate. It has been suggested that such a mechanism may have been operative in Protostigmaria, as it lacked the extensive, dichotomizing anchoring system of some of the other arborescent lycopsids (Jennings et al., 1983). Research with Protostigmaria also illustrates that developmental data can be determined from fossils that are preserved in ways other than permineralizations. DEVELOPMENT OF UNDERGROUND ORGANS Development of the underground parts of the arborescent lycopsids was probably quite similar to the epidogenic and apoxogenic stages described for the aerial stems. Despite the large number of Stigmaria specimens that have been collected, several features of these organs remain to be determined and interpreted. For example, the helical arrangement

294

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

of the lateral appendages is unlike the irregular arrangement of roots in most living plants. No root hairs have been identified on any specimens; perhaps fungi in some of the cortical parenchyma cells functioned as mycorrhizae. The monarch vascular bundle in these so-called rootlets is bilaterally symmetrical, that is, a collateral bundle. Typically roots have radially symmetrical vascular tissue in cross section, that is the phloem surrounds the xylem, while the vascular bundles in leaves are collateral. Finally, the lateral appendages apparently abscised from the parent axis in a regular manner, perhaps similar to the process of foliar abscission in many plants. Programmed abscission of laterals is unknown in the roots of living plants and has not been observed in any non-lycopsid fossil plants. For these reasons, it has been suggested that the rootlets of Stigmaria are actually homologous with leaves that have been modified for the functions of anchorage and absorption. This interpretation implies that the underground portions of these plants arose by evolutionary modification of aerial axes (Frankenberg and Eggert, 1969). Although some of the coal-swamp lycopsids in the Pennsylvanian grew to 40 m high, the stigmarian underground system was relatively shallow, and questions have arisen as to how much support it could have provided for these towering trees. Many of these plants grew in what must have been a supersaturated soil that also provided little stability. It may be, however, that the extensive horizontal development of the underground systems was sufficient to allow these plants to remain upright. Studies of extant trees suggest that the nature of the wood (e.g., strong, dense wood) and the density of the crown can have a pronounced effect on tree uprooting (Niklas, 1992). The arborescent lycopsids had very little secondary xylem, and this may have been an advantageous mechanical property in remaining upright in heavy wind or rain. In addition, they had somewhat bushy crowns. It has been suggested that the crowns of adjacent trees became entangled and thus provided mutual support for these trees in the Carboniferous swamps.

organization of lepidodendrid cones consists of a central axis with helically arranged sporophylls (FIG. 9.59). Sporangia are located on the adaxial (upper) surface of sporophylls which are upturned at their distal ends so that they overlap the sporophylls above. A portion of the lower surface of the sporophyll often extends downward to form a heel or distal extension. A ligule is present in a small pit just distal to the sporangium. Lepidostrobus is the most common designation for lycopsid cones of this type (FIGS. 9.60, 9.61). The name has been used, however, for cones demonstrating all forms of preservation, and for both monosporangiate (having only

Reproductive Biology

The reproductive units of the lepidodendrids consist of strobili or cones borne on distal branches in the crown of the tree. In Synchysidendron cones occur on late-formed crown branches, whereas in Diaphorodendron they are borne on deciduous lateral branches (DiMichele and Bateman, 1992). Lepidodendrid cones could reach considerable size (FIG. 9.58), for example cones assignable to Lepidostrobus goldenbergii could be more than 50 cm long. The basic

Figure 9.58 Compressed Lepidostrobus cone (Pennsylvanian).

Bar  2 cm.

CHAPTER 9

one type of spore) and bisporangiate (having two types of spores) forms, so that taxonomic problems within the genus are considerable.

LYCOPHYTA

295

MICROSPORANGIATE AND BISPORANGIATE CONES In an attempt to better define the taxonomic limits of arborescent lycopsid cones, Brack-Hanes and Thomas (1983)

suggested that the name Lepidostrobus be used for cones containing only small spores (monosporangiate) and Flemingites for bisporangiate strobili containing both microspores and megaspores (FIG. 9.62). Other authors have used cone morphology and in situ spores to more accurately circumscribe individual species within Lepidostrobus (Bek and Opluštil, 2004, 2006).

9.59 Compression specimen showing axis of Lepidostrobus cone with helically attached sporophylls (Pennsylvanian). Bar  1 cm.

Figure 9.60 Longitudinal section of distal end of permineralized Lepidostrobus cone with distal sporangia containing spores (Pennsylvanian). Bar  3 cm.

Figure

296

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 9.61 Transverse section of permineralized Lepidostrobus cone showing central axis and helically arranged microsporophylls (Pennsylvanian). Bar  2 cm.

Figure 9.62 Transverse section of Flemingites schopfii cone showing microsporangia (arrows) intermixed with megasporangia (Pennsylvanian). Bar  5 mm.

One of the oldest lepidodendrid cones is Lepidostrobus oldhamius, known from Lower and Middle Pennsylvanian deposits in both North America and Great Britain. Specimens range from 2–6 cm in diameter and exceed 30 cm in length. Sporangia are massive and, depending on the stage of development, may have an irregularly shaped pad of sterile tissue extending from the sporophyll into the lumen of the sporangium. All of the sporangia contain small (20–30 μm) spores that have a trilete suture on the proximal surface and delicate spines on the distal face. Dispersed spores of Lepidostrobus are often preserved, and one of the common generic names applied to these grains is Lycospora (Willard, 1989). Lepidostrobus shanxiense is a slightly smaller cone from the Xishan Coal Field in China (Wang et al., 1995). An interesting Lepidostrobus cone, L. xinjiangensis, also from China, occurs in Upper Devonian rocks (Wang, Q. et al., 2003a). The discovery of this species suggests that segregation into micro- and megasporangiate cones had already taken place in the lepidodendrids by the Late Devonian. Flemingites schopfii is a permineralized bisporangiate cone that resembles a massive modern Selaginella strobilus in general organization (Brack, 1970). Specimens are up to 8 cm long and 1.3 cm in diameter. The arrangement of parts and cone anatomy are identical with those of L. oldhamius, except for the presence of two types of spores in F. schopfii. Distal sporangia (FIG. 9.63) contain a large number of Lycospora-like microspores (FIG. 9.64), some still in tetrahedral tetrads; more basal sporangia contain 12–29 trilete megaspores (Brack-Hanes, 1978). The megaspores range from 700–1250 μm in diameter and are marked by an elongation of the proximal surface into an apical prominence or gula. Dispersed megaspores of this type are called Valvisisporites. Both compressed and structurally preserved lepidostroboid cones are known from the Fayetteville Shale of Arkansas (Upper Mississippian). These cones, which all appear to have been monosporangiate, may have attained lengths of 22.5 cm. Spores extracted from sporangia of both cone types provide a means of comparing different preservational types. In Lepidostrobus fayettevillense, the small, trilete miospores are characterized by a perforated flange (cingulum), that encircles the spore at the equator (FIG. 9.64) (Taylor and Eggert, 1968). It is probable that apparently monosporangiate cones such as L. oldhamius and L. fayettevillense represent the microsporangiate cones of heterosporous plants. In other instances, however, monosporangiate cones may have been produced by homosporous plants reproductively similar to Lycopodium, in which the spores germinate into free-living gametophytes. In this regard, the neutral term miospore

CHAPTER 9

LYCOPHYTA

297

Figure 9.63 Suggested reconstruction of sporangium of

Flemingites schopfii (Pennsylvanian). Figure 9.65 Megasporangium containing numerous Triletes sp. spores (Pennsylvanian). Bar  5 mm.

might be more appropriate than microspore, as the exact function of these spores in the life history of the plant is currently unknown. Mixostrobus givetensis, from the Middle Devonian of Kazakhstan, is a bisporangiate cone with an irregular arrangement of micro- and megasporangia. In some strobili, one type of sporangium is the dominant type, but this does not depend on the size of the cone (Senkevitsch et al., 1993). Cones assigned to the genus Kladnostrobus from the Pennsylvanian of the Kladno-Rakovník Basin, Czech Republic, differ from other lycopsid reproductive structures in that the in situ spores are reticulate and resemble the dispersed spore genera Convolutispora, Camptotriletes, Reticulatisporites, and Dictyotriletes (Libertín et al., 2005). Figure 9.64 Proximal surface of Lycospora-type spore show-

ing surface ornament, trilete suture, and equatorial cingulum (Pennsylvanian). Bar  7 μm.

MEGASPORANGIATE CONES Some lepidodendrid cones are monosporangiate, but produce only megaspores (FIGS. 9.65, 9.66). Possibly the

298

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 9.66 Megaspore of the Triletes-type showing proximal triradiate ridge and ornamentation (Pennsylvanian). Bar  1 mm.

Figure 9.67 Partial transverse section of Lepidocarpon palm-

erensis cone (Pennsylvanian). Bar  1 cm.

most common of these Lepidocarpon (FIGS 9.67, 9.68), which was borne on Lepidophloios stems. This cone type is regarded as the most highly evolved reproductive structure in the lycopsids, because the arrangement of the sporophylls closely approximates the function of integuments in seed plants. In Lepidocarpon, sporangia are borne on the adaxial surface of the sporophyll (FIG. 9.69), which consists of two lateral laminae (FIG. 9.70) and a distal extension or heel (Balbach, 1962, 1965) (FIG. 9.71) (as in Lepidostrobus and Flemingites) (Thomas, 1978). The lateral laminae partially envelop the sporangium with only a slit-like opening on the top. Within the sporangium is one large, functional, trilete

Figure 9.68 Compressed Lepidocarpon cone (Pennsylvanian).

Bar  1 cm.

megaspore and three aborted spores. The wall of the megaspore (called Cystosporites when found dispersed) is unique in that it is constructed of loosely arranged strands of sporopollenin. In some specimens, cellular megagametophytes are preserved, a few containing archegonia. Embryos have been described from Lepidocarpon cones and they are ellipsoidal, unvascularized, and characterized by an isoclinally folded epidermis (Phillips et al., 1975). As the embryo develops, there is a dichotomy of the vascularized axis; one branch of the dichotomy develops into the stigmarian rooting system and the other into the aerial stem. In more mature embryos, what has been termed secondary

CHAPTER 9

Figure 9.71

LYCOPHYTA

299

Margaret K. Balbach.

Figure 9.69 Longitudinal section of Lepidocarpon sporangium (below) and transverse section (above). Arrows indicate lateral laminae (Pennsylvanian). Bar  5 mm.

Figure 9.70 Cross section of Lepidocarpon lomaxi megasporangium showing lateral laminae, megasporangium, and multicellular megagametophyte (From Hirmer, 1927.)

tissues are present. When only the lepidostroboid sporophyll is preserved in the compressed state, the generic name Lepidostrobophyllum is sometimes used (FIGS 9.72, 9.73). Embryos have also been found in the cone Bothrodendrostrobus (FIG. 9.74) (Stubblefield and Rothwell, 1981). In the earliest stage of development, the embryo consists of an unvascularized globular structure embedded within megagametophyte tissue. In more mature specimens, two vascularized appendages extend through the trilete suture, one representing the first shoot, the other the first root. It was initially suggested that embryology in Lepidocarpon and Bothrodendrostrobus was sufficiently different to demonstrate two evolutionary paths within the lycopsids (Stubblefield and Rothwell, 1981). What was originally interpreted as bipolar development in Bothrodendrostrobus is now considered to be a derived condition of the unipolar developmental pattern demonstrated in Lepidocarpon (Rothwell and Erwin, 1985). Another megasporangiate cone that had an unusually ornamented megaspore is Caudatocorpus. This cone type was apparently monosporangiate, with helically arranged megasporophylls lacking lateral laminae (BrackHanes, 1981). The sporangium is large and the wall is constructed of columnar cells. Each sporangium contains a tetrad of megaspores with one large (FIG. 9.75), presumably

300

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

functional spore ( 4 mm long) and three smaller (200– 500 μm) aborted spores. If dispersed, these spores would be included in the genus Lagenicula (Scott and Hemsley, 1993). The sporoderm of the functional megaspore is 10 μm thick and has two layers. The outer surface is covered with numerous spines about 50 μm long; on the proximal surface is a conspicuous apical prominence. The four spores are enclosed in a granulose spongy structure that represents a

Figure 9.72 Detail of Lepidostrobophyllum showing sporangium attachment scar (arrow) (Pennsylvanian). Bar  1 cm.

Figure 9.74 Bothrodendrostrobus embryo showing two appendages extending from spore wall. Arrow indicates vascular element (Pennsylvanian). Bar  150 μm. (From Stubblefield and Rothwell, 1981.)

9.73 Compressed Lepidostrobophyllum showing the point of sporangium attachment (arrow) (Pennsylvanian). Bar  1 cm.

Figure

Figure 9.75 Section of Caudatocorpus arnoldii megaspor-

angium showing spiny megaspores of the Lagenicula type (Pennsylvanian). Bar  400 μm. (From Brack-Hanes, 1981.)

CHAPTER 9

distal, winglike attachment to the large functional spore. This extension may have been involved somehow in dispersal, or it may have functioned to orient the apical prominence of the trilete suture so as to enhance the possibility of fertilization. Achlamydocarpon is a monosporangiate cone with reduced lateral laminae and a single large, functional megaspore in each sporangium. Species of both Lepidodendron and Diaphorodendron are known to have borne this cone type (DiMichele, 1985). In Achlamydocarpon, the orientation of the trilete suture is toward the cone axis rather than away, as in Lepidocarpon. The suture of the functional spore in Achlamydocarpon is covered by a massa (FIG. 9.76) or cap of sporopollenin that may have functioned to protect the developing gametophyte and perhaps to help retain moisture in the region of the suture (Taylor and Brack-Hanes, 1976). In Lepidocarpon cones, this protection could have been provided by the conspicuous lateral laminae of the sporophyll, whereas in Achlamydocarpon, the developing megagametophyte may have been protected by both the reverse orientation of the proximal suture and the sporopollenin cap on the megaspore. Microsporangiate cones are assigned to A. varius on the basis of similarities in epidermal structure, pedicel alations, and other histologic details (Leisman and Phillips, 1979). The trilete microspores average 64 μm in diameter and exhibit scattered papillae over their distal surfaces, which may represent tapetal residues in the form of orbicules. If found dispersed, such grains would be included in the genus Cappasporites (Ravn et al., 1986). Achlamydocarpon pingquanensis is a megasporangium–sporophyll unit from the Lower Permian of China (Y. L. Zhou et al., 2006). The structure is 1.6 cm long and contains a large, presumably functional megaspore of the Cystosporites type, and three

Figure 9.76 Aborted megaspore of Achlamydocarpon varius with proximal massa (arrow) (Pennsylvanian). Bar  250 μm.

LYCOPHYTA

301

smaller abortive spores, each with a massa near the proximal suture. Based on the organization of A. pingquanensis, it is suggested that this disseminule adds additional support to the idea that there is a group of arborescent lepidodendrids from China that is distinct from those in Euramerican Carboniferous deposits. A single large megaspore within each sporangium is also a feature of the cone Suavitas imbricata collected from Upper Pennsylvanian marine deposits of Texas, USA (Rice et al., 1996). The cone is permineralized and the sporangium is located at the distal end of the sporophyll. Although the affinities remain conjectural, the analysis of characters suggests some relationship with the rhizomorphic lycopsids. Achlamydocarpon is believed to represent the cone of several different species of Diaphorodendron (DiMichele, 1981, 1985), based on an analysis of several hundred permineralized specimens of Diaphorodendron from Lower and Middle Pennsylvanian rocks. DiMichele recognizes different morphological groups among the Euramerican forms of lepidodendrids, each distinct relative to reproduction, habitat, and evolutionary history. One group, consisting of D. vasculare, D. scleroticum, and D. phillipsii, included trees 8–20 m tall (FIG. 9.77) that had deciduous lateral branches bearing cones of the A. varius type. Synchysidendron trees were smaller (from 10 to 15 m tall) and produced A. varius cones, but these occurred in large numbers near branch tips toward the end of the growing season. DiMichele (1981) suggested that the coal-swamp environments may have acted as evolutionary refugia for some of the arborescent lycopsids, such as Diaphorodendron and Lepidophloios, a habitat preference no doubt dictated by their reproductive biology, which was well adapted for aquatic dispersal. DiMichele (1980) also speculated that speciation in this group may have taken place outside the swamp habitat. Arborescent lycopsid megaspores of several types have been examined at the fine-structural level in an attempt to determine the affinities of the spores and also to investigate the development of the spore wall (T. Taylor, 1974). Wilson Taylor (1990) was able to correlate megaspore ultrastructure with the dispersal strategy in these lycopsids. His study indicates that some of the Carboniferous megaspores share both developmental and dispersal features with some modern species of Selaginella. Others appear to possess a uniquely organized sporoderm pattern (FIG. 9.78) that reflects the degree to which megaspores enlarge within the sporangium. W. Taylor, (1989) distinguished three basic types of construction (laminar, laterally fused spherules, and ordered units) in the walls of Selaginella. All megaspores possess an

302

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 9.78 Detail of sporopollenin units forming the wall of a functional Lepidocarpon megaspore (Pennsylvanian). Bar  3 μm.

N N N N

Figure 9.77 Suggested reconstructions of several arborescent lycopsids (left to right): Diaphorodendron vasculare, D. scleroticum, D. phillipsii, and Synchysidendron dicentricum. (From DiMichele, 1981.)

inner separable layer, which may be involved in regulating water balance. It is also important to characterize megaspore wall development at the ultrastructural level for its systematic value (Hemsley and Scott, 1989; Hemsley and Galtier, 1991). This type of study is useful in identifying the parent plants of dispersed spores, and the floral composition of particular assemblages where megafossils are absent or poorly preserved. Ultrastructural studies also provide important information about the development and evolution of lycopsid spore walls (Glasspool et al., 2000).

Figure 9.79 Polar view of Flemingites schopfii archegonium (arrow) showing four neck cells (N) (Pennsylvanian). Bar  100 μm.

GAMETOPHYTES Knowledge about the gametophyte generation of the lepidodendrid arborescent lycopsids is generally meager, and is based on only a few specimens (Renault, 1893; Gordon, 1908, 1910; MacLean, 1912). One interesting feature of Flemingites schopfii cones (discussed earlier) is the exquisite preservation of both the micro- and megagametophyte phases (Brack-Hanes, 1978). Within some of the megaspores near the trilete suture is a parenchymatous, cellular megagametophyte (FIG. 9.79). Some of the surface cells of the megagametophyte bear elongated tufts of rhizoids that extend from the trilete suture and actually penetrate the sporangium

CHAPTER 9

LYCOPHYTA

303

Figure 9.80 Polar view of several archegonia inside arbores-

cent lycopsid megaspore (Mississippian). Bar  80 μm. (Courtesy J. Galtier.)

wall. Other megagametophytes possessed archegonia at the time of fossilization, and several archegonial necks have been described interspersed among the rhizoids. Archegonia have from one to three tiers of neck cells (FIG. 9.82) and, in a few specimens, an enlarged cell, suggestive of an archegonial venter, occurs beneath the neck cells. Some microspores in the distal sporangia in F. schopfii reveal stages in the development of the microgametophyte, including partitions suggestive of the antheridial initial and prothallial cells. Some contain material that morphologically resembles chromosomes (Brack-Hanes and Vaughn, 1978) (FIG. 1.25). When compared with the gametophytes of existing lycopsids, F. schopfii has microgametophytes that are more similar to those of extant Selaginella, whereas the structure of the megagametophytes more closely corresponds to that of Isoetes. Other well-preserved lepidodendrid megagametophytes (FIGS. 9.80–9.82) have been discovered in spores assignable to Lepidodendron esnostense or L. rhodumnense that occur in late Viséan (uppermost Mississippian) cherts from central France (Galtier, 1964b, 1970a, b); these occur in dispersed spores as well as in specimens still within the megasporangium (FIG. 9.81). The megagametophytes are multicellular structures that develop inside the megaspore wall (endosporic gametophyte development). As they mature, however, they protrude from the spore in the region of the trilete suture, where they form a mass of tissue, within which several archegonia are produced (FIG. 9.82); rhizoids appear to be lacking. Archegonia are embedded in gametophytic tissue, with only the uppermost ring of neck canal cells protruding from the surface. The microgametophyte of this taxon remains unknown to date.

Figure 9.81 Gametophyte tissue inside arborescent lycopsid megaspore (arrows) still inside megasporangium (white arrow) (Mississippian). Bar  330 μm. (Courtesy J. Galtier.)

S

9.82 Megagametophyte tissue containing several archegonia (arrows) rupturing lycopsid megaspore wall (S). Bar  100 μm. (Courtesy J. Galtier.)

Figure

Sigillariaceae

Sigillaria is another important arborescent Carboniferous– Permian lycopsid, which did not branch profusely and was

304

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 9.84 Sigillaria mammillaris, stem surface with leaf

bases (Pennsylvanian). Bar  2 cm. (Courtesy BSPG.)

not as large as the lepidodendrids. Although some specimens have been reported to be 30 m tall (FIG. 9.83), it is probable that most sigillarians were 20 m tall. The absence of extensive branching and the structure of the leaf bases are the principal features that distinguish Sigillaria from other arborescent lycopsids. Sigillarian leaf bases are typically hexagonal in outline (FIG. 9.84), although some may be elliptical. Although they are helically arranged, they often appear to be aligned in vertical rows. The actual leaf scars are generally elliptical (FIG. 9.85), with a central leaf-trace scar (vascular-bundle scar) flanked by two large parichnos scars. The vascular bundle is v-shaped and may separate into two strands. A ligule scar is present above the leaf scar.

Figure 9.83 Suggested reconstruction of Sigillaria tree (Pennsylvanian). (Courtesy D. A. Eggert.)

LEAF BASES Subgenera of Sigillaria have been established to encompass the various configurations and arrangements of the leaf bases on the surface of the stems. Sigillaria subg. Eusigillaria includes forms with ribbed stem surfaces, and these have been divided into two sections. In the section Rhytidolepis, leaf bases and ribs are separated and the furrows between adjacent ribs are straight or nearly so. In the section Favularia, leaf bases and ribs are close together and the furrows have a zigzag configuration. Forms included in Sigillaria subg. Subsigillaria lack ribs. There are also two sections within this subgenus: Leiodermaria has widely

CHAPTER 9

LYCOPHYTA

305

VB P

L

Figure 9.85 Diagram of Sigillaria leaf bases showing posi-

tion of ligule (L), parichnos (P), and vascular bundle (VB) (Pennsylvanian). (From Taylor and Taylor, 1993.)

Figure 9.86 Sigillaria brardii, transverse section through a permineralized stem (Permian). Bar  5 cm. (Courtesy BSPG.)

separated vertical rows of leaf scars with no raised cushions, whereas leaf bases in Clathraria are closely arranged. LEAVES Most sigillarians must have been rather unusual looking plants with little distal branching and a large number of closely spaced, elongated leaves arising from the top of the trunk. Sigillariophyllum and Sigillariopsis leaves are similar to those of Lepidophylloides except that they may be vascularized by two laterally flattened strands instead of one. On the lower surface are two longitudinal grooves lined with conspicuous trichomes. Stomata are arranged in rows, and the guard cells are sunken. Cyperites is a morphogenus used for isolated linear leaves, usually 1 cm wide, which are butterflyor X-shaped in transverse section (Rex, 1983) and generally thought to represent compressed sigillarian leaves (Doubinger et al., 1995). Although the exact reason for the configuration in cross section is not clear, it is suggested that the two abaxial grooves may have contributed to this morphology. STEM STRUCTURE Although compressed remains of sigillarians are relatively common in Carboniferous rocks, structurally preserved stems are rare (FIG. 9.86). In Sigillaria approximata (FIG. 9.87) (Upper Pennsylvanian), the central portion of the stem consists of a parenchymatous pith surrounded by a continuous band of primary xylem (Delevoryas, 1957; Guo and Tian, 1994). In cross section, the outer edge of the exarch primary xylem appears sinuous, with leaf traces originating from the furrows. Metaxylem tracheids possess fimbrils between the scalariform bars. Relatively little secondary

Figure 9.87 Transverse section of Sigillaria approximata stem

(Pennsylvanian). Bar  1 cm.

xylem is produced, and it consists of scalariform tracheids and narrow vascular rays. Distribution of cortical tissues is similar to that described for Diaphorodendron, and tangentially banded periderm is common in these plants (FIG. 9.88); the periderm often contains concentric bands of presumably secretory cells. In addition, tangentially expanded cells form distinct clusters that appear spindle shaped in transverse section. Extending radially through the periderm are pairs of cylindrical to laterally flattened strands of parichnos tissue that can be related to the parichnos scars on the leaf bases and that may have functioned as an aeration tissue.

306

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 9.88 Periderm of Sigillaria in coal (Pennsylvanian).

Bar  1 mm. (From Winston, 1989.)

Sometimes Sigillaria is found in a partially decorticated state called Syringodendron. The outer surface exhibits vertical rows of large, often double elliptical scars that resemble rabbit tracks. These scars represent the parichnos strands as seen in tangential section, at a level of stem decortication. UNDERGROUND ORGANS Underground organs of Sigillaria are similar to the stigmarian system of the other lepidodendrids, but there are a few anatomic differences (Eggert, 1972). In stigmarian axes of Sigillaria, the pith is relatively narrow in proportion to the diameter of the stele, and consists of an outer zone of mixed tracheids and parenchyma and a central zone of parenchyma. The cortex is also narrow and consists of two primary zones. Secondary cortical development involves concentric rings of meristematic cells in the outer cortex, whereas in the underground parts of Diaphorodendron, periderm was produced from a single, central meristematic layer. In the lateral appendages (rootlets), the connective is continuous, unlike the organization in Stigmaria. The genus Stigmariopsis is known to have been the underground part of some subsigillarians and is distinguished from Stigmaria principally on the basis of unequal branching in which the smaller branch is directed downward (Eggert, 1972). One might be inclined to view the subtle differences between the underground parts of the sigillarians and Stigmaria as more apparent than real. To correlate underground parts with particular stem taxa, it is necessary to collect at sites where not all lycopsid taxa are present. Frankenberg and Eggert (1969) were able to characterize Stigmaria from several Middle Pennsylvanian coal-ball localities where Sigillaria was

Figure 9.89 Longitudinal section of Mazocarpon oedipternum cone. Arrow indicates megasporangium with megaspores (Pennsylvanian). Bar  5 mm.

absent. Eggert (1972), in turn, detailed the nature of the sigillarian underground system from an Upper Pennsylvanian site in which Sigillaria was present, whereas the other arborescent lepidodendrids were not, thus making it possible to correlate the aboveground and subterranean parts of the same plant even though they were not organically attached. REPRODUCTIVE BIOLOGY Sigillaria was heterosporous and produced monosporangiate cones. Evidence of the attachment of the cones to the parent plant comes not only from compressed specimens but also from structurally preserved stems that bear persistent cone peduncles interspersed among the leaf bases (Delevoryas, 1957). Mazocarpon oedipternum is a common Late Pennsylvanian taxon in North America (FIG. 9.89) (Schopf, 1941). The cones are 1.2 cm in diameter and frequently reach 10 cm in length. Sporophylls are arranged in a low helix or pseudowhorl, with the laminae forming conspicuous dorsal heels. The distal ends of sporophylls are relatively short. Megasporangia are roughly triangular as seen in a radial

CHAPTER 9

section and contain a central pad of parenchymatous tissue (subarchesporial pad) with eight megaspores around the periphery (FIG. 9.90). Megaspores are large and trilete, and have been described with short archegonial necks extending from the proximal suture. Microsporangiate cones (FIG. 9.91), also called Mazocarpon, contain trilete spores that average 60 μm in diameter. Mazocarpon villosum is a Late Pennsylvanian species that is 2.2 cm in diameter (Pigg,

S

S

Figure 9.90 Transverse section of Mazocarpon oedipternum

megasporangiate cone showing parenchymatous subarchesporial pad (S) in two sporangia. Note megaspores (arrows) between pad and sporangium wall (Pennsylvanian). Bar  3 mm.

LYCOPHYTA

307

1983). In this species, only immature megaspores have been found in apical sporangia. Megagametophytes have been described in several species of Mazocarpon; these consist of prothallial tissue with rhizoids and archegonia up to 65 μm long, each with three tiers of neck cells. Reproductively, the sigillarians are far more complex than originally thought. Pigg (1983) suggested three different types of megaspore dispersal in the genus. In one type, exemplified by Mazocarpon villosum, spores are believed to have developed rapidly and then been shed from the sporangium. In another type, for example M. oedipternum, the sporophylls are retained on the cone at maturity and the megaspores were dispersed (some with megagametophytes) when the sporangial wall broke down. This contrasts with M. pettycurense and M. cashii, which are hypothesized to have dispersed megaspores with their attached sporophylls by fragmentation of the cones. The presence of megaspores containing embryos embedded within the intersporangial tissue of the megasporangium has been used as evidence to suggest that apomixis occurred in this group (DiMichele and Phillips, 1985). A small amount of sterile tissue has been described in the microsporangia of M. bensonii (Feng and Rothwell, 1989). In this species the trilete microspores range from 48–54 μm and the sporoderm is constructed of a dense reticulum of interconnected rodlets. This sporoderm organization is similar to the fine structure of the megaspore wall reported in M. oedipternum (W. Taylor, 1990). Sigillariostrobus is a compressed cone believed to have been produced by sigillarian lycopsids. Many of the species described to date may represent preservational states of the Mazocarpon-type cone and have been correlated with the latter on the basis of size, organization of the sporophylls, and morphology of the spores. Some specimens of Sigillariostrobus may attain lengths of 30 cm. Despite the advances in recent years, there is still much to be learned about the biology of sigillarian plants. They represent one of the arborescent lycopsid groups that flourished during the Late Pennsylvanian. The sigillarians are believed to have inhabited near-swamp environments that were slightly drier than those dominated by the other arborescent lycopsids (DiMichele and Phillips, 1985). Other Lepidodendrid Genera

Figure 9.91 Transverse section of Mazocarpon oedipternum microsporangiate cone (Pennsylvanian). Bar  5 mm.

Some lycopsid stems have leaf bases that are inconspicuous or lacking, and these are sometimes difficult to distinguish from decortication stages. For example, in Bothrodendron (FIG. 9.92), an arborescent lycopsid that was 10 m tall, the leaf scars are not on raised cushions, but are flush with the surface of the stem (Wnuk, 1989). In some specimens,

308

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 9.92 Bothrodendron minutifolium, branched twigs (partially foliated) and a cone (arrow) (Pennsylvanian). Bar  2 cm. (Courtesy BSPG.)

Figure 9.93 Cross section of Paralycopodites (Anabathra) showing primary and secondary xylem, and departing branch trace (arrow) (Pennsylvanian). Bar  6.5 mm. (Courtesy W. A. DiMichele.)

the scars are arranged in an almost whorled pattern; in others, they are borne in a low helix. The ligulate microphylls are narrow and measure up to 25 cm in length. Cones, known from both compressions and petrifactions, consist of sporangia borne adaxially on sporophylls whose distal laminae are greatly elongated. The cones (Bothrodendrostrobus) are bisporangiate, with trilete microspores toward the apex and megasporangia toward the base, each with up to 20 megaspores. Megaspores of Bothrodendron have numerous spines that cover their surfaces. Paralycopodites (FIG. 9.93) differs from both Diaphorodendron and Lepidophloios in having reduced, persistent scalelike leaves that appear morphologically intermediate between these two genera (DiMichele, 1980). Stratigraphically, the genus extends from the Lower Carboniferous to the Middle Pennsylvanian (Westphalian D).

Paralycopodites has been reconstructed as a small tree that produced deciduous lateral branches. The diameter of the siphonostele in P. breviformis was 3 cm; the most distal branches were protostelic. Vegetative remains are consistently associated with strobili of the Flemingites type. Pearson (1986) has suggested that Paralycopodites is the same plant that was initially described by Witham (1833) as Anabathra. Another common lycopsid believed to be a progenitor of, or have affinities with, the lepidodendrids is Sublepidodendron songziense from the Upper Devonian of China (Q. Wang et al., 2002, 2003b). It has small, vertically oriented leaf bases arranged in a tight spiral arising from a stigmarian rhizomorph; branch scars are of the Ulodendron type. The stem has an ectophloic siphonostele with fibers in the pith. Strobili are of the Lepidostrobus type and borne at the tips of lateral branches (Q. Wang et al., 2003a). A second species in that genus, for which stems and branches (described as S. wusihense), and strobili (described as L. grabaui) have been reassembled from isolated parts, is Sublepidodendron grabaui from the Upper Devonian of the Wutung Formation (Famennian) of China (Y. Wang and Xu, 2005). A cladistic analysis of Sublepidodendron and related taxa suggests that members of this genus possess many synapomorphies with phylogenetically more advanced genera in the families Sigillariaceae, Lepidodendraceae, and Diaphorodendraceae, rather than the order Protolepidodendrales as previously thought (Q. Wang et al., 2003b). An interesting strobilar organization occurs in the Late Devonian genus Bisporangiostrobus. In this taxon, a branch dichotomizes and each dichotomy terminates in an eligulate, bisporangiate cone (Chitaley and McGregor, 1988). Sporangia are attached to the adaxial surface of sporophylls which are helically arranged on the cone axis. The cone is a diminutive Flemingites with apical microsporangia and basal megasporangia. Megaspores are 1 mm in diameter and of the Duosporites type, whereas the smaller microspores conform to the generic diagnosis of Geminospora. The affinities of this interesting cone are suggested as being close to Jurinodendron (Cyclostigma; see Doweld, 2001), a Late Devonian arborescent lycopsid that was up to 8 m tall. Jurinodendron stems often occur as decorticated axes of the Knorria preservation type (FIG. 9.43), and the plant is characterized by a distal crown of dichotomizing leafy branches. Underground organs are Stigmaria-like branch systems with appendages (Schweitzer, 1990). Sporophylls are arranged into distinct cones positioned terminally on ultimate branches (Chaloner, 1968). Jurinodendron appears to have been a widespread taxon, known from Upper Devonian and

CHAPTER 9

LYCOPHYTA

309

Figure 9.94 Suggested reconstruction of the distal branches

of Valmeyerodendron triangularifolium (Mississippian). (Based on Jennings, 1972; reproduced from Taylor and Taylor, 1993.)

Early Mississippian deposits throughout the world, including Germany (Mägdefrau, 1936), Bear Island, Spitsbergen (Schweitzer, 1969, 2006; Murašov and Mokin, 1979), China (Feng et al., 2004), Japan (Kimura et al., 1986), and the Himalayas (Pal, 1977). Leptophloeum rhombicum is an arborescent lycopsid which was distributed worldwide during the Late Devonian (see references in Q. Wang et al., 2005). Based on literature data and material from the Frasnian (Upper Devonian) Huangchiateng Formation of Hubei, China, Q. Wang et al. (2005) provided a reconstruction that depicts the plant as a monopodial tree, 10–25 m tall and 0.3–0.4 m wide at the base, which produced lateral branching systems by pseudomonopodial branching of the trunk, rather than by equal ramifications as formerly thought. Branches grew by means of isotomous dichotomies. The lateral underground portion is believed to have been Stigmaria like. Valmeyerodendron (FIG. 9.94) is a Mississippian form which is interpreted as transitional between the Devonian lycopsids, which generally lack leaf bases, and the

Figure 9.95 Suggested reconstruction of Eskdalia fimbriophylla leaf cushion (Mississippian). (From Taylor and Taylor, 1993.)

Carboniferous ones, which are ligulate and possess parichnos scars (Jennings, 1972). Only compressed specimens are known, and they include stems up to 3 cm in diameter. Helically arranged, quadrangular–hexagonal leaf cushions cover the stems. Each cushion bears a rhombic leaf scar at its apex; ligule and parichnos scars are absent. Unlike the majority of arborescent lycopsids, which possess narrow, linear leaves, Valmeyerodendron has leaves that are nearly triangular in outline with a constricted base and attenuated apex. Eskdalia is a Mississippian ligulate lycopsid with expanded leaf cushions (FIG. 9.95), each with a conspicuous keel (FIG. 9.95) (Thomas and Meyen, 1984a). Rowe (1988a) indicated that stomata are absent from the cushions in E. variabilis and suggested that they functioned in support, rather than as photosynthetic organs, as has been documented for certain arborescent lycopsids (Chaloner and Meyer-Berthaud, 1983).

310

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Lycopodiales The Lycopodiales includes homosporous, eligulate, usually dichotomously branched herbaceous plants. The group is represented today by four genera, Lycopodium (with 476 species; sometimes subdivided into two to several genera), Huperzia (439 species), Lycopodiella (41 species), and Phylloglossum (1 species). These are generally small herbaceous plants covered with scalelike microphylls; they have true roots, which arise adventitiously from a horizontal rhizome. All are homosporous and today the order is cosmopolitan in distribution. The fossil members are not well understood, and although the group itself is ancient, there is some ambiguity as to whether some extant members are of recent or ancient origin (Wikström and Kenrick, 2001). The generic name Lycopodites was first used to describe some Cenozoic axes bearing small, scalelike leaves. The fossils were later determined to be fragments of conifer shoots. Today the morphogenus includes axes with helically arranged or pseudowhorled scale leaves and, if present, sporangia on the adaxial leaf surface, in the axil of foliage leaves, or in monosporangiate strobili. Specimens of Lycopodites (FIG. 9.96) have been described from rocks ranging from Devonian to Pleistocene (Kräusel and Weyland, 1937; Harris, 1976a; Krassilov, 1978; Rigby, 1978b), and they include forms that are both isophyllous (one type of leaf) and anisophyllous (two types of leaves). There is a report of Lycopodites from the Paraná Basin in South America (Ricardi-Branco and Bernardes-de-Oliveira, 2002; Jasper et al., 2006) and if accurate, it represents the only occurrence of the genus in Gondwana, with the possible exception of the enigmatic Lycopodites amazonica that has been described from the Middle Devonian of Brazil (Dolianiti, 1967). One difficulty in dealing with fossils of the Lycopodites type is in distinguishing these remains from the distal twigs of members of Lepidodendrales. When sporangia are scattered along the stem in association with leaves resembling vegetative leaves, identification is easy, but most known specimens of Lycopodites consist only of vegetative remains. The absence of ligules is another feature that can be used to distinguish lepidodendrids, but even in exceptionally well-preserved specimens, these structures are often difficult to identify. In addition, as some extant species have strobili, it is conceivable that some of the small, apparently microsporangiate Lepidostrobus species may, in fact, represent a fossil herbaceous plant with a Lycopodium-type cone. The potential confusion in delimiting members of this group from the lepidodendrids will no doubt continue until structurally preserved specimens can be correlated with compression fossils

Figure 9.96

Lycopodites sp. (Pennsylvanian). Bar  2 cm.

or until more complete plants can be reconstructed based on compression specimens. Ultimately, epidermal features, including the distribution and type of stomata, may be useful in separating distal twigs of the arborescent lycopsids from axes of the herbaceous forms. Oxroadia gracilis includes small, dichotomously branched lycopsid axes that lack distinct leaf cushions but possess decurrent leaf bases (Alvin, 1965; Bateman, 1992). The genus is based on structurally preserved specimens from the Mississippian (Calciferous Sandstone Series, Scotland) and is thought to have represented an herbaceous lycopsid. The stem contains an exarch protostele with mesarch traces arranged in a helical manner. Microphylls are eligulate and vascularized by a single strand; parichnos is not present. Similar histology and association in the same block of material are used as the basis for assigning a small (4 cm long) cone to the stem remains. Sporangia are elongate and borne on sporophylls that have downward-projecting heels. A massive parenchymatous pad of tissue extends from the surface

CHAPTER 9

Figure 9.97 Synlycostrobus tyrmensis cone showing subtending bracts (Jurassic). Bar  5 mm. (Courtesy V. A. Krassilov.)

of the sporophyll and partially fills the sporangium cavity. Nothing is known about the spores. The genus is regarded as a herbaceous lycopsid, rather than the distal branches of an arborescent form, because there are no secondary tissues in the vascular system and cortex. Based on our current knowledge of stem development in the arborescent lycopsids, however, this distinction may be less certain. Synlycostrobus tyrmensis is an interesting lycopsid from the Late Jurassic to the Early Cretaceous of the Bureja Basin (Siberia) that has an unusual arrangement of cones (Krassilov, 1978). It is thought to have been a creeping plant, probably not unlike modern Lycopodium. The ligulate leaves are scalelike and anisophyllous. Cones are borne on what have been termed fertile shoots (FIG. 9.97), each located in the axil of a scalelike leaf or bract. The cones are small (5 mm long) and consist of 20 helically arranged sporophylls. Each sporophyll has a conspicuous distal lamina and a downward-projecting heel that partially covers the sporangium below. Only sporangia containing radial, trilete spores (20–22 μm in diameter) have been recovered from the cones, although a single megasporangium with four spores was isolated from the same matrix. Superficially, the fertile branches of Synlycostrobus resemble the primary axis and dwarf shoots (cones) of the

LYCOPHYTA

311

Cordaitales (Chapter 20). Associated in the same rocks are well-preserved vegetative axes that are placed in the genus Lycopodites. Although morphologic and cuticular features suggest affinities with that taxon, the presence of a ligule associated with each leaf may warrant inclusion within the Selaginellales (Skog and Hill, 1992). Onychiopsis psilotoides is a fossil from the Lower Cretaceous (Wealden) of England with a complex nomenclatural history. Originally thought to have affinities within the polypodiaceous ferns, Skog (1986) reinterpreted some specimens as herbaceous lycopsids and transferred these to a new genus, Tanydorus (Skog, 1986). She later noted, however (Skog, 1990), that the type specimen for O. psilotoides was missing and thus was unable to confirm that it conformed to the genus Tanydorus. Skog (1990) therefore designated a new type and named the lycopsid taxon Wathenia. This genus is characterized by helically arranged, decurrent, simple leaves, each with a single vein. Leaves have acute tips and sporangia occur in the axil of the leaves. Spores are trilete and foveolate (ornamented with small pits). Onychiopsis remains a genus for fossil ferns and is discussed in Chapter 11. Small (3 mm in diameter), eligulate permineralized cones from the Middle Pennsylvanian of Kansas are also thought to have been produced by herbaceous lycopsids (Baxter, 1971a). Carinostrobus has an exarch protostele and helically arranged sporophylls. The sporangium was attached to the adaxial surface by a delicate pedicel; spores were small (20– 22 μm), trilete, and covered by minute spines. Another permineralized Carbonifereous cone in which the sporangium is attached by a short stalk is Spencerites (Berridge, 1905) (FIG. 9.98). Spores of S. moorei are triangular in outline and characterized by an equatorial bladder that is also triangular in outline (Leisman, 1962; Leisman and Stidd, 1967). It is difficult to determine whether cones with Spencersporites spores are mono- or bisporangiate. Several types of fossil lycopsids are known from Gondwana. One of the better known taxa is Bumbudendron, a small, eligulate lycopsid with stems up to 3.5 cm wide (Archangelsky et al., 1981a). The specimens come from the Paganzo Basin in west-central Argentina and are Pennsylvanian in age. Leaf cushions are helically arranged with the leaf trace in the upper third of the cushion. Beneath the trace is an elongate structure that represents an infrafoliar bladder, a small depression or elongate mound just beneath the vascular-bundle scar. Topographically, it occupies the position of the two parichnos scars that characterize Diaphorodendron leaf cushions. The infrafoliar bladder is present on eligulate upper Paleozoic lycopsids and has been used as a systematic character in Thomas and Meyen’s

312

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 9.98 Emily M. Berridge.

(1984b) classification of impression–compression lycopsid stem genera. Fertile branches in the same rocks that contain Bumbudendron consist of reflexed, keeled sporophylls. From the same rocks these authors also described Malanzania. The stems of this lycopsid are narrow (1.3 cm wide) and contain widely spaced spines. Brasilodendron is an arborescent Permian lycopsid bearing persistent leaves up to 4 cm long (Chaloner et al., 1979; Jasper et al., 2006). On the abaxial surface of the leaf are two stomatal bands containing numerous, sunken stomata. Leaf bases are fusiform and no ligule has been observed. Several megaspores were recovered from the same matrix as the vegetative stems. They range from 800–1340 μm in diameter and are most similar to Lagenoisporites brasiliensis. Based on vegetative remains, Brasilodendron appears most similar to Ulodendron. As no ligule or ligule scar has been observed, Brasilodendron is discussed here in the Lycopodiales rather than with the other arborescent forms in the Lepidodendrales. Absence is not definitive evidence, however, as ligules may not have been seen due to preservation or to the persistent nature of the leaves. Brasilodendron is also comparable to Azaniadendron, an Early Permian, presumably eligulate, lycopsid from South Africa (Rayner, 1986). The vegetative axes in this taxon exhibit circular–oval leaf cushions with an elongate vascular-bundle scar. Cones are bisporangiate; each sporangium contains a single tetrad of megaspores (2 mm in diameter) of the Triletes type and microspores assignable to Zinjisporites.

Figure 9.99 Barry A. Thomas.

One of the few permineralized lycopsids known from late Paleozoic rocks of Gondwana is Eligodendron (Archangelsky and de la Sota, 1966). The single specimen comes from the Permian of Bolivia and possesses a parenchymatous pith and exarch primary xylem. The cortex is three-parted with large lacunae in the inner zone. The specimen lacks evidence of ligules and parichnos tissue.

Selaginellales Another herbaceous group that coexisted with the Carboniferous arborescent lycopsids is the Selaginellales (Carboniferous evidence reviewed in Thomas, 1997) (FIG. 9.99), today represented by a single genus (Selaginella) that includes around 500 species. Plants assigned to this order are herbaceous, ligulate, and heterosporous (FIGS. 9.100, 9.101), and are characterized by a creeping to erect habit. Modern members generally show leaves in four ranks, with two ranks of smaller leaves (anisophylly). Megagametophytes exhibit endosporic development (FIG. 9.102). Some extant species are capable of surviving extended periods of drought, for example S. lepidophylla,

CHAPTER 9

Figure 9.100 Selaginella selaginoides (Extant). (Courtesy

H. Bültmann.)

Figure 9.101 Partial longitudinal section of Selaginella cone showing microsporangium (left) and megasporangium (right) (Extant). Bar  1 mm.

LYCOPHYTA

313

the so-called resurrection plant. Recent studies based on gene sequence data suggest that the xeric and woodland species evolved from ancestors in the humid tropics, but today there is less resolution in classification schemes than there was earlier understood (Korall et al., 1999; Korall and Kenrick, 2004). The best-known fossil member, Selaginella fraipontii, represents an excellent example of the reconstruction of a complete plant based on isolated organs. For many years the generic name Paurodendron has been used for small (4 mm in diameter), anatomically preserved stems that are relatively common in coal balls from certain localities (Fry, 1954). The axes bear helically arranged, ligulate microphylls and are characterized by exarch protosteles that are stellate in cross section. The underground portion was subsequently discovered attached to a Paurodendron axis (Phillips and Leisman, 1966). It consists of an unbranched and unlobed, clavate rhizophore (root-bearing organ) from which helically arranged, monarch roots arose (FIG. 9.103). Despite the small size of the rhizophore stele, some secondary xylem is noted (FIG. 9.104). Reproductive parts of the plant are bisporangiate cones that were initially described under the binomial Selaginellites crassicinctus (FIG. 9.105). Cones of this type are 1.2 cm long and 5 mm in diameter. Sporophylls are ligulate and attached to the axis in alternating verticils (whorls), each with four sporophylls. Megasporangia are restricted to the basal region of the cone, with each sporangium containing four or occasionally up to seven megaspores of the Triletes type. Microspores are assignable to the sporae dispersae genus Cirratriradites. Ultrastructural features of the megaspores are shared by modern members of both the Selaginellales and Isoetales (Taylor and Taylor, 1988, 1990). There have been a number of studies that have focused on determining the developmental processes involved in the formation of the complex spore wall (Hemsley et al., 1994). Demonstration of organic attachment of Selaginellites cones to Paurodendron axes has made it possible to reconstruct the entire plant, which is now referred to as S. fraipontii (Schlanker and Leisman, 1969). The plant is reconstructed as herbaceous, sprawling, and sparsely branched; it produced cones terminally. There is some suggestion that the S. fraipontii was determinate in growth, much like arborescent lycopsids. The species is known throughout the Carboniferous, and is almost identical morphologically with many of the Selaginella species that inhabit relatively moist environments today. Extant Selaginella and the fossil differ, however, in the organization of their underground parts. In living Selaginella, roots are primarily adventitious, whereas in S. fraipontii, they are formed between adjacent older roots, resulting in a specific pattern of root formation. The

314

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Strobilus Microspore mother cells undergo meiosis, producing Ligule Young sporophyte

Microsporangium Sporophyte Megaspore mother cells undergo meiosis, producing Zygote 2n

Fertilization

Meiosis n Megaspore Megasporophyll Megaspores Archegonium containing egg

Sperm Microsporophyll

Female gametophyte

Microspores

Microspore Male gametophyte

Figure 9.102 Life history of Selaginella (Extant). (From Taylor and Taylor, 1993.)

organized production of laterals is a feature commonly associated with stems, not roots. The occurrence of this feature in S. fraipontii suggests homologies with the lobed rooting systems of a number of Paleozoic and Mesozoic lycopsids and may perhaps indicate a closer affinity with the Isoetales (Rothwell and Erwin, 1985; Bateman et al., 1992), or an intermediate status between the Selaginellales and Isoetales (Bateman, 1990). Modern-appearing Selaginella-like axes from the Rhaetian (Late Triassic) of Scania have been described as Selaginellites (Lundblad, 1950a). There are also numerous impression–compression specimens of anisophyllous lycopsids from Carboniferous rocks that have been described as either Selaginellites or placed in the extant genus Selaginella (Thomas 1992, 1997). Some of these, such as Selaginellites gutbieri, are represented as exquisite compressions that show the attachment of bisporangiate cones and details of

the microphylls (Rössler and Buschmann, 1994). In this species, there appear to be at least six to seven megaspores of the Triangulatisporites type in each megasporangium; microspores can be assigned to Cirratriradites. Isolated S. gutbieri bisporangiate cones have also been reported from Late Pennsylvanian deposits in the Czech Republic containing these same dispersed spore taxa (Bek et al., 2001). In Selaginellites primaevus (Selaginella primaeva of Thomas, 1997), from the roof shales in the Saar Coalfield, each megasporangium contains four megaspores of the Triangulatisporites type. Several small lycopsid cones from the Famennian (Upper Devonian) of Belgium exhibit an interesting collection of features unlike those of other known Devonian lycopsid cones (Fairon-Demaret, 1977). Barsostrobus cones are up to 14 cm long and bear helically arranged sporophylls and stalked sporangia. The sporophyll margins are evenly toothed, with the margins slightly enveloping the sporangium. The vascular

CHAPTER 9

R

LYCOPHYTA

315

X

Figure 9.103 Suggested reconstruction of Selaginella fraipontii (Pennsylvanian). (Based on Phillips and Leisman, 1966; reproduced from Taylor and Taylor, 1993.)

system is that typical of lycopsid cones, and the traces to the sporophylls have centrifugal and centripetal metaxylem. Spores are 240–320 μm in diameter, trilete, and evenly ornamented. The cones are thought to have been heterosporous, although no microspores have been discovered. Features of this cone suggest affinities with members of the Lycopodiales or Selaginellales; preservation prevents the recognition of ligules. The presence of Williamson striations between the scalariform bars of the metaxylem tracheids resembles some species of Drepanophycus; the Drepanophycales, however, lack heterospory and a strobilar organization of sporangia. Minostrobus chaohuensis is a Late Devonian cone from China that is believed to have been borne by a herbaceous lycopsid. There are four megaspores per sporangium with spores of the Lagenicula type (Y. Wang, 2001). Yuguangia ordinata is a ligulate lycopsid from the Givetian of China with bisporangiate terminal cones (Hao et al., 2007). Leaves of this permineralized form are spiny and arranged in pseudowhorls; megasporophylls contain four megaspores of the Triletes type, whereas the microspores are similar to Acinosporites. The occurrence of Y. ordinata suggests that small heterosporous, ligulate lycopsids had diverged from

Figure 9.104 Longitudinal section of Selaginella fraipontii

(Paurodendron) showing rhizomorph (R), secondary xylem (X) and numerous roots (Pennsylvanian). Bar  3 mm.

the homosporous ligulate grade by the Middle Devonian (Hao et al., 2007). Miadesmia membranacea (FIG. 9.106) is used for isolated cones known only from the Carboniferous of Europe. In some classifications of lycopsids, Miadesmia is included in its own order, Miadesmiales (Thomas and Brack-Hanes, 1984); but as nothing is known about the plant that bore this cone type, we will continue to include it within the Selaginellales until more information is available. The cones contain only megasporophylls which are attached to the axis at right angles (Benson, 1908). Each megasporophyll is 3 mm long and bears a megasporangium that is attached near the proximal end of the sporophyll. Lateral laminae completely envelop the sporangium, except in the distal region. The enveloping sporophyll is divided into elongate,

316

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

MI

tentacle-like extensions that project beyond the distal opening. The sporangium is somewhat flattened on the sporophyll so that the opening is directed away from the cone axis. A large ligule is present just distal to each megasporangium. Miadesmia is interesting in that the sporangium contains one large, functional megaspore and some are known with a cellular megagametophyte. In the original description, it was noted that M. membranacea occurs in the same coal balls as specimens of lepidodendrids. The small size of the sporophyll units has suggested to some that Miadesmia may represent a ligulate, heterosporous cone type within the Selaginellales that parallels the highly developed heterospory in the Lepidodendrales. This theory is based on the presence of one functional megaspore per megasporangium in each group and the integument-like morphology of the lateral laminae.

Pleuromeiales

ME

Figure 9.105 Longitudinal section of Selaginellites crassicinc-

tus showing both microsporangia (MI) and megasporangia (ME) (Pennsylvanian). Bar  2 cm.

Figure 9.106 Longitudinal section of Miadesmia membranacea (Pennsylvanian). Bar  1 cm.

The Pleuromeiales are an interesting group of lycopsids, and were originally thought to represent an evolutionary transition from fossil arborescent forms to herbaceous, cormose forms such as extant Isoetes. This hypothesis has been questioned, however, based on presumed Isoetes fossils in the Triassic (Retallack, 1997) and the discovery of cormose forms in the Paleozoic. The pleuromeialeans were smaller than the Lepidodendrales and believed to have been herbaceous or pseudoherbaceous. Plants with a similar growth habit have been described from the Paleozoic, for example Chaloneria (discussed below). Members of the Pleuromeiales have been included in the Lepidodendrales as well as the Isoetales, as they share features with both groups. They are generally unbranched and bear one or more terminal cones; their rooting structures are lobed, cormose, and bear stigmarian-type appendages. Some bear bisporangiate cones, whereas only one size of spore is known from other cones. Sporangia are generally somewhat sunken into the sporophyll. Most have trilete megaspores and monolete microspores like the isoetaleans, although Neuberg (1960b) showed that Pleuromeia rossica contained trilete microspores (Pigg, 1992). Pleuromeia is an exclusively Triassic genus known from localities around the world, including Germany, France, Spain, Russia, China, Japan, Argentina, and Australia. The discovery of Pleuromeia at many different localities throughout the world suggests that the plant may have

CHAPTER 9

inhabited varying habitats. For example, numerous specimens of Pleuromeia occur in Lower Triassic beds north of Sydney, Australia, and Retallack (1975) suggested, based on a detailed analysis of the lithology of the fossil beds, that the plant grew as a coastal halophyte. Specimens from China are believed to have grown in more xeric, inland sites, perhaps near desert oases (Z.-Q.Wang and Wang, 1982), and still others from Australia in habitats that were seasonally wet (Cantrill and Webb, 1998). From the Buntsandstein (Lower Triassic) of the Eifel region in Germany, Fuchs et al. (1991) described and illustrated spectacular large slabs with numerous densely spaced P. sternbergii stems in situ, which indicate that Pleuromeia formed extensive bitypic stands with the fern Anomopteris mougeotii. The seemingly sudden abundance of Pleuromeia in the Early Triassic is remarkable. Looy et al. (1999) suggested that the hypothesized dieback of woody vegetation at the very end of the Permian dramatically affected terrestrial ecosystems, and that lycopsids such as Pleuromeia played a central role in repopulating certain landscapes after the mass extinction. Pleuromeia has an unbranched, erect trunk, up to 2 m tall, and a four-lobed base (rhizomorph) from which helically arranged roots arise (FIG. 9.107). At the apex of P. longicaulis is a crown of elongate ligulate leaves, each with two vascular bundles (Retallack, 1975). Slightly below the attached leaves is a zone of persistent leaf bases that grades into an area of widely separated leaf scars. Decortication stages suggest that there was some secondary tissue production in Pleuromeia, although the absence of well-preserved petrified specimens makes it impossible to determine whether these tissues were vascular or cortical in origin. Reports by Mägdefrau (1931), Hirmer (1933b), and Grauvogel-Stamm (1993, 1999), among others, indicated that the axes of P. sternbergii, the type species of the genus from the Lower Triassic of Germany, were covered in leaf bases to near the base of the trunk (FIG. 9.124). This plant bore leaves of two types in a lax (subhorizontal) position. The rhizomorph is lobed and is characterized by a bilateral furrow system that divides the base typically into four lobes. The pattern of development of the rootlets appears to be like that in Isoetes. A species similar to P. sternbergii, P. obrutschewii (FIG. 9.108), has been reported from the Lower Triassic of the Russian Far East (Krassilov and Zakharov, 1975). At the apex of Pleuromeia is a single, relatively large cone, although it is possible that some species could have produced more than one cone. Support for this theory comes from the occurrence of large numbers of the small cones

LYCOPHYTA

317

Suggested reconstruction of Pleuromeia longicaulis (Triassic). (From Retallack, 1975.)

Figure 9.107

called Cylostrobus (FIG. 9.109) (sometimes misspelled as Cyclostrobus), which is thought to have been produced by Pleuromeia, at the same locality (Helby and Martin, 1965). Earlier reports, based on fragments of cones, suggested that Pleuromeia was a dioecious plant in which microsporangiate and megasporangiate cones were produced on different

318

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

9.108 Megasporophyll (arrow) and megasporangium of Pleuromeia obrutschewii containing casts of megaspores (Triassic). Bar  2 cm. (From Krassilov and Zakharov, 1975.)

Figure

Figure 9.109 Cylostrobus sp. (Triassic). Bar  1 cm. (Courtesy

D. Cantrill.)

plants. It is now known that some species of Cylostrobus were bisporangiate, with the megasporangia located in the basal portion of the cone (Helby and Martin, 1965). The sporophylls are circular, imbricate, and lack any downward extension in the form of a heel. The large megaspores ( 700 μm) are trilete and ornamented with numerous elongate spines; microspores are monolete and 30 μm in diameter. One bisporangiate cone is C. clavatus from the Early Triassic of Australia (Cantrill and Webb, 1998). Microspores are of the Lundbladispora–Aratrisporites type. Krassilov and Zakharov (1975) have suggested that megasporophylls

Figure 9.110 Compressed cone of Pleuromeia epicharis

(Triassic). Bar  2 cm. (From Z.-Q. Wang and Wang, 1990.)

of pleuromeids were dispersed by water, based on their shape and their abundance in the rocks. This hypothesis parallels that of Phillips (1979) for the dispersal of megasporophylls of many Carboniferous lycopsids (see section “Conclusions”).

CHAPTER 9

LYCOPHYTA

319

Figure 9.111 Fractured cross section of Pleuromeia epicharis

cone (Triassic). Bar  2 cm. (Courtesy Z. Wang.) Figure 9.112 Compressed cone of Skilliostrobus (Triassic).

Bar  1 mm. (From Ash, 1979.)

Other species of Pleuromeia are smaller (Meng, 1996). Pleuromeia jiaochengensis was 50 cm tall (Z.-Q. Wang and Wang, 1982). Specimens include compressions that come from the Early Triassic of Shanxi Province, China. On the unbranched stem are leaf scars to which were attached awl-shaped leaves 3 mm long. At the distal end of the stem is a large cone constructed of obovate sporophylls bearing discoid sporangia. Only megaspores are recorded, and these range up to 500 μm in diameter. Pleuromeia epicharis (FIG. 9.110) is one of the better-known species and is based on material from the Shiqianfeng Group in north China (Wang and Wang, 1990). The specimens include stems, leaves, basal parts, and micro- and megasporangiate cones (FIG. 9.111). Megaspores are of the Banksisporites type. Specimens of P. rossica were transferred to the new genus, Lycomeia (Dobruskina, 1985a). Microspores and megaspores of L. rossica reveal (ultra-)structural characteristics regarded as distinctive of the spores of Isoetales, which adds support to the hypothesis suggesting a close relationship between Pleuromeia and Isoetes (Lugardon et al., 1999, 2000). More recently, Grauvogel-Stamm and Lugardon (2004) have also demonstrated that the spores of P. sternbergii have a number of features characteristic of isoetalean spores. Some workers have included Pleuromeia longicaulis in a new genus, Cylomeia (White, 1981a). It is believed that this plant produced terminal cones of the Skilliostrobus type (FIG. 9.112) (Ash, 1979). This Early Triassic, pedunculate, bisporangiate cone is known from Australia and Tasmania.

It consists of helically arranged, wedge-shaped sporophylls with an adaxial groove containing obovate sporangia. The cone is up to 8 cm in diameter and only about half as long. Microspores are monolete (40 μm) and most similar to Aratrisporites, whereas the trilete megaspores range up to 1.1 mm and are most similar to Horstisporites. Petrified Triassic stems of Chinlea from North America were initially thought to represent osmundaceous ferns, but they are now regarded as lycopsids, possibly related to Pleuromeia (Miller, 1968). The stems contain an ectophloic siphonostele with a distinct perimedullary zone of thinwalled parenchyma. Leaf traces are numerous (up to 165 in one transverse section) and collateral. Pleurocaulis rewanense is a protostelic stem with circular to oval scars that may represent a decorticated stem stage (Cantrill and Webb, 1998). Ferganodendron is a Triassic genus that resembles Pleuromeia and Sigillaria in many respects (Dobruskina, 1974). The trunk of the plant varies from 20 to 30 cm in diameter and is covered with numerous, elliptical–rhombohedral leaf bases that are helically arranged. The leaves are small and are found only on the more distal portions of the plant. Nothing is known about the internal structure or reproductive parts. The genus name Lycostrobus is used for isolated, bisporangiate cones which are known from several Triassic deposits and thought to be allied with the Pleuromeiales, based partly on the presence of monolete microspores. The basic construction of these cones is similar to that of a bisporangiate Flemingites with helically arranged sporophylls

320

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

and adaxial sporangia. In L. scottii, the microspores occur in groups, and it has been suggested that the grouping may have been the result of sporangial trabeculae (partitions) that were not preserved. Lycostrobus chinleana, from the Triassic of Arizona, is now considered to be a member of the Equisetales (see Chapter 10). Annalepis is another Triassic lycopsid that was initially described based on isolated sporophylls (FIG. 9.113), but articulated cones upto 10 cm in diameter have more recently been reported. Specimens are known from China (Meng, 1998), and several sites in Europe (Kelber and Hansch, 1995; Kustatscher et al., 2004). Sporangia were adaxial and there is some suggestion that lateral laminae may have partially enclosed the sporangia (Grauvogel-Stamm and Lugardon, 2001). Microspores contained in A. zeilleri sporophylls are assignable to the dispersed spore genera Aratrisporites and Tenellisporites (Grauvogel-Stamm and Duringer, 1983). Although the parts are disarticulated,

Figure 9.113 Annalepis zeilleri, sporophyll with megaspores

(Triassic). Bar  1 cm. (Courtesy K.-P. Kelber.)

they suggest that Annalepis had a similar growth habit as Pleuromeia. Isolated lycopsid-like sporophylls from the Voltzia Sandstone (Middle Triassic) are called Bustia ludovici (FIG. 9.125) (Grauvogel-Stamm, 1991). The distal lamina is narrow and up to 3.5 cm long and vascularized by two vascular bundles. Sporangia were large and characterized by internal trabeculae. Spores are of the Aulisporites type. Austrostrobus ornatum (Triassic of Argentina) is a large, structurally preserved lycopsid cone that is believed to represent a petrified Cylostrobus (Morbelli and Petriella, 1973). The two taxa differ only in the size of the megaspores, and that difference may simply represent a combination of preservational phenomena and cone development.

Isoetales Isoetaleans are ligulate, heterosporous, and mostly herbaceous in habit. One or two genera of living plants are included in the Isoetales: Isoetes, which has an extensive distribution of almost 200 species ranging from the tropics to the sub-Arctic, and Stylites, which includes two species restricted to the high Andes of Peru. Most authorities regard Stylites as simply another morphologic form of Isoetes. Isoetes is characterized by a short, squat stem (usually less than a few centimeters long) that produces helically arranged, monarch roots from the lower surface and elongated, ligulate leaves in a dense rosette from the upper portion. Both micro- and megasporangia are produced on the same plant. Microspores of Isoetes are bilateral and monolete; megaspores are radial and trilete, although many of the fossils placed in this order have trilete megaspores and microspores; sporangia have trabeculae (sing. trabecula), sterile plates of tissue that extend into the sporangium. The Isoetales is now recognized as having an extensive fossil history, probably dating back to the Devonian (Pigg, 1992, 2001); the earliest forms with a morphology similar to that seen in the extant Isoetes have been reported from the Triassic (reviewed in Skog and Hill, 1992; Srivastava et al., 2004). Fossils from the Jurassic (e.g., Isoetites rolandii; see Ash and Pigg, 1991) are morphologically similar to extant forms in the presence of a bilaterally symmetrical rhizomorph. The modern genus exhibits great morphological and genetic uniformity (Schuettpelz and Hoot, 2006). Several Late Devonian and Early Mississippian forms have been interpreted as representing early members of the isoetalean lineage. These include Clevelandodendron

CHAPTER 9

ohioensis, a compressed, almost entire lycopsid plant, 1.25 m tall, from the Cleveland Shale member of the Upper Devonian Ohio Shale (Chitaley and Pigg, 1996). This plant consists of an unbranched, slender, monopodial axis, up to 2 cm wide, arising from a base bearing thick appendages. Terminally the axis bears a bisporangiate strobilus (FIG. 9.114), 9 cm long and 6 cm wide. The decorticated stem surface shows helically arranged elongate leaf traces and laterally compressed, slender leaves along the stem margin. Megaspores obtained from the cone are trilete and laevigate, and lack a gula; microspores are trilete, indistinctly punctate, and may be assignable to the dispersed spore genera Calamospora or Punctatisporites. Clevelandodendron demonstrates that lycopsids with a habit similar to the

LYCOPHYTA

321

Carboniferous genera Chaloneria and Sporangiostrobus (discussed below), and the Triassic Pleuromeia and related forms (see above), were present as early as the Late Devonian. Meyen (1987) suggested that several, of what he termed satellite genera, should be included in the Isoetales. One of these is Tomiodendron (FIGS. 9.27, 9.115) an unbranched, protostelic plant that may have reached 30 cm in diameter and had elongate leaf cushions on the stem surface. A second lycopsid in this group is Wexfordia hookense from the uppermost Famennian (Upper Devonian) type locality at Sandeel Bay, County Wexford, in southeastern Ireland. This lycopsid has a forked axis (3 cm in diameter) (Matten, 1989). The permineralized specimens consist of axes with medullated steles; tracheids with scalariform secondary wall thickenings contain fimbrils between the bars. Subsequent research on W. hookense indicates that secondary xylem with uni- to biseriate vascular rays is present in mature axes of Wexfordia (Klavins, 2004). Leaf bases are oval and the crowded leaves are each about 4 mm long. Although initial reports suggested that Wexfordia hookense shared features with anatomically preserved axes of other lycopsids, perhaps including the Protolepidodendrales and Carboniferous Lepidodendrales, the study by Klavins

Figure 9.114 Strobilus of Clevelandodendron ohioensis show-

ing sporangia and attachment of sporophylls to cone axis (Devonian). Bar  2 mm. (From Chitaley and Pigg, 1996.)

Figure 9.115 Stem of Tomiodendron peruvianum showing leaf cushions (Mississippian). Bar  1.5 cm. (Courtesy H. W. Pfefferkorn.)

322

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

hypothesizes that Wexfordia was a small tree with anatomical features closer to members of the isoetalean lineage. A third plant with possible affinities in the order Isoetales is Otzinachsonia beerboweri from the Famennian (Upper Devonian) of north-central Pennsylvania (Cressler and Pfefferkorn, 2005). Stems are up to 10 cm wide and the stem base forms a characteristic, four-lobed cormose rhizomorph with rootlets and circular rootlet scars arranged in orthostichies. A transitional zone between the rhizomorph (basal portion of the stem) and the distal part of the stem lacks evidence of rootlet scars. Distal stem portions display spirally arranged leaf scars in parastichies, but no leaf cushions. Carboniferous lycopsids with suggested affinities in the isoetalean lineage include Chaloneria cormosa, one of the most completely known fossil lycopsids within the group. This plant is known from both vegetative and fertile remains from the Upper Pennsylvanian of North America (Pigg and Rothwell, 1979, 1983a, b). The unbranched plant was 2 m tall, with small, ligulate leaves helically arranged about the stem. Leaf cushions were not produced. The base of the plant was cormlike. A limited amount of secondary xylem surrounds an exarch protostele, and in some specimens a thin band of periderm is produced (Pigg and Rothwell, 1985). These authors have been able to correlate various levels of decortication in C. cormosa with axis surface features of Knorria (FIG. 9.42), Asolanus, Bothrodendron, Pinakodendron, Jurinodendron (FIG. 9.43), and Stigmaria. A compressed, cormose lycopsid base of Middle Pennsylvanian age named Cormophyton also exhibits some of the same features as Chaloneria (Pigg and Taylor, 1985). Pigg and Rothwell (1983a,b) placed Chaloneria in its own family, Chaloneriaceae, in which they also included Sporangiostrobus and Polysporia. Chaloneria cormosa is heterosporous, with the fertile region (10 cm long) consisting of alternating megaand microsporophylls. Sporangia contain trabeculae. Microspores are monosaccate and of the Endosporites type, whereas the megaspores can be compared to Valvisisporites. Megagametophytes of Chaloneria, some with archegonia, have also been described (Pigg and Rothwell, 1983b). Their structure and embryology are similar to those known in Bothrodendrostrobus, suggesting perhaps that the latter should be allied with the Isoetales rather than the Lepidodendrales (Stubblefield and Rothwell, 1981). A Pennsylvanian fructification that may be related to Chaloneria is Porostrobus. This bisporangiate cone (2.5 cm long) possesses megasporophylls with hairlike tips, and produced Setosisporites-type megaspores (750–1150 μm in diameter) and Densosporites-type microspores (Leary and

Figure 9.116 Stem surface of Bodeodendron hispanicum

showing leaf cushions (Pennsylvanian). Bar  1 cm. (Courtesy R. H. Wagner.)

Mickle, 1989). The distal end of the megaspore is extended into a structure termed a gula that has been interpreted as a germ tube (Jha and Tewari, 2006). Polysporia includes compressed cones with the same types of spores as those in Chaloneria (Grauvogel-Stamm and Langiaux, 1995), but this genus has also been reported as a permineralized specimen (DiMichele et al., 1979). Sporangiostrobus encompasses relatively large, mono- and bisporangiate lycopsid cones, known from both compression and anatomically preserved specimens (FIG. 9.117); they are characterized by a massive axis occupying more than half of the total cone diameter (Bode, 1928; Chaloner, 1956, 1962; Remy and Remy, 1975a; Bek, 1996). The basic organization of this cone is similar to that of other arborescent lycopsid fructifications, but differs in having large, triradiate megaspores characterized by a broad equatorial flange formed by fused and anastomosing hairs (Zonalesporites type) and triradiate microspores (FIG. 9.118) that exhibit an extended range of morphologic variability that includes such dispersed spore taxa as Densosporites, Radiizonates, Cingulizonates, and Vallatisporites, sometimes found in

CHAPTER 9

LYCOPHYTA

323

Figure 9.117 Cross section of microsporangium of Sporangiostrobus kansanensis showing parenchyma tissue (Pennsylvanian). Bar  1.5 mm. (From Leisman, 1970.) Figure 9.119 Sporangiostrobus axis showing sporophyll bases (Pennsylvanian). Bar  2 cm.

Figure 9.118 Several microspores macerated from the sporan-

gium of Sporangiostrobus cone (Pennsylvanian). Bar  50 μm.

the same sporangium (Bek and Straková, 1996; Bek and Opluštil, 1998). One species from the Middle Pennsylvanian of Kansas, S. kansanensis, measures 16 cm long and is nearly 12 cm in width (Leisman, 1970). Wagner and Spinner (1976) have demonstrated that Sporangiostrobus was the cone of Bodeodendron (FIG. 9.116) based on their constant association and similar morphology. Exceptionally well-preserved specimens of Sporangiostrobus of late Stephanian age from Puertollano, Spain (FIG. 9.119) occur in large numbers in a tuff band at the base of a coal seam (Wagner, 1989) (FIG. 9.120). Some cones exceed 30 cm in length and are hypothesized to have been produced at the tips of vegetative branches. Microsporangia appear to have been intermingled with megasporangia in the cones. Sporangiostrobus has also been correlated with vegetative remains of Omphalophloios (Brousmiche-Delcambre et al., 1995). Mesozoic representatives of the isoetalean lineage include Nathorstiana, a Cretaceous genus that possesses the same root arrangement and stelar morphology as living members

324

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 9.120

Robert H. Wagner. Figure 9.121 Suggested reconstruction of Nathorstiana arborea (Cretaceous). (From Delevoryas, 1962.)

of Isoetes (Richter, 1909, 1910; Karrfalt, 1984). The genus is known from numerous specimens representing various developmental stages of the plant. Mägdefrau’s (1932) reconstruction of Nathorstiana arborea depicts the plant as being 20 cm tall, with elongate, grasslike leaves attached to the stem apex on conspicuous bases (FIG. 9.121). The underground parts were bulbous and produced helically arranged roots at the lobed bottom of the plant. Although nothing is known about the internal structure of the plant or the method of reproduction, Karrfalt (1984) hypothesized the developmental pattern of the fossil based on his studies of living Isoetes (Karrfalt and Eggert, 1977a, b, 1978). A detailed morphological study of this type in which fossils are used in association with a living developmental model (Isoetes) represents an excellent way to infer homologies among the various lycopsid rooting structures and thus more accurately determine phylogenies within the group. Nathorstianella is another Mesozoic fossil that has historically been included in this order based on cast specimens from the Lower Cretaceous of Australia (Glaessner and Rao, 1955). Although it differs in size, several features suggest a close correspondence with Isoetes. Karrfalt (1986) confirmed the isoetalean relationship, indicating that N. babbagensis was divided into five basal lobes, each bearing roots arranged

like those in extant Isoetes. He believes that the basal corm supported an arborescent aerial axis. Several fossils that morphologically resemble modern Isoetes have been described from rocks as early as the Triassic, and the generic name Isoetites has been used for many of these (Pigg, 2001), whereas others have been assigned to the extant genus Isoetes (see Srivastava et al., 2004). Isoetites serratifolius is the name used for compressed sporophylls from the Triassic of India (Bose and Roy, 1964); they extend up to 6.6 cm in length and are characterized by a serrate margin. A single vascular bundle extends the length of the sporophyll. The position of the sporangium is indicated by an elongated impression near the base of the sporophyll. In another species of the same age, I. indicus, sporophylls with megasporangia were preserved. The sporophylls are slightly smaller than those of I. serratifolius. Megasporangia contain up to 1500 trilete spores ranging 285–430 μm in diameter. Isoetites rolandii occurs in the Middle Jurassic of Oregon and Idaho (Ash and Pigg, 1991). Because the specimens are preserved as both mold–casts and impression–compressions they offer an excellent example of how different preservation types

CHAPTER 9

LYCOPHYTA

325

can be used to reconstruct an entire plant. The plants are 10 cm tall and consisted of linear leaves, up to 8 cm long. Casts of sporophyll bases contain structures interpreted as megaspores. What is especially noteworthy is that this discovery represents the oldest occurrence of Isoetes-like plants in western North America. The authors also noted that the occurrence of Isoetites provided the first evidence for an aquatic to semiaquatic environment within the Coon Hollow Formation. Isoetites serratus is a well-known form from Upper Cretaceous rocks (Frontier Formation) in Wyoming (Brown, 1939). The plant consists of rosettes of narrow, spathulate leaves with serrate margins; the leaves arise from the edge of a round corm that is 1.3 cm in diameter. The upper surface of the sporophyll has two rows of rectangular cavities interpreted as either wrinkles or the remains of collapsed internal air sacs. At the base of each sporophyll are compressed elliptical sporangia that contain impressions of either megaspores or microspores. Dichotomously branched roots are attached to the corm base beneath the outer rosette of sporophylls. Isoetes ermayinensis is a Triassic isoetalean from China that includes both vegetative and reproductive material (Z.-Q. Wang, 1991). The plant is small, with leaves up to 7 cm long, each with an elongate, lacunate band on either side of the vascular bundle; the corm is unknown. Megaspores are of the Dijkstraisporites type (50–200 megaspores per sporangium), whereas the microspores are assignable to Aratrisporites. Mature megaspores in stages of germination morphologically resemble Laevigatisporites.

Putative lycopsids There are always fossils that remain impossible to place systematically. The enigmatic Devonian–Carboniferous plant Barinophyton is one of these. Members of the genus consist of alternately arranged naked branches with sporangia organized in two rows on laterally born (FIGS 9.122, 9.123), spikelike fructifications. Structurally preserved specimens of B. citrulliforme from the Devonian of New York include an exarch protostele (Brauer, 1980). The tracheids have a continuous secondary wall that is plicated into the cell lumen, simulating annular secondary wall thickenings. Between the plications, the wall contains numerous delicate pits, each with a recognizable border. The sporangiferous appendages are alternate and two-ranked; each appendage is recurved and bears one large sporangium on the concave surface within the curve. An unusual feature of Barinophyton is that several thousand microspores and 30 megaspores

Figure 9.122 Suggested reconstruction of Barinophyton citrulliforme showing two rows of sporangia in a strobilus-like organization. (From Kenrick and Crane, 1997a.)

occur together in the same sporangium. A similar condition exists in Protobarinophyton pennsylvanicum described from the Upper Devonian of New York (Brauer, 1981). The small spores range from 30–42 μm in diameter, whereas the large spores extend from 410 to 560 μm. In both Barinophyton (Taylor and Brauer, 1983) and Protobarinophyton (Cichan et al., 1984) the wall structure of the large and small spores is different. This indicates that the small spores do not merely represent aborted megaspores and that these two plants were, in fact, heterosporous. Spores of two different sizes have also been reported in the same sporangium of the progymnosperm, Archaeopteris (Medyanik, 1982) (Chapter 12). In A. latifolia the number of megaspores per sporangium is highly variable, ranging 8–20 (Chaloner and Pettitt, 1987).

326

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

A

S Figure 9.123 Portion of Barinophyton citrulliforme axis with

appendages (A) and sporangia (S) (Devonian). Bar  1.5 cm. (Courtesy D. F. Brauer.)

Conclusions Despite the enormous amount of information that has been accumulated about fossil lycopods, there are still gaps in our understanding of their evolution. During the Carboniferous in Euramerica the lycopsids were well represented by at least four major orders. These included the large arborescent forms (Lepidodendrales) (e.g., Diaphorodendron, Lepidodendron, and Lepidophloios) with stigmarian rooting structures; the smaller woody types (Isoetales) (e.g., Chaloneria) with cormose rooting organs; and the herbaceous, ligulate or eligulate taxa included in the Selaginellales and Lycopodiales, respectively. It is becoming increasingly clear that there were other Carboniferous lycopsids, such as Hizemodendron serratum,

that were pseudoherbaceous, but produced mono-sporangiate cones like those of the Lepidodendrales, suggesting that perhaps this type of morphology evolved via reduction of the arborescent habit (Bateman and DiMichele, 1991). Historically, many comparative morphologists and paleobotanists believed that the fossil record, although admittedly incomplete, provided sufficient evidence to suggest that Isoetes represented the end of a transformational series beginning in the Carboniferous. According to this idea, the extant plant Isoetes represented a Flemingites-type cone seated on a stunted, stigmarian base (Stewart, 1947). Proponents of this concept, called the “lycopsid reduction series” (originally “lycopod”), suggested that a sparsely branched, heterosporous lycopsid like Sigillaria was the starting point of the series (Potonié, 1894). With a reduction in the number of aerial branches, a contraction of the branched, stigmarian rooting system to a cormlike base, and a reduction in the overall size of the plant, a plant like the Triassic Pleuromeia (FIG. 9.124) might represent an intermediate stage. Continued reduction of the main axis would produce a plant similar to Nathorstiana and eventually, Isoetes. The discovery of Carboniferous lycopsids with cormlike bases, such as Chaloneria, suggests that the cormlike lycopsids did not evolve from the arborescent forms by reduction, but existed at the same time. Additional specimens of rooting organs from which developmental data can be inferred, together with information on developmental stages of the embryos of two permineralized taxa, also support the idea that the lycopsid reduction series is too simple an explanation of the complexity and diversity of the fossil lycopsids. The discovery of lycopsids with cormlike bases older than Sigillaria, however, provides evidence that stigmarian rooting structures and lobed, cormlike bases coexisted. For example, Protostigmaria (Mississippian) possesses a cormlike base, as does the Pennsylvanian taxon Chaloneria, as well as several Mesozoic lycopsids (e.g., Pleuromeia and Nathorstiana) and a Late Devonian compression form, Otzinachsonia (Cressler and Pfefferkorn, 2005). In all of these taxa for which detailed information about the reproductive parts is known, the cones are bisporangiate, as is the production of micro- and megasporophylls in Isoetes. The inclusion of developmental information about the underground parts of fossil lycopsids, such as Stigmaria (Rothwell, 1984) and Nathorstiana (Karrfalt, 1984), indicates that, although the morphology of the organs may appear to be quite different, the patterns of growth and development are comparable. In their summary of homologies in the lycopsids, Rothwell and Erwin (1985) regarded the rooting organ (rhizomorph) as a modified shoot

CHAPTER 9

Figure 9.124 Suggested reconstruction of Pleuromeia sternbergii (Triassic). (Courtesy C. V. Looy.)

system that is fundamentally unlike the primary root system in seed plants. These authors suggested that the branched, stigmarian system is the most primitive type of lycopsid rhizomorph, and that the rooting structures of plants that

LYCOPHYTA

327

we have included within the Lepidodendrales, Isoetales, and Pleuromeiales are all homologous. Rhizomorph development in Selaginellites (Paurodendron) fraipontii, currently considered to be a member of the Selaginellales, is thus interpreted as being similar to that in Isoetes (Rothwell and Erwin, 1985). One fundamental difference between the Selaginellales and Lycopodiales and the other lycopsids concerns the nature of their rooting structures. Plants in these two orders are characterized by the production of adventitious roots, which are developmentally different from rhizomorphic types. Based on the single character of the anatomy and development of the rooting structures, Selaginellites fraipontii is mostly closely allied with the Isoetales. Although the morphology of the underground parts has been used to suggest homologies among lycopsid genera, it may more accurately reflect the ecological conditions of the plant. Detailed paleoecological information about the habitat and substrates of certain types of lycopsids is also contributing to a better understanding of the groups. Although many of the large arborescent lycopsids with stigmarian rooting structures are found in coal-swamp environments, some taxa with lobed bases have been recovered from rocks more indicative of levees, deltas, or sandy, lakeshore environments (Jennings et al., 1983). It has been suggested that the extinction of arborescent taxa with extensive stigmarian rooting systems may have come about through the loss of suitable swampy environments in the Late Pennsylvanian, whereas the smaller plants with cormose bases were perhaps better able to adapt as the swamp habitats shrank or disappeared. The reproductive strategies of fossil lycopsids are also important when considering the evolution and extinction of certain taxa. Phillips (1979) suggested that the morphology of the megasporophyll of Lepidocarpon and, to a lesser extent, of Achlamydocarpon made the units highly efficient, floating dispersal structures within the swamp communities. Another suggestion for dispersal is offered by Thomas (1981b), who believes that the winglike morphology of the megasporophylls made them adapted to wind dissemination. Both of these hypotheses may be correct. Experiments using reconstructions of Lepidocarpon suggest that the principal mode of dispersal may have been as suggested by Thomas, but that the ability of these structures to float in the swamp environment aided fertilization (Habgood et al., 1998). Although the lepidocarps were efficient dispersal agents in a constant aquatic environment, competing plants that produced an increased number of smaller, perhaps wind-dispersed megaspores may have been better able to survive in an environment with fluctuating water levels and to colonize

328

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 9.125 Bustia ludovici sporophyll (Triassic). Bar  5 mm.

(Courtesy L. Grauvogel-Stamm.)

newly emerging habitats. We still lack information about the frequency with which reproductive units were produced in these plants. Some have suggested that a plant like Chaloneria produced sporangia in a seasonal manner. Other

taxa, like Lepidophloios, may have been monocarpic and coexisted with species of Diaphorodendron, which produced propagules over an extended period of time (DiMichele, 1981). Continued specialization in reproductive strategies of the arborescent lycopsids culminated in the production of propagules, like Lepidocarpon, with a single large megaspore that included extensive food reserves. This development would have increased the ability of the plant to delay germination until the optimal environmental conditions occurred. Most agree that the lycopsids had their origin within the zosterophyll complex, and it is within this group, as well as in the Drepanophycales, that stages in the evolution of the microphyll can be found. The origin and subsequent evolution of major lycopsid families or orders continue to remain less resolved. The fossil record of the group during the Paleozoic is extensive and, as our knowledge of additional plants and their reproductive biology increases, traditional lineages are often difficult to separate because several taxa show unusual combinations of characters or characters that are shared with more than one group. By comparison, the systematics and diversity of lycopsids in the Mesozoic and Cenozoic are very poorly known even in instances in which reproductive organs have been discovered (e.g., Bustia; Grauvogel-Stamm, 1991) (FIG. 9.125). The major tasks in understanding lycopsid biology and evolution in the future will be to better resolve character states in some of the poorly understood taxa and to continue to test phylogenetic hypotheses. This will require more than merely the manipulation of data sets based on existing specimens, but rather will require a new focus on the discovery of additional, more informative specimens. The desired result of additional field work is the establishment of more fossil exemplar taxa that can be incorporated into phylogenetic analyses that use both extant and fossil species (DiMichele and Bateman, 1996). The continuing challenge in understanding lycopsid evolution is to comprehend these plants as members of dynamic ecosystems through time (DiMichele et al., 2001b).

10 SPHENOPHYTES PSEUDOBORNIALES ................................................................... 331

EQUISETALES ................................................................................ 342

Sphenophyllales................................................................... 332

Calamitaceae .................................................................................... 343 Tchernoviaceae and Gondwanostachyaceae .................................... 368

Devonian Sphenophyllales ................................................................333

Equisetaceae ..................................................................................... 371

Sphenophyllum ..................................................................................334

Forms with Uncertain Affinities....................................................... 376

Other Sphenophyllales ......................................................................338

SPHENOPHYTE EVOLUTION .................................................379

Ecology .............................................................................................341

It’s old equisetum, horsetail, out of dioramas at museums where in giant form it companied the treeferns, back when two-foot dragonflies rattled wings against horsetail’s glassy stems now long since coal. John Caddy, Morning Earth Poems

The geologic history of the sphenophytes (Sphenophyta) closely parallels the pattern of evolution and diversification exhibited by the lycophytes. Sphenophytes were first encountered in the Devonian, but attained their maximum diversity during the Carboniferous. From the Carboniferous to the recent, the group has experienced a gradual decline, until today they are represented by only ⬃15 species in a single genus, Equisetum (horsetails) (FIG. 10.1). In some treatments, the genus Equisetum is subdivided into two subgenera, subg. Equisetum and subg. Hippochaete, based on stomatal position and stem branching (Hauke, 1963; Des Marais et al., 2005). Although sphenophytes, along with ferns, are traditionally regarded as paraphyletic successive grades of increasing complexity between bryophytes and seed plants (Rothwell, 1999), some phylogenetic analyses (Pryer et al., 2001; Wikström and Pryer, 2005) suggest that sphenophytes and ferns are a monophyletic group (see also Nishiyama, 2007). Members of the Sphenophyta, sometimes termed the Arthrophyta or Equisetophyta, and in older treatments – Sphenopsida (Neuberg, 1964) are typically characterized

by clonal growth in the form of extensive, below ground rhizome systems from which arise adventitious roots and upright (also called aerial or orthotropous) axes or shoots with monopodial branching (FIG. 10.2). There are several reports of endomycorrhizae (AM) in the roots of modern Equisetum (Koske et al., 1985; Dhillion, 1993), but these are at best facultative associations (Read et al., 2000; Brundrett, 2002). Both rhizomatous and upright axes in the sphenophytes have distinct nodes and internodes, and the upright axes produce whorls of leaves or branches at the nodes. Regularly spaced, longitudinal ribs and furrows ornament the internodal regions. Leaves of modern Equisetum are typically small and thought to represent modifications of branching systems (Rutishauser, 1999), a hypothesis which is supported by the fossil record of the group. The vascular cylinder ranges from protostelic to siphonostelic, with primary xylem maturation either exarch or endarch. Secondary tissues are present in a few taxa, most notably some Carboniferous members of the Equisetales and the Devonian genus Pseudobornia. The reproductive organs are usually loosely arranged strobili or cones that consist of

329

330

PaleoBOtany: the biology and evolution of fossil plants

Strobilus

Strobilus Sporangiophore Branches

Sporophyte Sporangium Rhizome Root Gametophyte

Sporophyte Spore mother cells undergo meiosis, producing

Zygote

2n Meiosis

Fertilization n Sperm Spores

Antheridium Egg Archegonium Gametophyte

Figure 10.1 Life history of Equisetum. (From Taylor and Taylor, 1993.)

a central axis bearing whorls of modified branches, termed sporangiophores, that bear recurved, thick-walled sporangia. Some of the spores in this group are characterized by an extra-exinous layer that, in some genera, is elaborated into elongate structures termed elaters. All extant species are homosporous (Duckett, 1970), with the spores developing into unisexual gametophytes that produce either antheridia or archegonia. Sexual differentiation of the gametophytes is not genetically determined but is influenced by environmental conditions (Guillon and Raquin, 2002; Guillon and Fievet, 2003). Most fossil sphenophytes were homosporous, although a few heterosporous members are known; heterospory was not as extensively developed in the Sphenophyta

as in the Lycophyta. In this book, the Sphenophyta includes three orders:

Higher taxa in this chapter:

Pseudoborniales ( Devonian) Sphenophyllales (Devonian–Triassic) Equisetales (Devonian–recent) Calamitaceae Tchernoviaceae Gondwanostachyaceae Equisetaceae

CHAPTER 10

Figure 10.2 Habit of Equisetum pratense (Extant). (Courtesy

H. Bültmann.)

SPHENOPHYTES

Figure 10.3 Alfred Gabriel Nathorst. (With permission – Gustav Fischer, Verlag.)

One group that traditionally has been included in the Sphenophyta, or in their direct ancestry, is the Hyeniales (Kräusel and Weyland, 1926). Several recent studies suggest, however, that the Hyeniales are more similar to ferns and fernlike plants; consequently, a discussion of several of these (e.g., Calamophyton, Hyenia) can be found in Chapter 11.

PSEUDOBORNIALES The Pseudoborniales represents a unique order that is known from relatively few localities and, to date, is represented by a single species, Pseudobornia ursina. In 1894, the noted paleobotanist A. G. Nathorst (FIG. 10.3) collected a number of Late Devonian fossils from Bear Island (Bjørnøya), south of Spitsbergen. Included in the collection were a number of interesting compression and impression specimens consisting of axes bearing whorled, lateral appendages. The only other fossils that have been referred to this group were discovered in Devonian rocks in northeastern Alaska (Mamay, 1962). Since the original description by Nathorst, additional specimens have provided useful information about many previously unknown features of these plants. Pseudobornia ursina was monopodially branched and believed to have been 15–20 m tall (Schweitzer, 1967) (FIG. 10.4). The largest axis measures 60 cm in diameter and is

Figure 10.4 Hans-Joachim Schweitzer.

331

332

PaleoBOtany: the biology and evolution of fossil plants

thought to have represented the basal portion of the plant. It is not known what the underground parts were like, but the basal region has been reconstructed as a rhizome bearing roots. The primary stem of P. ursina consists of nodes

separated by internodes ⬃80 cm long (FIG. 10.5). In the basal regions of the plant, each node produces one or two first-order branches up to 3 m long and ⬃10 cm in diameter. Second-order branches are produced in a decussate arrangement (FIG. 10.5); each of these laterals bears axes in a pseudodistichous arrangement. Leaves (sometimes called meiophylls in older treatments, e.g., Benson, 1921) are produced only from the branches of the ultimate order and are arranged in superimposed whorls of four leaves each. A single leaf consists of an elongated, twice-dichotomized petiole to which are attached four leaflets. Each leaflet is up to 6 cm long and about 1 cm wide and consists of a lamina that is highly dissected along the margin. Nothing is known about the venation, but the configuration of the lamina suggests a dichotomous system of terete strands. Fertile branches up to 30 cm long are produced at the distal ends of primary branches. Each fertile unit consists of whorled bracts and sporangiophores, with the tips of the sporangiophores upturned and divided into two segments. Each sporangiophore produces ⬃30 sporangia. Nothing is known about the anatomy of P. ursina, although Schweitzer (1967) has suggested that the axes were hollow except at the nodes, where nodal diaphragms served to strengthen the stems. Mosbrugger (1990) suggested that the hollow stem of P. ursina, would have reduced the maximum bending stress in the periphery and at the base of the trunk, with only a minimum investment of cells; thus, this plant may have grown faster than plants with solid trunks. Moreover, the hollow stem may have served as an aeration system for the roots. This would have been particularly advantageous for this plant, as it lived in swamps in which the root systems were (temporarily) inundated. The affinities of the Pseudoborniales continue to remain obscure. The morphology of the vegetative organs is like that in the Equisetales, whereas the reproductive parts are more similar to members of the Sphenophyllales.

SPHENOPHYLLALES

Figure 10.5 Suggested reconstruction of Pseudobornia ursina. (From Schweitzer, 1967.)

Members of the Sphenophyllales were relatively small plants (probably
1 m tall) that formed a portion of the understorey in many Carboniferous forests. Sphenophyllaleans can be traced from the Devonian into the Triassic, but they are best known from petrified and compressed remains from the Carboniferous. Although the order was widespread in the late Paleozoic, Triassic representatives have only

CHAPTER 10

sporadically been documented from a few countries, including Japan (Asama and Naito, 1978; Asama and Oishi, 1980) and Korea (Kim, 1989). As the name of the order suggests, the leaves in this group are wedge-shaped and borne in whorls. Stems are protostelic, and secondary tissues are produced in some members. Reproductive parts are aggregated into cones or loosely constructed strobili; all members are apparently homosporous.

DEVONIAN SPHENOPHYLLALES

The record of Sphenophyllales from the Devonian is sparse. Several interesting fossils are known, however, that have been assigned to the Sphenophyllales, although they do not closely resemble the typical Carboniferous forms. These forms are based on macromorphological or anatomical features and may be significant with regard to resolving early sphenophyte evolution. Four of these fossils Hamatophyton verticillatum, Rotafolia songziensis, Eviostachya hoegii, and the morphogenus Xihuphyllum are given in detail. One of the best-known Devonian plants referred to the Sphenophyllales is Hamatophyton verticillatum from the Famennian (Late Devonian) to late Tournaisian–early Viséan (Mississippian) of southern China (Wu and Zhao, 1981; Feng and Ma, 1991; X. Li et al., 1995; D.-M. Wang et al., 2006a). A detailed reassessment of this plant was recently provided by D.-M. Wang et al. (2006a) based on both compressions and partially structurally preserved specimens. Axes of H. verticillatum are pseudomonopodial and possess axial trichomes or spines and nodal whorls of sterile leaves, which are dimorphic and sometimes also contain trichomes or spines. Hook-like leaves are up to 15 mm long and 0.7 mm wide and occur in whorls of 6–10 per node, whereas palmate leaves occur in whorls of 6, are up to 40 mm long, and dichotomize two to four times. The axes are protostelic, with primary xylem in a more or less triangular configuration and secondary xylem in radial arrangement. In contrast to other sphenophyllaleans, however, Hamatophyton lacks secondary xylem around the tips of the primary xylem arms (i.e., fascicular secondary xylem). Primary xylem maturation is exarch and the secondary xylem lacks parenchyma and rays. Several specimens showing fertile main axes and lateral branches have been reported. These axes usually lack vegetative leaves but instead bear nodal whorls of stalks up to 2.5 mm long that terminate in a single, elongate to ovoid, spiny sporangium. In some specimens, fertile whorls and vegetative leaves are interspersed. Some of the sporangia contain spherical–triangular trilete spores, 27–37 mm in diameter, with grani and coni in the equatorial and distal

SPHENOPHYTES

333

region. D.-M. Wang et al. (2006a) regard H. verticillatum as demonstrating the most primitive type of morphology and anatomy known in sphenopsids to date. A plant similar to H. verticillatum is Rotafolia songziensis from the Upper Devonian (Famennian) Xiejingsi Formation, southwestern Hubei Province, China (D.-M. Wang et al., 2005, 2006b). Specimens consist of branched axes, up to 20 cm long, that are characterized by slightly swollen nodes and internodes with ribs and narrow branches produced from the nodes. The axes are protostelic; cross sections show a more or less triangular or quadrangular organization of the primary xylem and radial arrangement of the secondary xylem. Primary xylem maturation is exarch, with protoxylem strands positioned around the tips of the primary xylem arms. There is no apparent differentiation between fascicular and interfascicular regions of the secondary xylem, and ray cells are rarely observed (D.-M. Wang et al., 2006b). Protoxylem tracheids have helical wall thickenings, whereas the tracheids of meta- and secondary xylem possess scalariform and/or bordered pits. All R. songziensis axes bear trichomes or spines, some up to 2.8 mm long. At the nodes are whorls of six leaves, up to 2.4 cm long, that regularly or irregularly fork two to four times. The leaves are positioned perpendicular to the axis in proximal portions of the plant but inserted at acute angles in distal portions. Fertile branches are produced in a terminal strobilus (8.5 cm long), which is subtended by whorls of normally developed leaves. It consists of a central axis and up to 16 whorls of fertile units, each of which consists of a bract and 6–10 sporangia (D.-M. Wang et al., 2005). The bracts are elongate–cuneate and have prominent marginal fringes; sporangia are attached to the abaxial surface at the base of each bract. Although R. songziensis closely resembles H. verticillatum with regard to external axis morphology, leaf shape, and structure of the primary xylem, the two forms differ in strobilus morphology. Eviostachya hoegii is a problematic Late Devonian cone that is sometimes included in the Sphenophyllales. The species is known from both compressed and structurally preserved specimens from Belgium (Stockmans, 1948; Leclercq, 1957) and China (Y. Wang, 1993). Compression specimens measure up to 5.5 cm in length (Stockmans, 1948). Permineralized material studied by Leclercq (1957) indicates that the strobilar axis in E. hoegii is three lobed. At the base of each cone is a whorl of highly dissected bracts; the remainder of the cone consists of whorls of sporangiophores. Each sporangiophore trichotomizes several times to produce 27 sporangia (FIG. 10.6). The triangular shape of the stele is one feature used to suggest affinities with the

334

PaleoBOtany: the biology and evolution of fossil plants

not only includes numerous species of leaves, but also impression–compression and structurally preserved stems, roots, and leaves, as well as whole plants (Baxter, 1948; Reed, 1949).

Figure 10.6 Suggested reconstruction of the sporangiophore of Eviostachya hoegii (Devonian). (From Taylor and Taylor, 1993.)

Sphenophyllales. The absence of sterile bracts between successive whorls of sporangiophores is a major distinction between E. hoegii and other sphenophyllaleans but parallels the condition in Hamatophyton. Affinities with the Hyeniales have also been suggested based on the organization of the sporangiophores. Xihuphyllum is a morphogenus used for vegetative, putatively sphenophyllalean, remains from the Xihu Formation (Upper Devonian) of Zhejiang, China (Q. Chen, 1988). These fossils consist of stout, longitudinally striate, articulated axes (up to 4 cm in diameter in X. megalofolium) with swollen nodes and internodes up to 11.5 cm long. The axes bear whorls of six to eight leaves that are flabellate, ellipsoidal, or elongate–cuneate with coarse parallel venation. Veins occasionally fork and anastomose. The reproductive structures of Xihuphyllum remain unknown. SPHENOPHYLLUM

The generic name Sphenophyllum was initially used for Carboniferous–Permian impression–compression specimens of wedge-shaped leaves borne in whorls. Today the generic name

LEAVES Impression–compression fossils assigned to Sphenophyllum include linear, spatulate, or fan-shaped leaves that extend up to several centimeters in length. The leaves are arranged along the axis in nodal whorls (sometimes termed verticils) of six or nine; the whorls are attached at various angles, but often at 90° to the axis. Whorls may be isophyllous (all leaves of the same size and shape) or anisophyllous (some leaves differ in size and/or shape from the others). Bilaterally symmetrical whorls with two pairs of large and one pair of distinctly smaller leaves are termed trizygoid. Each leaf is vascularized by a single bundle that enters the base and dichotomizes several times before terminating at the margin. Numerous Sphenophyllum species have been studied by means of cuticular analysis. The epidermis consists of thinwalled, irregularly isodiametric-to-elongate epidermal cells that usually possess sinuous anticlinal walls (FIG. 10.7). Stomata are confined to the abaxial surface of the leaf; the stomatal apparatus typically consists of two guard cells with polar and circumporal thickenings and two subsidiary cells (syndetocheilic-type stoma), one of which is distinctly larger than the other (FIG. 10.8) (Pant and Mehra, 1963; Barthel, 1997; Yao et al., 2000). Other species, however, appear to have been haplocheilic, with three to six subsidiary cells surrounding the guard cells (Abbott, 1958; Good, 1973). In some species, hairs up to 1 mm long occur along the margins (Abbott, 1958 (FIG. 10.9); Barthel and Müller, 2006). The identification of Sphenophyllum foliage species has been based on features such as the size, shape, and number of leaves per whorl, number of marginal teeth, and degree of dissection of the lamina between the teeth, which can range from entire to deeply lobed or even filamentous (e.g., Doubinger and Vetter, 1954 (FIG. 15.14); Storch, 1966, 1980; Batenburg, 1977; Cúneo et al., 1993a). Many of the species exhibit considerable heterophylly (FIG. 10.10) (Galtier and Daviero, 1999; Naugolnykh, 2003) and, consequently, species will continue to be reduced into synonymy as research continues on the genus. In the Permian Cathaysian floras of East Asia, a combination of two principal features, symmetry of the whorls (radial versus trizygoid leaf arrangement) and venation (e.g., straight versus curved veins), has been used to subdivide Sphenophyllum species into four morphogenera: Sphenophyllum, with radial leaf arrangement and straight veins; Parasphenophyllum, with radial leaf arrangement and curved veins; Trizygia, with trizygoid leaf arrangement and

CHAPTER 10

SPHENOPHYTES

335

Figure 10.7 Sphenophyllum cuneifolium, cuticle of a leaflet

(Pennsylvanian). Bar  200 µm. (Courtesy M. Barthel.)

Figure 10.9 Maxine L. Abbott.

Figure 10.10 Morphological variability of Sphenophyllum emarginatum leaves (Pennsylvanian). Bar  5 mm. (From Batenburg, 1977.)

straight veins; and Paratrizygia, with trizygoid leaf arrangement and curved veins (Asama, 1966, 1970; Kim, 1989). Sphenophyllum multirame is a structurally preserved stem with attached leaves from Middle Pennsylvanian coal balls in Illinois (Good, 1973). Two types of leaves were produced: small, linear leaves (1–4 mm long) in whorls of 5–12 and larger leaves borne in whorls of 6. The larger leaves are superimposed from one node to the next, that is, they overlap the node above, and their margin may be blunt or deeply incised, with up to nine veins present at the distal edge. In the highly dissected forms, longitudinal furrows on the abaxial surface correspond to each leaf lobe. Leaves of S. multirame are hypostomatic, with sunken and randomly oriented stomata. A distinct mesophyll is present, but no palisade layer identified to date.

Figure 10.8

Sphenophyllum emarginatum cuticle with stomata (Pennsylvanian). Bar  3 µm. (Courtesy H. Kerp.)

STEM ANATOMY Structurally preserved stems of Sphenophyllum are common in coal-ball permineralizations throughout the world.

336

PaleoBOtany: the biology and evolution of fossil plants

Specimens of S. plurifoliatum, the common Pennsylvanian form, may reach 2 cm in diameter and include several orders of branching (Baxter, 1948). The primary xylem of the protostele is triangular (FIG. 10.11) to subtriangular in outline with three concave sides. Xylem development is exarch (FIG. 10.12), although in a few cases mesarch development has been described. Thin-walled cells that make up the primary phloem alternate with the arms of xylem. In young

Figure 10.11

Distal end of Sphenophyllum plurifoliatum stem showing cortical tissues. Six arms represent bases of overarching leaves. Note triangular stele (arrow). (Pennsylvanian). Bar  1 mm.

Figure

10.12 Young Sphenophyllum plurifoliatum stem (Pennsylvanian). Bar  0.5 mm.

stems (FIG. 10.13), the remainder of the axis consists of a parenchymatous cortex bounded by an epidermis in which the cells are filled with opaque materials. Older stems have abundant secondary xylem, with fascicular tracheids (those opposite the protoxylem arms) that are considerably smaller in diameter than interfascicular tracheids (those between the arms) (FIG. 10.14). The secondary xylem tracheids are vertically elongated (FIG. 10.15), with tapering end walls and circular–elliptical bordered pits on the lateral walls. Some tracheids are up to 30 mm long (FIG. 10.15) and are among the longest known tracheids of any vascular plants, living or fossil (Cichan and Taylor, 1982a). Extending through the wood and into the phloem are narrow vascular rays (FIG. 10.14). The vascular cambium in S. plurifoliatum is considered determinate, as there is no evidence that the fusiform initials underwent anticlinal divisions (Cichan, 1985b). The increase in stem diameter was accommodated by an increase in the size of the fusiform initials. This pattern of secondary growth in Sphenophyllum parallels that found in several Carboniferous lycopsids but is unlike the developmental sequence in seed plants. There remains some uncertainty as to whether the vascular cambium in Sphenophyllum is unifacial (Cichan and Taylor, 1990; E. Taylor, 1990) or bifacial (Eggert and Gaunt, 1973) and, thus, whether phloem is primary or secondary in origin. Sieve elements have horizontal to oblique end walls and vary in diameter, depending on whether they are fascicular or interfascicular. Although sieve areas have been reported once (Stidd and Ma, 1978), they have not been studied in detail. Secondary cortical tissues are produced by a single persistent phellogen. Periderm cells are arranged in distinct radial files and appear tabular in transverse section.

Figure 10.13 Young Sphenophyllum plurifoliatum stem with surrounding leaves (arrows). (Pennsylvanian). Bar  1 mm.

CHAPTER 10

SPHENOPHYTES

337

The apical meristem of Sphenophyllum consists of a tetrahedral apical cell that includes a triangular upper surface and three, triangular internal cutting faces (FIG. 10.16) (Good and Taylor, 1972). The production of derivative cells (segment and sextant cells) occurs in a dextrorse direction similar to apical cell production in the genus Equisetum. These patterns appear to be consistent for a given species. In addition to apical growth, Sphenophyllum axes increase in length as a result of intercalary meristematic activity.

Figure 10.14

Mature Sphenophyllum plurifoliatum stem showing central triangular body and small diameter fascicular and large diameter interfascicular secondary xylem elements (Pennsylvanian). Bar  1 mm.

Figure 10.15 Longitudinal section of Sphenophyllum plurifoliatum stem showing extended length of tracheids (Pennsylvanian). Bar  2 mm.

ROOTS The roots of Sphenophyllum were adventitious and rarely produced laterals (Storch and Barthel, 1980). Anatomically, they possess the same complement of tissues as the stems, although the primary xylem in the roots is typically diarch. In some specimens, an extensive periderm was developed (Baxter, 1948). REPRODUCTIVE BIOLOGY Reproductive organs of the Sphenophyllales consist of aggregations of sporangiophores and bracts that form

Figure 10.16 Longitudinal section of Sphenophyllum apex showing triangular-shaped apical cell (arrow) and derivatives (Pennsylvanian). Bar  200 µm.

338

PaleoBOtany: the biology and evolution of fossil plants

cones or strobili. Many types are known and they are preserved in various ways. Historically, the morphogenus Sphenophyllostachys was used for all Sphenophyllum cones, but Hoskins and Cross (1943) argued that Bowmanites has priority and should be used instead. Several other genera of sphenophyllalean cones have later been introduced in recognition of the wide diversity displayed by these structures (see Good, 1978). Bowmanites dawsonii is a form known from both European and North American floras, and it will be used to characterize the basic strobilar organization in the group (FIG. 10.17). Specimens exceed several centimeters in length and consist of a central axis bearing whorls of 14–20 bracts; bracts alternate from whorl to whorl (FIG. 10.17). The stele of the axis varies from triarch to hexarch and the number of protoxylem points also varies; metaxylem tracheids have reticulate pits. Bracts are fusiform in outline and vascularized by a single trace. They are inserted at right angles to the

B

B

cone axis and fused laterally to form a shallow disk. Toward the margin of the disk, individual bracts are separated and upturned at their tips to overlap the bracts of several whorls above. In B. dawsonii, each bract subtends two sporangiophores of unequal length that arise from the adaxial surface of the bract close to the cone axis (T. Taylor, 1969). At the distal end, each sporangiophore is reflexed and bears a single large sporangium at the end. Sporangia are thick walled and ⬃2 mm in diameter; trilete spores range from 96 to 180 µm in diameter. Covering the surface of the spores is an extraexinous membrane that has been interpreted as homologous to the perispore of certain homosporous ferns (T. Taylor, 1970a; W. Taylor, 1986). Bowmanites moorei is a relatively small cone with a maximum diameter of ⬃4 mm (Mamay, 1959b). Three bracts are produced at each node and laterally fuse to form a disk. Each bract consists of a median fertile lobe flanked by two sterile lobes. Bract tips do not overlap the whorl above. Each fertile lobe produces two recurved sporangia. In B. fertilis (Scott, 1906a) from the Lower Coal Measures of Belgium, each node produces six adaxially oriented stalks, which branch at the tips to form 14–18 sporangiophores. Each sporangiophore bears two recurved sporangia. Further morphologic variability is present in B. trisporangiatus, a cone 6 cm long (Hoskins and Cross, 1943). A whorl consists of 18 bracts that produce three sporangiophores each; spores are large (100–150 µm) and trilete. Despite considerable variability in the size of some spores, which may, in part, reflect measurements that include the perispore, all species of Bowmanites were monosporangiate and the genus is regarded as homosporous. Reports of heterosporous Bowmanites-type cones (Arnold, 1944) have later been discounted or the affinities of the cones reinterpreted. It is interesting that not all species of Bowmanites had trilete spores. Some, like B. bifurcatus, contained sporangia with monolete spores (Andrews and Mamay, 1951). In this species the cone is small (1.5 cm long and 3 mm in diameter). Each whorl of bracts contains six bracts; a single sporangiophore arises from the adaxial surface of each bract and bears two sporangia. OTHER SPHENOPHYLLALES

Figure 10.17

Paradermal section of Bowmanites dawsonii cone showing whorls of sporangia containing spores and bracts (B) between (Pennsylvanian). Bar  3 mm.

There are several other types of articulated axes bearing leaves arranged in nodal whorls for which affinities with the Sphenophyllales have been suggested. One is Gondwanophyton from the Early Permian of India and Australia (Maithy, 1974; Srivastava and Rigby, 1983; McLoughlin, 1992), a plant characterized by slender, articulated striate stems and narrow cuneate, semicircular to

CHAPTER 10

reniform leaves arranged in nodal whorls or pairs (FIG. 10.18). Leaf venation is dichotomous and similar to Sphenophyllum. As the reproductive structures of Gondwanophyton remain unknown, however, the systematic affinities of this taxon continue to be elusive. Peltastrobus is a cone type that produced monolete spores (FIGS. 10.19, 10.20) and is thought to be associated with the Sphenophyllales (Baxter, 1950). Peltastrobus reedae is known from Upper and Middle Pennsylvanian deposits and includes cones ⬃4 mm in diameter (Leisman and Graves, 1964). Whorls consist of three sterile and three fertile bracts at each node (FIG. 10.21). Vegetative and sterile bracts alternate. Each fertile unit consists of a bract that subtends a cluster of five axillary sporangiophores. Two sporangiophores are directed upward and two downward. The remaining sporangiophore extends outward at 90° from the cone axis. Each sporangiophore bears up to eight sporangia arranged in an inner and an outer whorl (FIG. 10.22). The epidermal cells on the sporangiophores have thin walls with sinuous margins. Monolete spores also occur in Sentistrobus goodii (Riggs and Rothwell, 1985), a form known from the Upper

Figure 10.18 Suggested reconstruction of Gondwanophyton daymondii. (From McLoughlin, 1992.)

SPHENOPHYTES

339

Figure 10.19

Peltastrobus reedae sporangium with monolete spores (Pennsylvanian). Bar  80 µm.

Figure 10.20 Peltastrobus reedae spore. Note the opaque spherical structure inside the spore (Pennsylvanian). Bar  20 µm.

340

PaleoBOtany: the biology and evolution of fossil plants

Figure 10.22 Transverse section of Peltastrobus reedae cone

(Pennsylvanian). Bar  2 mm.

Figure

10.21 Longitudinal section of Peltastrobus reedae showing sporangiophores of several nodes and sporangia (Pennsylvanian). Bar  4 mm.

Pennsylvanian of the Appalachian Basin. These small (1.4 cm in diameter) cones have three appendages per node, each of which dichotomizes in a horizontal plane. Subsequent dichotomies of the upturned ends of the appendages produce spinelike tips. Both Peltastrobus and Sentistrobus have spores of the Columnisporites type. Cheirostrobus is an unusual cone that is often included in the Sphenophyllales because of its occurrence in Mississippian deposits rich with Sphenophyllum stems. The cone exceeds 10 cm in length and includes whorls of 12 bracts each. The bracts of C. pettycurensis (Scott, 1897) are more complex, in that each is divided horizontally to form two lobes, which are, in turn, subdivided into three segments. The upper segments produce elongated, flattened sporangia that extend inward toward the cone axis. The lower segments divide again to form three bracts; spores are ⬃65 µm in diameter. Lilpopia ranges from the Late Pennsylvanian to the Early Permian and also has an unusual fructification organization. It is known from only a few localities in Europe, including the Karniowice travertine near Krakow in Poland, Crock in eastern Germany, and Sobernheim in southwestern Germany (Remy and Remy, 1961; Conert and Schaarschmidt, 1970; Lipiarski, 1972a, b). The most complete specimens come from the Early Permian of Sobernheim (FIG. 10.23) and have been assigned to L. raciborskii (Kerp, 1981, 1984b).

CHAPTER 10

SPHENOPHYTES

341

Figure 10.24 Lilpopia raciborskii fertile region showing clusters of sporangia (arrow) (Permian). Bar  5 mm. (Courtesy H. Kerp.)

Figure 10.23

Lilpopia raciborskii, main shoot giving off a fertile lateral branch (arrow) (Permian). Bar  2 cm. (Courtesy H. Kerp.)

The lateral axes have normally developed vegetative whorls, both above and below the fertile region. The vegetative organs of the Sphenophyllales are relatively uniform throughout their geologic range, with the reproductive organs demonstrating greater diversity. All the taxa described to date are homosporous, with either monolete or trilete spores. An ultrastructural analysis of the spore wall of a number of taxa suggests that the group is a natural one (W. Taylor, 1986). ECOLOGY

This species is reconstructed as a vegetative axis with nodal whorls of six wedge-shaped leaves. Extending from the vegetative axis are loosely organized, lateral axes with intercalary fertile regions (FIG. 10.24). The fertile regions are characterized by whorls of laterals, consisting of three clusters of sporangia alternating with three bilobed leaves (FIG. 10.25).

Although the exact habit and paleoecology of Sphenophyllum are not known, Batenburg (1981; 1982) suggested that the plants had prostrate rhizomes from which aerial axes arose. Many of the compression species have been reconstructed as scrambling understorey plants in which some of the leaves (or entire leafy branches) were modified for attachment

342

PaleoBOtany: the biology and evolution of fossil plants

Figure 10.26 Sphenophyllum cuneifolium, linear leaflet with foliar climber hook (Pennsylvanian). Bar  200 µm. (Courtesy M. Barthel.)

(Bashforth and Zodrow, 2007). Sphenophyllum oblongifolium and S. cuneifolium from the Pennsylvanian–Early Permian of France and Germany, produced specialized shoots with narrow, elongated leaves with a single vein that extended through the leaf tip to form climber hooks (FIG. 10.26), indicating a scrambling or climbing habit (Barthel, 1997; Galtier and Daviero, 1999; Barthel and Müller, 2006). Leaves modified into climber hooks have also been recorded for S. miravallis from the lower Stephanian (Upper Pennsylvanian) of the Saar Basin in Germany (Hetterscheid and Batenburg, 1984) and S. biarmicum from the Lower Permian of the central Cis-Urals (Naugolnykh, 2003). It has also been suggested that Sphenophyllum was a hydrophyte (or hygrophyte) in which portions of the vegetative plant body were permanently or temporarily submerged in water. Further information about the various hypotheses concerning the ecology of Sphenophyllum can be found in Shchegolev (1991), and a brief summary in Naugolnykh (2003).

EQUISETALES Figure 10.25 Diagrammatic reconstruction of Lilpopia raci-

borskii. (From Kerp, 1984.)

Historically, the Equisetales included a small number of taxa that were thought to be related to one another on the basis of their presumed herbaceous habit and lack of secondary tissues. The order Calamitales was used for those genera that were arborescent and produced secondary tissues.

CHAPTER 10

Comprehensive studies of several equisetophyte genera, however, have suggested that such features as habit, spore morphology, bracteate versus nonbracteate cones, and the degree of fusion of the leaves can no longer be used to separate the two groups (Good, 1975). Accordingly, all the genera in these two groups are now included within a single order, the Equisetales. This combination is supported by the presence of similar anatomy and morphology in the two groups. Both calamites and equisetaleans produce elaters on their spores and possess similar primary vascular structure and meristematic organization of the primary axes. The Equisetales are characterized by the presence of distinct nodes, internodes, and longitudinal internodal ridges on their stems. Branches and leaves are borne in distinct whorls; leaves are connate or free and vascularized by a single, undivided vein. The underground portion of the plant usually consists of a rhizome bearing numerous adventitious roots. Apical meristems contain a single, large apical cell and intercalary meristems are present at the base of each internode which also produces growth in length. Vascular tissue is organized into bundles that alternate at successive nodes. Each bundle includes a protoxylem canal in the internode. Sporangia are recurved and organized into strobili that may or may not contain sterile bracts. Spores possess superficially attached elaters. The geologic history of the Equisetales can be traced from the Devonian to the recent, with the most conspicuous components of the flora present during the Carboniferous (Jongmans, 1911). In this chapter, the order is separated into four families: the Calamitaceae, Tchernoviaceae, Gondwanostachyaceae, and the Equisetaceae. CALAMITACEAE

The Calamitaceae are arborescent plants with considerable secondary growth. They first appeared in the Late Devonian and reached their peak diversity in the Pennsylvanian, when some forms attained heights ranging from 15 m to more than 20 m and were major constituents of the lowland equatorial swamp forest ecosystems. Although calamitacean diversity drastically declined during the Permian, a few forms persisted locally to the Late Permian (S.-J. Wang et al., 2003c, 2006). ARCHAEOCALAMITES Archaeocalamites, a morphogenus used for impression– compression fossils, casts and molds (steinkerns), and permineralizations of stems, belongs to a distinct group of calamites that extended from the Late Devonian into the Early Permian. The group has historically been regarded as a separate family for which the names Archaeocalamitaceae or

SPHENOPHYTES

343

Figure 10.27

Detail of calamite stem surface showing ribs and furrows (Pennsylvanian). Bar  1 cm.

Asterocalamitaceae have been used (Hirmer, 1927; Parihar, 1985). Considerable taxonomic confusion exists with regard to these family names, however, and both names appear to be illegitimate. The type genus of the Archaeocalamitaceae, Archaeocalamites, is based on the same basionym as the genus Calamites—C. radiatus. Asterocalamites scrobiculatus, the basionym for the genus Asterocalamites, is taxonomically identical with C. radiatus (for details, see Leistikow, 1959; Remy and Remy, 1978b). In this chapter, we disregard the wide-ranging taxonomic consequences resulting from these problems for reasons of simplicity and use Archaeocalamites and affiliated reproductive structures (i.e., Pothocites, Pothocitopsis, and Protocalamostachys) as genera in the family Calamitaceae. The genus Archaeocalamites displays an interesting stratigraphic range. It is first recorded from the uppermost Devonian and was abundant and widespread throughout Mississippian and earliest Pennsylvanian wetlands, especially in areas of high disturbance such as riparian zones and floodplains (DiMichele and Phillips, 1996b; DiMichele et al., 2001b). There is a considerable hiatus of records during the remainder of the Carboniferous, with the genus eventually reappearing in the Early Permian (Bateman, 1991). Archaeocalamites was arborescent, with siphonostelic stems several centimeters in diameter (16.5 cm in A. scrobiculatus; Hirmer, 1927). The principal feature that has been used to distinguish stems of Archaeocalamites from those of other calamites (FIG. 10.27) is the presence of ribs and furrows that do not alternate from node to node. In addition, the surface ribs are truncated at the node, whereas they are pointed in other equisetalean stems. Sandstone casts, with occasional molds still intact, have been described from the Mississippian (Chester Series) of

344

PaleoBOtany: the biology and evolution of fossil plants

Illinois as Archaeocalamites radiatus (Lacey and Eggert, 1964). The stems vary from 0.5 to 1.2 cm in diameter, with nodes ⬃3 cm apart. Longitudinal ribs are ⬃1 mm wide and do not alternate at the nodes. In one specimen, five circular scars are present, suggesting a highly branched axis. Near the base the axis rapidly tapers and, at this level, the ribs alternate from node to node. Structurally preserved Archaeocalamites stems have been described from several localities in Europe and North America (Solms-Laubach, 1897b; Walton, 1949b; Smoot et al., 1982; Bateman, 1991; Dunn, 2004). In A. esnostensis, which is preserved in a chert matrix from the Viséan of France, there is a central pith with inwardly projecting primary xylem strands (Galtier, 1970b). These strands are responsible for the furrows on the surface of pith casts. In a specimen from the Mississippian of North America, the protoxylem includes canals that occur near the tip of each wedge of primary xylem (Smoot et al., 1982). The remaining part of the stem consists of secondary xylem composed of elongated tracheids with scalariform to multiseriate circular pits and narrow vascular rays. The secondary xylem of a permineralized Archaeocalamites stem from the Mississippian of Kingswood near Pettycur (Scotland) shows well-developed growth rings characterized by size differences of the tracheids. Scott et al. (1986) suggested that these growth rings recorded environmental stress such as periods of drought, rather than regular seasonal or climatic variation. Cortical tissues are rarely preserved in Archaeocalamites stems. Leaves of Archaeocalamites are borne in whorls on the distal branches. They are slender (FIG. 10.28) and dichotomize one to three times; some are up to 10 cm long (Jennings, 1970). Each dichotomy of the leaf is vascularized by a single terete strand; nothing is known about the stomata and tissues of the leaf. Roots are borne on rhizomes in irregular whorls and branch infrequently. Uniseriate rays are present in the first-order roots but not in subsequent orders of branching. The reproductive parts are cones borne in whorls on the leafy branches. One cone morphotype attributed to Archaeocalamites is Pothocites (Kidston, 1883). Pothocites grantonii measures up to 9 mm in diameter and has up to 12 non-alternating whorls of sporangiophores (Chaphekar, 1965). The axis of the cone contains a ring of mesarch bundles and the number of bundles equals the number of sporangiophores. The distal tip of each sporangiophore is cruciate, with each arm containing a single, recurved globose sporangium. The sporangial walls exhibit delicate, peglike processes which extend inward toward the center of the sporangium. Spores range from 82 to 104 µm in diameter and are of the Calamospora type. The trilete spores have a

Figure 10.28 Archaeocalamites radiatus (Mississippian). Bar  5 cm. (Courtesy GBA.)

two-layered exine, suggesting that these spores may have possessed an extraexinous layer of elaters. Pothocitopsis (Nathorst, 1914) is the generic name of a cone described from the Devonian of Spitsbergen that may represent a poorly preserved specimen of Pothocites. Protocalamostachys is another cone known to have been borne on Archaeocalamites and regarded by some as identical with Pothocites. Specimens are at least a centimeter long and contain three pairs of sporangiophores per whorl (FIG. 10.29). The cone axis is vascularized by three mesarch strands that correspond to the position of the sporangiophores.

CHAPTER 10

SPHENOPHYTES

345

Figure 10.29 Diagrammatic cross section of Protocalamostachys pettycurensis (Missippian). (From Taylor and Taylor, 1993.)

The sporangiophores are divided into four segments distally, each of which bears an elongated sporangium (FIG. 10.29). Spores are trilete and up to 44 µm in diameter. Although Protocalamostachys cones have not been found organically attached to Archaeocalamites twigs, the similarity of the vascular bundles in both and their similar co-occurrence in the Pettycur limestone have been used to suggest their biological relationship. Chaphekar (1963) has suggested that P. pettycurensis was attached to petrified stems of A. goeppertii. The primary difference between these Mississippian cones and younger calamitaceans is the absence of sterile members alternating with the fertile whorls. CALAMITES The name Calamites was initially used for pith casts, but the genus now encompasses various preservational modes including impressions, compressions, and casts of the external surface (FIG. 10.30) of stems or of the central canal or pith (FIGS. 10.31, 10.40). The generic name is also used for structurally preserved stems. In the following discussion, the generic name Calamites will be used to describe the entire plant, as was done in the description of Lepidodendron (Chapter 9). Calamites extends from the Late Mississippian into the Permian and is reconstructed as an upright, arborescent plant that grew to a height of ⬃20 m; stems could reach a diameter

Figure 10.30 Compressed calamite stem showing internodal ribs and furrows (Pennsylvanian). Bar  2 cm.

of 60 cm, for example, Arthropitys ezonata (FIG. 10.32) (Rössler and Noll, 2006). Calamiteans were slightly smaller plants than many of the arborescent lycopsids or the cordaites (Chapter 20) and no doubt constituted the second story of the Carboniferous swamp forests. Calamite stems typically arise from subterranean rhizomes, or mature vertical stems that bud laterally to form new stems (Pfefferkorn et al., 2001) and are commonly reconstructed as freestanding trees with a multibranched crown (FIG. 10.33). Biomechanical evidence, however, suggests that some forms were not entirely selfsupporting (S.-J. Wang et al., 2006) and may have gained

346

PaleoBOtany: the biology and evolution of fossil plants

Figure 10.31

Arthropitys sp., segment of a permineralized calamite stem opened longitudinally to show pith/medullary canal and positions of transverse nodal diaphragms (Permian). Bar  5 cm. (Courtesy BSPG.)

Figure 10.33 Vegetative branch showing prominent axis with leaves and branch (Pennsylvanian). Bar  2 cm. Figure 10.32

Cross section of Arthropitys ezonata, the largest calamite stem discovered to date (Permian). Bar  5 cm. (Courtesy R. Rössler.)

mutual support by growing in dense stands (Spatz et al., 1998). Woody calamites from the Lower Permian of Tocantins, Brazil, developed woody adventitious roots (1–8 cm in diameter) in the proximal portion of the stem (FIG. 10.34) (Rössler and Noll, 2002; Rössler, 2006). These may have provided mechanical stability and may have also made the plants more independent from the rhizome during growth. Calamites gigas, used here for the whole plant composed of C. gigas stems (FIG. 10.35), Annularia carinata foliage, and Metacalamostachys dumasii cones (Kerp,

1984a), is from the Rotliegend (Upper Pennsylvanian–Early Permian) of Germany. It is believed to have been a succulent plant up to 2 m high. It apparently lacked an underground rhizome and instead had a deep-reaching taproot from which secondary roots extended (Barthel and Rössler, 1996; Barthel, 2004, 2006b). Naugolnykh (2005a), however, does not concur with this interpretation and regards C. gigas as a phreatophyte capable of obtaining water from a permanent ground supply or from the water table during periods of drought.

CHAPTER 10

SPHENOPHYTES

347

Figure 10.34 Cross section of Arthropitys sp. stem near the base showing large woody roots in cross section (arrows) (Permian). Bar  7.5 cm. (Courtesy R. Rössler.)

Several subgenera of Calamites have been established on the basis of size, position, and regularity of the branches. In the subgenus Calamitina (FIG. 10.36), numerous branches are produced at some nodes, whereas Calamites subg. Stylocalamites produces fewer, larger branches. A more regular branching pattern is present in Calamites subg. Diplocalamites (FIG. 10.37), where only two or three branches are produced per node. In this subgenus, however, the branches alternate from node to node in a distichous pattern, and internodal distances are smaller. The most profusely branched arborescent calamite is Calamites subg. Crucicalamites (FIG. 10.38). Numerous large branches are borne at each node, and the diameter of the stems suggests that these were among the largest of the arborescent calamites. As relatively few calamite stem specimens have ever been found with branches attached, the establishment of the preceding subgenera is principally based on the occurrence and number of branch scars present on stems and casts. It is highly probable that branches were shed as the plant continued to grow (Daviero and Lecoustre, 2000). The size of the scars is variable, but their general morphology is relatively constant and consisted of a circular to slightly oval structure with a central area that marked the position of the pith. In well-preserved specimens, the scars are ornamented by a series of wrinkles that extend out from the central scar like the spokes of a wheel. The presence of such features immediately distinguishes a specimen of the external surface of the stem from a pith cast. This becomes more

Figure 10.35

Calamites gigas, stem (steinkern) in situ (Permian). Bar  10 cm. (Courtesy M. Barthel.)

348

PaleoBOtany: the biology and evolution of fossil plants

Figure 10.37

Suggested reconstruction of Calamites subg. Diplocalamites. (From Taylor and Taylor, 1993.)

Figure 10.36

Suggested reconstruction of Calamites subg. Calamitina. (From Taylor and Taylor, 1993.)

difficult in those instances where branches were not produced. Some permineralized calamite stems from the Early Permian of Tocantins, Brazil, still have branches attached, and these indicate that in large stems with considerable secondary xylem development, the presence of branches is not evident from pith casts (Rössler, 2006). Rössler also noted that branching in calamites was generally more variable than previously thought, which casts doubt on the validity of classification systems based on branch types and their spatial arrangement. A different system of calamite classification was proposed, which was based on the number and arrangement of the primary cauline and foliar steles and includes the subgenera Mesocalamites, Calamitina, and Calamitopsis (Remy and Remy 1978b).

CHAPTER 10

SPHENOPHYTES

349

Figure 10.39 Portion of calamite pith cast showing closely spaced nodes and infranodal canals (arrow) (Pennsylvanian). Bar  4 cm.

Figure 10.38

Suggested reconstruction of Calamites subg. Crucicalamites. (From Taylor and Taylor, 1993.)

PITH CASTS. Pith casts of calamites were formed by sediments that filled the hollow central canal and solidified before the more resistant tissues of the stem, such as the primary and secondary xylem, were broken down by various biological agents. Following the decay of the remainder of the axis, additional sediment filled in around the cast (FIG. 10.39), thus resulting in a mold–cast preservation type. The structural organization of a calamite stem includes wedges of primary xylem that extended into the central canal, with broad channels of parenchyma representing the vascular rays between. On the surface of a pith cast, these appear as a series of ribs and furrows—furrows mark the former position of the primary xylem wedges, whereas ribs correspond to the vascular rays between them. In many pith

Figure 10.40 Detail of Calamites suckowi pith cast showing circular infranodal canals (Pennsylvanian). Bar  2 cm.

casts, it is possible to distinguish a small oval scar just below the node on the ribs. These scars mark the position of loose patches of tissue within the vascular rays and are termed infranodal canals (FIG. 10.40). On some specimens preserved in a very fine matrix, an additional scar may be seen just above the node at the bottom of a rib; this scar identifies the former position of a leaf trace. The presence of infranodal canals or leaf traces can also be of value in orienting the

350

PaleoBOtany: the biology and evolution of fossil plants

specimen. Many calamite pith casts exhibit a tapered end, which was initially thought to represent the growing point of the stem. It is now known that the tapered end represents the basal portion of branches (FIG. 10.38), where there was a rapid increase in the size of the pith. Some specimens have been described as pith casts but do not show infranodal canals, even though the surfaces show a series of ribs and furrows that alternate at the nodes. These casts may represent decorticated axes in which the extraxylary tissues were present at the time of fossilization and not pith casts; thus, they do not show branch or infranodal canal scars. In 1956, Arnold described a large Calamites trunk from the Pennsylvanian of Colorado. He was able to demonstrate that the vertical lines on the surface of C. huerfanoeinsis were the result of sediment deposited between the woody cylinder and the outer surface of the trunk where the cortical tissues were once located. The impression of the outer surface of the xylem cylinder and the pith cast looked almost identical, even in the alternation of ribs and furrows at the node, but the former did not show infranodal canals or branch scars. STEM ANATOMY. Numerous permineralized calamitean stems have been examined and details are known about the growth and development of the plant (Eggert, 1962). Cross sections of young stems consist of a central pith canal with a few parenchyma cells near the periphery (FIG. 10.41). Surrounding the pith are a variable number of collateral vascular bundles, each with a conspicuous protoxylem canal (carinal canal) (FIG. 10.42). Protoxylem canals occur only in the aerial axes and develop by the rupture of the first-formed protoxylem tracheids (FIG. 10.43). The maturation of the primary xylem is generally regarded as an endarch, although in Equisetum, which has the same basic anatomy, mature stems are endarch while developmentally young stems are mesarch. In Calamites, the number of collateral bundles increases by the addition of new strands at each node, so at higher levels of the stem there are a greater number of primary vascular bundles. Surrounding the primary vascular bundles is a zone of secondary xylem that may reach considerable thickness. Tracheids are thick walled and elongate; pitting on the radial walls ranges from scalariform to circular bordered. Large, interfascicular rays extend out between the primary xylem wedges and give the secondary xylem a sectored appearance in cross section (Arnold, 1956) (FIG. 10.43). Tangential sections of the secondary xylem provide a basis for distinguishing three basic structural types (FIG. 10.44) now recognized at the generic level (Andrews, 1952) (FIG. 10.45). Arthropitys is possibly the most common

Figure 10.41

Cross section of large Calamitea (formerly Calamodendron) stem showing wedges of secondary xylem with large rays (arrows) in pith region (Pennsylvanian). Bar  1 cm.

Figure 10.42

Cross section of Arthropitys hirmeri stem showing central canal (Pennsylvanian). Bar  3 mm. (Courtesy BSPG.)

Figure 10.43

Arthropitys hirmeri, transverse section showing carinal canals (arrow) (Pennsylvanian). Bar  1 mm. (Courtesy BSPG.)

CHAPTER 10

X (A)

ST R

X (B)

X

R

X

(C)

X

SPHENOPHYTES

R

351

X

Figure 10.44 Tangential sections of the secondary xylem of three calamite wood types. A. Calamitea. B. Arthropitys. C. Arthroxylon. X  xylem, ST  small tracheids, R  rays. (From Andrews, 1952; in Taylor and Taylor, 1993.)

Figure 10.46 Transverse section of Arthropitys communis

stem (Pennsylvanian). Bar  2.5 mm.

Figure 10.45 Henry N. Andrews.

petrified calamitean stem type and exhibits the simplest tissue organization (FIG. 10.44B). The secondary xylem in Arthropitys is segmented into distinct sectors of wood separated by large interfascicular rays that may vary considerably in size and shape (FIG. 10.41). The genus includes approximately 20 species (FIGS. 10.46, 10.47) and several varieties; the largest Arthropitys stem reported to date (FIG. 10.32) is

a specimen of A. ezonata (FIG. 10.32) from the petrified forest of Chemnitz (Germany) with a diameter of up to 60 cm (Rössler and Noll, 2006). Arthropitys yunnanensis from the Upper Permian of southwestern China shows distinct, uniformly thick growth rings that are continuous through the entire circumference of the secondary xylem (S.-J. Wang et al., 2006). These authors suggest that the growth rings are the result of environmental fluctuations perhaps due to variations in water availability. In Arthroxylon, the secondary xylem consists of two types of cells: elongate tracheids 0.5 mm long and ray cells (FIG. 10.44C). Interfascicular ray

352

PaleoBOtany: the biology and evolution of fossil plants

to fusiform initials. In A. deltoides, fusiform initials also increase in size, but there is also a significant increase in the size and number of interfascicular ray initials. The lack of tangential divisions in fusiform initials lends support to the belief that secondary growth was determinate. EXTRAXYLARY TISSUES. Extraxylary tissues have been reported from small branches and a few large stems. In Arthropitys, the tissues outside the secondary xylem are of two principal types based on the histology of the cells. Cells of the inner cortex include thin-walled parenchyma with some resinous materials. In older stems, this zone is often occupied by elongated resin canals. Cells of the outer cortex are also thin walled, but appreciably smaller than the cells of the inner zone. In A. deltoides (Pennsylvanian), the extraxylary tissues include a compound periderm forming a rhytidome (Cichan and Taylor, 1983). Figure 10.47 Silicified stem of Arthropitys sp. showing the central pith (red) and cylinder of wood with large interfascicular rays (Permian). Bar  2 cm. (Courtesy BSPG.)

cells are greatly elongated axially, some up to 3 mm long. Because of the length of these cells, it is sometimes difficult to distinguish the rays from the wood in tangential section. The third anatomically preserved genus of calamite stems is Calamitea (FIG. 10.41) (formerly Calamodendron; see Rössler and Noll, 2007). The wood of Calamitea is the most complex, consisting of bands of small tracheids, which were previously interpreted as bands of thick-walled fibers, that separate interfascicular rays from zones of secondary xylem (FIG. 10.44A). Some question the value of this artificial separation of taxa based on histologic features of the secondary xylem. Eggert (1962) has suggested that the organization of the wood rays and the type of pitting are the only reliable features useful in identifying woody specimens. Good (1975) suggested that the distinctions between the stem genera are partially developmental and further noted that several different stem genera may have produced the same cone type. There is no other group of fossil cryptogams that produced more secondary xylem than the Carboniferous members of the Calamitaceae. Cichan (1986a) analyzed the production of secondary xylem in two species of Arthropitys and demonstrated that fundamental differences exist in the pattern of cambial activity. In A. communis, the increase in cambial circumference is maintained by an increase in size of the fusiform initials and transformation of some ray initials

GROWTH AND DEVELOPMENT. Several calamitean apices have been found in coal-ball permineralizations (Melchior and Hall, 1961; Good, 1971a). The apical cell is a five-sided pyramid with a roundly rectangular upper surface and four internal triangular surfaces that cut off apical cell derivatives. This configuration is different from both Sphenophyllum and Equisetum, which are characterized by a tetrahedral apical cell with three internal cutting surfaces (Good and Taylor, 1972). Derivatives of the calamitean apical cell were produced in a dextrorse (counterclockwise) direction and apparently matured at a slower rate than those of Sphenophyllum. In addition, the cells produced toward the center of the apical cap in the calamites developed into the pith meristem, whereas the topographically identical cells in Sphenophyllum formed the procambium. In small twigs, the procambial vascular strands matured approximately at the third node, with metaxylem tracheids first observed between the third and fourth nodes. At the node, the vascular bundles fused in a ring. Surrounding the bundles was a cortex of elongated parenchyma and scattered areas of cells with dark contents, the so-called melasmatic tissue. Carinal canals were present at about the fifth node. In general tissue organization, the distal twigs of the calamites appear similar to twigs of Equisetum, except that vallecular or air canals are not present in the cortex of the calamites. In general, the manner in which the calamites grew tends to parallel the epidogenic and apoxogenic development present in the arborescent lycopsids (Eggert, 1962; Chapter 9). The underground rhizome system and attached major branches were characterized by a large pith surrounded by a large number of primary vascular strands (FIG. 10.48). The

CHAPTER 10

SPHENOPHYTES

353

Figure 10.49 Cross section of root of Myriophylloides williamsonii (Pennsylvanian). Bar  1.5 mm.

Figure 10.48 Diagrammatic three-dimensional reconstruction of the vascular system of a typical calamitean plant showing the progressive decrease in the number of primary vascular bundles. (From Eggert, 1962.)

maximum number of vascular bundles is believed to have occurred in aerial stems that developed directly from the rhizome. At these basal levels, there was also abundant secondary xylem. Lateral branches were smaller and had fewer vascular bundles than the primary axes at the level of their origin. The decrease in size of the primary body continued with each successive order of branching, resulting in minute branches with only a few vascular strands and no pith (FIG. 10.48). The number and size of the leaves are also related to the size of the stem, with larger leaves produced on the larger stems. A similar pattern of development was present in the roots, but the branch roots did not greatly increase in size as they grew. In all instances, the stele of the lateral roots was smaller than that of the parent root and had fewer primary vascular strands. ROOTS. Structurally preserved calamitean roots are described under the generic names Astromyelon, Myriophylloides

Figure 10.50 Cross section of the calamitean rootlet Zimmermannioxylon multangulare (Pennsylvanian). Bar  3 mm. (Courtesy BSPG.)

(FIG. 10.49), Asthenomyelon, and Zimmermannioxylon (FIG. 10.50) (Leistikow, 1962), with Astromyelon the most common permineralized form. Pinnularia and Myriophyllites (FIG. 10.51) are used for impression and compression roots (Jongmans, 1911). Morphologically, the roots differ from stems in lacking nodes and branching in a more irregular manner. Histologically, roots of Astromyelon (FIG. 10.52) appear somewhat similar to aerial stems. The root consists of

354

PaleoBOtany: the biology and evolution of fossil plants

a parenchymatous pith surrounded by a variable number of bundles that do not anastomose; maturation of the primary xylem is mesarch in Astromyelon, but mesarch or endarch in Myriophylloides, mesarch in Asthenomyelon, and exarch in Zimmermannioxylon (Leistikow, 1962). Carinal canals are not present in the roots (FIG. 10.52), and the vascular rays are greatly reduced. Secondary xylem is present, but in far smaller amounts than in the stems. In very small roots, the cortex is three parted, with a middle cortex composed of multiseriate trabeculae with lacunae between them. In more mature roots, two types of thin-walled tissue are present immediately outside the secondary xylem. Zones of presumably secondary phloem alternate with primary parenchyma that is opposite the primary xylem strands. Outside these areas are groups of cells with amorphous contents. Cells of the phloem are in radial files, suggesting that they represent sieve elements derived from fusiform cambial initials, although no cambial cells have been detected. Unlike other vascular cryptogams, it appears that the vascular cambium of Astromyelon was bifacial (Wilson and Eggert, 1974). A large amount of secondary cortical parenchyma was produced by a proliferation of cortical cells, and there is some evidence to suggest that a small amount of periderm was developed in very old roots as well. In some specimens it appears that branch roots were produced over a long period of time after a considerable amount of secondary xylem was present in the parent stele. It is not known whether they were produced from a pericycle or from the vascular cambium. In addition, branch roots appear to have maintained a relatively small primary body. Zimmermannioxylon-type roots lack a central pith and secondary xylem (Leistikow, 1962). LEAVES. Calamitean foliage is typically placed into two morphogenera, Annularia (FIGS. 10.53, 10.56) and Asterophyllites (FIGS. 10.54, 10.55). Leaves of Asterophyllites are needlelike and borne in whorls of 4–40. They arch steeply upward from their position on the stem and overlap the leaves of whorls above. Asterophyllites charaeformis is an impression–compression species that includes forms with 4–10 leaves per whorl. Each leaf is ⬃3 mm long and is wide at the base. Petrified leaves of A. charaeformis have been discovered in coal balls from Britain and North America, and these provide anatomical information about this species (Good, 1971a). In cross section, the leaf varies from three- to five-sided in outline, with a median ridge on the abaxial surface. Leaf shape appears to be determined in part by position on the plant. The single vascular bundle consists of a small group of tracheids surrounded by a sheath of thin-walled cells that may represent phloem. Thick-walled

Figure 10.51

Myriophyllites gracilis, compressed calamitean roots with lateral rootlets (Pennsylvanian). Bar  2 cm. (Courtesy BSPG.)

Figure 10.52 Cross section of Astromyelon. Note the absence of carinal canals (Pennsylvanian). Bar  2 mm. (Courtesy BSPG.)

fibers are present on the adaxial side of the vascular bundle; thin-walled parenchyma surrounds the vascular strand in other areas. This sheath has been suggested as a temporary storage site for assimilation products before they reached

CHAPTER 10

SPHENOPHYTES

355

Figure 10.54

Distal end of calamite stem (arrow) showing numerous attached, elongated leaves (Pennsylvanian). Bar  2 cm.

Figure 10.53 Axis with eight whorls of Annularia leaves (Pennsylvanian). Bar  1 cm.

the phloem. Columnar mesophyll parenchyma makes up the remainder of the leaf. Cuticles of some leaves possess stomata randomly arranged on all surfaces. The stomata are simple, consisting of two bean-shaped guard cells surrounded by two subsidiary cells. The long axis of the stoma

parallels the long axis of the leaf. Multicellular hairs are also present on some leaves. Calamites rectangularis includes anatomically preserved leaves of the Asterophyllites type from Upper Pennsylvanian coal balls (Good, 1971a). Complete leaves are ⬃4 mm long and rectangular in cross section. The stomata have

356

PaleoBOtany: the biology and evolution of fossil plants

conspicuous ridges that are associated with the common walls of the guard and subsidiary cells. It has been suggested that there was an increase in the size of Asterophyllites leaves (FIG. 10.55) through time in the Pennsylvanian, although such a trend has not been correlated with stem remains. The leaves of Annularia are lanceolate to spatulate in shape and sometimes appear to be fused proximally to form a shallow disk where they attach to the stem (FIG. 10.56). There are numerous species and the number of leaves ranges from 5 to 32 per node (Abbott, 1958). Unlike Asterophyllites, whorls of Annularia leaves do not overlap from node to node (FIG. 10.54). Individual leaves are variable in length, with those of A. stellata almost 8 cm long. Annularia hoskinsii is a Middle–Late Pennsylvanian form that is known from permineralized specimens from the Eastern Interior Basin of North America (Good, 1976). The leaves are borne on axes up to 4 mm in diameter, which are characterized by carinal canals, a multilayered cortex and an absence of secondary xylem. Leaves are ⬃3 mm wide and 0.8 mm thick. The shape of the leaf is convex abaxially, with the sides overhanging. The vascular bundle consists of numerous barrel-shaped tracheids surrounded by a bundle sheath that includes fibers on the adaxial surface. The remaining tissue consists of palisade parenchyma. Stomata are confined to the adaxial surface and aligned obliquely with the long axis of the leaf. Based on common occurrence and similar tissue organization, A. hoskinsii has been suggested as the foliage type of the plant that produced Calamocarpon cones (see below). An interesting characteristic of compression–impression specimens of Annularia is the consistent orientation of the leaf whorls, which typically are preserved and compressed into a single plane (FIG. 10.56). There has been much speculation as to whether such a configuration was the result of compression and deformation during the fossilization process, or whether the plants actually produced leaf whorls obliquely attached to the stems, possibly to maximize photosynthetic surface. The latter hypothesis is strengthened by the discovery of obliquely positioned nodal diaphragms in permineralized calamitean twigs which had not been deformed during preservation (Good, 1976). In order to truly maximize photosynthetic surface, however, leaves would be borne at varying angles, depending on the orientation of the branches on the plant. One would then expect to encounter fossil leaves preserved in different planes. As this does not appear to be the case, we must conclude that Annularia leaves are borne at a predetermined angle on the stem, regardless of their position on the plant.

Figure 10.55 Several whorls of Asterophyllites leaves (Pennsylvanian). Bar  2 cm.

CHAPTER 10

SPHENOPHYTES

357

Figure 10.57 Lobatannularia leaves (Permian). Bar  1 cm.

Figure 10.56

Numerous whorls of Annularia sphenophylloides leaves (Pennsylvanian). Bar  2 cm.

OTHER CALAMITEAN LEAVES. Dicalamophyllum (Florin, 1939a) is another calamitean foliage type that in many features is intermediate between Annularia and Asterophyllites. Structurally preserved forms contain two small furrows on the abaxial surface, with stomata confined to the furrows. Lobatannularia is principally a Permian genus found in the Cathaysia flora of East Asia (Kon’no and Asama, 1950; Sun, 2006), although it has also been reported in the Triassic (Ju and Lan, 1986; Kim and Kimura, 1988). There may be up to 40 oblanceolate leaves per whorl. Each whorl of leaves is divided into half whorls (FIG. 10.57), each containing up to

20 separate leaves. Leaves of Lobatannularia are anisophyllous. The leaf whorl at the apex of the stem is not divided into lobes. It has been suggested that some of the Mesozoic Lobatannularia specimens belong to Neocalamites (Kon’no and Naito, 1960). Lobatannularia-like foliage from the Permian of Australia and Tibet (Xizang) has been assigned to the genus Austroannularia (Rigby, 1989), and similar foliage from the Middle Triassic of the Basin Creek Formation in New South Wales, Australia, has been given the name Nymbolaria tenuicaulis (Holmes, 2000). However, there are few differences between these taxa and the Cathaysian Lobatannularia. Sun (2006) regarded Lobatannularia as able to adapt to a wide range of habitats (eurytopic) and thus a cosmopolitan element. Daubreeia is a morphogenus used for impression– compression fossils of isolated, medium- to large-sized leaf whorls from the uppermost Pennsylvanian of Europe and North America (Gillespie and Clendening, 1966; Van Amerom and Kabon, 2001; Van Amerom et al., 2003). Daubreeia whorls usually consist of four, entire-margined and laterally fused, fan-shaped leaves (FIG. 10.58), each up to 20 cm long. Leaves are sometimes distally incised and characterized by a

358

PaleoBOtany: the biology and evolution of fossil plants

Figure

10.59 Cluster (Pennsylvanian). Bar  1 cm.

of

attached

calamite

cones

Figure 10.58

Daubreeia, nodal whorl of fan-shaped leaves. (From Van Amerom et al., 2003.)

raised midrib and rectilinear to slightly curved lateral veins. In most specimens, the leaves are arranged around a small, more or less central point of attachment, whereas D. pateraeformis (originally described as Kahleria carinthiaca) shows leaf arrangement around a hollow stem (Boersma and Fritz, 1984). Based primarily on this, Van Amerom et al. (2003) suggested that the affinities of Daubreeia may be with the Equisetales. Much larger Daubreeia specimens have been compared to prophyllar sheaths (ochreoles) that occur around the branch buds in modern Equisetum (Hauke, 1987). REPRODUCTIVE BIOLOGY. Calamitean cones (FIG. 10.59) encompass various morphologic types and have been useful in developing whole plant concepts for the calamites. They vary considerably in size and degree of morphologic complexity, especially in features associated with the sporangiophore. Compression (FIG. 10.60) and impression specimens suggest that cones were borne in various ways,

Figure 10.60

Fertile calamite branch bearing several small cones (Pennsylvanian). Bar  1 cm.

CHAPTER 10

including singly, in clusters at nodes, or on specialized branches (Jongmans, 1911; Rössler and Thiele-Bourcier, 1999). Although much is known about some aspects of the reproductive biology of the calamites, nothing is known about how many cones were produced by a single plant, and exactly where the cones were positioned on the plant. Computer modeling of the architecture and growth stages of the Calamites multiramis plant (C. multiramis stems, Annularia stellata foliage (FIG. 10.61), and Calamostachys tuberculata cones (FIG. 10.62)) suggest that this calamite was monocarpic and produced reproductive structures only during the mature stages (Daviero and Lecoustre, 2000). All calamite cones consist of alternating whorls of sporangiophores and bracts (FIG. 10.63). Many are apparently homosporous, producing spherical spores with elaters of the Calamospora type. Heterospory in calamites was first documented by Hartung (1933) (FIG. 10.64) based on in situ spores isolated from 15 different cones, and has since also been reported from several other cone taxa (Baxter, 1963). Paracalamostachys cartervillei is a Middle Pennsylvanian impression–compression cone that reaches ⬃1.7 cm in length (Hibbert and Eggert, 1965). The cones are borne in clusters at each node of the plant. Approximately six sporangiophores are produced per whorl, with each bearing four sporangia; the number of bracts is approximately double the number of sporangiophores. Spores are of the Calamospora type, smooth-walled, and range from 40 to 100 µm in diameter. Despite the large size range of the spores, the cones are regarded as homosporous. An unusual specimen of Paracalamostachys from the Westphalian A or B of Britain is P. spadiciformis (Thomas, 1969). This species is ⬃9 cm long and contains 16 bracts. Sporangiophore number is variable, although six appears to be the most common. The position of the sporangiophores suggests that they arose from near the axil of the bract whorl. The individual cones are borne at the distal end of a leafy branch and are of different sizes, possibly reflecting different stages of cone development. In addition, the presence of Calamospora spores in two major size classes suggests that the cones may have been bisporangiate. If Paracalamostachys were structurally preserved, it would probably be assigned to the genus Calamostachys, which currently includes considerable morphologic variability (FIGS. 10.62, 10.63, 10.65), as well as forms that are both homosporous (monosporangiate) and heterosporous (bisporangiate). Another cone displaying marked spore size differences comes from the Early Permian of Thuringia, Germany, and has been named Paracalamostachys heterospora (Remy and Remy, 1958).

359

SPHENOPHYTES

Figure 10.61

Annularia stellata leafy branch (Pennsylvanian). Bar  6 cm. (Courtesy BSPG.)

5 mm

Figure

10.62 Calamostachys Bar  5 mm. (Courtesy M. Barthel.)

tuberculata

(Permian).

360

PaleoBOtany: the biology and evolution of fossil plants

Figure

10.63 Immature calamite cone (probably Calamostachys tuberculata) with closely spaced nodes bearing whorls of bracts and sporangia. Bar  1 cm.

Figure 10.64 Wolfgang Hartung. (Courtesy D. Remy.)

One of the largest species in the morphogenus Calamostachys is C. americana (FIG. 10.66), from the Upper Pennsylvanian of North America (Arnold, 1958). Specimens are at least 12 cm long and include both mono- and bisporangiate forms. The peduncle of the cone contains up to 30 primary bundles arranged in pairs. Sterile whorls contain 40–45 bracts that are laterally fused to form a cup. Whorls of sporangiophores are inserted on the cone axis midway between the bract nodes; and can include up to 30 sporangiophores. Spores (Calamospora) are spherical, trilete, and possess a perispore (Good and Taylor, 1975). Bisporangiate specimens

contain elater-bearing microspores that range from 70 to 118 µm in diameter; megaspores vary from 140 to 275 µm and are enclosed by an unornamented perispore. Probably the most commonly encountered and the bestknown Calamostachys species is C. binneyana (FIG. 10.67). This taxon is common in coal measures in Europe and is also known from the Pennsylvanian of North America (T. Taylor, 1967a). Specimens from eastern Kentucky are ⬃1.6 cm long and 3.1 mm in diameter (FIG. 10.68), whereas some of the European specimens are up to 3.5 cm in length.

CHAPTER 10

SPHENOPHYTES

361

S A

Figure 10.66 Cross section of Calamostachys americana cone showing central axis (A), sporangium (S) containing numerous spores, and outer bract tips (arrows) (Pennsylvanian). Bar  1 cm.

Figure 10.65 Compression specimen of Calamostachys (Pennsylvanian). Bar  1 cm.

The cone axis consists of a hollow pith surrounded by a ring of endarch vascular bundles and a narrow zone of secondary xylem tracheids. Regularly spaced, alternating whorls of 18–22 bracts and a variable number of sporangiophores are borne on the cone axis (FIG. 10.68). The bracts are basally fused to form a shallow disk. Stomata are confined to the adaxial surface of the bract. Sporangiophore number appears to be dependent on the level of the section (FIG. 10.69), but 12 is the most common number. The distal portion of each sporangiophore is flattened into a cruciate-shaped structure, with each arm bearing a large, pyriform sporangium (FIG. 10.70). The sporangium wall is a single cell layer thick, with the sporangial wall ornamented by inward-facing buttresses that appear to be a feature of many sphenophyte fructifications. Immature cones of C. binneyana indicate that during cone ontogeny, the shape of the sporangiophore was highly variable (Good, 1971b). In C. binneyana, the spores (Calamospora) are smooth walled (FIG. 10.71), trilete, and possess three coiled elaters (FIG. 10.80). Specimens of C. inversibractis are distinguished by adaxially convex bracts with two furrows on the abaxial surface (Good, 1975). In cross section, the bracts appear similar to leaves of Dicalamophyllum. This association is further strengthened by the consistent occurrence of these two genera in the same petrifactions. Morphologically aberrant (teratological) forms of Calamostachys cones have been reported sporadically. For example, Barthel (1980a) and Kerp and Fichter (1985) described specimens of C. tuberculata from the Rotliegend (Upper Pennsylvanian–Lower Permian) of Germany in which the

362

PaleoBOtany: the biology and evolution of fossil plants

A

A

Figure 10.68 Cross section of Calamostachys binneyana (Pennsylvanain). Bar  2 mm.

S

A

S

Figure 10.69

Cross section of Calamostachys binneyana axis (A) with attached sporangiophore showing sporangia (S). (Pennsylvanian). Bar  0.5 mm.

Figure

10.67 Slightly oblique longitudinal section of Calamostachys binneyana cone showing sporangiophore axes in cross (A) and longitudinal section and sporangia in between (Pennsylvanian). Bar  5 mm.

cone axis dichotomizes proximally to form a forked cone. Another aberrant C. tuberculata cone from the Westphalian D (Pennsylvanian) of western Germany is characterized by alternating whorls of sporangiophores and, instead of small bracts, normally shaped leaves of the Annularia stellata type (Krings and Sommer, 2000). Morphological aberrations resembling those seen in the fossil cones have also been described in modern Equisetum (Schaffner, 1933).

CHAPTER 10

SPHENOPHYTES

363

B

Figure 10.71

Calamospora spore showing well defined suture (Pennsylvanian). Bar  3 µm.

B

Figure 10.70

Paradermal section of Calamostachys binneyana cone showing sporangiophores in cross section and bract whorls (B) (Pennsylvanian). Bar  5 mm.

Palaeostachya is similar to Calamostachys but differs principally in that the sporangiophores are attached to the cone axis at an oblique angle (Fig. 10.72). Some have suggested that the Calamostachys cone type evolved from an ancestor with Palaeostachya type cones (Remy and Remy, 1975b). Palaeostachya andrewsii is a Middle Pennsylvanian species that includes both monosporangiate and bisporangiate types (Baxter, 1955, 1962). Specimens exceed 15 cm in length and include 6–10 pairs of endarch vascular bundles that extend through the cone axis. Bracts are rectangular in cross section and alternate from node to node. The number of sporangiophores per whorl in P. andrewsii ranges from 12 to 20 and traces of the sporangiophores originate just above the bract traces. From this level they extend through the cortex to about the middle of the internode above before recurving downward and extending into the base of the sporangiophore. Specimens of P. decacnema are ⬃6.5 cm long and characterized by a stele with four to five lobes (Delevoryas, 1955a). In this cone, sporangiophores and bract traces originate from paired bundles that traverse the length of the cone axis, with the traces to the sporangiophores extending directly out into the sporangium-bearing unit. Palaeostachya vera is known from the Mississippian (Lower Coal Measures) of England (Seward, 1898). The species may be distinguished by the degree of lateral bract fusion, low ratio of bracts to sporangiophores, length of bracts, and

364

PaleoBOtany: the biology and evolution of fossil plants

size of spores. All species of Palaeostachya possess trilete spores with three, distally attached elaters. The heterosporous species, P. andrewsii, has megaspores up to 345 µm in diameter with an encircling perispore. In P. dircei (Gastaldo, 1981), the number of bracts equals the number of obliquely inserted sporangiophores. This Middle Pennsylvanian taxon is preserved by authigenic cementation and is an excellent example of how structural information can be obtained from preservation types that are not permineralizations. The name Huttonia is used for relatively large heterosporous cones (Sternberg, 1837 (FIG. 10.73); Nˇemejc, 1950) that are included by some authors in Palaeostachya (Browne, 1927). Specimens of H. spicata from the Radnice basin (Middle Pennsylvanian, Czech Republic) are up to 20 cm long by 3 cm wide and characterized by nodal whorls of up to 18 lanceolate bracts (Libertín and Bek, 2004). A distinct zone of densely spaced sterile bract whorls occurs in the proximal portion of the cone. Bracts are proximally fused and have small projections oriented downward to form an umbrella-like foliar collar. Sporangiophores extend from the axils of the bracts and bear a single, terminal sporangium 1 mm in diameter. Microspores are spherical to subspherical, trilete, laevigate, 240 µm in diameter, and of the Calamospora type (Hartung, 1933). Megaspores may be correlated with the sporae dispersae taxon C. laevigata. They are spherical, up to 650 µm in diameter, and trilete, with a laevigate exine that is 3–10 µm thick (Libertín and Bek, 2004). Weissistachys is a structurally preserved cone known from a single locality in North America (Rothwell and Taylor, 1971a, b). The axis of this Pennsylvanian cone, originally described as Weissia—a name that was already occupied— consists of a ring of endarch bundles surrounding a parenchymatous pith (FIG. 10.74). Sporangiophores are diamond shaped in cross section, with the distal ends fused and sporangia are attached along the upper and outer surfaces. Spores range up to 75 µm in diameter and possess elaters. Pendulostachys is another structurally preserved cone type that further expands our knowledge of the morphology of calamitean cones (Good, 1975). These cones are large (14 cm long) and basically similar to other calamitean cones (FIG. 10.75). The sporangiophores of P. cingulariformis, however, are unusual, in that they are fused to the lower surface of the bract disk and hang pendantly. Each sporangiophore bears four sporangia and the spores bear elaters. Mazostachys pendulata is a calamite cone that shows an interesting combination of preservation types (Kosanke,

Figure 10.72 Palaeostachya Bar  5 mm. (Courtesy M. Barthel.)

thuringiaca

(Permian).

1955) (FIG. 10.76). The genus is known from an ironstone concretion collected from the famous Mazon Creek locality in Illinois. The specimen consists of a branching system with 15 cones attached at the nodes. Foliage associated with, but not organically attached to, the specimen is assigned to Annularia sphenophylloides. Individual cones are ⬃2.6 cm long and contain half as many sporangiophores (6) as bracts (12). Sporangiophores are inserted just below a whorl of bracts and are characterized by parenchyma tissue near the distal end. In this feature, M. pendulata is similar in organization to the sporangiophores of Weissistachys. Sporangiophores borne immediately below bract whorls also identify specimens of Metacalamostachys (Hirmer,

CHAPTER 10

SPHENOPHYTES

365

Figure 10.75

Cross section of Pendulostachys cingulariformis cone (Pennsylvanian). Bar  5 mm. (From Good, 1975.)

Figure 10.73

Kasper Maria Graf von Sternberg. (With permission – Gustar Fischer Verlag.)

Figure 10.76

Figure 10.74

Cross section of Weissistachys kentuckiensis (Pennsylvanian). Bar  1.25 mm.

1927). Cones referred to this genus have sporangiophores with a single sporangium, but M. (Calamostachys) dumasii apparently could have as many as four sporangia per sporangiophore (Jongmans, 1911 (FIG. 10.77); Barthel, 1989, 2004). Specimens of M. dumasii from the Rotliegend of

Robert M. Kosanke.

Thuringia (Germany) are ⬃4 cm long and 3 mm in diameter, with individual bracts 2–3 mm long. Sporangiophores are proximally fused with the bracts and bear two sporangia, each ⬃1 mm in diameter (Barthel, 2004). Metacalamostachys dumasii cones were borne on narrow axes assigned to the stem morphogenus Calamites gigas (see above), and occur in nodal whorls of up to eight cones. After spore liberation, the empty sporangia were shed, whereas the cones remained on the plant (Kerp, 1984a). Such de-sporangiated cones had

366

PaleoBOtany: the biology and evolution of fossil plants

Figure 10.77 Willem J. Jongmans.

been misinterpreted as sterile foliage in the past and given the name Asterophyllites dumasii (Zeiller, 1890). In the Pennsylvanian compression genus Cingularia, the bracts and sporangiophores have a relationship similar to that seen in Metacalamostachys (Weiss, 1870a) (FIG. 10.78). In Cingularia typica, the sporangiophores are flattened and strap shaped, with four large sporangia attached in a pendant manner to the lower surface (Remy, 1959). Sterile foliage believed to belong to C. typica has been described from the Carboniferous of Belgium as Annularia pilosa (Stockmans and Willière, 1945). In C. cantrillii, all sterile appendages are absent and the whorls of sporangiophores are borne at intervals of ⬃8 mm on the cone axis. SPORES An interesting example of heterospory within the Equisetales is found in the cone Calamocarpon (Baxter, 1963; Leisman and Bucher, 1971) (FIG. 10.79). The genus has a wide geographic distribution within North America and is known throughout the Pennsylvanian. Cones of C. insignis are both mono- and bisporangiate. In the bisporangiate cones, there is no transitional zone from one sporangial type to the other. Megasporangia appear to have been produced in the more distal regions. Specimens range from 4 to 12 mm in diameter and are at least 8 cm long. Sterile bracts alternate with sporangiophores. Microsporangia are a single cell layer thick with pegs extending into the lumen of each cell. Spores are 30–40 µm in diameter and characterized by three elaters attached to a pad on the distal surface (FIG. 10.80) (Good and Taylor, 1974; Kurmann and Taylor, 1984). Spores of this type were initially described under the sporae dispersae binomial Elaterites triferens (Baxter and Leisman, 1967) (FIG. 10.81). Megasporangia are rectangular and up to 3 mm

Figure 10.78 Christian Ernst Weiss.

Figure 10.79 Cross section of megasporangiate region of Calamocarpon insignis cone (Pennsylvanian). Bar  1 mm.

long (FIG. 10.82). The wall is a single cell layer thick, but peglike processes are absent. In some specimens, sterile tissue, especially at the proximal and distal ends, separates the sporangium wall from the megaspore. In C. insignis, each megasporangium bears a single, large (2.7 mm  0.7 mm) megaspore. Attempts to macerate the megaspore from the sporangium have been unsuccessful, with only small fragments released during acid treatment. Ultrathin sections of the megaspore indicate that it is constructed of several thin layers. Cellular megagametophytes have been observed (Baxter, 1964a).

CHAPTER 10

SPHENOPHYTES

367

Figure 10.82

Section of Calamocarpon insignis megasporangium (Pennsylvanian). Bar  1 mm.

Figure 10.80

Several Elaterites triferens spores macerated from sporangium of Calamostachys binneyana. Each spore has three elaters (Pennsylvanian). Bar  225 µm.

Figure 10.81 Robert W. Baxter.

Good and Taylor (1975) demonstrated that many of the Calamospora-like spores of the calamiteans possess distinctive elaters. In situ spore studies indicate that the elaters may assume various configurations according to the ontogenetic stage of the spore. At least three different spore types are known to have been produced by calamitean cones. The most common of these is Calamospora (FIG. 10.71). This type of spore is distinguished by its thin, smooth wall and thickened areas between the arms of the trilete mark. These spores have lost the elaters either through development or by mechanical separation. The second spore type commonly encountered in cone spore macerates is Vestispora. In this grain type, the elaters tightly surround the spore body and are visible only as oblique striations on the surface of the grain. These spores represent an immature ontogenetic stage in which the elaters are not fully developed. The third stage in calamite spore ontogeny is given the name Elaterites. This spore has three elaters that are attached to a triangular pad on the distal surface (FIG. 10.83). The fragile nature of the elaters is the principal reason why spores of this type do not constitute an important component of sporae dispersae assemblages (Turner, 1991). Another interpretation is offered by Ravn (1983), however, who suggested that Vestispora is a useful biostratigraphic microspore and that Calamospora, Vestispora, and Elaterites are not related ontogenetically but may be related phylogenetically. Serret and Brousmiche (1987) have examined a large number of sphenophyte fructifications from the Sarre– Lorraine coal field (western Europe) and have correlated foliage with the principal types. Cone specimens with spores in situ indicate that, although all produced spores of the Calamospora type, some sporangia also contained spores belonging to the dispersed spore genera Cyclogranisporites, Vestispora, and Auroraspora. Libertín and Bek (2006) recently analyzed the microspores obtained

368

PaleoBOtany: the biology and evolution of fossil plants

Figure 10.83 Elaterites triferens spore showing three coiled elaters (Pennsylvanian). Bar  100 µm.

from a single specimen of Calamostachys incrassata from the Kladno-Rakovník Basin (lower Bolsovian  Middle Pennsylvanian) in the Czech Republic. They found that most spores are sheathed with a delicate pseudosaccus-like structure and resemble taxa in the sporae dispersae genera Auroraspora, Remysporites, Callialasporites, Perotriletes, Phyllothecotriletes, and Diaphanospora. Only a few spores lack a pseudosaccus-like envelope and can be directly correlated with forms in the sporae dispersae genus Calamospora. What is perhaps most interesting is that none of the spores described by Serret and Brousmiche (1987) and Libertín and Bek (2006) were of the Elaterites type. It appears highly probable that all calamitean fructifications produced spores with elaters. The fact that elaters have not been reported in some cases may be a reflection of the developmental level of the spores at the time of fossilization or the failure to closely examine spores that appear wrinkled and folded, as some folds may actually represent elaters. One additional reason why elaters may have been overlooked concerns the ultimate function of these structures. Elaters in the living genus Equisetum remain attached to the sporoderm throughout the life of the spore, but the calamite spores may have possessed elaters that broke off during the dispersal phase. In this context, they may have functioned in a manner similar to that of sporangial elaters in certain bryophytes. Some support for this hypothesis is derived from the large numbers of calamitean sporangia that contain detached and broken elaters but no spores. The primary aim of paleobotany has always been the reconstruction of the complete plant. This process depends

on accurate morphologic and anatomic descriptions based on many specimens. An intriguing aspect of such reconstructions is the assignment of reproductive parts to vegetative remains. Good’s (1975) study of calamite cones is an excellent demonstration of how reproductive organs may be related to the probable parent plant. For example, the histology of the cone peduncle and anatomy of the sterile bracts of Calamostachys americana are almost identical with the leaves that have been described as Calamites rectangularis and stems of Arthropitys communis var. septata. Uniseriate, circular-bordered pits known to be present on Arthropitys stems, and foliage assigned to the genus Dicalamophyllum have been used to suggest the relationship of Calamostachys inversibractis with these vegetative organs. Calamocarpon cones are thought to have been produced on large vegetative stems assignable to the genus Calamitea, which had smaller aerial branches of the Arthroxylon type. Relationships that are based on certain similarities in the anatomy and morphology of the detached parts are, of course, strengthened when the different taxa can be identified from the same locality. TCHERNOVIACEAE AND GONDWANOSTACHYACEAE

The Tchernoviaceae (or Tschernoviaceae) and Gondwanostachyaceae are two groups of predominantly late Paleozoic Equisetales from the Carboniferous and Permian of Angara and Gondwana, respectively (Meyen, 1967, 1971, 1982a). Members of these families may resemble plants in the Calamitaceae and Equisetaceae with regard to certain aspects of the vegetative morphology but are set apart from them by the lack of reproductive structures arranged in cones or strobili, that is, a nonstrobilar organization. Although there has been some debate as to the nomenclatural validity, inventory, and status of these families, we use them here because they are generally suitable for equisetalean plants displaying distinct sporangiophore morphologies and spatial distribution patterns. VEGETATIVE BODY The vegetative parts of members of both families are linked to the widely distributed morphogenus Phyllotheca (FIG. 10.84), which has an extensive geologic range traceable from the Pennsylvanian through the Permian, where remains are most abundant, into the Cretaceous; there are a few other, less widely distributed morphotaxa as well. Phyllotheca has unbranched or monopodially branched axes, sometimes with swollen nodes. First-order branches bear whorls of up to 40 single-veined leaves that are fused basally to form a narrow sheath. Branches are ornamented with conspicuous ridges and furrows that do not alternate

CHAPTER 10

SPHENOPHYTES

369

whorl than in Phyllotheca); the leaves are variously fused together to form shallow, saucerlike sheaths extending from one-quarter to almost the full length of the leaves. In R. bengalensis the leaf sheaths are up to 10 cm in diameter and consist of 10–80 fused leaf segments. Raniganjia is known from throughout Gondwana, including India, Australia, South Africa, South America, and Antarctica. Although the affinities of the genus are generally regarded as being with the Sphenophyta, nothing is known about the reproductive parts. Koretrophyllites (Radchenko, 1955) is predominantly a Permian genus that is known principally from Russia but has also been reported from the uppermost Mississippian of Germany (Van Amerom et al., 1997). It includes forms with elongated leaves that are not fused into a cup at the node. Sporangiophores arise between whorls of leaves near the base of internodes. Meyen (1971) believed that all the species of this genus should be included in Phyllotheca.

Figure 10.84 Suggested reconstruction of two leaf whorls of Phyllotheca indica (Permian). (From Taylor and Taylor, 1993.)

from node to internode. Axial vascular bundles dichotomize and anastomose near the nodes to form the so-called nodal loops that supply the vascular bundles to the leaves. The diameter of the stems (2 cm) and abundant tracheids with bordered pits in P. indica suggest that some secondary xylem was produced (Pant and Kidwai, 1968). In P. australis, the leaves are fused to form a pronounced cup (Townrow, 1955). Umbellaphyllites is an Early Permian foliage genus that combines vegetative features of Phyllotheca and Annularia and is regarded by some as an intermediate between the two (Rasskasova, 1961). Another Permian (Lower Gondwana) sphenophyte foliage morphotype similar to Phyllotheca is Raniganjia (Pant and Nautiyal, 1967a; Anderson and Anderson, 1985; McLoughlin, 1992). The genus is known from shoots bearing whorls of numerous leaves (generally more leaves per

REPRODUCTIVE BIOLOGY Tchernoviaceae and Gondwanostachyaceae differ from Calamitaceae and Equisetaceae by the nonstrobilar organization of the reproductive structures. In members of the Tchernoviaceae, bractless peltate sporangiophores extend from the unbranched primary shoot or from normally developed leafy lateral branches. The sporangiophores occur in regular clusters and form distinct intercalary or terminal fertile zones that are internodal. For example, in Equisetinostachys sp. (FIG. 10.85) from the Permian (Kungurian Stage) of the middle Fore-Urals (Russia), the primary axes produce multilevel reproductive structures composed of several contiguous fertile internodes separated from one another by whorls of sterile leaves (FIG. 10.86) (Naugolnykh, 1998, 2004a). In Paracalamitina striata from the uppermost Lower Permian Pechora Cis-Urals (Russia), a single fertile zone,
2 cm long, is present in a near-terminal position in almost every leafy lateral branch of the penultimate and ultimate branches (Naugolnykh, 2002b). This fertile zone is positioned along a single internode and consists of six longitudinal rows of fertile units, each now is composed of 9–10 stalked sporangiophores. Sporangiophore heads are usually subrectangular and typically bear four sporangia on the abaxial side, but sporangiophores with more than four sporangia have also been recorded. Typically one whorl of sterile leaves occurs above the fertile internode. Naugolnykh (2002b, 2004b) provided a whole-plant reconstruction of the P. striata plant (FIG. 10.87) based on a reassessment of specimens originally described by Neuburg (1964) as Paracalamitina

370

PaleoBOtany: the biology and evolution of fossil plants

Figure 10.85 Equisetinostachys sp. showing fertile internodes and sterile leaves (arrows) (Permian). Bar  1 cm. (Courtesy S. V. Naugolnykh.)

striata (primary axes), Phyllotheca striata (foliage), Tchernovia striata (sporangiophores), and Sciadisca petschorensis (nodal diaphragms), as well as newly collected material. Paracalamitina striata is interpreted as a relatively small plant up to 70 cm high, with leafy laterals, 15–20 cm long and once-branched, given off from the distal portion of the main stem. The plant probably grew along the shores of lakes and freshwater lagoons. The reproductive structures of members of the Gondwanostachyaceae have been defined based on compression specimens of Gondwanostachys (Phyllotheca) australis (Townrow, 1955), the type species of the family and genus Gondwanostachys (Meyen, 1971). Bractless sporangiophores are borne in single whorls of six to eight in the lower part of fertile internodes and thus apparently are protected by the enveloping leaf whorl below. The sporangiophores are unusual in that each is twice dichotomized, with each of the branch ends bearing four recurved or pendant sporangia. Several plants have been described that resemble members of the Tchernoviaceae or Gondwanostachyaceae but

Figure 10.86

Hypothetical model of the evolution of the strobilus and peltate sporangiophore in modern Equisetum from a late Paleozoic ancestor. A. Equisetinostachys. B. Paracalamitina striata. C. Equisetites arenaceus. D. Neocalamites carreri. E. Equisetum arvense. (From Naugolnykh, 2004a.)

do not fit entirely into either of the families. For example, Cruciaetheca (FIG. 10.88), a genus of equisetalean sphenophytes from the Lower Permian of Argentina (Cúneo and Escapa, 2006), displays features intermediate between the families. Specimens of C. patagonica from the Río Genoa Formation, Chubut Province, consist of articulated, striate, and two- to three-times ramified axes with branches given off in whorls. Ultimate axes bear nodal whorls of 12–14 leaves, which are up to 6.1 cm long and 0.3 cm wide, singleveined, and laterally fused in their proximal portion forming a sheath. Reproductive structures occur on axes of the ultimate order and are arranged into series of contiguous

CHAPTER 10

SPHENOPHYTES

371

Figure 10.88 Cruciaetheca patagonica showing fertile internodal zones (arrow) (Permian). Bar 1 cm. (Courtesy N. R. Cúneo and I. Escapa.)

Figure 10.87

Suggested reconstruction of Paracalamitina striata (From Naugolnykh, 2002b.)

fertile internodes bounded by unmodified leaf whorls (fertile zones) (FIG. 10.88). Each fertile internode consists of three to five whorls of four to six cruciate sporangiophores that are positioned in the distal and middle portions of the internode, immediately below the leaf whorl above. Fertile internodes positioned distally on the axis generally have fewer sporangiophore whorls than more proximally positioned fertile internodes. Each sporangiophore produces four pyriform sporangia. In the second species, C. feruglioi, the fertile internodes consist of only a single whorl of sporangiophores attached immediately below the leaf whorl above. Another Permian equisetalean that has reproductive axes that are leafless and composed of distal internodes consisting of peltate sporangiophores is Peltotheca (Escapa and Cúneo, 2005). The foliage of P. furcata is of the Barakaria type, in which the leaves dichotomize at the tip. The organization of the reproductive parts in the internodal regions of the stems suggests a transitional stage leading to the more compact cone organization seen in geologically younger sphenophytes.

Another plant that does not fit completely into either of these two families is Giridia indica from the Permian of India (Pant et al., 1981) and possibly Argentina (Durán et al., 1997). This fossil consists of Phyllotheca indica-type leafy axes, which, if fertile, bear densely spaced sporangiophores on internodes between normally developed leaf whorls. Sporangiophores are inserted immediately below the leaf whorl above and are neither peltate nor cruciate, but rather several times forked, with single sporangia occurring terminally on the sporangiophore branches.

EQUISETACEAE

Included in the Equisetaceae are plants with fused whorls of leaves at each node, stems usually (but perhaps not always; see Schweitzer et al., 1997) lacking secondary tissues, and cones that generally lack bracts. Far less is known about the vast majority of fossil members of this family than the Calamitaceae, even though they are believed to have coexisted during the Carboniferous (Des Marais et al., 2005). Equisetites is a highly heterogeneous morphogenus that includes casts, impressions, and compressions of stems which morphologically resemble extant Equisetum (FIG. 10.89) (Weber, 2005). Specimens are distinguished from distal branches of calamites partially by the presence of small

372

PaleoBOtany: the biology and evolution of fossil plants

Figure 10.89 Equisetum lyellii. (Courtesy J. Watson.)

leaves fused into a sheath and the apparent absence of secondary xylem. The genus has been reported from numerous localities worldwide, including Europe (Halle, 1908; Kelber and Van Konijnenburg-Van Cittert, 1998), North America (Black, 1929; DiMichele et al., 2005b), Middle and South America (Villar de Seoane, 2005b; Weber, 2005), Antarctica (Cantrill and Hunter, 2005), China (Sze, 1933), and New Zealand (McQueen, 1954). Although the majority of specimens are known from Triassic or younger rocks, there have been several reports of the genus dating back to the Carboniferous. Equisetites hemingwayi comes from the Middle Coal Measures, Westphalian A–C (Early Pennsylvanian) of Yorkshire and consists of jointed stems, some of which bear oval, bractless cones (Kidston, 1892b). The cones are constructed of groups of hexagonal plates that have been regarded as the distal ends of the sporangiophores,

similar to extant Equisetum. Unfortunately, no cuticular remains or spores were preserved. Equisetites muensteri is known from numerous specimens from several Triassic and a few Jurassic localities in Europe (Semaka and Georgesco, 1967). Possibly the most detailed account is that of Harris (1931a), who described the taxon from the Late Triassic of Scoresby Sound in eastern Greenland, based in part on cuticular remains. Stems are ⬃1 cm in diameter, with variable internodal dimensions. Branching specimens are not known, although small scars that alternate with the leaves may indicate the former positions of branches or adventitious roots. Approximately 12 leaves are fused into a sheath at each node, with the number of leaves per node constant regardless of the size of the stem. Cuticle preparations exhibit numerous stomata at the base of each leaf sheath, with more stomata scattered on the stems. Guard cells have a beaded border and two longitudinal rows of pits. Cones occur at the tip of the stem and are constructed of whorls of peltate, hexagonal sporangiophores. Spores in the 40–50 µm size range, lacking elaters, have been recovered from the sporangia. In almost all features, E. muensteri is identical with species of Equisetum with the exception of the spores; however, this may be a reflection of preservation or the developmental stage at the time of fossilization. Perhaps the most completely known species in Equisetites is E. arenaceus from the Upper Triassic of Germany which is based on stems and lateral branches (FIGS. 10.90, 10.91), leaves, nodal diaphragms (FIG. 10.92), and reproductive organs (Kelber and Van Konijnenburg-Van Cittert, 1998; Kelber, 1999). In this form, erect stems arise from a creeping rhizome characterized by short internodes. The upright stems probably grew to 4 or 5 m, with diameters up to 20 cm, and terminated in a dome-shaped apex (FIG. 10.93) (Kelber and Hansch, 1995). Based on the considerable diameter of the stems, it has been suggested by some that E. arenaceus had secondary growth (Schweitzer et al., 1997). Whorls of more than 100 leaves are given off from the stem nodes and leaves are fused proximally to form a leaf sheath. Distal leaf portions are folded and leaf apices terminate in spinelike teeth. The stems produce massive branches and occasional whorls of slender sterile or fertile branches, 4–6 mm in diameter, although the slender sterile branches are usually found detached. In several specimens, narrow adventitious roots extend from the branches, and it has been suggested that these branches may have been effective in vegetative reproduction. Kelber and Hansch (1995) hypothesized that the branches broke off easily and fell onto the swampy substrate, where they eventually grew into a new plant. Fertile branches are relatively short and

CHAPTER 10

SPHENOPHYTES

373

Figure 10.91 Equisetites arenaceus, lateral branches in situ extending from a stem (Triassic). Bar  1 cm. (Courtesy K.-P. Kelber.)

Figure 10.90 Equisetites arenaceus, pith cast of stem (Triassic). Bar  6 cm. (Courtesy K.-P. Kelber.)

bear up to three strobili at the apex (FIG. 10.94) at different stages of maturity. Strobili are obovate (FIG. 10.95) and nonapiculate (not pointed at the apex), up to 35 mm long by 22 mm wide, and composed of up to nine whorls of stalked, peltate (FIG. 10.96) sporangiophores bearing elongate sporangia on the lower surface (FIG. 10.97). Spores are usually 50–60 µm in diameter, globose, often folded, and alete or with a small triradiate mark. The exospore is ⬃1 µm thick, faintly scabrate to microrugulate, and the possible perispore is smooth and ⬃0.5 µm thick. Equisetites aequecaliginosus, from the Upper Triassic Santa Clara Formation of Sonora, Mexico, was recently reassessed based on new specimens. The plant is reconstructed based on stems (6 to 15 cm in diameter) referred to E. cf. arenaceus and cones assigned to the genus

Figure 10.92 Equisetites arenaceus, nodal diaphragm sur-

rounded by the leaf sheath (Triassic). Bar  1 cm. (Courtesy K.-P. Kelber.)

Equicalastrobus (Weber, 2005). The morphogenus Equicalastrobus is used for cylindrical, pedunculate equisetalean cones that consist of a stout axis giving off whorls of peltate sporangiophores with hexagonal heads. A lanceolate, single-veined, leaflike appendage that is directed toward the cone apex extends from a slightly elevated umbo in the center of the head (Grauvogel-Stamm and Ash, 1999). Equicalastrobus

374

PaleoBOtany: the biology and evolution of fossil plants

Figure 10.93 Equisetites arenaceus, stem apex (Triassic). Bar  1 cm. (Courtesy K.-P. Kelber.)

Figure 10.95 Equisetites arenaceus, fertile lateral branch with three strobili (Triassic). Bar  1 cm. (Courtesy K.-P. Kelber.)

Figure 10.94

Suggested reconstruction of Equisetites arenaceus showing branches bearing terminal cones. (Courtesy K.-P. Kelber.)

chinleana was originally described as Lycostrobus chinleana and attributed to the lycopsids (Daugherty, 1941; Retallack, 1997), but reexamination of the cone has demonstrated that it is an equisetalean (Grauvogel-Stamm and Ash, 1999). The sporophylls are described as deciduous, with each consisting of a peltate scale or head attached to a four-angled stalk.

Figure 10.96 Strobilus of Equisetites arenaceus showing surface (Triassic). Bar  1 cm. (Courtesy K.-P. Kelber.)

CHAPTER 10

SPHENOPHYTES

375

Figure 10.97

Equisetites arenaceus, isolated strobilus showing strobilus axis and several peltate sporangiophores (arrows) (Triassic). Bar  1 cm. (Courtesy K.-P. Kelber.)

Other Mesozoic and most Cenozoic sphenophyte fossils have been assigned to the genus Equisetum (FIG. 10.98) for vegetative parts or the morphogenus Equisetostachys for isolated cones consisting of whorls of peltate sporangiophores. Equisetum clarnoi is the name used for petrified stems, 0.8 cm in diameter, from the Eocene of Oregon (Brown, 1975a). The external surface of the stems has ridges and furrows, but leaves are absent from the nodes. Cross sections show a central pith cavity surrounded by a ring of evenly spaced vascular bundles. A fibrous hypodermis, with alternating long and short, inwardly projecting bands, forms a continuous layer around the stem. Stomata are sunken in the hypodermis and occur in a single row. Numerous roots were found in the matrix with the stems, but none were organically attached. Roots are typically 2 mm in diameter and frequently possess root hairs. There can be little doubt that E. clarnoi is correctly assigned to the extant genus Equisetum. In fact, when compared with extant species, E. clarnoi appears almost identical to E. hyemale var. affine. Silicified equisetalean leaf-sheath fragments from the late Miocene–Pliocene Ash Hollow Formation in Nebraska show sunken stomata arranged in two regular rows on each sheath segment and thus were assigned to Equisetum subg. Hippochaete (Thomasson, 1980b). Equisetum fluviatoides is a Paleocene form that is nearly identical in both vegetative and reproductive features to extant specimens of E. fluviatile (McIver and Basinger, 1989a). Another Cenozoic sphenophyte that closely resembles extant species has been reported from the

Figure 10.98 Equisetum arcticum Bar  3 cm. (Courtesy M. Dolezych.)

(Eocene–Oligocene).

Miocene of Lühe, Yunnan Province (southwestern China) under the name Equisetum cf. pratense. This fossil consists of rhizomes with distinct internodes and nodes to which are attached bunches of elliptical, rounded, spindle-shaped, or mucronate tubers (Y.-L. Zhang et al., 2007). Tubers of an Equisetum-like sphenophyte have also been reported from the Paleocene of North America (FIG. 10.99). Compression and impression specimens and pith casts thought to be the stems of Equisetum have been described from rocks as old as the Carboniferous. The only diagnostic features that can be used to distinguish Equisetum or Equisetites from calamites are the presence of leaf sheaths or terminally positioned cones. Compression–impression specimens and pith casts lacking these features may be confused with the distal branches of a calamitean plant. Even if structurally preserved, small calamite branches may superficially resemble an Equisetum stem. If a large branch were fossilized while immature, the absence of secondary xylem and

376

Figure

PaleoBOtany: the biology and evolution of fossil plants

10.99 Tubercules (Paleocene). Bar  5 mm.

of

Equisetum

globulosum

the configuration of the primary body would closely resemble the herbaceous stem of Equisetum.

Figure 10.100

Schizoneura paradoxa, axis with nodal leaf groups (Triassic). Bar  4 cm. (Courtesy K.-P. Kelber.)

FORMS WITH UNCERTAIN AFFINITIES

Schizoneura (originally Convallarites, for nomenclature see Zijlstra et al., 2007) is a characteristic and important element of the Permian flora of Gondwana (Anderson and Anderson, 1985; Rohn and Rösler, 1986) but has also been reported from several Triassic and Jurassic sites in Asia, Africa, Australia, Europe, and North America (Hirmer, 1927; Vladimirovich et al., 1960; Grauvogel-Stamm, 1978; Kelber, 1983; Ash, 1985b). Vegetative parts of this probably herbaceous plant consist of slender striate articulated axes,
1 cm in diameter, that bear 2–4 linear to ovate leaf groups (aggregates of up to 11 laterally fused leaves) at each node (FIG. 10.100). Each leaf group contains several parallel veins that converge at the base (Ash, 1985b). In some specimens, the leaf groups are partly or completely divided into narrow (FIG. 10.101), individual single-veined leaves (Kelber, 1999). Reproductive structures consist of cones (strobili), originally described as Echinostachys (FIG. 10.102). These have been described in organic connection to S. paradoxa vegetative parts from the Grès à Voltzia Formation,

Figure 10.101 Schizoneura paradoxa, leaf groups (Triassic). Bar  2 cm. (Courtesy K.-P. Kelber.)

CHAPTER 10

SPHENOPHYTES

377

Figure 10.102 Schizoneura–Echinostachys paradoxa, reconstruction of a fertile branch. (From Grauvogel-Stamm, 1978.)

Lower Triassic (Buntsandstein) of France (Grauvogel-Stamm, 1978). The cones are ovate to elongate, up to 5 cm long, nonapiculate, stalked, and composed of whorls of peltate sporangiophores arranged around a central axis; sporangia are attached to the sporangiophore stalk in whorls. The S. paradoxa plant was heterosporous and produced microspores in ovate cones assignable to E. oblonga and megaspores in slightly more elongate cones of the E. cylindrica type, which have flat peltate heads on the sporangiophores. Spores are assignable to the sporae dispersae genus Calamospora keuperiana. Neocalamites (FIG. 10.103) is a morphogenus used for certain sphenophytes that were widespread and common elements in Triassic floras (Halle, 1908; Holt, 1947; Kimura et al., 1982; Kelber, 1983, 1999), but specimens have also been reported from Permian deposits (Banerji et al., 1987; Escapa and Cúneo, 2006). Neocalamites differs from other Mesozoic Equisetales such as Equisetites, Equisetum, and Phyllotheca in that the leaves are Annularia-like and do not form sheaths at the nodes. One of the best known species is Neocalamites merianii from the Triassic of southern Germany. Plants are up to 2 m high, with the main stems reaching 2–5 cm in diameter (FIG. 10.104), and lateral branches are produced at the nodes (Frentzen, 1933; Kelber

and Hansch, 1995). Stems and lateral branches have nodal whorls of grasslike, single-veined leaves, each up to 15 cm long. Neocalamites merianii morphologically resembles certain calamites; however, structurally preserved stem specimens indicate that the internal structure corresponds to Equisetum and lacks secondary growth (Kelber and Hansch, 1995). Nothing is known to date about the underground portions or reproductive structures of N. merianii. Vegetative remains of N. hoerensis from the Keuper (Triassic) of Madygen, Middle Asia, are associated with elongated cones that have tentatively been assigned to the morphogenus Neocalamostachys (Dobruskina, 1985b). Small strobili composed of whorled sporangiophores and lacking bracts co-occur with Neocalamites vegetative remains in the Upper Permian La Golondrina Formation (Santa Cruz Province) in Argentina and may have been produced by the same plant (Escapa and Cúneo, 2006). Spaciinodum collinsonii (FIG. 10.105) is one of the few permineralized equisetaceans (Osborn and Taylor, 1989). The specimens, from the lower Middle Triassic Fremouw Formation of Antarctica, consist of jointed stems up 3 mm in diameter with conspicuous vallecular canals in the cortex. There are 12–18 endarch collateral bundles in the internodal

378

PaleoBOtany: the biology and evolution of fossil plants

Figure 10.103

Neocalamites lehmannianus, portion of stem. Bayreuth, Germany (Late Triassic–Early Jurassic). Bar  2 cm. (Courtesy BSPG.)

regions; at the nodes the vascular tissue is a continuous ring. Several features, including the presence of superficial stomata, suggest affinities with Equisetum subg. Equisetum. Specimens described as fertile shoots consist of distal leafbearing nodes with a single whorl of sporangia borne at a

Figure 10.104

Neocalamites merianii, portion of a stem (Triassic). Bar  5 cm. (Courtesy K.-P. Kelber.)

basal node (Osborn et al., 2000). Spores are small (⬃10 µm), with a rugulate surface ornamentation. Nodal septations span the central pith and cortex, and thin fimbrils subdivide the internodal areas into smaller chambers. The vascular system consists of 31–33 continuous bundles that do not alternate in position between successive nodes and internodes.

CHAPTER 10

SPHENOPHYTES

379

Figure 10.105

Cross section of a bud of Spaciinodum collinsonii showing six branches extending above the apex (arrow) (Triassic). Bar  4 mm. (Courtesy P. Ryberg.)

SPHENOPHYTE EVOLUTION The origin of the sphenophytes continues to remain problematic. Most workers root the group with the trimerophytes, and several Devonian plants have been described that suggest the enigmatic fernlike Iridopteridales (Chapter 11) might represent an ancestral complex from which the sphenophytes evolved. Fossil taxa that have figured more or less prominently in a number of discussions concerning sphenophyte ancestry include the Middle Devonian genus Ibyka (Skog and Banks, 1973; Berry, 2005), as well as Anapaulia moodyi (Berry and Edwards, 1996b) and Compsocradus laevigatus (Berry and Stein, 2000) from the Middle to Upper Devonian of Venezuela. Ibyka amphikoma (Skog and Banks, 1973) is known from both compressed and permineralized specimens and has been reconstructed as a monopodial branching system bearing lateral (ultimate) appendages on several orders of branches (FIG. 10.106). Both ultimate appendages and branches are helically arranged and covered by delicate hairs; ultimate appendages, which Skog and Banks call leaves, are three dimensional and have curved tips. The vascular system

Figure 10.106 Suggested reconstruction of Ibyka amphikoma (Devonian). (From Kenrick and Crane, 1997a.)

consists of a three-lobed protostele with some of the arms bifurcated at higher levels to form terete leaf traces. Protoxylem tracheids and parenchyma occur near the tips of the xylem lobes in the form of peripheral loops. The branches that bear sporangia are considered homologous with the sterile appendages but have only been found on third-order axes. Sporangia are ovoid and 1 mm long. Features that suggest a relationship between the sphenophytes and Ibyka include the general morphology of the lateral appendages and fertile branches. Skog and Banks (1973) speculated on an evolutionary series leading from the trimerophytes via Ibyka and plants like Calamophyton and Hyenia to the Late Devonian and Mississippian sphenophytes. Stein et al. (1984), however, suggested that Ibyka may be in an intermediate position between the trimerophytes and Pseudosporochnus and related taxa in one lineage, and, in a second lineage, the Iridopteridales. Anapaulia moodyi (Berry and Edwards, 1996b) is a more or less monopodially organized plant known from compressions (FIG. 10.107). It is composed of four orders of axes,

380

PaleoBOtany: the biology and evolution of fossil plants

Figure 10.108 Edouard Boureau.

Figure 10.107 Suggested reconstruction of a branch system of Anapaulia moodyi. (From Berry and Edwards, 1996b.)

in which second-order branches and smaller, dichotomizing ultimate branching systems extend from first-order axes (up to 15 mm in diameter) in whorls or compressed helices separated by internodes (FIG. 10.107). The third-order branches are arranged in compressed helices and fourth-order branches occasionally extend from third-order branches. Ultimate branching systems, borne on third- and fourth-order branches, dichotomize several times. Sporangia occur in pairs on the recurved tips of dichotomizing axes. Anapaulia moodyi is an interesting fossil with regard to the ancestry of sphenophytes, as it displays a strong tendency toward a whorled arrangement of lateral branches and includes true whorls on some stems. The occurrence of sporangia on recurved fertile branching systems is less indicative but generally not inconsistent with development of the sphenopsid sporangiophore over time (Berry and Edwards, 1996b). Permineralized first-order branches of Compsocradus laevigatus (Berry and Stein, 2000) contain a mesarch actinostele with six primary xylem arms; a protoxylem strand occurs around the tip of each arm. First-order branches bear whorls of higher order branches and dichotomous ultimate appendages. A permineralized specimen partly confirms the presence of whorls and indicates that vascular traces are

derived from every other primary xylem rib in each whorl, with intervening arms producing traces in whorls above and below. The second-order branches give off ultimate appendages in a non-whorled, three-dimensional, or alternate arrangement. Fertile ultimate appendages bear paired sporangia distally. Similar to Anapaulia, C. laevigatus shows a whorled arrangement of lateral branches and thus may be interesting with regard to sphenophyte ancestry and origins. Archaeocalamites (Devonian) has been used as a starting point for the subsequent evolution of the Calamitaceae (Boureau, 1964) (FIG. 10.108). For this reason, and because of the stratigraphic position of the group, the Pothocites cone type is used as the starting point in discussions about calamitean cone phylogeny. Some workers believe that the near-axillary position of the sporangiophore trace in Palaeostachya represents the primitive condition, with the phylogenetic migration of the sporangiophore into an internodal position (Calamostachys) representing an intermediate evolutionary level. According to this series, cones such as Mazostachys and Cingularia, in which the sporangiophores are either immediately beneath the overlying bract or fused to it, constitute the most highly advanced forms. Another hypothesis views the Calamostachys cone organization as basic, with forms in which the sporangiophores are in an axillary or abaxial position being derived. Proponents of this transformational series view the Pothocites cone type as primitive, with forms in which the sterile bracts are intercalated between the whorls of sporangiophores being more highly evolved. Still another view suggests that vegetative whorls have developed phylogenetically as a result of the

CHAPTER 10

failure of some sporangiophore whorls to develop sporangia. The younger geologic age of many of the cone types in which sporangiophores are associated with the abaxial surface of the bracts supports this hypothesis. It has been suggested that this trend may have developed as a result of selective pressures for greater sporangial protection. Whether or not the late Paleozoic non-calamitacean sphenophytes and Mesozoic Equisetaceae evolved from the Calamitaceae, or all groups diverged from a common ancestor during the Carboniferous, is still not known. The question arises as to how the nonbracteate equisetacean cone developed from ancestors with bracteate cones. Based on a comparative analysis of the Late Triassic cone Equicalastrobus chinleana with other late Paleozoic and Mesozoic equisetalean cones, Grauvogel-Stamm and Ash (1999) suggested that the nonbracteate cone may have evolved from a bracteate calamite cone through reduction and fusion of the bracts and peltate sporangiophores. One step in this trend would be E. chinleana, in which a vascularized leaflike appendage (interpreted as a vestige of the bract), which they call an umbo tip, extends from each of the sporangiophore heads. Subsequent reduction of the appendage would ultimately lead to the condition seen in modern Equisetum. Another important question regarding the evolution of the modern equisetacean reproductive structures concerns the origin of the peltate sporangiophore. Traditionally, these structures were believed to have evolved only once in the evolutionary history of the Equisetales, that is, directly from an archaeocalamitean ancestor with primitive cruciate sporangiophores (Good, 1975). Cúneo and Escapa (2006), however, speculate that a second origin of the peltate sporangiophore may have taken place later in time. Based on a hypothesis originally advanced by Naugolnykh (2004a, b) (FIG. 10.86), these authors hypothesized a series of morphological changes and suggested that Cruciaetheca, which has contiguous fertile internodes bearing cruciate sporangiophores, may represent the starting point in a trend in equisetalean reproductive systems. From Cruciaetheca, nonstrobilar forms with contiguous fertile internodes bearing peltate sporangiophores, such as Equisetinostachys, may have developed (FIG. 10.86). Later, the number of sporangiophores became reduced to one, as in Paracalamitina. Further reduction would result in the compact, unbranched strobilus of the Mesozoic and Cenozoic Equisetaceae, such as Neocalamites, Equisetites, and Equisetum. Support for this hypothesis comes from teratological forms seen in extant Equisetum, in which there is a reappearance of the Permian and Mesozoic ancestors (Naugolnykh, 2004a). For example, Tschudy (1939) described aberrant plants of E. telmateia in

SPHENOPHYTES

381

which the primary shoots produced fertile internodes that parallel the condition seen in Equisetinostachys. Most of the Equisetales were at least morphologically homosporous. Although a few heterosporous forms are known, the variability in cone organization and sporangial contents within this group does not approach the level of heterospory present in the Lycophyta. The most highly evolved heterosporous system within the Equisetales is Calamocarpon. It might be argued that the Calamocarpon cone type is more highly evolved than the lycophyte forms, because current evidence suggests that the megaspore was not shed from the sporangium (Baxter, 1963). In this regard, the calamite reproductive strategy more closely approximates the seed habit than that of lycopsids such as Lepidocarpon, Miadesmia, and Achlamydocarpon. These are assumed to shed their single functional megaspore at maturity, based on the occurrence of isolated megaspores in coal. The presence of elaters on almost all equisetalean spores represents a modification of the sporoderm that may indicate a more highly evolved system than the simple sporoderm in lycopsid microspores. It would appear that much of the emphasis in the evolution of the lycopsid reproductive system was directed toward the megaspore-producing cone and features of the megaspore and megagametophyte. In the Equisetales, however, evolutionary modifications may have been occurring in both the microgametophyte and the megagametophyte phases. One additional comment needs to be noted regarding the spores of the Equisetales. As noted earlier, in many species the size range of the spores is highly variable, sometimes encompassing a range of several hundred micrometers. In Paracalamostachys spadiciformis, for example, the known range of the Calamospora spores is 55 to 350 µm in diameter. In this species, the larger spores have been designated megaspores, the cone is considered bisporangiate, and the plant heterosporous. The possibility exists that some of the presumed homosporous calamitean cones were in fact bisporangiate, with megaspores and microspores lumped together in the descriptions. Because of the difficulty in identifying immature elaters on spores, there may be no way to distinguish megaspores and microspores from developmental stages of isospores. Another problem concerns the role that each spore type may have played in the reproductive process. Microspore and megaspore are, in reality, functional terms that ultimately reflect whether or not a microgametophyte or a megagametophyte finally develops. Generally, larger spores develop into a megagametophyte, although in some living plants, the megaspores are the smaller of the two types early in spore development. The Equisetales may represent such a heterosporous system in which the

382

PaleoBOtany: the biology and evolution of fossil plants

ultimate function of the spore cannot be determined by size. This incipient form of heterospory is known in a number of living plants, including Equisetum, in which gametophytes produced may contain one or both types of sex organs. Although not as large as the lycopsids, some members of the Equisetales were arborescent plants that attained their tree-sized habit through a different method of producing supporting tissue. In the arborescent forms, the diameter of the stem was the result of a central pith and a large amount of secondary xylem produced by a cambium that may have been unifacial. In the lycopsids, stem diameter was the result of extensive development of secondary cortical tissues in the form of periderm. These two groups of tree-sized plants are an excellent example of convergent evolution in which the arborescent habit and many of the features associated

with the reproductive process have been achieved in different ways. Finally, the fact that Equisetum is the only surviving member of a formerly diverse and widespread group of plants parallels the evolutionary history of the Lycophyta, in which the arborescent members have also become extinct. We have already discussed ideas about the demise of some arborescent lycopsids as the environment became drier, in relation to the adaptation of their megasporophyllar units to wetland habitats (Chapter 9). Whether a similar situation existed for the calamites is not known and may be far more difficult to discern, as the majority of the arborescent members appear to have been homosporous. The fact remains, however, that although arborescent members in both groups flourished during the Carboniferous, only the herbaceous members persist to the present.

11 Ferns and early fernlike plants Evolution of the megaphyll..................................... 386

Botryopteridaceae .............................................................................443

Cladoxylopsida..................................................................... 387

Anachoropteridaceae.........................................................................449 Kaplanopteridaceae ...........................................................................451

Pseudosporochnales ..........................................................................388

Psalixochlaenaceae ...........................................................................452

Iridopteridales ...................................................................................398

Sermayaceae......................................................................................453

Phylogenetic Position of the Cladoxylopsids ....................................400

Tedeleaceae .......................................................................................454

Early fernlike plants ........................................................401

Skaaripteridaceae ..............................................................................457

Rhacophytales ...................................................................................401

Tempskyaceae ...................................................................................457

Systematics of the Rhacophytales .....................................................404

Schizaeaceae .....................................................................................459

Coenopterid Ferns .............................................................................405

Hymenophyllaceae ............................................................................462

Stauropteridales.................................................................................405

Gleicheniaceae ..................................................................................462

Zygopteridales...................................................................................408

Dicksoniaceae ...................................................................................464

Marattiales ................................................................................418

Cyatheaceae ......................................................................................465

Psaroniaceae: Vegetative Features .....................................................418

Matoniaceae ......................................................................................466

Psaroniaceae: Reproductive Features ................................................425

Loxsomataceae ..................................................................................469

Paleozoic Compression Taxa ............................................................431

Dipteridaceae ....................................................................................469

Mesozoic Marattialeans ....................................................................433

Polypodiales ......................................................................................470

Marattialean Evolution......................................................................434

Salviniales ................................................................................. 472

Ophioglossales.......................................................................435

Marsileaceae .....................................................................................472 Salviniaceae ......................................................................................473

Leptosporangiate ferns ..................................................436 Conclusions ........................................................................... 476

Osmundales .......................................................................................436

I was a seed again. I was a fern in the swamp. I was coal. Denise Levertov, A Tree Telling of Orpheus

A fern is sometimes described as a vascular cryptogam with foliar-borne sporangia (FIG. 11.1). That definition generally works for extant ferns, but defining a fossil fern is more difficult. Ferns first appear in the Devonian and today more than 10,000 species can be found in a wide variety of habitats. All

extant ferns are perennial and herbaceous and, although some are treelike, none produce secondary xylem. There is a wide variety of stem anatomy in the ferns, ranging from simple protosteles in most of the Paleozoic forms to complex dictyosteles. The leaves of ferns (Pteridophyta) are megaphylls

383

384

Paleobotany: the biology and evolution of fossil plants

(discussed below) and are called fronds; primary xylem maturation is endarch or mesarch. The immature fronds, also called croziers or fiddleheads, unroll in most members of this group, a developmental process called circinate vernation. Most fronds are compound leaves; the primary or first order of leaflets are called pinnae (sing. pinna) and the final order, pinnules. A number of fossil ferns may have basal pinnae on the frond that have a different morphology than the other pinnae; these are termed aphlebiae (FIG. 11.2) (sing. aphlebia). Sporangia are borne marginally or abaxially on the fronds and spores germinate to form free-living, photosynthetic gametophytes (prothallia) (FIG. 11.1). Two basic types of sporangial development and structure are found within the ferns. In the eusporangiate type (FIG. 11.3),

sporangia are relatively large and the sporangial wall develops from periclinal divisions of a group of superficial initials, whereas the sporogenous tissue (the cells that give rise to the spores by meiosis) develops from cells that are internal to the wall initials. The wall of the eusporangium is several cell layers thick, the sporangium is sessile, and the number of spores produced in each sporangium is large. Members of the Ophioglossales and Marattiales are the only modern ferns that possess sporangia of this type. In contrast, the leptosporangium (FIG. 11.3) is smaller and develops from a slightly oblique division of a single superficial initial cell. Subsequent cell divisions produce a stalked sporangium that is usually one cell layer thick. Sporangial dehiscence is either transverse or longitudinal, and the number of spores per sporangium is typically small compared to

Sporophyte Frond

Young sporophyte

Sorus Pinna “Fiddlehead” Rhizome Indusium

Adventitious roots

Zygote

2n

Sporangium containing spore mother cells which undergo meiosis, producing

Fertilization

Meiosis n Archegonium Egg

Sperm

Spores Antheridium

Germinating spore

Rhizoid

Gametophyte (prothallus)

Figure 11.1 Life history of a fern. (From Taylor and Taylor, 1993.)

CHapter 11

Ferns and early fernlike plants

eusporangiate types. Leptosporangia occur in filicalean ferns. The terms eusporangium and leptosporangium do not indicate levels of phylogenetic specialization, nor are they especially valuable in systematic studies. Several recent studies have suggested that the terms be abandoned because an examination of sporangial development in modern plants shows a high degree of variability in all sporangium-related features. For example, within the Ophioglossales, one to several cells may be involved in sporangial initiation. Because the eusporangium is produced by plants that are regarded as primitive on the basis of other features, this type has long been considered the more plesiomorphic form. Bierhorst (1971) (FIG. 11.4), however, suggested that a type similar to the leptosporangium

Figure 11.4 Figure

David W. Bierhorst. (Courtesy M. A. Millay.)

11.2 Aphlebia erdmannii (Permian). Bar  1 cm.

(Courtesy M. Barthel.) Eusporangiate development Spermatogenous cells

Leptosporangiate development

Wall layers

Spermatogenous cells Wall layer

Figure 11.3

Comparison of eusporangium and leptosporangium development.

385

386

Paleobotany: the biology and evolution of fossil plants

was primitive among most vascular plants. The author suggested that the ancestral filicalean sporangium was elongate, exhibited longitudinal dehiscence, and lacked an annulus (specialized cells involved in the opening of the sporangium). Although the fossil record generally provides little information about sporangial development, the two terms can be used for fossil material in a purely descriptive sense. As was the case for the Devonian sphenophytes and lycopsids, the features that are used to characterize most living ferns are less applicable for many fossils that represent various degrees of evolution within the ferns, especially those from the Devonian. In some treatments, these Devonian plants are referred to as fernlike or as preferns, since many lack planated fronds and the stelar systems do not exhibit typical trace and gap production. In spite of differing interpretation, there is general agreement that these Devonian plants probably gave rise to the more representative and geologically younger ferns. In effect, many of the fossil representatives possess characters that are no longer seen in modern ferns and, by definition, have not yet acquired all the features that characterize the later-appearing (so-called true) ferns. The classification that follows does not include examples of all fossil ferns, but is rather intended to provide a framework with which to discuss some of the plants that are believed to have a phylogenetic relationship with modern ferns. Detailed accounts of ancient- and modern-appearing fossil ferns that focus on biodiversity and ecology during the Mesozoic and Cenozoic (Collinson, 2001, 2002; Skog, 2001), and evolution of leptosporangiate ferns (Stockey and Rothwell, 2006, and papers therein) are important references that help to frame discussions about the evolution of these highly adaptive vascular plants and their role in ecosystems through time (Rothwell, 1996b).

Evolution of the megaphyll As noted in Chapters 9 and 10, all vascular plants except the lycophytes have leaves called megaphylls, and megaphyllous leaves have a different evolutionary origin than microphylls. Some fossil members of the Sphenophyta have conspicuous megaphylls, although modern Equisetum only has very reduced leaves. We introduce the evolution of the megaphyll here, however, because a number of fossil ferns exhibit various stages in the evolution of this leaf type. Megaphylls have complex venation (unless secondarily reduced) and a leaf gap occurs in the stele when a leaf trace is produced, unless a protostele is present. The megaphyllous habit is hypothesized to have arisen through the planation (flattening into two dimensions) and subsequent webbing of a branch system

Higher taxa in this chapter:

Pteridophyta Cladoxylopsida Pseudosporochnales (Devonian–Mississippian) Iridopteridales (Middle–Upper Devonian) Early Fernlike Plants Rhacophytales (Middle–Upper Devonian) [Coenopterids (Devonian–Permian)] Stauropteridales (Devonian–Pennsylvanian), Stauropteridaceae Zygopteridales (Lower Mississippian–Permian), Zygopteridaceae True Ferns—eusporangiates Marattiales (Carboniferous–recent) Psaroniaceae, Marattiaceae Ophioglossales (Jurassic–recent), Ophioglossaceae True Ferns—leptosporangiates Osmundales Osmundaceae (Permian–recent) Guaireaceae (Permian) Filicales (Carboniferous–recent) Botryopteridaceae (Carboniferous–Permian) Anachoropteridaceae (Pennsylvanian–Permian) Kaplanopteridaceae (Pennsylvanian) Psalixochlaenaceae (Carboniferous) Sermayaceae (Pennsylvanian) Tedeleaceae (Mississippian–Permian) Skaaripteridaceae (Permian) Tempskyaceae (Cretaceous) Schizaeaceae (Jurassic–recent) Hymenophyllaceae (Triassic–recent) Gleicheniaceae (Permian–recent) Dicksoniaceae (Triassic–recent) Cyatheaceae (Jurassic–recent) Matoniaceae (Triassic–recent) Loxsomataceae (Cretaceous–recent) Dipteridaceae (Triassic–recent) Polypodiales (Jurassic–recent) Dennstaediaceae, Pteridaceae, Onocleaceae, Blechnaceae, Polypodiaceae Salviniales (Cretaceous–recent) Marsileaceae Salviniaceae (FIG. 11.5), that is the veins of the leaf had their origin as branches. In recent years there has also been interest in leaf evolution of extant plants using a broad array of genetic, physiological, and molecular tools (Beerling and Fleming,

CHapter 11

(A)

(B)

Ferns and early fernlike plants

(C)

387

(D)

Figure 11.5 Hypothesized stages in the evolution of the megaphyll. A. Three-dimensional branching axis. B. Axial system flattened into single plane (planation). C. Webbing or filling in with laminar tissue. D. Megaphyll with complex venation pattern and expanded lamina. (From Taylor and Taylor, 1993.)

2007). In general these studies support the transformational hypothesis suggested by the Telome Theory (Zimmermann, 1930) (FIG. 11.5). Despite the fact that there is no single transformational series to illustrate the evolution of the megaphyll, the megaphyllous habit is well represented in several groups by the Late Devonian and Early Carboniferous. Some of the Devonian cladoxylopsid ferns show the beginnings of webbing and the Carboniferous Stauropteridales have many fernlike features, yet their branches are produced in four ranks, rather than being planated (laterals in two ranks). The morphology of the Devonian plant Rhacophyton suggests that the flattening of the ultimate segments in the frond took place prior to any major change in pinna arrangement and zygopteridalean ferns have petiolar symmetry that is often more stemlike than leaflike. Although ferns and seed ferns (Chapter 14) both exhibit similar leaf morphologies in the Paleozoic, Galtier (1981) has shown that these appear to have arisen separately in these two groups. One idea that has been suggested to account for the evolution of the megaphyllous habit relates to high concentrations of available CO2 prior to the Late Devonian (Beerling et al., 2001). Megaphyllous leaves became more widespread as CO2 levels and paleotemperatures dropped during the Middle and Late Devonian, and these authors hypothesized that the decrease in global temperature allowed planated leaves to develop without becoming overheated. Their development was coupled with a higher stomatal density, which is not seen in earlier plant forms. A more complex anatomy, that is more vasculature, would have been necessary to support a larger leaf both structurally and physiologically.

Cladoxylopsida The Cladoxylopsida at present includes two orders: the Pseudosporochnales and the Iridopteridales (Berry and Stein, 2000). These plants have a varied taxonomic history, due in part to the large number of homoplasies that they exhibit and their position at the beginning of the evolution of the ferns. Earlier authors included them in the order Cladoxylales, with the seed ferns, coenopterid ferns (Walton, 1953a), intermediate between the sphenophytes and ferns, somewhere between the coenopterid ferns and the Marattiales (Andrews, 1961), and within the Filicales (Gothan and Weyland, 1964). Most recently, Meyer-Berthaud et al. (2007) characterize the cladoxylopsids by a sequence of branching patterns that display a hierarchical architecture, with the production of a single type of lateral at each node; in other reports of this group, the branching is less well organized (Berry and Wang, 2006a). Traces to the laterals are supplied from several stelar ribs. Branches attached to stems indicate that some of these plants were large (Berry and Hilton, 2006). Anatomically the cladoxylopsids also have nests of sclereids in the cortex (Berry and Stein, 2000). The earliest representatives occur in the Middle Devonian, but several species extend into the Mississippian (Tournaisian) (Berry and Hilton, 2006). The group is thought to have evolved from the Trimerophytophyta, but probably was not directly involved in the origin of any later-appearing groups of plants. Cladoxylopsids are paraphyletic and characterized anatomically by a dissected vascular system consisting of a variable number of primary xylem segments (plates) and radially

388

Paleobotany: the biology and evolution of fossil plants

elongate primary xylem ribs (Berry and Stein, 2000). The position of protoxylem strands can be variable, but most are positioned at the end of the xylem plates. The term polystele was initially used to describe the vascular organization in some of these plants, but in light of current ideas on the nature of the vasculature in fossil and extant plants (Basinger et al., 1974; see Chapter 14), the term polystele for these plants is inappropriate. One of the characteristic features of the vascular tissue in this group is the presence of a cluster of small protoxylem elements at the periphery of the xylem segments. Structures that appear similar are present in several other groups of early fernlike plants, where they consist of thin-walled parenchyma cells surrounded by protoxylem elements. These structures have historically been termed peripheral loops (Leclercq, 1970). As noted later in this chapter, those in the Cladoxylopsida are better termed permanent protoxylem strands. In some species radially aligned metaxylem tracheids are uniformly distributed around the primary xylem or more extensively developed toward the center of the stem. A few specimens have been reported with secondary xylem. In many of the species, the number and position of the xylem segments and nature of the extraxylary tissues have been used in systematics. Since the vascular system is now known to be highly variable from level to level in a stem, subsequent studies of additional specimens will probably reduce the number of valid taxa. Pseudosporochnales

Many of the plants included in the Pseudosporochnales are now thought to have been arborescent. In a few specimens the ultimate stem segments consist of three-dimensional appendages; flattened leaflike forms have been reported from other taxa. Sporangia are terminal and often in pairs; however, in some they are aggregated and produced in clusters that are arranged in pseudowhorls. All species are thought to have been homosporous. The genus Cladoxylon was established in 1856 by Unger for anatomically preserved axes. After the discovery of C. scoparium (Kräusel and Weyland, 1926), the group was elevated to the ordinal level. Despite acceptance of these plants as fernlike, their taxonomic position with reference to other groups continues to be problematic. Today Cladoxylon is used by most authors as a morphotaxon for petrified axes that exhibit a particular type of anatomy involving at least partial primary xylem dissection into segments (Cordi and Stein, 2005). Several have cautioned that the dissection of the vascular system in these plants is not a reliable character, and thus the phylogenic position of these plants must await

Figure 11.6 Cross section of Cladoxylon radiatum (Devonian).

Bar  2 mm. (Courtesy B. Meyer-Berthaud.)

additional detailed studies (Cordi and Stein, 2005; Soria and Meyer-Berthaud, 2005). Cladoxylon radiatum is known from the Mississippian of Germany. Stems are up to 6.5 cm in diameter and contain 17 stelar segments that radiate out from the center; branches are helically arranged (FIG. 11.6). All xylem is primary and tracheids exhibit scalariform pitting. Thick-walled cells are present in the ground tissue of the stems. When laterals were produced, four to six arms of xylem extend into the petiole and subsequently divide to form a tangentially flattened ring of up to eight xylem strands. Smaller aphlebia traces occur at right angles to the petiole. The arrangement of the xylem strands differs in the smaller (3.5 cm in diameter) C. taeniatum (Solms-Laubach, 1896) (FIG. 11.7). In this species (Mississippian) there are two stelar zones in the stem. Around the periphery of the stem are numerous ( 20) radially elongate xylem strands that surround three to four centrally located, cylindrical xylem strands, each of which produces bipolar traces (traces with two protoxylem poles). In C. taeniatum (FIG. 11.7), the xylem segments produce tripolar traces (Meyer-Berthaud et al., 2007). Several specimens exhibiting cladoxylopsid anatomy have been described from the Mississippian (Lower Carboniferous) Calciferous Sandstone Series of Scotland (Long, 1968). Cladoxylon waltonii includes stems and petioles of two distinct size classes. One group includes radially symmetrical stems 1.2 cm in diameter with helically arranged petioles containing a clepsydroid-shaped (dumbbell-shaped) xylem strand. When petioles are produced, a band of xylem is formed by the fusion of two adjacent xylem strands. The second group includes smaller stems (2–7 mm

CHapter 11

Ferns and early fernlike plants

389

Figure 11.8 Cross section of Polyxylon australe axis (Devonian). Bar  2 mm. (Courtesy B. Meyer-Berthaud.)

Figure 11.7 Cross section of Cladoxylon taeniatum stem (below) with branch (Mississippian). Bar  6 mm. (Courtesy J. Galtier.)

in diameter) with petioles borne alternately. The larger stems contain 9–15 arms of primary xylem, whereas the smaller ones have 4–9 strands arranged in a U-shaped configuration. In both groups most of the tissue in each arm is composed of primary xylem; radially aligned metaxylem tracheids near the periphery of several arms were once interpreted as secondary in origin. Near the periphery of each segment are permanent protoxylem strands. Poorly preserved cells thought to represent phloem are present in the embayments between the xylem arms and, to a lesser extent, around the xylem. The ground tissue consists of a broad zone of parenchyma with occasional lacunae. In recent years there has been considerable attention directed at the anatomy and phylogenetic position of the Pseudosporochnales. What were once interpreted as systematic characters (e.g., number of vascular segments) useful in distinguishing taxa, are now understood to be highly variable based on the position within the stem and nature of the

vascular system. Nevertheless, there are a number of taxa that have been defined along more phylogenetic lines. One of these is Polyxylon australe (FIG. 11.8) from the Upper Devonian of Australia (Meyer-Berthaud et al., 2007). Axes are small (2.7 cm in diameter) and in cross section have 19 radially elongate, mesarch primary xylem arms. Each of the arms dichotomizes at the tip to form a minimum of six traces that extend into the whorled branches. The genus Hierogramma was established for small axes characterized by a distinctly bilateral vascular system of two elongate xylem bands, one of which may be T-shaped. At higher levels, the vascular system assumes a more U-shaped configuration. In H. mysticum, leaf traces are produced alternately as small terete strands. Bertrand (1935) suggested that H. mysticum was the petiole of Cladoxylon taeniatum. Syncardia, Voelkelia, and Arctopodium are additional petrifaction genera that anatomically resemble pseudosporochnaleans. Pseudosporochnus nodosus is a well-known member of the order and has been reconstructed as a small tree with a bulbous base (Leclercq and Banks, 1962). The genus provides an excellent example of how whole-plant concepts can change as additional specimens are discovered and earlier ones reinterpreted. These Middle Devonian fossils were initially described as algae, but later shown to contain vascular elements. At one time, Leclercq and Banks (1962) suggested that the lateral branches (FIG. 11.9) were fronds that terminated in ultimate, flattened pinnae. The discovery of additional specimens suggested that the pattern of traces to ultimate segments were bilateral. From these detailed anatomical (Leclercq and Lele, 1968) studies it became obvious that in

390

Paleobotany: the biology and evolution of fossil plants

Figure 11.10 Suggested reconstruction of Pseudosporochnus nodosus. (From Berry and Fairon-Demaret, 2002.)

Figure 11.9 Pseudosporochnus nodosus first-order branch divided into four second-order branches (Devonian). Bar  4 cm. (In Leclercq and Banks, 1962; courtesy M. Fairon-Demaret and P. Gerrienne.)

descriptions of these plants, more neutral terms, such as lateral branch and ultimate branching unit, were more appropriate until homologies among the units and taxa can be documented (Berry and Fairon-Demaret, 1997). These anatomical and morphological studies have changed ideas about the phylogenetic position of P. nodosus, which in turn has resulted in a new reconstruction of the plant (FIG. 11.10). Today the plant is interpreted as 2–4 m tall. Based on the

discovery of branch scars on specimens from Belgium, it is hypothesized that branches probably abscised during growth (Berry and Fairon-Demaret, 2002). Second-order branches dichotomized with the lateral branching system giving rise to ultimate branching units that are three dimensional (FIG. 11.11). There is little distinction between ultimate sterile and fertile branching units (FIG. 11.12), except that the latter produced pairs of sessile, ellipsoidal sporangia. Sporangia are up to 3 mm long, but nothing is known about the spores. One interesting feature of Pseudosporochnus is the presence of circular–elongate structures distributed on all parts of the plant. These represent nests of cortical fibers, and their presence on compression specimens has been used to determine

CHapter 11

Ferns and early fernlike plants

391

Figure 11.11 Detail of vegetative branch of Pseudosporochnus nodosus (Devonian). Bar  3 cm. (In Leclercq and Banks, 1962; courtesy M. Fairon-Demaret and P. Gerrienne.) Figure 11.12 Pseudosporochnus nodosus third-order branch

the biological affinities of detached parts (Berry and FaironDemaret, 1997). Pseudosporochnus hueberi was originally described as a species of Cladoxylon from the Middle and Upper Devonian (Givetian) of eastern New York (Matten, 1974), but subsequently transferred to Pseudosporochnus (Stein and Hueber, 1989). The partial compression–petrifaction specimens are up to 58 cm long and several centimeters in diameter. The number of xylem segments is highly variable (40–50) in the main axis and branches (20–25). Segments exhibit a variety of configurations in cross section (FIG. 11.13). Only primary xylem is present and metaxylem tracheids possess uni- to multiseriate bordered pits (Stein and Hueber, 1989). Production of traces to laterals involves the fusion of portions of three contiguous xylem strands in the stem that form bilateral trace groups. Ultimate branch segments are three dimensional; some trace patterns suggest that the ultimate appendages may have been planated.

with vegetative and fertile branches helically arranged (Devonian). Bar  3 cm. (In Leclercq and Banks, 1962; courtesy M. FaironDemaret and P. Gerrienne.)

Figure 11.13 Cross section of Pseudosporochnus hueberi

(Devonian). Bar  2 mm. (Courtesy W. E. Stein.)

392

Paleobotany: the biology and evolution of fossil plants

Calamophyton

Lorophyton

Pseudosporochnus

Wattieza

Figure 11.14 Selected examples of increasing complexity in vegetative (upper) and fertile frond morphology several Devonian plants.

(From Berry, 2000.)

Wattieza givetiana (FIG. 11.14) is a Middle Devonian plant from Belgium that was initially described by Stockmans (FIG. 11.15) as a plant with a habit similar to that of Pseudosporochnus. More recently discovered specimens from Venezuela consist of ultimate branches that were three dimensional (Berry, 2000); elongate sporangia lacking a stalk were produced on recurved distal tips. Lorophyton is an early Givetian genus characterized by bifurcating roots and aerial lateral branches that were sometimes dichotomous (FaironDemaret and Li, 1993). Ultimate branches also dichotomize. Fertile units consist of recurved ultimate segments, each bearing two sporangia. Sporangial dehiscence is lateral. In 1924, Goldring (FIG. 11.26) described some Middle Devonian trunk casts from Gilboa, New York, as Eospermatopteris (FIG. 11.16). This plant was believed to be 9–12 m tall, bearing massive fernlike fronds with seeds attached to the ultimate branch tips. The base of each trunk was cormlike (FIG. 11.16), with numerous radiating roots. For some time, Eospermatopteris was regarded as the oldest seed fern, although an examination of the presumed seeds showed them to contain spores (Kräusel and Weyland, 1935b), thus removing the taxon from association with seed plants. Recently Stein et al. (2007) demonstrated that the trunk of

Figure 11.15

François Stockmans.

Eospermatopteris produced a crown of short, erect branches that divided digitately with three-dimensional, terminal appendages (FIG. 11.17); there were no leaves. The discovery of Eospermatopteris stems attached to Wattieza branches in a

CHapter 11

11.16 Base of Eospermatopteris Bar  0.25 m. (From Stein et al., 2007.)

Figure

Ferns and early fernlike plants

393

(Devonian).

crown (FIGS. 11.18, 11.19) suggests that the plant was 8 m tall. What is especially interesting is that the branches were apparently regularly shed and replaced by new branches at the top of the tree. Thus, in the Middle Devonian there were already treelike forms that grew to some height, but they possessed different combinations of morphological and structural features (Meyer-Berthaud and Decombeix, 2007) than those present in the arborescent progymnosperms from the Late Devonian (Chapter 12). A genus that appears to be anatomically similar to Pseudosporochnus is Pietzschia (Gothan, 1927a). Stems are 2.5 cm in diameter and contain a cylinder of radially aligned xylem plates (FIG. 11.20). The number of xylem segments is large in P. polyupsilon (54), a stem from the New Albany Shale of Kentucky (Read and Campbell, 1939). At some levels, sclerenchyma plates occur between adjacent xylem segments. Four bundles appear to be involved in the development of each trace to the lateral appendages. At higher levels, these concentric strands form an inverted, U-shaped strand through lateral fusion. Permanent protoxylem strands are present in some Pietzschia specimens and absent in others, but, when present, they occur at either end of each xylem arm. The main axis of P. polyupsilon had determinate growth and a whorled arrangement of the laterals (Soria and Meyer-Berthaud, 2003). The presence of aerenchymatous

Figure 11.17 Fossil specimen (left) and reconstruction of

Eospermatopteris (Devonian). Bar  1.0 m. (From Stein et al., 2007.)

ground tissue suggests that Pietzschia levis, a specimen from the lower Famennian (Upper Devonian) of Morocco, lived in a wetland habitat (Soria et al., 2001). In another species, P. schulleri (FIG. 11.21) (Famennian of Morocco), all of the tissues are primary (Soria and Meyer-Berthaud, 2005). The stems often exceed 15 cm in diameter and include both

394

Paleobotany: the biology and evolution of fossil plants

Figure 11.20 Cross section of Pietzschia schulleri stem

(Devonian). Bar  2 mm. (Courtesy B. Meyer-Berthaud.)

Figure 11.18 Wattieza crown of ultimate appendages in place

(Devonian). (Courtesy W. E. Stein.)

Figure 11.21 Cross section of Pietzschia schulleri roots (Devonian). Bar  2 mm. (Courtesy B. Meyer-Berthaud.)

Figure 11.19 Wattieza crown with attached branches and trunk

(Devonian). Bar  10 cm. (Courtesy W. E. Stein.)

central and peripheral xylem strands (FIG. 11.20). Permanent protoxylem strands are not present. A liana habit has been proposed for the Middle–Late Devonian genus Xenocladia. Specimens of X. medullosina measure 10 cm in diameter and are reported to contain secondary xylem (Arnold, 1952a) (FIG. 11.22). Abundant secondary xylem and reduced primary xylem are features of another cladoxylopsid stem genus, Steloxylon. The type specimen of Steloxylon ludwigii is either Permian or Devonian from either Siberia or Kazakhstan. In S. ludwigii, the vascular segments anastomose to form a complex network of vascular tissue with

CHapter 11

Ferns and early fernlike plants

395

Figure 11.23 Diagrammatic reconstruction of a stem segment of Rhymokalon trichium showing branch at right (Devonian). (From Scheckler, 1975a.)

Figure 11.22 Chester Arnold.

numerous small petioles arranged helically around the stem. Trace emission includes the fusion of two or more strands, with the resulting petiole strands composed almost entirely of secondary xylem. Tracheid pitting is multiseriate bordered, with rays widest toward the center of the stem. In S. irvingense (New Albany Shale), secondary xylem development is more pronounced on one side of the xylem arm and the vascular strands are more circular in outline. Irregular strands of thickwalled fibers in the parenchymatous ground tissue further distinguish this species (Read and Campbell, 1939). Another plant with cladoxyl anatomy is Rhymokalon trichium (FIG. 11.23), which is known from the Upper Devonian (Frasnian) of New York (Scheckler, 1975a); the taxon is based on pyritized compressions. The largest axis is 3 cm in diameter and exhibits second- and third-order branches that are helically arranged. Longitudinal striations on the largest axes correspond to the positions of vascular strands within the axis. Trichomes up to 1 mm long are present on the surface of the more distal branches. Cross sections of Rhymokalon show a multiribbed strand of primary xylem 1.8 cm in diameter (FIG. 11.23). Histologically, the xylem consists of plates of tracheids interspersed with parenchyma. Metaxylem elements have scalariform to alternately arranged circular pits. Protoxylem tracheids are not identifiable in the first-order axes. Traces that supply second-order branches arise through division of cells at the outer margins

of the ribs of the stele (FIG. 11.23). At higher levels, each xylem strand is mesarch with one to three protoxylem strands. The cortex consists of elongate parenchyma cells with abundant lacunae. Nothing is known about the ultimate segments or fertile organs of the plant. Small, wedge-shaped leaves lacking leaf cushions characterize the Middle Devonian plant Duisbergia mirabilis (Schweitzer, 1966). The plant is thought to have been an upright, unbranched trunk about 2 m tall arising from a clubshaped base (FIG. 11.24). At the apex was a dense crown of leaves borne in a tight helix; the leaves appear to be in vertical rows. The ultimate segments consist of fan-shaped leaves, each 5 cm long, with striations on the surface of the lamina suggestive of a venation pattern. In the original description of D. mirabilis, the vascular system is described as polystelic (Kräusel and Weyland, 1929) (FIG. 11.25), and well-preserved petrified specimens from Germany exhibit up to 60 band-shaped strands of what was interpreted as secondary xylem arranged in a ring. There is some similarity between Duisbergia and certain lycopsids such as Sigillaria and Pleuromeia, but the stelar organization and numerous vascular bundles in the leaves are not characteristic of lycopsid. Berry and Fairon-Demaret (1997) suggested that there is some similarity between stem features of Pseudosporochnus and D. macrociccatricosus and hypothesize that Duisbergia may represent the lower trunk of P. nodosus. Polypetalophyton is a Late Devonian genus from China (Hilton et al., 2003c). Four branches are produced at each node with long internodes between. The vascular system

396

Paleobotany: the biology and evolution of fossil plants

Figure 11.25

Figure 11.24 Suggested reconstruction of Duisbergia mirabilis. (From Schweitzer, 1966; courtesy U. Schweitzer and R. Gossmann.)

consists of several peripheral strap-shaped primary xylem strands and in the center of the axis, a central, spherical bundle. Sporangia are terminal and produced on fertile appendages; ultimate sterile segments are flattened and of the Sphenopteridium type.

Richard Kräusel.

CALAMOPHYTON PLANT Calamophyton and Hyenia (discussed below) are two Devonian taxa that are now included in the Pseudosporochnales, but they have a long and changing taxonomic history (Bonamo and Banks, 1966). Both were at one time included in the Protoarticulatae (Hyeniales), a group established by Kräusel and Weyland (1926) for Devonian plants that were presumably herbaceous, and had features suggesting affinities with the sphenophytes. At that time, the Hyeniales were regarded as precursors of the articulates because of certain morphologic features and their Devonian age. In general, the Hyeniales were characterized by dichotomous branches arising from a rhizome system. Ultimate branches had sterile appendages, which were thought to have been borne in a pseudowhorl. Sporangia occurred on modified branches and were sometimes recurved. All forms were thought to have been homosporous. The Hyeniales were interpreted as occupying an evolutionary position that paralleled the Devonian (pre-)lycophytes, for example, the Drepanophycales and Protolepidodendrales. Today the concept of the Hyeniales is much different owing to the discovery of petrified axes of Calamophyton. This important discovery by Leclercq and Schweitzer (1965) (FIG. 11.26), and subsequent reports by Schweitzer (1972, 1973) suggested that both Calamophyton and Hyenia had a unique anatomy that was unlike that found in sphenophytes. Hyenia, originally the type genus of the order, is now known from both compressed and structurally preserved axes from the Middle Devonian. The rhizome described for

CHapter 11

Figure 11.26 Winifred Goldring (left) and Suzanne Leclercq. (Courtesy J. Galtier.)

Hyenia elegans is now interpreted as a series of branches (Fairon-Demaret and Berry, 2000). Extending from the branches are dichotomizing appendages, each of which produces three pairs of elongate sporangia. Sporangia open by a longitudinal slit, and the spores are large and trilete. The sterile appendages are three-dimensional structures that also dichotomize several times. There is relatively little uniformity about the vascular system in the ultimate segments. The suggestion that the reduced branches were leaves was initially based on the assumption that subsequent planation and filling in with laminar tissue would closely approximate the leaves in the articulates. Schweitzer (1972) described a specimen of H. elegans that consists of a rhizome almost 2 m long and 4.4 cm in diameter. It displays helically arranged aerial appendages that branch dichotomously, which contrasts with the whorled pattern originally thought to characterize the genus. A reinterpretation of H. elegans and H. complexa by Fairon-Demaret and Berry (2000) provided convincing evidence that these species are now congeneric with Calamophyton primaevum. The plant originally described as Cladoxylon scoparium by Kräusel and Weyland (1926) is now also included within

Ferns and early fernlike plants

397

the whole-plant concept of Calamophyton primaevum (Schweitzer and Giesen, 1980; Fairon-Demaret and Berry, 2000). This species was originally based on poorly preserved specimens from the Middle Devonian of Germany. It was reconstructed as a plant 35 cm tall with a digitately branched main axis bearing helically arranged, segmented leaves. Although initially interpreted as flattened laminar units bearing sporangia along the margin, the sporangial appendages of C. primaevum consist of short, dichotomizing fertile branches, each bearing a single, terminal sporangium (Schweitzer and Giesen, 1980). In transverse section, the number of xylem arms is variable, ranging from 10 to 33. Successive sections indicate that there were fewer xylem arms at more distal levels. Xylem strands are radially elongate in cross section and clusters of fibers are present in the ground tissue of the stem. Since Hyenia elegans, Cladoxylon scoparium, and Calamophyton primaevum were all named in the same publication (Kräusel and Weyland, 1926), the question of which name has priority is a difficult one. Fairon-Demaret and Berry (2000) suggested that Calamophyton primaevum is the most appropriate name because (1) the type material of Hyenia from Norway has not been reexamined and (2) the anatomy of C. primaevum axes does not correspond to the generic concept of Cladoxylon. With the removal of Cladoxylon scoparium from this genus, all the remaining species of Cladoxylon stems are Mississippian in age. Calamophyton was a plant at least 60 cm tall. Middle Devonian specimens of C. bicephalum (C. primaevum of some authors) (Leclercq and Andrews, 1960) consist of a central axis with branches produced in a digitate pattern (FIG. 11.27). Second-order branches, some up to 1 cm in diameter, branch dichotomously. In earlier reports, transverse striations on the axes were thought to represent nodes, but these are now interpreted as cracks formed during fossilization. In 1965, Leclercq and Schweitzer described a structurally preserved specimen of C. bicephalum. It is unbranched and 9.5 cm long; in cross section the axis has 14–16 distinct vascular segments that are either oval or consist of 4–5 radially disposed arms. All xylem segments are primary, with the ones near the periphery of the stem containing permanent protoxylem strands surrounded by mesarch xylem. In C. bicephalum there is a gradation in the size of the sterile appendages, with more basal ones being larger and more fully developed. Leaves of Calamophyton are three-dimensional structures, each dichotomizing once in a horizontal plane and once in a vertical plane. What are interpreted as young leaves in the distal region fork only once, whereas those in the more basal regions of the axis may dichotomize up to four times.

398

Paleobotany: the biology and evolution of fossil plants

Figure 11.28 Cross section of Arachnoxylon kopfii stem (Devonian). Bar  2 mm. (From Stein, 1981.)

Iridopteridales Figure 11.27 Calamophyton bicephalum showing digitate

branching (Devonian). Bar  9 cm. (Courtesy M. Fairon-Demaret and J.-P. Gerrienne.)

Some Calamophyton specimens suggest that there is a distinct separation between branches bearing only leaves and those producing sporangia. Sporangia are not aggregated into strobili or cones. The sporangium-bearing branches are borne oppositely along the fertile axes in a pseudowhorled arrangement, and sporangium-bearing branches do occasionally occur between the sterile whorls. Each fertile branch consists of a basal stalk that is divided into an upper and lower segment. Each segment is further subdivided into three shortened side branches that terminate in a reduced bifurcation, each bearing a single sporangium, so that in C. bicephalum, each branch produced a total of 12 sporangia (Leclercq and Andrews, 1960). The sporangia are cylindrical and pointed at the distal end; dehiscence is thought to have occurred on the ventral side. Calamophyton bicephalum spores exhibit considerable variability in size, ranging from 86–166 μm in diameter. Each is circular–subcircular in outline, with a distinct trilete suture and delicate spines covering the surface.

This is a small group of Middle and early Late Devonian, and perhaps Mississippian, plants that have several features in common with the Pseudosporochnales. Members of this order have highly dissected, ribbed protosteles with permanent protoxylem strands and whorled trace departure (Berry and Stein, 2000). Branching is iterative, that is the branching pattern is repeated in higher branch orders, and two types of laterals are produced at nodes, each supplied by a single trace originating from one rib of the stele. Xylem is mesarch, with tracheids largest toward the center of the xylem ribs. Meyer-Berthaud et al. (2007) indicated that the most definitive feature that separates iridopterids from pseudosporochnaleans is the single vascular trace from the xylem rib that supplies both the appendages and lateral branches. One of the best-known iridopterids is Arachnoxylon (Read, 1938a), a genus used for anatomically preserved stem fragments with a ribbed protostele (FIG. 11.28) of only primary tissues. At the end of each xylem arm is a strand of protoxylem or a cavity that resembles a peripheral loop (FIGS. 11.29, 11.30). Traces in Arachnoxylon were of three types that differed in size and position; laterals were produced in whorls. Primary xylem was mesarch, with secondary wall thickenings that were scalariform to circular elliptical bordered pits. In A. kopfii (Givetian of

CHapter 11

Figure 11.29 Cross section of Arachnoxylon minor (Devonian). Bar  1 mm. (Courtesy C. Beck.)

Figure 11.30 Cross section of Arachnoxylon minor xylem strand showing protoxylem in center (Devonian). Bar  60 μm. (Courtesy W. E. Stein.)

New York), the axes are 1.4 cm in diameter and include six to seven xylem strands (Arnold, 1935a). In specimens of A. minor (FIG. 11.31), the xylem strand is X- or H-shaped in transverse section (Stein et al., 1983). Projections are present on the

Ferns and early fernlike plants

399

Figure 11.31 Cross section of Arachnoxylon minor (Devonian). Bar  1 mm. (Courtesy W. E. Stein.)

surface of some specimens that may represent a type of epidermal trichome. Compsocradus is known from both impression–compression and permineralized specimens (Berry and Stein, 2000). This Middle–Late Devonian plant is similar to Arachnoxylon and is interpreted as pseudomonopodial with upright, first-order branches. Ultimate appendages are recurved, with the fertile ones terminating in pairs of sporangia. Both fertile and sterile appendages occur in whorls; fertile ones consist of up to 64 per node. Spores are not known. The stele consists of six primary xylem ribs, each with a permanent primary xylem strand toward the tip. Tracheids have uniseriate, circular-tooval bordered pits. Cortical cells are thick walled and sometimes contain opaque deposits. Another member of this order is Iridopteris (Arnold, 1940). The permineralized stems possess a five-ribbed protostele with mesarch xylem (FIG. 11.32). In I. eriensis, the axes are 5.5 cm in diameter and characterized by two types of traces (Stein, 1982a). Minor traces are circular in transverse section and produced by radial divisions at the periphery of the primary ribs in a regular, alternate pattern. Larger (major) traces are elliptical in cross section with a protoxylem strand at each end. Although Iridopteris and Arachnoxylon are similar, the bilateral symmetry and smaller size of Iridopteris are the main features that distinguish these two taxa. In Metacladophyton from the Givetian of China, first-order branches occur in whorls of two to seven and internodes are widely separated (Z. Wang and Geng, 1997). Stems are 1.5 cm in diameter and consist of lobed primary xylem arranged in a U-shaped stele, with each lobe possessing one to two peripheral protoxylem strands. Another species,

400

Paleobotany: the biology and evolution of fossil plants

Figure 11.33 Cross section of Asteropteris noveboracensis

(Devonian). Bar  3 mm. (Courtesy W. E. Stein.)

Figure 11.32 Cross section of Iridopteris eriensis stem (Devonian). Bar  1 mm. (Courtesy W. E. Stein.)

M. ziguinum, is now known to have produced whorled traces, peripheral protoxylem lacunae, and two patterns of stellar architecture (D.-M. Wang and Lin, 2007). Ibyka amphikoma is a Middle Devonian plant that has been transferred from the sphenophytes to the Iridopteridales (Skog and Banks, 1973). This taxon is known from both compressed and permineralized specimens, and has been reconstructed as a monopodial branching system bearing whorled, lateral branches (Berry et al., 1997). Ultimate appendages are produced on several orders of branches. Both recurved ultimate appendages and branches are covered by delicate trichomes. The vascular system consists of a three-lobed stele; some of the arms bifurcate at higher levels to form terete leaf traces. Protoxylem tracheids and parenchyma occur near the tips of the xylem lobes in the form of permanent protoxylem strands. Sporangia on the ultimate branches are ovoid and 1 mm long. A Middle Devonian whorled branching system earlier named Hyenia vogtii has been transferred to this genus as I. vogtii (Berry, 2005). Serripteris is a middle Tournaisian (Mississippian) plant from southern France that has tentatively been placed in the

Iridopteridales (Rowe and Galtier, 1989). The specimens are permineralized and consist of several orders of branching axes. The stele is four ribbed and mesarch. Asteropteris is a Late Devonian genus that has a stellate protostele with conspicuous permanent protoxylem strands (FIG. 11.33) (Bertrand, 1913). Traces are tetrapolar and arranged in a very shallow helix. Anapaulia moodyi (see Chapter 10) is a plant from the Middle and Upper Devonian of Venezuela that is tentatively assigned to the Iridopteridales (Berry and Edwards, 1996b). It is known from impression–compression specimens and interpreted as being monopodial, with branches produced in a pseudowhorl. Ultimate branches have recurved tips; elliptical sporangia are borne on some, and all branching systems are covered in delicate spines. Another Late Devonian iridopterid is Rotoxylon dawsonii (Cordi and Stein, 2005), initially called Cladoxylon dawsonii (Read, 1935). The xylem strands exhibit a variety of configurations in cross section. Xylem maturation is mesarch and one to three permanent protoxylem strands occur at the tip of each of the xylem arms. Tracheids have elliptical-toelongate bordered pits. Anatomical features do not resolve the placement of R. dawsonii in either the Pseudosporochnales or Iridopteridales (Meyer-Berthaud et al., 2007). Phylogenetic Position of the Cladoxylopsids

The Pseudosporochnales and Iridopteridales represent perplexing, albeit interesting groups of vascular plants that are

CHapter 11

united based on the structure of their vascular systems. In spite of this clearly artificial classification, the groups do possess several interesting features that suggest a degree of evolutionary specialization by the Middle Devonian. One of these is the highly dissected vascular system that appears to include an increasing number of xylem strands at more distal levels in the plant. It has been suggested that the cladoxylopsid vascular system may have evolved by the dissection of an actinostele. Although important details regarding the anatomy of the distal-most branch segments are lacking in the Pseudosporochnales, the unique configuration of the vascular system may represent a stage in the evolution of a complex frond system. An unusual axis with cladoxylopsidlike anatomy (Hapsidoxylon) is also known from the Triassic of Antarctica (McManus et al., 2002b). It remains to be seen whether it represents a Triassic member of this group or an example of a pattern of vascular system architecture that has evolved again in a different group of plants. The phylogenetic position of the Pseudosporochnales and Iridopteridales remains questionable. At one time, the presence of a so-called polystelic organization was the basis for suggesting affinities with the medullosan seed ferns. The Cladoxylopsida has also been regarded as a plexus of vascular plants from which a number of groups may have evolved, including the sphenophytes (Stein et al., 1984). The similarity in anatomy has been used as a basis for linking this group with members of the aneurophytalean progymnosperms and the coenopterid ferns. This latter relationship was based principally on the presence of parenchyma-filled peripheral loops in both groups. Scheckler (1975a) suggested that the so-called peripheral loops within the Pseudosporochnales were actually of two distinct types. In one type, longitudinally oriented rods of parenchyma are surrounded by mesarch xylem. This type of organization is exemplified by plants like Cladoxylon taeniatum and others. The other configuration was typified by Rhymokalon, in which the protoxylem strands disintegrate to form lacunae that resemble peripheral loops. These peripheral loops contain no parenchyma and only remnants of protoxylem tracheids. Scheckler suggested that these two conditions may provide clues to relationships with other vascular plant groups, that is the parenchymatous type is related to certain coenopterid ferns and the type lacking parenchyma is perhaps more closely allied with other groups. Detailed anatomical studies led Stein (1981) to suggest that the so-called peripheral loops found in many Devonian plant groups (e.g., Pseudosporochnales, Iridopteridales, and coenopterid ferns) are formed in a variety of ways and are not homologous, and therefore are not useful in phylogenetic analyses. Using extant plants as developmental analogs for

Ferns and early fernlike plants

401

certain fossils, Beck and Stein (1993) proposed the term permanent protoxylem strand for these areas in the iridopterids, cladoxylaleans, and early sphenopsids. A second designation, radiate protoxylem group, is used for certain plants in which diverging peripheral protoxylem strands occur in a one-toone relationship with the lateral appendages produced by the shoot. Included in this group are the aneurophytes, certain trimerophytes, and perhaps early seed plants. It will be interesting to see whether these protoxylem characterizations have phylogenetic validity. There are relatively few characters that can be used in a phylogenetic analysis of the Pseudosporochnales and Iridopteridales. Because of the limited length of many permineralized specimens, it is difficult to determine how vascular patterns change over an extended length of the stem. Cordi and Stein (2005) also cautioned that defining groups based on whorled phyllotaxis and internal protoxylem strands may not represent good examples of synapomorphies. Although there have been several ideas proposed regarding the relationships of these Devonian to Early Mississippian vascular plants (Berry and Stein, 2000; Cordi and Stein, 2005), there remains little agreement as to whether they are monophyletic or paraphyletic. One hypothesis places the cladoxylopsids, ferns, and sphenophytes within a monophyletic group termed the moniliformopses (Kenrick and Crane, 1997a; Doyle, 1998). A second analysis suggests that the cladoxylopsids form a monophyletic group together with the Rhacophytales and Zygopteridales (Rothwell, 1999). Some suggest that some of these Devonian plants have affinities with certain groups included in the sphenophytes (Berry and Stein, 2000; Soria and Meyer-Berthaud, 2003), whereas others see relationships within the true ferns. An accurate understanding of the phylogenetic placement of these plants is hampered by the low number of phylogenetically significant characters and the fragmentary nature of the fossils.

Early fernlike plants Rhacophytales

The Rhacophytales includes two genera of Devonian plants that encompass an interesting suite of morphological and anatomical characters. Like the Pseudosporochnales and Iridopteridales, their taxonomic position continues to remain speculative. They are characterized by unique branching systems, which include biseriate and quadriseriate units subtended by aphlebiae. The ultimate appendages may be dichotomously divided and in some they are planated (Cornet et al., 1976). The vascular system consists of a clepsydroid-shaped primary xylem strand surrounded by secondary xylem. Sporangia are borne in clusters

402

Paleobotany: the biology and evolution of fossil plants

Figure 11.34 Portion of a sterile frond of Rhacophyton sp.

(Devonian). Bar  4 cm.

and are exannulate. Current information suggests that all the plants assigned to the Rhacophytales are homosporous. RHACOPHYTON The most completely known member is Rhacophyton, and has been described from several Upper Devonian sites around the world. Rhacophyton ceratangium (Andrews and Phillips, 1968), probably the best-known species, is an upright plant 1 m tall. Branches, which have been called fronds (FIG. 11.34), consist of quadriseriate lateral branches (also termed tetrastichous, i.e. borne in four ranks or rows) that are attached to the main axis in two vertical rows (FIG. 11.35). Branch segments are 50 cm long. In R. ceratangium, branch morphology is highly variable, including a gradational series between two- and three-dimensional forms. Aphlebia-like appendages are attached to the main stem in pairs. Numerous pinnule-like structures are borne at the distal ends of the ultimate branch segments. Ultimate segments are of three basic types; all are nonlaminar and differ in the degree of flattening (Cornet et al., 1976). Stems of R. ceratangium are up to 2 cm in diameter and consist of a central, clepsydroid-shaped strand of primary xylem (FIG. 11.36) surrounded by radially aligned rows of scalariform-bordered tracheids that make up the secondary xylem. Extending through the secondary xylem are uniseriate vascular rays one to seven cells high (Dittrich et al., 1983). At each end of the narrow primary xylem bar is a so-called peripheral loop (FIG. 11.36) that may contain parenchyma at some levels. Additional information about the anatomy of Rhacophyton is known from Leclercq’s (1951) detailed study of

Figure 11.35 Rhacophyton ceratangium fertile unit (Devonian). (From Andrews and Phillips, 1968).

Figure 11.36 Cross section of Rhacophyton ceratangium stem

showing peripheral loop (arrows) at each end of the primary xylem bar, and radially aligned tracheids of secondary xylem (Devonian). Bar  2 mm.

R. zygopteroides. This species, from the Upper Devonian of Belgium, consists of an upright axis bearing helically arranged fronds. Pinnae are produced in two rows and contain small dissected pinnules up to 1 cm long. In cross section, the stem contains a star-shaped xylem strand with an outer zone of scalariform tracheids and an inner zone of poorly preserved cells. Traces arise from the arms of the stele as rectangular bars, each with opposite peripheral loops. At higher levels the traces become clepsydroid in transverse configuration. Pinna traces are crescent shaped in cross section, with subsequent pinnule traces C-shaped. Traces to the fertile pinnae are similar, except

CHapter 11

that two strands enter the bases of paired pinnae. A small amount of secondary xylem is present in R. zygopteroides. Another marked distinction between the two species is the apparent bilateral symmetry of R. ceratangium; the axis of R. zygopteroides is described as stellate in outline. Rhacophyton condrusorum is a species from the Upper Devonian of Germany (Schultka, 1978) that compares favorably with the specimens of R. ceratangium from North America. The so-called fertile fronds of Rhacophyton are common, but none have yet been found organically attached to the main axis. In general, the fertile frond is three dimensional with two elongate, sterile pinnae and two branched, fertile pinnae. This four-parted branch has been described as a nodal unit. Each fertile pinna consists of a basal stalk that dichotomizes several times in a relatively short distance to produce a dense, three-dimensional structure. Penultimate branches of each cluster are slightly curved toward the center of the unit and bear smaller branches along their inner surfaces. Smaller branches in turn dichotomize, each terminating in a sporangium. Considerable variation occurs in the number of parts present in each nodal unit of R. ceratangium. In some specimens, either one fertile or sterile pinna may be lacking. Similar variation in frond morphology is also noted in a specimen from West Virginia, in which each sterile frond branches at the midpoint to produce both a fertile and a sterile segment. Sporangia of R. ceratangium are fusiform, with an elongate, curved tip; they measure up to 2.4 mm long and 0.4 mm in maximum diameter. Dehiscence is longitudinal and apparently does not involve the tip of the sporangium. Spores are ovoid, 50 μm in diameter, and exhibit a faint trilete suture on the proximal surface. The sporoderm is often wrinkled and ornamented by delicate grana (small, wart-like structures); these spores appear similar to the dispersed graintype Perotriletes. OTHER TAXA Another plant that has been compared with Rhacophyton is Protocephalopteris. Specimens are known from the Middle Devonian of Spitsbergen and Siberia, and consist of large, bipinnate fronds with alternately arranged pinnae (Schweitzer, 1968, 1999). Sterile ultimate segments, similar to those of Rhacophyton, are also alternately arranged. The fertile unit consists of pairs of ultimate fertile segments bearing pairs of sporangia, each pair subtended by aphlebiae. A major distinction between Protocephalopteris and Rhacophyton is the apparent quadriseriate branching of the ultimate fertile appendages in the latter. Chlidanophyton is known from the Late Devonian into the early Tournaisian and includes

Ferns and early fernlike plants

403

Figure 11.37 Suggested reconstruction of fertile axes of Eocladoxylon minutum (Devonian). (Modified from Berry and Wang, 2006.)

plants with terminally borne, recurved, exannulate sporangia attached to branching systems (Gensel, 1973; Hilton, 1999). Like other rhacophytaleans, its phylogenetic relationships continue to remain problematic. Eocladoxylon is a Middle Devonian (Givetian) fossil from China that has both sterile and fertile branches resembling those of Rhacophyton (Berry and Wang, 2006b). The primary xylem is a clepsydroid-shaped protostele with scalariform to circular-bordered secondary wall thickenings; there is no secondary xylem. Ultimate segments are opposite to subopposite and planated (FIG. 11.37); fertile segments are three dimensional with paired sporangia. There is some similarity between the sporangia of E. minutum and the pollen sacs of younger putative seed fern reproductive organs like

404

Paleobotany: the biology and evolution of fossil plants

Telangiopsis. It has been suggested that E. minutum may be morphologically most similar to Rhacophyton, and possibly related to some members of the Iridopteridales. Protopteridophyton is a Devonian (Givetian–Frasnian) plant from South China that combines characters of trimerophytes, ferns, and progymnosperms (C.-S. Li and Hsü, 1987). The main axis produces alternately arranged, pinnate laterals that dichotomize multiple times; laminae are lacking; in P. devonicum the ultimate appendages are three dimensional (C.-S. Li and Hsü, 1987). The xylem strand of the main axis is V-shaped in cross section and characterized by mesarch primary xylem with protoxylem strands located at the periphery of each arm. Pitting ranges from simple scalariform to scalariform-pitted. Fertile branches in P. devonicum terminate in paired, recurved sporangia that contain spherical trilete spores. Ellesmeris sphenopteroides is a Late Devonian (Frasnian) plant with quadriseriate branching from Ellesmere Island in the Canadian Arctic, which exhibits features of both the Rhacophytales and Zygopteridales (S. Hill et al., 1997). The stele of the main axis is bipolar and clepsydroid in outline. Aphlebiae are produced at the base of each primary pinna and the ultimate foliar segments (pinnae) are laminar (FIG. 11.38) and of the Sphenopteris type (see Chapter 16). In comparing this species to zygopterid ferns, Ellesmeris is older than the other zygopterids and has fully laminar pinnules earlier than previously hypothesized (Viséan). Another relatively advanced feature is the presence of circularbordered pitting in the metaxylem tracheids. S. Hill et al. (1997) suggested that E. sphenopteroides may exhibit one way in which a biseriate frond could have evolved from a quadriseriate type—by the arrested growth of one pinna in a pair of pinnae (FIG. 11.38). Phillips and Galtier (2005) noted that Ellesmeris does not show a clear differentiation between leaf and stem, and therefore is better considered with the Rhacophytales rather than the zygopterid ferns. Systematics of the Rhacophytales

The systematic position of members of the Rhacophytales continues to remain unresolved. Rhacophyton was initially included in the Aneurophytales (Kräusel and Weyland, 1941), a group at that time reserved for plants believed to be the progenitors of the pteridosperms and true ferns. Rhacophyton has also been assigned to the Protopteridales, an order that was once used for fernlike plants characterized by clusters of terminal sporangia, but lacking laminar pinnules (Høeg, 1942). Others classified Rhacophyton in the Coenopteridales (Leclercq, 1954), within the Zygopteridales (Eggert, 1964), or as an ancestral group from which some of the coenopterids evolved.

Figure 11.38 Suggested reconstruction of Ellesmeris sphenopteroides showing ultimate pinna (right) and aphlebiae (left) (Devonian). (From Hill et al., 1997.)

Members of the Rhacophytales share features with certain taxa of the Iridopteridales (Berry and Fairon-Demaret, 2001) and progymnosperms (Matten, 1974), and some authors include them as a family within the Zygopteridales. Features in common with zygopterid ferns include the clepsydroidshaped xylem strand and quadriseriate branching. Characters shared with progymnosperms include the presence of secondary xylem and the repeated nature of the xylem strand from one order to the next, although the presence of laminate ultimate appendages (leaves) and the anatomy of the branches in the progymnosperms argue against such a relationship (Chapter 12). At the present time, members of the Rhacophytales are perhaps best interpreted as representing a level of evolution that is more advanced than the trimerophytes from which they evolved, but still lacking the organ differentiation seen in slightly younger fernlike plants, such as the zygopterids. The clepsydroid xylem strand in members of the Rhacophytales has also been used to suggest a relationship with some cladoxylopsids; however, as noted by

CHapter 11

Ferns and early fernlike plants

405

Berry and Stein (2000), not all of the structures with bipolar traces are necessarily homologous. Coenopterid Ferns

The Coenopteridales are no longer a formal order of fossil plants, but are important historically. The taxa that were once included in this order are now placed within other groups of ferns. The initial circumscription of the coenopterids included ferns and fernlike plants that ranged from the Devonian into the Permian. Many possessed well-organized megaphylls, circinate vernation, and shoot-borne roots. Some had foliar-borne sporangia; in others the sporangia were unknown or were produced on nonlaminar foliar organs. Vascular anatomy ranged from protostelic forms to laterappearing types characterized by siphonosteles. Only one genus, Zygopteris, possessed secondary xylem. The primary basis for classifying plants in this order was the symmetry and histology of the leaf traces, and the general morphology of the frond. Coenopterids were viewed as a heterogeneous collection of Paleozoic plants defined principally on anatomical characters, and at one time included four families (Eggert, 1964): Stauropteridaceae, Botryopteridaceae, Anachoropteridaceae, and Zygopteridaceae. As research with coenopterids progressed, principally with permineralized plants from the Carboniferous of France and North America, and as additional specimens were discovered in which the reproductive parts could be identified, many coenopterid taxa were recognized as true ferns (Phillips, 1974). This resulted in a modified classification in which taxa originally placed within the Botryopteridaceae and Anachoropteridaceae were transferred to families within the Filicales. The two remaining families, Stauropteridaceae and Zygopteridaceae, are known in some detail, and their reproductive parts and vegetative anatomy are unlike that of any filicalean ferns. Therefore, in this volume these two groups are treated at the ordinal level.

M

M

Longitudinal section of Stauropteris burntislandica with megaspores (M) (Mississippian). Bar  250 μm. Figure 11.39

Stauropteridales

The Stauropteridales includes small, bushy plants with quadriseriate or biseriate branching that lack planated laminae. In cross section the xylem strand is four lobed and slightly bilaterally symmetrical; branch anatomy is generally repeated in the higher orders of branching. Sporangia are terminal and both homosporous and heterosporous forms are known. The group is currently represented by three genera that range from the Devonian to Pennsylvanian of Europe and North America: Stauropteris, Gillespiea, and Rowleya. In Stauropteris oldhamia (Binney, 1872), branches were borne in pairs, with the configuration of the xylem strands

Figure 11.40 Longitudinal section of Stauropteris burntislandica with Megaspores (M) (Mississippian). Bar  250 mm

similar to that of the primary axis. Six orders of branches have been reported, with the vascular tissue becoming smaller at higher levels and finally appearing as a terete protostele. Xylem development is mesarch (FIG. 11.40);

406

Paleobotany: the biology and evolution of fossil plants

Figure 11.42 Section of Stauropteris oldhamia sporangium

(Pennsylvanian). Bar  350 μm.

Figure 11.41 Suggested reconstruction of Stauropteris oldha-

mia. (From Eggert, 1964.)

no secondary tissues have been found. Tracheid pitting is scalariform, and phloem occupies the furrows between the ridges of xylem (Smoot and Vande Wege, 1986). The threedimensional branching pattern in Stauropteris has been interpreted as demonstrating an early stage in the evolution of the frond. In the Pennsylvanian species S. oldhamia, pairs of highly dissected aphlebiae occur at the base of each branch (FIG. 11.41). Thick-walled cells and small vascular strands in the aphlebiae suggest that they were non-photosynthetic, rigid structures similar to thorns or spines in some extant plants. Sporangia have thick walls (FIG. 11.42) and occur terminally on ultimate axes. The number of wall layers near the distal end of the sporangium is reduced, suggesting that the exannulate sporangium dehisced through an apical pore or stomium. Spores are spherical and range from 32–40 μm in diameter. Based on the consistent occurrence of one type of spore, S. oldhamia is thought to have been homosporous. Stauropteris burntislandica (Bertrand, 1909) occurs in Mississippian rocks. The axes are similar to S. oldhamia,

although the aphlebiae show considerable morphologic variation. Each aphlebia is attached to the axis by an oblique ridge slightly below the insertion point of the branch. Aphlebiae branch irregularly, with some ultimate segments up to 4 mm long. One interesting feature of this species is the fact that it is heterosporous. This was initially recognized in 1952 by the noted Indian paleobotanist K. R. Surange, although the megasporangium had been described earlier as Bensonites fusiformis, but thought to represent a glandlike structure. Each megasporangium is spindle shaped and 1.3 mm long (FIG. 11.39). The lower half of the sporangium is parenchymatous and represents the tip of a branch. The upper (distal) half consists of a sporangial cavity with a wall that is only one cell layer thick. At maturity, the contents were released through the distal tip. The number of megaspores in S. burntislandica is typically two, but some specimens are known with three. One report suggested that up to eight megaspores per sporangium were present (Lacey et al., 1957), but this may have been the result of folded spores viewed in section. Chaloner (1958) described dispersed megaspores of the S. burntislandica type (Didymosporites) in which two smaller (presumably abortive) spores are associated with the two larger, functional megaspores. Both large and small spores are trilete and embedded in what appears to have been a series of tapetal membranes. The large spores range up to 580 μm in diameter, whereas the smaller ones are generally in the 45-μm-size class. Microsporangia have also

CHapter 11

Ferns and early fernlike plants

407

Figure 11.43 Cross section of Stauropteris biseriata axis (Penn-

sylvanian). Bar  1 mm.

been described in S. burntislandica (Scott, 1908). They are ovoid and 0.6 mm long, and differ from the sporangia of S. oldhamia in that the sporangial wall is one cell layer thick. Microspores are 30 μm in diameter and trilete. It is not known whether micro- and macrosporangia were borne on the same or different branches, or on different plants. Stauropteris biseriata (FIG. 11.43), from the Lower– Middle Pennsylvanian of North America, has distichous branches that are subtended by a pair of vascularized aphlebiae (FIG. 11.44) (Cichan and Taylor, 1982b). Nothing is known about the reproductive parts of the plant. It is interesting, however, that S. biseriata has a two-ranked branching pattern like that in more advanced ferns, as opposed to the quadriseriate pattern, that is, branches borne in pairs and arranged alternately on opposite sides of the main axis which is present in all other species of Stauropteris. Multifurcatus tenellus is a stauropterid-type plant reported from the Carboniferous of China (Y. Wang, 2003). It has a main axis up to 3 mm in diameter and second-order branches that are produced three per node. Dichotomously branched laterals, each with a sporangium, arise from the nodes of branches. Sporangium size is used to suggest that the plant had both micro- and megasporangia. Gillespiea is interpreted as a Devonian member of the Stauropteridales (Erwin and Rothwell, 1989). Gillespiea randolphensis includes smooth, limonitized axes that branch quadriseriately (FIG. 11.45). Larger axes are protostelic and appear triangular to four-sided in cross section; smaller axes have terete protosteles. Fusiform megasporangia, each

Figure 11.44 Suggested reconstruction of Stauropteris biseriata. (From Cichan and Taylor, 1982b.)

with a blunt tip, are borne singly or in pairs on small axes that occupy the position of aphlebia in other stauropterids. Megasporangia are 1 mm long and contain one to two trilete spores in the 160-μm-size range. The position and size of the megasporangia in Gillespiea is similar to that of S. berwickensis (Long, 1966). Nothing is known about the microsporangiate parts of G. randolphensis. Rowleya trifurcata from the Westphalian A of Lancashire, England, has a protostelic axis and triseriate lateral branches (Long, 1976). The tetrarch protostele is similar to the stelar structure in Stauropteris. At more distal levels, terete branches, interpreted as leaves, are arranged in pairs. The genus has been compared with Stauropteris, Psalixochlaena, and some species of Botryopteris.

408

Paleobotany: the biology and evolution of fossil plants

Figure 11.45 Suggested reconstruction of Gillespiea randolphensis (Devonian). (From Erwin and Rothwell, 1989; in Taylor and Taylor, 1993.)

Zygopteridales

The Zygopteridales once represented the largest of the coenopterid fern groups. They are now interpreted as true ferns and have a geologic history that extends from the Mississippian (Tournaisian) into the Permian. Members have elaborate, three-dimensional fronds bearing pinnae in four ranks. In cross section, the foliar members possess distinctive stele configurations (C, E, H, I, X, or Y) that have been used to subdivide the order. Sporangia were elongate and produced in large clusters on the abaxial surface of pinnules or at the tips of frond branches. The zygopterid ferns have a rachis (pl. rachides) that is radially symmetrical in cross section and often referred to as a phyllophore (Phillips, 1974). This term has been widely used in discussions of coenopterid ferns to emphasize an organization that differs from the typical C-shaped vascular strand of true ferns, which is bilaterally symmetrical in cross section. The phyllophore exhibits bipolar primary xylem strands and some have peripheral loops; traces to pinnae depart the xylem strand as crescent-shaped bundles that, in turn, produce lateral traces. Members of the Zygopteridales are sometimes separated into two groups on the basis of frond morphology and anatomy. The etapteroid type has four ranks of primary pinnae and so-called open peripheral loops, which are closed only during pinna trace production. The clepsydroid type exhibits two ranks of primary pinnae, as in living ferns, although some of the earliest forms have both biseriate and quadriseriate pinnae arrangement. In transverse section, the phyllophore of this group is hourglass shaped (clepsydroid), and the peripheral loops remain closed when traces to the

Figure 11.46 Jean Galtier.

pinnae are produced (FIG. 11.47). The discovery of fertile parts for many of these plants now makes it possible to consider the ecological parameters and systematic affinities of the group in more detail (Phillips and Galtier, 2005). Galtier (1966a,b) (FIG. 11.46), and Phillips and Galtier (2005) have reviewed the features of all zygopterids based on permineralized specimens. The 12 morphospecies included in Clepsydropsis (FIG. 11.48) have small, C-shaped pinna traces and peripheral loops with little parenchyma (Phillips and Galtier, 2005). In one species, C. leclercqii, from the lower Viséan (Middle Mississippian) of France, the clepsydroid anatomy is maintained, but the frond is quadriseriate, like members of the etapterid group (Galtier, 1966b). Underscoring the unusual nature of Clepsydropsis is the fact that the foliar members have been found attached to more than one type of stem. For example, Asterochlaena (Corda, 1845) is a Permian form with a deeply lobed cylinder of primary xylem. The stem is 8 cm in diameter and includes shoot-borne roots and numerous foliar members of the clepsydroid type. The stem stele contains a mixed pith surrounded by scalariform tracheids. Traces arise from the blunt arms of the stele, becoming hourglass shaped. Peripheral loops are located toward the abaxial surface. At higher levels, the petiole vasculature is C-shaped and adaxially directed.

CHapter 11

Ferns and early fernlike plants

409

Figure 11.47 Clepsydropsis sp. Bar  1 mm.

Figure 11.49 Dernbachia brasiliensis, transverse section

through a stem showing the cauline vascular system and leaf traces (Permian). Bar  5.5 cm. (Courtesy BSPG.)

Figure 11.48 Diagrammatic reconstructions of the petiolar or

phyllophore xylem of several zygopterid ferns. A. Clepsydropsis antiqua and Symplocopteris. B. Clepsydropsis sp. C. Clepsydropsis sp. 2. D. Clepsydropsis sp. 3. E. Clepsydropsis leclercqii. F. Clepsydropsis sp. 4. G. Clepsydropsis sp. 5. H. Metaclepsydropsis duplex. I. cf. Metaclepsydropsis. J. Diplolabis. (From Phillips and Galtier, 2005.)

Dernbachia is a small tree fern from the Permian of Brazil that is suggested to have a relatively close relationship to Asterochlaena but, that also exhibits similarities with certain members in the filicalean family Tedeleaceae (see below) (Rössler and Galtier, 2002a). Trunks of D. brasiliensis are

63–179 mm in diameter and possess a large actinostele (FIG. 11.49) (a rare feature in late Paleozoic tree ferns), surrounded by a relatively narrow parenchymatous stem cortex extending into the petiole bases. The stem is surrounded by a mantle of adventitious roots and petiole bases. Leaf traces depart the stele as oval, bipolar strands, as seen in cross section, and become pi shaped (π). At the point where the traces enter the petiole base, each consists of two lateral, enrolled strands and a small central bar. Petioles are arranged around the stem in pseudowhorls (FIG. 11.49), leading to a variable number of orthostichies as well as to two opposite sets of parastichies. Asterochlaenopsis kirgisica (Permian of Siberia) is a silicified stem 15 cm in diameter that contains a cylindrical stele with a mixed pith (Sahni, 1930). Another clepsydroid zygopterid is Austroclepsis (Sahni, 1932a), from the Tournaisian (Mississippian) of Australia. This plant was apparently a tree fern with a trunk 30 cm in diameter, which was constructed of numerous leaf-bearing stems and intertwined roots, also called a false stem. Each stem is radially symmetrical with petioles in a 2/5 phyllotaxy. Stem xylem is stellate in cross section, with a central pith and scalariform tracheids. Petiole traces are initially triangular in section view, but at more distal levels assume the characteristic clepsydroid configuration. Many aphlebiae are borne in pairs along the

410

Paleobotany: the biology and evolution of fossil plants

rachis. Although spores have been found associated with the root mantle of the false stem, nothing is known about the reproductive parts of this plant. Symplocopteris wyattii is a zygopterid tree fern from the Tournaisian of Australia (Galtier and Hueber, 2001; Hueber and Galtier, 2002). The plant is preserved as a false stem 50 cm in diameter that consists of branching stems and phyllophores (FIG. 11.50). Stems branch dichotomously, with each producing a phyllophore (FIG. 11.51) and a large, geotropic adventitious root (FIG. 11.52). Smaller roots are produced by the phyllophores and the larger roots; all roots are covered with hairs. These authors suggested that the larger roots were involved in absorption of water from the soil, whereas the smaller ones absorbed humidity from the air within the false trunk. Symplocopteris currently represents

the oldest example of a tree fern with a false trunk in the fossil record. Metaclepsydropsis duplex (Mississippian) is a genus with a horizontal rhizome that bears widely separated fronds. In transverse section, the stem stele is circular, with xylem consisting of an inner zone of long, narrow scalariform tracheids, and an outer zone of elongate, reticulate xylem elements. Based on the departure of leaf traces the ultimate frond was quadriseriate in architecture (FIG. 11.53). Specimens of M. duplex have been found in association with clusters of sporangia of the Musatea type (Chaphekar and Alvin, 1972). Diplolabis roemeri (Lower Carboniferous of Pettycur) has a solid protostele at least 13 mm in diameter with large petioles from 1 to 2 cm in diameter. Pinna traces are borne in pairs, and the configuration of the petiole trace is cruciate

Figure 11.52 Diagrammatic reconstruction of the xylem of Figure 11.50 Symplocopteris, false trunk (Mississippian). Bar  1 cm. (Courtesy J. Galtier.)

Figure 11.51 Symplocopteris, detail of one individual stem surrounded by petioles and roots (Mississippian). Bar  1 mm. (Courtesy J. Galtier.)

Symplocopteris A. and Diplolabis B. (Modified from Phillips and Galtier, 2005.)

Figure 11.53 Cross section of Metaclepsydropis duplex from Scotland (Mississippian). Bar  5 mm.

CHapter 11

(FIG. 11.54). Some believe that the lateral expansion of the peripheral loops of the Metaclepsydropsis petiole (FIG. 11.55) would result in a Diplolabis type of configuration. The culmination of such a series would have produced the Etapteris petiole (FIG. 11.62), in which there is a radial extension of the trace arms without a corresponding increase in the size of the now open peripheral loop. This series should not be interpreted as suggesting any relationships among the genera, but rather as a way in which frond morphology and the corresponding anatomy may have evolved among plants of this type. Both Metaclepsydropsis and Diplolabis

Ferns and early fernlike plants

411

(FIG. 11.56) had horizontal, rhizomatous stems and erect phyllophores. Diplolabis also bore large spines along the phyllophore and pinna rachides (Phillips and Galtier, 2005). Both Diplolabis and Metaclepsydropsis had sporangia corresponding to the genus Musatea (Galtier, 1968). Musatea globata has been found attached to pinnules that probably belong to Diplolabis (Galtier, 1981). This species includes banana-shaped sporangia with an elongate zone of conspicuous, thick-walled epidermal cells. Each sporangium contains trilete spores ranging from 30–35 μm in diameter. Small recurved, laminar structures suggestive of early pinnules are present in this zygopterid fern. Both Metadineuron (FIG. 11.57) and Dineuron (FIG. 11.58) from the Mississippian (Viséan) of France are known only from small petioles ( 5 mm in diameter) in which the xylem strand is large and ellipsoid (FIG. 11.59) (Galtier, 1964a). In Metadineuron, pinna trace emission is similar to that in Metaclepsydropsis, in which there is a temporary opening of

Figure 11.54 Cross section of Diplolabis roemeri petiole trace

(Mississippian). Bar  4 mm. Figure 11.56 Diplolabis, stem protostele and bipolar phyllophore trace (Mississippian). Bar  3 mm.

Figure 11.55 Metaclepsydropsis duplex. Transverse section through the stele of the frond rachis. Note beginning formation of pinna traces (arrow) (Mississippian). Bar  2 mm. (Courtesy BSPG.)

Figure 11.57 Metadineuron phyllophore with a single pinna trace

at left (arrow) (Mississippian). Bar  1 mm. (Courtesy J. Galtier.)

412

Paleobotany: the biology and evolution of fossil plants

Figure 11.58 Dineuron, phyllophore xylem strand with paired pinna traces (arrows) (Mississippian). Bar  0.5 mm. (Courtesy J. Galtier.)

11.59 Dineuron phyllophore (center) with trace (Mississippian). Bar  1 mm. (Courtesy J. Galtier.)

Figure

a small peripheral loop; the C-shaped pinna trace does not dichotomize. In D. pteroides, peripheral loops open, close, and then reopen to form C-shaped pinna traces that departed from the open loop. The pinna traces divide again in the cortex of the phyllophore to produce four appendages, which are thought to represent reduced aphlebiae (Phillips and Galtier, 2005).

Phillips and Galtier (2005) noted that the Zygopteris plant is perhaps the best known in this group, because it was a highly specialized, robust plant, which can be recognized in a variety of preservational modes. For example, Z. berryvillensis (FIG. 11.60) and Z. illinoiensis are common species in Middle and Upper Pennsylvanian coal ball localities in North America. Zygopteris illinoiensis has been reconstructed (Dennis, 1974; Phillips and Galtier, 2005) as consisting of a horizontal, rarely dichotomizing rhizome, bearing helically arranged, upright laminate fronds. The stem (Z. illinoiensis) was somewhat succulent, with a small amount of secondary growth. Vascularized aphlebiae were borne spirally on the rhizome and in two rows on the phyllophore (Etapteris scottii); a ramentum of scales was also sometimes present on parts of the frond, aphlebiae, and rhizome. Fertile fronds of Z. illinoiensis belong to the morphogenus Corynepteris (discussed below). Zygopteris berryvillensis is also known to bear E. scottii petioles, but has large sporangia assigned to Biscalitheca (discussed below). In contrast to the reconstruction of Dennis (1974), Phillips and Galtier (2005) noted that Z. illinoiensis has much more abundant and robust aphlebiae than does Z. berryvillensis. Following is a detailed description of the various morphogenera that make up these plants. The Carboniferous forms of Zygopteris plants are rhizomatous, whereas the Permian Z. primaria appears to have had an upright stem. In cross section (FIG. 11.61), the exarch protostele of Zygopteris is moderately stellate. One unusual feature of this taxon is the presence of secondary xylem but no secondary phloem. Although some have regarded this tissue as primary (Baxter, 1952b), others (Dennis, 1974) believe it arises from a lateral meristem and thus is secondary in origin. Uniseriate and multiseriate vascular rays divide the secondary xylem into wedges. Tracheid pitting consists of elongate, circular-bordered pits. Primary phloem surrounds the xylem and the cortex consists of a parenchymatous ground tissue with scattered clusters of sclereids, which are more abundant in Z. illinoiensis than in Z. berryvillensis. In older plants, a weakly developed periderm zone can be distinguished. The epidermis is covered with multicellular, branched peltate hairs which form a ramentum. Numerous shoot-borne, adventitious roots, each with a diarch stele, arise from the rhizome. The petiole trace of Zygopteris is H-shaped in cross section, and because it looks like the Greek capital letter eta (H), such petioles are called Etapteris (FIG. 11.62). Below the level of petiole trace formation, the rhizome stele is oval in cross section, with two conspicuous protoxylem points. At a slightly higher level, the trace is abaxially curved and bipolar; more distal sections show the development of sinuses between the arms (antennae) and the characteristic tetrapolar

CHapter 11

Ferns and early fernlike plants

413

Figure 11.60 Suggested reconstruction of Zygopteris berryvillensis. (From Dennis, 1974.)

Figure 11.61 Cross section of Zygopteris illinoiensis stem

(Pennsylvanian). Bar  5 mm.

11.62 Cross section of Etapteris scottii petiole (Pennsylvanian). Bar  3 mm.

Figure

structure of the Etapteris petiole. Pairs of aphlebiae are present at the lower levels of the petiole and may serve as protective structures for developing leaf primordia and later as photosynthetic structures on the mature leaf. Primary pinnae of Etapteris are produced in pairs (FIG. 11.63), so that the frond has a quadriseriate arrangement (Phillips and Galtier, 2005), while the pinnae and pinnules are flattened. Zygopterid fronds, such as Alloiopteris, are at least tripinnate. In one Middle Pennsylvanian specimen, the Etapteris petiole bears pairs of primary pinnae that give rise to second-order pinnae with alternately arranged laminar pinnules. Etapteris leclercqii is an early Middle Pennsylvanian species discovered in coal balls from eastern Kentucky (Smoot and Taylor, 1978). In this species, the lateral arms are attached broadly to the central bar of xylem (apolar) and taper to a point. Trace formation includes the development of a peripheral loop below the separation of pinna traces from the petiole. The surface of the petiole is covered with numerous multicellular hairs. Some believe that the thickened apolar, reduced number of pinnae, and lack of differentiation between apolar and lateral arms constitute primitive characters in the genus. The early Middle Pennsylvanian age of E. leclercqii provides some support for this hypothesis. Information about fossil phloem is relatively rare because the thin-walled cells are often destroyed during the fossilization process and because there are few suitable techniques with which to study the small sieve areas on the cell walls. In some Etapteris specimens (Smoot, 1979), however, a zone of phloem tissue approximately four cells thick surrounds the xylem core. This tissue is constructed entirely of sieve elements, each 120 μm long and 40 μm in diameter. End walls vary from horizontal to oblique and often exhibit swollen ends similar to those reported in extant ferns. Sieve areas (FIG. 11.64) consist

414

Paleobotany: the biology and evolution of fossil plants

Figure 11.63 Suggested reconstruction of Zygopteris illinoien-

sis with fronds. (From Phillips and Galtier, 2005.)

of regularly spaced, elongate depressions aligned perpendicular to the long axis of the cell (Smoot and Taylor, 1978). Alloiopteris is a morphotaxon used for sterile impression– compression foliage of Carboniferous zygopterids (FIG. 11.65; see also Chapter 16). The numerous species all have pinnae with alternate pinnules that are decurrent at the base and have prominent apical lobes or teeth. Specimens preserved as both compressions and permineralizations have been found in the Caseyville Formation (Lower Pennsylvanian) of southern Illinois and the anatomy suggests affinities with Zygopteris (Jennings, 1975b). The primary pinna vascular strand is butterfly shaped in cross section and pinnule traces are terete. Alloiopteris sternbergii has been found attached to foliar members with Etapteris anatomy. The fact that Etapteris foliar members produced several different foliage types underscores the difficulty in many Paleozoic ferns of characterizing frond morphology based only on petiole anatomy. Corynepteris is a morphotaxon for Zygopteris fertile fronds of the Alloiopteris-type, and is known both as compressions and permineralizations. Foliage of this type has long, slender, penultimate pinnae (FIGS. 11.66, 11.67) that bear alternate fertile pinnae at nearly right angles (Galtier and Scott, 1979). In C. scottii, a structurally preserved example, the vascular tissue of the pinna rachis consists of an oval bar with two protoxylem poles like that of E. scottii (FIG. 11.68) (Galtier and Holmes, 1976). Fertile and sterile pinnules are borne in two rows. Fertile pinnules are typically more dissected, and each bears a sorus attached by a short, vascularized pedicel. Each sorus has 5–10 slightly curved, ovoid sporangia, with a longitudinally oriented U- or V-shaped, multiseriate annulus. All species produce small trilete spores. Because various ornamentation patterns

Figure 11.64 Sieve areas (arrow) on the radial wall of

Etapteris phloem cell (Pennsylvanian). Bar  3 μm. (From Taylor and Taylor, 1993.)

Figure 11.65 Portion of Alloiopteris frond (Pennsylvanian).

Bar  1 cm.

CHapter 11

Ferns and early fernlike plants

415

have been observed, the spores are included in the sporae dispersae taxa Apiculatisporis, Verrucosisporites, or Cyclogranisporites. Another species of Corynepteris, C. australis, was discovered in Permian rocks of Chubut Province, southern

Argentina (Archangelsky and Cúneo, 1986). This represents the first report of the genus from the Southern Hemisphere, and extends the geographic range of the taxon from the Euramerican equatorial region into Gondwana. The stratigraphic range is also extended from the Pennsylvanian into the Permian. Corynepteris cabrierensis is a Mississippian form from southern France that has Alloiopteris-type foliage attached to Corynepteris fertile specimens (Galtier, 2004). This is the oldest Corynepteris species known to date. It has large fronds arising from a robust main rachis that ultimately

Figure 11.66 Portion of a frond of Corynepteris sternbergii (Pennsylvanian). Bar  2 cm.

Figure 11.67 Corynepteris angustissima frond portions (Penn-

sylvanian). Bar  2 cm. (Courtesy BSPG.)

Figure 11.68 Diagrammatic reconstruction of Corynepteris scottii fertile pinna (Pennsylvanian). (From Galiter and Holmes, 1976; in

Taylor and Taylor, 1993.)

416

Paleobotany: the biology and evolution of fossil plants

Figure 11.69 Aggregation of Biscalitheca sporangia (Penn-

sylvanian). Bar  1 mm.

produced paired pinnae. Although sterile pinnae are quadriseriate, the architecture of the fertile fronds is unknown. Biscalitheca is another type of fertile structure (FIG. 11.69) borne by zygopterid ferns. The sporangia of Biscalitheca contain trilete spores, similar to those in Corynepteris. The sporangia, however, are much larger: 3–4 mm long  0.7–1.1 mm in diameter versus only 1–1.5 mm long  0.5–0.8 mm in Corynepteris, and are characteristically banana shaped with a pair of elongate annuli. Biscalitheca musata (Mamay, 1957) includes elongate, slightly curved, annulate sporangia recovered from Upper Pennsylvanian coal balls. Ultimate pinnae bear small (3 mm long), alternately arranged, dissected pinnules. Vascularized aphlebiae arise alternately along the pinnae and bifurcate apically. Multicellular scales cover the pinnae and aphlebiae (Phillips and Andrews, 1968). Soral stalks arise close to the pinna near the leading edge of the pinnule margin and are regarded as an appendage of the pinna (Millay and Rothwell, 1983). Each soral stalk branches to produce laterals that terminate in clusters of up to 60 sporangia. Individual sporangia are banana-shaped and range from 3–4 mm long, each with a pair of longitudinally extended, multiseriate annuli. Patches of small, thick-walled cells occur between the annuli

near the distal end of each sporangium. The trilete spores of B. musata are spherical and range from 40–100 μm in diameter. Many spores contain preserved, endosporal gametophytes with up to 10 cells (Taylor and Millay, 1977). At the ultrastructural level, what are interpreted as cytoplasmic contents within the spores vary from homogeneous to vesiculate. Some contents appear similar to stages in the development of the gametophyte of some extant ferns. Biscalitheca kansana is a compression species from the Lawrence Shale (Upper Pennsylvanian) of Kansas (Cridland, 1966). It consists of a rachial lamina bearing subopposite fertile appendages. Sporangia occur in groups on the abaxial surface, each with paired, multiseriate annuli. Spores are 80 μm in diameter and ornamented by anastomosing ridges on the surface. They appear similar to the sporae dispersae genus Convolutispora. More than 20 sporangia are borne in each pedicellate aggregate in B. suzanneana (Mamay, 1972) from the Upper Pennsylvanian of Texas. Dispersed sporangia of the Biscalitheca type have been assigned based on the size and general histology of the sporangia. Sporangia associated with Etapteris lacattei occur as small tufts of three to eight sporangia, each attached by a small stalk (Galtier and Grambast, 1972); the sporangia are 2.5 mm long and slightly curved. The sporangial wall exhibits two opposite rows of elongate cells that mark the position of an axial annulus. Spores are 80 μm in diameter and trilete. Nemejcopteris feminaeformis is an uppermost Stephanian to Early Permian compressed form of zygopterid foliage known from Europe and China (Barthel, 1968). The plant that bore N. feminaeformis foliage has been reconstructed as a spiny rhizome from which arise phyllophores up to 2 m long; these produce proximally forked, twice-pinnate sterile fronds alternately. Each frond is subtended by a pair of aphlebiae, and the leaves exhibit circinate vernation. The fertile frond of N. feminaeformis has been described as Schizostachys spiciformis (Barthel, 1980b, 2005). It represents a strongly modified frond characterized by a stout, massive rachis and lateral pinnae bearing aphleboid leaves (FIG. 11.70). It remains unknown, however, exactly where these fertile parts occurred on the plant, for example fertile fronds on specialized phyllophores or on the same phyllophores as the sterile fronds. Banana-shaped sporangia ( 3 mm long) occur on the abaxial side attached below the pinna axes, and are characterized by a broad, median annulus; they contain trilete spores. Galtier and Scott (1979) suggested that two distinct groups of zygopterid ferns can be recognized based on the fertile parts. One group (late Namurian–early Stephanian) includes forms like Corynepteris in which the pinnules produced

CHapter 11

Figure 11.70 Nemejcopteris feminaeformis, fertile pinna with stalked sori (Permian). Bar  1 cm. (Courtesy M. Barthel.)

sessile sori with sporangia up to 2.5 mm long; in this form the annulus is V-shaped. In the second group (early Stephanian–Permian), characterized by Biscalitheca, the sori are attached directly to the pinna rachis and sporangia are 3–4 mm long. These sporangia possess two distinct, elongate, multiseriate annuli. Whether these groups represent stages in the evolution of a single zygopterid line or constitute parallel lineages is unknown; however, the stratigraphic ranges of the taxa might suggest the former. Unlike other species of Zygopteris, which have been reconstructed as rhizomatous, Z. primaria is thought to have been treelike, with a stem diameter up to 20 cm (Sahni, 1932a). Known from the Lower Permian of Saxony (Germany), this fossil plant contains primary xylem that is pentagonal in cross section and surrounded by abundant secondary xylem. The bulk of the stem consists of clepsydroid leaf traces and adventitious roots. Pinna traces are of the Etapteris type, and pinnae are covered with multicellular scales. The roots are diarch and also contain secondary xylem. ZYGOPTERID EVOLUTION According to Phillips and Galtier (2005), changes in the anatomy of the phyllophore of zygopterids can be traced from the clepsydroid type to the more elaborate forms present in Zygopteris (Etapteris). Within the Zygopteridales, several taxa possess clepsydroid anatomy and biseriate pinnae, suggesting that they may have evolved from the quadriseriate type. The few forms with quadriseriate fronds and clepsydroid anatomy may underscore their intermediate position. One interesting feature of a large number of zygopterid ferns

Ferns and early fernlike plants

417

is the highly dichotomous aphlebiae which occur along the petioles or arise from the stem. In some taxa, the aphlebiae are three dimensional, whereas in other taxa they are planated. Such structures may represent reduced branching systems that became planated, like the transformational series hypothesized for the evolution of the megaphyll. Laminar pinnules occur by Viséan time, but within the group as a whole they remain relatively small. Our knowledge of zygopterid ferns is largely based on detailed studies that have focused on their complex anatomy and the architectural patterns of sterile and fertile fronds. As a result it is possible to consider the morphological and reproductive aspects of these plants within the context of the ecosystem in which they lived. Phillips and Galtier (2005) suggested that many of the features seen in zygopterids suggest adaptations to dry conditions and high light intensities in coal-swamp habitats. Evidence supporting this interpretation includes the presence of a heavy ramentum that may have minimized moisture loss, and the presence of arrested apices suggesting stages of periodic dormancy, perhaps when there were large blow downs in the swamp forest. The complement of anatomical and morphological features displayed by the zygopterid ferns suggests that they were largely ecological opportunists (Galtier and Phillips, 1996). Phillips and Galtier (2005) also suggested that the ancestral form of the zygopterids was arborescent, based on the fact that the earliest forms (clepsydroid types) are arborescent. These are either unbranched (Asterochlaenopsis) or, as in the Australian zygopterids Symplocopteris and Austroclepsis, with a false trunk morphology. Although the fertile parts of many zygopterids remain to be discovered, those that have been identified typically consist of large sporangia with well-defined annular mechanisms in the form of one V-shaped or two elongate multiseriate bands of thick-walled cells. Sporangia are attached to flattened pinnae characterized by strongly reduced pinnule segments or none at all. From the evidence assembled to date, it appears that the fertile fronds of these plants were biseriate. Available evidence to date suggests that the zygopterids may have had dimorphic fronds, that is the fertile and sterile fronds, or parts of fronds, differed in their morphology. Rothwell (1999) presented a cladistic analysis of ferns, including fossil forms, that resulted in the recognition of three clades: (1) the extinct stauropterids; (2) extinct Pseudosporochnales, Rhacophytales, and Zygopteridales; and (3) the living and fossil members of the eusporangiate and leptosporangiate ferns (FIG. 11.3). Rothwell and Nixon (2006) reanalyzed these data and included molecular data from cladistic analyses of extant ferns (Pryer et al., 2001). They concluded that the relationships among the euphyllophytes are

418

Paleobotany: the biology and evolution of fossil plants

complex and still not completely understood. Inclusion of fossil taxa, not surprisingly, completely changed the proposed phylogenetic trees based only on living ferns. Their analyses do reveal, however, the gaps in our knowledge between early fernlike plants and zygopterids on the one hand, and the higher ferns on the other. It seems likely that ferns in the broadest sense (sensu lato), that is including the cladoxylopsids and zygopterids, are polyphyletic. Although new gene data sets continue to increase the resolution of relationships within extant ferns (Schuettpelz et al., 2006), whether or not the true ferns (including fossil members) are polyphyletic or monophyletic is still unresolved. The phylogenetic relationships of ferns and fernlike plants, both living and fossil, is far from complete and is perhaps one of the most interesting chapters in the evolution of plants. The ferns dealt with in the rest of this chapter are sometimes referred to as true ferns. All four orders (Marattiales, Ophioglossales, Filicales, and Salviniales) contain both fossil and extant representatives. Kenrick and Crane (1997a) suggested that the Marattiales, Ophioglossales, and leptosporangiate ferns represent a monophyletic clade. Others regard the Marattiales and Ophioglossales as monophyletic (Rothwell, 1996a), with leptosporangiate ferns as polyphyletic.

a crown of pinnate fronds. Shoot-borne adventitious roots (Barlow, 1994) emerge some distance below the apex of the stem and extend downward, forming a dense root mantle toward the base of this false stem. The generic name Psaronius was originally used for stems (Cotta, 1832), but today the genus encompasses an entire plant, including petrified stems. Specimens of Psaronius range from the Carboniferous (Pennsylvanian) into the Permian. Sedimentological data and ecosystem analyses have been used to suggest that Psaronius occupied poorly drained areas in seasonally dry landscapes (DiMichele and Phillips, 2002). The size and majesty of some of the plants can be appreciated by examining the exquisitely preserved massive forest of Permian plants, including Psaronius, that are housed in the Museum für Naturkunde in Chemnitz, Germany (Rössler, 2001a, 2003, 2006). Psaroniaceae: Vegetative Features

PSARONIUS PLANT Cross sections of Psaronius stems indicate a complex stelar organization (FIG. 11.71) like that found in modern tree ferns

Marattiales There are various ideas on the taxonomy of the Marattiales (Hill and Camus, 1986). Some delimit two to several families (Pichi-Sermolli, 1977; Stevenson and Loconte, 1996), whereas others include the living genera in a single family, the Marattiaceae. Most of these classifications, however, fail to include the extensive diversification of the group that can be traced back to the Carboniferous. Extant marattialean ferns are exclusively tropical and narrowly confined geographically. Living representatives typically consist of a short, unbranched trunk that produces massive, pinnately compound fronds. Leaves develop circinately and are characterized by a pair of fleshy stipules at the base of each rachis. The eusporangia are grouped into elongate sori or united into synangia on the abaxial surface of pinnules. Roots are polyarch and contain abundant mucilage. The vascular tissues of the stem are arranged as amphicribal bundles that may form a complex dictyostele. All of the genera are homosporous and gametophytes are bisexual. The Marattiales may be traced back to the Carboniferous, where numerous fossil specimens attest to the diversity of the order. One of the most common and best-known fossil marattialeans is Psaronius, a tree fern about 10 m tall with

Figure 11.71 Cross section of Psaronius stem showing

highly dissected dictyostele. Arrows indicate traces to petioles (Pennsylvanian). Bar  2 cm.

CHapter 11

(FIG. 11.72). Near the base of the plant, the stem consists of a small protostele surrounded by numerous adventitious roots forming a massive root mantle (Stidd and Phillips, 1968). At higher levels, the stem increases in diameter and the adventitious root mantle becomes reduced. Thus, the trunk of Psaronius is a false stem consisting of an obconical stele (smaller at the bottom) surrounded by a zone of adventitious roots (Morgan, 1959). The stele is a series of concentric, amphiphloic cauline strands separated by evenly spaced leaf gaps. Each concentric series of strands represents a cycle. Near the top of the plant there are at least 14 cycles with as many as 14 orthostichies of leaves. Leaf traces originate in the innermost cycle of the stem and progress upward and outward, uniting with successive cycles and closing gaps left by other traces as they move toward the base of the petiole. It is hypothesized that Psaronius had an open, unidirectional mode of growth in which the apex continually expanded. Mickle (1984) suggested that the stem and some of the roots eventually decayed in the basal portion of the plant.

Figure 11.72 Suggested reconstruction of Psaronius. (From Morgan, 1959; in Taylor and Taylor, 1993.)

Ferns and early fernlike plants

419

We will use P. blicklei as an example of stem histology (Morgan, 1959). The primary component of the stem is a ground tissue composed of thin-walled parenchyma cells (FIG. 11.73). In a few species cortical cells are loosely arranged with numerous intercellular spaces suggestive of aerenchyma tissue. Some cortical cells contain amorphous substances believed to be tannins. Interspersed in the cortex are elongate, lysigenous lacunae scattered throughout the ground parenchyma and petiole bases. These superficially resemble mucilage canals in living marattialean ferns. At the periphery of the Psaronius stem are thick-walled fibers that form a sclerenchyma sheath providing added structural strength to the stem. Secondary parenchyma cell proliferation occurs in the region of the petiole base. The dictyostelic vascular cylinder of Psaronius (FIG. 11.74) consists of primary xylem tracheids with xylem parenchyma intermixed among the tracheids. Protoxylem tracheids are clustered in groups along the inner surface of each strand (FIG. 7.37). Metaxylem elements have scalariform wall thickenings. Each vascular strand is surrounded by a phloem zone two to three layers deep, which is more extensively developed along the outer surface. The phloem consists of two zones of sieve elements distinguished by the cell types (FIG. 7.37). Sieve areas have been reported in two species (Smoot, 1984a). Stems of P. blicklei have 3 to 14 orthostichies of leaves (FIG. 11.75). As the number of rows of leaves increases there is also a change from a helical to whorled phyllotaxy. Sclerenchyma occurs in patches at the base of the stem, and occurs in tangential bands at higher levels (FIG. 11.76).

Figure 11.73 Cross section of Psaronius blicklei (Pennsylva-

nian). Bar  1 cm.

420

Paleobotany: the biology and evolution of fossil plants

Figure 11.74 Psaronius sp., transverse section. (Permian). Bar 

1 cm. (Courtesy BSPG.)

Figure 11.76 Cross section of Psaronius blicklei at high level in the trunk (Pennsylvanian). Bar  2 cm.

RM RM

Figure 11.75 Cross section near base of Psaronius blicklei

stem where stele consists of two cycles. Note massive adventitious root mantle (RM) (Pennsylvanian). Bar  3 cm.

Numerous lacunae are interspersed between the xylem strands in this species. Specimens of P. chasei (Late Pennsylvanian) are distinguished by the absence of ground tissue lacunae and the presence of bands of sclerenchyma between the second and third stelar cycles. The absence of cortical lacunae is also a feature of P. melanedrus. Stems of P. melanedrus (Middle Pennsylvanian) also have a thick band of fibers near the outer periphery of the cortex (FIG. 11.77). Psaronius magnificus (Late Pennsylvanian) is a permineralized stem thought to be at least 5 m long (Rothwell and Blickle, 1982). In this species the leaves are helically arranged and produced U-shaped leaf traces on the stem surface. Internally, the stem stele

Figure 11.77 Cross section of Psaronius melanedrus stem showing dissected stele and root mantle (RM) (Pennsylvanian). Bar  3 cm.

CHapter 11

Figure 11.78 Psaronius simplex, transverse section through the stem. Note distichous arrangement of bundles (Permian). Bar  2 cm. (Courtesy BSPG.)

consisted of three to six cycles of cauline bundles separated by aerenchymatous ground tissue. All species listed above have polycyclic dictyostelic stems, in which the vascular strands are arranged in concentric cycles, some of which may be distichously arranged (FIG. 11.78). A monocyclic Psaronius stem has also been described from Lower Pennsylvanian deposits in northern Illinois (DiMichele and Phillips, 1977). Psaronius simplicicaulis is 6.5 cm in diameter and contains a single stelar cycle. Maturation of the primary xylem is endarch, and numerous longitudinally oriented secretory ducts occur in the ground tissue. Leaf traces alternate in opposite orthostichies in a distichous pattern. The stratigraphic position of this species may suggest that the larger, polystelic psaronii were derived from smaller, monostelic ancestors through an increase in size and a concomitant increase in anatomical complexity. The earliest Psaronius species suggest that they grew in wet habitats, and from these sites migrated into the parts of peat swamps where there was no standing water. As is the case for some other fossil ferns (e.g., Pseudosporochnales), distinguishing taxonomic features from characters that are developmental has been especially difficult in Psaronius stems. Mickle (1984) suggested that a comparison of stem diameter versus number of leaves, cycles of vascular tissue, and number of vascular segments (meristeles), indicates that stem development may have occurred at different rates in certain species. These results imply that the various species of Psaronius that are based on stelar anatomy alone may only represent different levels within the same stem. One unusual feature of Psaronius is the massive adventitious root mantle (FIG. 11.79), which in some specimens at

Ferns and early fernlike plants

421

Figure 11.79 Detail of aerenchyma tissue of a Psaronius root.

(Pennsylvanian). Bar  1 mm.

C

Figure 11.80 Psaronius sp., adventitious root from the inner

part of the root mantle. Note proliferating epidermis and cortical tissues (C) that interconnect individual roots (Permian). Bar  3 mm. (Courtesy BSPG.)

the base of the trunk may have reached nearly 1.0 m in diameter. Psaronius roots (FIG. 11.80) originate in the stem and, after emerging, bend abruptly downward to parallel the stem surface. The roots have actinosteles with three to nine protoxylem points (Ehret and Phillips, 1977) and a distinctive

422

Paleobotany: the biology and evolution of fossil plants

aerenchymatous cortex, which is, constructed of chains of parenchyma cells with intervening air spaces. This zone is bounded by a massive sclerenchyma sheath several cells thick. The root epidermis and outer cortical layers apparently proliferated to produce a dense mass of parenchyma tissue that held the outer root zone together. The apices of some Psaronius roots have been identified and they include a multilayered root cap and up to four apical initials. The cross-sectional configuration of the leaf trace is variable in Psaronius, depending on the position of the petiole on the stem, variability of the vascular system of the stem, root mantle development, and stem features, possibly reflecting different species. Stipitopteris and Stewartiopteris are names applied to isolated, structurally preserved Psaronius petioles. Stipitopteris gracilis (Morgan and Delevoryas, 1952), a common species from the Pennsylvanian of North America, was several centimeters in diameter. The vascular strand varies from a horseshoe shape (FIG. 11.81) with enrolled ends to an almost complete ring with a W-shaped piece of vascular tissue in its center. Primary xylem maturation is endarch. The vascular tissue is embedded in a parenchymatous ground tissue surrounded by a sclerenchyma sheath with epidermal scales present on the surface of the petiole. Pinnae are borne in two rows along the adaxial surface of the rachis. Traces to the pinnae are C-shaped and leaf gaps are produced during their formation. In Stewartiopteris there is a thin zone of sclerenchyma beneath the epidermis (Morgan and Delevoryas, 1952). The rachis trace is similar to that in Stipitopteris, but lacks in the internal W-shaped segment. Pinna traces are produced from the enrolled edge of the xylem strand. Initially, the trace is cylindrical, but at higher levels it opens to form a C-shaped strand. Stems of P. chasei and P. melanedrus typically possessed petiole trace configurations of the Stewartiopteris type, whereas Stipitopteris is the petiole configuration most commonly found on P. blicklei stems (FIG. 11.73). At least one rachis specimen has been reported that includes both the stewartiopterid and stipitopterid types, depending on the position within the frond (Stidd, 1971). Psaronius croziers (FIG. 11.82) suggest that some species produced tripinnate fronds (FIG. 11.83). The croziers are covered by scales that are each 2 mm long. The most common morphotaxon for Psaronius foliage is Pecopteris (FIG. 11.84), but some marattialeans bore Sphenopteris foliage (Chapter 16). When fronds abscised from Psaronius, they left elliptical scars on the surface of the stem. The trunk surface can be preserved as either a mold–cast or impression– compression and morphotaxa have been delimited on the basis of size, shape, and arrangement of these leaf scars

Figure 11.81 Cross section of Psaronius (Stipitopteris) petiole

(Pennsylvanian). Bar  3 mm.

(FIG. 11.85). Megaphyton is the generic name applied to stem surfaces with leaf scars in two vertical rows on opposite sides of the stem (FIG. 11.86). Scars are variable in shape, ranging from oval to almost rectangular. The configuration of the vascular strand may be of the stewartiopterid or stipitopterid type. Hagiophyton (Corsin, 1948) is similar to Megaphyton, but is distinguished by a thick band of sclerenchyma surrounding the vascular strand (FIG. 11.86). Stem surfaces with petiole scars in more than two vertical rows, indicating a helical or whorled pattern (FIG. 11.86), are assigned to Caulopteris (FIG. 11.87) (Lindley and Hutton, 1831–1833). Leaf scars are variable in shape and may be solitary, overlapping, or covering the entire stem. In Artisophyton

CHapter 11

Ferns and early fernlike plants

423

Figure 11.82 Section of Psaronius frond crozier with attached

pinnae (arrow) (Pennsylvanian). Bar  4 mm. (Courtesy B. Stidd; in Taylor and Taylor, 1993.)

Figure 11.84 Pinna of Pecopteris unita (Pennsylvanian). Figure 11.83 Restoration (left) and diagrammatic reconstruc-

tion of Psaronius frond with Scolecopteris synangia on abaxial surface of pinnules. (From Stidd, 1971; in Taylor and Taylor, 1993.)

Bar  2 cm.

424

Paleobotany: the biology and evolution of fossil plants

there are two vertical rows of scars on opposite sides of the stem, but the vascular strand is closed, often with a deep indentation on the abaxial side (FIG. 11.86). Each half of the outer xylem strand contains a smaller, S-shaped trace (Pfefferkorn, 1976). The discovery of Artisophyton specimens in the Early Pennsylvanian of Nova Scotia confirmed that they were probably small trees a few meters tall (FalconLang, 2005b). Hydathodes are specialized structures on leaves that exude water. They have been observed in silicified, Stephanian and Early Permian marattialean foliage from France and Germany, but are not known from slightly older North American specimens (Lesnikowska and Galtier, 1991). The

presence of such structures may signal slightly different environmental conditions for these plants. This study illustrates how anatomical information can be used to develop hypotheses about habitat conditions.

Figure 11.85 Caulopteris leaf base scars (Pennsylvanian).

Figure 11.87 Detail of Caulopteris leaf base showing shape of vascular trace (arrow) (Pennsylvanian). Bar  1 cm.

Bar  2 cm.

Figure 11.86 Psaronius stem surface petiole base configurations. A. Caulopteris. B. Megaphyton. C. Hagiophyton. D. Artisophyton. (From Pfefferkorn, 1976; in Taylor and Taylor, 1993.)

CHapter 11

Rössler (2000) refered to Psaronius as an ecosystem unto itself, because there is evidence of several other plants, as well as animals, living on and within the root mantle of several Early Permian trunks from Chemnitz, Germany. Similar evidence occurs in Psaronius from North America (Rothwell, 1991). Plants entangled in the root mantle include the scrambling or climbing ferns Ankyropteris, Tubicaulis, Anachoropteris, Grammatopteris, and the seed fern Callistophyton. Also present are cordaitean roots and conifer stems. The presence of coprolites in some tissues suggests that the Psaronius root mantle ecosystem also supported certain herbivores, perhaps oribatid mites. OTHER STEM TAXA Tietea includes structurally preserved marattialean fern stems from the Permian of Brazil and can be distinguished from Psaronius by the arrangement of vascular bundles (meristeles) in the stem (Solms-Laubach, 1913; Derby, 1915; Herbst, 1986, 1992). Tietea singularis stems usually are 20 cm in diameter and bear four orthostichies of leaves in a decussate arrangement. The stem is surrounded by a continuous ring or sheath of sclerenchyma that separates it from the root mantle (Herbst, 1986). The stele consists of amphiphloic meristeles embedded in a homogenous parenchymatous ground tissue. The individual meristeles are round, ovoid, or C-shaped in transverse section; xylem maturation is endarch. In contrast to the regular pattern of meristele arrangement seen in the polycyclic dictyostele of Psaronius, the meristeles in Tietea are arranged irregularly or show a weak cyclic organization in the more peripheral parts of the stem. Although leaf traces in Psaronius are monomeristelic, they are polymeristelic in Tietea, that is composed of several irregularly arranged, rounded-to-elongate meristeles that originate from successive subdivisions of larger meristeles. As a group of meristeles are given off to form a leaf trace, they are enveloped by sclerenchyma. The Tietea root mantle is composed of polyarch roots embedded in a parenchymatous tissue that is produced both by the stem and the roots. Rössler and Galtier (2002a) estimate that T. singularis represents close to 90% of some assemblages in Brazil. Another structurally preserved fern stem described from the Late Permian of Paraguay and Uruguay is Tuvichapteris solmsii (Herbst, 1987). Silicified stems are bounded by a zone of sclerenchyma and embedded in a root mantle; specimens are up to 35  30 cm in diameter. Stems bear six orthostichies of leaves and petioles are also polymeristelic, as in Tietea. Although Tietea shows some organization of meristeles into a cyclic arrangement in the outermost meristeles, Tuvichapteris solmsii shows no such organization.

Ferns and early fernlike plants

425

Tuvichapteris petioles also show a mass of petiole parenchyma closely associated with the vascular strand in the base of the petiole. Psaroniaceae: Reproductive Features

Paleozoic marattialean sporangia are organized into synangia on the lower (abaxial) surface of Pecopteris (FIG. 11.88) and Sphenopteris-type pinnules. The most completely known petrified genera are from North American coal balls and include Scolecopteris, Cyathotrachus, and Eoangiopteris. Scolecopteris (FIG. 11.89), the most widely represented and best understood of these genera, is small, pedicellate and composed of rings

Figure 11.88 Cross section of Pecopteris pinnule with multi-

cellular trichome (arrow) (Pennsylvanian). Bar  1 mm.

Figure 11.89 Paradermal section of Scolecopteris altissimus synangium (Pennsylvanian). Bar  240 μm. (Courtesy M. Millay.)

426

Palestotany: the biology and evolution of fossil plants

Figure 11.90 Scolecopteris elegans, fertile pinnules in cross section (Permian). Bar  2 mm. (Courtesy M. Barthel.)

of laterally appressed sporangia (FIG. 11.90) that separate on dehiscence. The genus has also been reported from the Lower Permian of Germany (Barthel et al., 1995; Weiss, 2002), the Upper Permian of China (Hilton et al., 2004; X.-Y. He et al., 2006), and the Middle Triassic of Antarctica (Delevoryas et al., 1992). In his comprehensive study of Scolecopteris, Millay (1979) recognized 18 species (FIG. 11.91) which can be segregated into three groups based on synangium anatomy and pinnule morphology (FIG. 11.92). The Latifolia group has thin-walled synangia composed of a small number of sporangia which are partially enclosed by the enrolled fibrous margins of the pinnules (FIGS. 11.93, 11.94). The Minor group has pinnules that exhibit thin, lateral extensions of the lamina which also envelope the synangia. The sporangia of this type have thin (FIG. 11.95), outer-facing walls near the base. Members of the Oliveri group have uniformly thick-walled synangia which are borne on unmodified foliage. Although only three basic types of Scolecopteris are described in this volume, there are numerous species that show a continuum of forms. In S. saharaensis (Minor group), the synangium consists of four to five sporangia arranged around a reduced, vascularized pedicel (FIG. 11.96). Sporangial tips are curved toward the center of the unit (FIG. 11.97). Dehiscence is longitudinal and takes place along a line of thin-walled cells directed toward the center of the cluster. The pinnules that produce S. saharaensis have incised margins with teeth alternating with synangia. Spores are oval, monolete, and ornamented with spines. One of the oldest members of the Minor group is S. conicaulis (lower–Middle Pennsylvanian) (Millay, 1982). Based on the large size of the synangium pedicel, lack of zonation in the sporangium wall, and large size of the spores (27–62 μm), this species is regarded as one of the most primitive forms.

Figure 11.91 Section of Scolecopteris parvifolia synangia

(Pennsylvanian). Bar  425 μm. (Courtesy M. Millay.)

Figure 11.92 Cross section of pinnule of Scolecopteris alt-

issimus (Pennsylvanian). Bar  350 μm. (Courtesy M. Millay.)

Scolecopteris illinoensis (FIG. 11.98) is a member of the Oliveri group common in Upper Pennsylvanian coal balls. Four to six sporangia are basally attached to a short pedicel and dehiscence occurs by rupture of one to two rows of

CHapter 11

Ferns and early fernlike plants

427

Figure 11.93 Diagrammatic reconstruction of Scolecopteris

latifolia pinnule with synangia. (From Millay, 1979.)

Paradermal section of Scolecopteris minor var. parvifolia pinnule showing three synangia (Pennsylvanian). Bar  500 μm.

Figure 11.95

Figure 11.94 Paradermal section of several Scolecopteris lati-

folia pinnules showing synangia (Pennsylvanian). Bar  2 mm.

narrow cells along the inner sporangial midline (FIG. 11.99). Spores are monolete. In the Middle Pennsylvanian species S. iowensis (FIG. 11.100), spores are large ( 83 μm) and trilete (FIG. 11.101). Scolecopteris guizhouensis (Permian of

China) is similar to species in the Oliveri group in the possession of thick, outer-facing sporangial walls and large trilete spores (X.-Y. He et al., 2006). The ultrastructure of Scolecopteris spores indicates that all possess a three-layered sporoderm, with the ornamentation occurring on the sculptine layer (Millay and Taylor, 1984). In extant marattialean spores, the ornament occurs on the exine. Miospores of the sporae dispersae genus Thymospora have been recovered from sporangia of S. parkerensis. Two types of spores occur in the same sporangium of S. dispora: Torispora securis and Laevigatosporites globosus (Lesnikowska and Willard, 1997).

428

Paleobotany: the biology and evolution of fossil plants

11.99 Section of three Scolecopteris illinoensis synangia attached to abaxial pinnule surface (Pennsylvanian). Bar  375 μm. (Courtesy M. Millay.)

Figure Figure 11.96 Cross section of Scolecopteris saharaensis synangium (Pennsylvanian). Bar  300 μm. (Courtesy M. Millay.)

M

Figure 11.97 Cross section of Scolecopteris saharaensis pinnule midrib (M) with pinnule lamina (arrow) and synangium (Pennsylvanian). Bar  335 μm. (Courtesy M. Millay.)

11.98 Paradermal section of Scolecopteris illinoensis synangia. Compare with FIG. 11.99 (Pennsylvanian). Bar  325 μm. (Courtesy M. Millay.)

Figure

Figure 11.100 Diagrammatic reconstruction of Scolecopteris

iowensis pinnatifid pinna. (From Millay, 1979; in Taylor and Taylor, 1993.)

In S. calicifolia (FIG. 11.102) (Latifolia group) the synangia are borne in two rows and constructed of three to four exannulate sporangia each. This Middle Pennsylvanian species is characterized by a short synangial pedicel. Large pinnules, some 6.5 mm long, characterize S. incisifolia. The spores of S. mamayi are ornamented with delicate spines. Ironstone nodules from southern Illinois have provided both fertile and sterile foliage of S. macrospora (Minor group). The specimens are partially petrified and include pinnules with dissected margins. Synangia occur in a single row on each side of the midrib and are identical with other species of the group. Although the pinnules could not be related to any known species of Pecopteris,

CHapter 11

Ferns and early fernlike plants

429

Figure 11.103 Section of pinna axis of Acaulangium bul-

baceus (Pennsylvanian). Bar  1 mm. (Courtesy M. Millay.)

Figure 11.101 Two sporangia of Scolecopteris iowensis contain-

ing spores (Pennsylvanian). Bar  120 μm. (Courtesy M. Millay.)

Figure 11.104 Paradermal section of Acaulangium bulbaceus

synangium (Pennsylvanian). Bar  300 μm. (Courtesy M. Millay.)

Figure 11.102 Diagrammatic reconstruction of Scolecopteris calicifolia showing pinnule lobes and multicellular hairs. (From Millay, 1979; in Taylor and Taylor, 1993.)

the material is important in providing a method whereby compression taxa can be correlated with species that have been erected based on structurally preserved specimens. Acaulangium is a Late Pennsylvanian synangium that was

referred to the Marattiales (FIG. 11.103) (Millay, 1977). It has oval, sessile synangia, with the number of sporangia ranging from four to seven (FIG. 11.104). Stubblefield (1984), however, suggested that the genus is synonymous with Scolecopteris. Eoangiopteris (Mamay, 1950) is a Pennsylvanian form with distinctly bilateral synangia. This feature is more like the synangia of extant marattialeans than the radial organization of Paleozoic forms. Eoangiopteris goodii and E. andrewsii (Upper Pennsylvanian) exhibit a linear synangium of 10–19 sporangia (FIGS. 11.105, 11.106) (Millay, 1978). Sporangia are partially embedded in an elongate pad on the abaxial surface of pecopterid pinnules (FIG. 11.107). If found dispersed, the trilete, oval spores of E. goodii would be included in the

430

Paleobotany: the biology and evolution of fossil plants

Figure 11.105 Paradermal section of Eoangiopteris andrewsii

showing synangia M. Millay.)

(Pennsylvanian).

Bar  1 mm.

(Courtesy

genus Verrucosisporites. Eoangiopteris exhibits several features in common with the extant marattialean Angiopteris, which has laterally free sporangia arising from a common pad of tissue. Both forms are also similar in spore morphology and mode of sporangial dehiscence. Features of Grandeuryella include examulate sporangia that produce monolete spores borne on a heavily vascularized receptacle along lateral veins of highly recurved pecopterid pinnules (FIG. 11.108) (Lesnikowska and Galtier, 1992). Burnitheca pusilla (FIG. 11.109) is a synangium that shares features with both seed ferns and marattialean ferns (Meyer-Berthaud and Galtier, 1986a). The permineralized specimens are 2.6 mm long and produced trilete spores with what is interpreted as a perispore. If Burnitheca represents a marattialean synangium, it extends the geological record of the order back to the Tournaisian.

Figure 11.106 Paradermal section of two Eoangiopteris goodii pinnules (Pennsylvanian). Bar  500 μm. (Courtesy M. Millay.) Figure 11.108 Grandeuryella, paradermal section of fertile

pinnules showing sori (Pennsylvanian). Bar  1 mm. (Courtesy J. Galtier.)

Figure 11.107 Cross section of Eoangiopteris andrewsii pinnule. Arrow indicates midrib (Pennsylvanian). Bar  500 μm. (Courtesy M. Millay.)

Figure 11.109 Burnitheca synangium, transverse section (Mississippian). Bar  180 μm. (Courtesy J. Galtier.)

CHapter 11

11.110 Two fertile (Pennsylvanian). Bar  2 cm.

Figure

pinnae

of

Asterotheca

Ferns and early fernlike plants

431

sp.

Paleozoic Compression Taxa

Pecopteris unita is a distinctive species from the Francis Creek Shale (Middle Pennsylvanian) of Illinois (FIG. 11.84). Fertile pinnules have synangia in two rows, one on each side of the pinnule midvein. Each synangium has five to seven radially arranged, laterally appressed sporangia that are free at their tips on dehiscence. Spores are monolete and range from 14–20 μm in diameter. Synangia exhibiting a different morphology, as well as spores of a different type, have also been identified on P. unita foliage. The presence of these different fertile structures underscores the fact that many compressed foliage morphospecies actually represent different biological species. The genera Asterotheca and Ptychocarpus, among others, are used for compressed pecopterid pinnules (FIGS. 11.110–11.112) that bore one to three rows of synangia on the abaxial surface. Synangia are bilaterally symmetrical, elliptical in outline, and grouped in one or two rows on each side of the midrib. Acitheca is an arborescent marattialean fern believed to have been 3.5 m tall. The genus is worldwide in distribution, extending from the Middle Pennsylvanian into the Permian. Fronds were tri- to quadripinnate and left scars of the Caulopteris type on the stems (Zodrow et al., 2006). Fertile fronds are characterized by three to five (typically four) elongate sporangia (FIG. 11.113) partially covered by the edges of the lamina. In A. adaensis from the Middle Pennsylvanian of Iowa, 12–20 synangia are arranged in two rows on either side of pecopterid pinnules (Mapes and Schabilion, 1979). Permineralized specimens of Acitheca (A. polymorpha) are also known from the Stephanian of France (Lesnikowska and Galtier, 1991). Other forms (e.g.,

Figure 11.111 Mazon Creek nodule containing Asterothecatype pinna (Pennsylvanian). Bar  2 cm.

Remia pinnatifida; FIGS. 11.114, 11.115) are included in the Marattitales based on sporangial organization, spores, and frond morphology. Dizeugotheca is a fertile pecopterid borne on bi- or tripinnate fronds from the Permian of Patagonia, Argentina (Archangelsky and de la Sota, 1960). Synangia are elongate and arranged in groups of four, with two of the sporangia overlapping the ones below. In D. waltonii the pinnules are alternate and subopposite. Elongate synangia are borne at the distal ends of lateral veins in Gemellitheca and form an extended row over most of the pinnule surface (Wagner et al., 1985). Spores are alete to trilete and ornamented by minute spines. Marattialean ferns, including Scolecopteris, were also a conspicuous component of many Permian floras. Qasimia is a Late

432

Paleobotany: the biology and evolution of fossil plants

11.112 Asterotheca arborescens, fertile pinna of the penultimate order of a pecopterid frond (Pennsylvanian). Bar  3 cm. (Courtesy BSPG.)

Figure

Figure 11.114 Remia pinnatifida, frond segment with sterile and fertile pinnae (Permian). Bar  1 cm. (Courtesy M. Barthel.)

11.113 Two sporangia of Acitheca polymorpha (Pennsylvanian). Bar  650 μm. (Courtesy M. Millay.)

Figure

Figure 11.115 Remia pinnatifida, fertile pinna segments with synangia (Permian). Bar  5 mm. (Courtesy M. Barthel.)

CHapter 11

Permian form from Saudi Arabia consisting of partially permineralized synangia borne on taeniopterid pinnules (C. Hill et al., 1985). The large bipinnate fronds produced pinnules ranging from 11–110 mm long. Each synangium is bivalved and morphologically resembles the synangium of several Mesozoic and extant species of Marattia. The taeniopterid pinnules of Qasimia are in marked contrast to the pecopterid forms that dominate the majority of late Paleozoic marattialean ferns. It has been suggested that the pinnules of Qasimia may have evolved from the reduction of larger pinnae (C. Hill et al., 1985). Morphologically, Qasimia compares closely to the Cathaysian fertile fern Taeniopteris tabaensis (X. Li et al., 1982). In addition to the numerous synangial types borne on Pecopteris foliage, there are other marattialean foliage forms that were present during the Carboniferous. Radstockia kidstonii (Taylor, 1967b) is a Pennsylvanian compression fossil from the famous Mazon Creek locality in Illinois (FIG. 11.116). The largest specimen is an antepenultimate axis 9.5 cm long that bears 15 alternately arranged pinnatifid pinnules of the sphenopterid type. Pinnules are lanceolate and constricted at the point of attachment; the margins are deeply lobed and free from adjoining foliar members. Synangia are partially embedded in the abaxial surface of the foliar lobes. The synangium has a segmented appearance formed by a median furrow and

Ferns and early fernlike plants

433

secondary furrows that arise at right angles. The smaller subdivisions in R. kidstonii may correspond to individual sporangia. Although histologic features remain unknown, this fern contains smooth, trilete spores in the 40–60-μm-size range. Morphologically, the pinnules and synangia appear almost identical to those produced by the extant fern Marattia alata, although Brousmiche (1983) suggested that Radstockia has closer affinities with Crossotheca, the pollen organ of a lyginopterid seed fern (Chapter 14). Sydneia manleyi from the Westphalian D of Nova Scotia is similar to Radstockia, with the sporangia attached to the distal region of pinnule lobes (Pšenicˇka et al., 2003). Sporangia are reported as containing both trilete and monolete spores. Mesozoic Marattialeans

Several geologically younger foliage types with sporangia have been referred to the Marattiales on the basis of the apparent synangiate nature of the sporangia and, to a lesser degree, on pinnule morphology. One of these is Marattia, a name used by Harris (1961a) for specimens collected from the famous Yorkshire Jurassic localities. Marattia anglica consists of fertile and sterile pinnae up to 50 cm long. Synangia range from 4–7 mm long and are composed of 30 pairs of sporangia. Van Konijnenburg-Van Cittert (1975) (FIG. 11.117) indicated that, although the majority of the spores are monolete, at least 30% display a trilete suture. Marattia intermedia, from the Rhaeto–Liassic of northern Iran, includes entire-margined pinnules up to 50 cm long. Synangia are

Figure 11.116 Portion of a Radstockia kidstonii pinna showing

synangia (Pennsylvanian). Bar  4 mm. (From Taylor, 1976b; in Taylor and Taylor, 1993.)

Figure 11.117 Johanna H. A. Van Konijnenburg-Van Cittert.

434

Paleobotany: the biology and evolution of fossil plants

elongate and spores are monolete. Early Jurassic specimens from China are named M. asiatica. They include pinnae up to 15 cm long with an auriculate lobe at the base (Y. Wang, 1999). Fertile pinnae have linear synangia, each with 25–40 pairs of sporangia. Spores are monolete and 25 μm in diameter. Marattiopsis crenulatus (Lundblad, 1950b) (FIG. 11.118) is a Triassic species from Sweden that has some similarity to M. asiatica. The specimens are interesting in that a double set of veins is present, similar to the false veins that have been figured in a number of extant marattialean taxa. Each synangium contains 20 sporangia; spores are oval to elongate and monolete. Other Triassic marattitalean foliage types include forms such as Pseudodanaeopsis (FIG. 11.119).

Another marattialean from the Jurassic of Yorkshire is Angiopteris (C. Hill, 1987). The frond of A. blackii is thought to have been bipinnate, with each synangium composed of 6–12 sporangia. Approximately 3000 spores are produced per sporangium; each spore is trilete and up to 30 μm in diameter. Charcoalified synangia from the Cretaceous Dakota Formation are interpreted as members of the Marattiales based on spore ultrastructure and sporangial organization (S. Hu et al., 2006). The generic names Goolangia and Mesozoisynangia are proposed for these marattialeans. Fossil fertile pinnule segments similar to those of extant species of Danaea are called Danaeopsis. The Triassic species D. fecunda consists of pinnate fronds with lanceolate pinnules (Herbst, 1977). On the abaxial surface are numerous synangia, each consisting of a double row of ellipsoidal sporangia. At the apex of each sporangium is a small depression that probably represents the stomium. Spores are trilete and up to 70 μm in diameter. Danaea coloradensis is an Eocene impression species from the Green River Formation (Knowlton, 1922). Fertile pinnules contain numerous closely packed synangia partially embedded in the abaxial surface. Marattialean Evolution

Figure 11.118

Britta Lundblad.

Figure 11.119 Pseudodanaeopsis plana (Triassic). Bar  5 cm.

(Courtesy C. Pott.)

Several scholars have suggested that the bilateral synangium of extant marattialeans evolved by the lateral fusion of radial Paleozoic types, although transitional stages are not known in the Carboniferous. Unmodified pinnules and large spores, together with the absence of sporangial wall differentiation, have been regarded as plesiomorphic characters among the Paleozoic marattialean fructifications. Using these features, Millay (1978) has proposed two basic lines of Paleozoic marattialean synangial evolution. One, characterized by selected species of Scolecopteris, may be related to the extant synangiate genus Marattia (FIG. 11.120). The second

Figure 11.120 Section of Marattia pinnule showing synangium and sporangia with spores (Extant). Bar  325 μm. (Courtesy M. Millay.)

CHapter 11

pattern, which is exemplified by the synangial organization of Eoangiopteris, is more like that of extant Angiopteris, in which sporangia are free, but held together by a common pad of tissue. One pronounced difference between most of the Paleozoic marattialeans and the post-Paleozoic fossils, including the modern genera, is the organization of the foliage. The Mesozoic members possess foliage that is once pinnate and modern in appearance, whereas the Paleozoic foliage is at least four times pinnate and includes small pinnules. One exception is the larger taeniopterid foliage of Qasimia. If these trends have validity, the significance of the abrupt change in foliar morphology remains perplexing. It is interesting, however, that when the living marattialean Angiopteris lygodiifolia is grown at cooler temperatures, the foliage becomes less complex (Asama, 1960). The changing climate during the late Paleozoic may be a plausible explanation for what appear to be distinct differences in marattialean leaf morphology. DiMichele and Phillips (2002) suggested that the earliest monocyclic marattialeans grew near streams in wet habitats, and as speciation continued, habitats became more variable including non-swamp, floodplain environments. By the Middle Pennsylvanian, they suggested that the marattialeans were exploiting disturbed areas. It has also been suggested that in the late Paleozoic and early Mesozoic, after the vast wetlands along the continental margins had disappeared and, along with them, most of the arborescent lycopsids, marattialean ferns became the dominant elements of the flora that was now locally restricted to epicontinental coal-swamp vegetation (Kerp, 2000). Marattialeans continued occupying wetland habitats in the Permian in what has been interpreted as seasonally dry conditions. It will be interesting to see what relationships exist between the well-documented Permo-Carboniferous marattialeans, and those that appear sporadically in the Mesozoic, as we learn more about the younger fossil representatives of this order.

Ophioglossales The Ophioglossales are represented by three or four living genera (Ophioglossum, Botrychium, Helminthostachys, and Mankyua) and 70 species of temperate to boreal distribution, all placed in a single family, the Ophioglossaceae. The fossil record suggests that the group has remained small throughout its geologic history (Rothwell, 1996b). The living forms are relatively small and consist of a fleshy stem containing an ectophloic siphonostele embedded in

Ferns and early fernlike plants

435

parenchymatous ground tissue. Prominent leaf gaps are produced during leaf trace emission. Xylem maturation is mesarch or endarch, with secondary xylem and cortical periderm produced only in Botrychium. Tracheid pitting ranges from scalariform to circular bordered. All species lack sclerenchymatous tissues. Each plant consists of a single vegetative leaf and a fertile spike (dimorphic fronds). In Botrychium and Helminthostachys, the aerial portions are pinnate; in Ophioglossum, the leaf ranges from simple to dichotomously branched in some epiphytic species. The fertile spike is pinnate in Botrychium, dichotomously branched in Ophioglossum, and constructed of reduced branches radially arranged around the spike axis in Helminthostachys. In Mankyua the fertile spike, sometimes termed the sporophore, is borne from the adaxial side of the vegetative leaf (B.-Y. Sun et al., 2001). The eusporangia are massive and produce trilete, spherical–subtriangular spores that are morphologically similar in all species. All species are homosporous. The fossil record of the group is poor and has been historically restricted to a few reports of spores described from the Jurassic and Cretaceous of the former Soviet Union. Wellpreserved fertile and sterile segments of Botrychium, however, are known from the Paleocene of Canada (Rothwell and Stockey, 1989). The sterile segments are tripinnately compound and opposite to suboppositely branched; they appear most similar to the extant species B. virginianum. Fertile segments are also tripinnately compound, with both stalked and sessile sporangia. Sporangia range up to 2 mm in diameter and contain several hundred radial, trilete spores up to 67 μm in diameter. Although the modern appearance of these fossils does little to clarify the evolutionary history of the Ophioglossales, the discovery is significant in demonstrating that modern representatives were in existence by the Paleogene. Because of a poorly known fossil record, members of the Ophioglossales have been variously grouped with lycopsids, rhyniophytes, coenopterid and filicalean ferns, and progymnosperms (Kato et al., 1988). Stevenson and Loconte (1996), using anatomical and morphological characters, suggested that Ophioglossum and Botrychium are closely related as both have transverse sporangial dehiscence, and that Helminthostachys is more primitive because sporangial dehiscence is longitudinal. Mankyua has a mixture of characters found in the other genera, and may represent, at least morphologically, a member of the ancestral lineage of the group. Rothwell (1999) viewed the Ophioglossales as sister to the Marattiales and leptosporangiates, whereas A. Smith et al. (2006), combining analyses based on both molecular and morphological characters of extant forms only, placed the Ophioglossales together with the Psilotaceae in a group

436

Paleobotany: the biology and evolution of fossil plants

they term the Psilotopsida. Hauk et al. (2003), using rbcL sequences, supported monophyly of the Ophioglossaceae, but define two main clades within it. An interesting study of living Ophioglossales addresses how symbiotic interactions may affect molecular-based phylogenies (Davis et al., 2005). These authors noted that some gene regions indicated Botrychium virginianum as a fern, whereas others placed the species phylogenetically among certain parasitic angiosperms. These authors hypothesized that this discordance may be the result of horizontal gene transfer from a root parasite. One can only speculate as to whether horizontal gene transfer represents a significant evolutionary force in the fossil history of plants.

several families of extinct ferns. The reader is referred to the systems suggested by Bower (1923–1928), Copeland (1947), Holttum (1949), Pichi-Sermolli (1977), and Tryon and Tryon (1982) for additional details on the taxonomy and phylogenetic positions of extant families from a morphological viewpoint. The use of molecular markers has been employed to study both the narrow and broader phylogenetic relationships among ferns (Pryer et al., 2004; A. Smith et al., 2006); however, the lack of fossils in many of these large-scale phylogenies renders understanding the biodiversity and evolution of the group equivocal (Rothwell and Nixon, 2006). Not all recognized groups of ferns are included here, but the families discussed below demonstrate the large diversity of ferns in the fossil record.

Leptosporangiate ferns

Osmundales

The leptosporangiate ferns are represented today by 300 genera and some 11,000 species. They are widely distributed in the tropics, where they are common as both epiphytes and tree ferns, but many species occur in temperate regions as well. Although some ferns exhibit dichotomously branched fronds and some have simple leaves, the majority have pinnatifid or pinnate leaves, including bipinnate, tripinnate and so on. Frond size ranges from massive in some tree ferns to extremely small in aquatic ferns. Generally, the leaf is dorsiventral and epidermal cells contain chloroplasts. The arrangement of vascular tissue in the stems ranges from simple protosteles to complicated dictyosteles. In the higher ferns, the dictyostele is made up of segments, or meristeles, each of which is usually surrounded by a pericycle and an endodermis (Chapter 7). Filicalean leptosporangia are usually arranged in clusters termed sori that may be marginal or attached to the abaxial surface of the leaf. The position of the annulus has been used as an important character in delimiting families of filicalean ferns. Spore output is typically small, with both monolete and trilete spores produced in the group. Most ferns are homosporous, but the aquatic forms, that is, those in the Salviniales are heterosporous. It is important to note that the classification of filicalean ferns is not agreed upon by all pteridologists and vascular plant morphologists. Opinions differ relative to the importance of certain features that have been used as the basis of systematic comparisons. These features include the developmental sequence of the sporangium (FIG. 11.113), position and type of annulus, presence or absence of an indusium, and base chromosome number (Goldblatt, 1981). The classification scheme used here combines features of several previously recognized systems, but differs in that we also include

The Osmundales have historically been recognized as a primitive group intermediate between eusporangiate and leptosporangiate ferns, whether they are treated at the ordinal or familial level; they were once considered a separate subphylum, the Protoleptosporangiatae. Two families are currently recognized, Guaireaceae and Osmundaceae, and the order has a long evolutionary history. Based on phylogenetic analyses using rbcL sequences, the Osmundaceae are hypothesized to share a relationship with all leptosporangiate ferns. Using both molecular and morphological data sets, Osmundaceae and the remaining leptosporangiate ferns are regarded as monophyletic. The group has been suggested to be a possible ancestral stock for other fern families, although some believe that they represent an early offshoot from filicalean ancestors that had little or nothing to do with the evolution of other families (Miller, 1971) (FIG. 11.121). The family can be traced back to the Late Permian, where a number of structurally preserved stem morphogenera with osmundaceous features occur. Today the family is represented by three genera (Osmunda, Todea, Leptopteris) comprising 16 species, but it is estimated that there are 150 fossil species (Tidwell and Ash, 1994). A report by Metzgar et al. (2008) included a fourth genus Osmundastrum, in the family based on DNA sequence data, a suggestion proposed earlier by Miller (1967, 1971) based on the anatomy of fossil and extant specimens. The plants vary from arborescent forms to rhizomatous stems surrounded by tightly packed petiole bases; fronds have stipule-like wings at the base. In some species, dimorphic fronds are produced with sporangia attached along narrow, reduced foliar segments; in others, sporangia occur on the abaxial surfaces of fertile laminae, attached by short stalks. Sporangia are not grouped into sori and no indusium is present. Sporangial walls are thick and the

CHapter 11

Ferns and early fernlike plants

437

in the Osmundaceae be more accurately termed dictyoxylic (Rothwell et al., 2002). Secondary xylem is not produced, but sclerenchyma is common in the ground tissue. Most information about the fossil history of the order comes from structurally preserved rhizomatous stem segments. In some cases, foliage is known from the same stratigraphic level as the stem, but organic attachment has not been demonstrated. The diversity of stem morphotaxa is highest in the Permian, where both protostelic and dictyostelic forms are found, whereas the Mesozoic is dominated by three taxa that are relatively widespread geographically. The genus Osmunda is known from the Late Triassic and fossils assignable to O. cinnamomea, a modern species, occur in the Late Cretaceous. Isolated osmundaceous sporangia and spores are common in Mesozoic rocks, and are typically identical to extant forms.

Figure 11.121

Charles N. Miller. (Courtesy L. Matten.)

Figure 11.122 Cross section of Osmunda sp. rhizome showing

dissected xylem of stele and C-shaped leaf traces (arrows) (Extant). Bar  1 mm.

annulus may be rudimentary. The thickness of the sporangial stalk and the high spore output are characteristics more typical of eusporangiates. The stems exhibit a variety of stelar configurations, ranging from protosteles to highly dissected ectophloic or amphiphloic dictyosteles (FIG. 11.122). Since leaf trace departure results in gaps in the xylem, but not in the phloem, some suggest that the steles of the type found

PALEOZOIC STEM TAXA The earliest, structurally preserved osmundaceous stem genera are included in the subfamily Thamnopteroideae of the Osmundaceae. The oldest fossils are from the Late Permian and have either protosteles or ectophloic siphonosteles without leaf gaps (Miller, 1971). Thamnopteris, from the Upper Permian of Russia, includes arborescent stems that consist of a mantle of adventitious roots surrounding a protostele composed of a central and an outer xylem zone. Parenchyma cells are associated with the inner xylem tracheids in several species. It has been suggested that the stele of the Osmundaceae evolved from the conversion of tracheids to parenchyma in a partially protostelic axis like that in Thamnopteris. Zalesskya is another Late Permian stem with xylem organized into distinct zones and a cortex with numerous (100–150) leaf traces. Chasmatopteris principalis (Late Permian) may represent another exemplar in stelar transformation. In this form there are rudimentary invaginations of the stele, but no leaf gaps are produced. If this pattern were to be continued, with further indentation of the xylem cylinder above the point of leaftrace departure, it would result in a dictyoxylic siphonostele, the stele type present in geologically younger osmundaceous taxa. Such a sequence has been termed the intrastelar origin of a siphonostele (FIG. 7.40) and, within the Osmundaceae, is supported by the shortening of central xylem tracheids in the Permian thamnopterids. Thus, there is excellent fossil evidence, as detailed by Kidston and Gwynne-Vaughan (1907–1914) (FIG. 11.123) in a classic series of papers, to hypothesize that the osmundaceous siphonostele evolved through the phylogenetic conversion of central xylem tracheids to pith parenchyma, and the subsequent dissection of

438

Paleobotany: the biology and evolution of fossil plants

Figure 11.124 Cross section of Guairea carnieri stem with

leaf traces (Permian). Bar  4.5 cm. (From Tidwell, 2002.) Figure 11.123 Robert Kidston (left) and David T. GwynneVaughan. (Courtesy A. C. Scott.)

the xylem cylinder by leaf gaps. This series differs from the extrastelar origin, in which there is an intrusion of cortical parenchyma into the stele. An alternative interpretation was suggested by R. Hill et al. (1989) based on Late Triassic specimens of Australosmunda (now Millerocaulis). In M. indentata the ectophloic siphonostele has indentations where traces diverge, but no gaps are formed. Although the anatomy of living ferns has been relatively easy to characterize, the discovery of what is interpreted as a sympodial system in an Eocene fern may challenge this long-held assumption (Karafit et al., 2006). One of the problems with such transformational theories, however, is that there are other taxa with more derived stelar anatomy that existed at the same time (see Mesozoic stems discussed below). Well-preserved stem remains of osmundaceous ferns are known from the Upper Permian Coal Measures of Queensland as Palaeosmunda (Gould, 1970). This taxon has also been interpreted as representing an intermediate stage in the evolution of the siphonostele. Palaeosmunda differs from other Permian stem morphogenera (e.g., Bathypteris, Chasmatopteris, Iegosigopteris, Petcheropteris, Thamnopteris, and Zalesskya) in the possession of an ectophloic, dictyoxylic siphonostele with uniform metaxylem tracheids and a true parenchymatous pith. Leaf traces are endarch, and leaf bases are rhomboidal in transverse section with an adaxially curved, C-shaped strand. In P. williamsii, the cortex is differentiated into two zones and contains as many as 43 traces in a single transverse section. Preservation is so exceptional that in P. playfordii a zone of metaphloem elements is identified surrounding each xylem strand.

GUAIREACEAE Guairea is a silicified stem with osmundaceous anatomy from the Upper Permian of Paraguay (Herbst, 1981), which is placed in a separate family, the Guaireaceae. It is unclear whether some of these plants lacked a large mantle of petiole bases and adventitious roots or whether this feature is due to preservation (Herbst, 1981), as some exhibit a mantle of roots, but no persistent petiole bases. Guairea is characterized by a cortex that is not divided into two zones and petiole bases that lack sclerenchyma around the traces (FIG. 11.124). Miller (1971) noted that Itopsidema was unlike other members of the Osmundaceae, due to a mixture of parenchyma and tracheids in the xylem, different petiole trace configurations near the base, and a homogenous parenchymatous cortex lacking an outer sclerotic zone. This genus is also now placed in the Guaireaceae, along with Donwelliacaulis (Tidwell and Ash, 1994). MESOZOIC AND CENOZOIC STEM TAXA Stem morphotaxa assigned to the Osmundaceae are much more widespread in the Mesozoic and, like modern members of the family, are assigned to the subfamily Osmundoideae. Most cannot be placed in modern genera and thus have been classed in a number of morphogenera. Miller (1971) reviewed these forms and placed them in a new genus, Osmundacaulis, divided into three groups. Members of the O. braziliensis group have subsequently been placed in the genus Guairea, in a new family, Guaireaceae (Herbst, 1981). Species in the O. herbstii group were split into two genera, Millerocaulis (Tidwell, 1986) and Ashicaulis (Tidwell, 1994). The species in Miller’s original O. skidegatensis group remain in Osmundacaulis. This genus is characterized by arborescent forms with a wide xylem cylinder (25 tracheids

CHapter 11

Ferns and early fernlike plants

439

Figure 11.126 Cross section of Ashicaulis woolfei (Triassic).

Bar  2 mm. Figure 11.125 Ashicaulis sp. (Cretaceous). Bar  4.5 cm. (Cour-

tesy D. Cantrill.)

wide) dissected into a large number of strands; leaf traces are adaxially curved and depart from the stele with two to four protoxylem stands. Both Millerocaulis and Ashicaulis are generally rhizomatous to slightly upright forms with small stems and are known from the Triassic into the Lower Cretaceous. Both taxa have a wide, sclerotic outer cortex, and leaf traces exhibit only one protoxylem strand at departure from the stele. In Millerocaulis, leaf gaps are either lacking or occur sporadically, whereas Ashicaulis stems show numerous, definite leaf gaps (Tidwell and Ash, 1994). Ashicaulis (FIG. 11.125) currently contains 24 species (Matsumoto et al., 2006) and Millerocaulis more than 10 (Herbst, 2006; Cheng and Li, 2007); both genera are known primarily from the Southern Hemisphere. Ashicaulis exhibits an ectophloic dictyostele, whereas Millerocaulis is characterized by an ectophloic siphonostele. Other species of Millerocaulis come from the Upper Triassic of Argentina (M. stipabonettii; Herbst, 1995) and the Jurassic of New Zealand (M. dunlopii; Kidston and GwynneVaughan, 1907–1914). Ashicaulis beardmorensis (originally Osmundacaulis) is known from Middle Triassic rocks of the central Transantarctic Mountains (Schopf, 1978a). The xylem cylinder is 1.7 mm in diameter and surrounds a pith containing isolated tracheids, parenchyma, and secretory cells. Near the periphery of the cortex, leaf traces are

W-shaped in cross section and contain up to nine endarch protoxylem points. Permineralized specimens from the Middle Triassic of Antarctica are assigned to Ashicaulis woolfei (Rothwell et al., 2002). Stems are 2.5 cm in diameter and fronds are pinnate with alternate–subopposite, pinnatifid pinnae. The stele consists of a ring of eight to nine xylem segments and diverging leaf traces (FIG. 11.126). Late Cretaceous specimens of A. livingstonensis, also from Antarctica (Livingston Island) are interpreted as erect, mound-forming ferns (Cantrill, 1997a). A specimen of Aptian age, also from Livingston Island, is assigned to A. australis (Vera, 2007). Some modern members of the Osmundaceae produce croziers late in the season which abort, and the petiole bases of these aborted leaves (sometimes called cataphylls) may form a protective covering over the shoot apex. Similar structures occur in many fossil species; the oldest example is A. herbstii, a silicified specimen from the Upper Triassic of Argentina (Archangelsky and de la Sota, 1963; Miller, 1971). This species, which was found in association with Cladophlebis foliage, consists of a small stem (9 mm in diameter) surrounded by a zone of persistent petiole bases and roots. The stele is dictyostelic and ectophloic. Foliage impressions of Hausmannia, Coniopteris, and Cladophlebis were found in association with A. hebeiensis (originally Osmundaculis) from the Middle Jurassic of China (Z.-Q. Wang, 1983). This fern has an ectophloic dictyostele

440

Paleobotany: the biology and evolution of fossil plants

S

Figure 11.127 Cross section of Osmundacaulis jonesii show-

ing stem (S) surrounded by C-shaped petioles and adventitious roots (Cretaceous). Bar  1.8 cm.

organized into 9–14 vascular strands. Anatomically, the Chinese specimen shares the greatest number of features with extant Osmunda subgenus Plenasium. In Osmundacaulis skidegatensis from the Lower Cretaceous of British Columbia, Canada, the central portion of the axis contains a pith consisting of parenchyma and stone cells (Miller, 1971). The dictyostele is amphiphloic, and leaf traces include two endarch strands. Cataphylls are absent. In O. jonesii (FIG. 11.127) (Late Jurassic or Early Cretaceous of Australia) the axis is 30 cm in diameter (Tidwell, 1987), and the stele is ectophloic, with sclerenchyma absent in the outer pith and leaf gaps. Perhaps the oldest stem remains assignable to a modern genus are Osmunda (now Osmundastrum) cinnamomea from the Maastrichtian (Upper Cretaceous) of Drumheller, Alberta (Serbet and Rothwell, 1999). These small, permineralized stems (4–5 mm in diameter) show an ectophloic dictyostele, two-zoned cortex, and leaf trace anatomy typical of the modern genus. Trilete spores from the surrounding matrix are comparable to extant O. cinnamomea. These Late Cretaceous fossils are closely comparable to others known from the Paleocene and Neogene of North America, illustrating a continuous record of this species for at least the last 70 myr (Serbet and Rothwell, 1999). Aurealcaulis is an arborescent osmundaceous fern from the Paleocene of North America (FIG. 11.128) (Tidwell and Parker, 1987), in which the stem has an ectophloic siphonostele and C-shaped petiolar vascular strands. This taxon is unusual because leaf traces form by the fusion of

Figure 11.128 Suggested reconstruction of Aurealcaulis crossii (Paleocene). (From Taylor and Taylor, 1993.)

parts of two adjacent xylem strands and the primary xylem has exarch maturation. Associated with the silicified stem remains of A. crossii is compressed foliage assignable to Osmunda greenlandica. STERILE AND FERTILE FOLIAGE Several examples of osmundaceous foliage have been described from rocks ranging from the Triassic to the recent.

CHapter 11

Ferns and early fernlike plants

441

Figure 11.129 Osmundopsis sturii (Jurassic). Bar  15 μm. (Courtesy J. Van Konijnenburg-Van Cittert.)

The two most common osmundaceous foliage genera in the Mesozoic are Osmundopsis and Todites (Naugolnykh, 2002a). Osmundopsis (Harris, 1961a) includes dimorphic foliage, which is either bi- or tripinnate, and first appears in the Early Jurassic (Harris, 1961a). Osmundopsis sturii (Yorkshire) has delicate branches covered by obovate sporangia, each with a thickened distal cap. Dehiscence is longitudinal, and spores are subtriangular and trilete (FIG. 11.129). Anomopteris mougeotii is a Triassic foliage morphotype thought to be osmundaceous based on fertile remains (FIG. 11.130). The taxon is an important element of recovery floras immediately after the end-Permian extinction event (Fuchs et al., 1991). Todites first appears in the Late Triassic and is characterized by sessile sporangia borne on the abaxial sides of pinnae (Harris, 1961a). The Yorkshire Jurassic species T. thomasii includes widely spaced pinnules 2 cm long and 4 mm wide (Harris, 1961a). The pinnule midrib is pronounced, and the margin is dentate. Fertile pinnules are slightly shorter with a blunt apex. They bear oval clusters of sporangia attached over a lateral vein. The apical portion of the sporangium has a thickened patch of cells which is considered to represent an annulus. Spores are trilete, smooth, and 50 μm in diameter. Todites princeps (FIG. 11.131) is a widely distributed Jurassic species that includes fronds up to 100 cm long (Harris, 1961a). Pinnae are nearly opposite (FIG. 11.132), and pinnules vary from those with smooth margins to others that are deeply lobed. Sporangia are oval (250 μm in diameter) and tightly packed on the pinnule underside. Todites

Figure 11.130 Anomopteris mougeotii, portion of a fertile frond (Triassic). Bar  3 cm. (Courtesy BSPG.)

lobulatus from the Upper Permian of Russia has spores of the Osmundacidites type (Naugolnykh, 2002a). The spores of osmundaceous ferns from the Middle Jurassic of Yorkshire have been compared with other Jurassic spores and with those of living taxa. In general, fossil Osmundaceae have rather uniform spores differing only slightly in size, sporoderm thickness, and ornamentation (Van Konijnenburg-Van Cittert, 1978). Without attached sporangia, fern foliage is very difficult to accurately identify, and sterile, bipinnate frond material of Mesozoic age, similar to leaves of extant Todea, is referred to the foliage morphogenus Cladophlebis (see Chapter 16). Small permineralized osmundaceous trunks consisting of a branching stem with closely spaced, diverging frond bases and root traces are called Todea tidwellii (Jud et al., 2008). This Cretaceous plant has a stem with up to eight xylem bundles surrounding a sclerenchymatous pith. Leaf bases have C-shaped traces.

442

Paleobotany: the biology and evolution of fossil plants

Sporangia are stalked, 550–650 μm in diameter, and have a lateral annulus which is two to three cells deep. Some sporangia contain trilete spores which are 40–52 μm in diameter.

Figure 11.131 Suggested reconstruction of Todites prin-

ceps. (From Schweitzer, 1978; Courtesy U. Schweitzer and R. Gossmann.)

Figure 11.132 Todites princeps, small fronds (Late Triassic–

Early Jurassic). Bar  2 cm. (Courtesy BSPG.)

Permineralized fertile frond segments, from the Lower Cretaceous of Vancouver Island, British Columbia, are referred to as Osmunda vancouverensis (Vavrek et al., 2006). The fertile structures consist of branching, terete axes (pinnules) with large sporangia attached all around the axes in clusters.

OSMUNDALEAN EVOLUTION Although the fossil record of the Osmundaceae appears extensive, there are numerous gaps in our understanding of the origin of the family and the relationships among the taxa. The most comprehensive anatomical and morphological treatment of the family is that of Miller (1971), who examined 29 fossil and 14 extant species using multiple character correlation. The author distinguished nine groups of phylogenetically related species and suggested that the family originated from a Grammatopteris-like Paleozoic ancestor (discussed below), which rapidly radiated during the Permian. Since the Permian, the family has evolved slowly, with some specialized forms dying out near the end of the Mesozoic. The author recognized two subfamilies, the Permian Thamnopteroideae and the Osmundoideae, the latter of which includes the Mesozoic, Cenozoic, and recent species. The modern-appearing stele of Palaeosmunda (Upper Permian), however, suggests a far earlier geologic origin for the family and a separation of distinct lines in stelar evolution earlier than previously thought. Evidence for the early radiation of the family based on frond dimorphism can be seen in compression specimens of Osmunda claytoniites from the Upper Triassic of Antarctica (Phipps et al., 1998). Once-pinnate fronds (FIG. 11.133) consist of two sterile and four reduced fertile pinnae (FIG. 11.134) that arise from a single rachis. The fertile pinnae either possess a highly reduced lamina or entirely lack vegetative tissue and are borne below the sterile leaves. Fertile pinnae consist of numerous aggregated sporangia, each with a laterally oriented annulus like that in extant members of Osmunda. Although there are minor differences between the fossil and extant species, the position and organization of the fertile units is nearly identical to that in living O. claytoniana, the so-called interrupted fern, and to some extent with Osmundastrum cinnamomea (Yatabe et al., 1999). Frond dimorphism in the Osmundaceae is thought to represent a derived feature (Miller, 1971). Of the modern genera, this feature is found only in Osmunda and not in Todites or Leptopteris. Although the fossil O. claytoniites exhibits incompletely dimorphic fronds, that is, the fertile and sterile units are borne on the same rachis, Hewitson (1962) has suggested that the position of fertile pinnules in the family can be somewhat plastic. There is presently insufficient fossil evidence to decipher the origins of either Todea or Leptopteris. What can be noted, however, is that, like some Paleozoic groups of ferns, the Osmundaceae appear to have rapidly speciated during

CHapter 11

Ferns and early fernlike plants

443

S

Figure 11.133 Compressed frond of Osmunda claytoniites

(Triassic). Bar  2 cm.

the Late Pennsylvanian–Early Permian and then remained relatively unchanged to the present (Galtier et al., 2001). The discovery of Osmundastrum cinnamomea from the Upper Cretaceous of Canada is another example of the species longevity of certain ferns and supports the hypothesis that species turnover rates in certain pteridophyte lineages may have been small (Serbet and Rothwell, 1999). Botryopteridaceae

The Botryopteridaceae, an exclusively Paleozoic family, is one of the groups previously included in the Coenopteridales that is now classified within the Filicales (Phillips, 1974) (FIG. 11.135). The fronds range from two to three dimensional, and sporangia are borne in massive clusters on abaxial pinnule surfaces. The most characteristic feature is the omegashaped (ω), adaxially directed xylem strand of the foliar axis, Botryopteris. The family includes several structurally

Figure 11.134 Detail of fertile frond of Osmunda claytoniites arising from stipe (S) (Triassic). Bar  4 mm.

preserved genera, with Botryopteris being the most common and well known. Botryopteris is known from both vegetative and fertile specimens that range from the Mississippian into the Permian and has been recorded from 30 stratigraphic levels within the Pennsylvanian alone (Phillips, 1974). Today, however, Botryopteris is used not only for petioles exhibiting a particular anatomy but also for the entire plant. VEGETATIVE ORGANS Although some species of Botryopteris were prostrate, rhizomatous ferns with helically produced fronds, others (perhaps most) were organized into long foliar parts and short cauline units (Holmes, 1984). Branching was lateral and distichous. In some treatments the vegetative organs are termed

444

Paleobotany: the biology and evolution of fossil plants

Figure 11.136 Two sporangia of Botryopteris tridentata (Pennsylvanian). (From Rothwell and Good, 2000.)

Figure 11.135 Tom L. Phillips.

foliar members because in many cases it is not possible to determine whether an axis represents a rachis or a primary pinna (Phillips, 1970). All stems have terete protosteles, except in B. tridentata, which is siphonostelic (Phillips, 1974). Botryopteris tridentata is now known as an entire plant (Rothwell and Good, 2000). It consists of unbranched rhizomes with helically arranged fronds that are tripinnate with ultimate laminar pinnules. Fertile pinnae are interspersed among sterile frond segments and consist of sori on the abaxial surface of partially enrolled pinnules. The annulus is biseriate (FIG. 11.136) and spores are trilete. The oldest (Viséan) member of the family is B. antiqua, which exhibits several patterns of frond branching. Shoots are produced from foliar members and these in turn give rise to pinnae. Shoots have a terete protostele and also bear roots. Botryopteris antiqua (FIG. 11.137) exhibits the simplest foliar xylem strand; it consists of an oval strand of scalariform tracheids with one or two protoxylem points. Evolutionary modification of the foliar xylem configuration in the genus is thought to have involved a shift of the lateral protoxylem groups to form a median group, and the elaboration of the strand into three adaxial xylem arms, thereby forming the characteristic omega shape. Other suggested phylogenetic changes within the genus include a shift from scalariform to circular-bordered pitting, an increase in the

Figure 11.137 Botryopteris antiqua rachides of first to ulti-

mate orders and sporangia (Mississippian). Bar  1 mm. (Courtesy J. Galtier.)

size of stems and leaves, and the development of a lamina (Phillips, 1974). Galtier (1981) indicated that B. antiqua lacks pinnules with a flattened lamina and suggested that the three-dimensional, ultimate segments should more accurately be interpreted as rachides and, as such, are not phylogenetically far removed from a trimerophyte ancestor. Some of these segments suggest the first indications of lamina development in the form of webbing—perhaps the beginning of the earliest fern frond. It is interesting that the development of the fern leaf in the Viséan occurred after the development of the seed fern frond, which was completed by the Late Devonian–Early Carboniferous, thus supporting the independent origins of a

CHapter 11

Figure 11.138 Botryopteris hirsuta primary rachis (left) bearing an epiphyllous shoot (Pennsylvanian). Bar  1 mm. (Courtesy J. Galtier.)

fernlike frond in both of these groups (Galtier, 1981). Galtier considered B. antiqua as possessing both primitive (nonlaminar ultimate segments) and advanced (the presence of annulate sporangia) characters (Galtier, 1967). A specimen of B. hirsuta (FIG. 11.138) (Westphalian A; Lower Pennsylvanian) demonstrates the varied branching in the genus, as well as indicating that xylem-strand configuration may not be an especially reliable systematic character (Holmes and Galtier, 1976). The fossil is unusual because the stem bears opposite, twin lateral shoots, which, in turn, each produce numerous petioles. Although the axis is long, no other type of branching is present, and within the length of the axis the cross-sectional anatomy changes from ellipsoidal with barely detectable protoxylem points to the distinct trident form. Primary pinnae are produced at regular intervals, with a bud borne at the base of each pinna (Holmes, 1984). This is in contrast to the irregular production of pinnae in B. ramosa. Botryopteris dichotoma (FIG. 11.139) is an Early Pennsylvanian (Westphalian) species from Belgium (Holmes and Galtier, 1983) that produced leaves in a 2/5 phyllotaxy and, like many botryopterids, branched profusely. The stem in B. forensis (FIG. 11.140), a common Late Pennsylvanian species, is protostelic, with parenchyma usually absent in the xylem. Tracheid pitting ranges from uniseriate scalariform to multiseriate, oval-bordered pits. Leaves are produced in a helical, 1/3–2/5 phyllotaxis with pinnately compound fronds bearing laminar pinnules. Pinnules range from Pecopterislike with pinnatifid tips to others that are sphenopterid in outline (Galtier and Phillips, 1977). Stomata are abaxial and interspersed among multicellular, equisetiform hairs. Some

Ferns and early fernlike plants

445

Figure 11.139 Botryopteris dichotoma, dichotomizing stem (Pennsylvanian). Bar  225 μm. (Courtesy J. Galtier.)

Figure 11.140 Transverse section of Botryopteris forensis petiole with omega-shaped xylem strand (Pennsylvanian). Bar  1 mm.

trichomes are 300 μm long. Petioles and their pinnately arranged laterals may bear pairs of small, adventitious stems near their bases. These stems are oriented so that their first leaf occupies the same position as the lateral pinna of the frond member on which it was borne. As a result of the large number of stems and leaves that were produced in B. forensis, the plant had a false stem morphology. In transverse section, the foliar members are circular. Hair bases are uni- to multicellular, with the tips sharply pointed. The cortex is twoparted, with the outer zone constructed of thick-walled cells that grade into fibers near the epidermis. The inner zone consists of thin-walled parenchyma. Rothwell (1991) reported B. forensis rooted in the adventitious root mantle of a Psaronius specimen, thus establishing it as a trunk epiphyte.

446

Paleobotany: the biology and evolution of fossil plants

Figure 11.141 Botryopteris nollii (arrow) growing along the periphery of a Grammatopteris freitasii stem (Permian). Bar  2 cm. (Courtesy R. Rössler.)

Figure 11.143 Botryopteris nollii, partial reconstruction of the plant. (From Rössler and Galtier, 2003.)

protostelic stems and stipes that display a dorsiventral omega-shaped vascular system (FIG. 11.143). Botryopteris nollii is suggested to have formed false trunks of its own or thrived in the root mantles of tree ferns, such as Grammatopteris or Psaronius, with which it is associated.

11.142 Botryopteris nollii, longitudinal (Permian). Bar  1 cm. (Courtesy R. Rössler.)

Figure

section

One of the largest and also geologically youngest species in the Botryopteridaceae is B. nollii (FIGS. 11.141, 11.142) from the Permian of Tocantins in Brazil (Rössler and Galtier, 2003). The species is characterized by nearly circular

REPRODUCTIVE ORGANS Botryopteris globosa (FIG. 11.144) is a unique type of reproductive structure borne on certain foliar members. It consists of a fertile pinna that repeatedly branches within a very short space; the ultimate segments terminate in sporangia. As a result, the system forms a massive cluster of sporangia that is estimated to have contained 50,000 sporangia. Some specimens of B. globosa exceed 5 cm in diameter, with some of the axes displaying the omega-shaped botryopterid anatomy. In cross section the unit is circular, with a centrally positioned, omega-shaped trace (Murdy and Andrews, 1957). Externally, the fructification has a median groove that marks the position of the sporangial masses on each side of the central pinna axis. Two types of sporangia are present in the

CHapter 11

Ferns and early fernlike plants

447

Figure 11.145 Face view of Botryopteris globosa sporangium

(Pennsylvanian). (From Phillips and Andrews, 1965.)

Figure 11.144 Section of Botryopteris globosa showing portions of petioles and tightly packed sporangia (Pennsylvanian). Bar  1 cm.

aggregation. Sterile sporangia occur around the periphery of the mass and consist of radially elongated cells, often with parenchymatous contents (Galtier, 1971). Sporangia that contain spores are usually teardrop shaped and constructed of a single, outer layer of thick-walled cells overlying a zone of cells with thinner walls. These sporangia are annulate with an equatorial patch of cells that extends to near the base of the sporangium on one face (FIGS. 11.145, 11.146). Cells of the annulus are elongate and thick walled. The globosoid type of fructification has been separated into two species, B. globosa (Middle Pennsylvanian) and B. forensis (Late Pennsylvanian), based on the type of spores it contains (Phillips and Rosso, 1970). Spores of both species are subtriangular to oval, trilete, and range from 23–59 μm in diameter. Those of B. forensis exhibit a verrucate to rugulate ornament, whereas B. globosa spores are vermiculate to densely rugulate. Another Botryopteris from the lower Middle Pennsylvanian of Eastern Kentucky contains stalked sporangia borne on Sphenopteris-like pinnules (Good, 1979). The spores are small (23–34 μm in diameter), triangular shaped, and comparable to Acanthotriletes, Leiotriletes, and Lophotriletes. The morphology of the sporangia resembles that of modern osmundaceous ferns.

Figure 11.146 Botryopteris globosa sporangium showing

thick-walled cells of annulus on either side. Compare with face view in FIG. 11.145 (Pennsylvanian). (From Phillips and Andrews, 1965.)

Another non-globose type of sporangial organization is present in B. cratis (FIG. 11.147) (Middle Pennsylvanian) (Millay and Taylor, 1980). Here, branched frond members produce numerous spherical sporangia (FIG. 11.148), each with a bandlike annulus that extends transversely across the lower half of the sporangium (FIGS. 11.149, 11.150). This species is unlike the other fructifications in that it possesses an outer ring of slightly larger sterile frond

448

Paleobotany: the biology and evolution of fossil plants

Figure 11.147 Cross section of two Botryopteris cratis petioles (Pennsylvanian). Bar  1 mm.

Figure 11.150 Diagrammatic reconstruction of Botryopteris

cratis sporangium showing prominent annulus (Pennsylvanian).

Figure 11.148 Section of sporangium wall and spores of Botryopteris cratis (Pennsylvanian). Bar  4 μm.

Figure 11.151 Ultrathin section of Botryopteris globosa spore

showing inner separable layer (Pennsylvanian). Bar  10 μm.

Figure 11.149 Sporangium of Botryopteris cratis showing

annulus in section view (arrow) (Pennsylvanian). Bar  2 μm.

members that surround the sporangial cluster like a basket. Although the genus Botryopteris was initially delimited on the basis of a particular anatomy of the stem and foliar member, the addition of information on the diversity of reproductive structures may require a reevaluation of this filicalean genus. Ultrastructural studies have been completed on the spores of several species of Botryopteris (FIG. 11.151) (Millay and

CHapter 11

Ferns and early fernlike plants

449

Taylor, 1982). All spores possess a relatively uniform sporoderm; the one exception is B. cratis, which has a delicate sculptine layer on the surface of the spore. All species of Botryopteris are interpreted as homosporous. OTHER GENERA The genus Rhabdoxylon (Holden, 1960) was initially described based on specimens from the Lower Pennsylvanian coalfields of Britain. Rhabdoxylon dichotomum is a small, dichotomously branched fern densely covered by unbranched, multicellular hairs. The stem is 2 mm in diameter and contains a small, circular, centrarch protostele. A Late Pennsylvanian specimen, R. americanum, is slightly larger and produced numerous, helically arranged foliar units (Dennis, 1968). Additional specimens of R. dichotomum described by Holmes (1979a) suggest a prostrate habit, whereas the closely spaced leaves that characterize R. americanum suggest a plant that was more erect. Holmes noted, however, that such differences may simply represent ontogenetic variation. Another protostelic fern that has been included in the Botryopteridaceae on the basis of the adaxial position of the protoxylem is Catenopteris. There is some suggestion that a plant like Catenopteris is ancestral to the Osmundaceae (Rössler and Galtier, 2002b). Catenopteris simplex (Phillips and Andrews, 1966) is a small, structurally preserved protostelic fern stem 6 mm in diameter. Petioles were borne in a 2/5 phyllotactic sequence, each with adaxially curved, C-shaped traces with smaller (presumably protoxylem) tracheids present on the concave surface. Arising from near the base of each petiole is a diarch adventitious root (FIG. 11.152). Nothing is known about the foliar or fertile parts of this plant.

Diagrammatic reconstruction of Catenopteris simplex (Pennsylvanian). (From Phillips, 1974; in Taylor and Taylor, 1993.)

Figure 11.152

Anachoropteridaceae

The Anachoropteridaceae is another Paleozoic family of extinct filicalean ferns formerly included in the coenopterids (Phillips and Andrews, 1965). These plants bear petioles with abaxially curved C-shaped traces (FIG. 11.153), which are assigned to the morphogenus Anachoropteris (FIG. 11.154). Petiole vascular strands in this family are the reverse of that in other fern families, where the open part of C-shaped trace is directed toward the stem (adaxial curvature). In other anatomical features, the members of this group resemble living ferns more closely than other groups. Fronds are essentially planated, although in several genera the more distal levels are not known. Some stems are protostelic, including both centrarch and exarch forms, whereas others have mixed protosteles or siphonosteles. The family is

Figure 11.153 Cross section of Tubicaulis (Anachoropteris) petiole (Pennsylvanian). Bar  3 mm.

450

Paleobotany: the biology and evolution of fossil plants

Figure 11.154 Cross section of Anachoropteris involuta petiole (Pennsylvanian). Bar  2 mm.

S

Figure 11.155 Tubicaulis stem (S) surrounded by abaxially

directed petioles (Pennsylvanian). Bar  5 mm.

known from the Lower Pennsylvanian (Westphalian A) into the Permian. Anachoropterids that have been found with fertile parts have been assigned to either the Sermayaceae or the Kaplanopteridaceae (discussed below). Anachoropteris petioles are known to belong to stems of the Tubicaulis type (FIG. 11.155). This relationship was initially suggested by Hall (1961), based on a petiole and stem from a Middle Pennsylvanian coal ball. The stem of Tubicaulis is elliptical in outline, the result of numerous, closely spaced leaves arranged in a 3/5 phyllotaxy. Multicellular hairs cover the stem surface, together with regularly spaced adventitious roots. The xylem strand of the

stem is exarch with metaxylem tracheids and plates of radiating parenchyma. Tubicaulis multiscalariformis (Middle Pennsylvanian) lacks xylem parenchyma but is distinct because the prominent strands of protoxylem extend along the stem stele (Delevoryas and Morgan, 1952). These features contrast with T. stewartii (Late Pennsylvanian), which lacks prominent protoxylem strands and contains abundant parenchyma associated with the primary xylem of the stem stele (Eggert, 1959). Among the Tubicaulis species there are at least two different stelar types that produced the same type of petiole. One group had solid protosteles with circularbordered pits, whereas the second group contained mixed protosteles with metaxylem tracheids characterized by multiseriate, scalariform pitting (Phillips, 1974). Further attesting to the artificial nature of the genus are the various sizes of the recognized species. Some, such as T. scandens (Late Pennsylvanian), are small and believed to have been epiphytic (Mamay, 1952); others, such as the Permian form T. solenites, are estimated to have been at least a meter tall. One common Pennsylvanian species that extended into the Permian is A. involuta (FIG. 11.154) (Hall, 1961). Specimens possess a C-shaped vascular strand with the edges enrolled or involute in shape. Two to four protoxylem points are present along the adaxial side of the xylem strand. A Permian species, A. pulchra, exhibits conspicuously inrolled arms that are folded back on themselves (Corda, 1845). An unusual pattern of branching is seen in A. gillotii, in which a large lateral trace branches to become another rachis. Holmes (1979b) suggested that the protoxylem of the lateral trace originated from division of the persistent protoxylem poles, much like that in the formation of a pinna trace. The genus Grammatopteris (FIG. 11.156) (Stephanian– Permian) (Renault, 1896a) is sometimes included in this family but has recently been amended to include arborescent stems surrounded by a mantle of petioles and roots (Rössler and Galtier, 2002b). The stele in these small trees is either a solid protostele or a mixed protostele with parenchyma (FIG. 11.157). These are interpreted as a stage in the vitalization (addition of parenchyma cells) of a protostele (see Chapter 7). Protoxylem tracheids are present on the adaxial surface of the petiole strand near the margin. Metaxylem tracheids possess multiseriate pitting. In G. freitasii (FIGS. 11.156, 11.157) (Permian), the trunk is 35 cm in diameter with petioles produced in a 5/13 phyllotaxis. The petiole strand is oval at the base, becoming more bar-shaped distally. In the Permian species G. rigollatii, the stem is protostelic, with an outer zone of radially aligned tracheids initially interpreted

CHapter 11

Ferns and early fernlike plants

451

mixed protostele, histologic features of the cortex, and indentations of the stele that resemble the vascular configuration in the Osmundaceae when leaf traces are produced (Miller, 1971). Another fern that is hypothesized to be related to the Osmundaceae, based on stem anatomy, is Rastopteris (Galtier et al., 2001). This Early Permian form from China has a protostele surrounded by a mantle of roots and petiole bases. Leaf traces are reniform and become adaxially recurved in the petiole base. It has been suggested that Grammatopteris and Rastopteris may represent early stages in the evolution of the Osmundaceae. Another hypothesis is that these stem morphotaxa reflect growth patterns that accompanied the radiation of ferns within an expanding ecosystem during the Carboniferous-to-Permian transition. Until recently, Anachoropteris clavata (Delevoryas and Morgan, 1954a) was included within the Anachoropteridaceae. This fern is now the type species and genus of a new Paleozoic fern family (Kaplanopteridaceae) discussed next. Figure 11.156 Grammatopteris freitasii, transverse section through a stem (Permian). Bar  2 cm. (Courtesy R. Rössler.)

Figure 11.157 Grammatopteris freitasii, showing details of the stele and initial leaf and root traces (Permian). Bar  2 mm. (Courtesy R. Rössler.)

as secondary xylem (Corsin, 1937), but now interpreted as radial files of metaxylem tracheids (Galtier et al., 2001). It is suggested that Grammatopteris is closely related to members of the Osmundaceae based on the shape of the petiole trace,

Kaplanopteridaceae

This family, named in honor of the distinguished plant morphologist Donald R. Kaplan, is based on the Late Pennsylvanian species Anachoropteris clavata from Berryville, Illinois, coal balls (Tomescu et al., 2006). As a result of the discovery of additional specimens from Ohio, this filicalean fern is now defined by protostelic rhizomes bearing helically arranged fronds. In cross section, the petiole of Kaplanopteris clavata is about 2.5 mm in diameter and contains a C-shaped xylem strand with the ends of the arms expanded (FIG. 11.158). Protoxylem tracheids occur in four or five groups along the convex, adaxial surface of the trace. Some metaxylem tracheids are up to 200 μm in diameter and exhibit closely spaced, circular-bordered pits. Surrounding the xylem strand is a narrow band of phloem elements that is identical with those of modern vascular cryptogams (Smoot, 1985). The surface of the petiole is covered with numerous multicellular hairs. Traces to laterals are borne on the side of the C-shaped strand and, at higher levels, become terete. Fronds are planar with up to four orders of alternate pinnae. Sporangia are annulate, indusiate, and attached to the abaxial pinnule surface by a long stalk (FIG. 11.159). Spores are radial and trilete. Kaplanopteris clavata possesses iterative growth in which epiphyllous plantlets are formed and entire fronds are derived from primary pinnae (Tomescu et al., 2008). These developmental strategies optimize opportunistic growth and are interpreted as allowing K. clavata to exploit new habitats as a result of scrambling,

452

Paleobotany: the biology and evolution of fossil plants

Figure 11.158 Cross section of Kaplanopteris clavata showing

stem (left) and petiole (Pennsylvanian). Bar  1 mm. (Courtesy G. W. Rothwell.)

Figure 11.160 John C. Holmes. (Courtesy J. Galtier.)

S

Figure 11.159 Section showing pinnules (arrow) and sporangia (S) of Kaplanopteris clavata (Pennsylvanian). Bar  200 μm. (Courtesy G. W. Rothwell.)

climbing, and epiphytic growth. The presence of latent croziers in the place of basal pinnae is considered an adaptation to a liane-type habit (Trivett and Rothwell, 1988a). This fossil fern demonstrates that several so-called synapomorphic filicalean features (e.g., elongate sporangial stalks, indusiate sori, and gradate spore development) were already in existence in certain lineages by the Pennsylvanian (Tomescu et al., 2006). Psalixochlaenaceae

The Psalixochlaenaceae is a small family of extinct Carboniferous ferns with small cylindrical protosteles and mesarch primary xylem (Holmes, 1977) (FIG. 11.160).

Psalixochlaena cylindrica is a name initially established by Holden (1960) for fern axes that had earlier been referred to as Rachiopteris cylindrica and Botryopteris cylindrica. Traces to the foliar members are abaxially curved in the main rachis and pinnae, and the tripinnate fronds produce gradate sori of marginally attached annulate sporangia. The protostele (FIG. 11.161) is composed of large, more or less radially aligned tracheids surrounding a zone of smaller central ones (FIG. 11.162). Petiole xylem strands are initially circular, but in more distal regions of the leaf they become abaxially C-shaped with three protoxylem points along the adaxial surface. Changes in stelar size and changes in the number and position of the protoxylem tracheids at different levels suggest that P. cylindrica had a determinate type of growth habit. Epiphyllous buds occur on the fronds (Holmes, 1989). The profuse branching of these filicalean ferns suggests that they were scrambling plants and that vegetative reproduction may have played an important role in their ability to colonize sites. The planated frond was tripinnate with subopposite pinnules (1.1 cm long) and of the Sphenopteris type (FIG. 11.163). At the ends of some pinnule lobes, the vascular strand becomes swollen beneath the concave soral receptacle (Holmes, 1981). Each sorus is gradate and contains five sessile sporangia organized in a ring. The annulus is inclined at an angle of 20° from the horizontal axis of the sporangium

CHapter 11

Ferns and early fernlike plants

453

Figure 11.162 Psalixochlaena cylindrica, dichotomizing stem (Pennsylvanian). Bar  1 mm. (Courtesy J. Galtier.)

Figure 11.161 Psalixochlaena, stem showing lateral branching

(Pennsylvanian). Bar  1 mm. (Courtesy J. Galtier.)

and is typically biseriate. The organization of the fertile parts in this fern is similar to that seen in the Carboniferous compression genera Hymenophyllites and Zeilleria (FIG. 11.164) and to the Middle Pennsylvanian genus Norwoodia (Good and Rothwell, 1988). The reproductive parts of Psalixochlaena compare favorably with ferns in the extant family Hymenophyllaceae. In many features, the Late Pennsylvanian fern Apotropteris minuta (Delevoryas and Morgan, 1954b) is similar to Psalixochlaena. In the original description of A. minuta, the stele is described as protostelic, but additional, larger specimens indicate a few centrally positioned parenchyma cells that form a pith. Petioles are apparently produced in no specific sequence; diarch, adventitious roots are borne along the stem. Petiole traces are abaxially C-shaped and may or may not have produced a gap in the stele when they were formed. Like Psalixochlaena, when traces are produced, they result

Figure 11.163 Suggested reconstruction of Psalixochlaena

cylindrica (Pennsylvanian). (From Holmes, 1981; in Taylor and Taylor, 1993.)

in almost equal divisions of the stem stele. The genus is known from relatively few specimens and nothing is known about the reproductive parts. Sermayaceae

The Sermayaceae is an extinct family of Pennsylvanian ferns (Eggert and Delevoryas, 1967) that includes two genera: Sermaya and Doneggia, both from the Upper Pennsylvanian of North America. Nothing is known about the habit or

454

Paleobotany: the biology and evolution of fossil plants

Figure 11.165 Axis of the climbing fern Ankyropteris brong-

Pinnae of the ultimate order of Zeilleria frenzli, a foliage morphogenus used for forms similar to Sphenopteris (Pennsylvanian). Bar  2 cm. (Courtesy BSPG.)

Figure 11.164

stems of these genera; both are based on detached petioles of the Anachoropteris type that produced alternately arranged pinnae and bluntly lobed Sphenopteris-like pinnules. The vascular strand of the pinnae is also slightly C-shaped in the abaxial plane. Pinnules are relatively delicate and lack differentiation between mesophyll and palisade layers. One especially interesting feature of Sermaya biseriata (Eggert and Delevoryas, 1967) is the presence of pinnules bearing radial sori, each consisting of about four sessile sporangia. On the distal face is a horizontal-oblique annulus made up of two rows of thick-walled cells. Spores of S. biseriata are trilete, subtriangular, and average 30–40 μm in diameter. They are unevenly ornamented by delicate punctae and resemble the sporae dispersae genus Leiotriletes. In Doneggia complura, the sori are restricted to the pinnule lobes and consist of 25–35 stalked sporangia (Rothwell, 1978). The annulus is like that of Sermaya. Some pinnules bear stalked, peltate structures that may represent some type of gland, anomalous sporangium, or early stage of asexual plantlet. Sporangia of Sermaya and Doneggia are more similar to those of some living fern families than they are to other extinct groups of ferns. In the Gleicheniaceae, for example, sporangia are stalked and exhibit a single row of annulus cells. Although dehiscence and sporangial shape are similar in the two families, the anatomy of the frond is markedly different, thus justifying the classification of these two genera within a separate family.

niartii from the root mantle of a Psaronius stem (Permian). Bar  5 mm. (Courtesy BSPG.)

Oligocarpia (Abbott, 1954) includes sterile and fertile foliage from the Carboniferous. It was originally assigned to the Gleicheniaceae, but has been reinterpreted as a compressed specimen of Sermaya (Eggert and Delevoryas, 1967; Pšenicˇka and Bek, 2001). This comparison is based on similarities in the foliar morphology and the size, shape, and structure of the sporangium. Spore ultrastructure has been proposed as a way to confirm the affinities for Oligocarpia kepingensis (Permian) (Y. Wang et al., 1999) within the Gleicheniaceae; however, spore development was not considered and spore features used in this analysis are known to occur in many filicalean families, both fossil and living.

Tedeleaceae

The Tedeleaceae also consists entirely of extinct taxa (FIG. 11.165) and can be traced from the Upper Mississippian into the Lower Permian. These ferns were initially included in the Zygopteridales, but the discovery of fertile parts of some members suggests that the plants are more correctly allied with filicalean ferns (Eggert and Taylor, 1966). The best-known species is Ankyropteris glabra (A. brongniartii, according to Mickle, 1980), known from the Middle Pennsylvanian of North America and the Early Permian of Europe (Baxter, 1951; Rössler, 2001a). In cross section, the stele is angular, usually pentarch, with an inner group of parenchyma cells and small scalariform tracheids and an outer zone of large metaxylem tracheids. The cortex is bilayered: the inner cortex contains thick-walled sclerotic cells with dark

CHapter 11

Ferns and early fernlike plants

S AB

455

P

Figure 11.167 Cross section of Ankyropteris glabra stem

showing pentagonal stele (S), axillary branch (AB), and H-shaped petiole (P). Arrows indicate vascular traces (Pennsylvanian). Bar  5 mm.

Figure 11.166 Diagrammatic reconstruction of Ankyropteris stem, petiole and axillary branch (Pennsylvanian).

contents and an outer zone is made up of thin-walled parenchyma cells. On the outer surface are numerous scalelike aphlebiae, some of which are up to 2 mm long (FIG. 11.166) (Eggert, 1959). Each is vascularized by a mesarch trace that extends from a lobe of the pentarch stele. Multicellular hairs are also present on the stem. Adventitious roots are abundant on some stems and bear endogenously produced laterals. Root hairs at least 100 μm long have been described from some ankyropterid specimens, and are morphologically comparable to those of the extant fern Camptosorus rhizophyllus (Mickle, 1980). The leaves of A. glabra are borne in a 2/5 phyllotaxy. In transverse section, traces to the petiole are tangentially flattened, but at a higher level, they become H-shaped (FIG. 11.167). Peripheral loops are present at the four ends (antennae) of the xylem strand. Primary pinnae were produced from opposite sides of the petiole in two closely spaced, alternating series, with secondary pinnae appearing along the opposite sides of the primary member (Eggert,

Figure 11.168 Suggested reconstruction of a portion of the frond of Ankyropteris (Tedelea) glabra (Pennsylvanian). (From Eggert and Taylor, 1966; in Taylor and Taylor, 1993.)

1963). Ultimate frond segments are laminar pinnules (FIG. 11.168) with haplocheilic stomata. The basal pair of pinnules is slightly larger and possesses undulating margins. Scalelike aphlebiae occur on the adaxial side of the petiole at the base of each primary pinna. Thus, the frond of A. glabra may be compared with the fronds of many living ferns that are biseriate, pinnately compound, and have planated pinnae and small laminar pinnules. Compressed foliage

456

Paleobotany: the biology and evolution of fossil plants

believed to belong to A. brongniartii is also known and is of the Senftenbergia type. Ankyropteris glabra was a creeping, scrambling, or climbing plant. The species is most frequently found on or within Psaronius root mantles, but has also been documented growing on calamite stems (Rössler, 2001a, 2003). Specimens from the Permian of Chemnitz, Germany, demonstrate how these ferns came into the root mantles of tree ferns. The ferns initially grew on the surface of the root mantles, but subsequently became overgrown by new roots produced by the host, and thus entirely enclosed by the root mantle. The massive sclerenchyma in the cortex of A. brongniartii may have prevented the stems from being compressed by the expanding root mantle. One feature of Ankyropteris not present in any modern ferns is axillary branching (FIG. 11.167). At basal levels, the configuration of the branch is terete. Within a short vertical distance, two distichous, H-shaped petiole traces are produced from the axillary branch. The production of axillary branches resulted in several additional orders of stems and petioles that no doubt provided a rather bushy appearance to the plant. The fertile frond of A. glabra is known in some detail and given the generic name Tedelea (Eggert and Taylor, 1966). Fertile pinnules are lobed and morphologically similar to some forms of Pecopteris or Sphenopteris. Sporangia are attached in a submarginal position to the lower surface of the pinnule near the end of a lateral vein (FIG. 11.169). The

Figure 11.169 Suggested reconstruction of a portion of the

fertile frond of Tedelea glabra. (From Eggert and Taylor, 1966.)

number of sporangia per cluster is variable (two to seven) and, on the basis of sporangial features, soral maturation is thought to have been of the simple type. The stalked sporangia have a conspicuous annulus on the distal one-third of the sporangium. Spore masses extracted from sporangia suggest that 140 spores were produced per sporangium. Mature spores are of the Raistrickia type. Based on sporangium position and histology, T. glabra is most similar to certain primitive filicalean ferns included in the Osmundaceae and Schizaeaceae; vegetative features of the frond are, however, different. For now, it is best to regard T. glabra as a member of an extinct filicalean fern family that cannot be related to any extant taxon. Ankyropteris hendricksii is an interesting species that consists of a false stem constructed of numerous petioles tightly associated with a dense ramentum of shoot-borne roots (FIG. 11.170) (Read, 1938b). The taxon is known from a single specimen 30 mm in diameter, discovered in a silicified erratic boulder. The stratigraphic position of the fossil is believed to be Lower Pennsylvanian, although Mississippian may be equally possible since the precise source of the fossil remains unknown. In A. westphaliensis (Mississippian of Europe) (Williamson, 1874), the antennae in the petiole trace are greatly recurved in cross section, almost to the point of touching. Pinna traces do not leave a gap in the petiole trace upon departure (Mickle, 1980). Perhaps the anatomically simplest species in the genus is A. corrugata. This Early Pennsylvanian fern was small and had a prostrate habit. Petioles are distichously arranged, and at higher levels occasionally dichotomized. Aphlebiae are borne in two alternating series along the petioles. Radiating files of tracheids in some parts of the plant may be secondary in origin.

Figure 11.170 Transverse section of several petioles (H-shaped stele) and stem (arrow) of Ankyropteris hendricksii (Pennsylvanian). Bar  20 μm. (Courtesy F. Hueber.)

CHapter 11

The Carboniferous fern Senftenbergia is also assigned to the Tedeleaceae. At one time, it was considered to represent the earliest fossil evidence of the Schizaeaceae (discussed below). Features that suggested affinities with this family were the arrangement of the sporangia in a double row, their longitudinal dehiscence, and their attachment to pecopterid foliage. Jennings and Eggert (1977), however, substantiated that Senftenbergia was not schizaeaceous based on permineralized material. They suggested that the biseriate frond, features of the petiole trace, and sporangial organization indicate a much closer association with the Tedeleaceae. Compressed fertile and sterile foliage of S. plumosa from the Carboniferous of the Czech Republic (Pšenicˇka and Bek, 2003) include sporangia with in situ spores of the Raistrickia type, and also confirm the assignment to the Tedeleaceae. Senftenbergia foliage and sporangia represent the fertile frond of more than a single group of ferns as evidenced by the foliar trace in permineralized specimens, which is clepsydroid in some specimens, whereas in Tedelea the ankyropterid type is present. The origin of the axillary branch system in Ankyropteris has been discussed for many years. Opinions differ as to whether the branch represents a reduced member of a cauline dichotomy with an associated leaf or a member of a lateral pair of petioles. Most evidence supports the hypothesis that the axillary branch was initiated on the stem and that the leaf later became associated with the branch. Ankyropteris corrugata apparently did not produce axillary branches, based on the appearance of a protoxylem strand in the base of the leaf trace. This leaf trace may represent evidence of an old axis. It cannot be determined whether the basal strand represents a reduced axis or, as Eggert (1959) suggested, has no phylogenetic significance.

Ferns and early fernlike plants

457

bundles in petiolar traces, one at each end). Galtier and Taylor (1994) suggested that Skaaripteris may represent an intermediate stage between Carboniferous filicaleans and some modern fern families. Tempskyaceae

The Tempskyaceae are an extinct family of Mesozoic ferns known principally from silicified trunks common in Lower and mid-Cretaceous deposits (Jung, 1983). A specimen from the Santonian (Upper Cretaceous) of Japan is the youngest species known to date (Nishida, 1986, 2001). The genus was widely distributed at mid-to-high paleolatitudes and has recently been recorded for the first time from the Southern Hemisphere (e.g., Australia) (Clifford and Dettmann, 2005) and Argentina (Tidwell and Wright, 2003). Tempskya is characterized by false trunks up to 50 cm in diameter composed of up to 200 intertwined stems, petioles (FIG. 11.171), and adventitious shoot-borne roots (FIG. 11.172). Some trunk diameters suggest that Tempskya may have attained heights of 6 m (FIG. 11.173) (Andrews and Kern, 1947). Other species, for example T. judithae from Queensland, Australia, were smaller, only 4–10 cm in diameter. The stems within the false trunk branch dichotomously and have amphiphloic siphonosteles that produce traces to small petioles. Individual stems are variable in size in different species, with the largest stem only 1 cm in diameter. The stele is exarch and contains scalariform tracheids and parenchyma. The roots are profusely branched (FIG. 11.174) and diarch; numerous root hairs extend from the epidermis,

Skaaripteridaceae

This family is known from a single morphotaxon of permineralized ferns from the Late Permian of Antarctica (Galtier and Taylor, 1994). Stems of Skaaripteris minuta were protostelic, with xylem consisting of an outer zone of large tracheids surrounding the central core of conducting elements. Petiole traces are initially C-shaped with protoxylem at each end; at more distal levels the trace becomes W-shaped in cross section. Ultimate segments are nonlaminar. Sporangia are borne on short stalks and the large annulus is horizontal. Spores are trilete and of the Horriditriletes type. Like so many Paleozoic ferns, Skaaripteris possesses features that are well established in extant filicalean ferns (e.g., sporangia like those of the Osmundaceae), while also possessing characteristics that are rare in modern groups (e.g., two protoxylem

Figure 11.171 Detail of Tempskya stem stele with petiole trace

(arrow) (Cretaceous). Bar  2 mm.

458

Paleobotany: the biology and evolution of fossil plants

Figure 11.172 Cross section of a portion of a Tempskya axis showing stems (arrows) and numerous adventitious roots (Cretaceous). Bar  4 mm.

S

Figure 11.174 Tempskya sp., surface of a stem, showing numerous adventitious roots, which form the bulk of the stem (Cretaceous). Bar  2 cm. (Courtesy BSPG.)

Figure 11.173 Cross section of Tempskya false stem showing numerous stems with petioles (S) in matrix of adventitious roots (Cretaceous). Bar  1 cm.

even in areas that are quite mature. These persistent root hairs probably functioned for a considerable period of time in moving water into the aerial adventitious root zone of the false trunk. A comparison of numerous cross sections of Tempskya indicates that the largest axes contain the smallest number

of stems, whereas the smaller-diameter trunks contain a large number of stems. This has been used to suggest that the sporeling stage of a Tempskya plant consisted of a single stem that continued to branch at more distal levels. Numerous adventitious shoot-borne roots were produced that held the stem system together. As the plant continued to grow, the most basal stems began to decay, whereas the mantle of adventitious roots continued to provide support and water absorption for the mature plant. This is precisely the same developmental pattern hypothesized for the false stems in the marattialean Psaronius. One unusual feature about Tempskya is the manner in which leaves were produced. Although no foliage has ever been found attached to the trunks (FIG. 11.175), the size of the petioles and petiole traces close to the stele suggests that the leaves were small, abundant, and did not persist for an appreciable period of time. Petiole and stem relationships suggest that the leaves were borne along the trunk instead

CHapter 11

Ferns and early fernlike plants

459

of as a crown of fronds. Anatomical details of some species of Tempskya were used to subdivide the genus into two groups (Ash and Read, 1976). In one, the stems are larger (5–15 mm in diameter), possess short internodes, and contain abundant parenchyma in the xylem. The second group has small to medium stems (2.5–8 mm in diameter) and little parenchyma in the xylem. It is probable that many of the species that have been named merely represent different levels or growth stages of the same plant and thus features, such as internode length and tissue distribution, can be highly variable and are probably not useful taxonomic characteristics. Although a few isolated spores and several sporangial remains have been found associated with Tempskya stems, none has ever been reported in a position that might indicate attachment. The discovery and documentation of foliage, including the fertile remains, would appear to be an especially rewarding area of research within this fossil fern family. Schizaeaceae

Figure 11.175 Suggested reconstruction of Tempskya. (From

Andrews and Kern, 1947.)

The Schizaeaceae are, for the most part, a tropical and subtropical family consisting of four or five living genera, Anemia, Actinostachys (placed in Schizaea by some), Lygodium, Mohria, and Schizaea, and 170 species. Leaves vary from simple to pinnately compound and, in Lygodium, a climbing fern may be up to 10 m long. The organization of the stele ranges from protostelic to dictyostelic. Characteristics of the family include large, abaxial sporangia that are arranged singly and may be partially covered by an enrolling of the pinnule margin. Sporangia are often pyriform or flask shaped with an apical annulus and longitudinal dehiscence; spores are trilete (Dettmann and Clifford, 1991). As noted earlier, the Schizaeaceae were formerly considered an ancient family based on the presence of an apical annulus in Carboniferous Senftenbergia. The oldest evidence of the Schizaeaceae is now considered to be Jurassic, probably Middle Jurassic. The group is widely distributed by the Late Jurassic and is worldwide by the Early Cretaceous, with evidence from both spores and megafossils in Europe, North America, South America, China, and Japan. Some fossil forms can be assigned to an extant genus by the Early Cretaceous. Several species of schizaeaceous ferns are known from compression–impression remains of the Triassic and Jurassic age. Pekinopteris auriculata is a small herbaceous rhizomatous fern from the Triassic of the southeastern United States (Hope and Patterson, 1970). The plant is reconstructed with

460

Paleobotany: the biology and evolution of fossil plants

11.176 Partial frond of Pekinopteris auriculata (Triassic). Bar  2 cm.

Figure

simple pinnate fronds (FIGS. 11.176, 11.177) up to 35 cm long (Delevoryas and Hope, 1978). The fertile parts are believed to represent enlarged terminal pinnules bearing sporangia (Axsmith and Fraser, 2001). Klukia (Harris, 1961a) is a relatively common foliage morphotype which includes pecopterid pinnules borne on bipinnate fronds. Sterile pinnules are slightly convex, with the lower surface densely covered by delicate hairs. In K. exilis, from the Middle Jurassic of Yorkshire, the fertile pinnules contain 6–14 sporangia, each with a uniseriate annulus extending over the distal tip (Harris, 1961a). Spore counts range from 150–400 per sporangium, and include radial trilete grains with large sutures (FIG. 11.178). Specimens have also been reported from the Middle Jurassic of Iran (Vaez-Javadi and Mirzaei-Ataabadi, 2006). Dichotomous fronds with pinnately divided fertile foliar segments and clusters of sporangia borne at the tips of fertile pinnae are placed in Schizaeopsis (Skog, 1993). Schizaeopsis macrophylla, an Early Cretaceous species from the Potomac Group, has sporangia with an apical annulus and spores with prominent muri. The compression–impression fossils show characters found in extant Schizaea and Actinostachys, but spores which combine features of Mohria and Anemia (FIG. 11.179). Fertile foliage specimens now included in Anemia sphenopteroides (Skog, 1992) were initially believed to be osmundaceous. Schizaeopsis ekrtii from the

Figure 11.177 Suggested reconstruction of Pekinopteris auriculata. (From Delevoryas and Hope, 1978.)

Figure 11.178 Klukia exilis (Jurassic). Bar  25 μm. (Courtesy

J. Van Konijnenburg-Van Cittert.)

CHapter 11

Figure 11.179 Portion of an Anemia frond (Cretaceous). Bar 

2 cm.

Peruc-Korycany Formation (Upper Cretaceous) of the Czech Republic (Kvacˇek et al., 2006) is characterized by fronds that divide dichotomously more than seven times from base to tip. Both sterile and fertile fronds consist of narrowly lanceolate segments. The fertile segments terminate in sporangiophores that bear a single row of semicircular, tightly packed sporangia, which contain spores assignable to the Cicatricosisporites–Appendicisporites–Plicatella complex. Spores are tetrahedral and trilete with a triangular amb. Fertile spikes of Stachypteris consist of small axes bearing alternately arranged, reduced pinnule segments. The sporangia of this Jurassic (Yorkshire) fossil occur in two rows on the abaxial surfaces of pinnules. Details of the annulus are unknown, but the general morphology of the spores and organization of the fertile spike suggest affinities with Lygodium (Van Konijnenburg-Van Cittert, 1981). Cretaceous

Ferns and early fernlike plants

461

rocks from Wyoming contain excellent foliage specimens described as L. pumilum (Brown, 1943). Palmately lobed sterile pinnules have pinnule veins that dichotomize twice. Anemia quatsinoensis occurs in the form of permineralized sterile and fertile foliage remains from the Lower Cretaceous of Vancouver Island, British Columbia (Hernández-Castillo et al., 2006). Fertile pinnules are pinnately lobed, with abaxial sporangia and reflexed laminae. Pyriform sporangia occur in two rows on each pinnule lobe and have an apical annulus. Spores are tetrahedral to globose and ornamented by three sets of obliquely arranged parallel muri; they correspond to the sporae dispersae genus Cicatricosisporites. Associated with the foliage remains is an Anemia-like rhizome that contains an amphiphloic solenostele with exarch maturation. A number of Cenozoic schizaeaceous fossils have been assigned to modern genera. Lygodium kaulfussii includes both fertile and sterile foliage. The fertile pinnules of L. kaulfussii from Great Britain have axes bearing sporangia in clusters of three, each sporangium embedded in the pinnule surface and partially covered by what has been termed a bract or indusium. Spores (75–112 μm in diameter) are smooth and trilete. Specimens of L. kaulfussii from the Eocene of Wyoming show pronounced leaf dimorphism (Manchester and Zavada, 1987). The nonlaminar fertile pinnae produce trilete, psilate spores lacking a perispore. It is interesting to note that L. kaulfussii was widespread in the Eocene, but does not occur in the Oligocene of North America, although it persists in Europe to that time. Charcoalified sporophores of L. bierhorstiana are known from the Upper Cretaceous (Gandolfo et al., 2000). Permineralized specimens of Paralygodium vancouverensis have been described from the middle or upper Eocene of British Columbia (Trivett et al., 2006). Sporangia occur in two rows on the abaxial surface of lobed pinnules. Certain sporangial features, including a prominent apical annulus, support the inclusion of P. vancouverensis in the Schizaeaceae. Spores of P. meckertii are of the Deltoidospora type (Karafit and Stockey, 2008). Anemia poolensis is the binomial established for Cenozoic fertile foliage specimens from England (Chandler, 1955). Fertile pinnules are divided into four to seven segments that are recurved into a compact sporangial mass. Sporangia are globular and have a terminal annulus that occupies the upper third of the sporangium. Spores are trilete and 60 μm in diameter. Anemia fremontii is a common foliage type in the Late Cretaceous of North America. Another fossil that is morphologically similar to the extant genus Anemia is Pelletixia (formerly Pelletieria) from

462

Paleobotany: the biology and evolution of fossil plants

the Lower Cretaceous (Watson and Hill, 1982). Pelletixia amelguita (Skog, 1982) has two tightly enrolled pinnules bearing sporangia with spores of the Cicatricosisporites type. Skog (1982) suggested that Pelletixia and another probable schizaeaceous genus, Ruffordia, may be congeneric based on the similarity of pinnule morphology and spores. Lygodioisporites, Reticulosporis, Trilobosporites, Cicatricosisporites, and Appendicisporites are sporae dispersae taxa that date from the Cretaceous and have been suggested as being schizaeaceous (Prámparo, 1989). A phylogeny of the extant member of the family based on rbcL sequence data is used to suggest minimum ages for fossil taxa (Wikström et al., 2002). They hypothesize that the modern species of Lygodium evolved relatively recently, in the Neogene. This analysis also suggests that the earliest forms, like modern species, were perhaps climbers. Hymenophyllaceae

This family, sometimes termed the filmy ferns, includes more than 600 species in two major lineages that largely correspond to the traditionally recognized genera Hymenophyllum and Trichomanes (Schuettpelz and Pryer, 2006). They are found in tropical rainforests, although a few are known from temperate areas. The family occupies a near basal position in several phylogenetic analyses (Schneider et al., 2004). It has been suggested, based on extant taxa, that the family evolved in the Paleotropics, possibly in Asia (Dubuisson et al., 2003), although fossils are known from North America and China. The fossil record of the group is poor. Two reasons often cited for this are the membranous (filmy) nature of the frond which lacks preservational potential, and the possibility that some specimens are misidentified or grouped together with fossils that may be similar in overall appearance but represent entirely different groups of ferns (e.g., the dicksoniaceous foliage genus Coniopteris). Specimens from the Upper Triassic of North Carolina provide the most compelling evidence to date for the Mesozoic existence of the family (Axsmith et al., 2001). Hopetedia praetermissa has tripinnate fronds bearing lobed pinnules, often with funnel-shaped indusia (FIG. 11.180), each containing five to eight obovate sporangia. The annulus varies from oblique to vertical; no spores were recovered. Eogonocormus cretaceum and E. linearifolium from the Lower Cretaceous of northeastern China are small, thalloid plants with creeping rhizomes. These features, along with the marginal sori with in situ spores, borne on fanlike pinnule lobes, indicate convincingly that these ferns belong to the Hymenophyllaceae (Deng, 1997, 2002).

Figure 11.180 Detail of frond segments of Hopetedia praeter-

missa showing marginal sporangia (Triassic). Bar  4 mm.

Putative members of the family include Hymenophyllites quadridactylites, a Carboniferous foliage type with clusters of sporangia at the ends of the pinnule lobes, and Trichomanides laxum from the Jurassic of Australia. Gleicheniaceae

Extant members of the Gleicheniaceae are found in both the Old and New World tropics and include 130 species in three to five genera, including Gleichenia and Dicranopteris (Tryon and Tryon, 1982). Modern gleicheniaceous ferns possess long creeping rhizomes bearing large fronds that are several times pinnate. The branching of the fronds in the Gleicheniaceae is distinctive due to the presence of a so-called resting bud, which is formed by rachial dichotomies in which one part of the rachis is temporarily arrested in its growth. Stems are typically protostelic, with some steles containing parenchyma. Petiole vascular strands may be solid or C-shaped with enrolled ends. Sporangia occur in ring-shaped sori on the abaxial surface of the pinnule and there is no indusium (exindusiate). The number of sporangia per sorus is usually small, but in some species may be up to 15 sporangia. Pinnule segments in some species are greatly reduced and cup shaped. The annulus is transverse to oblique.

CHapter 11

Historically, the family occupies a position that parallels that of the Schizaeaceae, with the earliest presumed member, Oligocarpia from the Carboniferous, reinterpreted as belonging to the Sermayaceae (discussed above). Spore ultrastructure has been proposed as a way to confirm the gleicheniaceous affinities for O. kepingensis (Permian) (Y. Wang et al., 1999). There are several Paleozoic and early Mesozoic ferns that are believed to document the geologic range of the family beginning in the Carboniferous, but some of these are lacking certain diagnostic features of the family. Chansitheca is a morphotaxon of fertile foliage, similar to Cladophlebis, which was originally described from the late Paleozoic of Shansi Province, northeast China, and considered to be gleicheniaceous foliage, although others discount this assignment based on a lack of detail of reproductive parts (Collinson, 1996). Clusters of large sporangia are attached to the pinnules of both the Pecopteris and Sphenopteris types in C. palaeosilvana (Halle, 1927). The annulus is equatorial, but poor preservation makes the determination of this characteristic questionable. Chansitheca wudaensis, from the Lower Permian of Inner Mongolia, includes sori of 4–10 globose sporangia, each with a transverse annulus (Deng et al., 2000). Sori are arranged in two rows on either side of the pinnule midveins. Compressed, bipinnate fronds with pinnules of the cladophleboid-type believed to be those of gleicheniaceous ferns are assigned to Szea (X. Yao and Taylor, 1988). Szea sinensis (Permian of China) has round, exindusiate sori each containing 20–30 sporangia, with the sorus attached to a small receptacle. Sporangia have a transverse, uniseriate annulus in the upper third of the sporangium. The subtriangular, trilete spores are of the Triquitrites type and average 108–216 per sporangium. Based on the occurrence of these plants in situ in a roof-shale flora, it is suggested that Szea was growing in a coal-swamp habitat where acidic conditions existed. This type of environment is consistent with the distribution of the gleicheniaceous fern Dicranopteris today in China. Despite what appears to be a paucity of gleicheniaceous ferns in the Paleozoic, the family is well represented in the Mesozoic. Fertile and sterile frond members bearing lobed and pecopterid foliage have been assigned to both Gleichenia and Gleichenites, although Nagalingum and Cantrill (2006), among others, have pointed out that the latter name is invalid, as all the original specimens have been reassigned to seed fern groups. Nevertheless, the genus continues to be widely used to refer to both fossil foliage and spores. Specimens of Gleichenites coloradensis were

Ferns and early fernlike plants

463

Figure 11.181 Suggested reconstruction of Gleichenites coloradensis (Cretaceous). (From Andrews and Pearsall, 1941.)

discovered in the Frontier Formation (Upper Cretaceous) of Wyoming in an almost pure stand (FIG. 11.181) (Andrews and Pearsall, 1941). Although fertile remains were not encountered, the trichotomous branching pattern and pinnule morphology are almost identical with those of certain species of Gleichenia. Charcoalified rachides from the Lower Cretaceous with a C-shaped trace and pinnules with scales are described as Gleichenia chaloneri (Herendeen and Skog, 1998). Permineralized sporangia of Gleichenipteris antarcticus from the Middle Triassic (Phipps et al., 2000) are associated with, but not organically attached to, the stem Antarctipteris (Millay and Taylor, 1990). Cenozoic fossils of gleicheniaceous ferns include the permineralized rhizome Gleichenia appianensis from the Eocene of Vancouver Island (Mindell et al., 2006), which represents the first evidence of Gleicheniaceae in the Cenozoic of North America. Gandolfo et al. (1997) erected a new genus, Boodlepteris, for beautifully preserved charcoalified vegetative and reproductive organs from the Upper Cretaceous of New Jersey (Potomac Group). Boodlepteris turoniana includes rhizomes, petioles, pinnules, sori, and spores. Rachides are attached to the rhizome in a single row and the clavate sporangia are grouped into sori of 10–12 sporangia on the abaxial surface of the pinnules. Numerous dispersed spores have been referred to the family from as early as the Jurassic (Bolchovitina, 1967). Some of these, for example Plicifera and Gleicheniidites, are similar to extant forms in having a pronounced triangular morphology and conspicuous trilete suture.

464

Paleobotany: the biology and evolution of fossil plants

Dicksoniaceae

The Dicksoniaceae, along with the Cyatheaceae and the Loxsomataceae, consist a monophyletic clade of tree ferns. The Dicksoniaceae includes five extant genera and represents one of several tree-fern groups with large, pinnate fronds. Although tree ferns have been variously classified in recent years as clades or grades (Korall et al., 2006; A. Smith et al., 2006), we use the family in the traditional sense in the discussion that follows. Anatomically the stems vary from simple siphonosteles to dictyosteles. Hairs or scales (also known as paleae) are common on fronds and rhizomes. The position of the sorus is variable, ranging from marginal to covering the entire abaxial surface of the fertile segment. Soral development in some extant members can be used to distinguish certain tree-fern lineages (Churchill et al., 1998), and to a limited extent this should be possible with fossils. The annulus varies from oblique to longitudinal, with spores typically trilete. The Dicksoniaceae appears to have been well diversified by the Mesozoic and was already widespread in the Jurassic (Van Konijnenburg-Van Cittert, 1989). Foliage of Coniopteris from the Rhaeto–Liassic (Upper Triassic–Lower Jurassic) of Iran may be the oldest record of the family (Kilpper, 1964). In the Middle Jurassic Yorkshire flora, specimens of Coniopteris represent one of the dominant floral elements (Harris, 1961a). In C. hymenophylloides, the leaves are lanceolate and some may be 12 cm wide. Pinnae are alternate, with multilobed pinnules. Both fertile and sterile leaves are about the same size, but fertile pinnules bear pairs of sori at the ends of pinnule lobes. In some fertile pinnae, basal pinnules are vegetative. Sporangia are borne in a cup-shaped structure formed by the indusium. The annulus is vertical and extends about two-thirds of the way around the sporangium. A slightly different soral arrangement is present in C. bella, where the indusium is divided into upper and lower lobes. The number of spores per sporangium is relatively low ( 72), and all are trilete. Kylikipteris is another Yorkshire fossil assigned to the Dicksoniaceae (Harris, 1961a), but is also known from the Lower Jurassic of Romania. The fronds are large and pinnate. In some specimens of K. arguta they are at least 50 cm wide. The Pecopteris-like pinnules are relatively small, and both vegetative and fertile pinnules are borne on the same pinna. Fertile pinnules lack a lamina and bear a stalked apical sorus surrounded by a cup-shaped indusium. Cup-shaped sori also occur marginally along the strongly recurved pinnules of Eboracia from the Middle Jurassic of Yorkshire. The sorus is strongly curved along the pinnule margin in E. lobifolia and spores are trilete. Dicksonia (Jurassic) specimens

have marginal sori with bivalved indusia (Wilson and Yates, 1953). Rickwoodopteris hirsuta is a permineralized cyathean stem from the Upper Cretaceous of Vancouver Island (Canada), which is characterized by an amphiphloic dictyostele, sclerenchymatous and parenchymatous pith, and meristeles that produce 6- to 10-parted frond traces (Stockey and Rothwell, 2004). The cortex is surrounded by a homogeneous sclerenchymatous hypodermis, and clothed by a dense ramentum of large trichomes. This plant has an amphiphloic dictyostele with sclerenchyma sheaths associated with both foliar and cauline traces. Lophosoria cupulatus is an Early Cretaceous impression–compression fossil from Snow Island, Antarctic Peninsula, which could be placed within the extant genus Lophosoria (FIG. 11.182). The fossil shows some similarities

Figure 11.182 Lophosoria sp. (Cretaceous). Bar  5 mm. (Courtesy D. Cantrill.)

CHapter 11

to foliage of Gleichenites and Microphyllopteris (Cantrill, 1998). The ultrastructure of the spore wall of Lophosoria has been compared with that of the sporae dispersae taxon Cyatheacidites isolated from Cretaceous and Cenozoic compressions (Kurmann and Taylor, 1987). Specimens of L. quadripinnata from the Oligocene–early Miocene of Tasmania are nearly identical to their modern counterparts, with the exception of paraphyses being present in the sori of the fossils (R. Hill et al., 2001). It is hypothesized that during the Cenozoic the genus radiated from South America, perhaps via West Antarctica. A Cretaceous permineralized trunk segment (Conantiopteris) has features in common with Lophosoria and Dicksonia, and demonstrates that anatomical characters can be useful in combination with other data sets to identify relationships (Lantz et al., 1999). Friis and Pedersen (1990) reported on fertile material of Onchiopsis psilotoides from a peat horizon of Early Cretaceous age in Denmark, which was compared with extant Thyrsopteris. The material includes tripinnate fronds that are dimorphic, with alternately arranged, stalked fertile units. Approximately 70 sporangia, each with an oblique annulus, are enclosed in an overlapping, leaflike sorus. Spores are trilete and in the 45–60-μm-size range. Cyatheaceae

Included in the Cyatheaceae are 4–6 extant genera and more than 650 species of tree ferns in 5 major clades (Conant and Stein, 2001). The family dates from at least the Jurassic and includes forms in which the vascular system is a dictyostele with medullary bundles. Extending from the apex is a massive crown of pinnately compound leaves, some up to 5 m long. In some taxa, the sporangia are borne in a low, cup-shaped indusium; in others the indusium is absent. Sporangial development is gradate and the annulus is oblique. Unlike several other filicalean families, the fossil record of the Cyatheaceae includes a number of structurally preserved stem remains. Cyathocaulis is an anatomically preserved stem from the Cretaceous that resembles the extant Cyathea (Ogura, 1927). Stems are dictyostelic, with each meristele surrounded by a sclerenchyma sheath. On the surface of the stems are adventitious roots and multicellular hairs. In C. yezopteroides (Late Cretaceous of Japan), the stem is 14 cm in diameter (Nishida, 1989). Leaf trace anatomy is like that in the petioles of the frond of the Late Cretaceous fern, Yezopteris (Nishida, 1981a). Cibotiocaulis is a Cretaceous stem with many of the same anatomical characters as Cyathocaulis, but differs in having small, raised leaf scars and a thin root mantle (Nishida, 1989). Cyathorachis

Ferns and early fernlike plants

465

is used for Cretaceous rachides that anatomically compare with those of extant Cyathea species (Nishida, 1981b). Oguracaulis is a permineralized (Late Jurassic–Early Cretaceous) fern from Tasmania (Tidwell et al., 1989). Based on the dictyostelic anatomy and features of the leaf traces, O. banksii is suggested as representing an early cyatheaceous fern. Silicified specimens of Cenozoic dicksoniaceous ferns are also known from Argentina. The stem of Alsophilocaulis is 6 cm in diameter and includes abundant sclerenchyma associated with the xylem. In A. calveloi, the xylem is dissected into V- and W-shaped strands; the pith region contains numerous clusters of tracheids and fibers (Menendez, 1961) (FIG. 11.183). In the foliage genus Alsophilites, sporangia occur in clusters, each with thickened, oblique, annuluslike cells. The absence of soral details makes assignment to stems questionable. Some Cenozoic foliage specimens have been described as species of the extant genera, Cyathea and Alsophila, although in most instances features of the sporangia were not present. The ultrastructure of in situ spores has been reported from Alsophilites nipponensis, an Early Cretaceous species (Shuklina and Polevova, 2007). The first permineralized fossil sori assignable to Cyathea come from the Early Cretaceous of Vancouver Island (S. Smith et al., 2003). Sporangia of C. cranhamii have a vertical annulus and each sporangium contains 64 subtriangular spores. Sporangial and soral features in this fossil provide convincing evidence that Cyathea had evolved by the Early Cretaceous.

Figure 11.183 Carlos A. Menendez.

466

Paleobotany: the biology and evolution of fossil plants

Certain silicified ferns from the upper Eocene of Texas are called Cyathodendron texanum (Arnold, 1945). On the stem surfaces are helically arranged leaf bases and flattened multicellular hairs. On both sides of the xylem band is a narrow zone of sclerenchyma. Tracheids are scalariform, and medullary xylem bundles occur intermixed with the ground parenchyma of the pith. Each petiole contains numerous vascular strands. Cyathodendron is similar to the stems of certain extant members of the family, but differs in the absence of leaf gaps. Matoniaceae

Today the Matoniaceae is represented by two genera, Matonia and Phanerosorus, and confined to Indonesia, Borneo, and New Guinea. The frond of Matonia (FIG. 11.184) consists of pedate (palmately dissected) leaves; in Phanerosorus the leaves are pendulous and pinnate. The rhizome is covered by hairs and consists of two or three concentric amphiphloic siphonosteles, with leaf traces produced from the outer cycle; sporangia are massive and arranged in a ring around a

11.184 Secondary pinna of Matonia jeffersonii (Cretaceous). Bar  1.5 cm. (From Nagalingum and Cantrill, 2006.)

Figure

receptacle that is the stalk of a peltate indusium. The position of the annulus is oblique, and spores are trilete. This filicalean family was also widespread throughout the Mesozoic (Berry, 1919), but was apparently represented by a relatively small number of species. Permineralized sporangia from the Middle Triassic of Antarctica, Tomaniopteris katonii (FIG. 11.185), are thought to represent the oldest occurrence of the family (Klavins et al., 2004). The large number of sporangia per sorus, multiseriate indusium, and massive receptacle provide useful characters in identifying this Triassic fern. Phlebopteris (FIGS. 11.186, 11.187) (Laccopteris) is a foliage genus known from the Upper Triassic into the Cretaceous. In P. hirsuta (Middle Jurassic of Yorkshire), the pinnules are narrow but broadly attached to the rachis. The sorus is ringlike and contains up to 12 sporangia; spores are in the 50- to 60-μm-size range. Phlebopteris polypodioides occurs at several localities in Europe, as well as Greenland and Korea, and extends from the Upper Triassic well into the Jurassic. Specimens from Yorkshire have fronds with fused pinnule veins. Sporangia are crowded in a ring of 14 (Harris, 1980). Phlebopteris smithii is a palmate leaf with 5–15 pinnatifid pinnae (Ash et al., 1982). Exindusiate sori occur on either side of the pinnule midrib and contain 7–20 sporangia in a ring (FIG. 11.188). The annulus is broad and vertical to oblique; spores are triangular (FIG. 11.189), trilete, and smooth walled and are similar to the sporae dispersae taxon Dictyophiliidites. This species extends from the Upper Triassic to the Jurassic and is very common in the Upper Triassic Chinle Formation of the southwestern United States. Harris (1980) has transferred material previously included in Phlebopteris braunii to the genus Matonia based on the nature of the indusiate sori. An expanded placenta that partially covers the bases of the sporangia is a feature of Matonidium goeppertii. This species

Figure 11.185 Diagrammatic reconstruction of Tomaniopteris katonii (Triassic). (From Klavins et al., 2004.)

CHapter 11

Ferns and early fernlike plants

467

Figure 11.187 Phlebopteris angustiloba, frond of a young plant (Late Triassic–Early Jurassic). Bar  1 cm. (Courtesy BSPG.)

Figure 11.186 Phlebopteris angustiloba, frond (Late Triassic– Early Jurassic). Bar  2 cm. (Courtesy BSPG.) Figure 11.189 Phlebopteris muensteri (Jurassic). Bar  35 μm. (Courtesy J. Van Konijnenburg-Van Cittert.)

Figure 11.188 Several sori of Phlebopteris smithii (Triassic). Bar  220 μm. (From Ash et al., 1982; in Taylor and Taylor, 1993.)

468

Paleobotany: the biology and evolution of fossil plants

Figure 11.190 Suggested reconstruction of a partial Matonidium americanum frond (Cretaceous).

is known from the Jurassic to the Cretaceous and includes specimens that consist of a petiole bearing 20 pinnae, each 20 cm long (Harris, 1961a). The species has small pinnules (1 cm long by 2 mm wide) with veins that branch once or twice depending on the size of the pinnule. The trilete spores are in the 44-μm-size range. In M. americanum the number of pinnae may be 30 (FIG. 11.190). Some of the fossils in this family had massive stems that produced helically arranged fronds. Primary pinnae in Weichselia reticulata (FIG. 11.191) are bipinnately compound and borne in a palmate fashion; some are at least 1.0 m long. Stomata are confined to the lower surfaces in the intervein areas and each has large guard cells. Soral clusters occur in a cone-like unit on separate fertile fronds. The ultimate fertile units are nearly peltate and each contains 12 sporangia. Spores are large (80 μm in diameter) with distinct contact areas on their proximal surfaces. Specimens from the Albian of Spain suggest soral development was gradate (Diez et al., 2005). Extending from near the base of each petiole is an axis thought to represent an adventitious root. Transverse sections of partially petrified specimens from the

Figure 11.191 Suggested reconstruction of Weichselia reticulata. (From Alvin, 1971.)

Cretaceous of Belgium indicate that the petioles contain 12 rings of vascular tissue, which are termed meristeles (Alvin, 1971). Weichselia exhibits several features that suggest assignment to the Matoniaceae, including the sori, indusia, and spores. Other features, including bipinnate fronds, stems with presumed rhizophores, the polycyclic dictyostele, nonlaminar fertile pinnules, and certain histologic characters argue against matoniaceous affinities. This mosaic of characters suggests to some that Weichselia is an early member of a group that may have given rise to other members of the family. Another point of view is that the genus should be treated as an exemplar of a separate fossil family (Alvin, 1971). Delosorus heterophyllus is a matoniaceous fern from the Lower Cretaceous of Maryland (Skog, 1988). The frond is bipinnate and characterized by a ring of large (0.5 mm in diameter), indusiate sori. A principal feature used to assign Delosorus to the Matoniaceae is the morphology of the spores. It is important to note that heteroblastic leaf morphology, that is leaf form that changes along the shoot during development, is common to modern members of the family (Kato and Setoguchi, 1999), a fact which may be important

CHapter 11

Ferns and early fernlike plants

469

to consider when describing fossil morphotaxa assigned to the group. Loxsomataceae

The Loxsomataceae includes two living genera: Loxsoma (often misspelled Loxoma), an endemic from the North Island of New Zealand, and Loxsomopsis, which occurs in Costa Rica and several South American countries (Lehnert et al., 2001). The pinnate leaves may be up to 5 m long with marginal, indusiate sori. Stems are covered with characteristic, stiff trichomes that are enlarged at the base. Sporangia are stalked with a slightly elongate, oblique annulus. Few fossils have been assigned to this family. Loxsomopteris anasilla is a fusinized rhizome from the Lower Cretaceous of Maryland (Skog, 1976). It contains a solenostele with sclerotic pith. Hairs like those of extant Loxsomopsis cover the rhizome. The presence of these hairs is one of the distinguishing characters separating Loxsomopteris and Solenostelopteris. Solenostelopteris loxsomoides was transferred to Loxsomopteris when similar hairs were discovered on the surface (Nishida and Nishida, 1982). Solenostelopteris is a morphotaxon of anatomically preserved, solenostelic rhizomes, and is known from the Jurassic to the Late Cretaceous of North America, India, and Japan. There are approximately six to seven species of this morphogenus, but they may not all represent members of the Loxsomataceae (Little et al., 2006). Solenostelopteris skogiae, from the Lower Cretaceous of Apple Bay, British Columbia, includes small, solenostelic rhizomes, 1.3–1.6 mm in diameter. Leaf traces are C-shaped and primary xylem maturation is endarch (Little et al., 2006). Dipteridaceae

The Dipteridaceae has two extant genera, Dipteris and Cheiropleuria, which occur in the Indo-Malayan region, extending to China and Japan. Fronds branch dichotomously, with veins branching at right angles to form a reticulate mesh. Sori are exindusiate and arranged in a linear series on either side of the midrib. The annulus is oblique. Polyphacelus stormensis consists of permineralized petioles (FIG. 11.192) and fertile pinnules with sori from the Middle Jurassic of Antarctica (X. Yao et al., 1991). The vascular strand consists of two groups of vascular segments (meristeles). The abaxial strand is omega-shaped in cross section, whereas the adaxial vascular tissue is represented as a series of small bundles that become further dissected as pinna traces are produced (FIG. 11.192). Although the morphology of the entire frond is not known, P. stormensis exhibits

Figure 11.192 Cross section of Polyphacelus stormensis peti-

ole showing vascular bundle configuration. Note numerous ellipsoidal coprolites (Jurassic). Bar  1 mm.

the characteristic reticulate mesh pattern of tertiary veins. Pinnules are hypostomatic with slightly sunken guard cells. Sori contain 10–13 sporangia interspersed with trichomes and are randomly scattered on the lamina. The annulus is oblique; spores appear to be monolete. Impression–compression foliage assigned to the Dipteridaceae includes Clathropteris (FIG. 11.193), Dictyophyllum, Goepertella, and Hausmannia. In Clathropteris meniscoides from the Rhaeto–Liassic (Upper Triassic–Lower Jurassic) of East Greenland, tertiary veins form rectangular meshes (Harris, 1931a), whereas in specimens reported from the Triassic of Sarawak (Kon’no, 1968) the veins are described as irregular. No rectangular meshes are present in Dictyophyllum, although the pinnae are deeply divided into lateral segments (Harris, 1961a). Pinnae also dichotomize twice to form a fan-shaped lamina in Hausmannia (FIG. 11.194); trichomes occur on the lower surface and are interspersed among the sporangia. Specimens of Hausmannia morinii from the Lower Cretaceous of Vancouver Island, Canada, have been reported that include impression–compression and permineralized forms (Stockey et al., 2006a). The fronds are fan shaped with toothed margins and the reticulate venation of the higher order veins is beautifully preserved. Delicate sporangia, not arranged in distinct sori, occur between veins on the abaxial surface; only 32 spores per sporangia were produced in this species. Sedimentological evidence and comparison with extant dipterids suggests that at least some of the Hausmannia

470

Paleobotany: the biology and evolution of fossil plants

Figure 11.193 Clathropteris lunzensis (Triassic). Bar  1 cm.

(Courtesy C. Pott.)

Figure 11.194 Hausmannia (Courtesy S. McLoughlin.)

sp.

(Jurassic).

Bar  1 cm.

species may have grown in disturbed habitats and along streams as a pioneer species (Cantrill, 1995; Stockey et al., 2006a). Polypodiales

It is difficult to define the characteristics of fossil polypod ferns because pteridologists working with living members

of this large and diverse group are not in agreement regarding the systematics of the group. These plants are sometimes treated at the ordinal level and sometimes at the family level (Mickel, 1973; Pryer et al., 2004; A. Smith et al., 2006). Mickel (1973) divided the Polypodiaceae sensu lato into four principal groups: the dennstaedtioids, adiantoids, aspidioids, and polypods. More recently, A. Smith et al. (2006) elevated many of the extant genera to the family level (15 total) within the order Polypodiales. It is estimated that there are more than 9000 species of polypod ferns. They differ from other ferns in the position and nature of the indusium, features of the spores, type of stele, habit, and habitat. The common genus Polypodium includes once-pinnate leaves, which often have anastomosing veins. The vascular system of the rhizome is a dictyostele. Soral position is variable and an indusium is generally absent. The sporangia are stalked and the position of the annulus is longitudinal. The families that are mentioned here include the Dennstaedtiaceae, considered a basal polypod, the Pteridaceae, which also has a basal position, and three families of more advanced eupolypods: Onocleaceae, Blechnaceae, and Polypodiaceae. Fossil polypodiaceous ferns include both compression remains and structurally preserved rhizomes beginning in rocks of the Upper Cretaceous. Petrified roots of a lindsaeoid fern (Lindsaeaceae) found within the root mantle of Tempskya from the Albian (Early Cretaceous), however, suggest an even earlier radiation of polypodiaceous ferns (Schneider and Kenrick, 2001). The Clarno Formation (Eocene) of Oregon contains silicified rhizomes of at least two distinct polypodiaceous genera. Dennstaedtiopsis aerenchymata (Dennstaedtiaceae) includes rhizomes with alternately arranged, distantly spaced pinnae (Arnold and Daugherty, 1964). The rhizome is 1 cm in diameter and contains a siphonostele (FIG. 11.195) with internal and external endodermis, pericycle, and phloem. Tracheid pitting is scalariform. The most diagnostic feature of the species is the aerenchyma tissue in the pith and cortex, which is arranged in a series of tangentially oriented partitions. On the outside of the rhizome are multicellular epidermal hairs. The vascular trace to the petiole is initiated adjacent to the formation of a gap in the stele of the rhizome and is U-shaped in cross section near the base (FIG. 11.196). At higher levels, the sides of the trace become indented and are almost identical to the petiole traces of the extant fern Dennstaedtia. Distal portions of the frond are not known, nor were any sporangia found associated with the specimens. Dennastra is a Paleocene taxon from Canada based on fertile fronds and sporangia that may represent a component of a Dennstaedtia-type plant (McIver and Basinger, 1993).

CHapter 11

Figure 11.195 Transverse section of Dennstaedtiopsis aerenchymata rhizome (Eocene). Bar  3 mm. (Courtesy T. Delevoryas.)

Figure 11.196 Cross section of rachis of Dennstaedtiopsis aer-

enchymata (Eocene). Bar  2 mm. (Courtesy T. Delevoryas.)

Another fossil from the Clarno chert is placed in the extant genus Acrostichum (Pteridaceae or pteroid ferns). Acrostichum preaureum has erect rhizomes with large petioles. The stele consists of a narrow band of closely packed tracheids. Petioles have a distinct adaxial groove, with the

Ferns and early fernlike plants

471

vascular bundles arranged in an omega-shaped configuration in cross section. In large petioles, the number of bundles in any one level may approach 1000. The orientation of the protoxylem elements is of two general types. Bundles located near the sloping sides of the adaxial groove have protoxylem directed toward the outside, whereas the remaining peripheral bundles have a reverse orientation of protoxylem tracheids. Laminar pinnules are borne at more distal levels of the frond, each with abaxially positioned sporangia with no definite soral arrangement. The annulus is vertical and extends about two-thirds of the length of the sporangium. Spores are trilete and smooth. Sporangia of Acrostichum containing spores have also been described from Eocene–Oligocene rocks from the Isle of Wight (Collinson, 1978) and the Eocene Geiseltal area of Germany (Barthel, 1976a). The Clarno chert is an excellent example of a fossil site from which ecological information can be determined based on the kinds of plants present combined with an analysis of the depositional environment. It has been suggested that the plant-bearing chert was an indurated soil of a marsh in which the primary inorganic component was reworked volcanic ash. Hot springs in the area are thought to have been responsible for transporting quantities of dissolved silica to the site and also, perhaps, for elevating the temperature in the area to make it subtropical, at least on a local scale. Cooler post-Eocene climates are thought to have been responsible for the elimination of a number of plants, such as Acrostichum, from the area. The Cretaceous Dakota Sandstone of east-central Utah and southwestern Colorado is the source of Astralopteris, a polypodiaceous fern that is similar to the extant genus Drynaria (Polypodiaceae), which is known from southeastern Asia. Astralopteris coloradica has large fronds with alternate to opposite coriaceous pinnules (Tidwell et al., 1967). Sori are round and arranged in a linear row on both sides of the midvein. Nothing is known about the spores. Numerous specimens of Onoclea (Onocleaceae) are known in Paleocene rocks of Canada (Rothwell and Stockey, 1991). These ferns are preserved as impression–compressions and are identical in all morphological features to extant specimens of O. sensibilis, the sensitive fern (FIG. 11.197). The presence of several complete specimens rooted in situ was used to suggest that the site was periodically flooded. Other Paleocene members of the Onocleaceae have also been described from the Rocky Mountains and Great Plains of North America (Brown, 1962), such as O. hesperia, which has sterile pinnae with entire margins. Permineralized rhizomes with attached petiole bases from Canada (Lower Cretaceous) have been included in

472

Paleobotany: the biology and evolution of fossil plants

Salviniales Marsileaceae

Figure 11.197 Portion of a fertile spike of Onoclea (Paleocene). Bar  1.5 cm. (Courtesy G. W. Rothwell.)

the Blechnaceae (Serbet and Rothwell, 2006) and represent the oldest anatomically preserved member of this family. The rhizome of Midlandia nishidae is a dictyostele made up of amphicribal meristeles, that is, the phloem in each meristele completely surrounds the xylem. The stipe bases include a complex trace composed of two large hippocampiform vascular bundles on either side of one-to-four smaller, oval bundles (hippocampiform  shaped like a seahorse). Woodwardia, an extant genus in the Blechnaceae, was widely distributed in the Paleogene (Collinson, 2001). A permineralized Miocene member is included in the extant species W. virginica (Pigg and Rothwell, 2001). The rhizome is dictyostelic and produced traces to the frond in the form of two large adaxial bundles that flank an arc of four to six smaller bundles, precisely like the vasculature in living plants.

The Marsileaceae is one of two families of heterosporous, aquatic filicalean ferns that is variously regarded as its own order, Marsileales, or is included in the Hydropteridales (Rothwell and Stockey, 1994) or Salviniales (A. Smith et al., 2006). Several phylogenetic analyses support the concept that the family is monophyletic and nested within the larger group of leptosporangiate ferns. The Marsileaceae includes three extant genera, Marsilea, Pilularia (FIG. 11.198), and Regnellidium, and 60 species of rooted, aquatic or amphibious plants with slender stems and leaves that occur at widely spaced nodes; venation is dichotomous. The plants often grow in dense mats that extend out over deeper water. Leaves are reduced and range from two opposite pinnae in Regnellidium to three pairs of pinnae in some species of Marsilea. In Pilularia, the leaves are filiform and lack a lamina. The stem has an amphiphloic siphonostele with internal and external endodermis; xylem maturation is mesarch. Reproductive structures are modified, globose or bean-shaped structures termed sporocarps that are specialized, stomatiferous foliar units that contain both mega- and microsporangia (FIG. 11.198). Some are known to be viable 40 years after being collected, suggesting that they represent an adaptation to survive desiccation. Megaspores of this group are large and characterized by an apical, cone-shaped gelatinous mass. A few megaspores and vegetative remains from the Cretaceous have been attributed to this group, but, in general, the fossil record is rather poor. Compressed foliage that is morphologically identical to the living genus Marsilea from the Cenomanian (Upper Cretaceous) of Kansas (Skog and Dilcher, 1992) has been assigned to the extant genus. Each petiole has four leaflets with closed dichotomous venation, that is, veins anastomose with the marginal vein near the edge of the leaf. Sporocarps have not been found attached in M. johnhallii, but are present in the surrounding matrix. Nagalingum (2007) suggested that these vegetative specimens, together with others from the Upper Cretaceous of Gondwana and Eocene of Wyoming, should be included in the genus Marsileaceaephyllum, since no data are available on their reproductive structures; Marsilea johnhallii, however, exhibits closed dichotomous venation, which the other members of Marsileaceaephyllum do not. Skog and Dilcher (1992) suggested that the family has changed little since the mid-Cretaceous. Another creeping rhizome bearing petioles, but with two leaflets, is Regnellites nagashimae, a form that extends from

CHapter 11

Ferns and early fernlike plants

473

ME

Figure 11.199 T. S. Mahabale.

MI

Figure 11.198 Section of Pilularia globulifera sporocarp showing megasporangia (ME) and microsporangia (MI) (Extant). Bar  250 μm.

the Upper Jurassic to Lower Cretaceous of Japan (Yamada and Kato, 2002). Large (1 cm long) hirsute sporocarps are produced singly at each node; details about the spores remain unknown. There are no fossil megaspores that are morphologically similar to the large spores produced in modern Marsilea (Lupia et al., 2000), with the exception of Molaspora, a megaspore with spirally twisted acrolamella from the Upper Cretaceous of northeastern Japan (Takahashi et al., 2001). An acrolamella is a leaflike extension on the proximal end of the megaspore, similar to a gula, which is characteristic of several heterosporous ferns. There are, however, isolated sporocarps from the Upper Cretaceous that are assignable to the modern genus Regnellidium. These independent data sources provide some support to the hypothesis that the diversification into modern marsileaceous lineages took place at least by the mid-Cretaceous. Rodeites is the generic name that has been applied to megaspores, microspores, and

sporocarps from Cenozoic silicified Deccan Intertrappean cherts of India (Surange, 1966; Sahni and Sitholey, 1943; Mahabale, 1956) (FIG. 11.199). These sporocarps contain seven sori, with megaspores that are 600 μm in diameter and trilete microspores in the 45 μm size range. Rodeites dakshinii compares most closely with modern Regnellidium. Salviniaceae

The Salviniaceae are represented by two extant genera, Azolla and Salvinia, consisting of 16 species, which are sometimes placed in two separate families, Azollaceae and Salviniaceae. The plants are generally small and free-floating in lakes and ponds. Salvinia has leaves borne in whorls of three on a rootless rhizome, with one leaf highly dissected and submerged. Leaves of Azolla are clustered on a rootbearing rhizome (FIG. 11.200). Azolla is able to live in nitrogen-poor environments, due to the presence of symbiotic cyanobacteria and other bacteria in special cavities within their leaves. The microorganisms convert inert N2 into a metabolically accessible form such as nitrate or ammonia (Bergman et al., 2007). Sporocarps occur on submerged leaves and differ from those of the Marsileaceae in that they are derived from a highly modified sorus. Microsporangia and megasporangia occur in separate sporocarps on the same leaf. The stem vascular system consists of a few tracheids and phloem cells embedded in a parenchymatous ground tissue. Unlike the Marsileaceae, the fossil record of the Salviniaceae is rather extensive, consisting of vegetative

474

Paleobotany: the biology and evolution of fossil plants

remains (FIG. 11.201), dispersed megaspores, and microspore massulae (Florin, 1940a). Microspores of Azolla and Salvinia do not normally occur singly, but rather are confined within a bubbly, gelatinous substance, the massula, which distally may have numerous hooked structures termed glochidia (FIGS. 11.202, 11.203), which serve to attach the massula to the megaspore. Salvinia megaspores are globose and characterized by a thick perispore, three distal chambers, and a three-lobed cavity, whereas those of Azolla have large floats at the distal end that provide a diagnostic feature for their identification as fossils (FIG. 11.204) (Hall and Swanson, 1968). The proximal region of the megaspore consists of a tripartite prolongation of the perispore termed the acrolamella. Arcellites is used for large (300–550 μm long) dispersed trilete salviniaceous megaspores from the Cretaceous which were produced by various heterosporous ferns. In A. disciformis there may be short projections ornamenting the spore wall (Lupia, 2004). The fossil record of the Salviniaceae begins in the Cretaceous, and is based principally on isolated megaspores (Collinson, 1980a). Other Cretaceous members of the order have been extensively studied, and include the following genera: Azolla, Azollopsis, Ariadnaesporites, Salvinia, Parazolla, and Glomerisporites. Megaspores of the Parazolla type have been macerated from the Upper Cretaceous heterosporous fern Hydropteris pinnata (Rothwell and Stockey, 1994). This species is known from compressed specimens of branching rhizomes with pinnate fronds and pinnules with anastomosing veins. Ellipsoidal, bisporangiate sporocarps are produced where the petiole is attached to the rhizome. Microspores are trilete and

Figure 11.200 Azolla showing submerged leaves and tightly

compressed floating leaves (Cenozoic). Bar  5 mm.

Figure 11.201 Salvinia mildeana showing two floating leaves and one submerged leaf, which looks somewhat like roots (Miocene). Bar  1 cm. (Courtesy BSPG.)

Figure 11.202 Detail of Azolla filicoides glochidium (Extant).

Bar  5 μm.

CHapter 11

Figure 11.203 Azolla filicoides massula with glochidia (Extant). Bar  10 μm.

embedded in massulae, which produce numerous filaments. Megaspore complexes have apical floats and glochidia with circinate tips. A cladistic analysis suggests that the combination of several characters in H. pinnata supports the hypothesis that heterosporous aquatic ferns are monophyletic. As a result, a new family was proposed, the Hydropteridaceae, which is considered basal to the Salviniaceae and the sister group to the Marsileaceae (Rothwell and Stockey, 1994). Glomerisporites is an oval megaspore covered with a hairy perispore (Hall, 1975) (FIG. 11.205). The megaspore wall is two layered, and this is an important systematic character when studied at the ultrastructural level. Microspores consist of a perispore with numerous hairs, especially prominent at the distal end. Based on the large number of small, simply constructed floats, Glomerisporites is regarded as the most primitive member of the family (Hall, 1974). In Azollopsis (Late Cretaceous), a large number of floats are arranged around the megaspore. Massulae are attached to the megaspores by multibarbed glochidia. Parazolla includes large megaspores in which the floats are confined to the proximal region in one or two tiers (Hall, 1969). These megaspores have been compared with Azolla, but the massulae are distinct. They are banana-shaped and bear glochidia with terminal knobs. Ariadnaesporites (Cenomanian) is believed to represent an extinct salviniaceous group adapted

Ferns and early fernlike plants

475

Figure 11.204 Azolla intertrappea megaspore with two distal massulae (Tertiary). Bar  125 μm. (From Sahni, 1941.)

Figure 11.205

John W. Hall.

476

Paleobotany: the biology and evolution of fossil plants

to semi- or fully aquatic habitats (Nowak and Lupia, 2005). The megaspores have proximal floats with associated hairs. As a result of a detailed ultrastructural study of specimens of A. varius, several unique characters were used to establish the new genus Hallisporites (Nowak and Lupia, 2005). Vegetative remains of the family were initially included in the genus Azollophyllum, but later transferred to Azolla. Azolla primaeva is known from the Eocene of southern British Columbia (Hills and Gopal, 1967). Rhizomes bear bilobed, ovate leaves and unbranched roots. Associated with these fossils are abundant massulae, some containing recognizable, anchor-shaped glochidia. Azolla schopfii is a Paleocene taxon that is known from bilobed leaves and fertile material (Sweet and Chandrasekharam, 1973). Microspore complexes and megaspores are nearly identical with those seen in extant species of Azolla subgenus Rhizosperma. In A. schopfii the sporocarps are borne on the first leaf of each branch as in modern species. Another Azolla from the Paleocene that also includes both vegetative and fertile compressed specimens is A. stanleyi (Hoffman and Stockey, 1994). These were small plants (2.5 cm) with some of the leaves overlapping others. Sporocarps were produced on leaves along each lateral branch. Megaspore complexes include 15 floats; glochidia are well preserved on the microspore massulae. Abundant spore production is thought to be a plesiomorphic character that may be linked to periodically dry conditions. Salvinia stewartii is a Late Cretaceous (Maastrichtian) species recovered from rocks near Edmonton, Alberta, Canada (Jain, 1971). The megaspores lack hairs and the large trilete microspores are not associated with massulae or glochidia. Smaller megaspores (240–320 μm in diameter) and distinct massulae containing about 32 microspores have been described as S. aureovallis (Eocene of western North America). Glochidia are not evident, but the specimens are similar to the extant species of Salvinia, differing only in their slightly smaller size. Salvinia coahuilensis is known from compressions from the Upper Cretaceous of Mexico (Weber, 1973). The fossils include stems, sterile leaves, and fertile leaves with sporocarps in organic connection. Salvinia mildeana (FIG. 11.201) is used for Miocene impression and compression fossils (8–17 mm long) from Europe that are similar to modern forms (Collinson et al., 2001). These fossils occurred in large mats on the bedding plane and possessed so-called inflated aerenchymatous segments, which are absent in modern forms. Salvinitis is a stem from the Upper Cretaceous Deccan Intertrappean beds that also contains an aerenchymatous cortex and dorsiventral leaves with hollow trichomes (Nambudiri and Chitaley, 1991). The

permineralized sporocarps of S. deccaniana have elongate hairs arising from the surface. Vegetative and reproductive parts assignable to Azolla are also known from the same deposit. Based on the stratigraphy and morphology of the spores, Hall (1974) suggested that there were two evolutionary trends within the Salviniaceae. The first involved a reduction in the number of floats per megaspore in successively younger taxa. The second trend involved microspores, initially produced singly and later in massulae. Intermediate stages in the evolution of the massula involved coiled hairs that may have functioned as glochidia and the clustering of microspores, which superficially resembles a massula. Molecular studies of all extant species of Azolla suggest that the genus arose 50 Ma, but do not address the fact that there is a pattern in float number reduction through geologic time (Metzgar et al., 2007). The morphology and ultrastructure of the megaspore wall have also been used as systematic characters in the classification of fossil salviniaceous megaspores (Hall and Bergad, 1971; Friis, 1977a, Batten et al., 1998; Nowak and Lupia, 2005). In some cases the megaspore wall structure is nearly identical to that seen in extant and Quaternary specimens of Azolla pinnata from China (Z. Zhou, 1983a). Detailed studies like those of Nowak and Lupia (2005), which utilize multiple imaging systems within a framework of consistent protocols, represent an important next step in using fine-structural details as character states in phylogenies that include heterosporous fern spores. Examining the fine structure of the microspores of these plants will also be valuable, but may be more difficult. An intriguing small fossil with some resemblance to the Salviniaceae and Marsileaceae comes from the Upper Cretaceous of Russia (Baluyeva, 1964). Azinia paradoxa consists of bisporangiate sporocarps that contain megasporangia producing megaspores with floats and microsporangia producing microspore massulae. This combination of features is in contrast to that seen in any modern or fossil member of either family, and thus it has been suggested that Azinia should be placed in a separate family, Aziniaceae.

Conclusions Despite an extensive fossil record for many fern families (Collinson, 2001), fossils have only rarely been incorporated into studies focusing on the evolutionary relationships and origins of fern families. As Rothwell has noted (Rothwell, 1987a, 1996b), there is good fossil evidence for three major

CHapter 11

radiations of the ferns: one in the Paleozoic, which gave rise to the Carboniferous forms; one in the Late Permian to early Mesozoic, which gave rise to many modern families; and one in the Late Cretaceous to Paleogene, where the more advanced modern families and many modern genera appeared. It is apparent that the filicalean ferns were well represented in the Carboniferous by the Botryopteridaceae, Anachoropteridaceae, Kaplanopteridaceae, Sermayaceae, Psalixochlaenaceae, and Tedeleaceae. Various reproductive and vegetative characters that are considered apomorphic were present in the Paleozoic and may have evolved more than once (Galtier and Phillips, 1996). Many Paleozoic ferns produced biseriate fronds with foliar-borne sporangia, although some of these fronds had stem and petiole anatomy different from that of extant ferns. There is also considerable variation in the architecture of the frond in many of these taxa. Despite the limited amount of data on sporangial details, it appears that some of the earliest types possessed an annulus that was terminal or subterminal in position. In some of the Paleozoic families, the protection of the sporangium is variable, ranging from forms in which the sori are naked to those in which a definite indusium is present (Rothwell, 1987a). What is uncertain, however, is whether these groups became extinct in the Permian or had limited speciation into the Mesozoic. It has been suggested that this extinction is linked to the evolution of gymnosperm communities and large scale climatic shifts (Rothwell, 1996b). Fossil evidence for the second major radiation of the filicalean ferns extended from the Permian into the Jurassic (Anderson and Anderson, 2008). Most of the basal fern families appeared at this time, as well as the first filicalean tree ferns. At first glance it appears that all these families arose in the Jurassic, but this may simply reflect the well-preserved nature of some Jurassic floras, especially the fossils from the Middle Jurassic of Yorkshire, which have been extensively studied. The discovery of Triassic fern stems with modern filicalean anatomy and morphology underscores the fact that at least a few filicalean families extended into the earliest Mesozoic (FIG. 11.206). Although the Schizaeaceae and Gleicheniaceae can no longer be traced with confidence back to the Carboniferous, they, together with matoniaceous and osmundaceous members, are regarded as basal filicaleans. With the exception of the Osmundaceae, whose fossil record begins in the Permian, all of these families arose during the Early Mesozoic radiation. The third radiation of the ferns occurred principally from the Cretaceous onward and included the heterosporous ferns together with the polypodiaceous families. This radiation is associated with the explosive speciation of the flowering

Ferns and early fernlike plants

477

Figure 11.206 Anomopteris mougeotii (Osmundaceae) show-

ing circinate pinnae (arrows) (Triassic). Bar  5 mm. (Courtesy L. Grauvogel-Stamm.)

plants that must have created new ecological niches for those reproductive strategies encompassed by leptosporangiate ferns. Many of the fossil ferns that are permineralized, and thus provide an increased set of characters that can be used in phylogenetic analysis, show combinations of reproductive and vegetative features from more than one modern family (Rothwell and Stockey, 2006). These combinations of characters are present in both homosporous and heterosporous forms. They represent compelling evidence not only for mosaic evolution and speciation in Late Cretaceous–early Paleogene ferns but also offer at least indirect evidence of polyploidy which is so common in many living ferns. The success of allopolypoidy in living ferns is widely recognized (Haufler, 1989, 2008) and, in the changing ecosystems

478

Paleobotany: the biology and evolution of fossil plants

of the past, this mechanism may have given ferns the selective advantage in such situations, and this scenario has been repeated several times since the origin of ferns in the Devonian (Lovis, 1977; Page, 2002). The fossil record provides a context for this hypothesis, since it offers the opportunity to examine sporophyte architecture across a broad spectrum of major fern groups and over an extended period of geologic time. Understanding the ecological framework that impacted fern evolution has been a major contribution in tracing the evolution of this group, especially in the Paleozoic (DiMichele and Phillips, 2002). Several studies have focused on the ecological aspects of fossil ferns and these underscore that many groups appear to be associated with disturbed habitats. Added to this paleoecological dimension is new information about the evolution of growth habits, such as the epiphytic habit, based on field observations of extant ferns (Tsutsumi and Kato, 2006) and community structure of fossil forms. Poole and Page (2000) described a fern rachis from the famous London Clay flora (early Eocene of southeast England) with anatomical features of the Polypodiaceae. Based on the petiole vascular strand architecture and diameter they suggest that this Eocene fossil represents an epiphyte, a life form that is common in modern polypods. Although this plant is described as

the first evidence of epiphytism in a fossil fern, as noted earlier in this chapter there is good evidence for epiphytic ferns in the Paleozoic, for example Botryopteris forensis. The fossil record continues to provide important information about the biogeography and evolution of various features of ferns (Cantrill, 1998). Such characters as sporangial position (Schölch, 2007), maturation pattern, number, types, and ultrastructure of the spores are important systematic features that can help to resolve relationships among ferns. Other characters, including the nature of the gametophyte, have been assembled for extant ferns. Although these are less likely to impact phylogenies using fossils, ultrastructure of one part of the gametophyte phase—the spores—may provide an additional set of characters. As recent phylogenetic analyses have indicated, the phylogenetic relationships within the ferns are a long way from solution. Deciphering these relationships will no doubt involve the discovery of new fossil specimens and will require analysis of the rich and extensive diversity of ferns in the fossil record (Stockey et al., 2006b). Phylogenetic studies of ferns at all levels of resolution would appear to benefit from the inclusion of multiple data sets, including fossils (Soltis et al., 2002), if the evolutionary patterns within these vascular cryptogams are to be adequately understood (Rothwell and Nixon, 2006).

12 Progymnosperms

Archaeopteridales ............................................................. 480

Triloboxylon..................................................................................... 491

Archaeopteris Leaves ....................................................................... 481

Rellimia ............................................................................................ 492

Archaeopterid Reproduction ............................................................ 483

Other Aneurophytes ......................................................................... 494

Callixylon Stems .............................................................................. 484

Protopityales ......................................................................... 496

Other Archaeopterids ....................................................................... 487 Noeggerathians................................................................... 497 Aneurophytales ................................................................... 489 Progymnosperm evolution ....................................... 501

Aneurophyton................................................................................... 489 Tetraxylopteris.................................................................................. 489

There are rich counsels in the trees. Herbert P. Horne, Amico Suo

Certain discoveries and research accomplishments in paleobotany have had an enormous impact on the discipline. One of these is the recognition of a group of vascular plants in the Devonian that possessed gymnospermous anatomy, but which reproduced by free-sporing, pteridophytic methods— the progymnosperms (Beck, 1960a, b) (FIG. 12.1).

Higher taxa in this chapter:

Progymnosperms (Devonian–Pennsylvanian) Archaeopteridales Aneurophytales Protopityales Noeggerathians (Mississippian–Triassic)

Currently the progymnosperms (Progymnospermophyta) are represented by three orders: the Archaeopteridales, Aneurophytales, and Protopityales. Although the archaeopterids no longer represent the oldest arborescent plants

(see Eospermatopteris, Chapter 11), they are the earliest trees with modern wood anatomy and growth habit. The group extends from the Middle Devonian (early Eifelian) to the Early Mississippian (Tournaisian) (Beck, 1976b), and Archaeopteris–Callixylon was widespread by the Famennian (Late Devonian). In general all of these plants have a shrubby to arborescent habit with a pattern of lateral branching in which no axillary buds are produced. The ultimate appendages are either dichotomously branched units, as in the Aneurophytales, or laminate leaves with dichotomous venation in the Archaeopteridales. The current consensus is that progymnosperms are paraphyletic, and the leaves of Archaeopteridales are potentially homologous with those of seed plants (e.g., Friedman et al., 2004). Boyce and Knoll (2002), however, suggest that the progymnosperms represent a monophyletic lineage in which some members acquired megaphyllous leaves independently from seed plants. The vascular system in progymnosperms consists of a eustele, a stele type that consists of strands of vascular tissue called sympodia embedded in a parenchymatous ground tissue.

479

480

paleobotany: the biology and evolution of fossil plants

FIGURE 12.1 Charles B. Beck.

Progymnosperm steles include mesarch primary xylem and secondary tissues, both xylem and phloem, which are produced by a bifacial cambium. The wood is pycnoxylic, that is, dense (like conifer wood), with little parenchyma, and consists of tracheids with circular bordered pits and rays (Cichan and Taylor, 1990). These free-sporing plants produced fusiform sporangia that were borne along the adaxial or lateral surface of small axes that have been termed fertile leaves. Both homosporous and heterosporous forms are known. Although as a group the progymnosperms continue to be poorly understood, they provide the most convincing evidence of a lineage ancestral to the seed plants.

Archaeopteridales The most important taxon in this order is Archaeopteris, a morphogenus that was originally introduced for vegetative and fertile branch systems, but is today also widely used for the whole plant. Beck’s (1960a) discovery of an Archaeopteris specimen attached to a Callixylon stem was the evidence used to establish the concept of the Progymnospermopsida (in this volume, the Progymnospermophyta). The rules of botanical nomenclature indicate that, in a case where two morphogenera are demonstrated to be different parts of the same biological

FIGURE 12.2 Suggested reconstruction (Devonian). (From Beck, 1962.)

of

Archaeopteris

species, the generic name that was validly published first (in this case, Archaeopteris) serves as the legitimate name of the plant, although, as noted in Chapter 1, in some cases a new name is proposed for the whole plant. Archaeopteris was a major component of many Devonian– Carboniferous ecosystems (Meyer-Berthaud et al., 1999). The genus is known from numerous Northern Hemisphere localities in North America, Russia, China, Morocco (MeyerBerthaud et al., 1999), and Europe, as well as from the Southern Hemisphere, e.g., Australia (Beck and Wight, 1988), Colombia (Berry et al., 2000). By the middle to late Frasnian, monospecific archaeopterid forests had become the dominant vegetation type in lowland areas and coastal settings over a vast geographic area (e.g., Algeo et al., 2001; Scott and Glasspool, 2006). Archaeopteris has been reconstructed as a tall tree that looked much like any one of a number of modern conifers (FIG. 12.2) (Beck, 1962). The realization that

CHAPTER 12

Progymnosperms

481

FIGURE 12.4 Portion of an Archaeopteris branch. (From Andrews, 1961.) FIGURE 12.3 Branch with leaves of Archaeopteris fimbriata

(Devonian). Bar  2.5 cm. (Courtesy BSPG.)

Archaeopteris possessed pseudomonopodial branching in the lateral branch systems, however, suggests that the crown of the plant was probably more diffusely branched (Beck, 1979), and that perhaps not all species of Archaeopteris possessed the same branch architecture (Chaloner, 1999) or were arborescent (Beck, 1981). Archaeopteris Leaves

The morphogenus Archaeopteris (FIG. 12.3) was established as a subgenus by Dawson in 1871, but not validly published at the generic level until 1880 (Lesquereux, 1880; FaironDemaret et al., 2001). The genus was described as vegetative and reproductive structures of a free-sporing plant and at that time was classified with the ferns. Since the vegetative parts of the plant were initially thought to be fernlike, the early literature dealing with Archaeopteris referred to the leaves as rachides bearing pinnae with ultimate pinnules. As research continued, especially with specimens showing anatomical detail, it became apparent that the rachis of the

so-called frond was actually the main axis of a lateral branch system (Carluccio et al., 1966), which produced helically arranged leaves and ultimate branches in the same ontogenetic helix (Scheckler, 1978). Leaves arising from the penultimate branches appear opposite and decussate (FIG. 12.4), because they alternate with branches in the spiral. Scheckler (FIG. 12.5) showed that sections at many branch nodes reveal a continuity between the xylem of the branches and the parental axes, suggesting that the branches of Archaeopteris are not produced by axillary buds (FIG. 12.6), but rather develop from primordia like those that produced the leaves. Leaves of Archaeopteris exhibit extensive variability, ranging from highly dissected (FIG. 12.7) forms (A. fissilis) to those with nearly entire margins (such as A. obtusa). Leaf shape varies from obovate to spatulate; some were sessile, while others were attached by a short petiole. In some forms, anisophylly occurs, for example A. roemeriana (FaironDemaret and Leponce, 2001). In this species small leaves are about one-third the size of larger ones and are attached in different positions on the ultimate axes, that is, smaller leaves are attached to the adaxial surface, whereas larger ones are abaxial. The combination of shoot dorsiventrality

482

paleobotany: the biology and evolution of fossil plants

FIGURE 12.7 Archaeopteris leaves attached to a branch (Devonian). Bar  10 mm. FIGURE 12.5

Stephen E. Scheckler. (Courtesy L. C. Matten.)

Main axis

Penultimate branch

Callixylon 0

5mm Leaf

Ultimate branch

Archaeopteris

FIGURE 12.6 Diagram of primary and secondary body of Archaeopteris/Callixylon. Shaded represents primary xylem; cross hatching

the secondary xylem. (From Scheckler, 1978.)

CHAPTER 12

and anisophylly in extant plants is interpreted as maximizing light interception (Dengler, 1999), and Fairon-Demaret and Leponce (2001) suggest a similar adaptation in Archaeopteris. Anisophylly may be more common than realized in fossil plants, especially in compression specimens where leaves appear to extend into the matrix.

Progymnosperms

483

two rows (FIG. 12.9), perhaps on the adaxial surface, with as many as 40 sporangia being present on a single sporophyll (FIG. 12.10). Each sporangium is fusiform, 2 mm long, and pedicellate; in A. macilenta sporangia are up to 3.4 mm long. The sporangium wall is thick and dehisced longitudinally (FIG. 12.11), but there are no discernable specialized

Archaeopterid Reproduction

The fertile, ultimate branches of Archaeopteris often occur as the basalmost units of a branch system; however, in some specimens, both fertile and sterile branches are intermixed (FIG. 12.8). The ultimate fertile branches, which at one time were referred to as strobili (Beck, 1981), consist of axes (or fertile leaves) that are helically arranged. Vegetative leaves may occur below and above the fertile axes, or may be intermixed. An excellent example of the detail needed to characterize the fertile parts of Archaeopteris is the study of Fairon-Demaret et al. (2001). This study showed that the fertile axes (sporophylls) in A. roemeriana are helically arranged and bifurcate at the tip. In A. macilenta the decussate sporophylls are laminate, with the sporangia attached to the lamina (Andrews et al., 1965). Sporangia are borne in

FIGURE 12.8 Suggested reconstruction of Archaeopteris fissilis showing variability in fertile and sterile leaves (Devonian). (From Andrews et al., 1965.)

FIGURE 12.9 Fertile branch of Archaeopteris showing adaxial position of sporangia. (Devonian). Bar  6 mm. (Courtesy C.B. Beck.)

484

(A)

paleobotany: the biology and evolution of fossil plants

(B)

(C)

FIGURE 12.10 A. Archaeopteris sporophyll bearing sporangia.

B. Aggregation of microspores. C. Aggregation of megaspores. (From Andrews, 1961.)

cells in the form of an annulus. Stomata similar to those present on the vegetative leaves have been identified among the epidermal cells of the sporangia. Although heterospory has been demonstrated in several species of Archaeopteris, the exact distribution of the megaand microsporangia on a fertile branch is unknown for certain. In A. roemeriana microspores range up to 70 μm in diameter. The laesurae of the trilete scar are long, extending approximately three quarters of the spore radius. The proximal surface, especially in the contact area, is psilate, whereas the distal surface is ornamented with closely spaced coni. The small spores compare favorably with sporae dispersae grains of Aneurospora. Although no comprehensive study of Archaeopteris microspores has been undertaken at the ultrastructural level, sections of one species indicate that the exine was thin and homogeneous (Pettitt, 1966). Megaspores range from 110 to 400 μm in diameter and are assignable to the genus Contagisporites, with 16–32 megaspores per sporangium. The trilete scar is typically elevated and strongly developed. On the distal surface are closely spaced rods or coni; the proximal surface is smooth, occasionally ornamented with folds. A few small spores have been described as aborted megaspores. In other Archaeopteris fertile specimens the dispersed megaspores are included in Biharisporites while the microspores are of the Cyclogranisporites or Geminospora type (e.g., Balme, 1995; Marshall, 1996; Wellman and Gensel, 2004). Megaspores of the Biharisporites type have also been reported from Tanaitis furcihasta, a Frasnian archaeopterid described from Russia (Krassilov et al., 1987). Although the compressed specimen has fertile branches like those of Archaeopteris, the vegetative appendages are more similar to those of the aneurophytes. It is possible that species of Archaeopteris described as homosporous may represent one region of a branch system that included only microspores, with megasporangia confined to another region of the plant or produced in far less frequency than the microspores. Callixylon Stems

FIGURE 12.11 Archaeopteris sporangium that has split open

(Devonian). Bar  350 μm. (From Andrews et al., 1965.)

The trunk Callixylon was initially described by Zalessky (1911) for woody stems of Devonian age (for details, see Lemoigne et al., 1983). Some specimens extend up to 150 cm in diameter, and logs have been reported that exceed 10 m in length. Several species of Callixylon from countries throughout the world have been distinguished based on wood anatomical features, e.g., North America (Matten, 1972; Chitaley and Cai, 2001), Germany (Kräusel and Weyland, 1937), Russia (Snigirevskaya and Snigirevsky, 2001), China (Cai, 1989), and Morocco (Meyer-Berthaud

CHAPTER 12

et al., 1997, 2004). It is unknown whether Callixylon produced only Archaeopteris branching systems. Callixylon is eustelic with a central pith surrounded by a ring of mesarch primary xylem strands (FIG. 12.12) (Beck, 1979). In some specimens, these strands appear to be embedded in the parenchyma of the pith; in others, the strands are in contact with the secondary xylem. Secondary xylem consists of thick-walled tracheids and generally narrow vascular rays (FIG. 12.13). In a few specimens there is evidence of growth rings. The dense, compact (pycnoxylic) wood is similar to the secondary xylem of most coniferophytes and contrasts with the loose, highly parenchymatous organization of the manoxylic wood of seed ferns and cycads. Wood rays of most

Progymnosperms

485

Callixylon species are typically narrow, but may be broader in the region of the pith. They are variable in height and constructed of ray tracheids and ray parenchyma (FIG. 12.14). Circular bordered pits occur on the radial walls of the tracheids (FIG. 12.15), and pit size and aperture shape can vary greatly (Schmid, 1967; Beck, 1970). A characteristic feature of Callixylon is that the pits are grouped together in radially aligned rows and are multiseriate in most species (FIG. 12.13). Between groups of pits are slightly thickened areas of the primary wall that were originally compared to crassulae of extant conifers (Beck, 1970), but which are now thought to represent taphonomic changes resulting in the separation of wall layers in the tracheids (Beck et al., 1982). In C. arnoldii the pits are the largest in the genus (average of 7.1 μm in diameter) and possess circular apertures and a pit membrane that shows differentiation into a central and a peripheral area (FIG. 12.16). Beck et al. (1982) suggest that this may represent a type of primitive torus or a precursor to the torus (Beck et al., 1982). As research continues it is becoming increasing clear that the anatomy of Archaeopteris–Callixylon is far more complex than previously thought (Trivett, 1993; Meyer-Berthaud et al.,

FIGURE 12.12 Detail of primary xylem strand of Callixylon newberryi (Devonian). Bar  500 μm. (Courtesy C. B. Beck.)

R

FIGURE 12.13 Radial section of Callixylon newberryi wood

showing circular bordered pits and vascular rays (R) (Devonian). Bar  200 μm. (Courtesy C. B. Beck.)

Bordered pits of a ray tracheid in Callixylon erianum wood (Devonian). Bar  3 μm. (From Beck, 1970.)

FIGURE 12.14

486

paleobotany: the biology and evolution of fossil plants

FIGURE 12.15 Pits on the tracheid walls of Callixylon newberryi (Devonian). Bar  25 μm. (Courtesy C. B. Beck.)

FIGURE 12.16 Detail of pit membrane in Callixylon tracheid (Devonian). Bar  1 μm. (From Beck et al., 1982.)

1999). Based on an extensive collection of permineralized specimens from Morocco, three types of vascular trace patterns have been identified (Meyer-Berthaud et al., 2000). These include: (1) Type A traces are ephemeral and represent traces of short-lived apical branches that arise from the pseudomonopodial division of the trunk apex. (2) Type B traces supply larger branches that are interpreted as the major architectural units of the tree, which may have been adventitious. (3) Type H traces arise later, lack secondary xylem, and may have formed roots or branches if they became developmentally activated. One of the especially noteworthy aspects of this work is the fact that branching in Archaeopteris compares closely with the axillary branching seen in the earliest seed plants (Galtier, 1999). The branching innovations represented by the Type H traces may suggest a mechanism to increase tree survival (Trivett, 1993; Meyer-Berthaud et al., 2000). Recently Rothwell and Lev-Yadun (2005) demonstrated that within the wood of Callixylon there are unusual circular patterns of tracheary elements and associated parenchyma (FIGS. 12.17, 12.18). These patterns are identical to those seen in living seed plants where they form in response to a disruption of axial polar auxin flow caused by the presence of buds on branches. This report is noteworthy because it establishes that there was a regulatory mechanism associated with auxin transport in place at least 375 mya within the progymnosperms, and also because it underscores the potential for using certain types of fossil evidence in determining developmental and physiological mechanisms in ancient plants. Several authors have hypothesized that the morphologic and anatomic variability in Archaeopteris and Callixylon may indicate the presence of more than a single biological genus (e.g., Beck and Wight, 1988; Chitaley and Cai, 2001). The presence of both heterosporous and possibly homosporous taxa lends support to this suggestion. It is often difficult to recognize immature stages of sporophytes in the fossil record. One potential immature sporophyte is Eddya sullivanensis, a Late Devonian (Frasnian) plant interpreted as a juvenile form. Eddya has a maximum height of 30 cm (Beck, 1967) and consists of a slender axis that produced helically arranged, flabelliform leaves (6 cm long) with slightly undulating margins and dichotomous venation (FIG. 12.19). The underground root system of the plant is extensive with a robust primary root from which extend numerous lateral roots. The vascular system of the small stem consists of a eustele with mesarch primary xylem, surrounded by a small amount of secondary xylem that consists of tracheids with circular-bordered pits and narrow rays. It is hypothesized that E. sullivanensis represents

CHAPTER 12

Progymnosperms

487

FIGURE 12.18 Detail of circular pattern of tracheary elements

and parenchyma distal to branch stele (Devonian). Bar  1.0 mm. (Courtesy G. W. Rothwell.)

wood showing diverging branch stele (arrow) surrounded by pattern of circular tracheids (Devonian). Bar  1.5 mm. (Courtesy G. W. Rothwell.)

that consist of a grooved petiole 30 cm, and fan-shaped lamina ( 14 cm long by 15 cm wide). In addition, attached to the axes are sporangiophores which arise from the midrib region of wing-shaped leaves with distal lobes. The sporangiophores consist of a central axis giving rise to helically arranged, numerous (25), helically arranged short lateral branches. Sporangia are not preserved, and thus it remains unclear as to whether this enigmatic plant belongs to the progymnosperms, ferns, or another group of plants. Schweitzer and Giesen (2002) tentatively accommodate the fossils in the artificial, informal order Palaeophyllales (see Chapter 16).

an immature Archaeopteris plant, and the ontogenetic data on progymnosperms in general tends to supports this conclusion (Scheckler, 1978). Another axis bearing flabelliform leaves has been described from the Middle Devonian of Wuppertal (Germany) as Fuellingia gilkinetii (Schweitzer and Giesen, 2002). These axes are up to 4 cm in diameter and up to 1 m long, and repeatedly ramify. They bear helically arranged, large leaves

Several additional genera are included within the Archaeopteridales, but none of these is as well known as Archaeopteris. One is Svalbardia, a Middle Devonian genus originally described from Spitsbergen (FIG. 12.20) (Høeg, 1942 (FIG. 1.67); Schweitzer, 1999), but today also known from several other localities (e.g., Kräusel and Weyland, 1960; Matten, 1981; Allen and Marshall, 1986; Marshall and Stephenson, 1997). In the initial report of the genus the

FIGURE 12.17 Tangential section of Callixylon whiteanum

Other Archaeopterids

488

paleobotany: the biology and evolution of fossil plants

FIGURE 12.20 Suggested reconstruction of Svalbardia banksii

vegetative branch (Devonian). (From Matten, 1981.)

FIGURE 12.19 Suggested reconstruction of Eddy sullivanensis.

(From Beck 1967.)

branches were described as helical, but subsequent studies suggest a planated pattern (Beck, 1971). It may be that Svalbardia simply exhibits a high level of morphologic variability like that in Archaeopteris (Scheckler, 1978). Similar to Svalbardia, Actinopodium may also represent a portion of an Archaeopteris plant. Actinopodium nathorstii is known from

the Middle Devonian of Spitsbergen (Schweitzer, 1999). The genus is based on permineralized specimens that contain a medullated actinostele with wedges of mesarch xylem (Høeg, 1942); some secondary xylem is present. It has been suggested that Actinopodium represents the main axis of an Archaeopteris (e.g., Gensel and Andrews, 1984). Another Middle Devonian genus included in the Archaeopteridales is Actinoxylon (Matten, 1968). This plant has several orders of branching, with the ultimate branches bearing subopposite leaves in a decussate arrangement. The vascular strand consists of an actinostele with six ribs; each rib contains a protoxylem point at the end of the rib and another along the radius. Pitting ranges from helical to circular pitted and some secondary xylem is present. Langoxylon asterochlaenoideum from the Middle Devonian of Belgium is a peculiar permineralized fern-like stem of uncertain affinities that compares with archaeopterid progymnosperms such as Siderella, Actinopodium, and Actinoxylon (Scheckler et al., 2006). Siderella is a probable archaeopteridalean from Tournaisian (Lower Mississippian) rocks of the New Albany Shale in North America (Read, 1936a). Permineralized axes contain 7–10 lobes of primary xylem, with small patches of protoxylem, which superficially resemble peripheral loops, near the end of each lobe. Anatomically, Siderella is nearly identical with penultimate axes of Archaeopteris, except for the larger size of the former. One significant difference between the two is the decussate arrangement of the leaves in Archaeopteris,

CHAPTER 12

Progymnosperms

489

whereas in Siderella the leaves may have been distichous (Beck and Wight, 1988). This may not be a difference useful in separating the taxa, however, since detailed anatomical studies like those presented in Meyer-Berthaud et al. (2000) indicate that the arrangement of ultimate branches and leaves in the progymnosperms was highly variable.

Aneurophytales The Aneurophytales was initially proposed for plants that were thought to be the ancestors of both the ferns and seed ferns (Kräusel and Weyland, 1941). Today the order is considered to be the most primitive group within the progymnosperms (Beck, 1960). Aneurophytes are characterized by three-dimensional branching systems, with lateral axes helical or decussate in arrangement. The vascular system consists of a ribbed or lobed protostele with mesarch primary xylem; protoxylem strands occur at the tips of the ribs and toward the center of the stele. All members of the group have pycnoxylic wood with square-to-polygonal tracheids that sometimes branch. Tracheid pitting is circular bordered, and rays are tall; pits are crowded and occur on all walls, even the tapered end walls of the tracheids (Dannenhoffer and Bonamo, 2003). The fertile appendages consist of sporangia borne on ultimate pinnate divisions of a dichotomous branching system.

FIGURE 12.21 Aneurophyton germanicum, portion of a branch

system (Devonian). Bar  2 cm. (Courtesy BSPG.) Aneurophyton

One of the better-known members is Aneurophyton, described from Middle and Upper Devonian rocks from several localities in Europe, the former U.S.S.R., and the United States. The plant has at least three orders of helically arranged branches (FIG. 12.21). The three-dimensional, ultimate appendages (leaves) are one to three times dichotomized and unwebbed. Vegetative branches are determinate (Scheckler, 1976). The individual branches are usually naked, but a few specimens are known with prominent enations on the branches (e.g., Schweitzer and Giesen, 2002). Fertile branch systems of A. germanicum consist of three orders of branches that are helically arranged. The fertile units are represented by a central stalk 6 mm long that dichotomizes; each end curves inward and terminates in a blunt, lyre-shaped structure (FIG. 12.22) (Serlin and Banks, 1978). Arising from each arm of the unit are two rows of stalked, elliptical sporangia, each about 2.5 mm long. The number of sporangia per fertile unit is about 18 in A. germanicum (Schweitzer and Matten, 1982). In other

fertile organs of A. germanicum, which are interpreted as having been altered during preservation, the ultimate arms are slightly recurved away from one another, with the sporangia borne on the concave side. Spores have been macerated from only a few specimens of Aneurophyton. They are trilete, 40 μm in diameter, and correspond to the sporae dispersae genus Aneurospora. The primary xylem of Aneurophyton is triangular in cross section (Schweitzer and Matten, 1982). Xylem maturation is mesarch with four protoxylem strands, one near the tip of each lobe of the stele and one in the center. Secondary xylem occurs between the lobes of primary xylem and completely surrounds the primary body in older stems. Secondary xylem tracheids have multiseriate pits with elliptical borders and vascular rays are uniseriate. Tetraxylopteris

An equally well-known member of the Aneurophytales is Tetraxylopteris from the Upper and Middle Devonian of the

490

paleobotany: the biology and evolution of fossil plants

FIGURE 12.22 Fertile unit of Aneurophyton germanicuim

(Devonian). Bar  3 mm. (From Serlin and Banks, 1978.)

Northern Hemisphere (Beck, 1957); more recently, specimens have also been discovered in Gondwana (Hammond and Berry, 2005). Tetraxylopteris has a pseudomonopodial branching system with three orders of branches. First and second order branches are decussate (FIG. 12.23; 12.30), while ultimate appendages are helically arranged (Hammond and Berry, 2005). The ultimate appendages are highly variable in morphology and have been interpreted as leaves (Scheckler and Banks, 1971a). In cross section, the primary xylem of Tetraxylopteris consists of a cruciform mesarch strand (FIG. 12.24) that is maintained in successive orders of branches; in the leaves the xylem strand is terete. With a decrease in the size of the branches distally, there is a corresponding decrease in the number of protoxylem strands from as many as 18 in larger branches to 15 in fourthorder branches, and 2–4 in the ultimate appendages. Some of the secondary xylem tracheids have been described as branched and are over 3 mm long; pitting is circular bordered. The cortex contains anastomosing bands of fibers and parenchyma near the periphery (FIG. 12.25), a so-called Dictyoxylon (or dictyoxylon) type of cortex.

FIGURE 12.23 Tetraxylopteris schmidtii axis showing arrangement of branches. (Devonian). Bar  2 cm.

The fertile organs of Tetraxylopteris are complex, consisting of opposite and decussate sporangial complexes that occupy the position of penultimate branches (Bonamo and Banks, 1967). The fertile branching system consists of an axis that dichotomizes twice, with each of the four resulting branches three times pinnate (Scheckler, 1982). Each ultimate unit terminates in an elongate sporangium 5 mm long (Bonamo and Banks, 1967). Dehiscence is longitudinal, and in T. schmidtii there is a dark zone of cells suggestive of an annulus extending along the face of the sporangium opposite dehiscence. Spores are radial, trilete, and range from 73–173 μm in diameter. The exine separates to form an equatorial pseudosaccus that is ornamented by minute grana (Taylor and Scheckler,

CHAPTER 12

Progymnosperms

491

FIGURE 12.24 Cross section of Tetraxylopteris schmidtii axis

(Devonian). Bar  1.5 mm. (Courtesy C. B. Beck.) FIGURE 12.26 Proximal view of Rhabdosporites spore (Devonian). Bar  30 μm. (Courtesy S. E. Sheckler.)

FIGURE 12.25 Cross section of Tetraxylopteris schmidti showing fibers in cortex (arrows) (Devonian). Bar  1.2 mm. (Courtesy C. B. Beck.)

1996). Spores macerated from Tetraxylopteris sporangia are comparable to the sporae dispersae genus Rhabdosporites (FIG. 12.26) (e.g., Marshall, 1996). It may be that the large size range of the spores reflects stages in spore development, rather than heterospory. Tetraxylopteris reposana is interpreted as a relatively small bushy plant that may have grown in dense thickets with the three-dimensional branches interlocking (Hammond and Berry, 2005). Biomechanical analyses of T. schmidtii stems add some support to this interpretation, because they suggest that the overall development of the stems was not mechanically adapted for an entirely self-supporting architecture (Speck and Rowe, 2003, 2004). Another specimen was described as Sphenoxylon, but later demonstrated to

FIGURE 12.27

Lawrence C. Matten.

represent a poorly preserved Tetraxylopteris axis (Matten and Banks, 1967). Triloboxylon

Some researchers include Triloboxylon (Matten and Banks, 1966) (FIG. 12.27) in the Aneurophytales, because of the

492

paleobotany: the biology and evolution of fossil plants

FIGURE 12.29 Cross section of Triloboxylon ashlandicum stem

with cortical fibers (arrow) (Devonian). Bar  1 mm. (Courtesy C. B. Beck.)

FIGURE 12.28 Cross section of Triloboxylon ashlandicum

stem (above) with dichotomizing appendages below (Devonian). Bar  1 mm. (Courtesy C.B. Beck.)

spores. In T. arnoldii, traces are produced in pairs from the ribs of primary xylem (Stein and Beck, 1983). Rellimia

similarity in branching pattern and anatomy to Tetraxylopteris. In Triloboxylon ashlandicum, branches range from 1–4mm in diameter, with the primary xylem consisting of a threearmed mesarch protostele (FIG. 12.28); ultimate appendages contain a terete trace. Trace formation involves a small amount of parenchyma in the form of looplike areas at the ends of the lobes. Secondary xylem and extraxylary tissues, including a periderm, are similar to those of Tetraxylopteris, except that the arrangement of the cortical fibers is of the Sparganum (or sparganum) type, where vertical bands of cortical fibers are present (FIG. 12.29), but do not anastomose. Secondary phloem consists of sieve cells, fibers, and parenchyma (Stein and Beck, 1983). The fertile organs are twice dichotomized and borne in place of second-order vegetative branches along the main axis. Unlike other aneurophytes, the fertile organs of Triloboxylon are borne on the stem between regions of vegetative lateral branches (Beck and Wight, 1988), similar to the arrangement in the ultimate branches of Archaeopteris. Each fertile unit is 2 cm long, with the sporangia borne singly or in pairs. The xylem strand of the penultimate branch of the fertile axis is three armed, with the trace to the fertile organ dividing into two strands (Scheckler, 1975b). Sporangia possess apiculate tips and are about 3.5 mm long. Dehiscence is longitudinal and may involve a patch of cells functioning like an annulus. Nothing is known about the

Another genus included in the Aneurophytales is Rellimia (Leclercq and Bonamo, 1973), previously known as Protopteridium and Milleria (see a summary of the nomenclature in Leclercq and Bonamo, 1971, 1973). The genus has a long and complex history dating from 1871, when the first specimen was described by Dawson. Specimens of R. thomsonii (FIG. 12.31) are known from Middle Devonian (lower Eifelian to upper Givetian) rocks of several geographic regions (United States, Europe, Russia) (Bonamo, 1977; Dannenhoffer et al., 2007). Recently discovered specimens, including impressions, compressions, and partial petrifactions, indicate that the plant was shrubby, consisting of helically arranged branches in four orders. Terminating these branches were ultimate appendages, either leaves or fertile organs. In Rellimia the primary vascular system is three ribbed. Some of the largest branches are 2.5 cm in diameter. They contain lobed, mesarch primary xylem, surrounded by pycnoxylic secondary xylem composed of tracheids and high, narrow vascular rays. Details of the tracheid pitting indicate that the bordered pit pairs have crossed pit apertures (Dannenhoffer and Bonamo, 2003). Rays are homocellular. Surrounding the secondary xylem is a welldeveloped cortex with the sparganum arrangement of sclerenchyma bands. The presence of secondary phloem in Rellimia, as well as in Tetraxylopteris, indicates that the vascular cambium in the aneurophytes was bifacial. Dannenhoffer and Bonamo (1989) have described specimens of R. thomsonii in

CHAPTER 12

Progymnosperms

493

FIGURE 12.31 Vegetative branch of Rellimia thomsonii (Devonian). Bar  3 cm. (Courtesy P. Gerrienne.)

Branch of Rellimia thomsonii showing massive terminal fertile units (arrows) (Devonian). Bar  3 cm (Courtesy P. Gerrienne.)

FIGURE 12.30 Suggested reconstruction of Tetraxylopteris schmidtii fertile branch. (From Bonamo and Banks, 1967.)

FIGURE 12.32

which clearly defined growth layers are present in the secondary xylem. This information, when combined with sedimentological and paleomagnetic data, suggests that Rellimia grew in a seasonal wet–dry tropical climate. The fertile branch of Rellimia contains vegetative segments positioned basally and fertile structures above. The

axis of the fertile unit divides once into two first-order pinnae, each with three or four pinnate subdivisions that contain sporangia (Bonamo, 1977). Sporangia are often attached in pairs at three or four levels on the pinna, forming a massive (FIG. 12.32), overlapping structure. Sporangia are about the same size as those of Tetraxylopteris and produce spores of

494

paleobotany: the biology and evolution of fossil plants

the Rhabdosporites type (e.g., Marshall, 1996). Although the taxon was originally thought to be heterosporous, recent studies confirm the homosporous nature of the plant. At present, Rellimia and Tetraxylopteris may be distinguished by the oncedichotomized fertile organ and helical branching in the former. Other Aneurophytes

Polythecophyton from the Lower Devonian of China has a lateral branching system like that in some aneurophytes, but thus far there are no details about the vascular system (Hao et al., 2001). Eophyllophyton is another Early Devonian (Siegenian) plant from China that superficially resembles several progymnosperms (Hao, 1988; Hao and Beck, 1993). In E. bellum the axes bear lateral, flattened structures that are interpreted as leaves. They are small and fan-shaped with the lamina highly dissected (Hao et al., 2003). Some of these contain rows of sporangia that give the fertile unit (FIG. 12.33) the appearance of the ultimate fertile segments in Rellimia and Tetraxylopteris. What is especially interesting is that the laminate leaves in E. bellum are small (5 mm long) and anatomically complex. They appear to have been able to tolerate ecological conditions at a time of apparently very high CO2 levels (Hao et al., 2003). Several additional taxa are included within the Aneurophytales based on anatomical similarities with Aneurophyton, Tetraxylopteris, and Rellimia; none have fertile parts that are known. One of these is the genus Proteokalon (Scheckler and Banks, 1971b). This Late Devonian plant is known from permineralized axes displaying two orders of branching, the second of which reveals planated ultimate appendages. The stele of P. petryi is four-armed in cross section (FIG. 12.34), with the number of mesarch protoxylem strands as high as 36 in first-order branches. In second-order branches, the primary xylem is three-lobed (FIGS. 12.34, 12.35). Secondary xylem tracheids are typically four-angled, with crowded pits on the radial walls. Like most progymnosperms, Proteokalon contains secondary phloem, which includes fibers, tanniniferous cells, rays, and phloem parenchyma. Although thin-walled cells, presumably representing sieve elements, are present in the primary phloem, they were not observed in the secondary phloem. The bilateral symmetry of the penultimate axes has been suggested as evidence for an early stage in the evolution of the frondlike leaf and represents the most advanced branching pattern within the Aneurophytales (Scheckler and Banks, 1971b). Another plant tentatively placed in the aneurophytes is Cairoa lamanekii, known from a single petrified axis (penultimate branch) 1 cm in diameter containing a three-lobed,

FIGURE 12.33 Fertile units (arrow) of Eophyllophyton bellum (Devonian). Bar  3 mm. (From Hao, 1988.)

FIGURE 12.34 Cross section of Proteokalon petryi axis (left) and branch (right) (Devonian). Bar  1 mm. (Courtesy C. B. Beck.)

CHAPTER 12

Progymnosperms

495

FIGURE 12.36 Cross section of Reimannia aldenense axis (Devonian). Bar  500 μm.

FIGURE 12.37 Cross section of shoot of Cecropsis luculen-

ryi showing ultimate appendages. (From Scheckler and Banks, 1971.)

tum showing laminar organization of sporophylls and attached sporangia (arrows) (Pennsylvanian). Bar  1 mm. (Courtesy G. W. Rothwell.)

mesarch protostele to which are attached three ultimate branches (Matten, 1973, 1975). The morphology of the appendages attached to the ultimate branches is unknown, but the dichotomizing, terete traces in the cortex suggest that the appendages were forked close to the insertion point of the branch that produced them. Secondary xylem was produced in the main axis, and the outer cortex is characterized by alternating bands of parenchyma and sclerenchyma. A three-lobed protostele is also seen in Reimannia (FIG. 12.36) (Arnold, 1935b; Matten, 1973). This Middle Devonian plant is sometimes included in the Cladoxylales (Gensel and Andrews, 1984); in other treatments it is placed in the Aneurophytales (Stein, 1982b). Triradioxylon (Barnard and Long, 1975) possesses several features that suggest affinities with the progymnosperms, while other characters are more like those found in seed ferns. The stem contains a three-lobed protostele with protoxylem poles at the ends

of the ribs and near the center of the stele. Secondary xylem tracheids contain circular-bordered pits and uni- to biseriate rays. The trace to a lateral is also three-ribbed and the cortex of the petiole and stem contains alternating parenchyma and fibers. Several of these plants possess a mosaic of characters which makes assignment to major groups of plants difficult. An excellent example is Triradioxylon, which may be classified as an advanced progymnosperm or primitive pteridosperm (Gensel and Andrews, 1984). Cecropsis luculentum is a heterosporous pteridophyte from Upper Pennsylvanian coal balls (Steubenville, Ohio) that is also thought to be a progymnosperm (Stubblefield and Rothwell, 1989). The protostelic woody shoot bears helically arranged leaves and sporophylls (FIG. 12.37) containing adaxially borne clusters of both micro- and megasporangia (FIG. 12.38). Microspores range up to 55 μm in diameter, whereas the megaspores are 500 μm. Although C. luculentum shares

FIGURE 12.35 Diagrammatic reconstruction of Proteokalon pet-

496

paleobotany: the biology and evolution of fossil plants

FIGURE 12.39 John Walton.

FIGURE 12.38 Suggested reconstruction of Cecropsis luculen-

tum. (Redrawn from Stubblefield and Rothwell (1989); in Taylor and Taylor, 1993.)

a number of features found within the three orders of progymnosperms, no single order has the complement of characters found in this Carboniferous plant.

Protopityales The Protopityales are Mississippian in age and represented by a single genus containing two species, Protopitys buchiana and

P. scotica (Walton, 1957) (FIG. 12.39). Although the plants are not well known, they are included as an order of progymnosperms based on the presence of gymnospermous wood, a bifacial cambium, and pteridophytic reproduction. Nothing is known about the ultimate, photosynthetic appendages. In cross section the axes contain an elliptical pith with a pair of endarch primary vascular bundles located at either end. Secondary xylem tracheids in P. buchiana have scalariform bordered pits, while in P. scotica they are described as multiseriate circular bordered (Smith, 1962). Vascular rays are uniseriate. Longitudinal sections of Protopitys sp. from the Tournaisian (Lower Mississippian) of Thuringia, Germany, display secondary xylem tracheids with long, narrow tapering end walls suggestive of apical intrusive growth (Decombeix et al., 2005). Roots associated with a stem of P. scotica are diarch and lack secondary xylem (Walton, 1969). In P. scotica from the Tournaisian of Scotland, the pitting of the metaxylem tracheids is scalariform, while secondary xylem tracheids have transversely elongate to multiseriate bordered pits (Galtier and Scott, 1990). Specimens of P. buchiana have also been reported from the Viséan (Middle Mississippian) of the Vosges region of France, and these typically have uniseriate, ellipticalscalariform bordered pits; there is no evidence of multiseriate, circular bordered pits (Galtier et al., 1998a). Although there are no true growth rings, i.e., cessation of cambial activity, followed by reactivation of growth (Chapter 7), tracheid radial

CHAPTER 12

Progymnosperms

497

T

T

FIGURE 12.40 Longitudinal section of Protopitys buchiana

wood showing cells filled with tyloses (T) (Mississippian). Bar  100 μm. (Courtesy S. E. Scheckler.)

diameter varies in parts of the wood, suggesting a slowdown in growth, perhaps due to wet-dry cycles. Scheckler and Galtier (2003) demonstrate the presence of tyloses (FIG. 12.40) in the secondary xylem near these growth layer boundaries in P. buchiana. In modern woods tyloses are formed by xylem parenchyma cells which bulge through vessel or tracheid pits and subsequently fill xylem conducting cells (FIG. 7.10). They are thought to form mostly during dormancy, when tracheary elements are under no transpirational tension or have cavitated; thus, there is little to no difference in cell osmotic pressure between the parenchyma cells and the tracheary elements (Scheckler and Galtier, 2003). The discovery of tyloses in progymnosperm wood indicates that this type of ecophysiological response evolved early in the evolution of woody plants. Scheckler and Galtier (2003) were able to show that the relationship of structure to function in these fossils is the same as it is in extant woody dicots. This is an important discovery and it will be interesting to see if other detailed anatomical studies reveal additional signals that can be used to investigate some of the physiological processes in ancient plants.

FIGURE 12.41 Fertile region of Protopitys scotica (Mississippian). Bar  6 mm. (From Walton, 1957.)

The fertile organs of Protopitys are not well known. They consist of sporangia borne terminally on twice-dichotomized branching systems (FIG. 12.41) (Walton, 1957). Sporangia are elongate and apparently dehisced longitudinally. The trilete spores of P. scotica range from 75 to 355 μm in diameter and are associated with an extraexinous membrane, perhaps organized in the form of a pseudosaccus (Smith, 1962). The large size range of the spores suggests that heterospory may have been present in the Protopityales.

Noeggerathians There are some groups of fossil plants that do not fit conveniently into an overall classification scheme despite the fact that some details about the reproductive organs are known. One of these is the noeggerathians, sometimes also called the Noeggerathiales, a group of sporangiate plants that can be traced from the latest Mississippian into the Triassic. Historically they have been included in the cycads, sphenophytes, various ferns, and the progymnosperms. Max Hirmer (FIG. 12.42) and Paul Guthörl were perhaps the first to suggest some relationship existed between the noeggerathians and Archaeopteris (Hirmer and Guthörl, 1940). As Archaeopteris was better understood, Beck (1981) also supported this relationship. Despite the discovery of additional specimens (e.g., DiMichele et al., 2004), the group remains highly artificial. We have tentatively included the noeggerathians with the progymnosperms based on the two-ranked laminate sporophylls, organization of the reproductive organs with adaxially borne sporangia, and degree of heterospory exhibited by both groups of plants.

498

paleobotany: the biology and evolution of fossil plants

FIGURE 12.42 Max Hirmer. (Courtesy Hirmer Verlag.)

Noeggerathia (Pennsylvanian–Early Permian) includes leafy shoots bearing two opposite rows of leaves (FIG. 12.43) (Sternberg, 1822). Leaves are obovate and obliquely attached to the stems. The foliar organs may be several centimeters in length and are vascularized by numerous dichotomizing veins that terminate near the margin. Reproductive organs are produced at the ends of vegetative branches and have been termed strobili. The basic organization consists of semicircular sporophylls that are arranged in two rows (biseriate) on either side of the cone axis; sporophylls are decurrent where they attached to the axis (Remy and Remy, 1956) and sporangia are borne on the adaxial surfaces of these disks in rows. At least some cones are interpreted as having sterile regions along the axis (Šetlík, 1956). In Noeggerathiaestrobus vicinalis (Noeggerathiostrobus vicinalis), two types of spores are known (FIG. 12.44). Microspores and megaspores are both trilete; megaspores are up to 670 μm in diameter. In N. bohemicus (FIG. 12.45) from the Pennsylvanian of Bohemia, Czech Republic, microspores are 35 μm in diameter, while megaspores are in the 700-μm-size range (Šimu˚ nek and Bek, 2003). Although the cones are presumed to be bisporangiate, the arrangement of the two sporangial types within the cone is not well known.

12.43 Leafy shoot of Noeggerathia intermedia (Pennsylvanian). Bar  5 cm (Courtesy Z. Šimu˚ nek.)

FIGURE

Noeggerathia foliosa is interpreted as being a small plant with a rosette of leaves and cones borne at the distal end of the stem (Šimu˚ nek and Bek, 2003). Discinites is perhaps the best-known bisporangiate cone that is often included in the noeggerthians (FIG. 12.46) (Bek and Šimu˚ nek, 2005). Specimens contain whorls of sporophylls borne at closely spaced intervals along an axis (Mamay, 1954a). The margin of each sporophyll is dissected into elongated, tapering segments. Some cone sporophylls, like those in D. baculiformis (Permian of China), are more than 30 cm long and contain 80–90 stalked sporangia per sporophyll (J. Wang

CHAPTER 12

Progymnosperms

499

12.44 Noeggerathiostrobus vicinalis microspore (Devonian). Bar  15 μm. (From Remy and Remy, 1956.)

FIGURE

et al., 2004a). Spores in this species range from 50–80 μm in diameter. Leaves with alternate to subopposite pinnules, each about 5 mm long, that have been described as Yuania occur in the same rocks. The ultimate segments (pinnules) of Yuania have parallel venation; stomata have not been observed. Other species of Discinites are clearly heterosporous (FIGS. 12.47, 12.48) (J. Wang, 2005). Dispersed spore genera recovered from Discinites cones include Calamospora, Punctatisporites, Cyclogranisporites, and Deltoidospora (J. Wang, 2005). Operculate spores assigned to Discinispora were discovered in petrified Discinites sinensis cones, and these are reminiscent of spore morphology in certain sphenophyte cones (J. Wang et al., 2004b). Isolated flabelliform sporophylls from the Namurian A (uppermost Mississippian) of Germany and Belgium have been described as Archaeonoeggerathia gothanii (Remy and Remy, 1986; Leary, 1988). The sporophylls are up to 19 mm wide and 20 mm long, with up to 30 adaxially attached sporangia. They represent the earliest bona fide fossil evidence for the Noeggerathiales, but differ from geologically younger forms by the presence of distinct furrows on the sporophyll lamina. There are several other morphogenera used for noeggerathialean cones and sporophylls, such as Lacoea, Heninia, and Tongshania, but most of these have been variously interpreted as synonyms of each other or of Discinites (e.g., Stockmans and Willière, 1962; Leary, 1973; reviewed in Bek and Šimu˚ nek, 2005). Large pinnate or bipinnate fronds of Russellites (FIG. 12.49), sometimes named Yuania, which morphologically resemble leaves of cycads and bennettitaleans, are often

FIGURE 12.45 Noeggerathiaestrobus bohemicus (Pennsylva-

nian). Bar  5 cm. (Courtesy Z. Šimu˚ nek.)

found associated with cones of the Discinites type. Some leaves have pinnae up to 11.5 cm long (Mamay, 1968). Pinnae possess clasping bases and parallel veins that rarely dichotomize or anastomose. Nothing is known about the cuticular anatomy. Mamay (1968) compares Russellites to the noeggerathiopsid foliage genera Noeggerathia and Plagiozamites, and suggests that the occurrence of Russellites and Discinites at the same locality provides evidence that they are organs of the same plant. Another foliage form that bears some resemblance to the noeggerathialeans is Charliea from the Pennsylvanian of Utah and New Mexico, southwestern United States (for details, see Chapter 16).

500

paleobotany: the biology and evolution of fossil plants

FIGURE 12.48 Megaspore of Discinites (Pennsylvanian). Bar  120 μm. (From Remy and Remy, 1956.)

FIGURE 12.46 Cones of Discinites bohemicus (Pennsylvanian). Bar  5 cm. (Courtesy Z. Šimu˚ nek.)

FIGURE 12.49 Frond fragment of Russellites taeniata

(Permian). Bar  2 cm. (From Mamay, 1968.)

FIGURE 12.47 Microspore of Discinites (Pennsylvanian). Bar  10 μm. (From Remy and Remy, 1956.)

Tingia is a foliage type similar to Noeggerathia, but this genus has longer, anisophyllous leaves borne in four vertical rows (Halle, 1925; Kon’no and Asama, 1951). Leaves on the upper surface (FIG. 12.50) of the axis are large

and decurrent at the base, whereas those on the lower surface are narrow and less variable in morphology. Leaves of T. elegans from the Lower Permian of China are 10 mm long with rounded, denticulate apices (Gao and Thomas, 1987). Venation consists of a single bundle that enters the leaf base and then dichotomizes several times within a short distance to produce numerous parallel veins. Reproductive

CHAPTER 12

FIGURE 12.50 Tingia carbonica (Permian). Bar  1 cm. (Courtesy H. Kerp.)

organs (cones) attached to leaves of the Tingia type are called Tingiostachys (Kon’no, 1929). The cones dichotomize once at the base and consist of helically arranged sporophylls. Each sporophyll consists of a pedicel attached to the cone axis at right angles. Sporangia are attached to the adaxial surface in the proximal portion of the pedicel. In T. tetralocularis the cones are approximately 13 cm long and contain spores in the 20 μm size range. Tetraphyllostrobus is a Pennsylvanian fructification from the Sydney Coal Field (Nova Scotia, Canada), which was tentatively assigned to either the noeggerathians or the sphenophytes (Gao and Zodrow, 1990). The cone is 3 cm long and contains decussate bracts, but unlike the other noeggerathians, the spores are described as being monolete.

Progymnosperm evolution Like all other vascular plants, except for the lycopsids, the progymnosperms are thought to have their origin in the trimerophytes (Chapter 8). Since the initial establishment of the group, there has been widespread agreement that the progymnosperms are the ancestral plexus from which the gymnosperms evolved.

Progymnosperms

501

Opinions vary, however, as to whether the gymnosperms all evolved from a common ancestor or can be traced to more than a single ancestor. In a very real sense, the issue revolves around whether the seed habit evolved once or multiple times. One hypothesis suggests that the Aneurophytales are the ancestral progymnosperm group from which the gymnosperms evolved. In particular, it has been suggested that the aneurophytes can be traced to the Mississippian seed ferns, and from one or more of these groups, the remaining gymnosperms evolved (Rothwell, 1982a). Support for this hypothesis includes the discovery of Late Devonian cupulate seeds (Elkinsia) together with vegetative axes that possess the frond architecture of Mississippian pteridosperms (Rothwell and Erwin, 1987; Rothwell et al., 1989; see Chapter 13). The vascular system of these axes is protostelic like that of the aneurophytes, suggesting that these Late Devonian plants may be transitional between progymnosperms and early seed ferns. A different view of the origin of gymnosperms is presented by Beck in a series of papers summarized by Beck and Wight (1988). According to this interpretation, gymnosperms are polyphyletic, with the seed ferns evolving from an aneurophyte ancestor, while the cordaites and conifers had their origin within the Archaeopteridales. While still plausible, this theory is at odds with all phylogenetic analyses that view seed plants as monophyletic. Although these analyses perhaps provide more support for a monophyletic origin for the gymnosperms, evolving the simple leaves and reproductive organs of cordaites and conifers from the compound fronds and cupulate seeds of seed ferns continues to remain a major obstacle. Meyen (1984) has suggested that the gymnosperms are derived initially from the archaeopteridalean progymnosperms. Although the progymnosperms represent a paraphyletic group, they can be used to demonstrate some major trends in organ evolution (Beck and Wight, 1988). The three- or fourribbed stele in the aneurophytes is considered to represent the starting point in the evolution of a eustele via longitudinal dissection of the stele to form the sympodial strands seen in seed plant eusteles (FIG. 7.43) (Namboodiri and Beck, 1968c). Beck and Wight (1988) also suggest that within the aneurophytes there is a trend of increasing complexity of the fertile appendages and planation of the lateral branching systems and ultimate appendages. Despite all of the ideas presented earlier, the ancestral relationships of the major groups of gymnosperms remain elusive. In part this reflects the paucity of information during the Late Devonian–Early Mississippian transition, but also reflects the difficulty of interpreting certain fossils at the earliest grade of gymnosperm evolution. Although it is anecdotal, it seems that there is less paleobotanical work directed

502

paleobotany: the biology and evolution of fossil plants

at the origin of modern gymnosperm groups than in the past. What is needed now is a major breakthrough on the magnitude of Beck’s Archaeopteris–Callixylon discovery, which is now more than four decades old, and which dramatically

changed ideas about gymnosperm evolution. To accomplish this, paleobotanists must search out new collecting sites and utilize skill and patience in reconstructing this new fossil evidence.

13 ORIGIN AND EVOLUTION OF THE SEED HABIT HOMOSPORY, HETEROSPORY, AND THE SEED

CARBONIFEROUS SEEDS ........................................................ 518

HABIT.............................................................................................. 503

Pollen Chamber Function..................................................................523

Homospory....................................................................................... 503

Microgametophytes ...........................................................................524

Heterospory .......................................................................................504

Diversity of Early Seeds....................................................................525

Seed Habit .........................................................................................508

Paleozoic Seeds With Embryos.........................................................526

CUPULATE DEVONIAN SEEDS ...............................................511 Reproductive Biology .......................................................................517

Everything that exists is in a manner the seed of that which will be. Marcus Aurelius, Meditations

The seed habit is the most complex and evolutionarily successful method of sexual reproduction found in vascular plants, and development of the seed habit represents one of the most significant evolutionary events in the history of vascular plants. Today, seed plants are found in a wide variety of habitats owing to the selective advantages that this type of reproduction provides over that found in pteridophytes, including independence from liquid water for fertilization and the capacity for embryo dormancy in a changing environment. The origin and subsequent evolution of the seed habit is a fascinating subject that is well documented in the fossil record, beginning in the Middle to Late Devonian where the first seeds are found. Before considering the earliest seeds, however, it is important to trace the steps involved in the evolution of other reproductive systems that preceded the seed habit.

HOMOSPORY, HETEROSPORY, AND THE SEED HABIT HOMOSPORY

Homospory is regarded as the most primitive reproductive system. The first vascular plants bore sporangia with spores that were all morphologically identical and assumed to have functioned in an identical manner. In living homosporous plants, all the isospores germinate to develop free-living gametophytes, which produce male gametes in antheridia and female gametes in archegonia. This type of development is termed exosporic. Within a population of gametophytes, however, there may be various reproductive patterns, some being the result of environmental influences, which change the timing of sex organ development and thus may affect sex ratios within

503

504

Paleobotany: the biology and evolution of fossil plants

the population. The gametophytes of many homosporous taxa are relatively long-lived structures, capable of photosynthesizing and absorbing water and nutrients through rhizoids. Such gametophytes may contain both types of sex organs (bisexual) or only one type (unisexual). In bisexual gametophytes the sex organs may develop simultaneously, or more commonly sequentially, depending on metabolic resources and the type of environment in which they develop. Gametophytes of vascular plants are rare in the fossil record, and those from the Lower Devonian Rhynie chert are the most completely known (Chapter 8). The Rhynie chert land plants are regarded as homosporous, although some taxa, such as Aglaophyton major, are known to have produced spores that were morphologically identical but which developed into either male or female (dioecious) unisexualfree-living gametophytes (Taylor et al., 2005c). It is believed that each upright gametophyte was developed from a single spore. Although sex organs are segregated on different gametophytes in these fossils, the presence of both types in a population would insure the next generation of new sporophytes. Gametophytes are also known for other Rhynie chert plants, but it is unknown whether each spore produced a unisexual or bisexual gametophyte. To date we simply do not know what was produced when the spores of the earliest land plants, such as Cooksonia, germinated. HETEROSPORY

Heterospory is the production of spores of generally two sizes (microspores and megaspores), each developing into a particular type of gametophyte—either a microgametophyte that produces antheridia and sperm or a megagametophyte that produces archegonia and eggs. A review of the literature dealing with isolated spores (sporae dispersae) recovered from the maceration of Silurian and Devonian rocks suggests that not only were spores of varying sizes present early in the development of a land flora, but these spores also exhibited different shapes and patterns of ornamentation. Chaloner (1967) (FIG. 13.1) plotted the stratigraphic ranges of 74 genera of spores in stages of the Devonian and Silurian and found that there was an increase in the diversity of spores (number of taxa) from the Silurian through the Devonian. In addition, his data show an increase in the size of the spores from the Early to Middle Devonian and a shift to a bimodal distribution of small and large spores by the Emsian (late Early Devonian) in several clades of vascular plants (Chaloner and Pettitt, 1987; FIG. 13.2). The presence of these size classes of spores may indicate the first appearance of heterospory. It must be emphasized that the shift from small spores to large spores was probably not an abrupt

Figure 13.1 William G. Chaloner. (Courtesy P. R. Crane.)

event, but one that gradually evolved, and cladistic analyses suggest that heterospory evolved many times (Bateman and DiMichele, 1994a). Such a shift would also involve a reduction in the number of spores produced per megasporangium, possibly based on nutritional requirements, so that a smaller number of larger spores would be produced. As spore size was changing, there were also developmental and genetic changes taking place that resulted in the segregation of sex organs into separate gametophytes. Unfortunately, the fossil record does not tell us whether the changes in spore size preceded or followed the shift from exosporic to endosporic gametophyte development, but it appears almost certain that the protection afforded by the spore wall in endospory was associated with changes in nutritional aspects of the newly encapsulated megagametophyte. SPORANGIA Traditionally the presence of large and small spores in the fossil record has been used as evidence for the presence of a heterosporous reproductive system. Studies of homosporous and heterosporous extant plants, however, demonstrate that spores may ultimately function in different ways that are not directly linked to size. Heterospory in the fossil record includes intersporangial heterospory, in which there are two types of sporangia, mega- and microsporangia, and intrasporangial heterospory, in which there is a bimodal distribution of spore sizes within the same sporangium. A further complexity in fossils is that intrasporangial heterospory has

CHAPTER 13 ORIGIN AND EVOLUTION OF THE SEED HABIT

505

100 (Westphalian)

10



1 100 Famennian

10 1 100

Frasnian

Number of species: Logarithmic scale

10



1

10

Givetian

1 10

Eifelian

1

Emsian

10 1 10

Siegenian

1 10 1

Gedinnian

2

(Silurian) 200

500

1500 1000 Mean spore size ( )

2000

Figure 13.2 Histogram showing the diversity of spore number compared to spore diameter during the late Paleozoic. (From Chaloner, 1967.)

Figure 13.3 Portion of a fertile axis of Protobarinophyton obrutschevii (Devonian). Bar  1 cm. (Courtesy H. Kerp.)

not yet been correlated with the production of two types of gametophytes. There are numerous instances of intersporangial heterospory in the fossil record. Examples that are pertinent to the seed habit will be discussed below, but heterospory also occurs in the Lycophyta (Chapter 9) and Sphenophyta (Chapter 10), as well as in fossil and modern aquatic ferns (Chapter 11). In the heterosporous lycopsids (Chapter 9), intersporangial heterospory includes monosporangiate cones such as Lepidocarpon (megaspores only) and Lepidostrobus (microspores), as well as bisporangiate cones such as Flemingites schopfii, in which the megaspores produce cellular megagametophytes with archegonia and rhizoids.

In a few fossil examples, different-sized spores occur in the same sporangium, with each size class presumably producing a different type of gametophyte—microgametophyte or megagametophyte. This intrasporangial heterospory or anisospory is interpreted as an intermediate stage leading to complete heterospory, that is, segregation into separate mega- and microsporangia (Taylor and Brauer, 1983; Bateman and DiMichele, 1994a). Even when there is a clear-cut morphological and size difference among the spores within a single fossil sporangium, as in Protobarinophyton (FIG. 13.3) and Barinophyton (Chapter 9), it is difficult to determine whether the small spores functioned in the production of microgametophytes or are merely aborted large spores (Taylor and Brauer, 1983; Cichan

506

Paleobotany: the biology and evolution of fossil plants

et al., 1984). For example, in B. citrulliforme each sporangium contains thousands of trilete microspores 50 μm in diameter (FIG. 13.4) and around 30 trilete megaspores (FIG. 13.5), which are up to 900 μm in diameter. In B. richardsonii, a thin, wrinkled membrane (possibly a perine) is present on some spores (Pettitt, 1965). Anisospory also appears to exist in Protobarinophyton pennsylvanicum from the Upper Devonian of New York (Brauer, 1981). The small spores are 30–42 μm in diameter and large ones 410–560 μm. In both Barinophyton and Protobarinophyton the ultrastructure of the large and small spores is different (FIG. 13.6), suggesting that the small spores do not merely represent aborted megaspores and that these two plants were, in fact, biologically heterosporous (Taylor and Brauer, 1983; Cichan et al., 1984). Spores of two different sizes have also been reported in the same sporangium of Archaeopteris (Medyanik, 1982), and in A. latifolia the number of megaspores per sporangium is quite variable (8–20) suggesting possible intrasporangial heterospory (Chaloner and Pettitt, 1987). Another Devonian (Emsian) plant that appears to represent an early stage in the evolution of heterospory is Chaleuria (Andrews et al., 1974). The sporangia of this plant contain either small (30–48 μm) or large (60–156 μm) spores or, in a few instances, both types. Although all spores are trilete, the larger ones are circular, whereas the small ones are triangular. The selective advantage of heterospory over homospory must have been important in the early evolution of land plants, since heterospory was well established in the lycopsids, progymnosperms, and perhaps other groups as well by the Late Devonian (C.-S. Li et al., 1997). By the Late Devonian, some plants were producing megaspores that exceeded 1 mm in diameter. One of these is Cystosporites devonicus (FIG. 13.7), a tetrad of megaspores from the Escuminac Formation of Canada (Chaloner and

Pettitt, 1964). The tetrad consists of a large, saclike spore, 2.2 mm long, with a trilete mark on the proximal surface (FIG. 13.8) and three smaller, presumably abortive spores that are 100 μm in diameter and positioned so that they sometimes obscure the haptotypic mark. The sporoderm of the functional spore is 15 μm thick and relatively smooth; extending from the distal pole is a stalklike process. Ornamentation of

Figure 13.5 Megaspores (Devonian). Bar  500 μm.

of

Barinophyton

citrulliforme

M

13.4 Two Barinophyton citrulliforme megaspores surrounded by numerous small spores in the same sporangium (Devonian). Bar  25 μm.

Figure

Figure 13.6 Ultrathin section of Barinophyton citrulliforme

megaspore wall (M) and microspores on upper surface (Devonian). Bar  5 μm. (From T. Taylor and Brauer, 1983.)

CHAPTER 13 ORIGIN AND EVOLUTION OF THE SEED HABIT

the abortive spores consists of conical projections. The size of the C. devonicus tetrad and presence of a delicate membrane (?tapetal membrane) surrounding the four spores suggest that each tetrad represents the contents of a single sporangium.

507

Nothing is known about the nature of the sporangium that produced C. devonicus, although the presence of a single functional megaspore and three smaller abortive spores parallels the organization in some Carboniferous lycopsids. For example, each megasporangium of Lepidocarpon (FIG. 9.70) and Achlamydocarpon contains a single, large, functional megaspore (FIG. 9.76) more than 1 cm long and three smaller abortive spores attached to the proximal surface. When found dispersed, these megaspores are assigned to Cystosporites. Surrounding the megasporangium in these forms are various outgrowths of the megasporophyll that some have likened to the integuments of seed plants (see Chapter 9). ENDOSPORY Spore size alone is only one stage in the evolution of heterosporous reproductive systems. Encapsulation and retention of the megagametophyte, endospory, is also critical and has been interpreted as evolving rapidly (DiMichele et al., 1989). An intermediate morphological stage in the evolution of endospory might be similar to the condition in some species of extant Selaginella (FIG. 13.9) or the fossil Flemingites schopfii, in which megagametophyte tissue containing chloroplasts and rhizoids protrudes from the trilete suture of the megaspore, demonstrating that the megagametophyte had some photosynthetic potential. In extant heterosporous plants

Figure 13.7 Megaspore of Cystosporites devonicus with three

aborted spores (arrow) at proximal end (Devonian). Bar  100 μm. (Courtesy C. B. Beck.)

Figure 13.8 Proximal surface of Cystosporites devonicus spore showing small trilete suture (arrow) (Devonian). Bar  200 μm. (Courtesy C. B. Beck.)

Figure 13.9 Longitudinal section of Selaginella cone showing megaspores in sporangia on the right and microspores in sporangia on the left (Extant). Bar  650 μm.

508

Paleobotany: the biology and evolution of fossil plants

that are completely endosporic, the megagametophyte typically develops inside the large spores and the microgametophyte inside the small spores. Although complete details about the microgametophyte development in fossil heterosporous taxa are not well known, there have been several reports of exceptionally well-preserved megagametophytes in many of the major groups of fossil vascular plants. Some of these have been discussed in earlier sections, especially those dealing with arborescent lycopsids (Chapter 9). One of the functional differences between homosporous and heterosporous systems is the selective advantage that results from endosporic development. The fact that gametophytes become encapsulated means that the control of reproduction is transferred to the sporophyte in the development of spores (Bateman and DiMichele, 1994a). Further innovations in heterosporous reproductive strategies may be represented by spore size and ornamentation patterns like those seen in dispersed spores thought to have been produced by archaeopteridaleans (Marshall, 1996). Endosporic development in a large spore also means an increase in stored metabolites, which would lengthen the potential viability of a spore as well as the length of time the spore could survive until water was available to complete the life cycle. This adaptive strategy would be important in environments that were periodically dry. Many early land plants were clonal and appear to have reproduced vegetatively, for example, by fragmentation, and this strategy was also important in the colonization of new, potentially more disturbed habitats. LYCOPSID HETEROSPORY The lycopsids encompass a variety of reproductive strategies ranging from homospory to heterospory, and in some cases evolved reproductive structures that in many ways were functionally similar to seeds (Bateman, 1996b). Although there is no apparent modification of the distal end of the megasporangium into a microspore-receiving structure, the megasporangiate units in Lepidocarpon and Achlamydocarpon apparently functioned like seeds in that the megasporophyll and megasporangium were shed as a unit. It has been suggested, however, that the megaspore in these lycopsids, unlike the megaspore of a seed, was shed from the sporangium at maturity in the typical free-sporing manner. This assumption has been based on the occurrence of numerous isolated Cystosporites megaspores in coals, which may simply be related to preservation. The very resistant nature of the megaspore wall could have prevented alteration during diagenesis, whereas tissues of the sporangium were degraded. However, the enormous size of the lycopsid megaspores and the fact that no Lepidocarpon specimens

have ever been found lacking a megaspore seem to argue that the megaspore was not released from the sporangium at maturity. Thus, at least in the case of the arborescent lycopsids, retention of a single large functional megaspore was widespread by the beginning of the Carboniferous and probably evolved much earlier, some time in the Devonian. Both the functional and the abortive megaspores of the arborescent lycopsids possess small trilete sutures, indicating that they were produced in a tetrahedral tetrad. This is of interest because, with only a few exceptions, there appears to have been a distinct shift from tetrahedral to linear tetrads with the establishment of the seed habit. One exception occurs in the lagenostomalean ovule, Conostoma anglogermanicum, which has been described with three aborted spores situated between the trilete arms of the functional megaspore (Schabilion and Brotzman, 1979). The fossil record provides no evidence as to whether megaspore retention preceded or followed the other important changes in the evolution of the seed, such as the reduction to a single functional megaspore per sporangium, the modification of the distal end of the nucellus for pollen capture, and the evolution of integuments and accessory structures such as a cupule. But in spite of these limitations, the continued investigation of Devonian floras has provided important clues in documenting what might be theorized as transitional stages between homospory, heterospory, and the seed habit. SEED HABIT

Many regard the origin of the seed habit as the end member of a logical progression that began with homospory, was followed by various forms of heterospory (Sussex, 1966) and culminated in the structure termed the seed (Pettitt, 1970). As noted by several authors, however, there is no a priori reason that the seed habit could not have evolved directly from a homosporous system (Thomson, 1927; DiMichele et al., 1989). There is even less agreement as to when these events took place, and what developmental and structural modifications were necessary for seeds and the seed habit to evolve (Haig and Westoby, 1989; DiMichele et al., 1989). Evidence from dispersed megaspores interpreted as seed megaspores suggests that the seed habit was established at least by the Middle Devonian (Arkhangelskaya and Turnau, 2003; Marshall and Hemsley, 2003; Turnau and Prejbisz, 2006). Morphologically, a seed consists of an indehiscent megasporangium surrounded by one or two sheathing integuments termed the seed coat. In seed plants the megasporangium is called the nucellus and contains a single functional megaspore. In gymnosperms a cellular megagametophyte, which is

CHAPTER 13 ORIGIN AND EVOLUTION OF THE SEED HABIT

haploid or 1n, develops inside the megaspore and serves as a food source for the developing embryo. In angiosperms, the food-storage function of the megagametophyte is lost, but nourishment for the developing embryo is provided by the endosperm, which is generally triploid (3n) or pentaploid (5n), and is the result of a second fusion event (sperm  a non-egg cell). In discussions of fossil plants, the terms seed and ovule are often used interchangeably, since in most cases it is impossible to determine whether or not fertilization has taken place, but technically an ovule is unfertilized and a seed has been fertilized. The evolution of the seed habit involves several structural and functional changes that differ from those in either the homosporous or heterosporous systems. These include the retention of a single, functional megaspore (sometimes called monomegaspory), modification of the distal end of the nucellus (megasporangium) for pollen reception, a reduction in the number of cells in the endosporic gametophyte, and the evolution of protective structures (integuments) surrounding the nucellus. Concomitant with changes in the megagametophyte half of the life cycle, there were similar changes in the microgametophyte, including retention of the microgametophyte within the spore (pollen) wall, a reduction in the number of cells in the microgametophyte, change in structure of the pollen wall, and a change in the position of the aperture on the microspore-pollen grain. All of these changes, in either the mega- or microgametophytes, did not take place at the same time. EVOLUTION OF THE INTEGUMENT The morphologic unit that has received the most attention in discussions on the evolution of the seed is the integument. Several theories about the origin of the integument are of historical interest and worthy of note in light of current paleobotanical evidence. The synangial hypothesis was based on fossil evidence and was proposed by Benson (1904) (FIG. 13.10), who believed that the integument evolved from the sterilization of the outer ring of sporangia in a radial synangium. According to this theory, the number of spores in the central sporangium was reduced to one, which ultimately produced the female gametophyte. Other early ideas hypothesized the evolution of the integument from a cuplike indusium (Oliver and Scott, 1904) or from some portion of a frond or a leaf that became enrolled (Walton, 1953a). The coenopterid fern megasporangium Stauropteris burntislandica (Chapter 11) has been used to support a theory termed the nucellar modification concept (Andrews, 1961). According to this idea, the initial stage in the evolution of a seed was a reduction in the number of megaspores from two

Figure 13.10

509

Margaret Benson.

(the number usually found in S. burntislandica) to one. To evolve a simple seed from the S. burntislandica, megasporangium would require a further proliferation of the sporangial wall (increase in cellularization) and the subsequent division of the basal vascular bundle to extend into the newly formed integument. The existence of many Paleozoic seeds in which the integument and nucellus are adnate, such as Lagenostoma and Conostoma (Chapter 14), lends some support to this idea. A similar origin of the integuments was proposed by Walton (1953b), who envisioned a single spore in a young sporangium enlarging and becoming embedded in the pedicel of the sporangium. This concept derives the integuments via a developmental mechanism and was initially proposed to explain the presence of vascular tissue in the nucellus of some Paleozoic seeds. One of the most attractive hypotheses used to explain the origin of integuments is the telome concept (Zimmermann, 1938, 1952) (FIG. 13.11). It is interesting to note that this theory was suggested many years before there was an adequate fossil record of early seeds. The telome theory envisions a dichotomously branched axial system bearing terminal sporangia as the evolutionary starting point (FIG. 13.12A, B). The ultimate axes are termed telomes, either sterile or fertile, and intermediate axes are mesomes. Through time there is a gradual reduction of some of the axes so that a single sporangium becomes surrounded by an aggregation of sterile processes (FIG. 13.12C). A number of Devonian–Mississippian seeds exhibit some of the structural modifications (FIG. 13.12D) suggested by the telome theory,

510

Paleobotany: the biology and evolution of fossil plants

Figure 13.11 Walter Zimmermann.

as the earliest integuments are formed of unfused axes surrounding the nucellus (see later). Both the telomic and the nucellar modification theories appear, in light of the available fossil evidence, to provide the best explanation of how the seed coat or integument may have evolved in gymnosperms (Smith, 1964). The most obvious adaptive significance of an integument appears to be an increased protection of the megagametophyte and embryo, but the free integumentary lobes of some of the earliest seeds may also have served to increase pollen capture. EVOLUTION OF POLLEN CAPTURE Another important innovation in the evolution of the seed habit, perhaps the most important, was a method to capture pollen at the distal end of the nucellus (megasporangium). In many but not all early seeds, the pollen-receiving mechanism appears to be a hollow pollen chamber formed by the distal end of the nucellus, which is surmounted by a funnellike projection termed the salpinx or lagenostome. Inside the pollen chamber is a parenchymatous central column attached to a membranous pollen chamber floor. Hydrasperman reproduction is defined as the combination of pollen capture by the elaborated distal end of the nucellus with the presence of pollen grains that exhibit a proximal trilete suture where the

Figure 13.12 Suggested stages in the evolution of the integumentary system. A. Telome system with terminal sporangia. B. Single sporangium becomes enclosed by other telomes. C. Single sporangium becomes more enclosed by telomes. D. Outer telomes fuse to form integument. (From Andrews, 1961, in Taylor and Taylor, 1993.)

grain breaks open to release the microgametophyte or sperm inside (see section “Pollen”) (Rothwell, 1986). Although initially defined as a structural and morphological feature, hydrasperman reproduction now includes a functional aspect as well. Evidence suggests that once pollen grains filtered into the pollen chamber, growth of the cellular megagametophyte pushed the central column into the narrowed base of the lagenostome to effectively seal the chamber. As hydrasperman reproduction is currently interpreted (Rothwell and

CHAPTER 13 ORIGIN AND EVOLUTION OF THE SEED HABIT

Scheckler, 1988), sealing the pollen chamber may also have functioned to rupture the pollen chamber floor, thus providing a passage for sperm to enter the archegonium below. In these early seeds the integument consists of branches (telomes) that are loosely arranged around the apex of the nucellus, and because the integumentary lobes are generally free at the apex, there is no well-defined micropyle in the integument as occurs in later seeds. Because both the pollen and the ovular structures differ from those typically found in younger and extant gymnosperms, the terms preovule and prepollen are often used in plants with hydrasperman reproduction. Not all paleobotanists agree with the use of these terms, as these structures clearly functioned as ovules and pollen in the life cycle of a seed plant, even though structurally they represent early, more primitive forms. It has been suggested that the open morphology of the integumentary lobes in many of the early seeds indicates that pollination in these plants was abiotic and most certainly anemophilous (via wind). Based on models of several Mississippian seeds, Niklas (1981b, 1983, 1985) argued that the loosely organized integumentary lobes would serve to deflect wind-borne pollen to the funnel of the lagenostome, by creating turbulent flow which would thus decrease pollen velocity and cause settling. Although assigning functional roles based on structural features can be problematic, there appears to be another pollen-receiving system in early seeds that consists of a solid parenchymatous core arising from the distal end of the nucellus. Exactly how this structure may have functioned is not known, but it is interesting that this apparent pollen capturing mechanism is found in the earliest suggested seed Runcaria heinzelinii (Gerrienne et al., 2004). POLLEN It is interesting that pollen grains extracted from Late Devonian and Mississippian pollen organs differ little, either in their morphology or wall ultrastructure, from the isospores of homosporous plants or the microspores of certain heterosporous plants. The concept of prepollen is one that is ingrained in paleobotanical literature; the term was initially coined by Renault (1896a) to describe cordaite and seed fern pollen that he interpreted as containing a multicellular microgametophyte. Since that time the concept has been elaborated to include pollen grains that are believed to have germinated from the proximal surface, rather than from the distal surface as is typical in modern gymnosperms (Poort et al., 1996). Some fossil pollen grains do offer exceptions, where germination might have taken place from either the proximal or the distal surface (Taylor and Daghlian, 1980; Gomankov, 2000). It is generally assumed that prepollen grains did not produce

511

pollen tubes from the distal surface (Poort and Veld, 1997), but rather the flagellated gametes were released directly into the pollen chamber. Ultrastructural evidence from the majority of fossil pollen grains of early seed plants indicates that gametes were indeed liberated from a preformed proximal aperture; whether or not a pollen tube was produced is simply not known. Whether one uses the term pollen or prepollen, the functional significance of the structure is the same— transfer of male gametes to the egg cell (or to the region of the archegonium). The term prepollen has little significance if one is dealing with fossil seed plants because later stages in the development of the microgametophyte are not known. In this context, we will refer to the container in which the male gametes are transferred to the megagametophyte in seed plants simply as pollen. CUPULES Many of the earliest seeds are surrounded by an accessory structure formed of second-order branches that may be fused to some degree and may contain one to several ovules. Such structures have been termed cupules, as this term does not imply homology with other extraovular structures, such as arils and carpels. The cupule segments are often slightly swollen at their base and narrow distally; in many taxa, the ovules are borne in the basal regions of the cupule (Rothwell and Scheckler, 1988). In Moresnetia, for example, the branching system consists of a series of overtopped units that exhibit an open geometry, whereas in others the cupule lobes are arranged so as to suggest that they may have also directed pollen flow. Some cupules may have abscised from the plant as a natural process of development (Pettitt and Beck, 1968), whereas in others the cupules and fertile branches appear to have become disarticulated during dispersal (Rothwell et al., 1989). The most obvious explanation for the function of the cupule is one of protection. Another is that the cupule was an attractant structure associated with an early entomophilous pollination syndrome. Proponents of this idea suggest that the various glandlike structures that sometimes ornament the cupule surface may have functioned to attract potential pollinators or served to discourage predation.

CUPULATE DEVONIAN SEEDS To date the oldest structure described as an ovule or a seed precursor is Runcaria heinzelinii, an integumented megasporangium from the late Middle Devonian (middle Givetian) of Belgium. It is about 8 mm long and has a four-parted, cupshaped cupule at the base (Gerrienne et al., 2004). What

512

Paleobotany: the biology and evolution of fossil plants

is interpreted as an integument consists of up to 16 lobes that are twisted in a spiral pattern (FIG. 13.13); the nucellus extends above the integument at the distal end, and this extension is interpreted as the pollen-receiving device. Although R. heinzelinii was initially labeled as a preovule, but because it apparently does not have hydrasperman reproduction, the authors proposed the term proto-ovule for this structure. They define proto-ovules as the megasporangia of stem groups which are more complex than those of heterosporous plants but do not exhibit the features of hydrasperman reproduction (Gerrienne and Meyer-Berthaud, 2007). Biologically, however, Runcaria is assumed to have functioned as an ovule. Elkinsia polymorpha, an early seed plant from the Late Devonian (Famennian 2c) of West Virginia (Rothwell et al., 1989), consists of loose aggregations of cupules produced on a cruciately forked branching system (FIG. 13.14). Cupules occur either singly (FIG. 13.15) or in pairs, with each cupule containing 16 sterile branch tips surrounding a total of four orthotropous ovules (FIG. 13.16). Ovules are isodiametric in

cross section and up to 7 mm long with the integument consisting of four to five lobes fused to one another only in the basal region. Each integumentary lobe is vascularized by a small terete strand. In the distal region the integumentary lobes are widely separated and surround a cellularized extension of the nucellus that presumably functioned as a pollenreceiving mechanism (FIG. 13.17). Each ovule contained a large, functional megaspore, as well as smaller aborted megaspores. Moresnetia (Fig. 13.18) is another Famennian plant that produced cupules much like those of Elkinsia (Fig. 13.17), although in the initial description the cupules were interpreted as leaves (Stockmans, 1948) (FIG. 13.19). In M. zalesskyi from Belgium, each cupule contains up to four ovules. Individual ovules were organized much like those of Elkinsia, with 8–10 free integumentary lobes (FaironDemaret and Scheckler, 1987). Each ovule is 4 mm long

Figure 13.14 Suggested reconstruction of Elkinsia polymorFigure 13.13 Suggested reconstruction of Runcaria heinzelinii

(Devonian). (From Gerrienne et al., 2004.)

pha showing helically arranged dimorphic fronds. (From Serbet and Rothwell, 1992.)

CHAPTER 13 ORIGIN AND EVOLUTION OF THE SEED HABIT

513

and contains a single functional megaspore and three aborted megaspores. The cupule-bearing branch system is reconstructed as a forked axis giving rise to successive dichotomies (FIG. 13.20), which extend beyond the clusters of cupules (Fairon-Demaret and Scheckler, 1987). The stem is a three-ribbed protostele with C-shaped bundles produced in a helical pattern, with terete strands vascularizing the fertile axes (Prestianni et al., 2007). A slightly younger seed is Xenotheca devonica described from North Devon, southern England (Famennian 2d– Tournaisian 1b) (Arber and Goode, 1915; Hilton and Edwards, 1999), and the Upper Devonian Old Red Sandstone near Cardiff, South Wales (Hilton, 2006 and references cited

Figure 13.16 Diagrammatic reconstruction of Elkinsia polymorpha cupule system and seed. (From Serbet and Rothwell, 1992.)

I

pc

m Figure 13.17 Longitudinal section of Elkinsia polymorpha

of Elkinsia polymorpha (Devonian). Bar  3 mm. (Courtesy G. W. Rothwell.) Figure

13.15 Detail

cupule

ovule showing integument (I), pollen chamber (PC), and megaspore (M) (Devonian). Bar  750 μm. (Courtesy R. Serbet.)

514

Paleobotany: the biology and evolution of fossil plants

Figure 13.18 Suggested reconstruction of Moresnetia zalesskyi

(Devonian). (From Fairon-Demaret and Scheckler, 1987.)

Figure 13.19 Detail of Moresnetia cupule (Devonian). Bar  1.5 mm. (Courtesy P. Gerrienne.)

Figure 13.21 Diagrammatic reconstruction of Archaeosperma

arnoldii cupule showing four seeds. (From Pettitt and Beck, 1968.)

Figure 13.20 Branching axis of Moresnetia showing position of cupules (Devonian). Bar  8 mm. (Courtesy P. Gerrienne.)

therein). The cupules are asymmetrical with integumentary lobes that are fused only at the base (Fairon-Demaret and Scheckler, 1987). Archaeosperma arnoldii (Pettitt and Beck, 1968) is another cupulate seed from the Famennian 2d (Upper Devonian). It consists of two cupules borne terminally on a dichotomizing branch system (FIG. 13.21). Each cupule is 1.5 cm long and contains two short pedicels, each bearing a small seed, a total of two seeds per cupule. The seeds are flask shaped and 4.2 mm long. The integumentary lobes are fused for much of their length but are free at the apex, where the micropyle would be (FIG. 13.22). Because the

CHAPTER 13 ORIGIN AND EVOLUTION OF THE SEED HABIT

515

Figure 13.23 Five ovules of Pullaritheca longii within cupule wall (Mississippian). Bar  2.5 mm. (From Rothwell and Wight, 1989.) Figure 13.22 Archaeosperma arnoldii seed showing functional megaspore with aborted spores and enclosing integumentary lobes (Devonian). (From Pettitt and Beck, 1968.)

specimens are compressed, it is not possible to determine any histologic features of the integument, nor is preservation sufficient that details of the distal end of the nucellus could be determined (dotted lines, FIG. 13.22). Thus nothing is known about the pollen-receiving mechanism of Archaeosperma. The integument encloses a tetrahedral tetrad of one large, functional megaspore with three smaller, abortive megaspores at the micropylar end. On the proximal surface (micropylar end) of the functional spore is a small trilete suture, and the entire tetrad is covered with a delicate nucellar membrane. Dispersed seed megaspores comparable to those in A. arnoldii are known as early as the Middle Devonian. One example from the Givetian of East Greenland has been described as Spermasporites allenii (Allen, 1972; Marshall and Hemsley, 2003). The fossil record of Paleozoic dispersed seed megaspores has been reviewed by Hemsley (1993). There are other Devonian cupules that contained more than a single seed. One of these is Hydrasperma tenuis, a permineralized Famennian seed (Matten et al., 1975). Each of the symmetrical cupules has 24 lobes, with the number

of seeds per cupule ranging from two to six (Matten et al., 1980a). A larger number of ovules per cupule (up to 16) were reported in H. longii (Matten et al., 1980a). Small (42–50 μm) trilete pollen grains are preserved in the pollen chamber. Cellularization of the megagametophyte proceeded in a centripetal pattern like that in some extant gymnosperms. Rothwell and Wight (1989) showed that H. tenuis ovules are produced by at least two different kinds of plants. As a result they continue to use the name Hydrasperma for isolated ovules, but proposed Pullaritheca (FIG. 13.23) for ovulate cupules from the Mississippian and Kerryia for ovulate cupules from the Upper Devonian. Recently, Ruxtonia (FIG. 13.24) was proposed for small bilaterally symmetrical ovulate cupules up to 5 mm long from the Mississippian of Australia (Galtier et al., 2007). Although these taxa have Hydrasperma-type ovules, they differ in cupule morphology and other details of ovule anatomy. Matten et al. (1984a) have described three archegonia at the distal end of Hydrasperma (Kerryia) tenuis (FIG. 13.25), a seed that reflects the structural features of hydrasperman reproduction. Although no embryos have been described in any of these early seeds, differences in the integument morphology and the large number of seeds on an individual bedding plane suggest that in some of the early seed plants, for example, Elkinsia, dispersal may have been environmentally controlled

516

Paleobotany: the biology and evolution of fossil plants

Figure 13.25 Longitudinal section of Kerryia mattenii (Mississippian). Bar  1 mm. (From Rothwell and Wight, 1989.)

Figure 13.24 Diagrammatic reconstruction of Ruxtonia minuta. (From Galtier et al., 2007.)

(seasonal) and perhaps simultaneous (Rothwell and Scheckler, 1988). Not all Devonian seeds and seedlike structures were borne in cupules. Taxa without cupules include Aglosperma and Warsteinia, and it has been suggested that these forms may have been adapted for wind pollination (Prestianni, 2005). Almost all Devonian seeds described to date appear to have been radially symmetrical, although this is difficult to ascertain in impression–compression specimens, and some forms, such as Warsteinia paprothii (Rowe, 1992, 1997), appear to have possessed integumentary wings (see also, for example, Higgs and Streel, 1984). A note on seed symmetry is needed here. Seed symmetry generally refers to the shape of the seed in cross section, that is, radially symmetrical seeds are circular in cross section and bilaterally symmetrical or platyspermic ones are flattened or winged in cross section. These simple definitions, however, do not adequately cover the range of seed symmetries that are known (see Rothwell, 1986). One exception to the radial symmetry of early seeds

may be Spermolithus devonicus, a compressed seed from the Famennian of southern Ireland (FIG. 13.26) (Chaloner et al., 1977). The seed appears to be platyspermic (bilaterally symmetrical), with the integument fused all the way to the apex, unlike the free integumentary lobes of other Devonian seeds. Nothing is known about the pollen-receiving mechanism, nor whether there was even a megaspore within the nucellus. Although the possibility exists that the bilateral symmetry of these structures resulted from flattening during diagenesis (Rothwell and Scheckler, 1988), the uniform morphology of the specimens suggests that perhaps they were bilateral when living. If this is accurate, then within Early Devonian gymnosperms, two distinct evolutionary seed types were present: one characterized by cupulate, radial seeds with free integumentary processes, and a second form that is platyspermic with an integument constructed of two flattened units. As Prestianni (2005) has noted, early seeds and seedlike structures exhibit a great deal of diversity and nucellar complexity by the Famennian (Late Devonian). Several forms are known from the Givetian, including Runcaria and Spermasporites and possibly Moresnetia, suggesting that seeds may have arisen earlier than the Middle Devonian.

CHAPTER 13 ORIGIN AND EVOLUTION OF THE SEED HABIT

Figure 13.26 Compressed specimen of Spermolithus devoni-

cus (Devonian). Bar  1 mm. (Courtesy W. G. Chaloner.) REPRODUCTIVE BIOLOGY

The discovery of a number of new Devonian seeds in recent years has provided a more complete understanding of the structural and morphological organization of the seed early in the evolution of the seed habit. In addition, the discovery of permineralized seeds provides the opportunity to investigate some of the features of early seed plant reproductive biology in more detail. For example, hydrasperman reproduction has been considered a plesiomorphic character and used to suggest that the gymnosperms are monophyletic (Doyle and Donoghue, 1986). As we learn more about early seeds, however, this character may not be universal. For example, in Runcaria the pollen-receiving structure appears as a solid column extending from the tip of the nucellus (FIG. 13.13), and in Coumiasperma, a Mississippian (Tournaisian) seed from the Montagne Noire of southern France, the tip of the nucellus is a solid core of parenchyma and there is no welldefined pollen chamber (FIG. 13.27) (Galtier and Rowe, 1989, 1991). In C. remyi the seed is radial in cross section and 6.5 mm long (FIG. 13.28). The integument is composed

517

13.27 Diagrammatic cut-away reconstruction of Coumiasperma remyi showing integumentary lobes and parenchymatous pollen-receiving structure at the distal end of the nucellus (Mississippian). (From Galtier and Rowe, 1989.)

Figure

L

N L

Figure 13.28 Oblique cross section of Coumiasperma remyi showing integumentary lobes (L), nucellus (N), and central, cellular megagametophyte (Mississippian). Bar  5 mm. (Courtesy J. Galtier.)

518

Paleobotany: the biology and evolution of fossil plants

of eight thickened lobes (FIG. 13.29) that are fused to the nucellus only at the base, and curve in toward the center at the apex. Elongate hairs extend from the inner and lateral surface of the lobes. The organization of the integumentary lobes in this seed forms a rudimentary micropyle (FIG. 13.27). Within the nucellus is a large (2.2 mm long) megaspore (FIG. 13.28) and three smaller aborted spores. Galtier and Rowe (1989) suggested that Coumiasperma may have possessed hydrophilous (via water) pollination and dispersal. It might be argued that Coumiasperma represents an immature ovule that would have produced features like those found in Moresnetia and Elkinsia with continued development. The presence of a cellular megagametophyte in Coumiasperma, however, confirms that the seed was mature at the time it was deposited. At the present time, it appears that there were two basic types of reproductive processes in existence early in the evolution of the seed habit—one represented by the elaborated salpinx or lagenostome of hydrasperman reproduction and the other without this method of pollen capture. Does this mean that the seed habit may have arisen more than once in seed plants? If heterospory provides evidence of transitional stages leading to the seed habit, then the answer is yes. One of the sometimes frustrating, yet exciting aspects of paleobotany is the fact that a single specimen can greatly modify prior hypotheses. Our ideas about seed-plant evolution have greatly changed during the last several decades, and it will be interesting to see if subsequent discoveries support more than a single ancestral progymnosperm giving rise to seed plants.

CARBONIFEROUS SEEDS The fossil record indicates that by the Tournaisian (Early Mississippian), the seed habit was well established in several groups. Unfortunately, the evidence is principally in the form of detached seeds that range from a few millimeters to 10 cm in length. There are a few instances in which the seeds are attached, but in the majority of cases information about the parent plant has been inferred by association with other organs or the presence of specialized structures on several organs that have been used to establish affinities. Historically, one of the principal criteria used to classify detached late Paleozoic seeds was symmetry (Seward, 1917) (FIG. 13.30). This system relied on structural and morphological features of the integument and classified seeds into three orders; for the most part, this classification and the orders based upon it are no longer in use. Large, radially symmetrical seeds with the nucellus attached to the integument only at the base were placed in the Trigonocarpales. Some members of this group also had vascular tissue in both the integument and the nucellus. Seeds of this type (e.g., Pachytesta, Stephanospermum, Hexapterospermum) were produced by members of the medullosan seed ferns (Chapter 14). Platyspermic or bilaterally symmetrical seeds with simple pollen chambers (FIG. 13.31) and vascular tissue only in the integument were included in the Cardiocarpales. Many of these seeds (e.g., Cardiocarpus, Mitrospermum) were

Figure 13.29 Cross section of Coumiasperma remyi showing

eight integumentary lobes surrounding distal parenchymatous beak (Mississippian). Bar  3 mm. (Courtesy J. Galtier.)

Figure 13.30 Albert Charles Seward.

CHAPTER 13 ORIGIN AND EVOLUTION OF THE SEED HABIT

produced by cordaites and conifers (Chapters 20, 21). One seed that combines features of both the Trigonocarpales and the Cardiocarpales is Menaisperma, a compression specimen discovered in Middle Mississippian (early Late Viséan) rocks in North Wales (Pettitt and Lacey, 1972). The seed is about 7 mm long, with an entire integument (i.e., completely fused) that is attached to the nucellus only at the base. The nucellar apex consists of three lobes. Stomata are present on the nucellus of M. greenlyii, a feature not found in other Paleozoic seeds, although they have been reported in modern cycad seeds. Menaisperma is interesting in that it occurs during a time period when other ovules exhibited free integumentary lobes and elaborate lagenostomes. The third category of late Paleozoic seeds is the Lagenostomales, the only ordinal name of the three proposed by Seward that is still in widespread use (Chapter 14). Lagenostomalean seeds are typically small, radially symmetrical, and the integument is adnate to the nucellus except at the seed apex (FIGS. 13.32, 13.33). Vascular tissue is confined to the inner portion of the seed coat, and in many, the pollenreceiving lagenostome or salpinx is elaborate (FIG. 13.34). A large number of lagenostomalean seeds were produced in cupules, and many are known from Mississippian deposits. Although this classification system for seeds is highly artificial, it has provided a convenient means of classifying detached seeds and, in some instances, a basis for suggesting

Figure 13.31 Longitudinal section of Tyliosperma orbiculatum showing simple pollen chamber (Pennsylvanian). Bar  1 mm.

519

P

Figure 13.32 Longitudinal section of Physostoma calcaratum.

Note pollen (P) grains in pollen chamber. Arrow indicates micropyle (Pennsylvanian). Bar  1 mm.

Figure 13.33 Cross section of Physostoma calcaratum show-

ing lobed integument (Pennsylvanian). Bar  1 mm.

520

Paleobotany: the biology and evolution of fossil plants

N IL

Figure 13.34 Distal end of Salpingostoma dasu showing

inwardly directed epidermal hairs and salpinx (Mississippian). (From Taylor and Millay, 1979.)

evolutionary trends (Meyen, 1984). Rothwell (1986) has pointed out some of the pitfalls in using symmetry as a diagnostic character in classification. These include the presence of various shapes within specimens assigned to the same taxon and the fact that shape may vary with other factors, such as the developmental level. Symmetry can also be affected by fossilization and subsequent diagenesis, and in some seeds neither radial nor bilateral symmetry is applicable in describing seed shape. In the following section, a number of Mississippian seeds are discussed in relationship to morphological and structural diversity and where appropriate, features that document potential evolutionary stages in the reproductive biology of the ovules are noted. In some classification schemes these seeds would all be included within the Lagenostomales. Only a few of the more important ones are considered here; others will be treated in Chapter 14. One of the simplest Mississippian seeds is Genomosperma kidstonii (FIG. 13.35) (Long, 1960a) (FIG. 13.36). This permineralized seed is known from several localities in

Figure 13.35 Longitudinal section of Genomosperma kidstonii

showing integumentary lobe (IL) and nucellus (N) (Mississippian). Bar  2 mm. (Courtesy Hancock Museum.)

Scotland, all within the famous Calciferous Sandstone series, Berwickshire Cementstone Group (Tournaisian). The seeds are circular in cross section, 1.5 cm long and attached to a small pedicel. The integument consists of 8–11 free lobes that surround the nucellus, and thus there is no micropyle (FIG. 13.37). Each integumentary lobe contains a single,

CHAPTER 13 ORIGIN AND EVOLUTION OF THE SEED HABIT

521

Figure 13.36 Albert G. Long.

Figure 13.37 Suggested reconstruction of Genomosperma kidstonii (Mississippian). (From Taylor and Millay, 1979.)

terete vascular strand. The nucellus and the integument are fused only at the base; distally the nucellus forms a ringshaped pollen chamber with a central parenchymatous core. Small, trilete pollen grains are present in the pollen chambers of some specimens. In another species, G. latens, the integumentary lobes are fused more distally (FIGS. 13.38, 13.39). Although no cupulelike structures have been found associated with these seeds; the large number of specimens and their size suggest that they may have been borne in some form of cupule. Genomosperma kidstonii shares a number of anatomic and morphologic features with several Famennian seeds such as Elkinsia and Moresnetia. The organization of the integumentary lobes and distal end of the nucellus also suggest that G. kidstonii may have possessed hydrasperman reproduction. Several Mississippian seeds can be used to suggest ways in which the evolution of the integument and the

concomitant formation of the micropyle may have evolved. In Conostoma and Lagenostoma, for example, the micropyle is well defined by the apical fusion of the integumentary lobes. In most Mississippian seeds, there is a direct relationship between the degree of integumentary fusion and the organization of the nucellar pollen-receiving mechanism (Andrews, 1963). In seeds with free integumentary lobes, and by definition no micropyle, the pollen-receiving mechanism is often a highly ornate structure (e.g., lagenostome, salpinx) that may extend a considerable distance above the pollen chamber (FIG. 13.34). As has been postulated for seeds with hydrasperman reproduction, this structure probably functioned to direct wind-borne pollen into the pollen chamber. In a seed such as Eurystoma, also from the Calciferous Sandstone Series, the salpinx is conspicuous and up to 1 mm in diameter (FIGS. 13.40, 13.41) (Long, 1960b, 1965). Many seeds

522

Paleobotany: the biology and evolution of fossil plants

S

I

Figure 13.39 Suggested reconstruction of Genomosperma latens (Mississippian). (From Taylor and Taylor, 1979.)

Figure 13.38 Longitudinal section of Genomosperma lat-

ens showing integument (I) and salpinx (S) (Mississippian). Bar  1 mm. (Courtesy Hancock Museum.)

in the Pennsylvanian possess a simple pollen chamber with a small nucellar beak that is directly associated with the micropylar canal, as in modern cycads (FIGS. 13.42) a result of the continued fusion of the integument in these seeds, pollen capture, which was initially accomplished by the free end of the nucellus, became the function of the integument, perhaps via a pollination droplet (Rothwell, 1977a;

Taylor, T., 1982b). The evolutionary pressures for this shift may have included a response to predation, differing pollination syndrome, or the evolution of seed dormancy. It is important to point out that not all early seeds had elaborate pollen-receiving structures. Some, such as Genomosperma latens, lack a highly differentiated nucellus (Long, 1960a) (FIG. 13.38). As these seeds also lack a micropyle, pollen grains may have been trapped by other methods, such as pollination droplets, hairs, or other processes that may have aided in directing pollen flow. In Salpingostoma dasu, hairs lined both the inner and the outer surfaces of the integument lobes (FIG. 13.34) (Gordon, 1941). in Tantallosperma setigera, which is thought to have been produced on stems of Buteoxylon (Buteoxylonales; see Chapter 14), pollen grains have been found associated with the bristlelike processes that extend from the integument (Barnard

CHAPTER 13 ORIGIN AND EVOLUTION OF THE SEED HABIT

523

S

PC

Figure 13.41 Distal end of Eurystoma angulare showing

pollen-receiving mechanism (Mississippian). (From Taylor and Millay, 1979.)

These authors suggested that in several ovules with hydrasperman reproduction, integumentary lobes were open at pollen-receptive stages and perhaps closed once pollination was effected. Addressing problems of this type will require multiple specimens so that developmental stages can be analyzed. POLLEN CHAMBER FUNCTION

Figure 13.40 Longitudinal section of Eurystoma angulare

showing pollen chamber (PC) and salpinx (S) (Mississippian). Bar  2 mm. (Courtesy Hancock Museum.)

and Long, 1973). A slightly different method of pollen capture has been suggested in Physostoma (Mississippian– Pennsylvanian) (Chapter 14). Here the outer surface of the integument is covered with tentaclelike projections which are thought to have produced a sticky substance that aided in pollen capture (FIG. 13.43). It has also been hypothesized that the open or the closed nature of the integumentary lobes in some of the earliest seeds may simply reflect stages in the maturity of the ovules, rather than examples of integument evolution and pollen capture (Rothwell and Scheckler, 1988).

Once pollen entered the pollen chamber in many of these early seeds, the chamber was sealed in one of two ways, both related to growth of the developing megagametophyte. In seeds with hydrasperman reproduction, the central column, which is part of the nucellus, was pushed up into the neck of the lagenostome by the growth of the megagametophyte below. In some seeds with micropyles, the chamber was sealed by the growth of a small cellular pad of tissue termed the tent pole (FIG. 13.44), which is part of the megagametophyte, as occurs in the Pennsylvanian seed Nucellangium. The tent pole is located at the distal end of the megagametophyte within the megaspore wall. As mitotic cell divisions continued, the tent pole not only forced the closure of the pollen chamber, but may also have functioned to rupture the megaspore membrane, thus increasing the opportunity for male gametes to reach the archegonia. It is thought that a similar rupture occurred in hydrasperman-type seeds via growth of the central column. It is apparent that in many of these

524

Paleobotany: the biology and evolution of fossil plants

L

Figure 13.42 Longitudinal section through the apex of a cycad seed showing pollen tube (arrow) digesting cellular nucellus. Note two archegonia (Extant). Bar  2 mm.

Figure 13.44 Longitudinal section of Nucellangium glabrum showing tent pole (arrow) and lagenostome (L) (Pennsylvanian). Bar  3 mm.

early seeds the nucellus (FIG. 13.45) played a pivotal role in both pollination and post-pollination events. In some younger seeds, there is evidence suggesting that the micropyle was sealed by differential growth of the integument, as is the case in some extant gymnosperms. MICROGAMETOPHYTES

Figure 13.43 Distal end of Physostoma elegans showing pollenreceiving mechanism (Pennsylvanian). (From Taylor and Millay, 1979.)

Relatively little is known about the microgametophytes of these early seed plants except that pollen grains have been found in the pollen chambers and micropyles of some seeds (FIG. 13.46). These grains are small and trilete and, as noted above, morphologically resemble the isospores of homosporous plants. It is believed that germination occurred through

CHAPTER 13 ORIGIN AND EVOLUTION OF THE SEED HABIT

525

Pollen chamber of Physostoma calcaratum containing pollen grains (Pennsylvanian). Bar  225 μm.

Figure 13.46

Figure 13.45 Suggested reconstruction of Eurystoma angu-

lare. Note cuplike conspicuous pollen-receiving mechanism (the nucellus) (Mississippian). (From Taylor and Millay, 1979.)

the proximal trilete suture, but it is not known whether germination resulted in the formation of a pollen tube, or whether male gametes or sperm were simply released in the pollen chamber close to the neck of the archegonium. Although there is evidence that some Mississippian seeds continued to produce megaspores in a tetrahedral tetrad based on the presence of a triradiate scar on the distal end of the functional spore, the majority produced megaspores in linear tetrads. With just a few exceptions, e.g., Conostoma, linear tetrads are found in all Pennsylvanian seed plants, as they are in extant gymnosperms. DIVERSITY OF EARLY SEEDS

It is important to point out that not all Mississippian seeds contain the same complement of structures. In some, free integumentary lobes are associated with a relatively unspecialized nucellar apex, whereas in others the nucellar apex

is greatly elaborated. Some seeds show features related to pollination syndromes, such as anemophily (lack of surrounding cupules) or entomophily (presence of glands on the surface of the integument). Some have an integument in which the lobes are almost completely fused but still retain an elaborated nucellar apex. In others the structure and morphology of certain features may represent a developmental stage at the time of fossilization, including the presence of a megagametophyte (FIG. 13.47). One unifying feature of these seeds, however, is the small size (typically 1 cm in length), which has led some to suggest that many may have been borne in a cupule. One Mississippian seed that is distinctly cupulate is Eurystoma angulare from the Calciferous Sandstone Series (Long, 1960b, 1965). The seed measures about 8 mm long and is square in transverse section. At the apex are four vascularized integumentary lobes (FIG. 13.45), each corresponding to an angle of the seed. The nucellus and integument are adnate, except at the distal end, where the nucellar beak of the pollen chamber flares into a bell-shaped structure. The cupule is a dichotomizing system of branches, with the largest number of sterile lobes about 15. The maximum

526

Paleobotany: the biology and evolution of fossil plants

Figure 13.47 Longitudinal section of Deltasperma foulde-

nense with well-developed megagametophyte (Mississippian). Bar  1 mm. (Courtesy Hancock Museum.) Figure 13.48 Suggested reconstruction of Stamnostoma hut-

size of a single cupule system is 1.5 cm long by 1.1 cm in width. A variable number of ovules were produced (2–10), each associated with a sterile member of the cupule at the end of a dichotomy. Eurystoma trigona is triangular shaped, with each seed 4.5 mm long; in one specimen, more than 3000 ovules were found attached to a petrified axis (Long, 1969). The branching system consists of an outer ring of sterile, flattened branches surrounding an inner set of highly dichotomous axes that produced the seeds. The anatomy and the branching have been compared to a young, unrolled Alcicornopteris frond segment, commonly found associated with the seed fern stem Stenomyelon. A similar association has been made between the Mississippian stem Tristichia and seeds and cupules of Stamnostoma huttonense (FIG. 13.48). These cupules consist of dichotomizing axes, each producing four seeds. In S. oliveri (Tournaisian) each

tonense (Mississippian). (From Taylor and Taylor, 1979.)

cupule produces one to three ovules with hydrasperman reproduction (Rothwell and Scott, 1992). Long (1966) suggested that Mississippian cupules may be divided into three general groups based on organization. One type contains numerous ovules borne on what he interpreted as an entire frond segment. In a second type, which is also multiovulate, the cupule is regarded as homologous to part of a frond system. The third type is uniovulate and radially symmetrical. PALEOZOIC SEEDS WITH EMBRYOS

Megagametophyte tissue has been described from the Mississippian seed, Deltasperma (FIG. 13.47) (Long, 1961a), and there are numerous examples of exquisitely preserved

CHAPTER 13 ORIGIN AND EVOLUTION OF THE SEED HABIT

527

E

Figure 13.49 Cross section of the seed Plectilospermum

elliotii showing two embryos (arrows) (Permian). Bar  850 μm.

embryos in the seeds of Mesozoic gymnosperms (Stockey, 1975) and an embryo in the lycophyte Lepidocarpon (Phillips, 1979). There are relatively few reports, however, of embryos in Paleozoic seeds. One of these is from an Early Permian conifer seed collected in Texas. Within the megagametophyte of a small (1.2 cm long), flattened seed, Miller and Brown (1973a) described an aggregation of cells including several tracheids. This structure is histologically different from the surrounding megagametophyte and is interpreted as an embryo. An immature embryo was identified in the seed Nucellangium glabrum from Middle Pennsylvanian coal balls from North America (Stidd and Cosentino, 1976). In this seed the distal end of the megagametophyte contains two clearly differentiated archegonia and a small cellular mass, the one described as a proembryo. Embryos have also been reported in seeds from Permian rocks in Antarctica. In several specimens of Plectilospermum elliotii the distal end of the megagametophyte contains two large archegonial chambers (FIG. 13.49). In each is an elongate-to wedge-shaped proembryo that is estimated to be composed of 50 cells (Smoot and Taylor, 1986b). Attached to one end of the embryo is a coiled suspensor (FIG. 13.50). Multiple embryos in these seeds indicate that two important biological processes had evolved in certain seed plants by the late Paleozoic. One is the presence of a

Figure 13.50 Detail of archegonial chamber of Plectilospermum elliotii seed showing embryo (E) and suspensor (arrow) (Permian). Bar  250 μm.

suspensor, a structure that functions in almost all extant gymnosperm seeds to maintain continuity between the developing embryo and the nutritive tissue of the megagametophyte. The second process as demonstrated is Plectilospermum, is the presence of simple polyembryony, the fertilization, and development of more than one embryo per seed. Information about embryo development in Paleozoic fossil seeds is equally rare. However, Mapes et al. (1989) described well-preserved embryos in conifer seeds from the uppermost Pennsylvanian of North America. Each mature embryo is elongated, with about six cotyledons surrounding the epicotyl. These authors suggested that the presence of cotyledons implies that these plants had a quiescent period prior to germination (seed dormancy). The evolution of seed dormancy was an important physiological adaptation that provided the opportunity to colonize new habitats that may have been periodically dry. One idea that has been offered to explain the paucity of Paleozoic embryos is the suggestion that embryo development was delayed after pollination or fertilization (Arnold, 1938). The possibility also exists that some early seed plants produced embryos that were rapidly formed and escaped the protective integuments. It has also been suggested that Paleozoic embryos have not been properly identified to date.

This page intentionally left blank

14 PALEOZOIC SEED FERNS CALAMOPITYALES ....................................................................... 531

Pollen Organs ....................................................................................581

BUTEOXYLONALES ......................................................................539

Pollen ................................................................................................590 Medullosan Evolution .......................................................................591

LYGINOPTERIDALES ................................................................. 540 CALLISTOPHYTALES ...................................................................593 Lyginopteris Plant .............................................................................540

Vegetative Organs..............................................................................594

Other Lyginopterids: Vegetative Remains .........................................546

Reproductive Structures ....................................................................595

Other Lyginopterids: Seeds and Cupules ..........................................555

Callistophytalean Evolution ..............................................................598

Other Lyginopterids: Pollen Organs..................................................560 Incertae Sedis ....................................................................................563

GLOSSOPTERIDALES ................................................................. 598

Lyginopterid Evolution .....................................................................565

Leaves ...............................................................................................599

MEDULLOSALES .......................................................................... 566

Stems and Roots................................................................................605 Ovulate Reproductive Structures ......................................................606

Stems .................................................................................................566

Pollen Organs ....................................................................................616

Leaves (Fronds) .................................................................................570

Glossopteris Habit and Habitat .........................................................618

Roots .................................................................................................572

Phylogenetic Position ........................................................................618

Growth Habit.....................................................................................572 Seeds .................................................................................................573

Green how I want you green. Green wind. Green branches. Federico García Lorea, Romance Sonámbulo The history of the establishment of the seed ferns (Pteridospermophyta) is an interesting one, which demonstrates the presence of a previously unknown group of vascular plants by correlation of a number of isolated plant organs. In the late 1800s, the famous French paleobotanist F. C. Grand’Eury suggested that various Paleozoic foliage types such as Alethopteris (FIG. 14.1), Neuropteris, and Odontopteris may have been produced on petioles that are today known as Myeloxylon. Some time later, Myeloxylon was demonstrated to be the petiole of Medullosa stems. At about the same time, Stur (1883) suggested that certain foliage types previously regarded as fern leaves may represent leaves of an unknown group of plants, on the basis of the consistent absence of sporangia on the pinnules. In 1887, W. C. Williamson (FIG. 14.2) recognized that the anatomical

features of several stems combined structural characteristics of both cycads and ferns. The studies by these workers were incorporated into the concept of the Cycadofilices by H. Potonié (1899) (FIG. 16.5), who, on the basis of strictly anatomical evidence, suggested the existence of a group of vascular plants that was transitional between ferns and seed plants. Finally, in 1904 two British paleobotanists, F. W. Oliver (FIG. 14.3) and D. H. Scott (FIG. 14.4), established the existence of the seed ferns, or Pteridospermae, by a remarkable piece of detective work that united stem, petiole, foliage, and most importantly seeds of the same plant, Lyginopteris. The initial identification of the group, however, was not based on organic attachment of the parts, but rather on the common occurrence of a peculiar epidermal appendage in the form of a gland on all the

529

530

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.1 Paradermal section of Alethopteris pinnule show-

ing venation pattern (Pennsylvanian). Bar  3 mm.

Figure 14.3 Francis W. Oliver. (Courtesy H. N. Andrews.)

Figure 14.2 William C. Williamson. (Courtesy Gustav Fischer

Verlag.)

plant parts (FIG. 14.5), together with the frequent association of the vegetative remains and seeds at the same fossilbearing localities. Today, the group includes a number of well-defined orders in which both vegetative and reproductive features are known in considerable detail; in others, assignment to the group is largely based on anatomical features of the stems. Although

Figure 14.4 Dukinfield Henry Scott. (Courtesy Gustav Fischer

Verlag.)

CHAPTER 14

Figure 14.5 Suggested reconstruction of Lagenostoma lomaxii showing cupulate lobes with capitate glands (Pennsylvanian). (From Millay and Taylor, 1981a.)

the seed ferns extend into the Mesozoic, the Mesozoic orders (Chapter 15) are not known in as much detail as the Paleozoic forms. Nevertheless, the Mesozoic seed fern groups are especially significant, as many believe the progenitors of the angiosperms can be found in one of these groups. Some pteridosperms have been reconstructed as small trees with upright trunks bearing helically arranged, massive fernlike fronds (Retallack and Dilcher, 1988; Zodrow et al., 2007). Others were probably more similar to lianas or vines and possessed a scrambling or climbing habit (Baxter, 1949; Dunn et al., 2003a; Krings et al., 2003b,c). Various stelar configurations occur in the group, ranging from simple protosteles to eusteles. The wood is generally constructed of parenchyma and thin-walled tracheids that give it a spongy organization; this type of wood is termed manoxylic. Longitudinally oriented bands of sclerenchyma characterize the cortex of many taxa. Both the pollen-bearing structures and seeds were borne on leaves. Seeds were large and solitary in some forms, whereas in others they were small and produced in multiovulate cupules. Pollen organs were aggregated into clusters and in some taxa were arranged into large synangiate organs. The following orders will be discussed: Higher taxa in this chapter:

Calamopityales (Mississippian) Buteoxylonales (Upper Devonian–Mississippian) Lyginopteridales (?Upper Devonian–Carboniferous) Medullosales (Carboniferous–Permian) Callistophytales (Pennsylvanian–Lower Permian) Glossopteridales (Permian–?Triassic)

PALEOZOIC SEED FERNS

531

Figure 14.6 Cross section of Calamopitys americana stem

(Mississippian). Bar  4 mm (From Beck, 1970.)

CALAMOPITYALES The calamopityaleans are an interesting group of predominantly permineralized or petrified plants known from Mississippian rocks throughout North America and Europe. They have been classified with the pteridosperms principally on the basis of manoxylic wood in the stems (Galtier and Meyer-Berthaud, 1989; Galtier and Meyer-Berthaud, 2006). Although some petioles and foliage remains have been (tentatively) assigned to the group (Long, 1964b; Galtier, 1974, 1975; Knaus et al., 2000; Orlova and Snigirevsky, 2003; Orlova, 2007), there is virtually nothing known to date about the seeds and the pollen organs. As originally defined by Solms-Laubach (1896) and subsequently characterized by Lacey (1953) the order includes stems characterized by manoxylic wood, and an unusually thick cortex with complex patterns of cell size and shapes. These patterns have been shown to result from differential cell growth and proliferation of different cell types and tissue regions (Hotton and Stein, 1994). Some calamopityaleans have sparganum cortical sclerenchyma and petiole bases with several to many vascular bundles. A sparganum cortex is characterized by vertically aligned, hypodermal fiber strands that do not anastomose (FIG. 14.13). The type genus of the order is Calamopitys (FIG. 14.6). As emended by Galtier and Meyer-Berthaud (1989), Calamopitys includes slender stems, usually 2–3 cm in diameter (Rowe and Galtier, 1988), that range from

532

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

protostelic with abundant parenchyma to eustelic. Axes 3 cm in diameter are rare, but in C. embergeri they may be as large as 4 cm in diameter (FIG. 14.7) and 5–5.5 cm in C. schweitzeri (FIG. 14.8) from the mid-Tournaisian (Lower Mississippian) of France (Rowe and Galtier, 1988; Galtier et al., 1993). Xylem is mesarch (FIG. 14.9), with the strands either interconnected or discrete. Leaf traces divide in the cortex and may extend for several nodes before entering the petiole. When present, secondary phloem consists of alternating bands of parenchyma and sieve cells. Calamopitys americana (Scott and Jeffrey, 1914; Galtier and Beck, 1995) is a relatively common element in North America and will be used to characterize the type of plant typically encountered in this genus (FIG. 14.6). Some specimens have been found to exceed 4 cm in diameter, but no appreciable stem lengths have been discovered. The central portion of the stem consists of a mixed pith with many more parenchyma cells than tracheids (FIG. 14.6). In transverse section, the stele is roughly triangular and contains five mesarch primary strands. Surrounding the primary body is

Figure 14.8 Calamopitys schweitzeri stem surrounded by petioles (Mississippian). Bar  5 mm. (Courtesy J. Galtier.)

Figure 14.9 Calamopitys schweitzeri stem showing details Figure 14.7 Cross section of Calamopitys embergeri stem

(Mississippian). Bar  5 mm. (Courtesy J. Galtier.)

of pith and primary and secondary xylem (Mississippian). Bar  0.5 mm. (Courtesy J. Galtier.)

CHAPTER 14

a relatively broad zone of radially aligned, secondary xylem tracheids and numerous multiseriate, parenchymatous rays. Rays are 1–12 cells wide and variable in height. The cortex is parenchymatous and contains radial plates of sclerenchyma near the periphery. Petioles are apparently produced in a 2/5 phyllotaxy. Leaf traces are given off from the stele singly and then divide to produce four bundles in the base of the petiole, each lacking secondary wood. Another species C. foerstei, differs in that it possesses widely separated primary xylem sympodia and axial vascular strands of unequal size (Read, 1936b). In the type species, C. saturni (midupper Tournaisian), the protostele contains parenchyma, and the vascular system to the petiole consists of up to 15 leaf traces (Galtier and Meyer-Berthaud, 1989).

Figure 14.10 Cross section of Calamopitys solmsii (Mississippian). Bar  5.0 mm. (Courtesy B. Meyer-Berthaud.)

PALEOZOIC SEED FERNS

533

Not much is known about the growth habit of Calamopitys other than the fact that the stems were slender (FIG. 14.10) and, based on petiole specimens, bore large, bipartite fronds with naked petioles (Galtier, 1974). For C. embergeri specimens from France (Galtier, 1970b), a lianescent (vine-like) growth habit has been suggested based on the stem diameter (2 cm) (FIG. 14.7), the large proportion of soft cortical tissue compared to woody tissue (FIGS. 14.7, 14.11), and large fronds whose petiole bases were characteristically expanded at the point of attachment to the stem (Galtier, 1975, 1986). Larger forms, such as the eustelic C. schweitzeri, may have been upright, self-supporting plants (Galtier et al., 1993). Biomechanical analyses of Calamopitys specimens from Germany suggest that slender stems 1.5 cm in diameter were semi-self-supporting (Rowe et al., 1993). Moreover, they indicate that the outer cortex of alternating bands of fibers and parenchyma was the principal contributor to flexural stiffness, followed by the multifascicular petiole bundles, whereas the xylem core in the center was of relatively minor structural importance. Stenomyelon is considered by many to be the most primitive member of the calamopityaleans; S. tuedianum (FIG. 14.12), the type species of the genus, is known from the Calciferous Sandstone Series of Britain (Kidston and Gwynne-Vaughan, 1912). The stem is beautifully preserved and consists of a solid protostele dissected into nearly equal thirds by thin plates of parenchyma. Metaxylem tracheids exhibit multiseriate bordered pits. Secondary xylem is similar to that of other taxa in the order, differing only in the width of the vascular rays. Petioles are attached in a 2/5 phyllotaxy, with traces containing

C

Figure 14.11 Calamopitys embergeri stem showing detail of

cambium (C) zone and secondary phloem (Mississippian). Bar  0.3 mm. (Courtesy J. Galtier.)

14.12 Cross section of Stenomyelon tuedianum (Mississippian). Bar  1.5 mm.

Figure

534

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.13 Cross section of Stenomyelon petiole showing multiple vascular bundles and sparganum cortex (Mississippian). Bar  2 mm.

Figure 14.14 Cross section of Stenomyelon heterangioides

(Mississippian). Bar  3 mm.

Figure 14.15

Charles B. Read. (Courtesy D. S. Chaney and

W. A. DiMichele.)

two exarch primary xylem groups (FIG. 14.13). The surface of the stem is covered with multicellular projections. A slightly different stelar organization is present in S. heterangioides (FIG. 14.14), in which clusters of parenchyma cells, in addition to plates of parenchyma, divide the protostele (Long, 1964b). Meyer-Berthaud and Stein (1995) regarded S. tuedianum and S. heterangioides as conspecific and present a refined concept of S. tuedianum, which includes a unique pattern of petiolar bundle development and the presence of adventitious roots. Changes from a heterangioides-type anatomy with medullated protostele and mesarch protoxylem strands to a tuedianumtype with a solid protostele and exarch protoxylem are postulated to correspond to apoxogenetic phases of growth. Possibly the simplest vascular arrangement occurs in S. primaevum, another species from Britain. The stele is a trilobed protostele that lacks parenchyma; traces to the petioles divide repeatedly to form about eight mesarch bundles (Long, 1964b). In S. bifasciculare, each rib of the stele gives rise to a pair of leaf traces (Meyer-Berthaud, 1984b). The cortex in all Stenomyelon species contains radial plates of

sclerenchyma. A large compressed stem with several fronds in organic connection from the Tournaisian of Foulden, Scotland, has been suggested to represent a compression equivalent to Stenomyelon (Long, 1964b). Long interpreted the plant as a sprawling shrub bearing large, bipartite fronds. Diichnia is an Early Mississippian stem morphotaxon established by Read (FIG. 14.15) based on nodule material from the uppermost New Albany Shale of Kentucky (Read, 1937). Diichnia kentuckiensis and D. readii are characterized by a pentagonal eustele consisting of five mesarch primary xylem bundles located at the angles of a broad, five-angled, parenchymatous pith containing rare to centrally abundant medullary tracheids (Read, 1936b; Beck et al., 1992). Diichnia has helically arranged leaves, each supplied by two traces (FIG. 14.16) that diverge radially from two adjacent axial bundles in the stele. Axial bundles are typically sinuous, and an accessory bundle is produced below the level of leaf trace divergence, which then fuses to the axial bundle above trace divergence. Leaf traces divide tangentially in the cortex

CHAPTER 14

PALEOZOIC SEED FERNS

535

Figure 14.17 Cross section of Galtiera bostonensis (Mississip-

pian). Bar  6 mm. (Courtesy W. E. Stein.)

Figure 14.16 Cross section of Diichnia kentuckiensis stem with two leaf traces (arrows) (Mississippian). Bar  5 mm. (Courtesy J. Galtier.)

and supply the petiole bases with six, circular-to-elongate and obliquely oriented vascular bundles. Secondary xylem tracheids display multiseriate pitting restricted to the radial walls. The secondary phloem contains axial elements and dilated rays which extend into a peripheral zone of parenchyma. The cortex consists of an inner, parenchymatous ground tissue containing sclerotic clusters and an outer sparganum zone (Beck et al., 1992). Sparganum cortex differs from the dictyoxylon type in that the vertically oriented hypodermal fiber strands do not anastomose. The genus Galtiera (FIG. 14.17) is also known only from the New Albany Shale (Beck and Stein, 1987). Like Stenomyelon, the primary xylem in cross section appears as a triribbed protostele with persistent protoxylem strands located near the periphery. Surrounding the primary body is a zone of secondary xylem containing tracheids with elliptical pits and uni- to biseriate rays up to 30 cells high. The secondary phloem is characterized by tangential bands of macrosclereids. Leaf bases are of the Kalymma-type (FIG. 14.18) (see below). As in many seed ferns, the cortex of G. bostonensis contains a large number of sclerotic clusters (FIG. 14.17).

Figure 14.18 Cross section of Kalymma grandis (Mississip-

pian). Bar  5 mm.

Beck and Stein (1987) suggested two possible functions for this tissue system. One idea views the sclerotic tissue as important in mechanical support of the axis, whereas the other suggests that the sclerotic clusters represent a response by the plant to some external disease agent, such as a virus. Triichnia is an anatomically preserved stem from the Tournaisian of the Montagne Noire, France, which shows a unique combination of features as well as similarities in stelar organization and petiole anatomy to some species of Calamopitys, Stenomyelon, Diichnia, and Galtiera (Galtier and Beck, 1992). The stem of T. meyenii is 6 cm in diameter and characterized by a eustelic primary vascular

536

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.20 Cross section of Bostonia perplexa (Devonian–

Mississippian). Bar  6 mm. (Courtesy W. E. Stein.)

Figure 14.19 Cross section of the central portion of a Triichnia stem showing broad pith (Mississippian). Bar  3 mm. (Courtesy J. Galtier.)

system composed of a nearly continuous ring of mesarch primary xylem bundles around a wide parenchymatous pith that contains rare tracheids (FIG. 14.19). Leaves are helically arranged, and each leaf is supplied by three traces that diverge from two adjacent axial bundles. Leaf trace divisions in the stem cortex reveal a relatively asymmetrical pattern; petioles are of the Kalymma-type and characterized by up to 14 bundles. Another stem type from the New Albany Shale is Bostonia perplexa (FIG. 14.20) (Stein and Beck, 1978, 1992). This taxon is known from only two stem fragments 3 cm long and 2.1 cm in diameter. Stems are characterized by a parenchymatous ground tissue with nests of sclerotic cells and multiple segments of primary xylem surrounded by secondary vascular tissue. The plant is fundamentally protostelic with a deeply three-ribbed column of primary xylem. Each rib consists of a semi-discrete bundle of tracheids at the outside edge, which are intermittently connected to the stelar center by an extensive primary xylem parenchyma. The appearance of separate vascular segments at some levels is associated

with the departure of paired leaf traces. Between the levels of trace departure, the three-ribbed protostele is reconstituted with primary xylem ribs, following a helical course through the stem and supplying a regular phyllotaxis that conforms to the Fibonacci series. Surrounding the primary body is a zone of secondary xylem, which is 0.6–6 mm thick. The radial walls of the tracheids reveal bordered pit pairs with crossed, elliptical apertures, and multiseriate rays divide the manoxylic wood. Traces contain secondary xylem and divide at least once as they pass out to the laterals. The secondary phloem contains radial files of probable parenchyma and longitudinally elongated clusters of macrosclereids. The cortex is subdivided into an inner and an outer region. Attached petiole bases are broadly of the Kalymma-type (FIG. 14.18), but distinctly exhibit a three-ribbed medial petiole bundle. Stein and Beck (1992) interpreted Bostonia perplexa as a protostelic member of the Calamopityales with Medullosa-like vascular segments, thus reinforcing the long-standing view of a calamopityalean origin of the Medullosales. One of the most common and widespread genera in this order is Kalymma (FIG. 14.18), a morphotaxon for petioles that are considered characteristic of the Calamopityales (Long, 1964b; Barnard and Long, 1975; Matten and Trimble, 1978; Braun and Wilde, 2002). Anatomic features used to separate the numerous species include the distribution and shape of the vascular bundles, the level of tangential fusion between adjacent bundles, and the presence or absence of secretory canals. Kalymma petioles bifurcate near the base; above this division, the petiolar vascular bundles are arranged in a C shape; at

CHAPTER 14

PALEOZOIC SEED FERNS

537

PT

X

P

Figure 14.22 Cross section of Faironia difasciculata showFigure 14.21 Diplopteridium teilianum with fertile third rachis

(Mississippian). Bar  1 cm. (Courtesy B. Bomfleur and H. Kerp.)

higher levels, the number of bundles decreases. Long (1964b) suggested that petioles of K. tuediana represent the base of the Stenomyelon tuedianum frond. This suggestion is based on the position and number of vascular bundles and the common occurrence of both taxa in the same petrifactions. The petiole is slightly flattened on one side, 1 cm in diameter, and contains about nine exarch vascular bundles. In one specimen, 20 vascular bundles are apparent in cross section, many in the process of splitting. Foliage assignable to the morphogenera Diplotmema (often incorrectly spelled Diplothmema; see Zeiller, 1879–1880; Van Amerom, 1975) and Sphenopteridium has been suggested as belonging to Stenomyelon on the basis of its occurrence at the same stratigraphic level as K. tuediana (Long, 1964b). Another petiole taxon, initially described as Arnoldella minuta, is now regarded as displaying features of Kalymma (Sebby and Matten, 1969). Kalymma resinosa consists of terete axes with a xylem strand which is C shaped in cross section (Matten and Trimble, 1978). The abaxial surface

ing large pith (P), secondary xylem (X), and petiole (PT) (Mississippian). Bar  5 mm. (Courtesy A.-L. Decombeix.)

of the trace contains several lobes of xylem, each containing a mesarch protoxylem strand. Tracheids are scalariform and the cortex contains sclerenchyma plates that anastomose from level to level. Decombeix et al. (2006) recently reported on Faironia difasciculata (FIG. 14.22) from the Montagne Noire in France. This plant is characterized by a broad eustele, dense wood with multifascicular leaf traces that arise from two nonadjacent axial strands and Kalymma-type petioles. Although the Kalymma-type petioles suggest affinities within the Calamopityales, the anatomy of the stele, wood, and phloem compares more closely to that seen in early arborescent seed ferns such as Pitys (FIG. 14.23; see below) or Eristophyton (see section “Lyginopteridales”) and Megalomyelon. Chapelia campbellii is another taxon in which the difference between stem and leaf is indistinct. The specimen consists of a section of a petiole just below a branching point in the frond. In cross section, the vascular system consists of a four-lobed protostele with mesarch xylem, surrounded by a

538

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

core of manoxylic secondary xylem (Beck and Bailey, 1967). The cortex contains both secretory cells and clusters of presumed fibers. Branching of the stele results in the production of two outer, papilionoid vascular strands (butterfly-shaped tristichia in cross section) of distinct bilateral organization and two inner, clepsydroid-shaped strands that are oriented with their long axes at right angles to the plane of the frond. The flattened nature of the frond and the presumed pattern of branching are similar to the configuration of the compression foliage genus Diplopteridium (FIGS. 14.21, 14.24). Despite the uncertainty regarding the systematic position of the group, the Calamopityales are important in demonstrating significant intermediate stages in the evolution of the pteridosperm eustele. Namboodiri and Beck (1968c) and subsequent authors presented convincing evidence that several calamopityalean taxa could be used as examples in a series leading to the eustele in seed ferns such as Lyginopteris or Callistophyton. Beginning with a protostelic species of Stenomyelon such as S. primaevum as a starting point (FIG. 7.43A), there was a gradual dissection of the solid protostele through the addition of parenchyma in the form of both plates and clusters (FIG. 7.43B), as occurs in the tuedianum-type and heterangioides-type of S. tuedianum (Meyer-Berthaud and Stein, 1995). The subsequent step in the evolution of the eustele, leading from Stenomyelon to Calamopitys, involved the tangential separation of the primary xylem into distinct sympodia through the phylogenetic increase in the amount of parenchyma in the pith region, a process termed vitalization of the protostele (FIG. 7.43B, C) (Stein and Beck, 1992). According to these authors, the establishment of the eustele was followed by a single origin of paired leaf traces leading from Calamopitys to Triichnia and Diichnia. The discovery of the eustelic C. schweitzeri (Tournaisian), with leaf traces

originating in pairs from a single sympodium not only adds support to this phylogenetic hypothesis, but also shows that the character of paired leaf traces evolved earlier than was previously thought in the genus Calamopitys (Galtier et al., 1993). Concomitant with a vitalization of the protostele leading to the gymnosperm eustele, there was a change in leaf trace divergence. In Calamopitys, each sympodium divides tangentially to produce a trace directly outside the sympo-

Figure 14.23 Cross section of Pitys antiqua stem showing

Figure 14.24 Diplopteridium teilianum (Mississippian). Bar  1 cm. (Courtesy B. Bomfleur and H. Kerp.)

broad pith (Mississippian). Bar  1 cm. (Courtesy J. Galtier.)

CHAPTER 14

dium itself (FIG. 7.43C). In the lyginopterids (see later) and younger pteridosperms, each sympodium divides radially to produce a trace which is alongside the sympodium (FIG. 7.43D) (Beck, 1970; Galtier, 1973). Although no reproductive organs are known for the calamopityaleans, Long (1964b) has suggested that the seed Lyrasperma (FIG. 14.25) was borne on Stenomyelon stems, and Galtier

PALEOZOIC SEED FERNS

539

et al. (1987) have described several additional seed types from the Tournaisian of France in association with numerous stems of Calamopitys (Rowe and Galtier, 1990). Many believe that the most primitive seed plants evolved from an aneurophytalean ancestor (see Chapter 12). To date, however, no calamopityalean stems have been found associated with late Famennian seeds such as Moresnetia, Elkinsia, or Archaeosperma (Chapter 13) (Galtier and Meyer-Berthaud, 1989). Of particular interest will be the relationship of the fertile parts of the calamopityaleans to those of the later pteridosperms when the former are finally discovered or recognized.

BUTEOXYLONALES The order Buteoxylonales was established by Barnard and Long (1973) for protostelic stems with manoxylic secondary xylem containing high, narrow vascular rays from the upper Tournaisian Calciferous Sandstone Series, Cementstone Group of Oxroad Bay, Scotland. The principal feature of the order is the trilobed or T-shaped configuration of the petiolar vascular bundle in cross section (FIG. 14.26). Buteoxylon gordonianum is known from a permineralized stem fragment 12 cm long and nearly 2.5 cm in diameter (Barnard and Long, 1973); an additional specimen was described from the uppermost Devonian of Ireland (Matten et al., 1980b). Petioles up to 1.4 cm in diameter are attached to the stem in a helical pattern (2/5 phyllotaxy). Both stems and petioles contain a sparganum outer cortex. Elliptical bordered pits are present on the radial walls of the tracheids. Another stem genus placed in this group is Triradioxylon (Barnard and Long, 1975). As the name suggests, this genus possesses a small (1.5 mm in diameter), three-lobed

Figure 14.25 Longitudinal section of Lyrasperma scotica (Mississippian). Bar  2 mm. (Courtesy A. G. Long Collection, Hancock Museum.)

Figure 14.26 Cross section of Buteoxylon gordonianum peti-

ole showing trilobed configuration (Mississippian). Bar  0.5 mm. (From Barnard and Long, 1973.)

540

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

protostele. The presence of a central protoxylem group in the protostele as well as near the ends of the three lobes recalls the organization in several early fernlike plants. Petioles are around 1 cm in diameter and produced in a 1/3 phyllotaxy. The wood of T. primaevum is loosely organized and contains tracheids with bordered pits and high, narrow vascular rays; the cortex shows a sparganum organization. Of the several early seed plants and progymnosperms, Triradioxylon is believed to have the closest affinities with some members of the Aneurophytales (Barnard and Long, 1975). Petioles with trilobed vascular strands from the same locality as Triradioxylon are given the name Lyginorachis whittaderensis (Barnard and Long, 1975). Although slightly larger and differing in a few anatomical features, they are considered to belong to this order (Galtier, 1988). The Buteoxylonales today are probably the least understood Paleozoic seed fern group. Since they are known from such a small number of specimens, there has been no opportunity to examine variability within the taxa described to date. The configuration of the vascular strand in the petiole, which has been used as the principal feature to delimit this group, recalls the early assignment of specimens to the coenopterid ferns. Although the utilization of fern petiole anatomy has been replaced by information on sporangial organization in that group, the initial system provided a convenient method of organizing taxa. It remains to be seen how the discovery of pollen organs and seeds attached to some of these buteoxylonalean petiole taxa alters our understanding of the systematics and evolution of these early pteridosperms.

Despite the historical importance of this group in the recognition of the pteridosperms, many plants referred to the group today are still relatively poorly known. LYGINOPTERIS PLANT

Lyginopteris, like many fossil plants, was initially designated as the generic name of a particular organ (stem) (FIG. 14.27) but today is also regarded as the name for the entire plant. The name Lyginopteris (Potonié, 1899) is a substitute for the original name assigned to the stems, Lyginodendron, which is widely seen in the older literature (see Zimmermann, 1958). Usage of Lyginodendron is illegitimate because this name is preoccupied, having been previously assigned to a cast of a lepidodendrid stem, that is, L. landsburgii, by W. Gourlie in 1842. In the Carboniferous of Europe, Lyginopteris is well known and occurs widely (Patteisky, 1957). Its presence in North America, however, was only confirmed more recently, when Tomescu et al. (2001) described Lyginopteris royalii, a eustelic permineralized stem with rare secretory or sclerenchyma cells in the pith and middle cortex, and a Dictyoxylon-type outer cortex (see below) from the Upper Mississippian Fayetteville Formation in Arkansas. VEGETATIVE ORGANS Information about the type species Lyginopteris oldhamia comes from specimens collected from the Coal Measures of Britain. The largest eustelic stem is about 4 cm in diameter, with a central parenchymatous pith that contains scattered clusters of sclerotic cells. Around the pith are 5–10 mesarch

LYGINOPTERIDALES In many features, the lyginopterid seed ferns represent a highly artificial and heterogeneous group of plants. The information to date indicates that for the most part they were plants with relatively narrow stems and a scrambling or climbing type of habit. Among the pteridosperms about which we know more than just vegetative anatomy, they are the oldest group and considered to be the most primitive. They produced cupulate seeds, often with elaborate pollen-receiving structures and an incompletely fused integument, and had systems designed to seal the pollen chamber after pollination. The pollen organs consisted of clusters of sporangia that dehisced centrally; pollen was small, trilete, uniformly ornamented, and probably wind borne. Petiole morphology is commonly of the Lyginorachis type (below), but species vary from bearing unmodified branches to truly laminar pinnules.

Figure 14.27 Cross section of Lyginopteris oldhamia stem

showing dictyoxlon cortex (arrow) (Pennsylvania). Bar  5 mm.

CHAPTER 14

primary xylem strands; metaxylem tracheids exhibit multiseriate pits. The amount of secondary xylem produced by Lyginopteris is relatively small in proportion to the total size of the stem. Secondary xylem consists of large tracheids with angular bordered pits on the radial walls. In some specimens, a small amount of secondary xylem is visible adjacent to the pith (Williamson, 1890). This anomalous wood apparently results from a vascular cambium that formed secondary xylem toward the inside between the primary bundles. Vascular rays, one to several cells wide, extend through the secondary xylem; some rays are up to 2 cm in height. In well-preserved stems, some thin-walled cells that occupy the position of the phloem have been observed, although sieve areas have not been reported. Outside the phloem is a zone of cells suggestive of a pericycle, which contains sclerotic nests like those of the pith. Near the outer limits of this zone is a narrow band of periderm. The cortex is divided into two zones: the inner zone is rarely preserved and consists primarily of parenchyma. The outer cortex is relatively extensive and includes parenchyma and, near the periphery of the tissue, bands of fibers that are radially aligned and extend longitudinally in the cortex (FIG. 14.27). These bands anastomose from level to level, exhibiting a netlike structure when viewed in a tangential longitudinal section (FIG. 14.30); this type of cortex is termed a Dictyoxylon-cortex (or dictyoxylon) (FIG. 14.30). In cross section, the bands appear similar to roman numerals. The presence of these anastomosing, fibrous bands has facilitated the assignment of compressed axes to the Lyginopteris group of the Lyginopteridales (Gothan and Zimmermann, 1938; Zimmermann, 1960; Van Amerom, 1975; Krings and Schultka, 2000). The epidermis surface of young Lyginopteris stems and petioles is covered with large multicellular trichomes and numerous multicellular capitate glands that are also present on all other parts of the plant except the roots (FIG. 14.28). It was the presence of the latter structures that allowed Oliver and Scott (1904) to reconstruct the entire plant and establish the existence of the pteridosperms (FIG. 14.5). The glands are 3 mm long and consist of a flared base and a glandular head; stomata are present on the stalk. The epidermis of the distal end contains small tubular cells that surround a secretory tissue. Traces to the leaves are formed by a tangential division of a primary xylem strand (Blanc-Louvel, 1966). In the cortex, the trace separates into a pair of strands that fuse in the base of the petiole to form a V- or W-shaped bundle. Petioles with this type of anatomy have been referred to as Lyginorachis (FIG. 14.28) (Meyer-Berthaud, 1990). Lyginopteris stems bore foliage assignable to the morphogenus Sphenopteris (FIG. 14.29),

PALEOZOIC SEED FERNS

541

for example, the frond of L. oldhamia has been described as Sphenopteris hoeninghausii. Information about the arrangement of fronds on the stems and about the frond architecture comes from impression–compression specimens. One of the

Figure 14.28 Cross section of Lyginorachis sp. showing

multicellular glands (arrows) (Pennsylvanian). Bar  2.5 mm.

Figure 14.29 Compression of Sphenopteris hoeninghausii

(Pennsylvanian). Bar  5 cm. (Courtesy R. Daber.)

542

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.30 Dictyoxylon cortex, tangential section (Pennsyl-

vanian). Bar 1.5 mm. (Courtesy BSPG.) Figure 14.31 Stem of Lyginopteris oldhamia showing numer-

most revealing fossils shows a stem portion with dictyoxylon cortex from the Carboniferous of Lower Silesia (Poland) to which are attached several fronds assignable to Sphenopteris linkii (Gothan and Zimmermann, 1938). The bipartite fronds are produced in a 2/5 phyllotaxy; are relatively large, conspicuously planated, and pinnae are oppositely to suboppositely arranged on the petiole (stipe) and rachides. Pinnules are lobed and borne suboppositely or alternately. Roots of Lyginopteris are adventitious and are borne on all sides of the stem (FIG. 14.31) (Krings and Schultka, 2000), occasionally in vertical rows. Some are 7 mm in diameter and contain secretory cells in the cortex. The stele varies from triarch to polyarch, and some secondary xylem is apparent (FIG. 14.32). The name Kaloxylon hookeri is often used for anatomically preserved roots of Lyginopteris (Williamson, 1875, 1887a). Despite the existence of many well-preserved fossils of the individual parts of Lyginopteris, there is still relatively little known about the growth habit of the plant (see T. Taylor and Millay, 1981a). Some have reconstructed L. oldhamia as

ous adventitious roots extending from stem and frond bases (Pennsylvanian). Bar  4 cm. (From Krings and Schultka, 2000.)

a freestanding shrub (Retallack and Dilcher, 1988), whereas others have suggested that it was a scrambling or climbing liana (Potonié, 1899). Biomechanical analyses support the latter view and indicate that L. oldhamia was not self-supporting and had a relatively sophisticated, climbing stem architecture (Speck, 1994; Masselter et al., 2007). REPRODUCTIVE STRUCTURES The best-known permineralized seed of Lyginopteris has been given the name Lagenostoma lomaxii. Lagenospermum is the name applied to compressed seeds that correspond morphologically to Lagenostoma (Arnold, 1939). Lagenostoma is ellipsoidal and measures 5.5 mm long and up to 4.2 mm in diameter (FIG. 14.33). The integument is composed of thick-walled cells in two different zones: those of the inner zone parallel to the long axis of the seed, whereas those of the outer layer are oriented in a radial direction. At the

CHAPTER 14

Figure 14.32 Cross section of Kaloxylon root (Pennsylvanian).

Bar  0.5 mm.

Figure 14.33 Longitudinal section of Lagenostoma lomaxii

(Pennsylvanian). Bar  1 mm.

micropylar end of the seed, the integument is dissected into nine lobes, each supplied by a vascular strand. Covering the sclerified portion of the integument is a relatively thin zone of parenchyma that is rarely preserved.

PALEOZOIC SEED FERNS

543

Longitudinal section of Lagenostoma lomaxii showing central column (arrow) (Pennsylvanian). Bar  500 μm. Figure 14.34

The inner part of the integument in Lagenostoma is fused to the nucellus except in the distal (micropylar) region of the seed (Oliver and Scott, 1903, 1904; Arber, 1905; Long, 1977c). Apically, the nucellus consists of an elongated collar or flaskshaped structure, the lagenostome, which loosely surrounds a parenchymatous central column (FIG. 14.34). The distal end of the lagenostome probably projected a short distance above the lobes of the integument, which were not completely fused apically. The central column in this seed may have served as a source of a pollination droplet that trapped the wind-borne pollen grains, or it may have somehow directed the grains to the peripherally located archegonia of the megagametophyte. Wellpreserved specimens of L. ovoides have been described with a cellularized megagametophyte within a delicate megaspore membrane (Long, 1944). The megagametophyte is differentiated into an outer zone of radially elongated cells and an inner area of isodiametric cells (FIG. 14.35). Up to three ovoid archegonia, which lack neck cells, are present in the distal end of the megagametophyte. A triangular-shaped pad of tissue termed the tent pole occurs just below the central column in the megagametophyte in specimens judged to be mature, as determined by the presence of a cellular megagametophyte with archegonia. The tent pole is a common feature of the megagametophyte in several Paleozoic seeds. It may have functioned to seal the pollen chamber, elevate the necks of the archegonia, or rupture the megaspore membrane to allow gamete entrance into the archegonia. Lagenostoma was produced in a uniovulate cupule (FIG. 14.36). If found without seeds, this tulip-shaped cupule is

544

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.35 Longitudinal section of Lagenostoma lomaxii showing cellular megagametophyte (Pennsylvanian). Bar  1 mm. (From T. Taylor and Millay, 1979.)

Figure 14.36 Cupules of Calymmatotheca haueri (Mississip-

pian). Bar  1 cm. (Courtesy GBA.)

Figure 14.37 Calymmatotheca bifida (Mississippian). Bar  400 μm. (From Taylor and Taylor, 1993.)

given the name Calymmatotheca (FIG. 14.37). It consists of a series of enclosing lobes that are fused at the base and wrinkled on the outer surface (Benson, 1935b). The upper portion of the cupule consists of free lobes, each with a single, terete vascular strand. The outer surface of the cupule bears the same capitate glands (FIG. 14.5) that are present on the vegetative parts of Lyginopteris. The cupulate fructification Calathiops bernhardtii from the Upper Namurian of Germany has been interpreted as the immature form of Calymmatotheca by Benson (1935a). Gothan (1935), however, attributed C. bernhardtii to the foliage morphogenus Mariopteris (see Chapter 16); other species of Calathiops have been found in association with foliage assigned to Cardiopteris and Sphenopteridium (Gothan, 1927b). The large number of detached Lagenostoma seeds in the fossil record has led some to suggest that they were shed from the cupules at the same stage of development. Several dichotomizing branching systems terminating in similar cupules, each with a single seed, have been described from Lower Mississippian rocks (Price Formation) of Virginia as Lagenospermum (Gensel and Skog, 1977). These cupules range up to 3 cm long and are divided into six apical lobes. The seeds are preserved as casts and measure 6 mm long

CHAPTER 14

PALEOZOIC SEED FERNS

545

Figure 14.40 Suggested reconstruction of Crossotheca sagittata synangium (Pennsylvanian). (From Millay and Taylor, 1979.)

14.38 Lyginopterid cupules with seeds (arrows) (Mississippian). Bar  1 cm. (Courtesy B. Bomfleur and H. Kerp.)

Figure

Figure 14.41 Proximal surface of Crossotheca sp. pollen grain

(Pennsylvanian). Bar  10 μm.

Figure 14.39 Compression of Crossotheca sp. showing attachment of synangia (Pennsylvanian). Bar  1 cm.

and 2 mm in diameter. Cupules (FIG. 14.38) are relatively common among many of the earliest seed plants and may have functioned initially as protective structures for the developing ovules. From time to time, the lyginopterid cupule has been suggested as a precursor to the angiosperm carpel. Despite a considerable amount of information about the seeds and vegetative parts of Lyginopteris, the exact

pollen-producing organs have not been identified conclusively. The genus Crossotheca is used for compression–impression specimens and is often regarded as the microsporangiate unit of Lyginopteris (FIG. 14.39), but this relationship has been suggested only because of the association of the organ with Sphenopteris foliage and their occurrence at the same stratigraphic level (Stubblefield et al., 1982). Brousmiche (1982), however, has offered an alternative view, suggesting that Crossotheca is a eusporangiate fern. Crossotheca sagittata is a common North American species (Arnold and Steidtmann, 1937). It consists of a fertile pinna that bears opposite, reduced fertile pinnules and overall is shaped much like an arrowhead (FIG. 14.40). Extending from the lower surface are numerous (30) sporangia, each 4 mm long. The pollen grains (prepollen of some authors) are 70 μm in diameter and trilete (FIG. 14.41).

546

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

14.44 Synangium of Feraxotheca (Pennsylvanian). (From Millay and Taylor, 1978b.)

Figure

Figure 14.42 Several compressed synangia (arrows) of Crossotheca kentuckiensis (Pennsylvanian). Bar  2.3 mm. (From Stubblefield et al., 1982.)

Figure 14.43 Suggested reconstruction of Feraxotheca culcitaus (Pennsylvanian). (From Millay and Taylor, 1978b.)

It is suggested that the unusual shape of the synangial unit may have resulted from the fusion of three oval synangia on a small, pinnate frond segment. Basal pinnules of Pecopteris or Sphenopteris are attached to Crossotheca fertile units.

culcitaus

If Crossotheca were the pollen organ of Lyginopteris, the absence of Lyginopteris in the North American Carboniferous suggests that Crossotheca may represent the pollen organ of several different pteridosperm genera (FIG. 14.42). Feraxotheca is a permineralized lyginopterid pollen organ that may represent a different preservational state of Crossotheca (Millay and Taylor, 1977). Both fertile and sterile frond segments have been described from Pennsylvanian coal balls from eastern Kentucky, USA (Millay and Taylor, 1978b). The fertile axis includes three orders of branching, with ultimate pinnae alternately arranged (FIG. 14.43). Each fertile unit consists of a basal parenchymatous cushion that supports a variable number (6–10) of closely appressed, elongated, exannulate sporangia (FIG. 14.44). The upper surface of the parenchymatous pad is vascularized (FIG. 14.45), but no vascular tissue extends into the individual sporangia. The central region of the synangium is hollow, and dehiscence takes place along the inner-facing sporangial walls (FIG. 14.46). Pollen grains are small (40–60 μm), trilete, and ornamented with delicate coni. Hypostomatic laminar pinnules of F. culcitaus are lobed and possess a dichotomous venation pattern. Present on both vegetative and fertile frond segments are canals containing a yellow material that appears frothy in peel section. These structures ultimately may provide a clue to identify the other portions of this plant, similar to the association of parts demonstrated in Lyginopteris. OTHER LYGINOPTERIDS: VEGETATIVE REMAINS

Trivena arkansana is a permineralized stem type from the Upper Mississippian Fayetteville Formation in Arkansas. Specimens are up to 30 mm by 19 mm in diameter with a pentagonal eustele of five cryptic sympodia and a protoxylem pole located in each of the five stelar angles (Dunn et al., 2003b; Dunn, 2006). Secondary xylem is extensive and characterized

CHAPTER 14

PALEOZOIC SEED FERNS

547

Figure 14.47 Cross section of Heterangium americanum

stem showing mixed protostele and cortex with trace (arrow). (Pennsylvanian). Bar  2 mm.

Figure 14.45 Longitudinal section of synangium of Feraxotheca culcitaus. Arrow indicates parenchymatous cushion (Pennsylvanian). Bar  1 mm.

Figure 14.46 Cross section of Feraxotheca culcitaus synan-

gium (Pennsylvanian). Bar  0.5 mm.

by numerous wide rays; it is surrounded by a vascular cambium and secondary phloem. The phloem is surrounded by a narrow inner cortex, followed by a periderm and a parenchymatous middle cortex with abundant sclerotic clusters; these are randomly dispersed or arranged in discontinuous rows.

Sclerenchyma bands form a dictyoxylon pattern in the outer cortex. Leaf traces diverge in a 2/5 phyllotaxy and originate by division from a single protoxylem pole. Traces, accompanied by concentric secondary xylem, increase in size as they extend through the secondary xylem of the stem. The trace assumes a flattened C shape at the outer margin of the secondary xylem, and in the cortex, it divides into three discrete bundles, each surrounded by secondary xylem. Dictyastrum chesteriensis is a seed fern from the Upper Mississippian Chester Series of the Illinois Basin that is based on compressions, impressions, and pyrite–marcasite petrifactions. The taxon is characterized by a mixed protostele with indistinct protoxylem, a narrow zone of secondary xylem containing numerous small rays, and prominent anastomosing fiber bundles forming a dictyoxylon-type pattern in the cortex (Jennings, 1987). Petioles are massive and fork distally into unequal subdivisions. The fronds bear rounded pinnules corresponding to the foliage morphotaxon Sphenopteris stricta. Dictyastrum chesteriensis is interpreted as being morphologically intermediate between Lyginopteris and Heterangium. HETERANGIUM Heterangium is a stem morphotaxon, somewhat similar to Lyginopteris which is known from Carboniferous and possibly Permian (see Hirmerc, 1933; Galtier, 1996) rocks in both North America and Europe (Williamson, 1887a; Kubart, 1914; Andrews, 1942; Bertram, 1989). Stems typically measure 2 cm in diameter and are rarely branched. The stele consists of a vitalized or mixed protostele consisting of clusters of tracheids and parenchyma (FIG. 14.47); at some

548

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.48 Cross section of Heterangium americanum stem (Pennsylvanian). Bar  1 mm.

Figure 14.49 Cross section of Heterangium americanum stem. Note extensive secondary xylem and extraxylary tissues (Pennsylvanian). Bar  1 cm.

levels, the parenchyma is arranged in longitudinally oriented plates (FIG. 14.48). The large metaxylem tracheids reveal multiseriate pitting and grade into smaller tracheary elements (probable protoxylem) near the periphery of the stele and are arranged into about 20 mesarch strands in H. grievii from the Mississippian of Pettycur (Williamson, 1873). Surrounding the primary xylem is a band of secondary xylem (FIG. 14.49) with pitting on the radial walls. Surprisingly, phloem is preserved in a large number of Heterangium stems. In H. americanum (Late Pennsylvanian), both primary and secondary phloem are preserved (Hall, 1952). The primary phloem

consists of small cluster of cells near the periphery of the secondary phloem zone. The secondary phloem is extensively developed and may have exceeded the amount of secondary xylem in width. It consists of elongated sieve cells, phloem parenchyma, and phloem rays that extend out from the secondary xylem. Roots are triarch and may have produced some secondary xylem. The cortex of Heterangium includes an inner zone of parenchyma with evenly spaced horizontal plates of thick-walled cells and an outer zone of longitudinally oriented fibers embedded in a parenchymatous ground tissue. Sclerotic plates are absent in H. lintonii (Stidd, 1975). Species of the genus Heterangium generally lack a dictyoxylon-type cortex (Van Amerom, 1975). In Heterangium grievii (sometimes spelled H. grievei), the petiole is supplied by a single trace. Fronds are bipartite and have laminar foliage of the Sphenopteris elegans (adiantoides)-type; other species were also associated with Sphenopteris-type foliage (see Chapter 16). A large compression fossil showing a slender (1 cm in diameter), branched stem with attached petioles assignable to S. elegans (adiantoides) from Lower Silesia (Poland) has been used to interpret H. grievii as a small liana, perhaps similar in growth habit to some members of the extant genus Clematis (Ranunculaceae) (Gothan and Zimmermann, 1938). Adding support to this interpretation is the fact that several of the foliage remains associated with the stem have distal pinnules modified into climber hooks comparable to the hooks seen in the foliage taxa Karinopteris (FIG. 14.50), Mariopteris, and Pseudomariopteris (Chapter 16). Roots of H. grievii are adventitious, monopodially branching, and of endogenous origin (Benson, 1933). The petiole of Heterangium americanum from Late Pennsylvanian coal balls from Illinois contains 7–10 vascular bundles, depending on the level (Shadle and Stidd, 1975). The frond dichotomized at the base (FIG. 14.51) and vascular bundles in the region of the dichotomy are W-shaped in cross section. Laminar pinnules similar to Eusphenopteris obtusiloba are borne alternately along the secondary pinna axis (FIG. 14.52). Pinnules contain a prominent palisade layer and trichomes on the outer surface. This leaf type is similar to the frond morphology of Lyginopteris. One of the best-known Heterangium species is H. kentuckyensis from Middle Pennsylvanian coal balls of North America (Pigg et al., 1987). In addition to stems and petioles (FIG. 14.53), the taxon is known from frond members, foliage, and roots, several in attachment. Some developmental stages are known for the stems, which range up to nearly 6 mm in diameter. Development of the sclerenchyma plates within the cortex is believed to have arisen from discontinuous cambia. A limited amount of periderm was also produced in this

CHAPTER 14

PALEOZOIC SEED FERNS

Figure 14.52 Pinna of (Pennsylvanian). Bar  1 cm.

Eusphenopteris

549

obtusiloba

Figure 14.50 Climber hooks of Karinopteris sp. (Pennsylvanian). Bar  2 mm. (Courtesy H. Kerp.)

Cross section of Heterangium kentuckyensis petiole (Pennsylvanian). Bar  2.2 mm. (From Pigg et al., 1987.)

Figure 14.53

Figure 14.51 Diagrammatic reconstruction of Heterangium

frond (Pennsylvanian). (From Shadle and Stidd, 1975.)

species. The fronds are at least twice pinnate, with the primary pinnae alternate and attached at right angles. Pinnules of the Sphenopteris type possess stomata with overarching papillae. Heterangium stems and petioles from Upper Mississippian rocks of Illinois are of interest because they demonstrate another type of foliage produced by this plant (Jennings,

550

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.54 Cross section of Microspermopteris aphyllum stem showing parenchymatous plates in pith (Pennsylvanian). Bar  1 mm.

Figure 14.55 Cross section of Microspermopteris aphyllum

stem showing petiole trace (arrow) (Pennsylvanian). Bar  1 mm.

1976). The petioles are forked and bear foliage of the Rhodea type. This foliage lacks laminar pinnules and the distal ends of the pinnae are composed of dichotomous sterile segments. The presence of this frond morphology during the Mississippian supports the concept that the geologically younger Heterangium frond evolved through the planation and webbing of a leaf such as that of Rhodea. The sporangium-bearing fronds of this Mississippian Heterangium species reveal the same anatomy as the sterile axes, but they are terminated by clusters of Telangiopsis– Telangium-type sporangia (see below). MICROSPERMOPTERIS Microspermopteris (FIG. 14.54) shares a number of anatomical features with Heterangium (Baxter, 1949, 1952a). Information about the genus has largely been based on material from North America, but specimens have also been described from the Carboniferous of Germany (Bertram, 1989). The genus consists of a small stem with a pentarch protostele of metaxylem wedges separated by radially oriented plates of parenchyma (FIG. 14.54) (T. Taylor and Stockey, 1976). Protoxylem strands are located on either side of the parenchyma plates near the periphery of the xylem. Metaxylem tracheids are long (up to 1 mm) and contain multiseriate, simple reticulate pits. In some of the larger stems (1.1 cm in diameter) with abundant secondary xylem, the symmetry of the stem is apparently altered as a result of gaps in the wood where branches were formed. Petioles in Microspermopteris are produced in a 2/5 phyllotaxy and are attached to the stem by clasping, V-shaped bases (FIG. 14.55). Primary pinnae are alternately arranged and bear

pinnules of the Sphenopteris type; branches were produced in the axils of the fronds. Some branches produced helically arranged, scalelike leaves. The cortex of Microspermopteris is composed of thin-walled parenchyma cells, a small amount of periderm in mature stems, and axially aligned secretory cells. The presence of cortical fibers depends on the age of the stem, the degree of preservation, and the level of the section. Extending out from the stem as lateral extensions of the cortex in some specimens are irregular, winglike structures (FIG. 14.56) (T. Taylor and Stockey, 1976), some of which may be several millimeters long. In addition to the cortical flaps, irregularly shaped, multicellular trichomes are most common at the nodes and on the distal parts of axes. Adventitious roots with well-developed secondary tissues are borne at various levels along the stems. They range from triarch to polyarch and contain abundant secretory ducts. Since reproductive parts of Microspermopteris remain unknown, protoxylem architecture, nature of the leaf traces, and the pinnate organization of the frond are used as the basis for separating Microspermopteris from Heterangium (Pigg et al., 1986). SCHOPFIASTRUM Another stem that has been included within the Lyginopteridales is Schopfiastrum (FIG. 14.57). This genus was initially described by Andrews (1945) from two fragments collected in Iowa but today is known from a number of Middle Pennsylvanian localities (Rothwell and Taylor, 1972). In general, the stems are slightly larger than those of Heterangium. In cross section, the protostele is oval

CHAPTER 14

PALEOZOIC SEED FERNS

551

Figure 14.57 Cross section of Schopfiastrum decussatum stem

(Pennsylvanian). Bar  1 mm.

In S. decussatum, leaves are borne in a two-ranked, alternate arrangement (Stidd and Phillips, 1973). Leaf traces divide repeatedly to produce six to seven protoxylem strands on the abaxial face of a lobed metaxylem band. At higher levels of the frond, the primary pinnae exhibit a Y-shaped vascular trace from which pinnule traces depart (FIG. 14.58). Laminar pinnules are morphologically similar to Sphenopteris and Mariopteris and contain mesophyll plates that extend between the adaxial and the abaxial surfaces.

Figure 14.56 Suggested reconstruction of Microspermopteris aphyllum showing adventitious roots and epidermal appendages (Pennsylvanian). (From T. Taylor and Stockey, 1976.)

in outline, and the primary xylem consists of two to four protoxylem strands that surround the central metaxylem. Metaxylem consists of large, angular tracheids and a few scattered parenchyma cells. Surrounding the primary body is a large zone of secondary xylem and wood rays. Tracheids exhibit multiseriate-bordered pits with crossed, slit like apertures on the radial walls. Extending longitudinally through the inner cortex and phloem are large (1 mm in diameter) canals typically filled with an amber-colored material regarded as resinous by some authors. In the outer portion of the cortex are longitudinally oriented sclerenchyma strands.

PITYS Pitys (or Pitus; for details see Doweld and Reveal, 2002) was a large forest tree during the Mississippian with some specimens up to several feet in diameter. Of the several species that have been described to date, P. dayi is perhaps the best known (Gordon, 1935). The center of the stem was occupied by a parenchymatous pith with up to 50 mesarch primary xylem strands located around the periphery. Secondary xylem consisted of tracheids with hexagonal– circular bordered pits and wide, high vascular rays. On the basis of anatomy, Long (1963) suggested that Lyginorachis papilio was the petiole of Pitys and Tristichia ovensii the ovule-bearing axis arising in the bifurcation of the L. papilio rachis. Galtier (1977), however, proposed retaining the genus Tristichia for those axes whose attachment to Pitys has not been demonstrated. Foliage is believed to have been of the Sphenopteris or Sphenopteridium type. The pycnoxylic wood of Pitys and the large size of the stems are features that are unlike other lyginopterid seed ferns, and the

552

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.59 Cross section of Tetrastichia bupatides (Mississi-

ppian). Bar  2 mm.

Figure 14.58 Suggested reconstruction of Schopfiastrum decussatum showing vasculature in representative cross sections of the frond. (From Rothwell and Taylor, 1972.)

genus is assigned to the lyginopterids based primarily on the Lyginorachis petiole anatomy. Biomechanical analyses suggest that, in contrast to many other lyginopterids, P. dayi was self-supporting and showed trends in mechanical parameters during ontogeny similar to those of extant, self-supporting plants (Speck and Rowe, 1994). DEVONIAN–MISSISSIPPIAN TAXA Although much of the evidence for the existence of Mississippian pteridosperms has been based on the presence of seeds, several stem genera that demonstrate a set of characteristics suggesting affinities with the lyginopterid pteridosperms are known. Rhetinangium is a stem known

from the Mississippian of Scotland (Gordon, 1912) and the Upper Mississippian of Arkansas, USA (Dunn, 2004). The stem has a mixed protostele with clusters of tracheids surrounded by thin-walled parenchyma and associated secretory cells. Xylem is exarch and arranged in a series of axial bundles. The band of secondary xylem is narrow and tracheid pitting is reticulate. The inner cortex consists of parenchyma and numerous secretory canals, whereas the outer cortex contains thick-walled fibers aggregated into radial bands. Leaves are helically arranged on R. arberi and, when found isolated, are referred to as Lyginorachis arberi (Long, 1964a). The petioles are 1.5 cm in diameter and dichotomize to form alternate, pinnae-bearing cylindrical pinnules. The petiole vascular bundle consists of numerous U-shaped leaf traces when viewed in transverse section. At a higher level, these traces fuse to form a bandlike strand. Another Mississippian stem that has been included with the pteridosperms is Tetrastichia. The stems of T. bupatides are 1 cm in diameter and contain a protostele that is cruciform in transverse section (FIG. 14.59) (Gordon, 1938). The secondary xylem contains numerous vascular rays with reticulate pits present on both radial and tangential tracheid walls. Outside the vascular tissue in well-preserved specimens is a three-parted cortex. The outer zone consists of plates of hypodermal fibers and mucilage cells that surround a middle region of parenchyma and nests of fibers. The inner region contains a thin-walled parenchyma. Petioles are borne in a four-ranked arrangement, with opposite pairs alternating (decussate). Each petiole dichotomizes

CHAPTER 14

PALEOZOIC SEED FERNS

553

Cross section of Tristichia ovensii petiole (Mississippian). Bar  1 mm. (Courtesy J. Galtier.)

Figure 14.61

Figure 14.60 Tristichia stem showing primary xylem configu-

ration (Mississippian). Bar  400 μm. (Courtesy J. Galtier.)

a short distance from the stem; subsequent levels of frond morphology are not known. Tetrastichia is provisionally placed in the pteridosperms, but the exact affinities remain unresolved. Bertrand (1941), however, noted striking similarities between the anatomical characters of T. bupatides and certain ferns. In Tristichia (uppermost Devonian–Mississippian), the protostele is triangular in cross section (FIG. 14.60) (Long, 1961b; Galtier, 1977; Galtier and Meyer-Berthaud, 1996). Primary xylem is mesarch and secondary xylem with multiseriate-bordered pits is sometimes present. When present, the secondary xylem is characterized by narrow rays and tracheids with multiseriate pitting. Petioles of T. ovensii are large and borne in a 1/3 phyllotaxy; they bifurcate and at higher levels dichotomize several times (Long, 1961b). The petioles contain butterfly-shaped, triarch-tetrarch traces and nests of cortical fibers (FIG. 14.61). Numerous compression remains of small (1.5 mm long) pinnules and bifurcating axes are associated with the petrifactions. They are similar to those of the vegetative branches and bear cupules containing seeds. The cupules are borne in pairs at the ends of a dichotomized stalk and each cupule is multiovulate. Morphologically, the cupules and

macerated megaspore membranes appear identical with those of the seed Stamnostoma huttonense (Chapter 13). Kerryoxylon hexalobatum is a permineralized stem described from the uppermost Devonian of Ireland (Matten et al., 1984b). The taxon is characterized by an outer sparganum-type cortex, sclerotic nests in the inner cortex, an unusual six-ridged protostele, and laterals that have approximately the same diameter at their base as the stem; petioles resemble Lyginorachis. Although the exact affinities of Kerryoxylon remain uncertain, the taxon is probably best placed within the Lyginopteridales on the basis of the available anatomical characters. Another large, arborescent stem that shares some features with Pitys is Eristophyton waltonii (Lacey, 1953) (FIG. 14.62), from the Mississippian of Scotland. The taxon consists of a stem showing dense wood with alternate, crowded bordered pits on the radial walls of tracheids and numerous, small pits in the cross fields. Additional specimens from the Tournaisian of Scotland indicate that the wood of the trunk was at least 25 cm in diameter (Galtier and Scott, 1990). Eristophyton beinertianum (FIG. 14.63) is a second species from the Tournaisian site but differs in several anatomical features from E. waltonii. Lyginorachis waltonii is suggested to be the leaf of Eristophyton (Long, 1987; Galtier and Scott, 1990). Compared to all other arborescent genera of lignophytic affinities (seed plants or progymnosperms), known for the Mississippian, Eristophyton has the widest geographical extent and stratigraphic range, that is, Mississippian to Permian. The Mississippian forms are reviewed in Decombeix et al. (2007; 2008).

554

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

R

Figure 14.62 William S. Lacey. (Courtesy T. Delevoryas.)

Figure 14.64 Cross section of Stanwoodia kirktonensis stem Figure 14.63 Cross section of Eristophyton beinertianum

stem (Mississippian). Bar  5 mm. (Courtesy A.-L. Decombeix.)

Stanwoodia kirktonensis (FIG. 14.64) is a Mississippian stem that resembles Eristophyton (Galtier and Scott, 1991, 1994). The protostelic stem is 4 mm in diameter with sympodial strands in a continuous ring at the periphery of the xylem. Protoxylem is mesarch with leaves borne in a subopposite pattern. One of the characteristics of the genus is the extensively developed periderm that formed a rhytidome or outer bark. Bilignea is a woody stem from the Mississippian of Scotland in which the parenchyma of the pith is replaced by short tracheids (Scott, 1924). Scott suggested that these tracheids represent an adaptation to water storage. Rays are uniseriate and short, and pitting in B. solida (FIG. 14.65) is uniseriate bordered, whereas in B. resinosa (FIG. 14.66) the pits are multiseriate. Leaf traces are helically arranged, simple, and similar to those of Eristophyton. Many of these Early Carboniferous plants are believed to have produced foliage of the Rhacopteris type (FIG. 14.67) (Galtier and Meyer-Berthaud, 2006).

showing the rhytidome (R) (Mississippian). Bar  5 mm. (Courtesy J. Galtier.)

PROBLEMATIC LYGINOPTERIDS Two additional genera that are tentatively included in the Lyginopteridales are Lyginopitys and Laceya (Galtier, 1970b). In Lyginopitys puechcapelensis, the stem is 1 cm in diameter and characterized by a pentagonal primary body with mesarch bundles (Galtier, 1970b). Secondary xylem is manoxylic and petiole traces range from V- to W-shaped in cross section. The Late Devonian genus Laceya possesses a sparganum-type cortex in both the stems and the frond parts (May and Matten, 1983). The protostele is trilobed, and vascular rays are prominent in the secondary xylem. The frond of L. hibernica from the uppermost Devonian of Ireland is known in some detail. Fronds were bipartite and at least twice compound (Klavins and Matten, 1996). Alternate pinnae occur below the bifurcation of the primary rachis and along the secondary rachides. At the base, the primary rachis exhibits a V-shaped vascular strand with 8–12 protoxylem groups; a little higher up (more distal), it has a W-shaped strand

CHAPTER 14

PALEOZOIC SEED FERNS

555

14.65 Cross section of Bilignea solida stem (Mississippian). Bar  3 mm. (Courtesy A.-L. Decombeix.)

Figure

Figure 14.67 Rhacopteris lindseaeformis pinnate leaf (Mississippian). Bar  6 mm. (Courtesy B. Bomfleur and H. Kerp.)

protoxylem groups. Pinna traces are bilobed to arcuate with 2–4 protoxylem groups. Protoxylem maturation is mesarch. A well-developed sparganum-type outer cortex occurs in the rachides but is reduced or absent in the pinna axes. Figure 14.66 Cross section of Bilignea resinosa stem

(Mississippian). Bar  6 mm. (Courtesy A.-L. Decombeix.)

with 10–16 protoxylem groups, and immediately below the bifurcation, there are two V- or U-shaped vascular strands with 3–8 protoxylem groups each. Secondary rachides are characterized by a V- or J-shaped vascular strand with 4–8

OTHER LYGINOPTERIDS: SEEDS AND CUPULES

Numerous seeds, in fact many more types than there are stems, have been referred to the lyginopterids based only on their relatively small size. Several of these were discussed in Chapter 13, and a few additional ones will be considered here. Although Lagenostoma is the only anatomically preserved seed definitively attributed to vegetative organs in this

556

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

completely free from the integument except at the chalazal end of the seed. SALPINGOSTOMA Another Mississippian seed with a lagenostome is Salpingostoma dasu (FIG. 13.34), known from the volcanic ash beds of Oxroad Bay, Scotland (Gordon, 1941). It is elongated (5 cm long by 6 mm in diameter), and its greatest diameter is near the distal end. In transverse section, the integument shows six prominent ribs that are organized into free lobes near the apex. The pollen chamber and the lagenostome are similar to those of Lagenostoma, except that the distal end is elongated into a conspicuous funnel or salpinx that extends between the integumentary lobes. Numerous multicellular hairs extend from the inner surface of the integument and may have functioned to maintain the position of the pollenreceiving device. The pollen in this seed is large (104 μm) and trilete. Salpingostoma prinsii from the Westphalian A–B of southern Limburg, The Netherlands, is a geologically younger member of the genus (Van Amerom, 1990).

Figure 14.68 Oblique longitudinal section of Sphaerostoma ovale (Mississippian). Bar  1 mm. (Courtesy BSPG.)

order, several taxa have been suggested as the seeds of certain stem genera. The assignment of seeds and stems has been made principally on their association at the same locality. SPHAEROSTOMA Sphaerostoma ovale (FIG. 14.68) is known from the Mississippian of Pettycur, Scotland, and is associated with stems of Heterangium grievii (Benson, 1914). The seed is 3.5 mm long and up to 2.2 mm in maximum diameter. Integument and nucellus are fused from the base to near the micropyle, where the tip of the nucellus is modified to form a low pollen chamber and cellular central column. The integument is divided into three parts, with the middle zone (sclerotesta) constructed of elongated fibers. The outer layer, or sarcotesta, is conspicuous at the apex, where a crest of large, axially oriented cells formed a frill around the micropyle. The cells of this structure may have broken down to form the pollination droplet. The inner layer of the integument, the endotesta, is often preserved only as a thin cuticular layer. One interesting feature of this seed is the presence of a delicate, vascularized cupule that tightly encloses it. Although this structure has been termed an outer integument, it is

CONOSTOMA Conostoma (FIG. 14.69) contains a large number of species and is known throughout the Pennsylvanian of Europe and North America (Long, 1977; Rothwell et al., 1979; Stubblefield and Rothwell, 1980; Stubblefield et al., 1984b). The seeds are radially symmetrical and less than a centimeter in maximum length. In C. anglo-germanicum, which is known from Great Britain, Germany, and the United States, the outer surface of the integument is ornamented with four longitudinally oriented ribs that extend to the distal end of the seed (Rothwell, 1971a). Between these ribs are smaller, secondary extensions of the sclerotesta. Four vascular bundles extend through the endotesta (inner integumentary layer) to near the apex of the seed. The lagenostome in Conostoma is a small, doughnut-shaped extension at the distal end of the nucellus (FIG. 14.69). The central column is thought to have formed during the ontogeny of the pollen-receiving mechanism. According to this idea, as the pollen chamber differentiates, the central column is withdrawn but is later pushed into the base of the lagenostome to seal pollen grains in the pollen chamber. Such a force could have been exerted by the development of the tent pole during the later stages of megagametophyte ontogeny. Evidence that the lagenostome and central column once were organically attached may be seen in the unusual scalariform thickenings on the cell walls of these structures. A rather unusual integumentary structure is present in C. kestospermum, a species known only from coal balls

CHAPTER 14

PALEOZOIC SEED FERNS

557

C PC

L

Figure 14.70 Longitudinal section of Conostoma kestosper-

mum showing conspicuous integumentary processes (Pennsylvanian). Bar  500 μm.

Figure 14.69 Median longitudinal section of Conostoma

ovoides showing lagenostome (L) central column (C), and pollen chamber (PC) (Pennsylvanian). Bar  1 mm.

from the Middle Pennsylvanian of Illinois (T. Taylor and Leisman, 1963). In this seed, the sclerotesta is composed of two distinct zones. Cells of the outer zone are aligned at right angles to the long axis of the seed and form prominent, irregular bands that girdle the seed (FIG. 14.70). A slightly different integumentary anatomy is present in C. villosum (Upper Pennsylvanian), where the outer layer of the integument (sarcotesta) consists of numerous radially elongated, multicellular processes (FIG. 14.71) (Rothwell and Eggert, 1970). Despite the large number of specimens known, it is surprising that no cellular megagametophytes have ever been reported in this genus, although megaspores have been described. Pollen grains have been found in the pollen chamber of several species, however, and they are typically in the 50to-80-μm size range and reveal reticulate markings on their walls. Many authors have suggested that Conostoma was produced by Heterangium, although this speculation is based only on the common association of the seeds and stems at some collecting sites.

CORONOSTOMA Coronostoma quadrivasatum is the name of a small (4.5 mm long), Late Pennsylvanian seed that superficially resembles Conostoma (Neely, 1951). The integument is constructed of an outer uniseriate layer of radially elongated cells and an inner region of axially aligned fibers. Four vascular bundles traverse the inner layer of the integument. The most distinctive feature of this species is the tubular elongation of the lagenostome which extends to the base of the micropylar orifice. PHYSOSTOMA Physostoma has an integument constructed of relatively thinwalled cells. Vascularized integumentary segments are free above the level of the pollen chamber, which is bellshaped. In P. elegans, known only from the Pennsylvanian of England (and perhaps the Donets Basin of Russia; see Snigirevskaya, 1989), the surface of the seed is covered with large, quill-like unicellular hairs that are believed to have been mucilaginous (Oliver, 1909). Physostoma calcaratum (Middle Pennsylvanian of Kansas) has 6–10 integumentary lobes that are fused in the lower half of the seed. Each lobe is vascularized by a single, terete strand. The extension of the lagenostome (salpinx) is relatively short, and a central

558

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.72 Cross section of the apex of Tyliosperma orbiculatum showing integumentary lobes (Pennsylvanian). Bar 1 mm.

Figure 14.71 Diagrammatic cutaway of Conostoma villosum.

(From Rothwell and Eggert, 1970.)

column is apparent near the top of the pollen-receiving structure (Leisman, 1964). Another type of pteridosperm ovule, Oclloa (Early Pennsylvanian of the Paracas Peninsula, Peru) closely resembles Physostoma but is acupulate (Erwin et al., 1994). Oclloa cesariana is 1.1 cm long by 0.4 cm wide and borne singly on slender forked, ultimate branchlets of laterals which are alternately arranged along a central axis. The integumentary lobes are fused laterally and to the nucellus in the basal one-half to two-thirds of the seed. At levels just below the pollen chamber, the integument separates distally into a whorl of five to nine free, tentacle-like lobes. The megasporangium displays hydrasperman features, including a pollen chamber, short lagenostome, and remnants of a central column. Found exclusively associated with Oclloa are microsporangiate branch systems named Obandotheca laminensis that bear groups of two to four unfused, bananashaped sporangia. TYLIOSPERMA A cupule of unusual structure characterizes the Middle Pennsylvanian seed Tyliosperma (Mamay, 1954b). The seed

is 3.7 mm in maximum length and has almost the same diameter. At the distal end, the integument is divided into seven apical lobes that form a canopy over the pollenreceiving mechanism (FIG. 14.72). The nucellus and the integument are fused to near the apex, where a simple extension of the nucellar tip extends to the base of the micropyle. A central column is present in the pollen chamber of T. orbiculatum. The seed developed in a fleshy cupule consisting of seven to eight vascularized segments fused only at the base. The cupule segments are irregularly thickened. CALATHOSPERMUM Calathospermum is a cupule from the Mississippian of Scotland that is 4.5 cm long and composed of six, apically free segments (Walton, 1953a). It has been estimated that the cupule of C. scoticum produced about 70 stalked ovules. The ovule-bearing stalks are attached to the base of each cupule segment and contain numerous unicellular hairs. The ovules are 3 mm long and have been regarded as developmentally younger specimens of Salpingostoma dasu. Extending from the apex of the integument are several elongated processes. The position of the ovule stalks suggests that the organ is bilaterally symmetrical, and this feature, together with the C-shaped vascular strand of the cupule pedicel, has been used to support the concept that the Calathospermum cupule is homologous with a frond or part of a frond (Barnard, 1960). Such a structure may have evolved by the enrolling or folding of a frond segment, rather than by the fusion of a whorl of branch segments (telomes) as is suggested by other cupule types.

CHAPTER 14

PALEOZOIC SEED FERNS

559

C

Figure 14.73 Two seeds of Gnetopsis elliptica within a cupule (C) (Pennsylvanian). Bar  1 mm. (Courtesy J. Galtier.)

GNETOPSIS Multiovulate cupules characterize Gnetopsis (FIG. 14.73), but these are appreciably smaller (6.4 mm long) than those of Calathospermum. Gnetopsis elliptica, from the Pennsylvanian of France, consists of a cupule that is almost entire at the distal end; cupule segments are identified by very shallow indentations (Renault and Zeiller, 1884). Each cupule bears two to four seeds, each ~2.5 mm long. The integument and the nucellar beak are similar to those of Conostoma, except that three to four elongated, hair-covered plumes extend from near the margin of the micropyle distally to the end of the cupule. Gnetopsis hispida is a small (8 mm long), compressed seed from the Lower Mississippian that possesses four to six elongated extensions from the micropylar region of the integument (Gensel and Skog, 1977), and similar structures occur in G. robusta (FIG. 14.74). Along the surface of these structures are numerous hairs. The specimens differ from the anatomically preserved Gnetopsis seeds in that they possess two slender, winglike extensions of the seed coat. These hairlike

Figure 14.74 Gnetopsis robusta showing elongated hair covered plumes extending from near the micropyle (Mississippian). Bar  2 mm. (Courtesy H. W. J. Van Amerom.)

structures may have served to trap wind-borne pollen, directing the grains to the region of the micropyle. MEGATHECA In addition to seed-bearing cupules, several cupules lacking seeds have been described from Carboniferous rocks, suggesting that at some stage of their development the seeds were shed. Megatheca thomasii is a cupule from Mississippian oil shales of Scotland (Andrews, 1940a). It is large, measuring up to 6.2 cm in length. The distal end

560

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.76 Suggested reconstruction of Simplotheca silesi-

aca (Mississippian). (From Millay and Taylor, 1979.)

Figure 14.75 Suggested reconstruction of Zimmermannitheca

cupulaeformis showing nonsynangiate organization (Mississippian). (From Millay and Taylor, 1979.)

of this compression consists of six free lobes composed of fibers that are much like the sclerotesta of most Paleozoic seeds. Nothing is known about the contents of Megatheca, but the size of the unit suggests that it might have been multiovulate. OTHER LYGINOPTERIDS: POLLEN ORGANS

The pollen organs of the lyginopterids are less well known than the seeds (FIGS. 14.75, 14.76). On the basis of the branching pattern and arrangement of the sporangia, two distinct groups are identified. One type, exemplified by Crossotheca and Feraxotheca (see above), is characterized by synangia borne near the ends of ultimate pinnae on a planated frond segment. The other type includes pollen organs borne on monopodial branching systems that may have replaced a portion of a vegetative frond or an entire leaf (see Van Amerom, 1975). This type includes the permineralized form Telangium (Benson, 1904) and the compression genus Telangiopsis (Eggert and Taylor, 1971). Telangium scottii is a rarely encountered species from Westphalian A (Early Pennsylvanian) rocks of Great Britain. The synangia are 1.7 mm long and composed of eight thick-walled, basally fused sporangia. In cross section, there is a distinct difference in thickness between the outward-facing sporangial walls and the thinner, inward-facing walls. This suggests that

dehiscence took place toward the center of the unit. Several different preservational states of Telangium fertile structures have been found in rocks from the upper Chesterian series (Mississippian) in Illinois. This Telangium is a small, cupshaped synangium composed of eight sporangia arranged radially (in a ring) (Jennings, 1976). Individual sporangia are basally attached to a vascularized parenchymatous core. Dehiscence is similar to that of T. scottii. The pollen is unusual in that grains with both trilete and monolete marks are found among the sporangial contents. Telangiopsis was instituted for synangiate clusters that appear morphologically similar to Telangium but are preserved as compressions. Telangiopsis arkansanum (Mississippian) consists of radial clusters of five to six sporangia borne terminally on either dichotomously or monopodially branched axes lacking planated foliar structures. Some of the compressed synangia suggest that during dehiscence the sporangia spread apart. The pollen grains up to μm in diameter and are trilete. Like those of the Late Mississippian petrifaction specimens, they conform to the generic circumscription of the sporae dispersae genus Punctatisporites. Meyer-Berthaud (1989) has reviewed the reproductive organs of Mississippian progymnosperms and gymnosperms. She identified several basic types, with the simple synangium type Telangium first identified in the upper Tournaisian (Long, 1979). These are followed in younger Mississippian rocks by several other synangiate types. One of these is Phacelotheca pilosa, a fructification of late Visean age from Scotland (Meyer-Berthaud and Galtier, 1986b). The synangium is stalked and consists of eight elongated sporangia fused only at the base. Dehiscence was accomplished through the thinner, ventral walls of the sporangia. Pollen was small, 29–45 μm, and trilete.

CHAPTER 14

Melissiotheca is another late Visean synangium described from Scotland (Meyer-Berthaud, 1986). In M. longiana, the synangium contains from 50 to 150 sporangia embedded within a parenchymatous cushion. Structurally, this fructification is organized like several compound forms associated with the medullosan seed ferns (Millay and Taylor, 1979). Radial and trilete pollen is a characteristic of many lyginopterid pollen organs. Similar grains occur in Schopfiangium, a simple synangiate organ from the Pennsylvanian of North America (Stidd et al., 1985). In this genus, however, the pollen wall is alveolate, whereas in all other lyginopterid pollen the sporoderm is homogeneous and lacks internal ornamentation (T. Taylor and Millay, 1981a). The first-appearing lyginopterid pollen organs consist of simple synangia of the Telangium type, such as Zimmermannitheca (FIG. 14.75). By Visean time, several larger forms are present that include simple synangia aggregated into multiple units and compound types such as Melissiotheca. All three of these types are present in Pennsylvanian age rocks. Although these forms indicate a particular level of lyginopterid pollen organ evolution, they contribute little to our understanding of how and where the organs were produced on the frond system that bore them. Dichotangium is a Mississippian lyginopterid pollen organ that is attached to an axial system (Rowe, 1988b). In D. quadrothecum, the branch system is isotomous, with the synangiate organs attached to branch tips. Each fructification consists of a flattened pad bearing 12 or 24 unfused sporangia. Sporangia are 2 mm long and possess acuminate tips that curve toward the center of the unit. Nothing is known about the pollen, since the specimens consist principally of impressions. Rowe (1988b) suggested that the synangiate organs were borne on the frond system Diplopteridium holdenii (FIG. 14.77), which consisted of bifurcate or trifurcate fronds. Pinnules are planated and highly dissected. Clusters of cupules are attached in the region of the first bifurcation of the primary frond axis. The earliest lyginopterid pollen organs (Telangium type) are not morphologically far removed from the clusters of terminal sporangia that characterize several progymnosperms. The change from a progymnospermous fertile region to the form present in the lyginopterids could have been effected by the evolution of a planated frond system, which would bear the synangia on its abaxial surface. Planated, pinnate fertile branch systems occur in the Pennsylvanian genus Canipa (FIG. 14.78) (Skog et al., 1969). The foliage morphotaxon Rhacopteris is used for Mississippian age pinnate fronds bearing asymmetrical pinnules with open dichotomous venation. The genus is interesting because several specimens are known with

PALEOZOIC SEED FERNS

561

Figure 14.77 Suggested reconstruction of Diplopteridium holdenii. (From Rowe, 1988.)

Figure 14.78 Suggested reconstruction of Canipa quadrifida. (From Skog et al., 1969.)

microsporangiate reproductive structures in organic connection, for example, R. paniculifera (Stur, 1883). In these specimens, the rachis is distally forked and gives rise to two fertile branches where each bears pinnately arranged clusters of sporangia (FIG. 14.79). Fern, progymnosperm, and seed fern affinities for Rhacopteris have been suggested. Permineralized specimens from the Dinantian of Scotland, however, suggest affinities within the lyginopteridalean pteridosperms (Galtier et al., 1998b).

562

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.79 Rhacopteris paniculifera with microsporangiate

structures (Mississippian). Bar  1.5 cm. (Courtesy GBA.) 14.80 Suggested reconstruction of Eremopteris zamioides. (From Delevoryas and Taylor, 1969.)

Figure

Eremopteris zamioides is an interesting Pennsylvanian foliage type thought to have represented the leaves of a seed fern, principally because of the association of seeds and pollen-producing structures (FIG. 14.80) (Delevoryas and Taylor, 1969). The leaves of this plant are 20 cm long and consist of alternately to oppositely arranged pinnae (FIG. 14.81) (White, 1943). Pinnules are narrow and lack a midvein. The base of the rachis is sharply truncated, suggesting that the leaves abscised. Associated with the leaves are numerous, flattened, bilaterally symmetrical seeds with two prominent distal spines or horns (FIG. 14.81). The seeds range up to 1.5 cm long and 6 mm wide in the primary plane. On some specimens, the integument is split, revealing a granulose megaspore membrane. Morphologically, the seeds

appear similar to the structurally preserved Mississippian seed Lyrasperma (Long, 1960b). Numerous star-shaped structures are associated with the foliar and seed remains, and these represent either clusters of whole sporangia or portions of longitudinally dehisced synangia that may have been the pollen organs of E. zamioides. These structures may have been produced on associated foliar structures, and they have the same basic pinnate arrangement as the leaves; but exhibit short projections on the laterals (pinnae), which may have been the attachment points of either seeds or sporangia.

CHAPTER 14

PALEOZOIC SEED FERNS

563

Figure 14.82 Cross section of Stenokoleos sp. stem (Mississip-

pian). Bar  1 mm.

Figure 14.81 Eremopteris zamioides. Arrow indicates bicor-

nate seed (Pennsylvanian). Bar  2 cm.

INCERTAE SEDIS

A number of Late Devonian–Permian petrified axes and impression–compression fossils have been described as incertae sedis, because their affinities cannot be accurately determined. Some of these may represent pteridosperms, but others may be progymnosperms or early fernlike plants. Stenokoleos is used for permineralized axes that bear alternately arranged pinnae with elliptical to clepsydroid-shaped traces (FIG. 14.82). The primary xylem of the stem consists of a two- to five-lobed stele with mesarch protoxylem and weakly developed peripheral loops; tracheids are scalariform or have uni- or biseriate circular-bordered pits. In cross section, S. simplex (Late Devonian) reveals triangular traces in the rachis (Beck, 1960c). Other species, for example, S. bifidus, are distinguished based on variability in the configuration of the xylem (Matten and Banks, 1969; Matten, 1975). The ribbed primary xylem columns of S. holmesii from the Givetian Cairo flora of New York have been demonstrated to be important in the development of the trifurcate branching pattern of that plant (Matten, 1992). In many features, Stenokoleos is similar to Tristichia and Tetrastichia. Rothwell and Erwin (1987) and Serbet and Rothwell (1992) reported on Stenokoleos-like

Figure 14.83 Cross section of Elkinsia polymorpha stem showing three lobed stele and traces (arrows) (Devonian). Bar  1 mm. (Courtesy R. Serbet.)

anatomy in basal gymnosperms such as Elkinsia (FIG. 14.83), which suggests that some Stenokoleos may represent progymnosperms (see Scheckler et al., 2006). The discovery of reproductive parts associated with these stems will demonstrate conclusively whether these axes represent lyginopterid seed ferns or some progymnosperm ancestor. Two Early Mississippian (New Albany Shale) taxa that may ultimately turn out to be the petioles of seed ferns are Periastron (Beck, 1978; Gastaldo, 1995) and Aerocortex (Beck, 1978). Specimens of P. reticulatum vary from circular to elliptical in cross section and contain a nearly median row of 5–10 slightly curved amphicribral vascular bundles (FIG. 14.84).

564

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.84 Cross section of Periastron sp. showing aer-

enchymatous cortex (Mississippian). Bar  5 mm. (Courtesy J. Galtier.)

Tracheids have helical–reticulate secondary wall thickenings. The cortex is aerenchymatous with longitudinal lacunae (air canals) and peripheral sclerenchyma associated with secretory canals or ducts. Aerocortex kentuckiensis has many of the same features as Periastron, but with fewer vascular bundles (two to four) and centrally located secretory ducts. The Austrocalyxaceae is a family of gymnospermous plants with uncertain affinities that has been erected based on compression fossils from the Carboniferous of the Jejenes Formation, San Juan province, Argentina (Vega and Archangelsky, 2001). The family includes two taxa that represent unforked frond portions with terminal, multiovulate cupules that may have basal, vegetative structures. These basal structures can be of the rhacopteroid type, as in Austrocalyx jejenensis, which includes specimens with a single terminal cupule, or of the triphyllopteroid type, as in Polycalyx laterale, which includes specimens with two drooping terminal cupules (FIG. 14.85) and P. tetramera, where there are four erect terminal cupules (Vega and Archangelsky, 2001). Cupules are relatively large, that is, up to 32 mm long and 57 mm wide in P. laterale and composed of two valves that are laterally fused in the proximal portion to form an entire base; each valve bears multiple long, free tentacles. Ovules contained in the cupules are sessile and orthotropous. Also tentatively assigned to this family are triphyllopteroid foliage remains with apical sporangia, that is, Rinconadia archangelskyi (Vega, 1995, 2000), and isolated seeds/ovules of Jejenia alata (Vega and Archangelsky, 1997, 2000), which are similar to the ovules observed in the cupules. The seeds are characterized by a long, apical integumentary extension with a putative micropylar channel that is well defined. The cupulate fructifications of the Austrocalyxaceae

Figure 14.85 Diagrammatic reconstruction of Polycalyx laterale fertile pinna showing drooping cupules. (From Vega and Archangelsky, 2001.)

suggest affinities in the Lyginopteridales. Arguing against lyginopteridalean affinities, however, is the fact that the taxa have rhacopteroid or triphyllopteroid unforked foliage, whereas lyginopteridalean foliage typically is forked (see above). An anatomically preserved stem from the Jejenes Formation, Amosioxylon australis, is 13 cm long and 6 cm in diameter (Césari et al., 2005). It consists of at least four strands of primary and secondary xylem that vary in size and shape and are embedded in ground tissue. Secondary xylem tracheids have alternate bordered pits and oblique elliptical apertures. The systematic affinities of A. australis remain elusive, but the fossil is tentatively related to a group of primitive seed ferns probably represented by the Austrocalyxaceae (Césari et al., 2005). Nystroemia reniformis (pectiniformis) is a compression taxon from the Upper Permian of China (Hilton and Li, 2003; J. Wang et al., 2003). Specimens from the Upper Shihhotze Formation, Shanxi (Shansi) Province, consist of branching systems comprising at least four orders of branching. Firstorder axes branch to one side only and give off second-order branches either singly or in pairs, which are of two types: one fertile and bearing characteristic ovulate branching systems and the other presumably vegetative. Ovulate axes produce multiple, closely spaced slender lateral branches in two

CHAPTER 14

Pollen chamber Pollen chamber floor

Salpinx Central column Tent pole

Megaspore membrane

Megagametophyte

Archegonium Nucellus

Integumentary vascular bundle

Integumentary lobe nucellar vascular tissue

Pedicel

Figure 14.86 Longitudinal section of typical hydrasperman gymnosperm ovule showing the component structures.

alternate to subopposite rows which bear ovulate branching systems. The ovulate branching systems produce planated ultimate axes of different lengths. Each ultimate axis bears a terminal ovule with 180° rotational symmetry and two distal hornlike integumentary projections. Additional parts of the Nystroemia plant from Shanxi include pollen organs originally described by Halle (1927, 1929) and detached Chiropteris reniformis foliage. The organic connection between N. pectiniformis ovulate axes and Chiropteris reniformis has been documented based on material from Dengfeng in Henan (J. Wang et al., 2003). Hilton and Li (2003) suggested that Nystroemia represents a late stratigraphic occurrence of a plesiomorphic hydrasperman-type seed plant (FIG. 14.86) with affinities closely allied to members of the Lyginopteridales. LYGINOPTERID EVOLUTION

One feature that appears to have been relatively constant in the lyginopteridaleans is the organization of the stele and presence of cortical sclerenchyma. The steles of the various taxa are of interest in that they demonstrate stages in the evolution of a true eustele from early types such as those in the Calamopityales. Such a sequence is suggested to involve the longitudinal dissection of a protostele (FIG. 7.43), followed by further dissection of the columns of vascular tissue into discrete, sympodial bundles arranged in a cylinder. Some of the lyginopterids demonstrate the dissection of a protostele by increasing amounts of parenchyma, for example, Microspermopteris and Heterangium. These stele forms demonstrate an intermediate stage between the protostelic, Devonian calamopityaleans and the Pennsylvanian eustelic forms. Galtier (1988) indicated that a basic difference exists between calamopityaleans and lyginopterids based on the

PALEOZOIC SEED FERNS

565

ratio of the diameter of primary vascular tissue to the diameter of the stem. In various specimens, he found that the ratio in the lyginopterids ranges from 1:2 to 1:4, whereas the ratio in the calamopityaleans is closer to 1:10. Although these differences are no doubt of biomechanical importance, the biological significance is unknown. These data may, however, prove useful in distinguishing other putative seed fern stems. There is probably little disagreement that the early pteridosperms had their origins among the aneurophytalean progymnosperms and that the pollen of the early seed ferns was only functionally different from the spores of progymnosperms. Hilton (2006), however, pointed out that the most recent cladistic analysis of lignophytes (progymnosperms plus seed plants) by Hilton and Bateman (2006) shows that the change from heterosporous progymnosperms to the earliest seed plants equates to at least nine unambiguous character changes at the node leading up to seed plants. This transition between progymnosperms and seed plants currently lacks viable intermediates that can be included in morphological cladistic analyses and suggests that either key taxa have still to be identified from the fossil record or that this was an episode of macroevolution through saltation or punctuated equilibrium (Eldredge and Gould, 1972) and that intermediates do not exist (Bateman and DiMichele, 1994b). Research with both early seed ferns and progymnosperms during the last decade has greatly narrowed the gap between the two groups, and the forthcoming years will likely result in additional fossils which provide the level of resolution needed to reconstruct reproductive organs and vegetative parts necessary to accurately trace the origins of the major groups of seed plants. An interesting aspect of the Lyginopteridales is their biogeographic distribution of the group across Euramerica during the Middle Pennsylvanian. On the basis of foliage fossils, Cleal (2008) documented that, during most of the middle Westphalian (Middle Pennsylvanian), the lyginopteridalean seed ferns had a very similar distribution to the Medullosales. This biogeographical pattern progressively breaks down, however, during uppermost Westphalian and lowermost Stephanian (early Late Pennsylvanian) times, along with a drastic decline in species diversity. Cleal (2008) suggests that the decline in the number of species was a result of lyginopteridaleans being out competed for resources by the reproductively more sophisticated Callistophytales. Phillips (1981), who documented a similar pattern of demise based on a quantitative analysis of anatomically preserved lyginopteridalean remains, speculated that the geologically abrupt loss of these plants across the Euramerican Floral Province was perhaps caused by some kind of changing interaction with animals.

566

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

MEDULLOSALES The medullosan seed ferns are known from the Late Mississippian into the Permian (Stidd, 1981). They were the largest of the seed ferns with unusual stem anatomy characterized by several segments of vascular tissue, each surrounded by secondary xylem. They are distinguished from other pteridosperms by the mode of leaf trace production in which terete strands are produced by more than one cauline protoxylem sympodia separated by several centimeters of stem (Mapes and Rothwell, 1980; Dunn, 2006). The petioles are massive and contain a large number of scattered vascular traces. Seeds are acupulate, typically large—some of the largest known— and radial in transverse section. Pollen organs are synangiate, and many produce large, monolete pollen grains. STEMS

MEDULLOSA Like a number of other well-known fossil plants, the generic name of the stem has been adopted in discussions of the entire plant. The genus Medullosa (FIG. 14.87) was initially established by B. von Cotta (1832) (FIG. 14.88) for Early Permian stem remains. In cross section, medullosan stems are characterized by one to many segments of vascular tissue (FIG. 14.89), and in earlier treatments they were considered to be polystelic. The stele of the plant is now regarded as a series of sympodia which constitute a single stele (Basinger et al., 1974). According to Dunn (2006), 15 species of Medullosa are currently recognized, and these show an interesting geographical distribution pattern: five are known from North America, that is, M. endocentrica, M. noei, M. olseniae, M. steinii, and M. thompsonii, and eight from Europe, that is, M. centrofilis, M. geriensis, M. gigas, M. stellata (FIG. 14.90), M. leuckartii (FIG. 14.91), M. porosa, M. pusilla, and M. solmsii. Only two species have been reported from both Europe and North America, M. primaeva and M. anglica. Medullosa noei (Steidtmann, 1944; Delevoryas, 1955b) is one of the most common Pennsylvanian species in North American coal balls and will be used to describe the basic histology of the stem (FIG. 14.89). In cross section, the stem contains two to four segments of vascular tissue embedded in ground parenchyma, as well as occasional rings of secondary parenchyma that superficially resemble additional vascular segments. The number of vascular segments varies from level to level as a result of branching and fusion of segments and has resulted in the naming of over 40 species and a number of varieties within the genus. Many of these taxa have been reported from relatively small stem sections and no doubt represent variations within the same biological

Figure 14.87 Suggested reconstruction of Medullosa noei (Pennsylvanian). (From Taylor and Taylor, 1993.)

Figure 14.88

Bernhard von Cotta.

CHAPTER 14

Figure 14.89 Cross section of Medullosa noei stem showing endocentric development of secondary xylem (Pennsylvanian). Bar  1 cm.

Figure 14.90 Medullosa stellata var. incrassata, portion of

stem in cross section (Permian). Bar  2 cm. (Courtesy BSPG.)

Figure 14.91 Medullosa leuckartii, portion of stem in cross

section (Permian). Bar  2 cm. (Courtesy BSPG.)

PALEOZOIC SEED FERNS

567

species. Each vascular segment is elliptical to band-shaped in cross section, with the primary xylem containing abundant parenchyma. Protoxylem is oriented toward the outer margin of the stele. Surrounding each primary vascular segment is a cylinder of secondary xylem of variable thickness that is characteristically more extensively developed toward the center of the stem or centripetally (Delevoryas, 1955b). Secondary xylem tracheids are very long, extending up to 24 mm (Andrews, 1940b), with 12 rows of bordered pits on their radial walls. Rays are multiseriate rays, two to eight cells wide, and extend through the wood about every three rows of tracheids. Surrounding the xylem is a cambium (Cichan, 1986b) and a zone of secondary phloem, consisting of alternating tangential bands of sieve cells, fibers, and axial parenchyma (Smoot, 1984b). Periderm is common in many stems of Medullosa. It is seen in the inner cortical zone and may be up to 1 cm or more in thickness. The cells are isodiametric, thick walled, and arranged in radial files. The disposition of the periderm in Medullosa suggests that the phellogen produced a phelloderm principally toward the inside of the stem, which differs from the typical method of periderm formation. In addition, it appears that the periderm tissue was living and continued to divide as in Lepidodendron. The remaining tissues of M. noei are composed of cortical parenchyma and abundant, elongated secretory canals filled with amorphous contents. These are several millimeters long and common in many of the plant parts of this genus. At the periphery of each canal is a layer of small epithelial cells. Patches of sclerenchyma fibers form interrupted bands around the periphery of the stem. On the basis of permineralized stem and associated foliage specimens from the Upper Pennsylvanian of the Appalachian Basin, Beeler (1983) suggested that Medullosa noei-type stems produced fronds assignable to Macroneuropteris scheuchzeri and Neuropteris ovata. Medullosa thompsonii is a smaller stem (FIG. 14.92) (4 cm in diameter) that contains three vascular segments surrounded by a narrow band of periderm (Andrews, 1945). Like M. noei, the species is known from Middle and Upper Pennsylvanian deposits. Medullosa endocentrica has slender stems up to 2.3 cm in diameter with two to three vascular segments making up the stele; secondary xylem development is conspicuously endocentric (centripetal). This Late Pennsylvanian species is known from coal balls collected near Berryville, Illinois (Baxter, 1949), and Steubenville, Ohio (Hamer and Rothwell, 1988). The small stems bear fronds in a 3/8 phyllotaxy, with axillary buds present at the base of the fronds. Each frond bifurcates at the base and is tripinnately compound above the dichotomy. Pinnules are comparable to

568

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

P P

Figure 14.92 Cross section of Medullosa thompsonii showing large petiole (P) (Pennsylvanian). Bar  1 cm.

Figure 14.93 Cross section of Medullosa primaeva showing central vascular tissue and three attached petiole bases. (Pennsylvanian). Bar  1 cm.

the Eusphenopteris type. Internode length in this species is quite variable, ranging from 0.1 mm to 12 cm long. Hamer and Rothwell (1988) suggested that M. endocentrica was a liana, based on the presence of long internodes, slender stems, and the absence of adventitious roots attached to the stem. Medullosa primaeva is a Middle Pennsylvanian species which contains quite a variable number of vascular segments that changes from level to level in a relatively short distance. The stems are 8 cm in diameter, with individual vascular segments averaging 7 mm by 2.5 mm (FIG. 14.93). Delevoryas (1955b) demonstrated an ontogenetic sequence for M. primaeva in which the number of vascular segments changes from 2 to 20. This species is also important in that it demonstrates both the presence and the absence of secondary xylem associated with the leaf traces, a feature previously considered taxonomically important. In M. anglica, a species known from coal balls of the Lower Coal Measures of Lancashire and the Middle Pennsylvanian of Kentucky, leaves were produced in a 2/5 phyllotaxy (FIG. 14.94) (Smoot and Taylor, 1981). Some secondary xylem is associated with the leaf traces and accessory vascular strands occur in the cortex. One of the geologically oldest species in the genus Medullosa is M. steinii from the Chesterian Series (Upper Mississippian) of Arkansas, USA (Dunn et al., 2003a). Medullosa steinii stems are only 2–3 cm in diameter and contain two to eight cauline vascular segments. Primary xylem bundles consist of one to several protoxylem poles with metaxylem and abundant parenchyma; each bundle is surrounded by secondary vascular tissue. Leaf traces typically diverge from a vascular segment as pairs, separated from one another by a wedge of secondary vascular tissue. Frond bases

P

Figure 14.94 Cross section of Medullosa anglica showing large petiole (P) partially surrounding vascular segments (Pennsylvanian).

Bar  1 cm.

CHAPTER 14

are of the Myeloxylon type, decurrent, and separated from one another by a discontinuous row of sclerotic bundles and from the stem by a prominent periderm. Other Mississippian medullosan stems have been described from the Imo Formation of Arkansas (T. Taylor and Eggert, 1967) and from the Upper Silesian Coal Basin in Poland (Brzyski and Stuchlik, 1992). The majority of the Early Permian medullosans differ markedly from the Carboniferous forms, principally in the arrangement and number of vascular segments (Rössler, 2001a,b). Medullosa stellata (FIG. 14.90) is one of the largest stems known (Weber and Sterzel, 1896), with partially decorticated specimens up to 48 cm in diameter. In some specimens, the stems contain 43 separate central vascular segments, sometimes referred to as star rings. Surrounding these is a cylinder of primary and secondary xylem. The secondary xylem occurs on both the sides of the primary growth, that is, centripetal and centrifugal xylem, and the centrifugal xylem is more developed. Medullosa porosa is similar to M. stellata but differs in that it contains two zones of small central strands (Cotta, 1832). Vascular segments of the outer zone are tangentially flattened, and secondary xylem occurs principally on the inner surface. A slightly different xylem configuration is present in M. solmsii. The segments near the periphery are numerous and arranged in two rings (Schenk, 1889). The central ground tissue contains several very small star rings that are poorly developed. The peripheral strands of M. leuckartii are tangentially elongated and sinuously arranged (Göppert and Stenzel, 1881). In the center of this stem are a variable number of small vascular segments with uniformly developed tissues (Solms-Laubach, 1897a) (FIG. 14.104). The largest specimen ever discovered of M. leuckartii was described and illustrated by Sterzel (1918) (FIG. 14.95). It is from Chemnitz, Germany, and measures 160 cm long and 26 cm  18 cm in diameter; this specimen shows that the stem bore densely positioned fronds, that is, it had short internodes, with long petioles (up to 60 cm) in a helical arrangement. OTHER STEM TAXA Additional stem genera referred to the Medullosales include Colpoxylon and Sutcliffia (Scott, 1906b). Colpoxylon is a monospecific genus known only from the Permian of France; it is very similar to a number of species in Medullosa except that there are no vascular strands (star rings) in the pith of Colpoxylon (Seward, 1917; Dunn, 2006). Sutcliffia insignis is known from both North America and Great Britain. Stems are 15 cm in diameter and contain a large, central vascular segment from which arise other vascular segments (FIG. 14.96). Sutcliffia stems may be further distinguished

PALEOZOIC SEED FERNS

Figure 14.95

569

Johann Traugott Sterzel.

Figure 14.96 Cross section of Sutcliffia insignis (Pennsyl-

vanian). Bar  4 mm.

from Medullosa by the presence of leaf traces that appear as miniature vascular segments lacking secondary xylem, the concentric nature of the petiole bundles, and the distribution of sclerenchyma surrounding the peripheral vascular traces in the petiole (Phillips and Andrews, 1963). Histologic features

570

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

of the vascular segments, cortex, and periderm are similar to those described for Medullosa. The presence of conspicuous, moundlike emergences on the surface of the stem and petiole bases has been used to identify S. insignis var. tuberculata. The frond of Sutcliffia is thought to have been 3–4 m long and 4–5 times pinnate. In cross section, the petiolar/rachial bundles radiate in curved rows from the center of the axis, with those of the outer ring embedded in the outer sclerenchymatous cortex. Pinnules assignable to the foliage morphotaxon Linopteris are thought to have been produced by S. insignis var. tuberculata because of their consistent association in coal balls containing Sutcliffia petioles (Stidd et al., 1975). Morphologically, they appear similar to pinnules of Paripteris but can be distinguished by their reticulate venation (Chapter 16). Laveine et al. (1993b) and Scott et al. (1997) noted that the characteristic, moundlike emergences seen in S. insignis var. tuberculata support the suggested affinities with the foliage taxa Linopteris and Paripteris, in which comparable emergences are known (Laveine 1967). Although the preceding features have been used to separate Medullosa and Sutcliffia, the variation known among medullosan species may ultimately demonstrate Sutcliffia as a species of Medullosa or a developmental stage of a currently recognized taxon. Quaestora amplecta is a Mississippian permineralized stem that demonstrates features suggestive of affinities with the medullosan pteridosperms (Mapes and Rothwell, 1980). The specimen is 4.5 cm in diameter and consists of an exarch, cruciform (FIG. 14.97), vitalized protostele surrounded by dense secondary xylem and secondary phloem, with a narrow cortex. Leaves are produced in an opposite decussate pattern (FIG. 14.98), with numerous peripheral traces surrounding a smaller number of central ones; rachis bases are highly decurrent and of the Myeloxylon type. Leaves and reproductive structures are currently unknown, but it was likely an understorey tree (Dunn, 2004). Quaestora is the most primitive type of medullosan stem currently known. Although it shares a number of features with the calamopityaleans and medullosans, including manoxylic wood, sparganum cortex, large metaxylem tracheids interspersed with xylem parenchyma, vascular architecture, leaf-trace production, and the presence of secretory cells and resin rodlets, the peripheral and internal traces in the petioles of Quaestora are most similar to petiole anatomy in Medullosa (Stein and Beck, 1992). Another pteridospermous stem that shares some features with the medullosans is Eoguptioxylon antiqua from the Upper Permian La Antigua Formation of Argentina (Crisafulli and Lutz, 2007). The axes consist of a number of meristeles irregularly dispersed in parenchymatous ground tissue and surrounded by sclerenchyma. Each meristele is composed of

Figure 14.97 Cross section of Quaestora amplecta (Missi-

ssippian). Bar  1 cm. (Courtesy G. W. Rothwell.)

Figure 14.98 Cross section of Quaestora amplecta stem

showing opposite petioles. (From Mapes and Rothwell, 1980.)

primary and secondary xylem and characterized by a central diaphragm pith with lacunae. Other features of E. antiqua, such as pycnoxylic secondary wood with uni- to biseriate pits and araucarioid cross-field pits, more closely correspond to conifer wood. LEAVES (FRONDS)

Three of the more common late Paleozoic foliage morphogenera thought to have been produced by the medullosans are Neuropteris s.l., Odontopteris, and Alethopteris (Chapter 16).

CHAPTER 14

PALEOZOIC SEED FERNS

571

Figure 14.99 Cross section of Myeloxylon showing numer-

ous scattered vascular bundles in ground tissue (Pennsylvanian). Bar  1 cm.

Among these types, Neuropteris is particularly interesting with regard to the earliest occurrence of the medullosan seed ferns, since this taxon has been described from the Culm sequences (Tournaisian-Visean) in the Silesian and Saxothuringian basins of central Europe (Hartung, 1938; Havlena, 1969) and thus predates the earliest unequivocal anatomically preserved medullosan stem (Medullosa steinii) from the lower Namurian A (uppermost Mississippian) of the Fayetteville Shale, Arkansas (Dunn et al., 2003a). Medullosan fronds typically branch dichotomously, or slightly unequally, in the lower portion of the frond, and are regularly pinnately compound (Beeler, 1983). Anatomically preserved petioles and compression remains indicate that some fronds were massive structures. In some forms, the primary rachis or stipe may have attained lengths of nearly 2 m before undergoing the bifurcation, and Laveine (1986) reported on an Alethopteris specimen observed in a coal mine in northern France that was 7.4 m long and 2 m wide. Numerous stem specimens with attached petioles suggest that a 2/5 or 1/3 phyllotaxy was common. The detached petioles of Medullosa are termed Myeloxylon (FIG. 14.99) and were up to 20 cm in diameter; some of the earliest records for Myeloxylon-type petioles come from the Namurian A (middle Mississippian) of Belgium (Gerrienne et al., 1999b) . In cross section, Myeloxylon petioles are circular or elliptical (FIG. 14.99) and consist of a parenchymatous ground tissue with a ring of peripheral strands and scattered, centrally located vascular bundles (FIG. 14.100). These are

Figure 14.100 Diagrammatic reconstruction of Alethopteris/ Myeloxylon type medullosan frond. Arrow indicates position of Bernaultia-type pollen organ. (From Ramanujam et al., 1974.)

more abundant near the periphery of the rachis, decreasing toward the center. Surrounding the axis is a zone of hypodermal fibers similar to those of the stem. Scattered throughout the ground tissue are numerous secretory canals. A wounded Myeloxylon specimen indicative of arthropod herbivory, along with evidence for wound healing, has been described by Holden (1910). A leaf base on the stem is first recognizable as a small triangular, unvascularized area of cortex delimited by sclerenchyma strands. Numerous divisions of two or more primary xylem sympodia produce traces that pass through the secondary xylem. Continued divisions produce the large number of bundles apparent in the Myeloxylon petiole. At the level

572

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

of the penultimate pinnae, the axis lacks sclerenchyma fibers near the periphery, and the vascular bundles are arranged in a C-shaped configuration. ROOTS

Roots of Medullosa are adventitious and represent a common component of many petrified Carboniferous remains. Specimens attached to stems of M. anglica are typically triarch and appear in vertical series on the stem. Older roots contain a wide band of periderm. Some Middle Pennsylvanian roots of Medullosa are known in detail and have provided a basis not only for determining stages in tissue differentiation but also for suggesting characteristics of potential taxonomic significance (Rothwell and Whiteside, 1974). Some of the larger specimens are 2.5 cm in diameter and exhibit abundant secondary tissues (FIG. 14.101). In these roots, the center of the axis consists of an exarch actinostele with up to five protoxylem points. The secondary xylem is continuous around the primary body, a situation that is different from that of other Medullosa roots, which have wedges of secondary xylem opposite the metaxylem (Steidtmann, 1944). Secondary xylem tracheids are angular in cross section and exhibit 4–12 rows of crowded pits on the radial and oblique tangential walls. Secondary phloem is well developed in larger roots and consists of sieve cells and rays. Lateral roots originate from a thin-walled pericycle opposite to the protoxylem strands. Surrounding the pericycle is a narrow endodermis and parenchymatous cortex. Histologic variations, including the number of protoxylem strands and the presence or absence of such tissues as endodermis, may reflect species differences or constitute stages of development.

Figure 14.101 Cross section of Medullosa root (Pennsyl-

vanian). Bar  3 mm.

GROWTH HABIT

Hypotheses have been advanced about the growth habit of individual medullosan seed ferns based on both permineralized specimens (Stewart and Delevoryas, 1956; Hamer and Rothwell, 1988) and impression–compression material (Bertrand and Corsin, 1950; Krings and Kerp, 1999; Zodrow et al., 2007). They suggested that medullosan seed ferns were a diverse group of plants that included freestanding trees, semi-self-supporting leaners, and climbing vines or lianas. Forms such as M. stellata from the Upper Carboniferous–Lower Permian of Germany were most likely self-supporting trees, based on stem diameters of up to 50 cm (see above). Narrow stemmed forms, however, are more difficult to interpret, as biomechanical analyses suggest that the stem architecture of narrow-stemmed medullosans, as measured by the proportional abundance and spatial arrangement of ground, conducting, and mechanical tissues, is generally characteristic of nonself-supporting plants, rather than selfsupporting, treelike forms (Mosbrugger, 1990). Most specimens of medullosan stems do not provide information on the absolute height of the plant, which is an important parameter required in determining mechanical stability. For example, Wnuk and Pfefferkorn (1984) described a compressed Middle Pennsylvanian medullosan stem segment from the Bernice Basin (Pennsylvania, USA) that is 4.88 m tall and up to 5.3 cm in diameter. Since a freestanding medullosan trunk of this diameter is believed to be mechanically stable only up to a maximum height of 3 m (Mosbrugger, 1990), this stem would be interpreted as produced by a non-self-supporting, leaning or lianescent plant. Different growth habits of narrow-stemmed medullosan seed ferns can be visualized, depending on the absolute height of the plant. If mature individuals were short-statured, for example, up to 1.5–2.5 m high, and produced only a relatively small number of fronds, they may have been mechanically stable and self-supporting. Conversely, if the plants were 5–6 m tall and produced a larger number of relatively large fronds, it would be difficult to envisage a self-supporting habit, and these plants were probably leaners or lianas (Andrews, 1945; Krings et al., 2003c). It is also possible that leaning and lianescent, narrow-stemmed medullosans began their lives as self-supporting seedlings but, as growth continued, became mechanically unstable and relied on external support for their continued vertical growth (Krings and Kerp, 2006), a growth sequence not uncommon in some modern plants. Plant parts modified into specialized attachment devices, which are similar or identical to those seen in extant vines and lianas, are critical in delimiting lianescent forms among the seed ferns. Unfortunately, most of the evidence gathered to date on the

CHAPTER 14

attachment devices produced by leaning and lianescent seed ferns comes from compression specimens and cuticle preparations (FIG. 14.102) (Chapter 16) that cannot normally be correlated with permineralized stems. It is interesting that the oldest known representative of the genus Medullosa, M. steinii from the Upper Mississippian of Arkansas (see below), displays a maximum stem diameter of only 3.5 cm (Dunn et al., 2003a). These authors suggested that the large-stemmed, treelike medullosans represent derived forms that attained greater stem stability by various deviations from the typical stem construction, including enlarged and persistent frond bases, for example, in M. noei, or concentric rings of exocentric xylem, as in M. stellata. SEEDS

To date, information about the attachment of seeds to the medullosan fronds has come principally from compression–impression fossils. For example, Kidston (1905) described large (3 cm long) seeds of the Trigonocarpus type

Climbing organ (tendril) of the medullosan seed fern Blanzyopteris praedentata; reconstruction based on evidence from cuticles (Pennsylvanian). Bar  2.5 cm (From Krings and Kerp, 1999).

PALEOZOIC SEED FERNS

573

(FIG. 14.103) from the Middle Coal Measures of Britain. The seeds are longitudinally striate and radially symmetrical. Two pinnules of Neuropteris heterophylla are present at the base of the seed, and Kidston suggested that the ovule may have occupied the position of a terminal pinnule on the rachis. However, a trigonocarpalean seed has been documented from the Pennsylvanian of the Sydney Coalfield, Nova Scotia (Canada), which is attached to a pinna axis of Neuropteris flexuosa, but does not replace a pinnule (Zodrow and McCandlish, 1980). Several examples of seeds attached to foliage were described by Halle (1929) from the Permian of China. Alethopteris norinii is a tripinnate frond that produced large seeds (4 cm long by 1.2 cm diameter) attached to the midrib of the pinna, with typical alethopterid pinnules both above and below the point of seed attachment. Smaller seeds (7.5 mm long) that are probably bilaterally symmetrical are attached in the same position to pinnae of the Lonchopteris-type. Additional information about the modes of attachment of ovules or seeds in medullosan seed ferns can be found in Zodrow (2004).

Figure 14.102

Figure 14.103 Trigonocarpus sp. (Pennsylvanian). Bar 1 cm.

574

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.105 Harriette V. Krick Bartoo.

Figure 14.104

Hermann von Solms-Laubach.

Isolated seeds assignable to the medullosan pteridosperms represent a relatively common type of fossil and are known from various preservation states. Medullosan seeds are the largest of any seed fern group and, when found permineralized, are structurally almost identical to the seeds of extant cycads (Harper et al., 1970). The basis of classifying the structureless forms with the medullosan seed ferns has been their relatively large size and the presence of three longitudinal ribs that divide the integument into three equal valves. Seeds of this type are common as sandstone casts and may exhibit variable shapes, depending on their method of formation. The morphogenus Trigonocarpus was originally instituted by Brongniart in 1828 for such casts, but it has also been applied to compression forms (Krick, 1932) (FIG. 14.105) and was used earlier for structurally preserved specimens. Several other morphogenera of medullosan ovules or seeds preserved as casts or compressions have been discriminated from Trigonocarpus on the basis of morphological features, geologic age, and provenance (e.g., Suchov, 1969; Naugolnykh, 1997). Casts of T. leeanus (FIG. 14.103), a Middle Pennsylvanian form, are 10 cm long (Gastaldo and Matten, 1978). This species is interesting because it not only demonstrates features of the integument, including the extent of the ribs, but also contains a cast of the nucellar region. Trigonocarpus casts typically

represent the external features of the ovule, but in the case of T. leeanus, it was possible to make a latex transfer of the nucellar cast, and this revealed features of the inner surface of the integument. This technique, especially with very finegrained casts, may hold promise in more accurately defining the large number of seeds preserved as compressions and casts. In some instances, it is difficult to determine whether a seed cast was formed from the outer surface of the integument (sarcotesta) or the inner surface. Similarly, it remains a difficult task to determine whether the longitudinal striations on some casts represent the ribs of the integument or the positions of vascular strands that extend through the nucellus. In one particular type of cast, fine-grained sediment filled the inside of the nucellus, forming what have been termed nucule casts. Fossils of this type rarely show any longitudinal striations that mark the former position of the integumentary ribs. There are a number of radially symmetrical, structurally preserved seeds thought to belong to the medullosan seed ferns (FIGS. 14.106, 14.107). These are generally large (1 cm long), with a ribbed sclerotesta, and contain a nucellus that is attached to the integument only at the chalaza (FIG. 14.108). Pachytesta is the most common and largest genus of anatomically preserved medullosan seeds. No doubt many of the species of Trigonocarpus represent different preservational states of Pachytesta. The genus currently consists of more than a dozen species

CHAPTER 14

PALEOZOIC SEED FERNS

575

Figure 14.107 Diagrammatic reconstruction of Polypterospermum renaultii (Pennsylvanian). (From Combourieu and Galtier, 1985.)

Figure 14.106 Diagrammatic reconstruction of Polylophospermum stephanense (Pennsylvanian). (From Combourieu and Galtier, 1985.)

from the Lower Coal Measures (Pennsylvanian) of Britain, the Upper Pennsylvanian of France (T. Taylor, 1965) and Spain (Martín-Closas and Martínez-Roig, 2007), the uppermost Pennsylvanian–Lower Permian of Bulgaria (Tenchov, 1977), and the Lower Permian of China (Hilton and Cleal, 2007). Specimens range from slightly 1 cm in P. pusilla to 11 cm in length. In some large species of Pachytesta such as P. gigantea (FIG. 14.109), the cuticle of the endotesta (integument) and the nucellus closely conform with one another, suggesting that the integument and nucellus may have been fused along the length of the ribs (Smoot and Taylor, 1983b). The pollen chamber in Pachytesta is

simple, that is, there is no salpinx, and both the integument and the nucellus contain vascular tissue (Taylor and Delevoryas, 1964). Pachytesta illinoensis, a common species in North American permineralizations, is ovoid and 2.5 cm in diameter (Stewart, 1954). Some specimens range up to 4.5 cm long (FIG. 14.110) and consist of a three-parted integument that includes a parenchymatous outer layer (sarcotesta), a middle zone of fibers (sclerotesta), and a uniseriate, innermost endotesta. Secretory canals identical with those of Myeloxylon and Medullosa are common in the seed coats of Pachytesta. The sclerotesta is ornamented by three prominent ribs that extend from the base to near the micropylar opening. Equidistantly arranged between these primary ribs may be smaller secondary and tertiary extensions of the sclerotesta (FIG. 14.111). Although the nucellus is attached to the integument only at the base of the seed, it was tightly appressed to the inner surface of the integument when the seed was alive. At the distal end, the nucellus forms a simple, bell-shaped pollen chamber with a small nucellar beak that communicates directly with the micropylar opening. The vascular system of the integument of P. illinoensis consists of up to 42 distinct strands; the nucellus contains 25 bundles, some of which are laterally fused.

576

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

MI

PC

I

Figure 14.108 Median longitudinal section of Pachytesta stewartii showing conspicuous chamber in base. (From Taylor, 1965.)

A reduced number of secondary and tertiary ribs distinguish P. composita (Stewart, 1958). One of the largest anatomically preserved members of the genus is P. gigantea (FIG. 14.109), initially described from the Stephanian of France and now also known from Middle and Upper Pennsylvanian rocks of North America. Some specimens exceed 7 cm in length, but unlike other species, the sclerotesta of the seed is not ornamented with ribs (T. Taylor, 1965). A cross section at midlevel of this seed shows that the integument consists of three parts marked by indentations on the inner surface of the integument. A small number of secondary ribs are present only in the apex. Fibers of the sclerotesta contain simple pits on their walls. Fiftyone peripherally located vascular strands are present in the

Figure 14.109 Composite longitudinal section of Pachytesta gigantea showing micropyle (MI), integument (I), and pollen chamber (PC) (Pennsylvanian). Bar  2 cm.

integument, whereas the nucellus contains about 40 tangentially flattened bundles made up of scalariform tracheids. Pachytesta vera (Middle Pennsylvanian) is about the same size as P. gigantea but differs in having three prominent ribs (FIG. 14.112) (Hoskins and Cross, 1946). A slightly different sculpturing of the integument is present in P. berryvillensis, a species described from a large number of specimens differing in stages of pollen chamber development (T. Taylor and Eggert, 1969). The ovoid seeds are 7 mm long and contain a short micropylar tube. The positions of the three primary ribs are marked by slight indentations on the inner integument surface. The integument includes sarco-, sclero-, and endotestal tissues, and the inner surface of the sclerotesta contains fibers that extend into the seed cavity a short distance to form a reticulum of lacunae (FIG. 14.113). In

CHAPTER 14

PALEOZOIC SEED FERNS

577

S

SL

Figure 14.112 Cross section of primary rib of Pachytesta vera showing sarcotesta (S) and sclerotesta (SL) (Pennsylvanian). Bar  3 mm.

Figure 14.110 Longitudinal section of Pachytesta illinoensis.

Arrow indicates collapsed pollen chamber (Pennsylvanian). Bar  1 cm.

Figure 14.113 Cross section of Pachytesta berryvillensis showing ornamented inner surface of the sclerotesta (Pennsylvanian). Bar  1 mm.

Figure 14.111 Cross section of Pachytesta illinoensis showing

conspicuously ribbed sclerotesta (Pennsylvanian). Bar  1 cm.

the pollen chamber of P. berryvillensis, numerous saccate grains of the Florinites type are seen, but they may represent anemophilous contaminants. In most of the other species, pollen grains of the Monoletes type have been identified and are thought to represent the pollen biologically associated with Pachytesta. Structurally preserved megagametophytes have been identified in only a few species of Pachytesta (FIG. 14.114). The tissue inside the megaspore membrane shows what

578

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

M

Figure 14.114 Cross section of Pachytesta sp. with well-developed megagametophyte tissue (M) (Pennsylvanian). Bar  2 mm.

A

Figure 14.115 Longitudinal section of apex of Pachytesta hexangulata showing position of pollen grain (arrow) in pollen chamber and archegonium (A) (Pennsylvanian). Bar  3 mm.

are interpreted as stages in cellularization. Cells near the periphery of the megaspore are arranged in radial files, whereas those more deeply positioned within the megagametophyte are polygonal in outline. The only record of clearly distinguishable archegonia in Pachytesta is from

Figure 14.116 Wilson N. Stewart.

P. hexangulata (FIG. 14.115) and was reported by Stewart (1951) (FIG. 14.116). At the distal end of the megagametophyte are three oval archegonia, each ~1 mm in diameter. Histologic features of the archegonia were not preserved, but in one a large fimbriate mass of material has been suggested as representing the coalified remains of the egg cytoplasm. This seed is even more remarkable in that the pollen chamber contains a number of Monoletes pollen grains in proximity to the archegonia (FIG. 14.117). In one grain, two amorphous ovoid structures are preserved, which are morphologically similar to the flagellated sperm of certain extant cycads. Hexaloba finisensis is an ovule type from the Gzhelian (Upper Pennsylvanian) of Texas, which is similar to Pachytesta but differs in exterior shape, morphology of the integument, and vascular system (Dunn et al., 2002a). Ovules are up to 30 mm long and 20 mm in diameter; a hexagonal symmetry is created by six longitudinal ribs extending from the chalazal to the micropylar end. The unvascularized integument is three layered, with the sclerotesta delimited into three equal valves by commissured ribs. The stalked nucellus is attached to the integument only at the base. Nucellar bundles persist for 60–75% of the ovule length and consist of tracheids with scalariform wall thickenings. Another common Carboniferous seed that may have been associated with the medullosan seed ferns is Stephanospermum (FIG. 14.118). Specimens are 1 cm long and possess an apical integumentary crown that encircles

CHAPTER 14

PALEOZOIC SEED FERNS

579

Figure 14.117 Monoletes grain in the pollen chamber of the seed Pachytesta hexangulata showing two opaque bodies interpreted as male gametes. See FIG. 14.115 (Pennsylvanian). Bar  150 μm.

an elongated micropylar canal; the apical crown is absent in the Early Pennsylvanian species S. costatum (Good et al., 1982). Three primary ribs are present, but the sclerotesta is not ornamented externally. In S. elongatum (Middle–Upper Pennsylvanian), the apical crown is toothed, with each segment corresponding to a rib (FIG. 14.119) (Leisman and Roth, 1963; Serbet and Rothwell, 1995). Between the crown and the beak is a so-called premicropylar trough filled with parenchymatous cells of the sarcotesta (Hall, 1954). Vascular tissue is present in both the integument and the nucellus; nucellar bundles are tangentially flattened and fused to form a continuous ring, called the tracheal sheath. As in Pachytesta, the nucellus is free from the integument except at the base, and the pollen chamber is simple. Stephanospermum konopeonus is an ovule 1.5 cm long preserved in sideritic concretions from the Francis Creek Shale (Westphalian D) of the famous Mazon Creek area in northeastern Illinois (Drinnan et al., 1990a). At both the apex and the base, the integument is flared into a pair of bifid, winglike extensions (FIG. 14.120). What is most interesting about these ovules is that several are attached to a branching axis and thus were apparently not borne on fronds, as is generally depicted for the medullosan seed ferns. Of additional significance is the presence of Monoletes grains in the pollen chamber of these seeds, strengthening the relationship of Stephanospermum with the medullosan seed ferns. Several pollen grains of the Florinites-type have been reported in the pollen chamber of S. ovoides, but they may represent contaminants (T. Taylor, 1962).

Figure 14.118 Longitudinal section of the type material of

Stephanospermum akenioides. Note pollen grains in the pollen chamber (arrow) (Pennsylvanian). Bar  2 mm. (Courtesy J. Galtier.)

The presence of six rather than three longitudinal ribs in the sclerotesta was a feature used by Brongniart to delimit the genus Hexapterospermum (T. Taylor, 1966). The seed has a flattened base and a tapered apex (FIG. 14.121). The ribs are most conspicuous at the distal end, become gradually reduced toward the base, and extend into keel-like structures that project below the base of the seed. The seed coat consists of sarcotesta, sclerotesta, and endotesta, with the cells of each layer grading into those of the next zone. The six sclerotestal ribs (FIG. 14.122) are equidistantly spaced around the seed and may be identified in cross section by a thin plate of cells that divides the integument radially. One interesting aspect of H. delevoryii from Middle Pennsylvanian coal balls of Illinois (Matten and Hopkins, 1967) is the presence of a small parenchymatous cone

580

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

W

W

Figure 14.119 Longitudinal section of Stephanospermum

elongatum showing collar (arrow) at seed apex (Pennsylvanian). Bar  5 mm.

of tissue projecting through the nucellus toward the micropylar opening (FIG. 14.123) (T. Taylor, 1966). Topographically, this structure resembles the persistent central column often present in some lagenostomalean seeds. Such a structure may also represent a portion of the cellularized megagametophyte projecting through the distal end of the pollen chamber, or it may represent some form of prepollination remnant. Whatever the function of this structure, seeds such as Hexapterospermum typically possess simple pollen chambers that lack the morphologic complexity of the smaller, geologically older lagenostomalean seeds. Rhynchosperma quinnii is a Mississippian (Chester Series) seed with radial symmetry that is characterized by a threelayered integument (T. Taylor and Eggert, 1967; Dunn et al., 2002b). The seed apex is attenuated and sculptured by eight to ten sclerotestal ribs; below the midlevel the seed is smooth. Some specimens are at least 2.2 cm long. The connection between integument and nucellus is variable, with the degree of fusion varying between 25% and 85% of the length of the seed cavity. The pollen chamber possesses an apical beak and pollen chamber floor characterized by a thickened central column. Hollow lobes of the sclerotesta form a stellate micropylar

Figure 14.120 Longitudinal section of Stephanospermum konopeonus showing integumentary wings (W) and wide micropylar tube (arrow) (Pennsylvanian). Bar  2 mm. (From Drinnan et al., 1990.)

canal. In R. quinnii, a single primary bundle enters the base of the ovule and branches to vascularize integument and nucellus. The chalazal bundle extends into the base of the nucellus and distally forms a cup-shaped sheath of anastomosing tracheids, resembling the condition in Stephanospermum. This feature strongly suggests medullosan affinities for R. quinnii (Dunn et al., 2002b). Well-preserved megagametophyte tissue in one of the specimens described by Dunn et al. (2002b) is suggestive of alveolar cellularization of the gametophyte. Nucule and integumentary casts displaying similar morphologic features have been referred to the morphogenera Boroviczia and Rhynchogonium. One species, Rhynchogonium fayettevillense, is a fairly common seed cast of Chester age from southern Illinois and probably represents Rhynchosperma in a different state of preservation (Lacey and Eggert, 1964).

CHAPTER 14

PALEOZOIC SEED FERNS

581

Figure 14.123 Longitudinal section of Hexapterospermum delevoryii showing pollen chamber region with central parenchymatous column. (From T. Taylor, 1966.)

Figure 14.121 Longitudinal section of Hexapterospermum delevoryii showing flattened base (Pennsylvanian). Bar  5 mm.

Chilbinia lichii includes compressed, seed-bearing foliage displaying characters found in several groups of Paleozoic seed ferns. The pinnate fronds bear pinnules with reticulate venation similar to Lonchopteris-type foliage and have pedicellate cordate pinnule bases as seen in Sphenopteris. The fronds bear tiny (4 mm long by 1 mm wide), narrow, slightly ovate seeds apparently attached to the pinna rachis (Ash, 2006). Except for the small size, these seeds resemble one described by Halle (1927) as attached to a pinna of Alethopteris norinii from the Permian of China. What is most intriguing about this taxon is that it has been discovered in the Upper Triassic Chinle Formation of Arizona, USA. On the basis of the co-occurrence of C. lichii and other taxa considered archaic, such as Sphenopteris and Pelourdea, in the Chinle Formation, Ash (2006) suggested that the extinction of plants following the end-Permian crisis occurred in a stepwise manner and apparently continued for at least 20–25 myr after the end of the Permian in western parts of the Triassic paratropical belt. POLLEN ORGANS

Figure 14.122 Cross section of Hexapterospermum delevoryii

showing prominent ribs (Pennsylvanian). Bar  5 mm.

The pollen organs of the medullosan seed ferns are also rather large, with some up to several centimeters in diameter. Numerous forms have been described, with the compression forms assigned to the medullosans based principally on the type of pollen they produced. All are synangiate and range from simple and solitary forms to others that are organized

582

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.124 Suggested reconstruction of Paracalathiops

stachei (Pennsylvanian). (From Millay and Taylor, 1979.)

into compound fructifications and borne as pairs, possibly along the side of a common axis, such as Codonotheca and Paracalathiops (FIG. 14.124), or alternately along a fertile pinna, as in the Permian genera Schuetzia (FIG. 14.125) (Geinitz, 1863) and Dictyothalamus (Göppert, 1864, 1865). They are constructed of elongated, tubelike sporangia embedded in a parenchymatous ground tissue. Sporangial dehiscence occurs toward the center of the organ. Pollen grains are large (up to 600 μm long) and monolete and conform to the dispersed genus Monoletes. Possibly the simplest of the anatomically preserved medullosan pollen organs is the Late Pennsylvanian species Halletheca reticulatus (T. Taylor, 1971), a pyriform synangium 1.5 cm long that consists of five sporangia arranged around a central zone of sclerenchyma (FIG. 14.126). In the distal region of the organ, the central area is hollow (FIG. 14.127). A vascular bundle is associated with the outer wall of each sporangium and sporangial dehiscence took place toward the center of the organ (T. Taylor and Millay, 1981b). A second species, H. conica from the Middle Pennsylvanian of Oklahoma, differs from H. reticulatus in size (H. conica is only 0.5 cm long), number of sporangia (12 in H. conica), and histology of the central column and dehiscence tissue (Mapes, 1982). Although Halletheca has not been found attached, the gradually tapered end of the

Figure 14.125 Suggested reconstruction of Schuetzia anomala (Mississippian). (From Millay and Taylor, 1979.)

Figure 14.126 Cross section of Halletheca reticulata showing

five pollen sacs (Pennsylvanian). Bar  1.5 mm.

unit suggests that it may have been borne singly at the end of a small branch. Schopfitheca boulayoides is a stalked, clavate pyriform pollen organ 20 mm long by 10 mm wide, which is solid for most of its length. It is slightly concave distally with five to

CHAPTER 14

PALEOZOIC SEED FERNS

583

Figure 14.128 Suggested reconstruction of Schopfitheca boulayoides (Pennsylvanian). (From Delevoryas, 1964.)

Figure 14.127 Diagrammatic reconstruction of Halletheca

reticulata showing pollen within a sporangium (Pennsylvanian). (From Millay and Taylor, 1979.)

six lobes around the rim. The surface is covered with densely spaced hairs (FIG. 14.128). Pollen is of the Monoletes type. The taxon was defined on the basis of a single Mazon Creek nodule and has been placed within the medullosan seed ferns because of the Monoletes pollen grains that were macerated from the sporangia (Delevoryas, 1964b). Additional specimens, two of which had previously been illustrated by Langford (1958) as Sporangites and Rhabdocarpus sp., were documented and placed in synonymy with S. boulayoides by Drinnan and Crane (1994). Another simple synangium that is structurally homologous with Halletheca is Sullitheca. This pollen organ is slightly larger (2.5 cm long) than Halletheca and contains 40 elongated sporangia embedded in a parenchymatous ground tissue (Stidd et al., 1977). In the center of the organ, as seen in cross section, is an H-shaped zone of fibers similar to the

Figure 14.129 Cross section of Sullitheca dactylifera pollen

organ. Arrows indicate the position of dehiscence slits (Pennsylvanian). Bar  2 cm.

central sclerenchyma zone of Halletheca. Morphologically, the Sullitheca dactylifera pollen organ (FIG. 14.129) has been folded symmetrically, thereby providing room for an increased number of sporangia (FIG. 14.130), while providing

584

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.131 Cross section of Rhetinotheca tetrasolenata

(Pennsylvanian). (From Leisman and Peters, 1970.)

Figure 14.130 Diagrammatic reconstruction of Sullitheca dactylifera pollen organ. (From Stidd, 1977.)

a structural organization that allows internal sporangial dehiscence. Each sporangium is associated with a vascular bundle, and another ring of bundles is present near the outer margin of the organ. On the surface of Sullitheca are numerous multicellular hairs. Rhetinotheca tetrasolenata is a Middle Pennsylvanian form that consists of a central fibrous area surrounded by four sporangia (FIG. 14.131); the distal end of the central area is apparently hollow (Leisman and Peters, 1970) (FIG. 14.132). Each synangium is small (2 mm long) and is borne at the end of a greatly telescoped branching system, resulting in a complex mass of simple synangia. Slightly smaller, more spherical sporangia characterize R. patens (Rothwell and Mickle, 1982). The surface of each synangium is covered

Figure 14.132

Gilbert A. Leisman.

with numerous peglike processes that probably provided the mechanical support that held the compound unit together. Dolerotheca is the name that was used historically for large, complex, and compound synangiate pollen organs, regardless of preservational state (FIG. 14.133); the entire unit is called a campanulum. The organs consist of multiple synangial units made up of numerous elongated sporangia embedded in a ground tissue (FIG. 14.134). Some time ago, however, the genus Bernaultia was established for Dolerotheca specimens that consist of four synangial units united together (FIG. 14.135) (Rothwell and Eggert, 1986). The name Dolerotheca is retained for similar synangial units for which the external morphology and internal organization are not known in detail, often compression and impression

CHAPTER 14

PALEOZOIC SEED FERNS

585

Figure 14.135 Diagrammatic reconstruction of Bernaultia for-

mosa (Pennsylvanian). (From T. Taylor and Millay, 1979.)

Figure 14.133 Compressed specimen of Dolerotheca sp. (Pennsylvanian). Bar  2 cm.

Figure 14.136 Longitudinal section of Bernaultia formosa

showing organization of pollen sacs (Pennsylvanian). Bar  1 cm.

Figure 14.134 Cross section of Bernaultia (Dolerotheca) sp.

showing organization of pollen sacs (Pennsylvanian). Bar  1 cm.

specimens. Some specimens of Bernaultia are 4 cm in diameter and constructed of what appear to be radiating pairs of elongated sporangia (pollen sacs) (FIG. 14.136). In B. formosa, the common Pennsylvanian species, the pollen organ is constructed of four radial synangial units that have been symmetrically folded. Figure 14.137 is an idealized reconstruction of the entire unit showing the folded configuration of the component synangial units. Like the other medullosan pollen organs, dehiscence of each pair of sporangia is directed inwardly through the breakdown of specialized cells, and pollen is of the Monoletes type. The attachment of Bernaultia (Dolerotheca) to a Myeloxylon frond (FIG. 14.100) bearing Alethopteris pinnules has been

suggested on the basis of the arrangement and number of vascular bundles and common histologic features in the petiole and pedicel of the fructification (Ramanujam et al., 1974). These authors suggested that the campanulum was borne in the position of a penultimate pinna just above the main dichotomy of the frond. Stewartiotheca is a Late Pennsylvanian campanulate synangium 1 cm in diameter that superficially resembles a small Bernaultia. In S. warrenae (FIG. 14.138), there are about 80 elongated pollen sacs that are vascularized in pairs (Eggert and Rothwell, 1979). The central portion of the organ contains a cone-shaped hollow area that is surrounded by a narrow zone of sclerenchyma. A narrow pedicel is attached asymmetrically at the apex of the unit. Morphologically, the unit is highly plicated and similar to a single unit of the compound Bernaultia fructification. Pollen is of the Monoletes type (Millay et al., 1980).

586

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.137 Diagrammatic reconstruction of Bernaultia for-

mosa showing folded organization of pollen organ (Pennsylvanian). (From Millay and Taylor, 1979.)

Figure 14.139

Figure 14.138 Diagrammatic reconstruction of Stewartiotheca warrenae (Pennsylvanian). (From Eggert and Rothwell, 1979.)

James Morton Schopf.

The evolution of a complex pollen organ such as Bernaultia has been the topic of several detailed studies. One of the first studies was done by J. M. Schopf (FIG. 14.139) who proposed that the campanulum resulted from the fusion of a large aggregation of terminally borne sporangia (Schopf, 1948). Another theory views the Bernaultia pollen organ as a series of fused, smaller pollen organs such as Codonotheca or Halletheca (Stidd, 1978) (FIG. 14.140). This hypothesis was later refined (Stidd, 1990) to interpret the Bernaultiatype organ as consisting of a radially arranged series of bifurcating units bearing pinnately arranged, pendulous pollen sacs. Still another model hypothesizes that the Bernaultia is homologous with one to four highly plicated synangia that have been fused evolutionarily (Dennis and Eggert, 1978; Rothwell and Eggert, 1986). According to these authors, each unit of the compound structure was initially a ring of elongate sporangia that has become folded. Drinnan and Crane (1994), using Dolerotheca specimens from Mazon Creek, documented a radiating series of once- or twice-dichotomized partitions (interpreted as the remains of sclerenchyma), which

CHAPTER 14

PALEOZOIC SEED FERNS

587

Figure 14.141 Thore G. Halle.

Figure 14.140

Benton M. Stidd. Figure 14.142 Cross section of Potoniea illinoiensis (Pennsylvanian). Bar  2 mm.

delimit a series of small depressions. These depressions are interpreted as the positions of microsporangia in the campanulum. Since the partitions represent a continuous system connected through the center of the pollen organ, these authors suggested that the campanulum is of simple or single unit construction. Potoniea is a pollen organ initially described from European compression specimens (Halle, 1942) (FIG. 14.141) but which is now known from petrified remains, as well (Stidd, 1978). The campanulum is bell-shaped, up to 1 cm in diameter, and composed of numerous elongated sporangia embedded in a common ground tissue (FIG. 14.142). The sporangial tubes protrude from the distal surface for 1 mm (FIG. 14.143). The entire structure has been interpreted as five concentric rings constructed of clusters of four to six radial synangia. Each sporangial group is vascularized like those in Halletheca. There is only sparse information about the manner in which Potonieatype campanula were borne on the plant; several specimens, however, suggest that they were produced on skeletonized fronds or frond portions: A specimen of P. carpentieri from

Figure 14.143 Diagrammatic cutaway reconstruction of Potoniea illinoiensis pollen organ showing organization of synangia. (From Millay and Taylor, 1979.)

588

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.145 Distal surface of Parasporites pollen grain showing vestigial sacci (Pennsylvanian). Bar  125 μm.

14.144 Synangiate pollen organs of Potoniea bechii showing pinnate organization and attachment to main axis (Pennsylvanian). Bar  1 cm. (Courtesy S. Schultka.)

Figure

the Pennsylvanian of England displays a branched axis portion with four Potoniea campanula in organic connection (T. Taylor, 1982c). Axial systems bearing Potoniea-type campanula have also been documented on the basis of specimens identified as P. adiantiformis from the Upper Mississippian of northern China by S. Zhang et al. (1992) and Laveine et al. (1993b). A beautiful specimen of P. bechii comes from the lowermost Pennsylvanian (Namurian B) of western Germany (Schultka, 1995). This fossil consists of an axis of the penultimate order (9 cm long) that gives off helically arranged axes of the ultimate order, which, in turn, give off alternately positioned small branches. Each small branch terminates in a bell-shaped synangiate pollen organ 1 cm in diameter (FIG. 14.144). The pollen of Potoniea is radial and trilete and, because it is so unlike other medullosan forms, the assignment of this reproductive structure to the group remains tentative. The constant association of Potoniea with foliage of the Linopteris and Paripteris types, however, is believed to strengthen the relationships with the medullosans, perhaps with the stem Sutcliffia (Laveine et al., 1989, 1993b; Zodrow et al., 2007). Parasporotheca leismanii is another pollen organ that contains nonmonolete pollen (Dennis and Eggert, 1978). These grains are 275 μm long and characterized by two vestigial sacci (FIG. 14.145), one at either end of the grain (Millay et al.,

1978). The suture is monolete and slightly bent at the midlevel. Parasporotheca is also unusual in that the pollen organ lacks the radial symmetry characteristic of other medullosan pollen organs. In P. leismanii from the Upper Pennsylvanian, the synangium is constructed of several scoop-like synangia (FIG. 14.146), each of which consists of alternating sporangial lacunae in the ground tissue. The whole pollen organ measures 20 cm in length and 3 cm in width. Each of the scoop-shaped synangia is covered with peglike hairs that may have held the individual units together. It has been hypothesized that such a unit may have evolved from a laminar, fertile telome system that underwent lateral fusion, or it may have been initially radial, becoming bilateral when a small area of the enlarging, coneshaped synangium failed to develop (Millay and Taylor, 1979). Several compressed pollen organs are referred to the medullosan seed ferns on the basis of the presence of Monoletes pollen in them. Whittleseya (FIG. 14.147) is a form that has been interpreted as a ring of fused, elongated sporangia surrounding a central hollow (White, 1901; Halle, 1933; Jongmans, 1954). Each synangium is 5 cm long and 3 cm in maximum width. At the apex, the unit bears small teeth that correspond to the position of longitudinal grooves on the surface. In their review of Paleozoic pollen organs, Millay and Taylor (1979) offered an alternative interpretation of Whittleseya’s morphology. They viewed each synangial unit as flattened and aggregated together. Another compressed pollen organ that has been interpreted as a ring of fused sporangial tubes is Codonotheca (FIG. 14.148) (Sellards, 1903, 1907). Specimens are smaller than the synangia of Whittleseya and consist of a slightly swollen, sterile proximal portion with six elongated segments distally (FIG. 14.149). Each segment contains a

CHAPTER 14

PALEOZOIC SEED FERNS

589

Figure 14.146 Pinnate branching system of synangiate pollen organ Parasporotheca leismanii. (From T. Taylor and Millay, 1979.)

Figure 14.148 Compressed pollen organ Codonotheca caduca (Pennsylvanian). Bar  1 cm.

Figure 14.147 Whittleseya elegans (Pennsylvanian). Bar  1 cm.

single sporangial cavity flanked on either side by a strand of resistant tissue. A distinctive feature of Codonotheca is the longitudinal striations that depart from a central core in the sterile proximal portion, bifurcate near the sinus between the free segments, and then extend for the full length of the structure (Drinnan and Crane, 1994). Compressed specimens of Aulacotheca collicola (FIG. 14.150) from the Lower Pennsylvanian of Illinois, USA, suggest that these pollen organs may have been attached in clusters on the axis (Mickle and Leary, 1984). Aulacotheca

iowensis consists of a frond segment organized into a primary axis that bears six, alternately arranged laterals (Eggert and Kryder, 1969). Planated foliar structures are absent and the stalked synangia are borne in groups at the tips of small dichotomizing axes along the laterals. An individual synangium consists of three to four pollen sacs, each 5 mm long. Aulacotheca and Whittleseya pollen organs are known in association with Neuralethopteris- and Cardioneuropteristype foliage, but they also occur with Alethopteris (Leggewie and Schonefeld, 1954; Laveine, 1997; Goubet et al., 2000). Murielatheca delicata (FIG. 14.151) is a permineralized synangiate pollen organ of Middle Pennsylvanian age from eastern Kentucky coal balls that is very similar to Aulacotheca (Serbet et al., 2006). The synangium is 1 cm long and 2.5 mm in diameter and composed of 10–12 thin-walled, elongated sporangia containing Monoletes-type pollen grains. Vascular bundles occur in the periphery of the ground tissue.

590

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.150 Diagrammatic reconstruction of Aulacotheca hemingwayi showing central area surrounded by sporangia (Pennsylvanian). (From T. Taylor and Millay, 1979.) Figure 14.149 Suggested reconstruction of Codonotheca caduca pollen organ (Pennsylvanian). (From T. Taylor and Millay, 1979.)

on the basis of the co-occurrence of vegetative and reproductive organs, Serbet et al. (2006) suggested that M. delicata may have been borne on Medullosa anglica stems. POLLEN

All the pollen organs assigned to the Medullosales except Potoniea and Parasporotheca contain bilateral, monolete pollen assignable to Monoletes (FIG. 14.152). The grains of this genus range from 100 to 600 μm in length and exhibit a bent monolete suture on the proximal surface (FIG. 14.153) (T. Taylor, 1978). On the distal surface are two longitudinal grooves (FIG. 14.154) that extend almost the entire length of the grain. The sporoderm is thick (10 μm) and constructed of two layers: an inner homogeneous layer and an outer zone of sporopollenin units that are fused to form a meshlike or alveolate layer (FIG. 14.155). Some grains of Monoletes extracted

from sporangia contain small (0.3–0.8 μm), hollow, spherical units termed orbicules or Ubisch bodies (T. Taylor, 1976). These structures are associated with a series of membranes (tapetal membranes) and represent products of the tapetum. The microgametophyte in the medullosans is morphologically similar to that found in certain members of the Cycadales, except for the larger size of the sperm. Additional information about the putative microgametophyte in the medullosan seed ferns has been reported by Stidd from the pollen organ Saharatheca lobata (FIG. 14.156) (Stidd, 1991). Several of these Monoletes grains possess internal walls that subdivided the grain lumen into 10–14 compartments, each of which is believed to represent an individual cell. It is suggested that the gametes were produced in a central cell and exited through the proximal monolete suture. To date there is no information to suggest that haustorial or siphonogamous pollen tubes were present. Rather, it might be speculated that the large size of the pollen grains provided an adaptive advantage of adequate

CHAPTER 14

PALEOZOIC SEED FERNS

591

Figure 14.151 Longitudinal section of Murielatheca delicata

pollen organ (Pennsylvanian). (From Serbet et al., 2007.)

stored metabolites during the development of the microgametophyte. Medullosan pollen grains may have been some of the first transported through biotic methods, perhaps by some Carboniferous arthropod (Scott and Taylor, 1983), as they seem much too large for wind dispersal (Schwendemann et al., 2007). The alveolate wall organization of Monoletes also could have provided space within the wall for physiologically active materials involved in germination, attraction, or compatibility interactions (T. Taylor and Zavada, 1986). MEDULLOSAN EVOLUTION

There can be little doubt that the medullosans provide some evidence for the evolution of the partially dissected eustele from a protostele. Within the order, some taxa, such as Sutcliffia, contain a single protostele; others of a younger geologic age contain a larger number of vascular segments. Variation also exists within stems of comparable age, in part reflecting that many of the so-called species probably represent different levels of the same biologic species. Since stems with a small number of vascular segments are

Figure 14.152 Cluster of Monoletes pollen grains macerated from Dolerotheca pollen sac (Pennsylvanian). Bar  1.5 mm.

known from both the oldest and youngest geologic strata, the fossil record of the medullosans is difficult to interpret. One theory suggests that evolution among the medullosans was directed toward increased complexity of the stem, with the more primitive forms similar to the lyginopterid stem Heterangium. Another view interprets the forms with more vascular segments as primitive, suggesting that the trend

592

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.155 Alveolate organization of Monoletes pollen

wall (Pennsylvanian). Bar  10 μm.

Figure 14.153 Proximal surface of Monoletes pollen grain

showing median deflection of suture (Pennsylvanian). Bar  100 μm.

14.154 Distal surface (Pennsylvanian). Bar  100 μm.

pollen

Figure 14.156 Diagrammatic reconstruction of Saharatheca lobata pollen organ (Pennsylvanian). (From Stidd, 1991.)

was toward fewer vascular segments through phylogenetic fusion. Forms with a peripheral ring of secondary xylem evolved through the tangential fusion of peripheral vascular segments. Proponents of this latter theory point to older

Devonian plants included in the Cladoxylales as potential progenitors because of their similar stelar sympodial organization. Others view the multisegmented Permian forms as having stelar arrangements similar to those found among

Figure

of

Monoletes

CHAPTER 14

PALEOZOIC SEED FERNS

593

is unknown to date in medullosan seed ferns. Another conspicuous disparity between the two groups involves the marked differences in pollen size; the pollen of Medullosa may reach 0.5 mm in some forms, whereas the distally germinating cycad pollen grains average 30 μm. The occurrence in late Paleozoic rocks of relatively modern, true cycad reproductive structures that are contemporary with medullosan remains (Chapter 17) indicates that the two groups may have become separated by the Mississippian, and that the Permian medullosans played no part in the subsequent radiation of the cycads.

CALLISTOPHYTALES Figure 14.157 Suggested evolution among selected medul-

losan pollen organs. A. Cluster of terminal sporangia of a progymnosperm. B. Pollen sacs of the Codonotheca-type pollen organ. C. Incomplete fusion of pollen sacs to form a bilateral pollen organ like that of Parasporotheca. D. Fused ring of pollen sacs like that of Halletheca. E. Pollen organ showing plication like that of Sullitheca. F. Compound pollen organ like that of Bernaultia. (From T. Taylor, 1988b.)

the extant Cycadales, the group most workers believe is phylogenetically related to the medullosans (Worsdell, 1906). Both the cycads and medullosans produced similar ovules (Norstog and Nicholls, 1997), and a case can be made for the fact that at least some medullosans bore their seeds in a manner similar to the Cycadales. The pollen organs continue to present the biggest problem in relating the two groups. According to Millay and Taylor (1979), pollen organs of the medullosans can be organized into three basic types (FIG. 14.157). One is the simple, solitary form characterized by taxa such as Halletheca and Codonotheca. The aggregated type includes simple synangia that are topographically clustered but not fused together. These would include Rhetinotheca, Parasporotheca, and Whittleseya. A third group includes simple or complex types, with the complex forms, such as Bernaultia and Dolerotheca, interpreted by most scholars as composed of multiple, simple units (but see Drinnan and Crane, 1994) that are highly plicated and embedded in a common ground tissue (Serbet et al., 2006). The pollen sacs in the Cycadales are usually regarded as being attached to the abaxial surface of simple sporophylls. A morphogenetic analysis of the male reproductive structures of the cycad Zamia amblyphyllidia (Mundry and Stützel, 2003), however, indicates that the microsporophylls are pinnately organized, with the pollen sacs borne on reduced leaflets and thus consistent with the basically pinnate pollen organs of the late Paleozoic pteridosperms. Nevertheless, cycad male reproductive organs are aggregated into strobili, a feature that

Although the Callistophytales are the most recently discovered seed ferns (Stidd and Hall, 1970a, b), they are probably the best known of the Paleozoic pteridosperms (Rothwell, 1975, 1981). The order includes plants with eustelic stems that produce pinnately compound fronds with axillary buds or branches at each node. Pollen organs are synangiate and borne superficially on pinnules of the frond; pollen is small and saccate. The seeds are slightly flattened (platyspermic), attached to the abaxial side of reduced pinnules, and noncupulate; they contain a nucellus and integument that are free except at the base. The Callistophytales are known principally from the Middle and Upper Pennsylvanian of the United States (Rothwell, 1975, 1976, 1981) and the Upper Pennsylvanian of central Europe (Rössler, 2000), and to a lesser extent from the Lower Permian of China (Hilton et al., 2002), and possibly the Lower Permian of Brazil (Rössler and Noll, 2002). Although the generic name Callistophyton was initially intended for anatomically preserved axes, the concept is today also used for the entire plant (FIG. 14.158) (Rothwell, 1975). Moreover, correspondences in pinnule size and shape, along with similarities in ovule/seed morphology, strongly suggest that the widespread foliage taxon Dicksoniites (Chapter 16) represents the compression equivalent of Callistophyton foliage (Stidd and Barthel, 1979; Meyen and Lemoigne, 1986; Galtier and Béthoux, 2002). Callistophyton was a relatively small, scrambling or climbing plant that probably represented one of the understorey elements of the Carboniferous forest (DiMichele et al., 2006; Willard et al., 2007). It consists of branching stems with large, widely separated bipartite or monopodial fronds, some of which are equipped with axial tendrils (Krings et al., 2003a). Pinnules were laminar and entire margined, lobed or dissected (Galtier and Béthoux, 2002). Numerous adventitious roots were borne at many of the nodes. DiMichele et al. (2006) interpreted the overall aspect

594

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.158 Suggested reconstruction of Callistophyton (Pennsylvanian). (From Rothwell, 1975.)

of this plant as being one of an opportunist, based on features such as (1) probable wind pollination (small saccate pollen), followed by the production of many, highly dispersible, small seeds (2) patchy, often dense occurrence within wetland landscapes, and (3) frequent preservation as charcoal, as if growing in areas prone to fires. These plants were likely able to rapidly colonize available physical space such as treefall gaps and thus garner light and mineral resources. VEGETATIVE ORGANS

The largest stem sections known for Callistophyton measure 3 cm in diameter. The stem has a parenchymatous pith that is slightly angular and surrounded by up to 13 axial bundles of primary xylem (FIG. 14.159). Bundle number is variable depending on the level of the stem and the leaf traces that are being produced. The primary bundles of the Late Pennsylvanian species C. poroxyloides are mesarch (Delevoryas and Morgan, 1954c; Delevoryas, 1956), those of the Middle Pennsylvanian taxon C. boyssetii are exarch. Pitting on the metaxylem tracheids varies from reticulate to bordered. Surrounding the primary body is an extensive zone of secondary xylem up to 70 cells in radial thickness. The secondary xylem tracheids are arranged in files one to five cells wide that are separated by biseriate rays. Tracheid pitting is confined to the radial walls and shows crossed, slitlike apertures. A zone of fusiform and ray initials two to four cells in thickness is situated along the outer edge of the wood.

Figure 14.159 Cross section of Callistophyton poroxyloides stem (Pennsylvanian). Bar  5 mm.

The vascular cambium initially produced only secondary xylem, later producing equal amounts of wood and secondary phloem. The position and extent of the primary phloem are not known with certainty, but secondary phloem consists of sieve cells, phloem parenchyma, and associated phloem rays (Russin, 1981; Smoot, 1984c). Amber inclusions within the sieve areas have been suggested as representing callose plugs. The cortex is composed of an inner parenchymatous zone and an outer area of longitudinally oriented fibers interspersed with parenchyma. The degree of fusion among the fibrous bundles of the cortex varies from level to level in the plant and also among stem specimens. Scattered throughout the cortex are enlarged cavities, some with an epithelial lining, that contain an amber-colored substance (Delevoryas and Morgan, 1954c). Histologically identical secretory cavities are known to have occurred in the primary ground tissues of all plant organs of the genus, and they have also been recorded for compressed foliage of the Dicksoniites-type (Krings, 2000b). Periderm is apparent in the inner cortex, and in older stems it becomes the outer limiting tissue of the stem. The vascular system of Callistophyton has been interpreted as composed of axial sympodia that extended through the stem and produced leaf traces which were initially double at the petiole base and then single at higher levels. Small axillary buds are borne at each node and are enclosed by two oppositely placed leaves or cataphylls that extend beyond the apex of the bud. Axillary branches have the same histologic features as the stems already described. The roots of Callistophyton are diarch, and older specimens contain abundant secondary

CHAPTER 14

PALEOZOIC SEED FERNS

595

Figure 14.160 Cross section of Johnhallia lacunosa showing stem and petiole (arrow) (Pennsylvanian). Bar  1 mm. (Courtesy B. Stidd.)

tissues, including periderm. They are adventitious and attached to the stems in the axils of buds or branches. Johnhallia lacunosa is a permineralized stem described from Middle Pennsylvanian coal balls of Indiana that shares features with Callistophyton (Stidd and Phillips, 1982). These eustelic stems are 1 cm in diameter and possess a primary body composed of five vascular strands (FIG. 14.160). Associated with each bundle is a core of parenchyma that superficially resembles the protoxylem lacunae of some sphenophytes. Secondary xylem tracheids possess scalariform and multiseriate bordered pits on both the tangential and radial walls. Rays are uniseriate. Fronds are flattened and attached to the stems in a 2/5 phyllotaxy. In cross section, petiole traces appear slightly curved toward the adaxial surface (FIG. 14.160), and at higher levels give rise to subopposite laterals. The foliage lacks internal differentiation of palisade and spongy mesophyll and appears to be similar to the nearly nonlaminate compression taxa Rhodea (Rhodeopteridium), Diplotmema, and Palmatopteris (DiMichele et al., 2006). The leaves of Callistophyton were helically arranged, with the larger leaves estimated at 30 cm long based on permineralized specimens (Rothwell, 1975). Fronds range from bito quadripinnately compound and bear laminar pinnules. The pinnules are entire-margined to deeply lobed with a somewhat constricted base. Multicellular hairs are present on the abaxial surfaces of some pinnules. Morphologically, the pinnules of Callistophyton correspond most closely to the compression morphogenus Dicksoniites (Chapter 16). Another putative callistophytalean foliage taxon is Pseudomariopteris (see below and Chapter 16).

Figure 14.161 Longitudinal section of Callospermarion pusillum ovule showing two archegonia (arrows) (Pennsylvanian). Bar  1 mm.

REPRODUCTIVE STRUCTURES

Callospermarion is a platyspermic seed that was produced by Callistophyton (Eggert and Delevoryas, 1960). Where the seeds were produced on the plant is not known for sure, but the occurrence of pollen organs attached to the lower (abaxial) surfaces of laminar pinnules suggests a similar mode of attachment for the ovules. In the putative compression equivalent Dicksoniites, platyspermic ovules/seeds similar to Callospermarion occur on the abaxial side of slightly reduced pinnules (Langiaux, 1986; Meyen and Lemoigne, 1986). Abaxially positioned platyspermic seeds resembling Callospermarion have also been documented for the Late Pennsylvanian–Early Permian foliage morphogenus Pseudomariopteris (Krings and Kerp, 2000; Castro Martínez, 2005), suggesting affinities of this taxon with the Callistophytales. Permineralized specimens of Callospermarion range from 0.8 to 5 mm long, up to 3.8 mm wide in the primary plane (FIG. 14.161) and 2 mm in the secondary plane. When sectioned transversely, lateral extensions of the integument show blunt wings in the primary plane, each containing a vascular bundle (Neely, 1951). The integument is three parted, with the major part of the seed coat composed of cells of the

596

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.162 Longitudinal section of Callospermarion pusil-

lum ovule with pollination droplet (arrow) (Pennsylvanian). Bar  1 mm.

outer sarcotesta, including large secretory cavities. Utilizing 125 specimens of Callospermarion pusillum from Upper Pennsylvanian coal balls from Berryville, Illinois, Rothwell (1971b) was able to demonstrate several ontogenetic stages of the ovule. Vascular tissue is present as two prominent strands in the wings of the integument and as a small pad of tracheids at the base of the nucellus. The nucellus in C. pusillum is fused with the integument only at the chalazal end; the distal end is modified into a flask-shaped pollen chamber. In many ovules that were judged to be immature on the basis of the differentiation of the integument, the pollen chamber contains numerous pollen grains of the Vesicaspora type. The ovules apparently were aided in trapping pollen by the production of a resinous pollination droplet that extended from the micropylar orifice (Rothwell, 1977a), in a manner similar to those seen in many extant gymnosperms. In modern gymnosperms, these secretions function as trapping devices, and as a germination medium for pollen, but may also play a significant part in reducing pollen germination by foreign species (Gelbart and von Aderkas, 2002; Labandeira et al., 2007b). In one specimen of C. pusillum, the fossilized droplet actually contained several pollen grains (FIG. 14.162). Another species, C. undulatum, is slightly larger

Figure 14.163 Diagrammatic reconstruction of synangium of

Idanothekion glandulosum. (From Millay and Taylor, 1979.)

than C. pusillum and is known from Middle Pennsylvanian rocks. Two pollen organs have been described for Callistophyton, and these differ only in minor features, although they are of different ages. Both are known to have been borne on the abaxial surface of small laminar pinnules of a tripinnate frond. Idanothekion glandulosum is known from the Middle Pennsylvanian of Illinois and consists of a ring of six to nine exannulate sporangia that are proximally fused around a vascularized central column (FIG. 14.163) (Millay and Eggert, 1970). Each synangium is ~1 mm long and radial in cross section. The wall of the sporangium is thick, except on the side facing the interior of the synangium, where dehiscence took place. Vascular tissue is present as a broad band in the outer layer of each sporangium. Krassilov et al. (1999) noted that sporangial and pollen morphology in Idanothekion is similar to that seen in Permotheca, a putative peltaspermalean pollen organ from the Lower Permian of Tschekarda, the Urals, and that this may indicate phylogenetic relationships between these two groups of seed ferns. The structural features of the second type of callistophytalean pollen organ, Callandrium callistophytoides (FIGS. 14.164, 14.165), are almost identical with those of Idanothekion, differing only in the absence of vascular tissues in the sporangial wall of Callandrium (Stidd and Hall, 1970b). Callandrium is

CHAPTER 14

PALEOZOIC SEED FERNS

597

Figure 14.166 Pollen grains in Callandrium callistophytoides pollen sac (Pennsylvanian). Bar  100 μm.

Figure 14.164 Longitudinal section of several Callandrium callistophytoides pollen sacs (Pennsylvanian). Bar  0.5 mm.

Figure 14.167 Distal view of Vesicaspora pollen grain show-

ing central body and endoreticulations in sacci (Pennsylvanian). Bar  12 μm. Figure 14.165 Callandrium callistophytoides pollen sacs attached to leaf. (From Millay and Taylor, 1979.)

known from the Upper Pennsylvanian Calhoun Formation near Berryville, Illinois, a site which has also yielded Callistophyton poroxyloides stems. Both pollen organs bear small (40 μm) monosaccate grains (FIG. 14.166) of the type generally associated with coniferophytes. If found in the dispersed state, these grains would be included in the genus Vesicaspora (FIG. 14.167). Immature sporangia have provided evidence that both tetrahedral and isobilateral tetrads were formed in the group (Hall and Stidd, 1971). The grains are bilateral and elliptical in equatorial outline. The central body (corpus) is surrounded by an equatorial saccus that is lined with a delicate ornamentation (FIG. 14.168). On the

Equatorial view of Vesicaspora pollen grain showing large central cell in corpus (Pennsylvanian). Bar  15 μm.

Figure 14.168

598

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.169 Vesicaspora pollen grain showing three-celled stage (arrows) in the microgametophyte (Pennsylvanian). Bar  15 μm.

proximal surface is a thickened cap (cappa); between the two lobes of the saccus is an elliptical sulcus. A series of stages in the development of the microgametophyte is known for Vesicaspora (Millay and Eggert, 1974). Grains have been found in Idanothekion and Callandrium sporangia that contain two-, three-, and four-celled stages of the microgametophyte (FIG. 14.169), with the cells arranged in an axial stack within the corpus. Those closest to the proximal face of the grain represent prothallial cells, with the largest cell representing either the embryonal cell, if further prothallial cells were produced, or an antheridial cell if prothallial cell production had ended. In one grain observed in polar view, two ovoid bodies may have represented the generative cell and protoplast of the tube cell. Pollen was not shed in this group until the microgametophyte had developed at least to the four-celled stage, and judging from the immature nature of the pollen exine, the microgametophyte may have remained in the sporangium for a considerable time. The morphologic nature of Vesicaspora pollen was well suited to dispersal by wind and, as already noted, the grains were trapped by the resinous pollination droplet extruded from the micropyle of the Callospermarion ovule. It is not known whether the trapped pollen grains floated into the pollen chamber of the ovule or were pulled in by the shrinking of the droplet. Vesicaspora pollen germinated from the distal sulcus and is the first pollen grain discussed thus far to germinate from the distal surface of the grain. A grain with a branched pollen tube (FIG. 14.170) has been found in the nucellus of a C. undulatum ovule, further strengthening the biological relationship of the seeds and pollen grains of this group. In extant plants, branched pollen tubes are typically haustorial, for example, in the Cycadales, and within the

Figure 14.170 Pollen tube (arrow) extending from the distal surface of Vesicaspora grain (Pennsylvanian). Bar  30 μm.

Callistophytales, there may be evidence that the haustorial function of the pollen tube preceded siphonogamy. CALLISTOPHYTALEAN EVOLUTION

We currently know more about the biology of the callistophytalean seed ferns than any other group of pteridosperms (Rothwell, 1980). Available information details not only the structure and development of the vegetative parts of the plant but also the ontogeny of the seeds and microgametophytes. All that is really missing from our understanding of this taxon is the exact manner in which the seeds were attached to the parent plant and stages in the development of the embryo. What is most surprising about these seed plants is their relatively modern, saccate pollen type and the equally modern appearance of the microgametophyte. The number of primary prothallial cells (three) present in the Callistophytales is greater than that in extant cycads and closely approximates the upper number found in living coniferophytes. These fossil microgametophytes do not support the long-held belief that the microgametophyte of seed plants represents the reduced sexual phase of the life history that has evolved through the loss of almost all of the vegetative and sterile cells of an antheridium (Florin, 1936a).

GLOSSOPTERIDALES The plants included in the Glossopteridales are all extinct representatives of a flora that once dominated the

CHAPTER 14

PALEOZOIC SEED FERNS

599

supercontinent Gondwana (including parts of New Zealand, Australia (Rigby, 1972a), Africa, South America, Antarctica, and peninsular India) during the Permian. Historically, the glossopterids have been included with the cycads, seed ferns, gnetophytes, cordaites, and angiosperms, and more recently have figured prominently in phylogenetic analyses in which they have been treated as a single terminal (Nixon et al., 1994; Doyle, 2006). The most common leaf morphogenus, Glossopteris, is historically important as its presence on several of the southern continents represented some of the early evidence for the existence of continental drift (Wegener, 1924). The discovery of structurally preserved, seed-bearing megasporophylls substantiates beyond any doubt that at least some of the glossopterids should be classified as seed ferns (Gould and Delevoryas, 1977; E. Taylor and Taylor, 1992; Nishida et al., 2007). As work continues on this fascinating group, it is becoming apparent that we are only now beginning to fully appreciate the diversity that existed among these plants. LEAVES

GLOSSOPTERIS The most commonly found remains of glossopterids are leaves of the morphogenus Glossopteris (FIG. 14.171), a lanceolate to tongue-shaped, entire leaf, which is characterized by a midrib made up of multiple veins and reticulate secondorder venation (FIG. 14.172). Since the original description of Glossopteris by Brongniart (1828), numerous species from throughout Gondwana have been erected. Chandra and Surange (1979) (FIG. 14.173) included more than 70 taxa in their revision of the Indian species of Glossopteris, and it is quite possible that several hundred species have been described worldwide. Despite a number of attempts to develop a standardized classification for glossopterid leaves, little progress has been made in defining species on the basis of morphology, secondary venation, cuticular anatomy, or even reproductive organs, in part no doubt because many species are based on poorly preserved impression material. Glossopteris is also a component of several mixed floras that existed between the typical Cathaysian and Gondwana floral provinces from the Middle to Late Permian. These floras no doubt reflect changing climatic regimes during this period with elements found as far north as Turkey (Archangelsky and Wagner, 1983) and Oman (Berthelin et al., 2006). The venation of Glossopteris leaves consists of a midrib made up of several parallel vascular strands that extend to near the leaf tip (FIG. 14.174). The outer bundles of the midrib give off laterals that repeatedly dichotomize and anastomose to form a uniform reticulate pattern of veins in which

Figure 14.171 Glossopteris stricta (Permian). Bar  1 cm.

there is no hierarchical pattern (Trivett and Pigg, 1996). In paradermal view this pattern of veins results in oblong– polygonal meshes that assume an oblique course to the entire margin (FIG. 14.172). In G. wilsonii from the Lower Permian of Argentina, however, veins are decurrent and lack anastomosing smaller veins (Archangelsky et al., 1981b).

600

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.172 Paradermal section of Glossopteris schopfii leaf showing venation pattern (Permian). Bar  1.5 mm.

Figure 14.174 Glossopteris leaf showing reticulate venation

(Permian). Bar  2.5 cm. (Courtesy H. Kerp.)

Figure 14.173

Krishna Rajaram Surange.

Some species, such as G. browniana, exceed 30 cm in length and are characterized by a rounded apex (FIG. 14.175). Veins in Glossopteris fibrosa, a slightly smaller species collected from the upper coal measures of Tanganyika, have an obtuse apex (Pant, 1958). The upper epidermis is typically devoid of stomata; stomata on the lower surface are

haplocheilic with four to eight subsidiary cells arranged in an irregular ring. Cuticular and epidermal features have been used to distinguish some species of Glossopteris (Pant and Singh, 1974). In some species trichomes are present on the lower surface of the leaf. Since most of the species have been described from detached leaves, it is not known for certain whether the leaves were attached to modified branches such as short shoots or borne directly on the stems (FIGS. 14.176, 14.177). In Glossopteris maculata, the leaves were borne in what are interpreted as whorls, with the individual leaflets up to 14 cm long (Pant and Singh, 1974). Leaf whorls of G. recurva contained up to 11 leaves per whorl. It appears that most glossopterids

CHAPTER 14

PALEOZOIC SEED FERNS

601

Figure 14.176 Glossopteris sp. leaves attached to axis (Permian). Bar  2 cm. (Courtesy S. McLoughlin.)

Figure 14.175 Glossopteris browniana (Permian). Bar 

2 cm.

were arborescent plants (FIG. 14.178) and some had alternately arranged leaves borne in tight helices that simulate whorls (Pant, 1977). The presence of numerous leaf mats (FIG. 14.179) in autumn–winter varved sediments suggests that many, if not all glossopterids were seasonally deciduous. An important breakthrough in our understanding of the glossopterids was the discovery of anatomically preserved specimens in silicified peat deposits (FIG. 14.180) in the Bowen Basin of Queensland, Australia, and the

Figure 14.177 Immature Glossopteris sp. leaves attached to axis (Permian). Bar  1 cm. (Courtesy S. McLoughlin.)

602

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.178 Suggested reconstruction of a Glossopteris plant and associated organs. (From Gould and Delevoryas, 1977.)

Figure 14.180 Glossopteris schopfii leaf on surface of permineralized peat block. Note midrib and venation (Permian). Bar  1 cm. Figure 14.179 Stacked permineralized Glossopteris leaves from Antarctica (Permian). Bar  1 mm.

Transantarctic Mountains of Antarctica (Schopf, 1970; Gould and Delevoryas, 1977; E. Taylor et al., 1989, E. Taylor and Taylor 1990). Two species of leaves have been described from Antarctica, and these also appear to be present in Australia, in addition to a third species (Pigg, 1990a; Pigg and Taylor, 1990). One of the Antarctic species, G. schopfii, is characterized by leaves thought to be 12 cm long that are helically attached to small branches in a 2/5 phyllotaxy (Pigg and Taylor, 1993). The midrib consists of four to five vascular strands with narrow meshes (FIG. 14.181) (Pigg, 1990a).

Vascular bundles in the leaves are exarch to mesarch in primary xylem maturation and surrounded by a prominent bundle sheath of thick-walled fibers (FIG. 14.182). The cuticle is thin and stomata appear to be borne in rows. Glossopteris schopfii resembles one of the narrow mesh, impression–compression forms such as G. angustifolia (Pigg, 1990a). The second leaf type from Antarctica (and Australia) is G. skaarensis. It is characterized by larger meshes and a more prominent midrib; like G. schopfii the mesophyll is not well defined (Pigg, 1990a). The bundle sheath in this species is composed of thin-walled cells that sometimes include dark contents. Stomata are sunken. Leaves of the G. skaarensis type

CHAPTER 14

PALEOZOIC SEED FERNS

603

Glossopteris leaves from Lower Triassic rocks (Pant and Pant, 1987), although other authors have disputed the identification as Glossopteris. One of these is G. sidhiensis from the Sidhi District, Madhya Pradesh, India. This species is petiolate and contains veins with broad meshes. Other reports of glossopterid leaves in Triassic rocks are noted in McManus et al. (2002a), which also includes descriptions of permineralized leaves, roots, and wood from the Late Permian–Early Triassic of Antarctica. Another barrier to identifying Glossopteris in the Early Triassic is the difficulty in precisely defining the Permian–Triassic boundary, especially in continental rocks.

Figure 14.181 Paradermal section showing Glossopteris schopfii vein meshes (Permian). Bar  1 mm.

Figure 14.182 Cross section of Glossopteris schopfii leaf.

Arrows indicate vascular bundles (Permian). Bar  0.5 mm.

were attached to small twigs 1 cm in diameter (Pigg and Taylor, 1993). Wood is of the Agathoxylon (=Araucarioxylon) type and the eustelic stems have uni- to biseriate rays and oval to hexagonal bordered pits. The broader-meshed G. skaarensis leaves are most similar to the impression–compression species G. conspicua. The third species of anatomically preserved leaves known to date is from Australia (Pigg and McLoughlin, 1997). Leaves of G. homevalensis have narrow meshes and a well-defined palisade and spongy mesophyll. Although Glossopteris has been used as a biostratigraphic marker for the Permian, there are a few reports of

GANGAMOPTERIS Gangamopteris (FIG. 14.183) is a common element of Permian Gondwana floras that in general morphology is nearly identical to Glossopteris; in fact, some researchers include both genera in Glossopteris (Kovács-Endrödy, 1979). Gangamopteris is generally found in Lower Permian rocks (Høeg and Bose, 1960), whereas Glossopteris is believed to have its greatest distribution in the Upper Permian (Rigby, 1984a). The genera have also been distinguished by the larger size and the absence of a well-defined midrib in Gangamopteris; in general the vein meshes are also more uniform (Tybusch and Iannuzzi, 2008). Some species of Glossopteris exhibit a relatively inconspicuous midrib, but according to Pant and Singh (1968), the two taxa may be distinguished by cuticular features of the midrib area. Leaves with stomatiferous areas bounded by nonstomatiferous areas similar to those on the rest of the leaf warrant assignment to Gangamopteris. Other cuticular and epidermal features seem to show the same range of variability exhibited by leaves of Glossopteris. OTHER LEAF TYPES Other common Gondwana leaf types include Belemnopteris (Pant and Choudhury, 1977), Rhabdotaenia (Pant, 1958), and Surangephyllum (FIG. 14.184) (Chandra and Singh, 1986). In Belemnopteris the secondary veins are similar to those in Glossopteris and Gangamopteris, but the leaf is typically sagittate at the base (Pant and Choudhury, 1977; KovácsEndrödy, 1990). Some species, such as B. pellucida, may be nearly 20 cm long. A primary feature used to distinguish Rhabdotaenia is the presence of nonanastomosing lateral veins (Pant, 1982). Haplocheilic stomata are confined to the lower surface and consist of unspecialized subsidiary cells. Other, less common Gondwana leaf morphogenera historically included within the Glossopteridales are Rubidgea, Palaeovittaria, and Euryphyllum. The last taxon, however,

604

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.184 Surangephyllum sp. leaf (Permian). Bar  1 cm. (Courtesy S. McLoughlin.)

Figure 14.183 Gangamopteris sp. leaf (Permian). Bar  1 cm. (Courtesy S. McLoughlin.)

is today mostly regarded as belonging to the Cordaitales (Chapter 20) (Singh, 2000; Singh et al., 2007). Before concluding the discussion of foliage attributed to the glossopterids, Glossopteris-like foliage from younger rocks should be mentioned. Leaves of Mexiglossa are known from the Middle Jurassic of Oaxaca, Mexico (Delevoryas and Person, 1975) (FIG. 14.185). Associated with the leaves are typical Jurassic foliage taxa, including Zamites, Otozamites, Pterophyllum, Ptilophyllum, several ferns, and some cones assignable to the genus Williamsonia (see Chapter 17). These floral elements suggest a Jurassic age for this assemblage. The leaves demonstrate the same range of morphologic

characters that exist among species of Glossopteris, including forms that may be up to 26 cm long. Except for the geologic age and the other floral elements, the Oaxacan leaves are almost identical to Glossopteris. Specimens of Mexiglossa have also been reported from the Late Triassic or Early Jurassic of Honduras (Ash, 1981). An important future area of research will be to determine whether these Jurassic leaves represent a remnant of the earlier Gondwanan floral province that migrated northward during the Triassic or Early Jurassic, perhaps as the climate changed, or were simply a basic leaf type that was produced by different groups of seed plants during the late Paleozoic and Mesozoic. Scale leaves of many sizes and shapes are also commonly found associated with Glossopteris and Gangamopteris in Gondwana facies. Those with rhomboidal shapes are possibly the most common type (Lacey et al., 1975) and have been interpreted as bud scales, sterile scales from reproductive organs, and organs that are intermediate between typical photosynthetic leaves and the foliar parts of reproductive organs.

CHAPTER 14

PALEOZOIC SEED FERNS

605

Figure 14.186 Terminal roots of Vertebraria (Permian). Bar  1 cm. (Courtesy S. McLoughlin.)

Figure 14.185 Mexiglossa varia (Jurassic). Bar  2 cm.

STEMS AND ROOTS

The stems that produced Glossopteris have pycnoxylic wood with conspicuous growth rings and are included in the morphogenus Agathoxylon (Araucarioxylon) (Philippe, 1993). It is characterized by multiseriate pits on the radial walls of the tracheids; some pits are hexagonal in outline. Cross-field pits are of the cupressoid type, and xylem rays are uniseriate. The distinctive roots of Glossopteris plants are called Vertebraria (FIG. 14.186) and are recognizable in both compression–impressions and petrified fossils. Initially illustrated by Royle (1833), the genus has been the subject of several studies (Singh and Chandra, 1995) and is now known from anatomically preserved and compressed specimens from a number of localities (Gould, 1975; Neish et al., 1993; Srivastava, 1995).

Figure 14.187 Cross section of Vertebraria root showing well-

developed xylem wedges with lacunae between (Permian). Bar  1 mm.

In cross section, the axes include a central zone of exarch primary xylem surrounded by four to seven radiating arms of wood separated by hollow areas (FIG. 14.187). The secondary xylem can be continuous near the periphery of the axis (but see Neish et al., 1993) and typically contains distinct growth rings (FIG. 14.188). Surrounding the zone of secondary xylem is a narrow band of periderm. In longitudinal section, the secondary xylem wedges are connected at varying intervals by transverse segments of wood, termed platforms, that often contain a trace to a lateral root. A few crushed parenchyma cells have

606

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.188 Cross section of large Vertebraria axis with

secondary xylem (Permian). Bar  1.5 cm.

been identified in the hollow areas between the xylem arms and platforms, and it remains to be seen whether these areas were devoid of cells during the life of the plant. Protoxylem tracheids have annular secondary wall thickenings, whereas the secondary xylem tracheids exhibit up to six vertical rows of opposite or alternate bordered pits. Extending through the pycnoxylic wood of Vertebraria are uniseriate vascular rays with cross-field pits. OVULATE REPRODUCTIVE STRUCTURES

The reproductive organs of the glossopterids have been a continuing source of controversy since the first report of “linear sori” on some Glossopteris leaves (Feistmantel, 1886). Since that time, more than 50 different types of ovulate reproductive structures have been described from Glossopteris-rich rocks around the world (Holmes, 1974, McLoughlin, 1990a), and additional forms are being described today as modern detailed studies are undertaken (Adendorff, 2005). A few reproductive organs have been found attached to the parent plant in association with vegetative leaves (Pant and Singh, 1974) (FIG. 14.189). In other instances, the reproductive organs are attached to highly modified leaves that have an anastomosing, mesh-like pattern of veins and occur at the same localities as vegetative organs. Because the vast majority of these organs are known from impression–compression and mold–cast remains (White, 1978) their morphology and subsequent interpretation have remained a source of speculation and controversy (Pant, 1982). The morphogenus Hirsutum (now Bijariala) has been the focus of a particularly controversial debate, as it formed the basis for a now-rejected theory that the glossopterid fructifications were bisexual (Plumstead,

Figure 14.189 Glossopteris taenioides showing attachment of

leaf and reproductive structures (Permian). (From Pant and Singh, 1974.)

1956, 1958). Plumstead (1956) was the first to describe the ovulate fructifications of the glossopterids and thus prove that they were gymnosperms and not ferns. She erected the genus Hirsutum for shield-shaped, “bivalvate” fructifications, which bore what she considered to be hairlike, pollen-producing organs on one of the two “valves.” She also interpreted flat, bractlike structures in Scutum as pollen bearing (for details, see Prevec et al., 2008). Thus far no glossopterid reproductive organs have been conclusively demonstrated as containing both pollen and seeds (Pant, 1987). Despite the large number of glossopterid ovulate structures that have been described, many, and perhaps all, represent variations on a basic theme, although some are interpreted differently (Rigby, 1984a). Each appears to consist of a megasporophyll bearing seeds; this megasporophyll has been variously termed a capitulum, cupule, fertiliger, or cladode. In instances where the reproductive structure is attached, it appears in an axillary position or fused or adpressed to the petiole or lamina of a Glossopteris leaf. PERMINERALIZED FORMS There are numerous examples in paleobotany of a single fossil discovery that greatly alters the interpretation of previously

CHAPTER 14

PALEOZOIC SEED FERNS

607

Figure 14.190 Megasporophyll of Homevaleia gouldii showing several seeds (Permian). Bar  1 mm. (Courtesy P. Ryberg.)

described specimens and, in a few instances, influences the interpretation of the affinities of a particular group of plants. One of these is the report of anatomically preserved Glossopteris remains from the Late Permian of Queensland (Gould and Delevoryas, 1977). This report included information on silicified layers of plant material that contained not only Glossopteris vegetative leaves, but also megasporophylls with attached ovules. Although the megasporophylls were not found organically attached to axes or leaves, anatomical similarities between the megasporophylls and vegetative leaves of Glossopteris suggest that they were borne on the same plant (see also Nishida et al., 2007). Attached to one surface of the fleshy megasporophyll are numerous ovoid seeds; the edges of the megasporophyll are enrolled and partially cover the seeds (FIG. 14.190). The megasporophyll, now named Homevaleia gouldii (Nishida et al., 2007), lacks the sclerenchymatous hypodermal fibers found in the vegetative leaf of Glossopteris homevalensis. Each orthotropous seed is 1.2–1.3 mm long by 0.8–0.9 mm wide and attached to the megasporophyll by a small stalk (FIG. 14.191). The integument is thickened in the micropylar region, and the outer layer of the integument (sarcotesta) extends out from the seeds to form a meshwork of filaments between the seeds. Gould and Delevoryas (1977) suggested that this meshwork (FIG. 14.190) may have been involved in some way in directing pollen flow, possibly aided by a pollination droplet. Nishida et al. (2007) hypothesized that perhaps the spongy meshwork served to insulate the ovules in the cool temperate environment in which this plant lived. The nucellus is free from the integument. Many of the seeds have well-preserved megagametophytes and a single archegonium (FIG. 14.191). Bisaccate pollen of the Protohaploxypinus type has been reported in the pollen chambers of these seeds (Gould and Delevoryas, 1977; Nishida et al., 2004). Especially significant is the report of pollen tubes

Figure 14.191 Seed of Homevaleia gouldii with archegonium

in megagametophyte tissue (Permian). Bar  0.25 mm. (Courtesy P. Ryberg.)

(FIG. 14.192) in various stages of releasing flagellated sperm in the region of the archegonium (Nishida et al., 2003). Sperm are top-shaped, small (12 μm in diameter), and characterized by spiral bands of dark dots near one end (FIG. 14.193). The dark spots are interpreted as corresponding to the positions of the basal bodies of numerous flagella aligned along a multilayered structure (MLS). The MLS is characteristic of charophycean algae and land plants with motile gametes (Nishida et al., 2004). The glossopterid sperm are reminiscent of those of cycads and Ginkgo but are much smaller. In general organization, the Homevaleia megasporophyll shows some similarities to the impression–compression ovulate structure Dictyopteridium, and Nishida et al. (2007) correlated the arrangement of ovules in the permineralized and compressed forms. Structurally preserved glossopterid megasporophylls with attached ovules are also known from permineralized peat in the Late Permian Buckley Formation of Antarctica (E. Taylor and Taylor, 1992; E. Taylor et al., 2007). One megasporophyll is 6 mm wide and 1 mm thick (FIG. 14.194) and is thought to have been produced by the same plant that bore Glossopteris schopfii leaves, based on similar anatomy

608

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.194 Section of megasporophyll (arrow) showing adaxial attachment of several seeds (Permian). Bar  1 mm.

MG

Figure 14.195 Suggested reconstruction of a Glossopteris megasporophyll with seeds attached to the adaxial surface. (From E. Taylor and Taylor, 1992.) Figure 14.192 Micropylar end of Glossopteris homevalensis ovule showing pollen tubes (arrows) and megagametophyte tissue (MG) (Permian). Bar  0.25 mm. (Courtesy H. Nishida.)

Figure 14.193 Sperm from Glossopteris homevalensis ovule.

Arrow indicates the position of possible basal bodies (Permian). Bar 10 μm. (Courtesy H. Nishida.)

(E. Taylor and Taylor, 1992). An interesting aspect of this structure is that the seeds are attached to the adaxial surface of the megasporophyll (FIG. 14.195) (E. Taylor and Taylor, 1992) based on the position of the vascular bundles. Nishida et al. (2007) also noted adaxial attachment of the ovules in Homevaleia gouldii. This position contrasts with the suggested abaxial attachment of ovules described from impression fossils (see below and Adendorff, 2005). A permineralized specimen from the Upper Permian of Antarctica appears to be similar to several of the so-called cupulate glossopterid reproductive structures (Surange and Chandra, 1975) such as Arberia or Rigbya (see below). A preliminary description suggests that it is a branching structure bearing at least four uniovulate cupules. Each cupule is 3 mm long and contains a sessile ovule with two flattened wings extending from the integument. The stalk at the base of the organ contains a C-shaped vascular strand similar to those that are produced in some leaves (E. Taylor et al., 2007). The ovules have a pad of tissue that surrounds the micropyle, and striate, bisaccate pollen grains have been found in the pollen chamber (FIG. 14.196). Dispersed ovules have been described from the Antarctic permineralized peat which show details of embryos

CHAPTER 14

PALEOZOIC SEED FERNS

609

Figure 14.196 Distal end of seed showing several pollen grains (arrow) in pollen chamber (Permian). Bar  150 μm.

development. Although they occur in close association with Glossopteris leaves, details of attachment are not known. They are considered to represent glossopterid seeds due to the overwhelming dominance of this group in the peat and in nearby floras of comparable age (Cúneo et al., 1993b). Seeds similar to Plectilospermum (FIG. 14.196) are 5 mm long and platyspermic with the nucellus and integument attached only at the base (T. Taylor and Taylor, 1987). Many seeds have a cellular megagametophyte with two archegonia, each showing evidence of a multicellular embryo, and confirm the existence of polyembryony (FIG. 14.197). Development had progressed in some embryos so that it is possible to distinguish the presence of a suspensor (Smoot and Taylor, 1986), a spirally coiled structure that is responsible in extant seeds for maintaining the continuity of the embryo with the nutritive tissue of the megagametophyte. Choanostoma is another isolated seed from the same Permian peat deposit in Antarctica (Klavins et al., 2001). It is bilateral with the integument and nucellus fused from the base to near the micropyle. At the apical end of the seed, two integumentary appendages overtop a funnel-shaped pollenreceiving structure containing several bisaccate pollen grains. The megagametophyte is known for C. verruculosum, but archegonia have not been observed. IMPRESSION–COMPRESSION SPECIMENS As mentioned earlier, there is an extraordinary diversity of ovule-bearing reproductive structures preserved as impression and compression specimens. Discussion of the interpretation and significance of these structures has been greatly hampered

Figure 14.197 Longitudinal section of Plectilospermum seed

with two archegonial chambers (arrows) (Permian). Bar  650 μm.

by the use of nonstandard and often nonhomologous terms to describe their morphology, for example, fertiliger, polysperm, and capitulum. Some authors have used an artificial system of morphofamilies based on macromorphological features, for example, Arberiaceae, Dictyopteridiaceae, Lidgettoniaceae, and Rigbyaceae, to classify and sort these structures (see Adendorff, 2005, for a history of glossopterid classification). Others have used different names for higherlevel taxonomy in the mistaken belief that families of fossil plants should be named after an ovulate organ rather than the oldest validly published genus in the group. Scutum (Dictyopteridiaceae) was one of the first ovulate structures described for the glossopterids (FIG. 14.198) and demonstrates a wide range of morphologies within the general shield shape of the megasporophyll (Plumstead, 1952). In the initial description of Scutum, the megasporophyll was thought to be bilaterally symmetrical, in the form of a twosided cupule borne on a pedicel attached to the midrib of a leaf. Specimens that were split open revealed ovules on one surface and bracts that were interpreted as microsporophylls on the other. On the basis of this interpretation, Scutum was initially considered to be a bisexual reproductive structure (Plumstead, 1958). Subsequent authors suggested that one half of the unit was a receptacle that contained a large number of ovules borne on elevated projections; the other half of the

610

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.199 Plumsteadia semnes showing numerous seeds

(Permian). Bar  1.7 cm. (From Rigby 1978; Courtesy P. Ryberg.)

Figure 14.198 Scutum rubidgeum (Permian). Bar  5 mm. (Courtesy S. McLoughlin.)

structure was a fertile scale that protected the ovules when they were young, detaching later to expose the micropyles to the presumably wind-borne saccate pollen. Rigby (1978a) and Schopf (1976) considered the ovules of Scutum as being borne on a single, fleshy structure, rather than on a two-parted receptacle and fertile scale. These authors were the first to suggest that the appearance of a two-parted structure was the result of splitting the rock through the middle of the ovulebearing organ. In subsequent studies the receptacle is simply the region of the sporophyll where the seeds are attached and we will use that term here only in a positional sense. Like many glossopterid ovulate structures, S. leslii shows numerous scars on the receptacle, which are interpreted as the former positions of seeds, but no seeds have been found attached. Extending out from the receptacle is a prominent wing with the margin slightly scalloped or dentate. Another glossopterid ovulate structure that shares some similarity to Scutum is Gladiopomum (Adendorff et al., 2002). This Early Permian ovulate organ, preserved as an impression, has a receptacle and winglike structure, but also extending from the distal end of the receptacle is a prominent

spine. Specimens of G. dutoitoides are up to 5.5 cm long with the number of seed scars exceeding 100. Although the shape of Plumsteadia (FIG. 14.199) (Plumstead, 1952) is similar to that of Scutum and Gladiopomum, it lacks the prominent spine, and the receptacle is generally more elongated. One of the earliest Permian glossopterid ovule-bearing structures described from India is Dictyopteridium (FIG. 14.200) (Feistmantel, 1886). Although the structure was initially interpreted as a fern leaf with sporangia, it is now known to represent a glossopterid ovule-bearing structure (Chandra and Surange, 1976). Morphologically, it represents a flattened organ, with a receptacle bearing seeds or seeds scars on one surface (FIG. 14.200) and a winglike extension (McLoughlin, 1990a,b). Differences between Plumsteadia and Dictyopteridium are based on the features of the seed scars (McLoughlin, 1990a,b). Bifariala intermittens (originally Hirsutum) is an isobilateral, dorsiventral, pedicellate organ from the Lower Permian Vryheid Formation of the Karoo Basin, South Africa, that is comprised of an ovule-bearing receptacle flanked by two superposed, peripheral wings (Prevec et al., 2008). This double-winged structure is unlike that of other glossopterid fertile organs. The primary wing is contiguous with the marginal seed scars on the fertile surface of the receptacle and the secondary wing is flush with the sterile surface of the receptacle. The fructification is attached near the base of a petiole of a Glossopteris leaf.

CHAPTER 14

PALEOZOIC SEED FERNS

611

Figure 14.201 Ottokaria reproductive structure showing the outlines of seeds (arrows) beneath the sterile portion of the head (Permian). Bar  6 mm. (From White, 1978; Courtesy P. Ryberg.)

Figure 14.200 Dictyopteridium sp. Pits represent former position of seeds (Permian). Bar  1 cm. (Courtesy S. McLoughlin.)

A peripheral winglike structure is also a feature of Ottokaria (FIG. 14.201). This ovulate organ, as originally described by Zeiller (1902–1903), was interpreted as a leaf with a fluted margin attached at the base to a leaf of Glossopteris (FIG. 14.202). The receptacle is round and the winglike extension is lobed, a character that can be used to distinguish Ottokaria from other glossopterid ovulate structures. Numerous species of Ottokaria have been described

Figure 14.202 Suggested reconstruction of Ottokaria zeilleri

showing relationship between vegetative leaf and megasporophyll. (From Pant et al., 1984.)

from throughout Gondwana (e.g., Pant and Nautiyal, 1984; Anderson and Anderson, 1985). Some glossopterid reproductive structures from the Upper Permian are interpreted as being compound. This morphological type was described by Surange and Chandra (1975)

612

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.203 Partha belmontensis showing attached seed bearing structures (arrows) (Permian). Bar  5 mm. (Courtesy S. McLoughlin.)

as cupulate, which they contrasted with the multiovulate type such as Scutum that bore many ovules on the surface of a megasporophyll. One of these so-called cupulate types is Lidgettonia (FIG. 14.204) (Lidgettoniaceae), a morphogenus that consists of one to several pairs of pedicels, each bearing a flattened megasporophyll (capitula of some authors; cupule of Surange and Chandra, 1975) that in turn bore seeds (H. Thomas, 1958). The pedicels arise from the petiolar region of a glossopterid scale leaf and are arranged in opposite pairs (FIG. 14.204). Each seed-bearing structure produced seeds on one surface, and was surrounded by a peripheral, winglike structure. The margin is variously lobed

Figure 14.204 Lidgettonia elegans (Permian). Bar  5 mm.

(Courtesy S. McLoughlin.)

and exhibits distinct crenulations or teeth. In L. africana the seeds are just a few millimeters long and circular in outline. The integument extends into a wing in the primary plane, which is uniform in width except at the micropyle, where it is notched. As the specific epithet implies, Mooia lidgettonioides can be compared to Lidgettonia, differing only in the smaller number of seed-bearing units and the radial morphology of the seeds (Lacey et al., 1975). The seed-bearing units

CHAPTER 14

PALEOZOIC SEED FERNS

613

Figure 14.206 Rigbya arberioides (Permian) Bar  5 mm.

(Courtesy S. McLoughlin.)

Figure 14.205 Diagrammatic reconstruction of Denkania indica. (From Taylor and Taylor, 1993.)

are borne on unbranched pedicels from the adaxial surface of the fertile leaf. The seeds are 4.2 mm long and nearly radial in symmetry. Anderson and Anderson (1985) suggested Mooia should be included in Lidgettonia. Another interesting ovulate structure that has been described from the Raniganj Stage of India is Denkania (FIG. 14.203). In this ovulate structure, approximately six seed-bearing cupules are attached to long pedicels borne on the midrib of a Glossopteris scale leaf. Each unit is ~1 cm long. Apparently, only one seed was produced per cupule. Partha is a cupulate structure from India that shows the same basic morphologic form of pedicels arising from the base of a scale leaf (FIG. 14.203) (Surange and Chandra, 1973). As additional information is accumulated about many of these seed-bearing organs, it will be interesting to see if the genera and species that have been instituted simply represent variations of the same pedicel-bearing, compound organs.

Figure 14.207 Rigbya arberioides showing lobes (arrow) and cupules (Permian). Bar  5 mm. (From White, 1978; Courtesy P. Ryberg.)

Two ovulate glossopterid genera that consist of flattened, variously lobed, simple organs are Rigbya (FIG. 14.206) (Rigbyaceae) (Lacey et al., 1975) and Cometia (McLoughlin, 1990a). They differ principally in the degree of lobing, as Rigbya has been described with between five and nine lobes (FIG. 14.207), whereas Cometia has two, although these characters are probably highly variable. Cometia biloba (Late Permian) exhibits what are interpreted as two slightly raised ovule scars (4 mm long) (FIG. 14.208). Arberia (FIG. 14.209) is another presumed glossopterid seed-bearing structure, which may have branched in various planes and

614

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.208 Cometia biloba (Permian). Bar  4 mm. (Courtesy S. McLoughlin.)

Figure 14.209 Arberia hlobanensis (Permian). Bar  5 mm.

(Courtesy S. McLoughlin.)

produced a single ovule at the end of each dichotomy (Rigby, 1972b). Isolated seeds preserved as compression–impressions of several types are common in rocks containing Glossopteris leaves (Pant et al., 1985). Two of the more commonly encountered morphotaxa are Pterygospermum and Stephanostoma (Pant and Nautiyal, 1960). Cuticle macerations of the nucellus of both types of seeds have yielded striate, bisaccate pollen typical of the glossopterids. Similar seeds from the Triassic of Nidpur, India, have been described under the generic name Rugaspermum (Pant and Basu, 1977). Although polyembryony is known for the group (Smoot and Taylor, 1986b), the only other postfertilization stage reported is that of Pant and Nautiyal (1987) who described Diphyllopteris, an early Permian axis bearing two pairs of opposite foliar organs that they interpret as being a glossopterid seedling. The lower pair of leaves is interpreted as the cotyledons, whereas the upper pair is considered to be the first true leaves of the Glossopteris type, based on the characteristic reticulate venation. WHAT IS THE GLOSSOPTERID OVULATE STRUCTURE? It is obvious that there is considerable morphological diversity among the ovule-bearing organs of the Glossopteridales and that within this paraphyletic group there are various interpretations regarding the nature of the reproductive structures. Although the gymnospermous affinities of the group have long been recognized, information about the reproductive biology of several members of the group has reinforced this concept (Smoot and Taylor, 1986b; Nishida et al., 2004, 2007). Nevertheless, the vexing problem regarding the glossopterids continues to be the homology of the ovule-bearing structure (FIG. 14. 210) and its morphological relationship to the subtending vegetative leaf (FIGS. 14.211, 14.212) (Doyle, 2006). By morphological and anatomical convention, the glossopterid ovulate structure is either a leaf (megasporophyll) or a stem system that bears megasporophylls. As mentioned earlier, one of the problems in discussing the ovulate structures in the glossopterids is the use of nonstandard terms to describe their morphology. Schopf (1976) suggested the term fertiliger for the megasporophyll, which Pant (1987) also used. Schopf (1976) proposed that the ovulate complex in the glossopterids represented an axillary shoot and suggested a distant relationship with the cordaites, termed the cladode concept by Retallack and Dilcher (1981a). The permineralized reproductive structures that have been described to date clearly indicate that the seeds are borne on the adaxial surface of a megasporophyll, although reconstructions based on impression fossils

CHAPTER 14

PALEOZOIC SEED FERNS

615

Figure 14.210 Suggested reconstruction of Austroglossa walkomii. (From Holmes, 1974.)

14.212 Detail of FIG. 14.211 showing attachment of cupules to leaf-bearing branch. Bar  5 mm. (Courtesy S. McLoughlin.)

Figure

Figure 14.211 Austroglossa walkomii (Permian). Bar  1 cm.

(Courtesy S. McLoughlin.)

suggest attachment to the abaxial surface. What remains to be determined is whether the megasporophyll is attached to a reduced and modified axillary branch system or to the base of the subtending vegetative or scale leaf (Pigg and Trivett,

1994). It may be that the structures which make up the glossopterid ovulate complex do not involve the same homologies in every case. The presence of various morphologic forms, including those that have multiple ovules (FIGS. 14.213, 14.214) and those with apparently solitary seeds, further underscores that the glossopterids were a diverse group. Seed morphology is also variable, with both radially and bilaterally symmetrical forms known. The presence or absence of wings on some seeds may be related to seed dispersal syndromes. The situation may in part parallel the extraordinary variation demonstrated among the seeds of Paleozoic seed ferns, which were produced on leaves, in cupules, and on naked branching systems. Thus, it would appear that the glossopterids, despite the apparent uniformity of their vegetative organs, represent a rather diverse group of late Paleozoic seed ferns that may ultimately rival the Carboniferous taxa in the variability of their ovulate reproductive structures. Although at least some of the glossopterid ovulate structures are homologous with a

616

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.213 Austroglossa walkomii showing seeds (arrows) attached to cupule (Permian). Bar  4 mm. (Courtesy P. Ryberg.)

megasporophyll, with seeds attached to the adaxial surface, exactly how that megasporophyll is related to the subtending vegetative leaf is currently very poorly understood. We echo the comments of many who have examined and discussed the glossopterid reproductive structures in noting that additional permineralized specimens will be critical in understanding the homologies of these unusual reproductive organs. One interesting feature is the apparent absence of any winglike structure on the megasporophyll in the permineralized specimens. It is doubtful that the enrolled edges of the megasporophyll of Homevaleia gouldii could create such a structure if compressed. The absence of the wing may simply reflect a developmental stage of the megasporophyll or a difference in taxa between South Africa and Australia, for example. The presence of this structure on many of the impression specimens suggests that there is some function for the wing. Protection of the developing ovules may be one obvious adaptation. Another adaptation could be dispersal, with the entire megasporophyll being shed as a unit (Adendorff et al., 2002), although this is very unlikely, since most impressed specimens appear to have already shed their seeds. Also, the presence of some seeds with integumentary wings may not be consistent with shedding the entire unit. POLLEN ORGANS

The glossopterid pollen organs described to date show little of the morphological diversity present in the seed-bearing organs. One of the most commonly encountered forms is

14.214 Vegetative leaf and megasporophyll of Lanceolatus lerouxides (Permian). Bar  1 cm. (From Anderson and Anderson, 1985.)

Figure

Glossotheca (Surange and Maheshwari, 1970). It consists of two pedicels that extend from the petiole of the leaf. At the end of each pedicel is a cluster of 100 elongated pollen sacs. On the surface of the pollen sacs are longitudinal

CHAPTER 14

PALEOZOIC SEED FERNS

617

Figure 14.216 Detail of several Arberiella pollen sacs showing

characteristic striations on the surface (Permian). Bar  1.5 mm. (Courtesy P. Ryberg.) Figure 14.215 Eretmonia natalensis with aggregation of pol-

len sacs (arrow) near stalk (Permian). Bar  6 mm. (From White, 1978; Courtesy P. Ryberg.)

wrinkles. Nothing is known about the pollen of this fructification. Another common pollen-producing organ believed to have been produced by the glossopterids is Eretmonia (FIG. 14.215) (Du Toit, 1932). The pollen sacs are arranged in two pedicellate clusters attached to a structure termed a fertile leaf. Individual pollen sacs are 1 mm long and variable in overall morphology (Surange and Maheshwari, 1970). The fertile leaves of Eretmonia are also highly variable in size, shape, and numbers of pollen sacs (FIG. 14.215). In general, the leaf is rather small and characteristically rhombohedral in outline, similar to some so-called scale leaves. Lacey et al., (1975) demonstrated that several morphologic forms defined as species merely represent gradational stages of a single type, E. natalensis. Pant (1987) suggested that the differences between Glossotheca and Eretmonia are so minor that they should be placed in the same genus. Arberiella (FIG. 14.216), Lithangium, and Polytheca are used for isolated pollen sacs that are believed to have been produced by the glossopterids (Pant and Nautiyal, 1960). In A. africana, the uniloculate sporangia are occasionally attached by slender stalks that closely approximate the arrangement in Glossotheca. Each pollen sac is 3 mm long and exhibits twisted cells at the apex that may have served as a dehiscence mechanism. Pollen grains are bisaccate and have endoreticulations extending from the inner surfaces of their saccus walls (Zavada, 1991). Folds or striations extend across the cappa (FIG. 14.217) of the central body,

Figure 14.217 Pollen grain with proximal striations. Bar  10 μm. (Courtesy S. McLoughlin.)

suggesting affinities with Protohaploxypinus. Lindström et al. (1997) extracted pollen from a single Arberiella sporangium (FIG. 14.218) and demonstrated that the pollen grains represent four different morphogenera and six morphospecies of sporae dispersae. In one Arberiella species, A. vulgaris, the grains are up to 85 μm long and 55 μm wide. Arberiella-type sporangia have also been reported by Gould and Delevoryas (1977) in the same peat that contains the permineralized glossopterid megasporophyll Homevaleia gouldii. Arberiella pollen sacs are also known in the compressed pollen organ Nesowalesia (Rigby and Chandra, 1990). Lithangium and Polytheca are additional unilocular sporangia, but they have been described as containing monolete grains in Lithangium and

618

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 14.219 Suggested reconstruction of Perezlaria oaxacensis fertile structure. (From Delevoryas and Gould, 1971.) Figure 14.218 Monosaccate (arrow) and bisaccate pollen grains in a glossopterid pollen sac. Bar  50 μm. (Courtesy S. McLoughlin.)

monocolpate pollen with a small triradiate mark on the proximal face, in the case of Polytheca. Kendostrobus is a pollen organ that is suggested to have glossopterid affinities (Surange and Chandra, 1975). It is interpreted as conelike, with a central axis of helically arranged, naked, exannulate sporangia arranged in groups. On the pollen sac walls are minute pits. Pollen of K. cylindricus is elliptical–subcircular and 100 μm long. Surface ornamentation is a series of parallel ridges, and on one face, there is a monolete suture. Although the organization of Kendostrobus a is unlike the other presumed Glossopteris pollen-producing structures, the monolete grains are like those macerated from Lithangium pollen sacs. Associated with Mexiglossa leaves from Mexico is an interesting putative pollen-producing organ. Perezlaria is reconstructed as an axis with laterals that bear whorls of saclike bodies that are interpreted as pollen sacs (FIG. 14.219) (Delevoryas and Gould, 1971). Each sporangium is 3 mm long. Morphologically, this paniclelike reproductive structure bears some resemblance to the pollen organ Caytonanthus. It is not known whether Mexiglossa foliage and Perezlaria represent organs of the same biological species or merely demonstrate a fortuitous association in the rock record. GLOSSOPTERIS HABIT AND HABITAT

Some glossopterids have been interpreted as small herbaceous plants (Singh, 2000), or even some form of liana (Pant, 1999). Today, however, the consistent co-occurrence of woody axes and the records of attachment of leaves to stems indicate that many were probably woody trees of one size or another (FIG. 14.178) (Schopf, 1970; Gould and Delevoryas, 1977; Pant, 1977). Most appear to have been

seasonally deciduous, especially those growing at high paleolatitudes (E. Taylor et al., 2000). Growth rings in both the trunks and roots indicate levels of periodic growth, (E. Taylor et al., 1992) and the presence of aerenchyma in the roots has been used to suggest that Glossopteris grew in soils with a fluctuating water table. Although Glossopteris was no doubt a coal-swamp plant in many localities, it grew in more environments than this single habitat. Cúneo et al. (1993b) examined multiple floras and depositional environments in the Upper Permian of the Transantarctic Mountains, Antarctica, and showed that Glossopteris occurred in all types of environments. Remains were found in coarse- and fine-grained fluvial deposits, as well as in lacustrine environments. These high-latitude biotas consisted of relatively dense, but low diversity, forest growth. On some bedding planes, glossopterid remains accounted for up to 80% of the plant remains. PHYLOGENETIC POSITION

Like many other Paleozoic and Mesozoic pteridosperms, glossopterids have, at one time or another, been suggested as possible ancestors to the angiosperms. Perhaps the earliest advocate of this relationship was Plumstead (1956), who interpreted the glossopterid ovulate organs as bisexual. The glossopterid–angiosperm connection has periodically gained support as additional information about the ovulate organs has been published and hypotheses advanced that attempt to homologize these organs with those of other seed plants (Melville, 1983). For example, Retallack and Dilcher (1981a) championed the idea put forward by Stebbins (1974) of glossopterids as flowering plant progenitors. They proposed a series in which the number of ovules on the megasporophyll is reduced to one. The megasporophyll would then enclose this ovule to form a second integument and the subtending leaf would form the carpel. Doyle (2006) also discussed this

CHAPTER 14

idea as a way to evolve the angiosperm carpel, with perhaps Caytonia (Chapter 15) as an intermediate. The glossopterids are also an interesting group for ecological reasons. They must have moved into open niches as the glaciers retreated in the Pennsylvanian–Early Permian (E. Taylor, 1996) and continued to dominate most Gondwanan ecosystems, especially those at high latitudes, for almost 50 myr—a remarkably well-adapted group of plants! Glossopterid stratigraphic distribution is particularly noteworthy, as most reconstructions suggest a steadily warming climate throughout the Permian (Kidder and Worsley, 2004). Phylogenetic analyses have treated the group as a single, composite terminal, constructed of various organs, thus providing little resolution of the group’s diversity in the analysis (Nixon et al., 1994; Doyle, 2006; Hilton and Bateman,

PALEOZOIC SEED FERNS

619

2006). There has been some attempt to relate isolated organs based on associations and anatomy, and in some cases, the vegetative parts of plants have been reconstructed, for example, Glossopteris schopfii and G. skaarensis from Antarctica (Pigg and Taylor, 1993) and G. homevalensis from Australia (Pigg and Nishida, 2006). E. Taylor and Taylor (1992) were able to attribute the ovulate structures from Antarctica to G. schopfii on the basis of anatomy. These studies represent a first step in taking what is probably a heterogeneous group and developing a framework that can be used to test subsequent hypotheses of relationships, not only within the glossopterids, but with other seed plants. Until homologies can be established, there is currently insufficient evidence to link the glossopterids with any other major group of plants, either as progenitors or descendants.

This page intentionally left blank

15 Mesozoic seed ferns Caytoniales ............................................................................. 622

Pollen Organs ................................................................................... 631

Sagenopteris ..................................................................................... 622

Ovulate Structures............................................................................ 634

Caytonanthus .................................................................................... 623

Petriellales ................................................................................637

Caytonia ........................................................................................... 624 Peltaspermales........................................................................ 639

Ruflorinia and Ktalenia .................................................................... 626

Foliage .............................................................................................. 639

Corystospermales ................................................................627

Reproductive Organs and Whole-Plant Concepts ............................ 643

Foliage .............................................................................................. 627 Conclusions ........................................................................... 648

Stems ................................................................................................ 630

The most beautiful thing we can experience is the mysterious. It is the source of all true art and science. Albert Einstein, What I Believe

A number of predominantly Triassic and Jurassic gymnosperms have historically been termed Mesozoic seed ferns. They do not form a natural group, but rather represent a collection of either orders or families in which the relationships among the various taxa are uncertain (E. Taylor et al., 2006). These extinct gymnosperms are known principally from impression and compression specimens and therefore not understood with the same degree of resolution as the Carboniferous pteridosperm groups. Nevertheless, they possess some morphological features that have been used to suggest affinities with some of the Paleozoic seed ferns (Petriella, 1981) (FIG. 15.1). As research has continued, even the characters initially used to unite the groups, such as seeds attached to megasporophylls, are in some cases no longer applicable, resulting in orders that are not resolved based on phylogenetic analyses (Nixon et al., 1994; Axsmith et al., 2000). In addition, some of these groups are now known from the Paleozoic, so the name Mesozoic seed ferns is perhaps no longer applicable. These seed plants are the subject of considerable interest as potential angiosperm progenitors, based on a number of features, most notably the enclosure of

621

Figure 15.1

Bruno Petriella. (Courtesy S. Archangelsky.)

622

paleobotany: the biology and evolution of fossil plants

the ovules in some type of leaf like structure, usually termed a cupule. We have followed the convention of treating these plants as four orders (E. Taylor et al., 2006).

Higher taxa in this chapter:

Caytoniales (Triassic–Cretaceous) Corystospermales (Upper Permian–Cretaceous) Petriellales (Triassic) Peltaspermales (Pennsylvanian–Triassic)

Sagenopteris

The most common foliar element of the Cayton Bay flora is Sagenopteris (FIG. 15.2), a palmately compound leaf consisting of three to five leaflets. Thomas (1925) associated this leaf with Caytonia on co-occurrence and cuticular similarities. Sagenopteris (FIG. 15.3), however, is a widespread taxon known from a number of widely separated geographic localities, including South and North America, Greenland, Europe (Alvarez Ramis, 1982; Barale, 1982 (FIG. 15.4);

Caytoniales The Caytoniales are a small group of Mesozoic seed plants (Krassilov, 1977a) that was erected in 1925 by the British paleobotanist H. H. Thomas, based on compression specimens from the Middle Jurassic plant-bearing beds along the coast of Cayton Bay in Yorkshire, Great Britain. So striking and unusual were some of the features of these plants that Thomas (1925) regarded them as a new group of angiosperms in his initial description. Since the initial description of the seed-bearing organs Caytonia and Gristhorpia, which was later merged into Caytonia, paleobotanists have been fascinated by the possible link between this cupulate organ and an angiosperm carpel (Doyle, 1978, 1996, 2006). Thus the Caytoniales have figured prominently in a number of phylogenetic analyses, although they are one of the most poorly known groups covered in this chapter.

Figure 15.3 Sagenopteris colpodes leaf. (From Crane, 1985a.)

Figure 15.2 Sagenopteris sp. (Jurassic). Bar  1 cm. (Courtesy

S. McLoughlin.)

Figure 15.4 Georges Barale.

CHAPTER 15

Zaton et al., 2006; Kustatscher et al., 2007), Japan (Kim and Kimura, 1987), and Antarctica (Cantrill, 2000a; Rees and Cleal, 2004), ranging from the Middle Triassic (Anisian; see Kustatscher et al., 2007) to the Cretaceous (Albian; Sender et al., 2005). Caytonialean reproductive organs, however, are not known from all of these areas and have never been described from North America. The lanceolate leaflets of S. serrata are up to 7 cm long; each contains a prominent midvein and anastomosing laterals that form a reticulate venation pattern (Harris, 1932a). Leaflets show some variability in morphology ranging from entire to those with lobed margins (Rees, 1993). In S. phillipsii, the cuticle is thick and epidermal cell walls are straight; stomata are of the anomocytic type (Barbacka and Bóka, 2000). Leaflets from the Early Jurassic of Alaska are up to 10 cm long and have papillate structures on the upper cuticle (Barbacka et al., 2006). Scoresbya is a palmately organized leaf generally similar to Sagenopteris but with individual leaflets that are once or twice forked (Cao, 1982; Weber, 1985; Schweitzer and Kirchner, 1998). This foliage type has been variously assigned to the ferns, that is, Dipteridaceae (Kräusel and Schaarschmidt, 1968; Herbst, 1979; Tidwell and Ash, 1994), and the seed ferns or Caytoniales.

Figure 15.5 Several Caytonanthus pollen sacs (Jurassic).

Bar  3 mm.

mesozoic seed ferns

623

Caytonanthus

Caytonanthus (originally Antholithus) is a pollen-bearing structure that was also found in the Middle Jurassic Yorkshire flora in association with Sagenopteris leaves. Thomas (1925) included it in the Caytoniales on the basis of the pollen, which was also found in Caytonia. It consists of a slender axis bearing flattened, pinnate lateral branches; each branch bears from one to three elongate synangia (FIG. 15.5) (Osborn, 1994; Osborn and Taylor, 1994). Each synangium is up to 1 cm long, pointed at the distal end (FIG. 15.6), and contains one to four pollen sacs (originally termed locules) that are arranged around a central zone of tissue (Harris, 1941a). These structures (which have been termed anthers in some treatments) are radially symmetrical, with dehiscence taking place toward the center of the synangium. The epidermis is composed of delicate fusiform cells, perhaps with thickerwalled fibrous cells beneath. Pollen grains of Caytonanthus are small and bisaccate, and if found dispersed are included in Vitreisporites. The pollen is 25–40 μm in diameter, and contains endoreticulations lining the interior of the sacci, thus making these grains eusaccate (Zavada and Crepet, 1985; Osborn, 1994). Others interpret the grains as protosaccate, in which the endoreticulations are continuous between saccus wall and surface of the corpus (Pedersen and Friis, 1986). The ultrastructure of the exine indicates that the sexine is alveolate (FIG. 15.7) and there is a conspicuous sulcus on the distal surface. Pollen from C. kochii is about 30 μm in diameter (Zavada and Crepet, 1986).

Figure 15.6 Suggested reconstruction of Caytonanthus arberi axis. (From Crane, 1985a.)

624

paleobotany: the biology and evolution of fossil plants

Caytonia

The ovule-bearing structure of the Caytoniales is named Caytonia, and consists of an axis 5 cm long bearing stalked, multiovulate cupules in subopposite pairs (FIG. 15.8) (Thomas, 1925; Harris, 1964). Scars along this axis, which some interpret as a megasporophyll, suggest that the cupules were shed. Each cupule is nearly circular in outline and up to 4.5 mm in diameter. They are borne along the axis in such a way that the cupule is recurved (FIG. 15.9), with a liplike projection directed toward the point of attachment (Harris, 1940). A single cupule (FIGS. 15.10, 15.11) contains 8 to

Figure 15.7 Ultrathin section of Caytonanthus pollen grain

showing alveolate organization of the wall (Jurassic). Bar  1 μm. (Courtesy J. M. Osborn.)

Figure 15.9 Caytonia fertile branch showing position of several cupules. (From Taylor and Taylor, 1993.)

Figure

15.8 Three stalked Caytonia cupules (Jurassic).

Bar  5 mm.

Figure 15.10 Suggested reconstruction of Caytonia cupule showing attached seeds. (From Dilcher, 1979.)

CHAPTER 15

30 seeds, depending on the species. Each seed is attached by a delicate stalk in an orthotropous position along the midvein of the cupule (FIG. 15.10). Seeds are 2 mm long and radially symmetrical. The integument consists of an outer, uniseriate epidermis that covers a row of radially aligned, thick-walled cells (Harris, 1958). An inner cuticle has been isolated from C. sewardii seeds. Harris (1951a) suggested that perhaps the outer portion of the integument was fleshy, much like a berry. Beneath the epidermis are several rows of

mesozoic seed ferns

625

longitudinally directed fibers; the nucellus is attached only at the base. Macerations suggest that the seed integuments are vascularized. Dispersed seeds of this type are sometimes placed in the morphogenus Amphorispermum (Harris, 1931b, 1932a). In the initial description of Caytonia, pollen grains were found on the liplike portion of the cupule, a structure that Thomas (1925) (FIG. 15.12) termed the stigmatic surface. This was no doubt the primary reason that the author described the cupule as an angiospermous fruit containing numerous seeds. According to his interpretation, fertilization would have taken place via a pollen tube that grew from the pollen grain on the stigmatic surface to the pollen chamber of an ovule. Fine strands of cuticle observed extending from the cupule to the seeds were thought to be the remnants of pollen tubes or an extension of the seed micropyles. In a subsequent study, however, Harris (1940) (FIG. 15.13) demonstrated that pollen grains were present inside the cupule, and suggested that they were probably drawn in by pollination droplets that originated at the micropylar end of each seed. The next step in determining the relationship between the ovules and cupules was provided by Maria Reymanówna (1973) (FIG. 15.14). Her detailed work demonstrated that each seed was associated with an elongate canal extending from the seed micropyle to the outer lip of the cupule. The

Figure 15.11 Caytonia sewardii cupules (Jurassic). Bar  5 mm. (Courtesy LPPU and H. Kerp.)

Figure 15.12

H. Hamshaw Thomas.

Figure 15.13 Tom M. Harris. (Courtesy T. Delevoryas.)

626

paleobotany: the biology and evolution of fossil plants

Figure 15.14 Maria Reymanówna (left) and Jeanne Doubinger.

(Courtesy T. Delevoryas.)

number of these canals was consistent with the number of seeds produced in each cupule, for example C. sewardii, 8; C. nathorstii (originally Gristhorpia nathorstii; Thomas, 1925), 15; and C. thomasii, 30. Today Caytonia is interpreted as a multiovulate cupule and not an angiospermous fruit as originally thought. The structure does, however, represent at least one method whereby seeds may have become enclosed and in this way does approach the morphology, although not the function, of an angiosperm carpel (Krassilov, 1977a). Doyle (1978, 2006) has supported the idea, proposed earlier by Gaussen (1946), that the Caytoniales represent angiosperm ancestors. This theory suggests a reduction in the number of ovules in Caytonia to one per cupule. The cupule wall would then be homologous with the second integument in an angiosperm ovule and the cupule-bearing axis would give rise to the carpel. Reymanówna (1973) provided additional information about the histology of Caytonia based on specimens collected in the Jurassic of Poland. In C. harrisii the cupule is vascularized by a flattened plate of tracheids with bordered pits that extend up the middle of the cupule and give off traces to the seeds. This species produced a cluster of seeds in the center of the cupule, and the seed cluster was covered by a cutinized membrane, rather than being separated by cupule tissue, as has been suggested in other species. The large number of isolated seeds found in the matrix when Caytonia specimens

Figure 15.15 Diagrammatic section of Ktalenia circularis cupules. (From T. Taylor and Archangelsky, 1985.)

are uncovered suggests that the seeds were probably dispersed from the cupules at some point in time. The Caytonia plant has been reconstructed as a small tree, based on discoveries of woody axes with attached Sagenopteris foliage (Harris, 1971). LaPasha and Miller (1985) reconstructed the plant that bore Sagenopteris williamsii leaves from the Kootenai Formation as growing in a swamp environment based on its association with lignites.

Ruflorinia and Ktalenia

Ruflorinia is a morphogenus for bi- and tripinnate leaves from the Lower Cretaceous (Aptian) Anfiteatro de Ticó Formation (formerly Baqueró Formation) of Argentina (Chapter 16), whereas isolated cupules from the same site are called Ktalenia (FIGS. 15.15, 15.16) (Archangelsky, 1963). The cupules of K. circularis are morphologically almost identical to those of Caytonia; however, Ktalenia has only two seeds per cupule (T. Taylor and Archangelsky, 1985). Ultrastructural studies of the cuticle structure suggest that some species of Ruflorinia lived in xeric habitats (Villar de Seoane, 2000).

CHAPTER 15

mesozoic seed ferns

627

Figure 15.17 Dicroidium odontopteroides leaf showing characteristic bifurcation.

Figure 15.16 Suggested reconstruction showing Ruflorinia

frond with Ktalenia cupules. (From T. Taylor and Archangelsky, 1985.)

the same beds, the similarity in epidermal anatomy, and the identical pollen found in both pollen organs and seeds have been used to indirectly associate the various organs (Petriella, 1983). When permineralized, the various parts have been put together, like the Lyginopteris oldhamia plant (Chapter 14), on the basis of unique secretory cells. Foliage

Corystospermales Like the Caytoniales, the Corystospermales are a relatively small group of plants. They have primarily a Gondwanan distribution and are known from Triassic localities in Antarctica, South Africa, Australia, Argentina, Tasmania, and India; the stem morphotaxon Rhexoxylon has been reported from the Lower Jurassic of southern Africa (Anderson and Anderson, 1983), and the cupulate organ Umkomasis has been reported from the Permian of India (Chandra et al., 2008). Corystosperm reproductive structures and foliage are also known from the Jurassic of Europe (Harris, 1964; Kirchner and Müller, 1992) and Triassic of China. The family was instituted by Thomas in 1933 for helmet-shaped, uniovulate cupules borne on a branching system. Based on the morphological diversity within the corystosperms, Archangelsky (1996) suggested that they were a varied and rapidly evolving group during the Triassic. The plants were probably small to large woody shrubs and trees and bore pinnate leaves with open dichotomous venation. Although the various plant parts have not been found attached, their consistent occurrence in

Several foliage types are believed to have been produced by the corystosperms. The most common of these is Dicroidium, a frond characterized by a basal bifurcation of the rachis (FIG. 15.17), which includes a large number of species that range from pinnate (FIG. 15.18) to tripinnate (FIG. 15.19) to entire (FIG. 15.20). Pinnae are usually subopposite and decurrent (Townrow, 1957; Anderson and Anderson, 1983), and pinnules (FIG. 15.20) are pinnatifid (FIG. 15.19) to needlelike. Venation varies from sphenopteroid to taeniopteroid. The cuticle is variable in thickness, with trichomes present in a few species; most species are amphistomatic. Permineralized specimens of Dicroidium are known from the Middle Triassic of Antarctica (Pigg and Taylor, 1987). Dicroidium fremouwensis consists of a portion of a leaf 9 cm long (Pigg, 1990b) (FIG. 15.21). Venation is of the odontopteroid type and stomata are dicyclic. Near the basal bifurcation, the petiole contains two groups of vascular bundles; eight bundles are arranged in a ring near the abaxial surface and six more occur laterally near the adaxial surface. Tracheids are scalariform-reticulate and associated with smaller, reticulate transfusion cells. On the adaxial side of the leaf are palisade parenchyma cells.

628

paleobotany: the biology and evolution of fossil plants

Figure 15.18 Dicroidium odontopteroides leaf with cupulate

structure, Umkomasia uniramia (arrow) (Triassic). Bar  1.5 cm.

Figure 15.20 Dicroidium dutoitii leaf (Triassic). Bar  2 cm.

Figure 15.19 Suggested reconstruction of Dicroidium zuberi.

(From Taylor and Taylor, 1993.)

Because Dicroidium is the dominant foliage type in the Triassic of Gondwana, it has been used to delimit megafossil zones for biostratigraphy (Anderson and Anderson, 1983) and was used in the past as a stratigraphic marker for the Triassic. It is interesting, however, that the earliest representatives of the genus do not come from Gondwanan deposits, but rather have been documented from the Upper Permian Um Irna Formation (Dead Sea region, Jordan) of the Arabian Peninsula based on large compression fossils (FIG. 15.22) with excellently

CHAPTER 15

mesozoic seed ferns

629

Figure 15.21 Cross section of Dicroidium fremouwensis leaf (Triassic). Bar  1 mm.

Figure 15.23 Pinnule cuticle of Dicroidium jordanensis

(Permian). Bar 5 mm. (Courtesy H. Kerp.)

Figure 15.22 Dicroidium jordanensis (Permian). Bar  1 cm. (Courtesy H. Kerp.)

preserved cuticles (FIGS. 15.23–15.25) (Kerp et al., 2006; Abu Hamad et al., 2008). These authors suggested that Dicroidium originated in the paleotropics during the Permian. During the climatic warming of the Late Permian–Early Triassic, the plants migrated southward to eventually colonize

Figure 15.24 Abaxial cuticle of Dicroidium irnensis showing stomatal complexes (Permian). Bar  50 μm. (Courtesy H. Kerp.)

630

paleobotany: the biology and evolution of fossil plants

Figure 15.25 Stomatal complex of Dicroidium jordanensis (Permian). Bar  50 μm. (Courtesy H. Kerp.)

the entire Gondwana region by the Middle and Late Triassic. Retallack (1977) reconstructed Dicroidium as one of the major plant types in a composite of Middle Triassic vegetation in eastern Australia; in this approach the author correlates different species of Dicroidium with particular paleoecological settings. Xylopteris and Johnstonia are two additional Triassic foliage types that traditionally have been included with the corystosperms (Pattemore and Rigby, 2005). Both are characterized by uni- to tripinnate fronds with a bifurcating rachis. Pinnules of Johnstonia may be entire or lobed and characterized by taeniopteroid or odontopteroid venation. In Xylopteris the vasculature includes a single main vein that runs the length of the foliar segments (Petriella, 1979). The use of infrared spectroscopy has been applied to cuticles of various species of Johnstonia as a potential chemotaxonomic application (D’Angelo, 2006). Xylopteris is amphistomatic with the lower cuticle typically thicker (Baldoni, 1980). Two additional Southern Hemisphere foliage genera included in the corystosperms are Diplasiophyllum and Zuberia (Gnaedinger and Herbst, 1998, 2001). Some authors, however, regard Diplasiophyllum, Johnstonia, Xylopteris, and Zuberia as synonyms of Dicroidium (Archangelsky, 1968; Anderson and Anderson, 1983; Kustatscher and Van Konijnenburg-Van Cittert, 2007). Pachypteris (Chapter 16) is a primarily Jurassic foliage type (FIG. 15.26) that is known throughout the world (Cleal and Rees, 2003), and is attributed to the corystosperms. Although most reports of Pachypteris are from the Northern Hemisphere, the genus has been reported from Gondwana, where it extends into the Cretaceous (Cantrill, 2000b). Leaves may be uni- or tripinnate, but the main rachis does not

Figure 15.26 Pachypteris indica (Cretaceous). Bar  1 cm.

(Courtesy D. Cantrill.)

bifurcate (Harris, 1964) (Chapter 16). The stem of Pachypteris papillosa from the Middle Jurassic of Yorkshire ranges from 10–50 mm in diameter (Harris, 1983). On the surface, there are numerous swollen leaf bases and epidermal blisters. Harris (1964) suggested that this corystosperm grew on the banks of tidal river channels. Stems

The corystosperms are considered to be woody plants based on the co-occurrence of Dicroidium foliage and axes of the petrified or permineralized woody stem Rhexoxylon (FIG. 15.27) in the Ischigualasto Formation of Argentina (Archangelsky, 1968), although the foliage and trunks are now known from other sites in South America. Some have suggested that Rhexoxylon stems were lianescent, based on the unusual stem anatomy (Walton, 1923; Kräusel, 1956). The large size (70 cm in diameter) of some of the Rhexoxylon stems, however, clearly argues against a lianescent habit for these plants (Archangelsky and Brett, 1961; Artabe et al., 1999). The stem of Rhexoxylon consists of a ring of primary vascular bundles that produce wedges of secondary xylem on both sides, that is, centrifugally and centripetally. There may be additional rings of xylem wedges outside the initial ring, and the configuration differs based on the species (Artabe

CHAPTER 15

Figure 15.27 Cross section of a segment of Rhexoxylon wood

from Antarctica showing growth increments in secondary xylem (Triassic). Bar  5 mm.

et al., 1999). The wedges of xylem are separated by parenchymatous tissue that is continuous with the pith and includes numerous bundles (many with secondary xylem) and sclerotic nests (FIG. 15.28). In R. piatnitzkyi the young stem has a ring of about 15 primary bundles separated by narrow pith rays (Archangelsky and Brett, 1961). The pith (FIG. 15.28) includes sclerotic nests and abundant secondary parenchyma organized into irregular masses (Brett, 1968). Tracheid pitting is circular bordered on the radial walls and growth rings are regular in the wedges of wood. On the outer surface of young stems, there are rhombic leaf bases and branch scars or buds surrounded by small scale leaves. Rhexoxylon is known from South America and South Africa (Bancroft, 1913; Kräusel, 1956; Herbst and Lutz, 1988) and also from Antarctica (E. Taylor, 1991). In recent years ideas about the diversity and habit of some of the corystosperms have changed. This has come about as a result of the discovery of additional permineralized stems associated with Dicroidium leaves from the Middle Triassic of Antarctica (Meyer-Berthaud et al., 1992, 1993). Specimens of Kykloxylon (FIGS. 15.29, 15.30) range from small leafy stems to axes with secondary xylem showing growth rings. Distinct secretory cavities in the Dicroidium leaves (D. fremouwensis) and stems (FIG. 15.30) indicate that they are parts of the same plant. Other woody axes in the form of in situ stumps (FIG. 1.54) are also known from Antarctica (Del Fueyo et al., 1995). Specimens of Jeffersonioxylon are up to 61 cm in diameter and are thought to have been at least 30 m tall (Cúneo et al., 2003). In the shales surrounding the stumps were numerous Dicroidium leaves, and there is anatomical evidence that supports the belief that these plants were seasonally deciduous (MeyerBerthaud et al., 1993).

mesozoic seed ferns

631

Figure 15.28 Cross section of Rhexoxylon piatnitzkyi stem showing wedges of vascular tissue with parenchymatous zones in between. Inset shows centripetal and centrifugal wood. (From Archangelsky and Brett, 1961.)

Figure 15.29 Cross section of Kykloxylon fremouwensis axis

with abundant secondary xylem (Triassic). Bar  2.5 mm.

From the foregoing, it is clear that the Corystospermales include different types of plants. The Dicroidium plant from South America was reconstructed by Petriella (1978) as having the appearance of a tree fern (FIG. 15.31), and includes the trunk Rhexoxylon. The Dicroidium plant from Antarctica, however, was probably a larger forest tree, which had coniferlike wood (FIG. 15.32), even though it also bore Dicroidium fronds (Taylor, 1996). The habit of a third type of Gondwana corystosperm may be represented by the stem Tranquiloxylon petriellai that has secondary xylem characterized by peripheral lobes separated by parenchyma (Herbst and Lutz, 1995). Pollen Organs

The most common pollen organ included in the Corystospermales is Pteruchus (FIG. 15.33). The genus is known

632

paleobotany: the biology and evolution of fossil plants

L L

L

Figure 15.30 Cross section of a young Kykloxylon fremouwensis showing secretory cells (arrows) and helically arranged leaves (L) (Triassic). Bar  2 mm.

Figure 15.32 Suggested habit of Dicroidium plant from

Antarctica. (From E. Taylor et al., 2006.)

Figure 15.31 Suggested reconstruction of Dicroidium plant

from the Ischigualasto Formation. (From Petriella, 1978.)

throughout Gondwana (see E. Taylor, 1996), but there are also a few reports from the Northern Hemisphere, for example P. septentrionalis from Rhaeto-Liassic (Late Triassic– Early Jurassic) of Germany (Kirchner and Müller, 1992). Most specimens have alternately arranged microsporophylls

that are attached to an axis about 4 cm long (Thomas, 1933; Townrow, 1962a) (FIG. 15.33). In one species they appear to be helically arranged (Pant and Basu, 1979), whereas in another they are described as being subopposite (Townrow, 1962a). Each microsporophyll terminates in a flattened head (FIG. 15.34) that bears numerous elongate pollen sacs on the abaxial surface, partially protected by the tissue of the head (FIG. 15.34). The number of sacs per head is variable, ranging from 20 to 200, and dehiscence is longitudinal. The cuticle on all parts shows epidermal cells with slightly undulate margins and few stomata. Pollen is bisaccate, with the sacci slightly inclined and partially covering the distal sulcus. Pollen of

CHAPTER 15

mesozoic seed ferns

633

Figure 15.33 Branch bearing several Pteruchus sp. microsporophylls (Triassic). Bar  5 mm. (Courtesy C. P. Daghlian.) Figure 15.35 Longitudinal section of Pteruchus fremouwensis axis showing attached pollen sacs (arrows) (Triassic). Bar  3 mm.

Figure 15.34 Pteruchus microsporophyll with attached pollen

sacs (arrows) (Triassic). Bar  2 mm. (Courtesy C. P. Daghlian.)

P. dubius ranges from 80–115 μm in the primary plane (T. Taylor et al., 1984). The sacci contain endoreticulations on their inner surfaces, but the outer surfaces are smooth (Zavada and Crepet, 1985). Permineralized pollen organs of Pteruchus are known from peat deposits from the early Middle Triassic of Antarctica (FIG. 15.35) (X. Yao et al., 1995). Pteruchus fremouwensis

Figure 15.36 Suggested reconstruction of Pteruchus fremouwensis fertile axis. (From Yao et al., 1995.)

634

paleobotany: the biology and evolution of fossil plants

consists of a branch bearing helically arranged microsporophylls (FIG. 15.36). Each microsporophyll has a flattened head bearing 20 sessile, unilocular elongate pollen sacs. Pollen sacs are 2 mm long (FIG. 15.37) and possess walls that contain secretory cells like those in leaves of Dicroidium from the same peat deposit (Pigg and Taylor, 1990). If found dispersed, the bisaccate pollen of Pteruchus would be assigned to Alisporites, Pteruchipollenites, or Falcisporites (FIG. 15.38). Another pollen organ believed to belong to the corystosperms is Pteroma. Pteroma is known to date only from the Middle Jurassic of Yorkshire, where it is associated with Pachypteris type foliage (Harris, 1964). In P. thomasii the microsporophyll is flattened and bears two rows of pollen sacs from what is interpreted as the lower surface. Each head is shield shaped, with the pollen sacs partially embedded in the tissue. Pollen is bisaccate and up to 107 μm long. It is not known whether these grains are protosaccate or eusaccate like those of Pteruchus (Osborn and Taylor, 1993, 1994). Other putative corystosperm pollen organs include Nidiostrobus from India (Bose and Srivastava, 1973) and Kachchhia from India and Antarctica (Bose and Banerji, 1984; Gee, 1989).

Figure 15.37 Section of Pteruchus fremouwensis pollen sac showing bisaccate grains (Triassic). Bar  250 μm.

Ovulate Structures

One of the interesting aspects of the corystosperms is that the reproductive organs are morphologically consistent within the group. The most common seed-containing structure is Umkomasia (FIG. 15.39) (Thomas, 1933), which is known widely from Gondwanan deposits (Taylor, 1996), but has also been reported from Europe (Kirchner and Müller, 1992) and China (Zan et al., 2008). It consists of a fertile branch (FIG. 15.40), more than 15 cm long in some taxa, for example U. quadripartita (Anderson and Anderson, 2003), which produces laterals. Each lateral bears one to several pairs of recurved, helmet-like, and in most species, uniovulate cupules. Some specimens show a pair of subtending bracts at the base of the branch; in others they are not present or may have been shed (Holmes, 1987). The branch bearing the cupules in Umkomasia was originally interpreted as a reduced and flattened pinna and the cupules as pinnules (Thomas, 1933). Permineralized specimens from the Middle Triassic of Antarctica, however, demonstrate that the cupulate branch has stem-like anatomy and produces paired traces to the laterals. The cupule-bearing branches of U. resinosa are helically arranged (FIG. 15.41) and cupules contain either one or two ovules attached to the abaxial surface (FIG. 15.42) of the cupule (Klavins et al.,

Figure 15.38 Bisaccate pollen grain extracted from Pteruchus fremouwensis pollen sac (Triassic). Bar  25 μm.

Figure 15.39 Diagrammatic reconstruction of Umkomasia uni-

ramia. (From Axsmith et al., 2000.)

CHAPTER 15

mesozoic seed ferns

635

Figure 15.40 Several Umkomasia cupules (arrows) attached to branch (Triassic). Bar  1 mm.

2002). Ovules of U. resinosa are small, possess a single integument, and are characterized by a bifid micropylar extension at the distal end. Cupules preserved as compressions and impressions often appear urn-like in morphology with two prominent lobes. The wrinkled nature of the cupule surface has suggested to some that these were fleshy structures. On well-preserved specimens stomata are present on both surfaces of the cupule. The seeds of U. macleanii are small (5 mm long) and borne either singly or in pairs in each cupule. Compressed cupules have also been found attached to short shoots that in turn arise from longer branches bearing D. odontopteroides leaves (FIG. 15.43) (Axsmith et al., 2000). Cupules of U. uniramia are borne in whorls of five to eight. It is important to factor in all aspects when considering the growth and development of fossil plants. The cupules attached to short shoots (FIG. 15.44) and leaves attached to long shoots from the Late Triassic of Antarctica have perplexed some who challenge whether long shoots would retain leaves (Artabe and Brea, 2003; Holmes and Anderson, 2005; Anderson et al., 2008). Modern Ginkgo, however, bears leaves on relatively large stems (long shoots) as well as

Figure 15.41 Suggested reconstruction of Umkomasia resi-

nosa cupulate branch. (From Klavins et al., 2002.)

on short shoots (Axsmith et al., 2007). Of equal importance is the fact that this Dicroidium plant grew at high polar latitudes and that the parameters of growth were no doubt much different than they are in temperate regions today (Axsmith et al., 2007). The recent report of Umkomasia cupules (FIGS. 15.45, 15.46) in association with foliage of the Thinnfeldia type from the Triassic of China (Zan et al., 2008) not only further expands the diversity within the corystosperms but also underscores the existence of the group in the Northern Hemisphere. Several additional cupulate structures are often included with the corystosperms. These include Spermatocodon (Thomas, 1933), Pilophorosperma (FIG. 15.47) (Thomas, 1933), and Karibacarpon (Lacey, 1976). Based on the number

636

paleobotany: the biology and evolution of fossil plants

C CS

S

Figure 15.42 Cross section of Umkomasia resinosa cupule

showing seed and cupule stalk (CS). Note distinct secretory cavities (arrows) (Triassic). Bar  1 mm. (From Klavins et al., 2002.)

Figure 15.44 Short shoot (S) of a corystosperm bearing sev-

eral pedicles terminating in cupules (C) of Umkomasia uniramia (Triassic). Bar  2 cm.

S

C C

Figure 15.43 Dicroidium axis showing leaf attached to stem

and short shoot (S) (Triassic). Bar  2 cm.

of species described, Pilophorosperma is perhaps the most common seed-producing member of this group (Thomas, 1933). Some cupules are less hemispherical in outline than are those in Umkomasia. Specimens of P. granulatum have three cupules at the end of each branch, whereas two are common in P. gracile. On the outer surface of the cupule are short papillae, and the inner surface is covered with long, pointed hairs. In P. crassum the cupules are borne in an overlapping fashion along the axis so that the fertile branch appears compact. Holmes (1987) placed Pilophorosperma and another corystosperm cupule, Karibacarpon (Lacey, 1976), in the

C

Figure 15.45 Axis of Umkomasia asiatica showing several

cupules (C) (Triassic). Bar  2 mm. (Courtesy B. J. Axsmith.)

genus Umkomasia. On the basis of consistent co-occurrence in the Middle Triassic of eastern Australia, the author suggested that Umkomasia feistmantelii was the cupulate organ of the plant that produced Dicroidium zuberi foliage.

CHAPTER 15

mesozoic seed ferns

637

Figure 15.47 Suggested reconstruction of Pilophorosperma geminatum. (From Taylor and Taylor, 1993.)

Figure 15.46 Suggested reconstruction of Umkomasia asiatica.

(From Zan et al., 2008.)

Seeds of the corystosperms possessed a slightly curved, bifid micropylar canal, which often extended beyond the cupule margin (Thomas, 1933). Macerations suggest that the integument contained no fibrous cells and the seed epidermis was covered with a thick cuticle. Embedded in the pollen chambers of several specimens are numerous saccate pollen grains similar to those found in the pollen-producing organ Pteruchus (Townrow, 1962b).

Petriellales This order has been defined based on a single genus of permineralized cupule containing seeds from the Middle Triassic

of Antarctica (T. Taylor et al., 1994a). Cupules in Petriellaea are bilateral, elongate, and borne in clusters on a dichotomizing axis (FIG. 15.48); they are 3 mm wide and constricted at the base to form a short pedicle. A vascular bundle extends along the midrib of the cupule with a trace given off to each ovule. The wall is thin but there is no evidence of a preformed suture. Each cupule contains from two to six small (1 mm in diameter), triangular ovules (FIG. 15.49) that are attached to the adaxial surface of the cupule surface at differing levels. Ovules are orthotropous with the integument of the seeds thickened only in the corners (FIG. 15.50). The nucellus and integument of P. triangulata are fused and the distal end of the seed extended into a short micropylar tube. Morphologically there is some similarity between the cupules of Petriellaea and those of Caytonia. Although the structure of the Caytonia cupules remains unknown, in Petriellaea the cupule organization suggests that the cupule has evolved by a proximal–distal folding of the megasporophyll (FIG. 15.51), rather than by enrolling or folding

638

paleobotany: the biology and evolution of fossil plants

Figure 15.50 Cross section of Petriellaea triangulata seed showing prominent corners of thick-walled cells (Triassic). Bar  250 μm.

Figure 15.48 Suggested reconstruction of Petriellaea triangulata. (From Taylor et al., 1994a.)

Figure 15.51 Diagrammatic cut away of Petriellaea triangulata

cupule showing seed attachment. (From Taylor et al., 1994a.)

Figure 15.49 Cross section of Petriellaea triangulata cupule

showing three seeds (Triassic). Bar  1 mm.

laterally along the midrib. To date nothing is known about the vegetative parts of the plant that produced these cupules. The Kannaskoppiaceae, a family within the order Petriellales, has been erected for compressed foliage of

Kannaskoppifolia (FIG. 15.52), pollen organs of Kannaskoppianthus, and ovulate structures of Kannaskoppia from the Upper Triassic of South Africa (Anderson and Anderson, 2003). Kannaskoppifolia leaves are cuneate to flabellate and lack a petiole (FIG. 15.52); the lamina is entire to deeply divided into segments. The venation is anastomosing. Similar leaves from the Upper Triassic of Argentina and Chile have been described as Rochipteris by Herbst et al. (2001).

CHAPTER 15

mesozoic seed ferns

639

Figure 15.53 Branches bearing numerous Kannaskoppia vincularis cupules (Triassic). Bar  2 cm. (Courtesy J. M. Anderson.) Figure 15.52 Leaf of Kannaskoppifolia sp. (Triassic). Bar 

2 cm. (Courtesy J. M. Anderson.)

Kannaskoppifolia leaves are known detached and in organic connection to axes. Moreover, several specimens show leafy twigs with Kannaskoppia strobili in organic connection. The strobili are up to 2 cm long, and are attached in groups of two or three in the axils of the leaves. They consist of a short proximal axis segment which then forks to form two secondary axes, each of which bearing two rows of megasporophylls. Each megasporophyll bears a single cupulate ovule (FIG. 15.53); the cupule splits into three lobes upon maturity of the seed. The male strobilus Kannaskoppianthus is between 8 and 45 mm long, and comprised of a forked axis bearing two rows of microsporophylls, each with five recurved, longitudinal pollen sacs in a distal concavity protected by an operculum.

Peltaspermales The Peltaspermales were initially described from the Upper Triassic of Greenland and South Africa based on associated foliage, pollen, and seed-bearing parts (Thomas, 1933; Harris,

1937). Today the order encompasses a far wider geographic region, including North America, Europe (Kerp, 1988; Kerp et al., 2001), and the Russian platform (Gomankov and Meyen, 1986; Naugolnykh and Kerp, 1996), and incorporates specimens ranging from Pennsylvanian (Kerp et al., 2001) to Triassic. Foliage

There are numerous late Paleozoic and Mesozoic foliage types assigned to the Peltaspermales with varying degrees of confidence, and it is impossible to address them all in detail here. The principal foliage type of the Mesozoic peltasperms is Lepidopteris, a generally bipinnate frond with pinnae attached suboppositely to alternately, and obliquely attached pinnules of the alethopterid type. The genus is known from both the Northern and Southern hemispheres and is common in Gondwana (McLoughlin et al., 1997; Gnaedinger and Herbst, 1998; Retallack, 2002). One characteristic of the genus is the occurrence of intercalary pinnules, or Zwischerfiedern, borne along the rachis between primary pinnae (Townrow, 1960). Pinnules are amphistomatic, and in some species epidermal papillae are present. On the rachis of L. strombergensis and other forms are small irregular blisters that are interpreted as subepidermal swellings. Karasev and Krassilov (2007), however, suggested that these structures represent reduced leaves of an initial scale-leaved shoot. Cuticular anatomy is

640

paleobotany: the biology and evolution of fossil plants

Figure 15.54 Suggested reconstruction of Permotheca showing helically arranged synangia. (From Krassilov et al., 1999.)

known for several species (Popa, 2000), and some ultrastructural features have been described (Labe and Barale, 1996). Lepidopteris callipteroides from the Triassic of Australia is associated with ovule-bearing peltate disks of Peltaspermum townrovii and peltasperm pollen clusters assigned to Permotheca (FIG. 15.54) (Retallack, 2002). Callipteris is a widely used historical morphogenus for late Paleozoic foliage attributed to the peltasperms. Callipteris fronds are relatively compact (FIG. 15.55), up to 80 cm long, and unforked (Barthel and Haubold, 1980; Kerp, 1988). Pinnules not only occur on the primary pinnae but also occur in an intercalary position, that is, directly attached to the rachis. Pinnules vary in size and shape, from tongue shaped to linear, and from profoundly lobed to entire margined; papillae are common on the epidermis (FIG. 15.56). In forms with tongue-shaped pinnules, such as C. conferta, the midvein is relatively well defined and persists for about three quarters of the pinnule length. Lateral veins depart the midvein at a high angle. They typically are slightly sinuous, the angle decreasing slightly immediately after departure from the midvein and inflecting upward again near the pinnule margin. Lateral

Figure 15.55 Sterile peltasperm frond Autunia ( Callipteris)

conferta (Permian). (From Taylor and Taylor, 1993.)

Figure 15.56 Autunia (Callipteris) conferta cuticle show-

ing papillae (arrow) on abaxial surface (Pennsylvanian–Permian). Bar  50 μm. (Courtesy H. Kerp.)

veins may fork between the midvein and margin. Kerp and Haubold (1988) divided Callipteris into a number of different genera, including Rhachiphyllum, Lodevia, Arnhardtia, Sphenocallipteris, and Dichophyllum (FIG. 15.57) based on pinnule morphology and epidermal anatomy.

CHAPTER 15

mesozoic seed ferns

641

Figure 15.57 Dichophyllum flabellifera sterile frond (Permian).

Bar  1 cm. (Courtesy M. Barthel.)

Supaia is a putative peltasperm foliage type recorded from China (Z.-Q. Wang, 1997), Europe (Gand et al., 1997), and North America (White, 1929; DiMichele et al., 2005a). The genus has been referred to the peltasperms by S.-Q. Wang (1997) due to associated peltaspermaceous reproductive organs in Late Permian rocks of China. The fronds are bipinnate (FIG. 15.58) like those of Dicroidium, but with the outer pinnules larger than those between the dichotomized rachides. Specimens of Supaia from the Permian of North America are associated with ovulate organs of the Peltaspermum type (DiMichele et al., 2007b). Another foliage genus with possible affinities in the Peltaspermales is Glenopteris, known almost exclusively from the Permian of Kansas (Sellards, 1900). Glenopteris fronds are unforked (FIG. 15.59), with pinnules attached directly to the rachis. Pinnules increase in length from the base of the frond to the center, from which point their length decreases. Pinnules are relatively wide and have midveins that persist to near the apex, with high-angle, lateral venation that is not “S” shaped.

Figure 15.58 Supaia sp. (Permian). Bar  5 cm. (Courtesy D.

S. Chaney and W. A. DiMichele.)

A few rachial veins enter the pinnule in a basal, basiscopic auricle. Venation is generally difficult to recognize because the specimens have a thick cuticle and because the laminae appear to have been succulent (pachymorphous) and xeromorphic. Along with Auritifolia (FIG. 15.60) (DiMichele et al., 2008), Glenopteris is the only putative peltasperm foliage taxon from

642

paleobotany: the biology and evolution of fossil plants

Figure 15.60 Auritifolia waggoneri leaf (Permian). Bar 

5 cm. (Courtesy D. S. Chaney and W. A. DiMichele.)

Figure 15.59 Suggested reconstruction of Glenopteris splend-

ens frond (Pennsylvanian). (From Krings et al., 2005a.)

North America that has yielded well-preserved cuticles. Fronds of G. splendens are heavily cutinized, amphistomatic, and glabrous (Krings et al., 2005b). Leaves in the foliage morphogenus Comia (FIG. 15.61) are known from the Permian of China, Russia, and North America (Mamay et al., 2008). They consist of unforked fronds with pinnately arranged pinnules borne directly on the main rachis. The distinguishing characteristic of Comia is its distinct venation pattern. Pinnules are decurrent with rachial

veins directly entering the basiscopic (decurrent) portion of the pinnule. The midvein is strongly marked and extends nearly the length of the pinnule. Lateral veins are grouped into fascicles, originating as a single vein. The originating vein bifurcates and all derivative veins, which may branch further, sweep upward concavely, relative to the pinnule midvein, and reach the pinnule margin. Between successive fascicles are several veins that depart directly from the midvein and either branch sparsely or remain unbranched, also reaching the pinnule margin. Tatarina is a peltasperm foliage type that demonstrates a variety of morphological forms from palmate to pinnate and is a common floral element of the Russian platform (Meyen, 1982a). In T. lobata the margin is lobed and the stomata are of the monocyclic type. Permophyllocladus is used for vegetative compression remains with possible peltaspermalean affinities from the Upper Permian (Tatarian) of the Vladimir Region (Russia). They consist of dorsiventrally flattened branching shoots with variously connate scale leaves; the shoots are gradually transformed into imparipinnate phylloclades (FIG. 15.62) by

CHAPTER 15

mesozoic seed ferns

643

Figure 15.61 Comia sp. (Permian). Bar  5 cm. (Courtesy D.

S. Chaney and W. A. DiMichele.)

extensive leaf cohesion (Karasev and Krassilov, 2007). Leaves are typically distinct on the abaxial side, but reduced to coalesced leaf cushions or marked by suture lines alone on the adaxial side. The phylloclades are amphistomatic and have massive cuticles; guard cells are sunken. Additional foliage morphogenera referred to the peltasperms are listed in E. Taylor et al. (2006) and Dobruskina (1975) with regard to the principally Angaran Triassic foliage genera Scytophyllum, Madygenia, Vittaephyllum, Madygenopteris, and Paratatarina. A more in-depth survey of the late Paleozoic putative peltaspermalean foliage morphotaxa recorded for North America can be found in DiMichele et al. (2005a). Reproductive Organs and whole-plant Concepts

There is not a great deal of detailed information known about the pollen organs of the peltasperms. In the more

Figure 15.62 Permophyllocladus polymorphus phylloclade

(Permian). Bar  1 cm. (Courtesy E. Karasev.)

primitive forms, it appears that the pollen organs are pinnately organized; fertile pinnae may be reduced but are essentially of the same architecture as the sterile pinnae. Perhaps the best known form is Callipterianthus arnhardtii (FIG. 15.63) from the Rotliegend (Pennsylvanian–Early Permian) of Germany (Roselt, 1962). This fossil displays frond-like organization in which pollen sacs are attached to reduced pinnules in the distal part of the frond. Sterile pinnae occur

644

paleobotany: the biology and evolution of fossil plants

15.63 Pollen organ Callipterianthus arnhardtii (Pennsylvanian–Permian). Bar  1.2 cm. (Courtesy H. Kerp.)

Figure

Figure 15.65 Frond of Autunia (Callipteris) conferta (Pennsylvanian–Permian). Bar  4 cm. (Courtesy H. Kerp.)

Figure 15.64 Suggested reconstruction of Antevsia zeilleri

pollen organ. (From Crane, 1985a.)

in the proximal portion of the frond, and intercalary pinnules are present on the rachis. Although the specimen is partly sterile, a correlation with sterile foliage is difficult. Callipterianthus arnhardtii has been suggested to represent the pollen-producing organ of either Autunia (Callipteris) naumannii or Arnhardtia (Callipteris) scheibei (see Kerp,

1996). Barthel (2006a), however, assigns pollen organs of the Pterispermostrobus type, which is composed of elongate pollen sacs attached to a peltate microsporophyll, to Arnhardtia scheibei. Antevsia includes branched axes which bear lateral groups of 4–12 elongate pollen sacs at their distal tips. Townrow (1960) described Antevsia as pinnate, with the main axis bearing alternate laterals in one plane and secondary branches produced irregularly (FIG. 15.64). Although these branches are described as microsporophylls, the branches are not expanded at their tips. Pollen sacs extend up to 5 mm in length

CHAPTER 15

mesozoic seed ferns

645

Figure 15.66 Detail of Autunia (Callipteris) conferta frond

showing venation (Pennsylvanian–Permian). Bar  2 cm. (Courtesy H. Kerp.)

and dehisce longitudinally. Like the majority of the Mesozoic Gondwanan peltasperms, pollen grains are small (23–40 μm) and oval, with a distal sulcus; if dispersed, these grains would be assigned to Cycadopites. The ovulate organs of the peltasperms are known in far greater detail than the pollen organs, and it has been shown that there is considerable morphological variability among them. Moreover, there is compelling evidence for a number of forms linking ovulate structures to sterile foliage. Based on repeated co-occurrences of material from the central European Rotliegend (Pennsylvanian–Early Permian), the well-known foliage type Callipteris conferta (FIGS. 15.65, 15.66) and the reproductive organ Autunia (FIG. 15.67) were reconstructed into Autunia conferta (Kerp, 1982, 1988). Autunia was originally introduced for strobili with fan shaped, suboppositely to alternately arranged megasporophylls (FIGS. 15.68, 15.69) bearing one or two ovules of the Rhabdocarpus type on the lower surface. Pollen-producing fructifications believed to belong to A. conferta consist of five to nine pollen sacs positioned on a peltate microsporophyll of the Pterispermostrobus type (Barthel, 2006a). Pollen sacs contain Vesicaspora-type bisaccate pollen up to 50 μm in diameter. Whole-plant concepts similar to that advanced for A. conferta have been proposed for several other Early Permian taxa, including Autunia (Callipteris) naumannii

Figure 15.67 Suggested reconstruction of Autunia conferta ovuliferous organ. (From Kerp, 1988.)

Figure 15.68 Suggested reconstruction of two Autunia conferta megasporophylls. Arrow indicates ovule. (From Kerp, 1988.)

and Arnhardtia (Callipteris) scheibei (Kerp, 1988; Kerp and Haubold, 1988; Barthel, 2001, 2006a). Stems assigned to A. scheibei, originally described as Kontheria striata, reveal close spacing of fronds in a helical 3/11 phyllotaxis (Barthel,

646

paleobotany: the biology and evolution of fossil plants

Figure 15.69 Ovulate organ Autunia milleryensis (Pennsylvanian–

Permian). Bar  1 cm. (Courtesy H. Kerp.)

2006b). Rhachiphyllum schenkii (FIGS. 15.70, 15.71) is a foliage type that is very similar in many respects to Callipteris ( Autunia) conferta, and Barthel (2006a) suggested that R. schenkii plants resembled A. conferta. Mamay (1975) described the enigmatic reproductive organ Sandrewia from Lower Permian rocks of north-central Texas, suggesting vojnovskyalean affinities. The morphology of Sandrewia (FIGS. 15.72, 15.73) is essentially the same as that of Autunia (FIG. 15.74); however, Mamay documented that the fan-shaped, lateral sporangium or ovule-bearing structures were directed upward on their stalks, whereas Kerp (1988) has shown the opposite positioning of the lateral structures in Autunia. Peltaspermopsis polyspermis is a natural-genus concept used for a Late Permian peltasperm from Russia. The plant includes vegetative stems, some with conspicuous nodes (FIG. 15.75), which are interpreted as showing evidence of seasonal growth interruptions. Pursongia-like lanceolate leaves are also considered to be part of the same plant, along

Figure 15.70 Suggested reconstruction of a frond of Rhachiphyllum schenkii. (From Kerp, 1988.)

CHAPTER 15

mesozoic seed ferns

647

Figure 15.71 Rachis and pinnae of Rhachiphyllum schenkii

showing pinnule venation (Permian). Bar  1 cm. (Courtesy H. Kerp.)

with reproductive organs in the form of seed-bearing disks and their “racemose” aggregations (Naugolnykh, 2001a). Peltaspermum includes peltaspermalean ovulate organs (Harris, 1937) that consist of elliptical to radial, umbrella-like disks, which are sometimes termed peltoids (Naugolnykh, 2001a). The disks or megasporophylls are up to 1 cm in diameter and are helically arranged into simple linear-cylindrical strobili; ovules are borne in a ring on the lower surface of the megasporophylls (FIGS. 15.76, 15.77) (Townrow, 1960). The number of ovules per disk is variable (FIGS. 15.78, 15.79). Based on association, a concept which was expanded by Meyen (1987) and termed assemblage genera, Poort and Kerp (1990) broadened the definition of the genus Peltaspermum (FIG. 15.77) to what they term a natural genus. This reconstructed taxon includes foliage of the Lepidopteris type (FIG. 15.80), the pollen organ Antevsia zeilleri (FIG. 15.81), and the seed-bearing organ Peltaspermum rotula. From a site in Morocco, however, Peltaspermum reproductive organs have been described that display certain features,

Figure 15.72 Sandrewia texana axis (Permian). Bar  1 cm. (Courtesy D. S. Chaney and W. A. DiMichele.)

that is, resinous bodies, in common with co-occurring Rhachiphyllum-type foliage (FIGS. 15.70, 15.71) (Kerp et al., 2001). The genus Meyenopteris is also thought to be a natural genus of peltasperms, using the concept of Poort and Kerp (1990). Lepidopteris matalensis foliage is typically bipinnate, with small blister-like swellings on the axes. The ovulebearing organs of Peltaspermum thomasii bear two ovules (3 mm long) on the lower surface of each peltate megasporophyll (FIG. 15.76). Lopadangium is used for peltaspermaceous ovulate organs that cannot be correlated with foliage or other organs and therefore do not represent a natural genus.

648

paleobotany: the biology and evolution of fossil plants

Figure 15.74 Autunia conferta ovulate organ showing the Sandrewia texana aspect (Pennsylvanian–Permian). Bar  1.8 cm. (Courtesy H. Kerp.)

Figure 15.73 Reproductive axis of Sandrewia texana showing fan-shaped sporophylls. Bar  1 cm. (Courtesy D. S. Chaney and W. A. DiMichele.)

Conclusions As noted above, the so-called Mesozoic seed ferns do not represent a natural group, and whether or not each of the four orders discussed here represents a monophyletic group is also not clear. No complete phylogeny has been done for any of these plants. In almost all phylogenetic analyses of seed plants to date, each order has been treated as a clade and is often represented as a single, composite terminal, even when reconstructions have only been based on association at the same locality.

Figure 15.75 Suggested reconstruction of Peltaspermopsis polyspermis. (From Naugolnykh, 2001a.)

CHAPTER 15

mesozoic seed ferns

649

Figure 15.79 Cuticle of Peltaspermum martinsii peltate head showing region of stalk attachment (arrow) (Permian). Bar  2 mm. (Courtesy H. Kerp.)

Figure 15.76 Suggested reconstruction of Peltaspermum tho-

masii. (From Taylor and Taylor, 1993.)

Figure 15.77 Peltaspermum martinsii peltate head showing scalloped margin (Permian). Bar  4.5 mm. (Courtesy H. Kerp.) Figure 15.80 Frond of Peltaspermum (Lepidopteris) martinsii (Permian). Bar  6 mm. (Courtesy H. Kerp.)

Figure 15.78 Suggested reconstruction of Peltaspermum rotula showing several ovules. (From Crane, 1985a.)

In the corystosperms, which may be the best-known group overall, the remarkable uniformity of both pollen and ovulate organs throughout Gondwana suggests that at least the Southern Hemisphere representatives may represent a single clade. Artabe and Brea (2003) concluded that the group was not a natural one; however, their analysis was based primarily on stem anatomy, which may vary with habitat. Although the Dicroidium plants reconstructed from South America and from Antarctica have

650

paleobotany: the biology and evolution of fossil plants

15.81 Suggested reconstruction of Lepidopteris frond with pollen organs of the Antevsia type at the tip. (Courtesy M. Zavada.) Figure

very different habits, their reproductive organs and foliage are essentially the same. With the exception of the Petriellales, the other two orders are known from compression–impression specimens only. As a result, we know a great deal about the morphology and epidermal anatomy of the Caytoniales and Peltaspermales, but no

internal anatomy, which would no doubt be useful in pinpointing homologies with other seed plant groups. The Caytoniales are perhaps the most poorly known of these younger Mesozoic seed ferns. Although the specimens from the Jurassic of Yorkshire and Poland have been studied in detail, Sagenopteris leaves have been reported from a variety of sites around the world, with no associated caytonialean reproductive organs. At the present time, the Peltaspermales are perhaps the most confounding of all of these plants. The morphology of the Autunia type of ovulate organ is very different from that of the Peltaspermum type known from Europe, Asia and Gondwana. Moreover, the pollen organs assigned to Autunia bear bisaccate grains of the Vesicaspora type, whereas those known from Antevsia are monosulcate and conform to Cycadopites. Significant advances in understanding all of these groups will come about when anatomically preserved specimens are found and when there are better reconstructions that either combine organs unique to a site or that link organs based on distinctive synapomorphies (E. Taylor et al., 2006).

16 LATE PALEOZOIC AND MESOZOIC FOLIAGE LATE PALEOZOIC FOLIAGE .....................................................652

Tinsleya .............................................................................................685

Adiantites ..........................................................................................655

Triphyllopteris, Genselia, and Charbeckia ........................................685

Alethopteris .......................................................................................656

MESOZOIC FOLIAGE ................................................................. 685

Aneimites ..........................................................................................657

Anomozamites ..................................................................................687

Aphlebia ............................................................................................658

Cladophlebis .....................................................................................687

Alloiopteris .......................................................................................658

Coniopteris ........................................................................................688

Botrychiopsis ....................................................................................659

Ctenis ................................................................................................689

Callipteridium ...................................................................................659

Deltolepis and Cycadolepis...............................................................689

Cardiopteridium ................................................................................660

Dictyophyllum...................................................................................689

Cardiopteris (Fryopsis) .....................................................................660

Dictyozamites....................................................................................689

Charliea .............................................................................................660

Doratophyllum ..................................................................................690

Cyclopteris ........................................................................................661

Macrotaeniopteris .............................................................................690

Dicksoniites.......................................................................................662

Matonidium .......................................................................................690

Discopteris ........................................................................................664

Mesodescolea ....................................................................................690

Eremopteris .......................................................................................664

Nilssonia ...........................................................................................690

Ginkgophytopsis ...............................................................................664

Nilssoniopteris ..................................................................................691

Kankakeea .........................................................................................665

Otozamites ........................................................................................693

Karinopteris, Mariopteris, and Pseudomariopteris ...........................665

Pachypteris, Komlopteris, and Thinnfeldia .......................................695

Lesleya ..............................................................................................669

Phlebopteris.......................................................................................696

Linopteris, Reticulopteris, and Barthelopteris ..................................669

Pseudoctenis......................................................................................696

Lobatopteris ......................................................................................671

Pseudocycas ......................................................................................697

Lonchopteridium and Lonchopteris ..................................................672

Pterophyllum .....................................................................................697

Megalopteris......................................................................................672

Ptilophyllum ......................................................................................698

Neuropteris sensu lato.......................................................................673

Ptilozamites .......................................................................................699

Nothorhacopteris ...............................................................................677

Ruflorinia ..........................................................................................699

Odontopteris and Lescuropteris ........................................................677

Taeniozamites....................................................................................700

Pecopteris ..........................................................................................679

Ticoa..................................................................................................700

Rhodea (Rhodeopteridium) ...........................................................680

Wingatea ...........................................................................................700

Sphenopteris......................................................................................680

Yabeiella ............................................................................................700

Spiropteris .........................................................................................683

Zamites..............................................................................................701

Taeniopteris .......................................................................................683

651

652

Paleobotany: the biology and evolution of fossil plants

Not knowing The name of the tree, I stood in the flood Of its sweet smell Basho

The most widespread and conspicuous plant fossils are impressions and compressions of leaves and fronds. In this chapter, some of the common groups and types of Paleozoic and Mesozoic foliage will be described and illustrated. Also included is consideration of permineralized specimens, where available, and other parts of the parent plant or the natural affinities of a particular foliage type, if known. The foliage of some plants is discussed in the chapter in which the parent plant is described (e.g., Glossopteris and Dicroidium).

LATE PALEOZOIC FOLIAGE The Carboniferous was historically referred to as the Age of Ferns, due to the diverse types of fernlike foliage that were found in abundance in rocks of this age. The designation Age of Ferns, however, was applied before it was known that some of the foliage was produced by certain types of gymnosperms, that is seed ferns (pteridosperms). Historically, all fernlike foliage was included under the term pteridophylls because the natural affinities were unknown. There have been several important attempts to classify fern and fernlike foliage types in an artificial system of form genera (today called morphogenera) based on the complements of morphological features such as size, shape, venation pattern, and the attachment of the ultimate foliar segments to the axis. Many of the morphogenera and species used today were initially delimited by noted paleobotanists such as Lindley and Hutton (1833–1835), Sternberg (1820–1838), Brongniart (1828–1837) (FIG. 16.2), Göppert (1836) (FIG. 16.1), Stur (1875, 1877), Grand’Eury (1877), Zeiller (1886, 1888a, b, 1906) (FIG. 16.3), Kidston (1892a, 1914, 1923–1925), Gothan (1913) (FIG. 16.4), Corsin and Dubois (1933), Jongmans (1960), and Potonié (1903– 1913) (FIG. 16.5). In North America, much of the early work on Carboniferous impression–compression foliage can be attributed to the detailed studies of Lesquereux (1879, 1880, 1884) (FIG. 16.6), White (1900, 1937a, b, 1943) (FIG. 16.7), and Noé (1925) (FIG. 16.8). Foliage morphotaxa have become widely established in paleobotanical literature, but they almost always consist of sterile specimens. In cases where reproductive parts are found attached to foliage, some authors have adopted the

Figure 16.1 Heinrich R. Göppert.

Figure 16.2 Adolphe Brongniart.

CHAPTER 16

Figure 16.3 Charles René Zeiller.

LATE PALEOZOIC AND MESOZOIC FOLIAGE

Figure 16.5

Henri Potonié. (Courtesy S. Schultka.)

Figure 16.6

Figure 16.4 Walther Gothan. (Courtesy S. Schultka.)

practice of assigning the generic name of the fructification in place of the name of the sterile foliage. In other instances, sterile foliage has been found attached to stems and, as a result, it has been possible to reconstruct the growth habit of

653

Leo Lesquereux.

the plant. In these cases, usually the name given for the foliage is applied to the entire plant. It is important to understand the biological implications of dealing with late Paleozoic foliage types. In several instances, a single morphogenus may represent the foliage of widely disparate plants, as in the case of Sphenopteris, which has been identified as the foliage of several different seed ferns as well as true ferns. Nowhere is the species problem more complex in paleobotany than in dealing with fern or fernlike foliage types, where many

654

Paleobotany: the biology and evolution of fossil plants

Figure 16.9

Figure 16.7 Charles David White. (Courtesy Oklahoma

Geological Survey.)

Figure 16.8 Adolphe C. Noé. (Courtesy H. N. Andrews.)

of the so-called species undoubtedly represent the natural variation inherent in foliage from different parts of the same frond or perhaps from different stages of development. Despite these limitations, numerous late Paleozoic foliage

Xing-Xue Li.

types have been used successfully as biostratigraphic markers to subdivide portions of the Carboniferous (biozonation). In some instances, stratigraphic levels can be recognized by a single index species, whereas in other cases an assemblage of species has proven more useful in identifying particular time stratigraphic units (assemblage zone). For example, several well-defined floral zones (three in the Mississippian, eight in the Pennsylvanian, and three in the Permian) have been designated in the late Paleozoic of the US by Read and Mamay (1964), and Gillespie and Pfefferkorn (1979) have provided detailed ranges for a number of taxa in the Pennsylvanian of North America. Wagner (1984) has successfully used megafloral zonation in the Pennsylvanian of the Equatorial Belt, and Li and Zhang (1983) (FIG. 16.9) have used megafloras in association with other fossil assemblages to correlate the Pennsylvanian of China. More recently, several approaches have been used in order to refine the traditional classification system for impression– compression foliage. In general, these approaches add a variety of criteria that are regarded as more reliable diagnostically (biological criteria) to basic characters of pinna and pinnule morphology on which genus and species discrimination is traditionally based. One such approach uses frond architecture in suggesting natural relationships (Laveine, 1967, 1997, 2005; Laveine et al., 1993a, 1998, 2005 Cleal et al., 1996; Laveine and DiMichele, 2001). This approach, however, is sometimes hampered by our currently limited knowledge about the range of intraspecific variability in the architecture of late Paleozoic fronds (Krings et al., 2006).

CHAPTER 16

Figure 16.10

Manfred Barthel. (Courtesy S. Schultka.)

A second approach focuses on the anatomy of the epidermis preserved in the overlying cuticle, that is, cuticular analysis (Chapter 1). Although long known as a tool in paleobotanical research, cuticular analysis was rarely used in the analysis of late Paleozoic fern and fernlike foliage until Barthel (1961, 1962a) (FIG. 16.10) published a comprehensive study on the cuticles of late Paleozoic foliage from central Europe. This work represents a starting point for the wider recognition and appreciation of cuticular analysis in the evaluation of late Paleozoic compression foliage fossils. Today, cuticles provide a wealth of information that is useful in the taxonomy of fossil foliage, but they are also useful in paleobiological and paleoecological applications (Kerp, 1990; Kerp and Barthel, 1993; McElwain and Chaloner, 1996; Uhl, 2006). In some instances, cuticular analysis has been successfully used in reassembling whole plant taxa from isolated parts such as foliage, stems, and reproductive structures. It is interesting to note that, in general, fern foliage typically has delicate cuticles that are rarely preserved, whereas seed fern cuticles are usually more robust, and thus more often represented in the fossil record. As a result, preservation of cuticles alone may, in some instances, provide a clue as to whether a leaf type belongs to the ferns or seed ferns.

LATE PALEOZOIC AND MESOZOIC FOLIAGE

655

The basic organization or architectural plan (Bauplan) of fossil fern and seed fern leaves, or fronds, is similar in many instances to the leaf of a living fern. The frond is attached to the primary shoot axis (creeping rhizome or erect stem) by the stipe or petiole; in some instances, the stipe was massive and its proximal portion (base) was important in maintaining the mechanical stability of the stem, for example in some medullosan seed ferns (see Chapter 14). Branching of the frond assumed a variety of forms; the main axis of the leafy portion (blade) of the frond is termed the rachis. Many seed fern fronds are characterized by a bifurcating stipe that results in the formation of a two-part blade with two rachides (bipartite frond). First-order subdivisions of the rachis or rachides are termed pinnae, and these may be opposite, subopposite, or alternately attached to the rachis. In modern ferns, pinnae generally are all borne in the same plane (biseriate), but in a few late Paleozoic ferns or fernlike plants (e.g., Zygopteridales; see Chapter 11), pinnae were borne at right angles to each other, a condition termed quadriseriate. The ultimate foliar segments of a frond are termed pinnules. If the main rachis lacks pinnae, only producing pinnules, the frond is said to be simply pinnate. When pinnules are borne on pinnae, the frond is bipinnate, tripinnate, or quadripinnate depending on the number of rachis subdivisions. Although many entire fern and seed fern fronds have been described, most late Paleozoic foliage is represented by fragmentary specimens, mostly isolated pinnae and pinnules, and it is principally on pinnule morphology that the various morphogenera have been delimited. Following are descriptions of some of the most commonly encountered late Paleozoic foliage morphogenera; in some instances, information on the epidermal anatomy, along with ideas about the frond architecture, growth habit, and affinities of the plants that produced these foliage types, are included. Descriptions of additional morphogenera can be found in Remy and Remy (1977). ADIANTITES

Pinnule morphology (FIG. 16.11) has been the principal feature used in the identification of Adiantites. Fronds of Adiantites may exceed 30.0 cm in length and are highly variable in branching. Specimens of A. antiquus (FIG. 16.12) from the Mississippian are at least three times compound and consist of wedge-shaped pinnule segments, each with a bluntly rounded apex (Jennings, 1985). Permineralized axes contain a C-shaped xylem strand with elliptical bordered pits on the tracheids. Anatomical evidence suggests that at least some specimens of the genus represent the vegetative parts of seed plants of the calamopityean type.

656

Paleobotany: the biology and evolution of fossil plants

Figure 16.13 Alethopteris grandinii (Pennsylvanian). Bar  2 cm.

Figure 16.11 Adiantites machanekii (Mississippian). Bar  1 cm.

(Courtesy GBA.)

Figure 16.12 Adiantites antiquus (Mississippian). Bar  1 cm.

(Courtesy GBA.)

ALETHOPTERIS

Pinnules of Alethopteris are variable in shape, roughly triangular, tongue-shaped or elongate, distally rounded or pointed, usually entire-margined or slightly lobed, and often quite large, with some species ranging up to 5.0 cm long (Wagner, 1968). They are borne at an acute or wide angle on the pinna axis (FIG. 16.13–16.15), with the base characteristically decurrent (flared) on the lower side. Each pinnule is vascularized by a single trace that generally forks to produce steeply arched secondary veins (subsidiary veins) (FIG. 16.14). In some species, the secondary veins arise from the vascular bundle of the pinna axis and directly enter the decurrent base. A relatively large number of species have been studied by means of cuticular analysis (Barthel, 1962; Kerp and Barthel, 1993; Šimuº nek, 1996a; Zodrow and Cleal, 1998). Alethopteris pinnules were hypostomatic and most were characterized by relatively large, uniseriate, multicellular trichomes. In some species, stomatal pores were partially covered by overarching papillae extending from the subsidiary cells. The stomatal apparatus of A. sullivantii is described by Reihman and Schabilion (1985) as actinocytic–haplocheilic; however, Stidd (1988) considered the stomatal organization to be different. He reconstructs each stomatal complex as consisting of four cells in parallel alignment but lacking flanking subsidiary cells, as in paracytic stomata. The stoma is formed by walls of the larger outer cells, and it is not known what role the smaller interior cells played in stomatal function.

CHAPTER 16

LATE PALEOZOIC AND MESOZOIC FOLIAGE

657

Figure 16.14 Two Alethopteris pinnules. (From Taylor and

Taylor, 1993.)

Permineralized alethopterid pinnules show a prominent midvein from which secondary veins emerge at sharp angles. In transverse section, the pinnules are often revolute (turned under along the margin), and the midrib contains one to four vascular strands. The vascular bundle is surrounded by a parenchymatous bundle sheath, including bundle sheath extensions. Other anatomical features of alethopterid pinnules include a hypodermis consisting of large, nearly square cells below the adaxial epidermis, and a mesophyll composed of palisade and spongy parenchyma (Franks, 1963; Baxter and Willhite, 1969; Reihman and Schabilion, 1976a, b; Mickle and Rothwell, 1982). Alethopteris fronds were probably bipartite, tri- or quadripinnate, and attained considerable size (Laveine et al., 1993a). Laveine (1986) reported on a specimen observed in a coal mine in northern France that was 7.4 m long and 2 m wide. All species of Alethopteris are believed to have been borne by medullosan pteridosperms, and there is at least one report of anatomically preserved Alethopteris pinnules from the Upper Pennsylvanian of the Appalachian Basin organically attached to axes exhibiting Myeloxylon petiole anatomy. This morphogenus is known to have been borne on a Medullosa noei stem (Pryor, 1989). Based on material from the Sydney Coalfield (Nova Scotia, Canada), Zodrow (2007) provided a tentative reconstruction of the A. zeilleri

Figure 16.15 Alethopteris pinna with pinnules (Pennsylvanian).

Bar  2 cm.

plant as a small tree, 5–7 m high, with relatively large fronds to which are attached Pachytesta incrassata ovules and Dolerotheca-type pollen organs (Chapter 15). ANEIMITES

Foliage of this type may be similar to Triphyllopteris and Genselia, common Mississippian morphogenera (see below). In Aneimites (FIG. 16.16), the pinnules are fairly large (up to 4 cm), stalked, pyriform or ovate, and typically divided into two or three lobes (White, 1943). There is no primary vein but rather a series of dichotomous veins that

658

Paleobotany: the biology and evolution of fossil plants

Figure 16.17 Aphlebia. (From Taylor and Taylor, 1993.)

Figure 16.16 Aneimites. (From Taylor and Taylor, 1993.)

extend into the lamina. Foliage of this type is often included with the seed ferns based on attached or associated reproductive structures (White, 1904). APHLEBIA

This genus is used for vascularized, anomalous foliar elements that are often borne at the base of petioles or at the point of pinna attachment in many ferns and fernlike plants. They may be entire-margined, fringed, (deeply) incised, lobed, or irregularly pinnately organized and range in size from a few centimeters to 20 cm long (FIG. 16.17). Aphlebia erdmannii from the Rotliegend (Late Pennsylvanian–Early Permian) of Germany differs from all other aphlebiae in the presence of densely spaced protuberances bearing spine- or scalelike appendages on the surface (Barthel, 2005). Similar protuberances bearing appendages have been observed on juvenile Scolecopteris oreopteridia fronds and a Caulopteris stem from the same locality, suggesting that A. erdmannii was produced by a marattialean fern (Marattiales). One specimen from the Mesozoic is believed to be related to the fern family Dipteridaceae (Boersma, 1985).

Figure 16.18 Alloiopteris sp. (From Taylor and Taylor, 1993.)

ALLOIOPTERIS

In this genus, the frond is at least three times pinnate, with the pinnae of the ultimate order bearing alternately arranged pinnules (FIG. 16.18). Attachment of pinnules to the axis is usually Sphenopteris- or, more rarely, Pecopteris-like (Josten, 1983). Pinnules are usually lobed, with lobes that

CHAPTER 16

are prominent and in some species have marginal teeth. Each pinnule is vascularized by a midvein that may divide several times to produce laterals that extend into the individual lobes and to the pinnule margin (Alvarez Ramis et al., 1979). In most species, the pinnules are strongly decurrent and fused at their bases. Similar foliage with sporangia on the abaxial surface is called Corynepteris (Chapter 11). A recent study has demonstrated that Alloiopteris pinnules were produced by plants with a Zygopteris type of anatomy. BOTRYCHIOPSIS

This Carboniferous–Permian foliage genus from Gondwana includes bipinnate fronds with broadly inserted pinnules in the proximal regions of the frond (Archangelsky and Cúneo, 1981; Jasper et al., 2003). Near the midlevel of the frond (FIG. 16.19), the pinnae are short with broadly attached, subcircular to elongate pinnules (Archangelsky and

LATE PALEOZOIC AND MESOZOIC FOLIAGE

659

Arrondo, 1971). Pinnules in the distal portion of the frond are shorter and lobed. Venation is of the open dichotomous type. Foliage of Botrychiopsis is believed to have been produced by pteridosperms, based on its association in strata of the Paganzo Basin in Argentina with pinnately organized pollen organs borne in clusters on ultimate branches (Artabe et al., 1987). Botrychiopsis foliage from the Permian of South Africa has been reported in association with pollen organs consisting of a central axis with spirally arranged microsporophylls bearing distal sporangia (Anderson and Anderson, 1985). Archangelsky (1996) suggested that the genus belongs to an ancestral stock of seed plants from which certain Mesozoic groups, perhaps the Corystospermales, evolved. The genus is biostratigraphically significant as it represents one of the character taxa of the Nothorhacopteris– Botrychiopsis–Ginkgophyllum Biozone in the Pennsylvanian of Argentina (Carrizo and Azcuy, 1995; Perez Loinaze and Cesari, 2003) and Brazil (Jasper et al., 2003). CALLIPTERIDIUM

Figure 16.19 Botrychiopsis plantiana (Permian). Bar  5 mm. (Courtesy I. Escapa.)

Specimens of this genus are similar to those of Alethopteris and, in some instances, have been confused with it. Pinnules of Callipteridium are typically smaller than Alethopteris and inserted at nearly right angles to the pinna midrib, with the base of the pinnule generally less decurrent than in Alethopteris (Leisman, 1960). A single vein enters the pinnule and becomes less distinct well below the pinnule apex, often forming subsidiary veins. Some secondary veins may enter the base of the pinnule directly from the rachis. The frond architecture of one species, C. pteridium, has been reconstructed based on material from the Stephanian of France and Germany. Wendel (1980) reconstructed the frond as bipartite and tripinnate, with circular to deeply incised cyclopterid elements on the stipe. Pinnae of the penultimate order positioned on the exterior side of the rachides are considerably larger than those on the interior side. Intercalary pinnae of the ultimate order are produced from both the interior and exterior side of the rachides. Mature fronds of Callipteridium may have been 3 m long (Laveine et al., 1977). Callipteridium-type foliage is believed to have been produced by seed ferns, based on the discovery of Emplectopteris-like ovules (Konnoa koraiensis; Asama, 1959) near the pinnule bases of C. koraiense. Ovulate reproductive structures associated with Callipteridium foliage have also been reported from Germany (Wendel, 1980). Because the flora that co-occurs with these ovules consists almost exclusively of C. pteridium foliage, it is reasonable to conclude that the ovules were produced by the C. pteridium plant.

660

Paleobotany: the biology and evolution of fossil plants

CARDIOPTERIDIUM

CHARLIEA

Foliage of this genus is known from the Mississippian and Lower Pennsylvanian of China, Europe, North America, and Russia (Gensel, 1988). Pinnules are small ( 2.0 cm), variously shaped (FIG. 16.20) (e.g., orbicular or oval, reniform or tongue-shaped, often asymmetrical), and borne on fronds that are at least bipinnate. Pinnules possess a constricted base as in Neuropteris and exhibit a pattern of radiating, dichotomizing veins. Cardiopteridium is believed to have been produced by pteridosperms, but Gensel (1988) suggested a progymnosperm affinity as well.

Cardiopteris (Fryopsis) foliage consists of pinnate fronds with axes characterized by nests of sclerotic cells in the cortex. Pinnules are round to tongue shaped and attached to the axis by a short, broad pedicle (FIG. 16.21). Venation is fan shaped (Remy and Remy, 1977).

Charliea is used for narrow axes to which are attached linear– oblong ultimate segments that are distally dissected into two to four nearly equal lobes (FIG. 16.22). The segments are characterized by parallel venation with a few dichotomies. The genus has been reported from the Lower Pennsylvanian of Utah and Upper Pennsylvanian of New Mexico. Charliea was originally interpreted as a pinnately compound leaf, perhaps with affinities in the cycadophytes (Mamay, 1990). Reevaluation of the type specimens and additional material, however, revealed that the fossils do not represent pinnate leaves but rather are plagiotropic long shoots with leaves arranged alternately in two rows (Tidwell and Ash, 2003). These authors suggested that Charliea may represent foliage branches produced by a stem that was either creeping or climbing, or both. Moreover, they note that the fossils bear some resemblance to the plagiotropic leaves of the noeggerathialeans (Chapter 12). The reproductive structures and anatomy of the plant that produced Charliea are unknown.

Figure 16.20 Cardiopteridium nanum. (From Taylor and

Figure 16.21 Cardiopteris (Fryopsis) frondosa (Mississippian).

CARDIOPTERIS (FRYOPSIS)

Taylor, 1993.)

Bar  2 cm. (Courtesy BSPG.)

CHAPTER 16

LATE PALEOZOIC AND MESOZOIC FOLIAGE

661

CYCLOPTERIS

Cyclopteris was originally established as a purely artificial morphogenus for isolated orbicular foliar elements (FIG. 16.23) (Brongniart, 1828) but is today widely used also for the morphologically distinct “pinnules” that occur in the proximal portions of some pteridosperm fronds, especially those of the neuropterid group (e.g., Laveinopteris, Neuropteris sensu stricto; Cleal and Shute, 2003; Laveine, 2005; Lyons and Laveine, 2005) (FIG. 16.24). These socalled pinnules (sometimes also termed cyclopterid leaves)

Figure 16.23 Cyclopteris probably from Laveinopteris tenui-

folia (Pennsylvanian). Bar  1.5 cm. (Courtesy J.-P. Laveine.)

Figure 16.22 Charliea manzanitana Bar  5 cm. (Courtesy S. H. Mamay.)

(Pennsylvanian). Figure 16.24 Jean-Pierre Laveine.

662

Paleobotany: the biology and evolution of fossil plants

are rounded (orbicular to reniform), entire margined, and up to 10 cm in diameter, with densely spaced veins radiating outward from the point of attachment in an open dichotomous pattern (Cleal and Shute, 1998). They do not represent true pinnules but rather are homologous to pinnae of the ultimate order (Laveine, 2005). Cleal and Shute (2003) proposed that the morphogenus Cyclopteris should be used only according to its original definition, that is, for isolated fossils that cannot be related to the foliage species that originally bore them. Concerning the function of cyclopterid leaves in Laveinopteris, Shute and Cleal (2002) hypothesized that juvenile Laveinopteris plants consisted of a monopodial sapling bearing only cyclopterid leaves, which were directly attached to the shoot axis, a pattern that optimized growth in shady conditions. On reaching the canopy level, the plant produced a crown of large, pinnate fronds with cyclopterid leaves borne only proximally. The interpretation of cyclopterids as leaves of a juvenile plant, however, has been challenged by Laveine (2005), who suggested that they represent protective leaflets (folioles) at an early stage of frond differentiation. Distinctly enlarged foliar elements that occur in the proximal portions of other seed fern fronds differ in morphology from typical Cyclopteris. These can be oval or elongate triangular in outline (e.g., in Blanzyopteris praedentata; Krings and Kerp, 1999), fringed, dissected or deeply incised (FIGS. 16.25, 16.26) (e.g., in Odontopteris brardii; Krings et al., 2006), or lobed (e.g., in Lescuropteris genuina; Zeiller, 1906). Some authors use the informal term cyclopteroid elements for these structures.

Figure 16.25 Cyclopteroid leaf of odontopterid seed fern

(Pennsylvanian). Bar  1 cm.

DICKSONIITES

Attachment of pinnules in Dicksoniites is similar to Pecopteris, that is, they are attached all along their base. Pinnules are triangular or elongate triangular and often faintly or profoundly lobed. The venation is marked and composed of a slightly flexuous midvein that gives off laterals; these may fork several times as they extend into the pinnule lobes and margin. Although several species have been described within Dicksoniites, only one form, D. pluckenetii, is known in sufficient detail and thus will be used here as a model for the characterization of the genus. Dicksoniites pluckenetii occurs in the Late Pennsylvanian– Early Permian of Europe and North America; the most complete specimens known to date come from the Stephanian of the Graissessac Basin in France (Galtier and Béthoux, 2002). Fronds of D. pluckenetii are bipartite (FIG. 16.27), more than 45 cm long, up to 40 cm wide, bipinnate, and characterized by a relatively long, naked petiole. In architecture, they resemble the fronds of Pseudomariopteris (see below).

Figure 16.26 Cyclopteroid leaf of odontopterid seed fern (Pennsylvanian). Bar  1 cm.

Pinna axes of the ultimate order bear pinnules, usually 1–2 cm long, that are triangular or elongate triangular in general outline and (extensively) lobed. The fronds are attached to flexuous stems that are about 1 cm in diameter; spacing of

CHAPTER 16

LATE PALEOZOIC AND MESOZOIC FOLIAGE

663

Figure 16.27 Dicksoniites pluckenetii, proximal portion of a bipartite frond showing the forking of the petiole (Pennsylvanian). Bar  2 cm. (Courtesy BSPG.)

Figure 16.29 Axial tendrils (arrow) of Dicksoniites pluckenetii (Permian). Bar  1 cm. (Courtesy H. Kerp.)

Figure 16.28 Portion of a monopodial Dicksoniites pluckentii

frond (Pennsylvanian). Bar  5 cm (From Krings et al., 2003.)

the fronds is irregular, and the internodes are relatively long, ranging from 10 cm to 20 cm. In addition to bipartite fronds, several late Stephanian–Early Permian specimens show pinnae with typical D. pluckenetii pinnules arranged on unbranched fronds. One of the largest of these comes from the Rotliegend (Upper Pennsylvanian– Early Permian) of Wiesweiler in southwestern Germany (FIG. 16.28). It is characterized by a slender, sinuous stem or rhizome to which are attached widely spaced, monopodially organized fronds or pinnae of the penultimate order bearing D. pluckenetii-type pinnules (Krings et al., 2003b). Similar fossils have been described from the upper Stephanian of France where they have been assigned to D. sterzelii (Langiaux, 1984).

Bedding planes at the Graissessac locality are sometimes densely covered with D. pluckenetii stems and fronds, suggesting that this species grew in dense stands or thickets in which stems and leaves of the individual plants provided mutual support. A specimen from the Rotliegend of Bad Sobernheim (Germany) demonstrates that this taxon could develop axial tendrils (FIG. 16.29) (Krings et al., 2003b). Dicksoniites pluckenetii was originally believed to represent fern foliage, until platyspermic ovules were found attached to the abaxial side of reduced pinnules. Because these ovules are structurally similar to the anatomically preserved ovule Callospermarion (Chapter 14), D. pluckenetii is today thought to represent foliage of callistophytalean seed ferns (Meyen and Lemoigne, 1986). The pollen organs of D. pluckenetii are less well known. A single fertile pinna from Graissessac bears alternating fertile pinnules with slightly reduced laminae (Galtier and Béthoux, 2002). Six synangia, each composed of six or more elongate pollen sacs, are attached to each pinnule, but there are four or fewer on the

664

Paleobotany: the biology and evolution of fossil plants

distal pinnules. Nothing is known about the contents of the pollen sacs. DISCOPTERIS

This genus is used for both sterile and fertile pinnules that superficially resemble those of Sphenopteris (Pfefferkorn, 1978). Pinnules are highly dissected and possess a prominent midvein. Fronds exhibit basal aphlebiae that become highly dissected at higher levels. In D. karwinensis, the fertile pinnules are rounded and characterized by a disk-shaped sorus at the end of the midvein, which consists of 50–70 annulate sporangia. Sporangial histology closely resembles a number of true ferns, including the genus Botryopteris (Chapter 11). EREMOPTERIS

Pinnules of Eremopteris are similar to some species of Sphenopteris but may be distinguished by the lack of a midrib and the presence of dichotomous venation. The margin may be variable, but most species are characterized by uneven teeth or lobes (FIGS. 14.80, 14.81). The genus is thought to have been produced by pteridosperms, with one species, E. zamioides, found associated with bilaterally symmetrical seeds and small synangiate pollen organs (Delevoryas and Taylor, 1969). GINKGOPHYTOPSIS

Impressions and compressions of leaves assigned to this morphogenus are relatively large and elongate flabelliform, that is, they have the form of a fan. One of the more widespread species is Ginkgophytopsis delvalii from the Westphalian (Middle–Upper Pennsylvanian) of central Europe (Gothan and Kukuk, 1933). Fossils of this foliage type are generally rare but locally known to occur in abundance on certain bedding planes (e.g., in the Ruhr district, Germany), whereas bedding planes above and below contain none. Ginkgophytopsis delvalii leaves are up to 40 cm long (FIG. 16.30), with a narrow leaf base that is somewhat arcuate. Lateral margins are slightly concave, and the distal margin is convex and may be irregularly lobed or fringed. The venation is delicate and composed of numerous parallel veins that fork several times in their course to the margin. Ginkgophytopsis belongs to an enigmatic group of Devonian–Permian foliage types that, in contrast to most other late Paleozoic foliage, were not pinnately organized. Some scholars accommodate Ginkgophytopsis and similar leaf morphotypes from Europe, Asia, and North America (e.g., Eddya, Enigmophyton, Platyphyllum, Psygmophyllum, Ginkgophyton, and Ginkgophyllum; for a note on taxonomy

Figure 16.30 Leaf of Ginkgophytopsis delvalii (Pennsylvanian). Bar  2 cm. (Courtesy BSPG.)

see Stone, 1973) in the artificial order Palaeophyllales, or they use the informal term palaeophyllalean forms for these leaves (Remy and Remy, 1977). The Palaeophyllales represent one of the great mysteries in paleobotany, because, to date, next nothing is known about the plants that produced the leaves. They are almost consistently found isolated and only one or two forms are understood in detail (e.g., Eddya; Chapter 12). The generic name Ginkgophytopsis translates to something along the lines of “plant that looks like a ginkgophyte,” but this does not mean that these leaves were produced by early ginkgophytes. Some have suggested that Ginkgophytopsis (and other palaeophyllalean) leaves represent fronds of a herbaceous fern or aphlebia of a tree fern, whereas others have associated them with seed ferns or cordaites; still others have suggested that some of them might actually represent foliage of an early ginkgophyte.

CHAPTER 16

Figure 16.31 Kankakeea showing attachment of a “bud”

to the tip of a pinna (Pennsylvanian). Bar  2 cm. (Courtesy H. W. Pfefferkorn.) KANKAKEEA

The genus Kankakeea has been used for small, thalloid-lobed tips of pinna rachides that are sometimes dichotomously branched (FIG. 16.31). They are often found attached to pinnae bearing other foliage types and have been suggested to be some form of vegetative reproductive structure. The genus has been reported from the Pennsylvanian of North America and France (Pfefferkorn, 1973). KARINOPTERIS, MARIOPTERIS, AND PSEUDOMARIOPTERIS

Mariopterids are some of the most common and widespread Pennsylvanian–Early Permian impression–compression fossils. The plants that bore these leaves were small to medium-sized scrambling or climbing plants with narrow stems and small, bi- or quadripartite fronds (Danzé-Corsin, 1953; Boersma, 1972; Krings and Kerp, 2000). In the Pennsylvanian to Early Permian, the mariopterids were represented by three major genera: Karinopteris (Namurian– Westphalian), Mariopteris (Namurian–Westphalian), and

LATE PALEOZOIC AND MESOZOIC FOLIAGE

665

Pseudomariopteris (Stephanian–Early Permian). These plants are well understood today based on numerous impression– compression fossils ranging from frond fragments and complete fronds to stem portions with attached foliage (Zeiller, 1886, 1888a, 1906; Huth, 1912a; Corsin, 1932; White, 1943; Danzé-Corsin, 1951, 1953; Boersma, 1972; Gastaldo and Boersma, 1983; Krings et al., 2001a) and well-preserved cuticles (DiMichele et al., 1984; Krings and Kerp, 2000; Krings and Schultka, 2002). The mariopterids were among the earliest late Paleozoic foliage fossils studied by means of cuticular analysis (Huth, 1912b). Karinopteris, Mariopteris, and Pseudomariopteris are thought to belong to two different orders of seed ferns. Krings and Kerp (2000) suggested that the affinities of Pseudomariopteris may be with the Callistophytales, whereas Karinopteris and Mariopteris have variously been referred to the Lyginopteridales, although persuasive evidence for this assignment has not yet been produced. Because the three genera were similar in morphology and growth habit, they are treated together here. Mariopterids produced narrow, flexuous stems that rarely exceeded 2 cm in diameter (FIG. 16.32). The most conspicuous feature of compressed stems of Karinopteris and Mariopteris is a surface pattern composed of short horizontal bars, which is due to transverse sclerotic plates (nests of sclerified cells) in the cortex. Similar cortical plates have been observed in the permineralized stems of Schopfiastrum decussatum (Lyginopteridales), which produced foliage with sphenopterid or mariopterid pinnules (Stidd and Phillips, 1973). The plates probably provided stiffness and rigidity to the stem and thus may have reduced deformation due to compression, bending, or torsion (Krings, 2003). Pseudomariopteris stems, however, do not display this surface pattern. Two large compressions from the Namurian of Hagen-Vorhalle (Germany) provide information about the structure of the root system. These specimens display portions of mariopterid stems (20 cm long and 1 cm wide) that produced elongate clusters of shoot-borne roots, some up to 40 cm long, with first- and second-order lateral roots (Krings, 2003). Mariopterid foliage consists of small fronds, usually 45 cm long, that are spirally arranged on the stem (Corsin, 1950); considerable variation with regard to the size and shape of the fronds within a single individual may occur (Kidston, 1925; Krings et al., 2001a). The mariopterid frond consists of two major parts, a naked stipe and a leafy blade; fully differentiated fronds are quadripartite in Mariopteris and bipartite in Karinopteris and Pseudomariopteris (Boersma, 1972). Quadripartite fronds (FIG. 16.32) are characterized

666

Paleobotany: the biology and evolution of fossil plants

3 2

4 1

Figure 16.32 Mariopteris muricata quadripartite frond attached to slender stem (arrow) (Pennsylvanian). Bar  5 cm. (Courtesy H. Steur.)

by two bifurcations of the stipe, the first being followed at a short interval by the second, which results in the formation of a four-part blade with four rachides; in bipartite fronds (FIG. 16.33) only one bifurcation of the stipe occurs, and a two-part blade with two rachides is formed. The individual rachides bear alternately positioned exterior and interior pinnae. Pinnae can either produce a second order of alternately positioned pinnae (in tripinnate blades) or pinnules (in bipinnate blades). In most Karinopteris and Mariopteris species, the proximal portion of the blade is tripinnate and the distal portion bipinnate, whereas in Pseudomariopteris (FIG. 16.33) the entire blade is bipinnate. Size and shape of the pinnules are variable between species, but pinnule morphology also varies within a single species, depending on the position of the pinnule in the frond. These genera do, however, possess one consistent feature—an enlarged, basal, basiscopic (pointing toward the base) lobe that occurs in the acroscopic and basiscopic pinnules positioned in the proximal

Figure 16.33 Pseudomariopteris busquetii, bipartite frond

attached to slender stem (arrow) (Pennsylvanian). Bar  1 cm.

portion of fully differentiated pinnae, which gives these pinnules an asymmetrical outline (FIGS. 16.34; 16.35). A conspicuous feature of many mariopterid fronds is the large number of specialized climber hooks (FIGS. 14.50, 16.36), some up to 4 cm long, which usually develop from apical prolongations (apical or terminal extensions) of pinna axes and pinnule midveins (Kerp and Krings 1998; Krings et al., 2003b, c). There is considerable variation among the species with regard to size and morphology of the climber hooks, ranging from simple prolongations with recurved tips to complex forms with secondary hooks on the abaxial side (DiMichele et al., 1984; Krings et al., 2003a, b). Krings et al. (2001b) hypothesized that these variations represent adaptations of individual species to certain morphological features of their most frequently utilized support plants. Stands of calamites or the fronds of larger tree ferns and pteridosperms were probably suitable support media for mariopterids (FIG. 16.37) since their structure provides an ideal support system for small fronds that used multiple hooks for attachment.

CHAPTER 16

LATE PALEOZOIC AND MESOZOIC FOLIAGE

667

Figure 16.34 Mariopteris nervosa, portion of the blade.

Note the typical mariopterid assymmetrically triangular pinnules (arrows) (Pennsylvanian). Bar  2 cm. (Courtesy BSPG.)

Figure 16.36 Climber hook of Pseudomariopteris cordatoovata (Pennsylvanian). Bar  2 mm. (From Krings and Kerp, 2000.)

Figure 16.35 Mariopteris. (From Taylor and Taylor, 1993.)

Evidence of a second mode of attachment utilized by mariopterids comes from a large fossil of Pseudomariopteris busquetii from the Lower Permian of Germany (Krings et al., 2001a). This specimen includes part of a stem with four fronds extending from one side, an indication that the stem

was closely attached to a support medium, perhaps by means of twining (FIG. 16.37). Also of special interest is a specimen from the Westphalian of northern England reported by Cleal and Thomas (1999), which shows a slender stem with mariopterid fronds still attached entwined around the stump of an arborescent lycopsid. Perhaps the mode of attachment in mariopterids changed, based on the stage of development or environmental factors (FIG. 16.37). These plants may also have used different modes of attachment simultaneously, which was no doubt particularly beneficial in rapid colonization of new habitats where the quantity and quality of supporting structures could vary considerably. If no suitable supports were available, mariopterids may have formed dense thickets (Krings et al., 2003b). Krings et al. (2001a) suggested that mariopterids represent part of a vigorously growing or sprawling type of vegetation structurally comparable to the vegetation found at the edges or in disturbed areas of modern tropical and subtropical forests.

668

Paleobotany: the biology and evolution of fossil plants

Figure 16.37 Suggested reconstruction of Pseudomariopteris busquetii using an arborescent as a support system. (From Krings

et al., 2001a.)

The reproductive biology of the mariopterids remains largely unknown. Gothan (1935) suggested that the cupulate fructification Calathiops bernhardtii from the Namurian of Hagen-Vorhalle (Germany) may have been produced by a mariopterid, and Danzé-Corsin (1957) briefly described

specimens of Mariopteris latifolia from the Westphalian of France that bear pollen sacs, arranged in verticils, on prolongations of the pinnule veins. Unfortunately, these latter fossils were lost without ever having been figured (Boersma, 1969). Thus, no convincing fertile specimens of Karinopteris

CHAPTER 16

LATE PALEOZOIC AND MESOZOIC FOLIAGE

Figure 16.39

Figure 16.38 Lesleya sp. (Pennsylvanian). Bar  5 cm. (Courtesy

R. L. Leary.)

or Mariopteris are known today. The only indisputable evidence for the morphology and spatial arrangement of reproductive structures in mariopterids comes from a species of Pseudomariopteris. A frond portion of P. busquetii from the Stephanian of France displays small platyspermic seeds positioned on the abaxial side of pinnules that otherwise do not differ from vegetative foliage (Krings and Kerp, 2000). LESLEYA

This genus is known from the Mississippian to Late Permian of Europe and North America, with less than 15 species described to date (Florin, 1933b; Leary, 1980; Šimuº nek, 1996b). Lesleya (FIG. 16.38) consists of individual, strap-shaped leaves, each up to 20 cm long and 5 cm wide, with entire or dentate margins and a massive midvein. Leaves superficially resemble

669

Richard L. Leary.

Taeniopteris but differ from the latter in that the secondary veins ascend to the margin at a relatively steep angle. The veins do not arch but rather remain more or less straight or become sinuous in their course to the margin. Leaves may be hypostomatic or amphistomatic (Šimuº nek, 1996b). The systematic affinities of Lesleya remain uncertain. Florin (1933b) regarded the genus as a primitive gymnosperm, not closely related to either late Paleozoic seed ferns or Mesozoic gymnosperms based on morphological and epidermal features. The genus has been suggested to be ancestral to cycads (Remy and Remy, 1978a; Leary, 1990) (FIG. 16.39). LINOPTERIS, RETICULOPTERIS, AND BARTHELOPTERIS

Pinnules of these three Pennsylvanian–Early Permian genera resemble those of Neuropteris sensu lato in all characteristics (i.e., size and shape, midvein, and attachment to the axis), with the exception that they have reticulate venation in which the secondary veins anastomose as they extend toward the pinnule margin (Wagner and Lemos de Sousa, 1982; Zodrow and Cleal, 1993). In permineralized pinnules of Reticulopteris muensteri, pore-like openings are recognizable near the vein endings, and these have been compared with hydathodes (Reihman and Schabilion, 1978). Linopteris, Reticulopteris, and Barthelopteris may be nearly identical with regard to pinnule morphology; as a result, many species in these genera have a long and complicated nomenclatural history. Some authorities suggest that Linopteris

670

Paleobotany: the biology and evolution of fossil plants

Figure

16.40 Linopteris neuropteroides (Pennsylvanian).

Bar  2 cm.

Figure 16.42 Barthelopteris germari (Pennsylvanian). Bar  1 cm.

(Courtesy H. Kerp.)

Figure 16.41 Reticulopteris pinnule. (From Taylor and Taylor,

1993.)

(FIGS. 16.40, 16.43) and Reticulopteris (FIG. 16.41) can be distinguished by the pseudoreticulate venation in the latter in which the veins bend toward one another but do not actually anastomose. Others suggest that the two taxa merely

represent variations of a single genus. This latter hypothesis has been challenged by Laveine and DiMichele (2001), however, based on a survey of the morphological characteristics of the two genera. Especially significant is that Linopteris is paripinnate (no terminal pinnule on the frond) and thus probably related to Paripteris (Parispermaceae sensu Laveine et al., 1993b), whereas Reticulopteris is imparipinnate and represents a mesh-veined form of Neuropteris sensu stricto (Cleal and Shute, 1995). A form that appears to be intermediate between Neuropteris and Reticulopteris is N. semireticulata from the Westphalian of Germany (Josten, 1962). This species displays strongly sinuous secondary veins bent so closely toward one another that they almost appear to anastomose. Barthelopteris (FIG. 16.42) corresponds to Reticulopteris with regard to overall morphology but can be distinguished from the latter based on differences in venation, epidermal anatomy, and stratigraphic range (Zodrow and Cleal, 1993). Other distinguishing features include the presence of peltate glandular trichomes on the abaxial pinnule surface (FIG. 14.102) and a distinct “surrounding vein” along the pinnule margin in Barthelopteris germani (Krings and Kerp, 1998). According to Cleal and Shute (1995), Barthelopteris

CHAPTER 16

represents the mesh-veined analog of Neurocallipteris. It has been suggested that reticulate venation patterns may have developed in seed ferns in response to increased physiological stress, due to reduced water availability (Zodrow and Cleal, 1993). The individual veins in pinnules with reticulate venation are longer, and, as a result, the area of interface between the veins and mesophyll is increased. Laveine and DiMichele (2001) suggested that Reticulopteris fronds were bipartite and at least 2.4 m wide. Specimens from the Stephanian of France (Langiaux, 1984; Krings and Kerp, 1998) indicate that Barthelopteris fronds also were bipartite but probably smaller than those of Reticulopteris. Fronds of Linopteris are believed to generally resemble those of Paripteris (Laveine et al., 1993b). Zodrow et al. (2007) provided evidence suggesting that the L. obliqua plant was a medium-sized tree with a trunk 0.2 m wide and complex fronds 7 m long. The fronds were not strictly planar structures as suggested by Laveine et al. (1993b) and Wnuk and Pfefferkorn (1984) but rather attained some degree of three-dimensional architecture. At least one Linopteris (FIG. 16.43) species is known to have been produced by the seed fern stem Sutcliffia (Phillips and

LATE PALEOZOIC AND MESOZOIC FOLIAGE

671

Andrews, 1963; Stidd et al., 1975; Laveine and DiMichele, 2001). Pollen organs of Linopteris are assignable to Potoniea and produce pollen grains of the Crassispora type (Laveine et al., 1993b). Zodrow et al. (2007) reported on ovules of the Hexagonocarpus type which were attached to the petiole near the base of an L. obliqua frond. LOBATOPTERIS

This Pennsylvanian genus consists of large fronds with pinnules broadly attached to the pinna axis at angles of 60°–90°. Pinnule morphology is similar to that in some species of Pecopteris. Pinnule venation consists of a primary vein with laterals that dichotomize three times (FIG. 16.44) (Wagner, 1958). Specimens with foliar-borne sporangia suggest affinities with marattialean ferns (Shute and Cleal, 1989; Pšenicka et al., 2008). Using simple fractal techniques to produce figures that closely resemble a Lobatopteris frond, Heggie and Zodrow (1994) have shown that the fractal geometry underlying the frond is simple, at least based on a mathematical model.

Figure 16.43 Linopteris pinnule. (From Taylor and Taylor,

1993.)

Figure 16.44 Lobatopteris micromiltonii. (From Wagner, 1958.)

672

Paleobotany: the biology and evolution of fossil plants

LONCHOPTERIDIUM AND LONCHOPTERIS

Pinnules of these Pennsylvanian foliage morphogenera are generally similar to those of Alethopteris, except that the venation of the former displays pseudoanastomoses or anastomoses of the lateral veins. The epidermal anatomy of Lonchopteridium has been detailed by Zodrow and Cleal (1998, for L. karvinensis) and that of Lonchopteris (FIGS. 16.45, 16.46) by Barthel (1961, for L. bricei). Pinnules are hypostomatic, with anomocytic stomata restricted to the intercostal fields. Fronds were probably bipartite and at least tripinnate (Purkynˇová, 1971; Zodrow and Cleal, 1998). It has been suggested that Lonchopteridium occupies a position between Alethopteris and Lonchopteris based on the intermediate nature of the lateral venation and similarities in epidermal anatomy. Specimens are believed to have been produced by seed ferns. MEGALOPTERIS

Megalopteris pinnules are long, strap shaped, and sometimes bifurcated nearly to the base (FIG. 16.47). The pinnule midrib is prominent, with many closely spaced laterals ascending to the margin at a steep angle. The end of the pinnule is often sharply pointed. Nothing is known about the plant that produced Megalopteris foliage; it is thought to be the leaf of a seed fern, a hypothesis strengthened by the similarity of the epidermis and stomata to those in Neuropteris and a general morphology similar to Alethopteris. Megalopteris represents a rather rare foliage type, but it has been reported in abundance from one Early Pennsylvanian flora in which a large percentage of the taxa are included in the noeggerathialeans (Leary and Pfefferkorn, 1977). Because of its rare occurrence with other Paleozoic foliage form genera, Megalopteris has been interpreted as an indicator of so-called upland habitats.

Figure 16.45 Lonchopteris. (From Taylor and Taylor, 1993.)

Figure 16.46 Lonchopteris rugosa (Pennsylvanian). Bar  2 cm.

(Courtesy BSPG.)

Figure 16.47 Megalopteris. (From Taylor and Taylor, 1993.)

CHAPTER 16

NEUROPTERIS SENSU LATO

Neuropterid foliage is one of the most common and widespread type of plant fossil in the Carboniferous. The genus Neuropteris was originally established by Brongniart (1822) for compound leaves or fronds in which pinnules have a constricted (cordate) base and non-anastomosing venation (FIG. 16.48). Neuropteris pinnules have rounded (FIG. 16.49) or pointed tips and each is supplied by a single midvein from which arise arched, dichotomously branched secondary veins. Neuropterid fronds are associated with medullosan seeds and pollen organs (Cleal and Shute, 1995). The vast majority of neuropterid foliage types are known only from impressions or compressions, but a few forms are known from permineralized specimens. Reihman and Schabilion (1978) and Oestry-Stidd (1979) reported on pinnules of N. rarinervis preserved in coal balls from the Middle Pennsylvanian of Iowa and Illinois. The pinnules display vascular bundles composed of tracheids with spiral–scalariform thickenings surrounded by parenchymatous bundle sheaths with bundle sheath extensions. Cells of the upper epidermis possess undulating anticlinal walls, and stomata are confined to the intercostal fields of the lower pinnule surface. Each pair

Figure 16.48 Neuropteris (From Taylor and Taylor, 1993.)

LATE PALEOZOIC AND MESOZOIC FOLIAGE

673

of guard cells is surrounded by four to six polygonal subsidiary cells. Many neuropterid pinnules have marginal thickenings of the lateral pinnule veins. These structures are very similar to the thickened vein endings present in extant ferns, such as species of Phyllitis or Polypodium, where they consist of groups of short, large-diametered tracheids that end below the upper epidermis and often terminate in distinct water pits. Vein endings terminating in water pits are regarded as functional analogs to hydathodes, which are specialized epidermal structures in leaves that function to exude water from the leaf. Neuropteris in its traditional, broad sense (i.e., Neuropteris sensu lato) represents a highly heterogeneous taxon comprised of several distinct clusters of species. Several groups of species that are technically neuropterid with regard to general appearance and pinnule outline, but which display deviations from Neuropteris sensu stricto, were placed into separate genera (Cleal et al., 1990). A synoptical evaluation of these genera based on both frond architectural and epidermal

Figure 16.49 Neuropteris sp. (Pennsylvanian). Bar  2 cm.

674

Paleobotany: the biology and evolution of fossil plants

features has resulted in a generic reclassification of the European species of Neuropteris sensu lato into nine more closely circumscribed and homogenous morphogenera (Cleal and Shute, 1995). These are discussed below.

mally that structurally resemble those found in some odontopterids (Laveine and Belhis, 2007). Distinctive epidermal features of M. macrophylla include brachyparacytic or cyclocytic stomata (Cleal and Zodrow, 1989) (FIG. 16.53).

LAVEINOPTERIS Species in this genus are characterized by large orbicular foliar elements (cyclopterid pinnules) in the proximal portion of the frond (FIG. 16.51). Fronds are bipinnate, relatively large and at least tripinnate (Laveine et al., 2005). Epidermal features of Laveinopteris loshii include the absence of multicellular trichomes and the weak differentiation of costal and intercostal fields on the adaxial pinnule surface (Cleal and Shute, 1992) (FIG. 16.52).

MARGARITOPTERIS Species included in this genus are characterized by broadly attached and lobed pinnules. Prior to the introduction of the genus Margaritopteris by Gothan (1913), these forms were typically assigned to Odontopteris or Sphenopteris. NEURALETHOPTERIS This taxon is used for foliage that displays a mixture of pinnule characteristics found in Alethopteris and Neuropteris, such

MACRONEUROPTERIS This foliage type is very similar to Neuropteris sensu stricto, except that the pinnules (FIG. 16.53) are larger and the fronds less divided. Although typical Neuropteris sensu stricto fronds are tri- or quadripinnate, those of Macroneuropteris are mostly bipinnate (Cleal et al., 1996, 1998). Macroneuropteris fronds are up to 2.5 m long and bear enlarged pinnules proxi-

Figure 16.51 Bifurcation of Laveniopteris frond showing fragmentary cyclopterids attached to the distal portion of the petiole (Pennsylvanian). Bar  2 cm. (Courtesy J. -P. Laveine.)

Figure 16.50 Large portion of Laveinopteris rarinervis frond

(Pennsylvanian). Bar  5 cm. (Courtesy J. -P. Laveine.)

Figure 16.52

Cedric Shute.

CHAPTER 16

as the co-occurrence of pinnules with cordate base and pinnules attached by the entire base in a single specimen (Josten, 1983). The venation pattern generally resembles that seen in Alethopteris (Goubet et al., 2000). Neuralethopterid fronds are bipartite, tripinnate, and lack intercalary elements, which separates this genus from most other neuropterids and suggests affinities closer to the alethopterids (Laveine et al., 1993a). NEUROCALLIPTERIS This genus includes more complex stomatal apparati than typical neuropterids (Reichel and Barthel, 1964 for N. planchardii and Barthel, 1976a for N. neuropteroides and N. planchardii). Frond architecture appears to have been similar to that in Neuropteris sensu stricto. NEURODONTOPTERIS This foliage type has pinnules that are intermediate between Neuropteris sensu stricto and Odontopteris. Fronds of Neurodontopteris are smaller and less divided than

LATE PALEOZOIC AND MESOZOIC FOLIAGE

675

Neuropteris sensu stricto, and pinnules tend to be fused to the axis along the basiscopic side (Langiaux, 1984). Some species, such as N. auriculata, are characterized by amphistomatic pinnules (Krings, 1999; Šimuº nek, 1999), a condition rarely observed in neuropterids. NEUROPTERIS SENSU STRICTO In this morphogenus, the pinnules are often partly fused to the pinna axis and have a relatively weak midvein. The most important feature separating this taxon from Laveinopteris is the absence of proximal orbicular foliar elements in Neuropteris sensu stricto. Instead, pinnae in Neuropteris sensu stricto bear pinnules that are enlarged or laciniate, that is, deeply divided into irregular segments (Cleal and Shute, 1991b). Fronds are bi- and tri-partite or rarely quadripinnate. Pinnules have anomocytic or brachyparacytic stomata and often produce numerous trichomes, especially on the abaxial side (Barthel, 1962a for N. ovata). PARIPTERIS This genus can be separated from all other neuropterid foliage by its distinctive, paripinnate frond architecture, that is the frond terminates with a pair of apical pinnules, and also has intercalary pinnules on the penultimate pinna axes. Fronds are sizeable (5 m long) and tri- or quadripinnate and pinnules are more or less distinctly curved in an upward direction, i.e., toward the pinna tip (FIG. 16.54). A reconstruction

Figure 16.53 Macroneuropteris scheuchzeri from Mazon Creek

(Pennsylvanian). Bar  2 cm.

Figure 16.54 Paripteris gigantea pinnules (Pennsylvanian). Bar  1 cm. (Courtesy J.-P. Laveine.)

676

Paleobotany: the biology and evolution of fossil plants

presented by Laveine et al. (1993b) depicts the frond as unbranched. SPHENONEUROPTERIS These fronds have large, relatively lax pinnules with a low vein density. One species assigned to Sphenoneuropteris by Cleal and Shute (1995), S. praedentata, is now included in Blanzyopteris (see below). NEUROPTERID GROWTH HABIT Most neuropterids are believed to have been small- to medium-sized, free-standing trees resembling tree ferns (Bertrand and Corsin, 1950; Shute and Cleal, 2002; Knaus and Lucas, 2004), whereas others have been interpreted as leaners, vines, or lianas based on features corresponding to those of extant leaning and lianescent plants. For example, Laveinopteris (Neuropteris) attenuata from the Pennsylvanian of Germany includes part of the compressed stem and an almost complete frond. Maximum stem diameter is only 4 cm, but internode lengths are 20 cm, and the frond is 90 cm long. These features strongly suggest that L. attenuata was either a leaning or lianescent plant (Krings and Kerp, 2006). We do not yet know how L. attenuata plants and other lianescent neuropterids attached to support plants; hook-like cortical spines that may have been effective as scrambling or climbing aids, however, have been documented on the petioles of several neuropterids (Beeler, 1983; Cross et al., 1996). BLANZYOPTERIS Perhaps, the most fully understood neuropterid with regard to growth habit and paleoautecology is Blanzyopteris praedentata from the Stephanian (Upper Pennsylvanian) of France (Krings and Kerp, 1999). Several stem portions, complete fronds, and large frond portions come from the Commentry and Blanzy-Montceau Basins in the Massif Central of France (Zeiller, 1888b, 1906). These fossils indicate that the plants were large vines with slender stems (FIG. 16.55) and relatively small (25–40 cm long), monopodial, bipinnate fronds bearing neuropterid pinnules. Cuticle preparations indicate that B. praedentata produced complex climbing organs (15 cm long) composed of a main axis or tendril that produced lateral branches up to 5 cm long. The laterals bore numerous branchlets up to 1.5 cm long that widened apically and terminated in adhesive pads. A ring of specialized epidermal cells just beneath the adhesive pads is interpreted as a shock absorbing structure that may have functioned like the bellows of a concertina, that is, extending under tension and contracting under compression (Speck et al., 2000). Moreover, the exceptional preservation of the cuticles on this plant has made it possible to fully reconstruct the

Figure 16.55 Suggested reconstruction of the Blanzyopteris praedentata plant. (From Krings and Kerp, 1999.)

indumentum (hair cover) of the fronds and tendrils (Krings et al., 2003c, 2004). Six distinct types of uniseriate, multicellular, non-glandular, and capitate glandular trichomes occur on B. praedentata, each of which displays a different spatial distribution on the frond or tendril. Especially interesting are filament-bearing glandular trichomes up to 1 mm long that consist of a multicellular stalk terminated by one to several enlarged secretory cells. The distal secretory cell bears a filament of several smaller cells. The filament is broken off in most samples, and the secretory cells are typically open at the point where the filament was once attached. These trichomes are interpreted as a specialized defense mechanism against arthropods in which the filament was a touch-sensitive trigger that opened the secretory cell(s) upon contact by arthropods moving on the plant (Krings et al., 2002, 2003c).

CHAPTER 16

LATE PALEOZOIC AND MESOZOIC FOLIAGE

677

The trichomes of B. praedentata are interpreted as functionally similar to the so-called explosive trichomes seen in certain extant Cucurbitaceae and Solanaceae. Studies of living aphids on Sicana odorifera (Cucurbitaceae) demonstrate the effectiveness of this defense (Kellogg et al., 2002). When touched and ruptured, the trichomes of S. odorifera rapidly release a sticky exudate that adheres to the animal’s legs. Eventually, the legs become encased within a thick layer of solidified secretion in which the tarsi and pretarsi (the prehensile tools with which aphids attach to the plant) are no longer effective. As a result, the aphids are unable to maintain contact on the plant surface. NOTHORHACOPTERIS

Nothorhacopteris is one of the most characteristic floral elements of Gondwana from the Late Mississippian to the Early Permian (Iannuzzi and Rösler, 2000; Iannuzzi and Pfefferkorn, 2002). The genus includes pinnate fronds with pinnules of varying shape (triangular–subcircular) that may be entire or slightly lobed. Venation is dichotomous with a few smaller veins extending between the larger ones. Based on the common association with silicified progymnospermlike logs, it has long been suggested that Nothorhacopteris may have affinities with the progymnosperms (Archangelsky, 1983). However, the discovery of fertile N. argentinica fronds bearing apical cupules (i.e., Austrocalyx jejenensis) from the Carboniferous of Argentina (Vega and Archangelsky, 1996) established that this foliage type was produced by pteridosperms (see Chapter 14).

Figure 16.56 Odontopteris lingulata pinnules showing venation

(Permian) Bar  5 mm. (Courtesy M. Barthel and R. Werneburg.)

ODONTOPTERIS AND LESCUROPTERIS

Odontopteris includes 200 Late Pennsylvanian and Early Permian foliage species (FIG. 16.56), predominantly from Europe and North America (Cleal and Shute, 1991a). The characteristic feature of this genus is pinnules that are entirely adherent to the axis and lack a typical midvein; rather one to several veins enter the pinnule and may dichotomize at irregular intervals in their course to the margin (Knight, 1983). With regard to pinnule morphology, the species in Odontopteris are divided into roughly two groups, one with elongate, asymmetrically triangular pinnules, for example O. brardii (FIG. 16.57) and the other with rounded pinnules (FIG. 16.58), for example O. subcrenulata. Based on these differences, Doubinger (1956) (FIG. 15.14) suggested that Odontopteris should be subdivided into two subgenera, which according to Wagner (1964) coincide with Weiss’s (1870b) genera Xenopteris (elongate, triangular pinnules) and Mixoneura (rounded pinnules). Most species in Odontopteris represent seed fern foliage, most likely with affinities in the Medullosales (Cleal and Shute, 1991a; Šimuº nek and Cleal, 2004).

Figure 16.57 Odontopteris brardii (Pennsylvanian). Bar 

4 cm.

678

Paleobotany: the biology and evolution of fossil plants

Figure 16.58 Odontopteris pinnule. (From Taylor and Taylor,

1993.)

The traditional concept of odontopterid frond architecture is based on a reconstruction published by Zeiller (1900). It depicts a relatively small (1 m long), bipartite, and asymmetrically tripinnate (i.e., tripinnate on the exterior and bipinnate on the interior side of the rachis) frond (FIG. 16.59) bearing dissected pinnules, termed cyclopteroid elements (FIGS. 16.25, 16.26), along the stipe. Krings et al. (2006), however, demonstrated that at least one species, O. brardii, was capable of producing entirely bipinnate and asymmetrically tripinnate fronds and thus displayed considerable frond variability. In addition, a small odontopterid frond from the Stephanian of France depicts a form intermediate between the bipinnate and asymmetrically tripinnate frond architecture. It is asymmetrically tripinnate in the proximal part of the frond and bipinnate in the distal portion. The epidermal anatomy of only a few Odontopteris species is known in detail (Barthel, 1961, 1962; Krings et al., 2000; Kerp and Krings, 2003; Cleal et al., 2007). Lescuropteris is very similar to certain species of Odontopteris with regard to pinnule morphology and venation (FIG. 16.60). The genus was originally included in Odontopteris but was later excluded by Remy and Remy (1975c), based on the presence of so-called intercalary pinnules that are missing in most members of Odontopteris. Most fronds of Lescuropteris are bipinnate, but material described by Round (1921) from North America suggests that this form may also have produced tripinnate fronds. The most completely known species in the genus is Lescuropteris

Figure 16.59 Suggested reconstruction of the asymmetrically tripinnate frond of Odontopteris brardii (O. minor-zeilleri). (From Zeiller, 1900.)

Figure 16.60 Lescuropteris pinnule. (From Taylor and Taylor,

1993.)

CHAPTER 16

LATE PALEOZOIC AND MESOZOIC FOLIAGE

679

Figure 16.62 Lescuropteris genuina showing the apical portion of a pinna with several pinnules modified into tendrils (Pennsylvanian). Bar  3 mm.

Figure 16.61 Lescuropteris genuina showing frond stipe and cyclopteroid basal leaves (Pennsylvanian). Bar  8 mm.

genuina from the Stephanian of France (Zeiller, 1906) and Spain (Knight, 1983). Fronds of L. genuina are bipartite, up to 70 cm long, bipinnate, and bear lobed, cyclopteroid elements in the proximal portion of the long stipe. With increasing proximity to the bifurcation point of the frond, cyclopteroid elements progressively differentiate into pinnae with individual pinnules (FIG. 16.61); in some specimens, however, lobed cyclopteroid elements are present below and immediately above the bifurcation (Laveine, 1997). Alternately positioned pinnae of the ultimate order are produced from the rachides; exterior and interior pinnae are more or less identical in morphology. Pinnules are broadly attached, generally oblong, and asymmetrically triangular in outline, entire margined usually with tapered tips, and up to 1.0 cm long and 0.5 cm wide.

Lescuropteris genuina has been interpreted as a vine or liana based on the presence of leaflet tendrils that developed from specialized apical portions of pinnae (FIGS. 16.62, 16.63). These consist of one or several pinnules that are variously modified into tendrils similar to leaflet tendrils produced by many extant Fabaceae (Krings and Kerp, 1997). In addition, one specimen indicates that tendrils also developed from apical extensions of normally shaped pinnules (i.e., prolongations of one vein through the pinnule apex). The epidermal cells of the tendrils have large papillae, not found in any other part of the plant, which were no doubt advantageous for affixing a young and developing tendril to the surface of its support. In tendrils of extant climbing plants, papillae are known to function in transmission of contact stimuli, which are important in directing tendril growth around the support, a process called thigmotropism (Krings et al., 2003a). The structure of the root system, shoot axis or stem, and reproductive biology of Odontopteris and Lescuropteris remain unknown to date. PECOPTERIS

The pinnules of this common foliage type are typically small and arise free from one another. They are attached along the

680

Paleobotany: the biology and evolution of fossil plants

Figure 16.64 Pecopteris pinnule. (From Taylor and Taylor,

1993.)

of its forked petiole. Partially, petrified axes bearing foliage of the Rhodea type have been found attached to stems with Heterangium anatomy (Chapter 14) (Jennings, 1976). SPHENOPTERIS

Figure 16.63 Lescuropteris genuina showing the apical por-

tion of a normal pinna (Pennsylvanian). Bar  4 mm.

entire width of the base and often have parallel or nearly parallel margins (FIG. 16.64). A single vein enters the base and extends almost to the rounded pinnule tip. Lateral veins are straight or slightly arched and are simple or forked near the margin. Pecopteris-type foliage is known to been borne by the Carboniferous marattialean fern, Psaronius (Chapter 11), several filicalean ferns, and at least one seed fern. Although the genus extends throughout the Carboniferous and into the Permian, few forms are known from the Lower Pennsylvanian (but see Gastaldo, 1984). RHODEA (RHODEOPTERIDIUM)

Rhodea pinnules usually consist of dichotomous subdivisions of narrow segments (FIG. 16.65) in which distinct laminae are almost extirely absent (FIGS. 16.66–16.68). The genus is similar to Diplotmema (FIG. 16.69) (often misspelled Diplothmema) or to some highly dissected species of Sphenopteris; the former can be distinguished on the basis

The morphotaxon Sphenopteris in its traditional, broad sense (sensu lato) represents a large and confusing assemblage of more or less well-known, predominantly late Paleozoic (uppermost Devonian–Permian) foliage types, generally characterized by pinnules constricted at the base. In some forms, pinnules may be oval in outline and almost entire margined, whereas others have pinnules that are lobed or variously toothed. Pinnules are usually decurrent, giving the pinna axes a distinctly winged appearance. The midvein is straight or flexuous (FIG. 16.70) and produces forking secondary veins that depart at a steep angle and extend toward the margin either singly or in small groups. Sphenopteris foliage is very common in the Carboniferous, where it is associated with several true ferns as well as seed ferns, although the genus appears to have survived into the Mesozoic. Mesozoic representatives of Sphenopteris are generally rare but seem to be widespread based on reports from several countries (e.g., China, Japan, Russia, the United States, South Africa [surveyed in Ash, 1999], and Antarctica) [Rees and Cleal, 2004]. Several attempts have been made to define clusters or groups within Sphenopteris or to subdivide the genus based on pinnule features or morphology of the reproductive structures, which may be attached or just associated with the foliage (surveyed in Van Amerom, 1975). The primary problem with these studies is that the individual species are inconsistently understood; some forms are based on small,

CHAPTER 16

LATE PALEOZOIC AND MESOZOIC FOLIAGE

681

Figure 16.66 Rhodea. (From Taylor and Taylor, 1993.)

Figure 16.65 Rhodea cf. moravica (Mississippian). Bar  1 cm.

(Courtesy GBA.)

isolated pinna fragments, whereas others are known from entire fronds or frond portions attached to the stem segments (e.g., Krings and Schultka, 2000). Van Amerom (1975) proposed that Sphenopteris sensu lato be used for foliage fossils for which reproductive structures are not known. Other forms, for which gymnospermous reproductive structures

Figure 16.67 Rhodea hochstetteri (Mississippian). Bar  1 cm.

(Courtesy GBA.)

682

Paleobotany: the biology and evolution of fossil plants

Figure 16.68 Rhodea (Pennsylvanian). Bar  1 cm.

Figure 16.70 Sphenopteris. (From Taylor and Taylor, 1993.)

have been demonstrated in organic connection or can be assigned with some degree of confidence, are assigned either to Sphenopteris sensu stricto or to Eusphenopteris (FIG. 16.71). Still other forms, for which the plant that bore them is known in detail, should be excluded from morpho-classification and referred to a whole-plant concept. The best example of the latter is S. hoeninghausii, a foliage species known to have been produced by Lyginopteris oldhamia stems (Chapter 14). SPHENOPTERIS SENSU STRICTO Fronds are bipartite and tripinnate with rachides and pinna axes characterized by transverse sclerotic plates in the cortex (Heterangium-type structure, Chapter 14). Pinnules have an inverted triangular or elongate ovoid form in cross section and are attached to the axis by a short pedicel; the lamina appears reduced in comparison to that in Eusphenopteris. Pollen organs are of the Telangium type, cupules belong to the Calymmatotheca or Calathiops type, and seeds are of the Lagenospermum type (Van Amerom, 1975). The plants that bore Sphenopteris sensu stricto foliage are believed to have been scramblers or lianas (Remy and Remy, 1977).

Figure 16.69 Diplotmema. (From Taylor and Taylor, 1993.)

EUSPHENOPTERIS Fronds are symmetrically bipartite and tripinnate (Laveine, 1989), with the rachides of some species characterized by transverse sclerotic plates. The trilobed pinnules are characteristically rounded and may have prominent stalks (Simson-Scharold,

CHAPTER 16

LATE PALEOZOIC AND MESOZOIC FOLIAGE

683

Figure 16.72 Spiropteris sp., coiled, immature frond (circinate vernation) of a fern (Rhato-Liassic). Bar  1 cm. (Courtesy BSPG.)

Figure 16.71 Eusphenopteris. (From Taylor and Taylor, 1993.)

1934). Pinnule margins are entire, and the basal pinnule of a pinna is often somewhat larger than the others (FIG. 16.71). Each pinnule lobe contains a vein that repeatedly dichotomizes close to the edge of the lamina. Pollen organs are of the Telangiopsis type, cupules are assignable to Calymmatotheca, and seeds are probably of the Nudospermum type (Van Amerom, 1975). Stem segments assignable to Eusphenopteris foliage lack a Dictyoxylon-type cortex, which consists of a series of vertical fibrous strands that form a netlike system. Shadle and Stidd (1975) provided evidence that one species, E. obtusiloba, was probably produced by stems of Heterangium americanum (Chapter 14). Specimens of E. sanjuanina have been described from the Paganzo and Calingasta-Uspallata Basins of Argentina, making it a common foliage type of Carboniferous floras of Argentina (Césari et al., 1988). SPIROPTERIS

Spiropteris is the generic name used for the coiled, immature leaves (circinate vernation) of various Paleozoic and Mesoszoic fossil ferns and seed ferns (FIG. 16.72).

Figure 16.73 Taeniopteris venation. (From Taylor and Taylor,

1993.) TAENIOPTERIS

Specimens assigned to Taeniopteris represent a heterogeneous group of foliage that includes at least two types of leaves. One type consists of individual, strap-shaped leaves with a prominent midrib from which depart arching secondary veins at an angle of approximately 70°; these branch once or several times before they reach the entire margin (FIG. 16.73). Also included are pinnately organized fronds with large, more

684

Paleobotany: the biology and evolution of fossil plants

Figure 16.74 Taeniopteris jejunata, sterile pinna (Permian).

Bar  1 cm. (Courtesy M. Barthel.)

or less petiolate pinnules displaying a venation pattern that is identical to that seen in the strap-shaped leaves (FIG. 16.74). The genus Taeniopteris extends from the Carboniferous into the Cretaceous and constitutes an excellent example of a morphogenus (Remy and Remy, 1975d). Late Paleozoic taeniopterid foliage is believed to have been produced by ferns or seed ferns, but Mesozoic forms were probably produced by several groups of plants, including cycads, bennettitaleans, and other gymnosperms such as Pentoxylales. The epidermal anatomy of T. jejunata, a pinnately organized Pennsylvanian–Early Permian form has been detailed by Barthel (1962a) based on material from the Rotliegend of Germany. Pinnules are hypostomatic and display differentiation of costal and intercostal fields on both the adaxial and abaxial surfaces. Stomata are haplocheilic and monocyclic and possess weakly cutinized guard cells. Fertile specimens of T. jejunata come from the Manebach Formation (Rotliegend) in Germany (FIG. 16.75). They consist of pinnules in which the distal portion of the lamina is reduced in width and bears numerous, small oval structures interpreted as ovules, each attached marginally and positioned over the vein endings (Barthel et al., 1975); these fructifications correspond to what was described as Ilfeldia (FIG. 16.75) by Remy (1953). Barthel (2006a) suggested that Manebachia polysporangiata, a pinnule with a distally reduced lamina which lacks secondary veins may represent the synangiate pollen organ

Figure 16.75 Taeniopteris jejunata, partially fertile pinnule

(Ilfeldia) (Permian). (From Barthel et al., 1975.)

of T. jejunata. This fossil from the Manebach Formation has groups of two to four basally fused pollen sacs (Remy and Remy, 1958) attached along the margin of the pinnule. Pennsylvanian and Early Permian Taeniopteris-type foliage from North America displaying small seeds attached to the abaxial surface has been described as Spermopteris,

CHAPTER 16

LATE PALEOZOIC AND MESOZOIC FOLIAGE

685

Phasmatocycas, and Archaeocycas (surveyed in Axsmith et al., 2003) and formed the basis for an influential theory on the evolution of the cycad megasporophyll (Mamay, 1976a, Chapter 17). Numerous Taeniopteris-type foliage fossils from China, such as Procycas densinervioides, Cladotaeniopteris shaanxiensis, and Lepingia emarginata have also variously been proposed as cycad ancestors or early cycads based on leaf morphology or epidermal anatomy (S. Zhang and Mo, 1981; Liu and Yao, 2002a). These forms are equivocal as possible cycads, and fertile specimens will be required before their systematic position can be demonstrated (Axsmith et al., 2003). TINSLEYA

This is an Early Permian foliage type thought to be related to Callipteris (Mamay, 1966). The fronds are large and bipinnate, with closely spaced pinnules up to 4.0 cm long (FIG. 16.76). Venation is odontopteroid with four to eight veins entering each decurrent base. Urn-shaped seeds 2.0 cm long are borne at the tips of some pinnules. Tinsleya texana is believed to be a seed fern. TRIPHYLLOPTERIS, GENSELIA, AND CHARBECKIA

Triphyllopteris is the generic name applied to Mississippian Sphenopteris-like foliage from Europe and North America that may have been produced by members of the Paleozoic seed fern groups Calamopityales or Buteoxylonales. Individual pinnules are highly variable in outline (FIG. 16.77), ranging from unlobed to fan shaped and occasionally dentate or pointed. Pinnules are obliquely inserted on the rachis and abaxially concave. Fertile specimens have been discovered from the Price Formation of Virginia and these consist of massive aggregations of elongate sporangia that contain trilete spores (Skog and Gensel, 1980). A reevaluation of the generic concept of Triphyllopteris, which was based exclusively on material from Europe, revealed that most North American forms assigned to that genus differ from the European fossils (Knaus, 1994, 1995; Knaus and Gillespie, 2001). As a result, the North American species were excluded from Triphyllopteris and placed in the genus Genselia (FIG. 16.78). Fronds of Genselia are generally bipinnate and have a non-bifurcate rachis. The petiole is basally widened. Pinnules are borne either individually or in clusters, usually of three. Pinnule attachment is Sphenopteris like, sessile to stalked, and often decurrent. Proximal pinnules of each pinna are either broad and infrequently lobed or lobed to the same degree as other pinnules. Distal pinnules appear deeply cleft in some species, but in others are closely spaced, entire-margined pinnules appressed to each other (Knaus and Gillespie, 2001). The pinnules are vascularized

Figure 16.76 Tinsleya texana (Permian). (Courtesy D. Chaney and W. D. DiMichele.)

Bar  5 cm.

by repeatedly forking veins radiating outward from the point of pinnule attachment, forming an open dichotomous pattern. A coeval form resembling Genselia in basic structure is Charbeckia (FIG. 16.79) (Knaus et al., 2000). Charbeckia fronds are distinctly larger in all parts (extending up to 1 m long), and pinnules are never lobed. Specimens of Triphyllopteris have been described from the Carboniferous of Argentina bearing erect, multilobed cupules (Vega and Archangelsky, 2001) and named Polycalyx. Specimens of Polycalyx have elongate seeds with apical projections extending from the micropyle and are interpreted as a member of the seed fern family Austrocalyxaceae (see Chapter 14).

MESOZOIC FOLIAGE The Mesozoic is sometimes referred to as the Age of Cycads because of the abundance of isolated foliage that is believed

686

Paleobotany: the biology and evolution of fossil plants

Figure 16.79 Charbeckia macrophylla (Mississippian). Bar  3 cm. (Courtesy M. J. Knaus.)

Figure

16.77 Triphyllopteris

rhomboides

(Mississippian).

Bar  1 cm.

Figure 16.78 Genselia compacta (Mississippian). Bar  1 cm. (Courtesy M. J. Knaus.)

to have been produced by members of the Cycadales or the Bennettitales. The leaves occur as impressions and compressions, varying from a few centimeters in length to over a meter. They are typically once pinnate, although some may be entire. The individual laminar parts are variously termed pinnae, leaf segments, or pinnules. Many leaf types appear to have been rather coriaceous with a thick cuticle. The importance of cuticular features in identifying these leaves was understood as early as 1856 by Bornemann. It was the comprehensive studies of Nathorst (1902), and later of Thomas and Bancroft (1913) and Florin (1933a), however, that were primarily responsible for demonstrating the existence of two major groups of Mesozoic cycadophytes: the Cycadales and Bennettitales. Thomas and Bancroft (1913) established two orders (Nilssoniales and Bennettitales) that correspond rather closely to the Cycadales and Bennettitales, respectively. In distinguishing the two basic foliage types, one of the most important epidermal

CHAPTER 16

Figure 16.80 Cuticle preparation of Nilssoniopteris angus-

tior showing typical syndetocheilic stoma (Triassic). Bar  20 μm. (Courtesy C. Pott.)

characters is the stomatal apparatus. In the nilssonialean type, the two guard cells have their origin from a common mother cell, whereas the subsidiary cells are derived from adjacent epidermal cells. This type of stomatal ontogeny is termed haplocheilic and regarded as plesiomorphic; it is typical of members of the Cycadales (Florin, 1933a). Leaves with haplocheilic stomata also exhibit a relatively thin layer of cutin, especially over the guard cells. Other characteristics that have been useful in identifying cycadalean leaves, but are less consistent, include the irregular orientation of the stomata and straight-walled epidermal cells that are not arranged in rows. The second (apomorphic) pattern of stomatal ontogeny, which is slightly more complex, is termed syndetocheilic and is found in the Bennettitales. In this pattern the two guard cells and the subsidiary cells all have their origin from a common epidermal cell. Syndetocheilic stomata (FIG. 16.80) have a thicker cuticle on the outer and dorsal walls (wall opposite the stoma) of the guard cells. Secondary characteristics include stomata aligned in distinct rows, with the stomatal axes oriented at right angles to veins and sinuous cell walls in the epidermal cells. Harris (1969) noted that one interesting feature of the bennettitalean cuticle is the rather consistent crumbly nature of the thicker, upper cuticle, which often makes it impossible to obtain satisfactory cuticle preparations. This may suggest some basic difference in the biochemistry of the cutin. These two patterns are not unique to cycadophytes; the haplocheilic type is also present in seed ferns, cordaites, conifers, ginkgophytes, Ephedra

LATE PALEOZOIC AND MESOZOIC FOLIAGE

687

(Gnetales, Chapter 19), and some angiosperms, whereas the syndetocheilic pattern occurs in some angiosperms, Gnetum, Welwitschia (both Gnetales), and the bennettitaleans. Syndetocheilic stomata have also been reported in the Devonian lycopsid, Drepanophycus spinaeformis (Stubblefield and Banks, 1978), and a stomatal type resembling syndetocheilic stomata also occurs in some sphenophytes (see Chapter 10). Although the two modes of stomatal development are important at the ordinal level and provide an important systematic character for fossil foliage types, studies with extant plant cuticles indicate that caution must be exercised when using such features exclusively. For example, syndetocheilic stomata have been found on the leaves of some species of the Magnoliaceae and a haplocheilic pattern on the floral parts (Paliwal and Bhandari, 1962). In other plants, both types may appear on the same plant at different stages of development. Following is a brief characterization of some of the more common Mesozoic foliage genera; for some of the widespread forms, data on the geographic distribution are provided as well. ANOMOZAMITES

These leaves are similar to Pterophyllum and are sometimes included in that genus. Harris (1932a, 1969) listed a few epidermal features, including the arrangement of the stomata, that can be used to distinguish the two. Information about the epidermal anatomy of Anomozamites, however, remains incomplete to date (Pott et al., 2007a). The leaf is lanceolate in shape with the lamina divided into segments, which are typically about as broad as they are long (FIG. 16.81). In addition, the midrib of the rachis is partially exposed on the upper surface of the frond. Anomozamites ranges from the Late Triassic to the Early Cretaceous and is included within the Bennettitales (Harris, 1969). A Middle Jurassic Anomozamites (A. haifanggouensis) with associated reproductive structures has been described from the Haifanggou Formation in the Nei Mongol Autonomous Region of China (Zheng et al., 2003). This fossil consists of a dichasial system of twigs bearing Anomozamites-type leaves (up to 10 cm long and 2 cm wide) in the axils of the dichotomies. Closely associated with these leaves are small bractoid leaves and liguliform (strapshaped) to club-shaped structures, 2 cm long and 4–6 mm wide, which are interpreted as microsporophylls. CLADOPHLEBIS

This foliage type was produced by osmundaceous ferns (Chapter 11). The fronds were bipinnate, with alternate

688

Paleobotany: the biology and evolution of fossil plants

Figure 16.82 Cladophlebis. (From Taylor and Taylor, 1993.)

Figure 16.81 Anomozamites thomasii. (From Taylor and

Taylor, 1993.)

pinnae attached at an angle of 70° (FIG. 16.82). The sterile leaves have lanceolate pinnules up to several millimeters long, which are slightly flared at the point of attachment. Specimens from the Upper Jurassic of northeast Japan have closely spaced, ultimate pinnae up to 5.0 cm long (Kimura and Ohana, 1988). Most species have entire margins, although some are toothed. CONIOPTERIS

Fronds of this common Mesozoic fern are at least twice pinnate, with the pinnules often wedge shaped or finely dissected (FIG. 16.83). Coniopteris typically is characterized by considerable sterile–fertile frond dimorphism, as well as dimorphic (aphlebioid) basal pinnules (Harris, 1961a). In addition, individual fertile fronds of Coniopteris have regions containing sterile pinnules. Fertile pinnules bear sori marginally at the ends of lateral veins. Sporangia are typically wedge shaped, and the indusium in most forms is

Figure 16.83 Coniopteris bella. (From Taylor and Taylor,

1993.)

cuplike and often possesses two flaps. In C. bella, the sporangia are small and possess a vertical annulus (Harris, 1961a). Although Coniopteris is not common until the Jurassic, there are rare Triassic records (Kilpper, 1964). Foliage of the Coniopteris type is usually included within the Dicksoniaceae.

CHAPTER 16

LATE PALEOZOIC AND MESOZOIC FOLIAGE

689

Figure 16.84 Ctenis kaneharai. (From Taylor and Taylor,

1993.)

CTENIS

Some of the leaves assigned to this genus exceed 1 m in length and are similar to those of some modern cycads in that leaf segments are attached broadly to a prominent rachis. Segments are long (14 cm) and attached at an acute angle in the middle of the frond, with several parallel veins entering the lamina; some veins may be interconnected (FIG. 16.84). Leaf segments may be marginally toothed or irregularly lobed (Rees and Cleal, 2004). Stomata are confined to the lower surface and are organized into broad zones. DELTOLEPIS AND CYCADOLEPIS

In addition to normal leaves (trophophylls), both Cycadales and Bennettitales produce small, scalelike leaves termed cataphylls. Cycadalean and bennettitalean scale leaves may be similar in gross morphology and thus often can only be distinguished by means of cuticular analysis. One of the more common types of Mesozoic cycadophyte scale leaves is Deltolepis (Harris, 1964, 1969). This form is broadly triangular in outline. Venation consists of numerous parallel bundles that terminate near the margin and apex. Stomata are sunken and haplocheilic. Harris (1964) suggested that Deltolepis may have affinities with the Cycadales. Cycadolepis is another Mesozoic scale-leaf type (Harris, 1964, 1969). These bracts and scales are lanceolate–circular in outline, 1–10 cm long, and interpreted as the sterile bracts of some bennettitalean cones (Harris, 1969; Barale, 1981b). Species have been circumscribed on the basis of various epidermal features, including longitudinal wrinkles on the adaxial surface (Boyd, 2004). There has been some debate as to where these structures were borne, either as bracts subtending the cones or as a series of temporary bracts or scales (Boyd, 2004). Cuticles of C. coriacea and C. involuta from

Figure 16.85 Dictyophyllum bremerense. (From Taylor and

Taylor, 1993.)

the Lower Cretaceous of Santa Cruz Province, Argentina, reveal that these scale leaves were produced by bennettitaleans. The leaves show xeromorphic characters, such as sculptured surfaces, sunken stomata, and thick, outer periclinal walls in the epidermal cells. These features may indicate high temperatures, low humidity, and arid soils, environmental conditions that were perhaps associated with volcanic activity (Villar de Seoane, 2001). DICTYOPHYLLUM

This Triassic–Jurassic foliage type is borne by members of the fern family Dipteridaceae (Oishi and Yamasita, 1936). The petiole dichotomizes into two segments, each of which gives rise to pinnae (Webb, 1982). Pinnae are lanceolate, up to 1 m long, and deeply divided into lateral segments, with the segments joined by a web of lamina. One characteristic feature of this genus is the polygonal shape of the vein meshes (FIG. 16.85). Fertile leaves, referred by some to Thaumatopteris, have scattered sori with one to four sporangia each. The annulus is oblique, and spores are trilete. DICTYOZAMITES

The morphology of this bennettitalean foliage type is reminiscent of Otozamites (Seward, 1903), in that leaf segments are attached to the upper surface of the rachis and are either

690

Paleobotany: the biology and evolution of fossil plants

stomatal morphology (see below). Doratophyllum is similar in gross morphology to Nilssonia but may be distinguished by the attachment of the lamina to the sides of the midrib. One late Paleozoic species, D. jordanicus, has been reported from the Upper Permian of the Dead Sea area (Mustafa, 2003). MACROTAENIOPTERIS

The main characteristic of this leaf type is its size, with some forms exceeding 1 m in length. The genus was introduced by Schimper (1869) for entire-margined, strap-shaped leaves previously accommodated in Taeniopteris. Schimper assigned the genus Macrotaeniopteris to the Marattiales, but subsequent studies (Florin, 1933a) revealed that some forms, including specimens described as M. gigantea, represent cycadalean foliage. Based on cuticular analyses, Pott et al. (2007a) demonstrated that fossils previously referred to Macrotaeniopteris from the Carnian of Lunz (Austria) represent bennettitalean foliage of the morphogenus Nilssoniopteris. MATONIDIUM

Figure 16.86 Dictyozamites hawellii. (From Taylor and Taylor,

1993.)

sessile or borne on a short stalk. Foliage segments often overlap (FIG. 16.86), and the segment bases are usually slightly auriculate. The genus may be distinguished by the anastomosing pattern of the veins (Schweitzer and Kirchner, 2003). Cuticle ultrastructure of bennettitalean foliage from the Cretaceous of Argentina suggests that Dictyozamites may be related to Otozamites and Zamites (Villar de Seoane, 2003). Anatomically preserved Dictyozamites rachides have been described from India (Bose and Kasat, 1972a; Bose and Zeba-Bano, 1978). The rachis of D. falcatus has a hypodermis composed of 6–7 layers of thick-walled cells; in cross section, the rachis exhibits 25–30 collateral vascular bundles that are arranged in a double series in the form of a U. Each bundle is surrounded by an incomplete layer of mechanical tissue. Arrangement of rachial vascular bundles in D. indicus is similar to that seen in D. falcatus, but the former species apparently lacks a distinct hypodermis. DORATOPHYLLUM

This morphotaxon is used for strap-shaped, Taeniopterislike leaves from the Mesozoic that are characterized by non-bennettitalean stomatal morphology (Harris, 1932b); Nilssoniopteris is a similar leaf type with bennettitalean

As the generic name implies, this foliage type is placed within the filicalean fern family Matoniaceae (Tidwell, 1975). The genus is quite similar to Phlebopteris but differs in having an indusium. Pinnae may be numerous (up to 20), each about 20 cm long (FIG. 16.87). Pinnules are broad, entire, and typically 1–2 cm long. Matonidium is a common element of some Jurassic and Cretaceous floras. MESODESCOLEA

This Early Cretaceous leaf type is known from Argentina and is believed to be bipinnate (Archangelsky, 1963). Pinnae are about 6.0 cm long and oppositely arranged. Pinnules are subopposite to alternate and decurrent at the base (FIG. 16.88). The pinnules are hypostomatic, with monocyclic stomatal apparati, and have hair bases present on the lower surface (Villar de Seoane, 2005a). In many features, the leaf appears similar to the modern genus Stangeria, and its affinities are considered to be with the cycads. NILSSONIA

Leaves of Nilssonia are linear to oblanceolate, with the lamina occasionally subdivided into leaf segments of various sizes and shapes (FIG. 16.89). The lamina is attached to the upper surface of the rachis and may cover it. Veins are simple and depart from the midvein at right angles (FIG. 16.90); resin bodies may be present between the veins. Leaves are hypostomatic or amphistomatic and papillae and trichomes may be present on the leaf surface. Haplocheilic stomata

CHAPTER 16

LATE PALEOZOIC AND MESOZOIC FOLIAGE

691

Figure 16.88 Mesodescolea plicata. (From Taylor and Taylor,

1993.) Figure 16.87 Portion of a Matonidium leaf. (From Taylor and

Taylor, 1993.)

suggest that Nilssonia foliage was produced by members of the Cycadales. A specimen of N. acuminata from the Liassic (Lower Jurassic) of southern Germany shows that Nilssonia leaves were arranged in dense clusters on the distal portion of the shoots (Schmeissner and Hauptmann, 1998). The earliest occurrences of Nilssonia have been recorded from the Upper Triassic of Europe (Broglia Loriga et al., 2002; Pott et al., 2007b) and North America (Ash, 2001). The genus is common and widely distributed in the Jurassic as well, including, for example, Europe (Nathorst, 1879, 1909; Lundblad, 1950b; Harris, 1964; J. Kvacˇek, 1995; Kelber, 1998; Watson and Cusack, 2005), Iran and Afghanistan (Sadovnikov, 1989; Schweitzer et al., 2000), India (Sahni and Rao, 1931, 1934), Japan (Yabe, 1925), and Greenland (Harris, 1932a, 1946; Boyd, 2000). An anatomically preserved Jurassic leaf assignable to Nilssonia has been described from northwestern Scotland (Dower et al., 2004). Stomata preserved in this fossil superficially resemble those seen in the extant cycad Macrozamia. The anatomy of the lamina is characterized by a series of oval cavities that are lined with dense cells and occur

between the veins on the upper surface of the lamina and within the spongy mesophyll layer. The cavities either represent secretory cavities, large storage cells, or idioblasts. NILSSONIOPTERIS

Nilssoniopteris foliage is strap shaped and petiolate (FIG. 16.91), with an undivided, lobed, or completely dissected lamina that is attached laterally or to the upper surface of the rachis. The veins are simple or forked and stomata are syndetocheilic, suggesting affinities with the Bennettitales (similar strap-shaped leaves with haplocheilic stomata are assigned to Doratophyllum); the anticlinal cell walls in the epidermis are usually sinuous but may, in some Triassic forms, also be straight. Typical representatives of Nilssoniopteris are entire margined (FIG. 16.92) and Taeniopteris like or lobed; fully pinnate leaves of similar outline have traditionally been assigned to Anomozamites. The epidermal anatomy of typical representatives of Anomozamites has not been documented to date, however, and there are a number of fossils that are morphologically intermediate between Nilssoniopteris and Anomozamites. Boyd (2000) and Pott et al. (2007a) proposed to include all these intermediate leaves within

692

Paleobotany: the biology and evolution of fossil plants

Figure 16.90 Nilssonia tenuicaulis. (From Taylor and Taylor,

1993.)

16.89 Nilssonia sp., cycadalean leaf (Jurassic). Bar  2 cm. (Courtesy BSPG.)

Figure

Nilssoniopteris if they are either fully pinnate (FIG. 16.93) or entire margined and strap shaped; all must also exhibit syndetocheilic stomata. Nilssoniopteris is a typical element of many Jurassic floras known from localities in, for example, Sweden, Great Britain

(Nathorst, 1909; Lundblad, 1950b; Harris, 1969; Watson and Sincock, 1992), Iran and Afghanistan (Sadovnikov, 1989; Schweitzer and Kirchner, 2003), China (Barale et al., 1998), and Greenland (Harris, 1932; Boyd, 2000). The fossil record of Nilssoniopteris from North America is generally poor (Ash, 1989). Taeniopterid foliage fossils resembling Nilssoniopteris have also been reported from the Triassic and Jurassic of the Southern Hemisphere (Anderson and Anderson, 1989; Gnaedinger and Herbst, 2004; Cantrill and Hunter, 2005). Epidermal anatomy, however, has not been

CHAPTER 16

LATE PALEOZOIC AND MESOZOIC FOLIAGE

693

Figure 16.92 Leaf of Nilssoniopteris angustior (Triassic).

Bar  2 cm. (Courtesy C. Pott.)

Figure 16.91 Nilssoniopteris major. (From Taylor and Taylor,

1993.)

documented for any of these, and thus the affinities remain speculative. OTOZAMITES

Leaves of Otozamites are once pinnate (FIG. 16.94); leaf segments have constricted, asymmetrical bases and are attached to the upper surface of the rachis (Watson and Sincock, 1992). They are broadly oval, with the anterior lobe pronounced. Veins entering leaf segments typically spread

out (FIG. 16.95), but may be almost parallel in some species. Epidermal cells of the adaxial surface have sinuous anticlinal walls; abaxial cells are straight or sinuous. Leaves are usually hypostomatic, and the stomata are syndetocheilic, often sunken, and surrounded by papillae. The stomatal complexes of O. kerae and O. takahashii from the Upper Cretaceous of Japan are unusual in that the guard cells and subsidiary cells are sunken and located in a stomatal pit covered by a dome- or sac-shaped roof with a distal, rounded, or stellate opening that is formed by 10–12 roof cells (Ohana and Kimura, 1991). Similar roofed stomatal apparati have been observed in other Otozamites species by Harris (1969) and Barale (1987) and are regarded as unique to some species

694

Paleobotany: the biology and evolution of fossil plants

Figure 16.93 Leaf of Nilssoniopteris lunzensis (Triassic).

Bar  1 cm. (Courtesy C. Pott.)

in the genus. Many Otozamites species produce papillae or trichomes on the abaxial leaf surface (Kilpper, 1968; Villar de Seoane, 2001). Anatomically preserved Otozamites leaves have been described from the Middle Jurassic of Bearreraig Bay (Skye) in northwestern Scotland (Dower et al., 2004) and the Upper Cretaceous of Hokkaido, Japan (Ohana and Kimura, 1991). The rachial vascular system of O. kerae (Japan) is composed of 12–13 endarch and collateral bundles arranged in a double series in the form of a U. The xylem in the outer and inner series of the U face in opposite directions. Leaf segments have a non-stomatiferous upper and stomatiferous lower epidermis, upper and lower hypodermal layers

Figure 16.94 Otozamites brevifolius frond (Rhaeto-Liassic). Bar  2 cm. (Courtesy BSPG.)

composed of thick-walled cells, one or two layers of palisade parenchyma, and spongy mesophyll with large intercellular spaces. Vascular bundles are integrated in the spongy mesophyll. Intercalated between the adaxial hypodermis and the vascular bundles is a wedge-shaped segment of

CHAPTER 16

LATE PALEOZOIC AND MESOZOIC FOLIAGE

695

Figure 16.95 Otozamites thomasii. (From Taylor and Taylor,

1993.)

mechanical tissue consisting of large sclerenchymatous cells. In O. mortonii (Scotland), the rachial vascular system is composed of several bundles that are almost completely joined to form an outer, horseshoe-shaped (omega-shaped) amphivasal and inner, cylindrical amphicribal bundle. PACHYPTERIS, KOMLOPTERIS, AND THINNFELDIA

Pachypteris is a foliage type known primarily from the Upper Triassic–Jurassic of the Northern Hemisphere (Popa, 2000) but which has also been documented from the Jurassic (Townrow, 1965) and Cretaceous of Gondwana (Cantrill, 2000b). Leaves are pinnate or bipinnate; each pinnule is decurrent at the base and has a blunt apex (FIG. 16.96). Venation consists of a prominent vein that produces inconspicuous laterals. Most species are hypostomatic with sunken, haplocheilic stomata. Pachypteris is referred to the seed fern order Corystospermales (Chapter 15). Based on the excellent preservation of cuticles in many species, Pachypteris has been used in several TEM studies of fossil plant cuticle ultrastructure (Labe and Barale, 1996; Bajpai and Maheshwari, 2000). A recent detailed ultrastructural analysis of the cuticles of P. gradinarui from the Jurassic of Romania (Guignard et al., 2004) has been used

Figure 16.96 Pachypteris papillosa. (From Taylor and Taylor,

1993.)

to develop a dichotomous key for cuticle types observed in that species. Komlopteris is a segregate of Pachypteris that is based on differences in leaf shape, pinnule size, distribution of stomata on the abaxial pinnule surface, and stomatal morphology (Barbacka, 1994). The most important delimiting feature is

696

Paleobotany: the biology and evolution of fossil plants

that the abaxial stomata of Komlopteris are arranged in distinct intercostal fields, whereas Pachypteris has stomata scattered over the entire lower leaf surface. The genus Thinnfeldia was historically used for Dicroidium-like foliage from the Northern Hemisphere (Gothan, 1912; Antevs, 1914a; Thomas, 1933; Jones and de Jersey, 1947). The holotype for the type species, T. rhomboidalis, however, from the Jurassic of Romania, was later shown to be a Pachypteris (Doludenko, 1971; Doludenko et al., 1998), so that today Thinnfeldia is generally regarded as a synonym of Pachypteris (Popa, 2000). However, Kirchner and Müller (1992) described the corystosperm reproductive structures Umkomasia franconica and Pteruchus septentrionalis from the Liassic (Early Jurassic) of Franconia, Germany and these are associated exclusively with Thinnfeldia foliage. Based on the close association in the bedding plane, along with similarities in morphology and epidermal anatomy, they suggest that the foliage and reproductive structures were produced by the same plant, and thus retain the generic name Thinnfeldia, since Pachypteris has been found associated with the pollen organ Pteroma (Harris, 1964), which was described as slightly different from Pteruchus.

Figure 16.97 Pseudoctenis lanei. (From Taylor and Taylor,

1993.) PHLEBOPTERIS

Phlebopteris fronds have a basal dichotomy with each division bearing 20 pinnae. Pinnules are borne at right angles, widely spaced, many times longer than wide, and have obtusely pointed apices (Ash, 1991). The pinnule venation consists of a prominent central vein that divides to form a reticulate pattern. Abaxial sori consist of about 14 sporangia (Harris, 1961a). Phlebopteris foliage was produced by matoniaceous ferns (Matoniaceae) and can be traced from the Late Triassic into the Early Cretaceous. PSEUDOCTENIS

Pseudoctenis morphologically resembles Ctenis, but the two taxa can be distinguished based on the occurrence of anastomoses in the venation of the latter taxon (Seward, 1911). Leaves are hypostomatic with scattered stomata. In P. lanei from the Jurassic of Yorkshire, the leaf is at least 1 m long and up to 30.0 cm wide (FIG. 16.97) (Harris, 1964). Assignment of Pseudoctenis to the Cycadales is based primarily on epidermal anatomy (Harris, 1932). The genus is quite common in the Rhaetian (uppermost Triassic) and Jurassic of Europe (Seward, 1911, 1917; Harris, 1932, 1964; Lundblad, 1950b; Schweitzer and Kirchner, 1998). Pseudoctenis cornelii (FIG. 16.98), a species from the Carnian (lowermost Upper

Figure 16.98 Pseudoctenis cornelii (Triassic). Bar  2 cm.

(Courtesy C. Pott.)

Triassic) of Lunz (Austria), represents the earliest unequivocal evidence for the genus from Europe (Pott et al., 2007c). This species differs from others in having slightly undulating epidermal anticlinal cell walls. One specimen from the Upper Permian of England was assigned to Pseudoctenis (Stoneley, 1958) but a lack of adequate information on epidermal anatomy makes the systematic placement uncertain. Although the oldest Northern Hemisphere Pseudoctenis fossils come from the upper Carnian, the earliest evidence in

CHAPTER 16

LATE PALEOZOIC AND MESOZOIC FOLIAGE

697

the Southern Hemisphere is lower Carnian in age (Anderson and Anderson, 1989). Reports of Pseudoctenis from other early Late Triassic floras (Leppe and Moisan, 2003; Nielsen, 2005) do not yield cuticles, and thus remain inconclusive. In North America, Pseudoctenis-type foliage is borne by the Late Triassic cycad Leptocycas gracilis (Chapter 17). PSEUDOCYCAS

Leaves of Pseudocycas are once pinnate with narrow leaf segments. Each segment is attached broadly at the base with the posterior margin slightly decurrent (Boyd, 1998). The segments are characterized by a so-called double midrib composed of two parallel, abaxial rims that border a central furrow containing stomata (Halle, 1915). Watson and Sincock (1992) interpreted this feature as indicating the xerophytic habitat of the genus. The syndetocheilic morphology of the stomatal apparati indicates that this foliage type belongs to the Bennettitales (Nathorst, 1907). PTEROPHYLLUM

The morphogenus Pterophyllum was originally defined for bipinnate, cycad-like foliage fossils (Brongniart, 1828) but is today used exclusively for bennettitalean leaves. Pterophyllum includes pinnate leaves (FIG. 16.99) with almost parallel-sided pinnae that are elongate, laterally inserted, and broadly attached. The fronds have a robust striate rachis and syndetocheilic stomata (Thomas, 1930; Florin, 1933a; Harris, 1969; Watson and Sincock, 1992; Van Konijnenburg-Van Cittert et al., 2001). The anticlinal epidermal cell walls tend to be straight and not sinuous like those in typical bennettitalean leaves; several forms are amphistomatic. In P. filicoides, the type species of the genus, from the Carnian of Austria, leaves range from 10 to more than 50 cm long and 4–20 cm wide (FIG. 16.100) (Pott et al., 2007d), thus indicating that considerable variability was a characteristic feature of the species and perhaps of bennettitaleans in general. Pterophyllum filicoides displays circinate vernation in the form of inwardly coiled leaf segments (FIG. 17.59), which are similar to immature leaf segments in extant Cycas species (Pott and Krings, 2007). A specimen of P. bavieri from the Rhaetian of Iran illustrated by Schweitzer and Kirchner (2003) indicates that Pterophyllum leaves were either compound and composed of pinnate units attached to a massive rachis, or were simple, pinnate, and arranged in groups of a few leaves or in clusters at the tips of narrow twigs (FIG. 17.64). The earliest bona fide Pterophyllum fossils come from the Carnian (Upper Triassic) of central Europe (Pott

Figure 16.99 Pterophyllum brevipenne (Triassic). Bar  1 cm.

(Courtesy BSPG.)

et al., 2007d), and the latest occurrence is recorded as Early Cretaceous (Kelber, 1998); the genus was most widespread during the late Late Triassic and Early Jurassic. The geographical distribution of Pterophyllum is largely restricted to North America, Europe, and Asia. Assignment of late Paleozoic foliage fossils to Pterophyllum, including P. cutelliforme (Sze, 1936; Sun, 2006), P. cotteanum (Geinitz, 1873; Barthel, 1976b), P. blechnoides (Geinitz, 1873), P. samchokense (Kawasaki, 1931), and P. daihoense (Kawasaki, 1931, 1934; Sun, 2006) from the Pennsylvanian and Permian of China, Korea, and Germany, remains problematic since cuticles are not preserved or have not been studied. Pterophyllum grandeuryi and P. fayolii from the Stephanian

698

Paleobotany: the biology and evolution of fossil plants

Figure 16.101 Ptilophyllum pectinoides. (From Taylor and

Taylor, 1993.)

Figure 16.100 Pterophyllum filicoides, frond (Triassic). Bar  3 cm. (Courtesy BSPG.)

(Upper Pennsylvanian) of France (Zeiller, 1906) resemble Pseudoctenis with regard to morphology. The abaxial cuticles of Pterophyllum grandeuryi, however, display haplocheilic stomata, a character that occurs in cycads, but also other plant groups such as seed ferns, and thus is not reliable. PTILOPHYLLUM

The leaf segments of Ptilophyllum (FIG. 16.101) are attached to the upper surface of the rachis and vary in their angle

of attachment (Watson and Sincock, 1992). They are characterized by an asymmetrical base composed of a decurrent proximal and constricted distal margin. Veins arise from the entire region of attachment and run parallel to the margin; forking occurs near the pinnule apex. Epidermal cells usually have sinuous anticlinal walls, and stomata are generally restricted to the abaxial side and display syndetocheilic morphology (Harris, 1969). Epidermal features are critical in the discrimination of species within Ptilophyllum. Sukh-Dev and Zeba-Bano (1977) provided a dichotomous key based primarily on epidermal characters for numerous Ptilophyllum species from India and several other countries. Anatomically preserved foliage assignable to Ptilophyllum has been described from India by several authors (Rao and

CHAPTER 16

Achuthan, 1967; Sharma, 1967; Bose and Kasat, 1972b). The rachis of P. amarjolense has a hypodermis of 6–7 cell layers and in cross section exhibits 11 collateral vascular bundles arranged in a kidney-shaped strand composed of five abaxial, two lateral, and four adaxial bundles. Each bundle is surrounded by an incomplete layer of mechanical tissue. Radial walls of the xylem tracheids show scalariform thickenings. Leaf anatomy is relatively simple and includes an epidermis, a single layer of thick-walled hypodermal cells, a one to two layered palisade parenchyma, and spongy mesophyll. Vascular bundles in the leaf segments are located above the lower hypodermal layer. The rachial stele of P. sahnii consists of 23–32 collateral vascular bundles arranged in a double series in the form of a U facing the adaxial surface.

LATE PALEOZOIC AND MESOZOIC FOLIAGE

699

Figure 16.102 Ruflorina sierra (Cretaceous). Bar  1 cm.

PTILOZAMITES

Leaves assignable to Ptilozamites are once pinnate, broadly or narrowly lanceolate to linear, with stout, often forked rachides. Pinnae are square, rhomboidal, linear, asymmetrically triangular or falcate, and are densely spaced or imbricate. They are completely attached laterally to the rachis, but are not decurrent. Numerous veins arise from the rachis and fork at least once in their course to the margin. Epidermal cells are often papillate and have straight or slightly sinuous walls. Segments are hypostomatic or amphistomatic and have sunken guard cells surrounded by irregularly shaped subsidiary cells that form a ringlike thickening (Antevs, 1914b). Ptilozamites has been attributed to various groups of plants, including Cycadales and Caytoniales, but is today commonly regarded as a pteridospermous morphotaxon with uncertain affinities (Kustatscher and Van Konijnenburg-Van Cittert, 2007). The genus was initially considered typical for the Rhaeto–Liassic (Late Triassic–Early Jurassic) of Sweden (Nathorst, 1886) but is today known to occur widely in same age rocks throughout the Northern Hemisphere and has also been reported from various Early and Middle Triassic localities of the southern Alps (Wachtler and Van Konijnenburg-Van Cittert, 2000a, b; Kustatscher and Van Konijnenburg-Van Cittert, 2005). Schweitzer and Kirchner (1998) suggested a close relationship between Ptilozamites and Ctenozamites based on the fact that the leaves are often terminally forked in both taxa; Ctenozamites leaves, however, are bipinnate (see Harris, 1964, for details on Ctenozamites). RUFLORINIA

This tripinnate leaf is characterized by alternate to subopposite, decurrent pinnae bearing obliquely attached pinnules

Figure 16.103 Ruflorinia sierra. (From Taylor and Taylor,

1993.)

(FIG. 16.102) and is known from the Early Cretaceous of Argentina (Archangelsky, 1963). The pinnules of R. sierra range up to 4.0 mm long and 1.5 mm wide (FIG. 16.103). Stomata are slightly sunken and occur only on the lower surface of the pinnule; surrounding each stoma is a number of papillae. Taylor and Archangelsky (1985) have demonstrated the attachment of the cupulate organ Ktalenia to Ruflorinia foliage, thus establishing the genus as a probable seed fern.

700

Paleobotany: the biology and evolution of fossil plants

Figure 16.104 Wingatea. (From Taylor and Taylor, 1993.)

TAENIOZAMITES

Taeniozamites was introduced by Harris (1932b) for entiremargined, Taeniopteris-like leaves from the Mesozoic that are characterized by bennettitalean epidermal anatomy; Doratophyllum is used for the equivalent foliage with nonbennettitalean epidermal features. Taeniozamites is today regarded as a synonym of Nilssoniopteris (Cleal et al., 2006). TICOA

Foliage placed in the genus Ticoa (Cretaceous of Argentina) consists of large, tripinnate fronds with a stout primary rachis. Primary and secondary pinnae are attached at nearly right angles; pinnules are morphologically similar to those of Pecopteris, with a well-developed vein entering the base of each pinnule (Archangelsky, 1963). The upper cuticle is generally thicker than the lower one. Stomata are sunken and protected by large or small epistomatal chambers; they are rare or absent on the adaxial surface but abundant on the abaxial side (Villar de Seoane, 2005a). The organization of the stomatal apparatus is more similar to cycads than to Mesozoic seed ferns. Trichomes may be present on one or both surfaces of the pinnules. The genus has also been reported from the Late Jurassic–Early Cretaceous of Hope Bay, Antarctica (Gee, 1989). WINGATEA

This Triassic genus, which is included within the Gleicheniaceae, has been reported from the Chinle Formation of western North America (Ash, 1969; Tidwell, 1975; Litwin, 1985). Fronds are relatively large (Fig. 16.50) and atleast tripinnate, with secondary pinnae divided into delicate pinnules with pointed tips (FIG. 16.104). Each

Figure 16.105 Zamites guiniae. (From Taylor and Taylor,

1993.)

pinnule bears a single sorus (Ash, 1969) that is composed of 7–20 sporangia and has been suggested to be indusiate (Litwin, 1985). The sporangia are attached to the receptacle by a short stalk and have a complete, oblique annulus lacking a stomium. Spores are triangular. YABEIELLA

This taxon is used for elongate, Taeniopteris-like leaves from the Triassic that are characterized by the presence of a marginal vein. The leaf is strap shaped, entire margined, and the lateral veins unite with the marginal vein. Yabeiella is thought to have been produced by bennettitaleans (Oishi, 1931). Based on the consistent co-occurrence of Yabeiella leaves with Fraxinopsis seeds, Axsmith et al. (1997) suggested that both structures were produced by the same plant.

CHAPTER 16

LATE PALEOZOIC AND MESOZOIC FOLIAGE

701

Figure 16.107 Zamites sp. (Jurassic). Bar  2 cm.

Figure 16.106 Zamites oaxacensis (Jurassic). Bar  4 cm.

ZAMITES

Pinnae of this bennettitalean foliage type are linear or lanceolate and attached to the upper surface of the rachis (FIGS. 16.105, 16.106). The pinna base is greatly constricted (FIG. 16.107), and the venation pattern includes parallel or slightly divergent veins; the points of pinna attachment are

sometimes slightly thickened and elevated (Schweitzer and Kirchner, 2003). Stomata are confined to the upper surface. Zamites can be distinguished morphologically from similar genera (e.g., Ptilophyllum) by the symmetrical nature of the pinna bases. Rees and Cleal (2004) hypothesized that the distinction between Zamites and Ptilophyllum is artificial and that both foliage types, along with Williamsonia and Weltrichia fructifications (Chapter 17), probably belong to a single natural genus.

This page intentionally left blank

17 CYCADOPHYTES CYCADALES .................................................................................. 703

Discussion: Cycad Evolution ............................................................721

Leaves and Petioles ...........................................................................706

BENNETTITALES...........................................................................722

Stems .................................................................................................707

Cycadeoidaceae.................................................................................725

Paleozoic Reproductive Structures ...................................................709

Williamsoniaceae ..............................................................................732

Triassic Cycads .................................................................................715

Discussion: Bennettitales .................................................................. 739

Jurassic Cycads .................................................................................718 Pollination Biology ...........................................................................721

Cycads are to the vegetable kingdom what Dinosaurs are to the animal, each representing the culmination in Mesozoic times of the ruling Dynasties in the life of their age. Lester Ward Traditional classifications include the true cycads, members of the order Cycadales, and the Bennettitales, an extinct group of plants, in an informal group termed the cycadophytes (see Watson and Sincock, 1992; Norstog and Nicholls, 1997). In some classifications the cycadophytes are included with the pteridosperms in the class Cycadopsida (Meyen, 1987), whereas in others they are represented as a basal order of gymnosperms that is sister to all other seed plants (Stevenson, 1990). Compressed fossil foliage of these two orders is often difficult to distinguish based on macromorphology, and as a result, morphogenera, for example, Nilssonia and Pterophyllum, have been erected. In some instances, a foliage genus may well include both cycadalean and bennettitalean foliage (see Chapter 16). In addition to normal leaves (trophophylls), both Cycadales and Bennettitales produce small, scalelike leaves termed cataphylls. Cycadalean and bennettitalean scale leaves may be very similar in gross morphology, and thus often can only be distinguished by means of cuticular analysis. Anatomically preserved stems have been described from both groups, and the Bennettitales are known in some details from studies of petrified reproductive organs. Regardless of how they are classified, the Cycadales are an ancient lineage of gymnosperms that can be traced back to the Paleozoic, and whose

modern members present a unique combination of features among extant seed plants (Brenner et al., 2003), whereas the Bennettitales are restricted to the Mesozoic.

Higher taxa in this chapter:

Cycadales Bennettitales Cycadeoidaceae Williamsoniaceae

CYCADALES The Cycadales includes both living and fossil members and can be traced back to the Pennsylvanian (Norstog and Nicholls, 1997; Pant, 2002) (FIG. 17.1). Fossil evidence indicates that the group has remained relatively unchanged morphologically for a long time. Today the group contains 12 genera in three families (according to Chaw et al., 2005) and 300 species that are distributed in tropical and subtropical regions of the Americas, Africa, southeast Asia, and Australia (Hill, 2004). Some extant Cycadales are characterized by tuberous, short, squat trunks, whereas others

703

704

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 17.3 Cross section showing the offset (O) extending from

the stem of Antarcticycas schopfii (above) (Triassic). Bar  1.0 mm.

Figure 17.1

Divya Darshan Pant.

Figure 17.4 Position of cycad leaves and pollen cones (Extant). (Courtesy P. Ryberg.)

Figure 17.2 Extant Cycas with numerous offsets arising from

stem. Bar  20.0 cm.

resemble tree ferns or palms. Most of the arborescent types are unbranched, with some, such as Macrozamia, reaching a height of 18 m. On the surface of the stems are persistent leaf bases surrounded by reduced, scalelike cataphylls; some species bear offsets, vegetative reproduction structures, on their stems (FIGS. 17.2, 17.3). Distally, cycads exhibit a crown of pinnately or bipinnately (in Bowenia) compound leaves (FIG. 17.4), a feature, among others, that distinguishes them from all other living gymnosperms (FIG. 17.5). Most cycad leaves show circinate vernation. The vascular system is an eustele with a massive pith and cortex containing mucilage canals. The narrow cylinder of secondary xylem is

CHAPTER 17

CYCADOPHYTES

705

Ovule cone Pollen cone (on another cycad plant) Microsporophyll and pollen sacs

Megasporophyll and tow ovules

Young leaf Ovule Stem Seed coat

Meiosis

Pollen sac

Root Meiosis

Seed coat

Embryo Gametophyte

Integument

Pollen grain

Megaspore Nucellus

Archegonium Haustorial pollen tube

Gametophyte

Pollen grain

Micropyle Pollen grain

Figure 17.5

Archegonium

Life history of cycad. (From Taylor and Taylor, 1993.)

of the manoxylic type with relatively thin-walled tracheids and wide vascular rays; a synapomorphy of the group is the presence of girdling traces in the cortex. After leaf traces are given off by the stele, they extend horizontally around the axis before vascularizing the petioles. All extant cycads possess both normal roots and apogeotropic, above-ground roots

or coralloid roots, which are generally infected by cyanobacterial symbionts (Milindasuta, 1975). The cyanobacteria occur in the coralloid roots as intercellular endophytes in a distinct zone composed of elongated cortical cells (FIG. 17.6), although for some cycad species, intracellular growth of the cyanobacteria has also been reported (Rai et al., 2000;

706

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 17.6 Cross section of Cycas coralloid root showing zone of cyanobacteria (arrow) (Extant). Bar  650 μm.

Costa and Lindblad 2002; Bergmann et al., 2007). Infection of the coralloid roots may occur at an early stage of root development (i.e., precoralloid), but the infection process is still not completely understood (Lindblad and Costa, 2002). Cycads are dioecious, with the seeds borne on megasporophylls which represent reduced, modified leaves; these are aggregated into cones or strobili (FIG. 17.7). One exception is Cycas, in which the megasporophylls are borne in the same phyllotactic spiral as the leaves and are loosely aggregated at the apex of the stem and borne in the same phyllotactic spiral as the leaves. Seeds in Cycas are produced along the lateral edges of the petiole, with the terminal expanded portion of the megasporophyll divided pinnately in some species. In other genera, the elongated, distal end of the megasporophyll is reduced or missing. Here the distal area is tangentially expanded to form a peltate structure, with two large ovules borne on the inner surface. The outer layer of the integument, the sarcotesta, is fleshy and may serve as an attractant for seed dispersal. Pollen cones are small and compact, with the pollen sacs attached to the abaxial (lower) surface of the microsporophylls. The number of pollen sacs per microsporophyll may range from 1000 to 20. Pollen grains are boat-shaped, thin walled, alveolate, and typically devoid of any ornamentation. In those species studied in detail, pollination is now known to be entomophilous. Male gametes are flagellated, a plesiomorphic character that cycads share with Ginkgo biloba (Chapter 18) and some pteridosperms (Chapter 14). Modern cycads produce some secondary compounds that serve to discourage predation. The reader is referred to Stevenson (1992) for a formal classification of the extant members. For additional details on extant

Figure 17.7

Extant cycad cone. Bar  4.0 cm. (Courtesy

P. Ryberg.)

cycads, see the comprehensive treatment of cycads in The Biology of the Cycads (Norstog and Nicholls, 1997) and The Cycads (Whitelock, 2002). LEAVES AND PETIOLES

There have been numerous leaves described from the late Paleozoic and Mesozoic that have presumed affinities with the Cycadales (see Chapter 16). In most instances these are known from impression–compression remains, such as Mesodescolea (Artabe and Archangelsky, 1992), Nilssonia (FIG. 17.8) (Harris, 1964), Kurtziana (Artabe et al., 1991), and Yixianophyllum (Zheng et al., 2005), but a few have been reported from structurally preserved specimens. For example, Cycadinorachis from the Lower Cretaceous of the Rajmahal Hills (formerly Jurassic; Sharma, 1973a) and Cretocycas from the Upper Cretaceous of Hokkaido, Japan (Nishida et al., 1996), are interpreted as cycadalean. Pterostoma is a morphogenus for large fronds with simple pinnae contracted near the base and with simple or dichotomizing veins (R. Hill, 1980). This Eocene cycad leaf has sinuous epidermal cell walls.

CHAPTER 17

CYCADOPHYTES

707

Figure 17.8 Pinna of Nilssonia sturii (Triassic). Bar  4.0 cm. (Courtesy C. Pott.)

Lyssoxylon grigsbyi tracheids showing bordered pits (Triassic). Bar  35 μm. (From Gould, 1971.)

Figure 17.10

Figure 17.9 Cross section of Lyssoxylon grigsbyi stem

(Triassic). Bar  2.0 cm. STEMS

There are several permineralized stems that are interpreted as cycads based on their anatomy. One of these is Lyssoxylon, a stem described from the Upper Triassic Chinle Formation of New Mexico and Arizona. Lyssoxylon grigsbyi was initially described as a williamsonian (Williamsoniaceae) axis (Daugherty, 1941), but the discovery of additional specimens demonstrated that the anatomy was more cycadalean (Gould,

1971). The stem consists of a central pith surrounded by a cylinder of manoxylic wood containing a few growth rings (FIG. 17.9). The tracheids are square in cross section and have biseriate alternate pits on the radial walls (FIG. 17.10). Extending through the wood are vascular rays with some ray cells containing structures interpreted as nuclei. Bicelled hairs arise from the epidermis and form a dense mat around the petiole bases. One characteristic that strengthens the cycadalean affinities of L. grigsbyi is the girdling pattern of the leaf traces. Although evidence suggests that Triassic and Jurassic cycad stems were relatively slender and often branched (Delevoryas and Hope, 1976) (FIG. 17.11), Lyssoxylon appears to be an exception. Although it was several times larger in diameter than other fossil cycad stems, L. grigsbyi is still considerably smaller than most extant cycads. There is more secondary xylem and less ground tissue, including cortex, than in extant cycads. If the modern

708

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 17.11 Theodore Delevoryas.

cycads are descendants of the Paleozoic seed ferns, as some have suggested, then there has been a gradual reduction in wood accompanied by an increase in ground tissue production in the stem. Vladiloxylon is a structurally preserved stem from the Triassic of Chile that is similar to Lyssoxylon, but with more secondary xylem (Lutz et al., 2003). Another silicified Triassic cycad stem is Michelilloa waltonii from the Ischigualasto Formation of northern Argentina (Archangelsky and Brett, 1963). Although the entire stem is not preserved, it is estimated to have been 10 cm in diameter. In the central portion is a massive pith with scattered mucilage canals. The secondary xylem consists of tracheids with multiseriate pits and numerous vascular rays. The structure of the leaf gap and the presence of long filamentous hairs on the surface of the stem are features that M. waltonii shares with a number of extant cycads. Charmorgia dijolii is an even larger stem from the Chinle Formation in the Petrified Forest National Park, USA (Ash, 1985a). The stem is 30 cm in diameter, with persistent, rhombic leaf and cataphyll bases. Primary xylem is endarch and secondary xylem tracheids have bordered pits on their radial walls. Up to 40 vascular bundles arranged in an omega shape enter the leaf base. Although C. dijolii shares a number of features with modern cycads, including the characteristic girdling leaf traces, there is no single taxon to which it can be closely related. Menucoa and Bororoa are two anatomically preserved fossil cycads from the Cenozoic of Río Negro and Chubut provinces in Argentina, respectively. Menucoa cazaui is up to

Figure 17.12 Menucoa cazaui, proximal stem portion showing leaf bases (Paleogene). Bar  2.0 cm. (Courtesy BSPG.)

60 cm in diameter and characterized by persistent leaf bases (FIG. 17.12) and numerous mucilage canals (Petriella, 1969). The wood is polyxylic, with broad vascular rays. Stems of Bororoa andreisii are slightly smaller (30–45 cm in diameter) and have mucilage canals associated with the wide vascular rays (Petriella, 1972). In addition to the general size and shape of the trunk, Bororoa differs from Menucoa in having medullary vascular strands that do not anastomose in the pith. Both morphogenera are related to extant cycads based on anatomical features (Artabe and Stevenson, 1999). A number of new taxa of cycads have been described from the Late Cretaceous of Patagonia, Argentina (Artabe et al., 2004, 2005). Worsdellia and Brunoa are both polyxylic and included in the subfamily of Encephalartoideae (Zamiaceae) which includes modern Encephalartos (Artabe et al., 2004). Worsdellia bonettiae has anastomosing medullary vascular bundles and Brunoa santarrosensis exhibits cone domes. Since cones are borne terminally in the cycads, after senescence of the cones, some cycads will continue growth by

CHAPTER 17

CYCADOPHYTES

709

Figure 17.14 Reconstruction of the megasporophyll of Spermopteris coriacea with attached seeds. (From Cridland and Morris, 1960.)

Figure 17.13 Suggested reconstruction of Nilssoniocladus nipponensis. (From Kimura and Sekido, 1975.)

development of a new vegetative apex, which grows beyond the old apex. The now-dormant reproductive apex forms a dome, which can be seen in a longitudinal section of the stem. In Chamberlainia pteridospermoidea (Artabe et al., 2005), the stem is covered by persistent rhomboidal leaf bases and a few adventitious branches in the form of bulbils (offsets). Fascisvarioxylon methae is a structurally preserved axis from the Rajmahal Hills of India that is regarded as a cycadalean stem (Jain, 1964). On the surface of the stem is an armor of leaf bases. Surrounding the pith are a series of mesarch primary xylem strands and a zone of secondary xylem 2.0 mm thick. Medullary rays are uni- or biseriate and up to 30 cells high. Nilssoniocladus is a morphogenus used for slender cycadalean stems (long shoots) with attached leaf-bearing short shoots. The specimens come from the Early Cretaceous of Japan, and consist of helically arranged short shoots up to 2.0 cm long (Kimura and Sekido, 1975). Attached to the apex of each spur shoot is a cluster of 3–7 leaves of the Nilssonia nipponensis type (FIG. 17.13); the lower portion of the spur shoot is covered by rhomboidal scars that mark the former position of leaves. Specimens of two species of Nilssoniocladus have also been reported from the Albian–Cenomanian (Lower–Upper Cretaceous) of the Arctic and the plant was interpreted as a

frost-tolerant, high-latitude cycad (Spicer and Herman, 1996). In these specimens, Nilssonia leaves are attached to short shoots that are interpreted as abscising from the parent plant, based on the presence of scars on long shoots. In this regard the fossils demonstrate an example of deciduousness in which the entire, leaf-bearing short shoot is abscised rather than the individual leaves, similar to the extant conifers Metasequoia and Taxodium (Chapter 21). Based on the generally slender habit of the stems, Takimats, et al. (1997) have suggested that Nilssoniocladus plants were climbers. PALEOZOIC REPRODUCTIVE STRUCTURES

As noted earlier, cycad ovules are borne on leaflike megasporophylls that can range from little modified (e.g., Cycas) to greatly modified. Based on the presence of megasporophylls in the fossil record, the Cycadales are believed to have originated in the late Paleozoic. Lesleya (Fig. 16.38) is an interesting leaf that shares some similarities with Taeniopteris foliage. The veins arise obliquely from a prominent midrib (Remy and Remy, 1975d). Some specimens from the Lower Pennsylvanian of Illinois contain a row of what are interpreted as ovules on either side of the petiole (Leary, 1990). Ovules are interpreted as occurring within a receptacle; however, an alternative interpretation is that this receptacle may also represent a portion of the seed integument. Nothing is known about the anatomy or morphology of the ovules. If the Lesleya leaves are, in fact, cycad megasporophylls, then perhaps the group arose even earlier in the Mississippian. Spermopteris was a genus established for taeniopterid leaves bearing seeds from the Upper Pennsylvanian of Kansas (Cridland and Morris, 1960). The seeds were described as being attached to the abaxial surface of the lateral margins of the leaf, one row on either side (FIG. 17.14).

710

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 17.15 Distal branch portion with numerous attached leaves of Phasmatocycas birdwellii (Pennsylvanian). Bar  4.0 cm.

Subsequent examination, however, showed that the seeds are attached to the midrib of the leaf (Axsmith et al., 2003). Because Spermopteris is known only from impressions, details about the seeds were lacking. Phasmatocycas is a genus established for compressed megasporophylls from the Lower Permian of Kansas (Mamay, 1973), which consist of an axis bearing two rows of sessile (FIG. 17.16), broadly attached gymnospermous ovules. In P. kansana, the fertile axis is 9.0 cm long. The ovoid seeds are up to 4.0 mm long (FIG. 17.17) and characterized by two cuticular membranes containing a thick megaspore. The presence of glandular bodies associated with both the ovule-bearing axis and associated Taeniopteris foliage (FIG. 17.18) led Sergius Mamay (FIG. 17.19) to suggest that the Phasmatocycas megasporophyll consisted of an expanded distal lamina (Taeniopteris) and an unexpanded basal portion (petiole) to which the ovules were attached in rows (FIG. 17.20). Gillespie and Pfefferkorn (1986) discovered additional specimens of Phasmatocycas and Taeniopteris and were able to confirm that both were part of the same plant. These authors suggested that the ovules were attached to the midrib of the leaf and partially covered by an abaxial flange extending from the midrib (FIG. 17.21). Based on an examination of all of the available Spermopteris and Phasmatocycas specimens, Axsmith et al. (2003) determined

17.16 Reconstruction of the megasporophyll Phasmatocycas kansana. (From Mamay, 1976.)

Figure

of

that they represent the same taxon. They erected a new species, P. bridwellii (FIGS. 17.15, 17.20, 17.22), for the Upper Pennsylvanian forms and retained P. kansana for the Lower Permian plants. Since the basionym of Spermopteris belongs

CHAPTER 17

CYCADOPHYTES

711

Figure 17.17 Several seeds of Phasmatocycas kansana showing

micropyle (arrow) (Permian). Bar  2.5 mm. (From Mamay, 1976.)

Figure 17.19 Sergius H. Mamay (Courtesy H. N. Andrews).

Figure 17.18 Glandular-like bodies associated with the leaf of Phasmatocycas kansana (Pennsylvanian). Bar  2.0 mm.

to a species of sterile foliage from Europe, it could not be used for the fertile structures. The two species differ in age and features such as leaf venation, lamina attachment, and fertile–sterile leaf dimorphism. Eophyllogonium is another taeniopterid leaf from the Permian of China with what are interpreted as ovules attached to the margin of the leaf (Mei et al., 1992). The Early Permian megasporophyll Sobernheimia (FIG. 17.23) from the area of Sobernheim, Germany, is somewhat similar to Phasmatocycas (Spermopteris) (Kerp, 1983). It is

described as a phylloid (leaf like) bearing a row of alternate, ovoid, seedlike structures (7.0 mm  4.0 mm) along each lateral margin (FIG. 17.24). The margin of the lamina is divided into lobes, each 1.0 cm long. Each lobe is vascularized by a single vein. Archaeocycas is another Early Permian cycadean megasporophyll originally described by Mamay (1973, 1976a). In A. whitei, ovules are interpreted as being borne in opposite pairs along the stalk of the megasporophyll, with the lamina of the sporophyll partially enclosing the ovules (FIG. 17.25). Molds of the ovules suggest that they were 3.0 mm long, but the quality of the specimens is poor and additional specimens are needed to better understand this genus. Other cycad megasporophylls similar to those of some extant cycads are known from the Lower Permian of China (J.-N. Zhu and Du, 1981). In Crossozamia (FIG. 17.26), the megasporophyll is 5.0 cm long and has five pairs of subopposite, small ovoid ovules along the stalk (Gao and Thomas, 1989). The foliage genera Taeniopteris, Tianbaolinia (FIG. 17.27) (Gao and Thomas, 1989), and Yuania (Du and Zhu, 1982) are found in association with the megasporophylls, but not in attachment. Morphological features of the leaves suggest that the plant that produced the Crossozamia megasporophylls had immature foliage of the Yuania type and mature leaves of Tianbaolinia (Gao and Thomas, 1989). Both Tianbaolinia and Yuania are pinnate.

712

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 17.21 Diagrammatic reconstruction of Phasmatocycas

megasporophyll showing position of seeds. (From Gillespie and Pfefferkorn, 1986.)

Figure 17.20 Phasmatocycas bridwellii megasporophyll showing rows of seeds (Pennsylvanian). Bar  2.0 cm.

Palmate megasporophylls, covered by small scales, are included in Primocycas chinensis (FIG. 17.28) (J.-N. Zhu and Du, 1981). The specimens are also from the Upper Permian of China and have two seeds attached to opposite sides of the sporophyll stalk, much like those of some species of Cycas and Crossozamia. Lasiostrobus is an unusual pollen cone that has been suggested to have affinities with the cycads. Specimens of L. polysaccii occur in coal balls from the Upper Pennsylvanian

Figure 17.22 Suggested reconstruction of Phasmatocycas bridwellii. (From Axsmith et al., 2003.)

Mattoon Formation near Berryville, Illinois (Taylor, 1970b). The cones consist of helically arranged microsporophylls that are broadly attached at the base and upturned distally (FIG. 17.29). On the abaxial surface of each sporophyll (FIG. 17.30) are seven to ten thick-walled pollen sacs

CHAPTER 17

CYCADOPHYTES

713

Reconstruction of Sobernheimia showing leaflike structure and abaxially attached seeds. (From Kerp, 1983.)

Figure 17.24

Figure 17.23 Sobernheimia jonkeri (Permian). Bar  2.0 cm. (Courtesy H. Kerp.)

(FIG. 17.31). The pollen of L. polysaccii is small (20–29 μm) and inaperturate, and a few grains have been found with trilete sutures. Extending from the surface of the grain are three to eight subequatorial sacci, a feature that is unknown

Figure 17.25 Reconstruction of Archaeocycas megasporophyll with lamina partially enclosing seeds. (From Taylor and Millay, 1979.)

714

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

17.28 Megasporophyll of Primocycas chinensis (Permian). Bar  1.0 cm. (From J.-N. Zhu and Du, 1981.)

Figure

17.26 Megasporophyll of Crossozamia (Permian). Bar  4.0 mm. (From Gao and Thomas, 1989.)

Figure

Figure 17.27 Tianbaolinia circinalis Bar  8.0 mm. (From Gao and Thomas, 1989.)

leaf

minor

(Permian).

in any cycads, living or fossil. The ultrastructural organization of the exine is tectate. One interesting feature concerning the pollen grains is the suggestion that they were shed in the microspore stage of development (Taylor and Millay, 1977), a feature regarded as advanced because of the reduced nature of the microgametophyte. An analysis of the fossil cones, including pollen, cone apex, vascularization, epidermis,

Figure 17.29 Diagrammatic reconstruction of Lasiostrobus polysaccii. (From Taylor, 1970b.)

pollen sacs, and overall histology, shows features in common with both the cycads and conifers. Lasiostrobus has also been suggested as being an early ginkgophyte on the basis of the unusual vascular system of the microsporophylls.

CHAPTER 17

Figure 17.30 Cross section of microsporophyll of Lasiostrobus polysaccii showing abaxial attachment of pollen sacs (Pennsylvanian). Bar  1.0 mm.

CYCADOPHYTES

715

Figure 17.32 Cross section of Antarcticycas schopfii stem

(Triassic). Bar  1.0 cm.

17.33 Cross section of Antarcticycas schopfii stem showing girdling leaf trace (arrow) in cortex (Triassic). Bar  1.0 mm.

Figure

Figure 17.31 Section of Lasiostrobus polysaccii pollen sacs (Pennsylvanian). Bar  500 μm.

TRIASSIC CYCADS

There are several Mesozoic cycads that are known in sufficient detail that most or all of the plant can be reconstructed with some degree of confidence. One of these is Antarcticycas

schopfii, known from permineralized stems, leaves, cataphylls, and probable pollen cones from Middle Triassic silicified peat from Antarctica (Smoot et al., 1985). The stems are roughly circular in cross section (FIG. 17.32), 4.0 cm in diameter, and consist of a wide parenchymatous pith and cortex, with some cells containing dark contents (Smoot et al., 1985). Separating these two zones is a narrow ring of vascular tissue that includes secondary xylem, cambium, and secondary phloem. Traces to the leaves demonstrate a girdling pattern typical of extant cycads (FIG. 17.33), and the traces change from endarch to exarch in the petiole base (Hermsen et al., 2007), another characteristic that can be

716

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 17.35 Cortical cells of Antarcticycas showing Giga-

sporites myriamyces arbuscule (Triassic). Bar  40.0 μm. Figure 17.34 Trichomes on a cataphyll of Antarcticycas

(Triassic). Bar  15.0 μm.

compared with living forms. Tightly arranged cataphylls, roughly triangular in outline, are present at the outside of the periderm. Covering both surfaces of the cataphylls is an indumentum of sinuous hairs (FIG. 17.34) (Hermsen et al., 2006). Numerous adventitious roots occur in the cortex and extend out from the stem surface. They range from diarch to polyarch, with the larger ones possessing secondary xylem. Endomycorrhizal fungi have been reported in the cortical cells (FIG. 17.35) (Phipps and Taylor, 1996) and the presence of a phi-layer in the roots suggests that the plant grew in an environment with changing water levels (Millay et al., 1987). Extending from one stem is an adventitious bulbil similar to those produced in modern cycads (Fig. 17.3) (Hermsen et al., 2008), which are sometimes termed offsets or suckers; these represent a method of vegetative reproduction. Leaves of Antarcticycas are once pinnate and the petiole trace consists of an inverted omega shape like all modern members of the Cycadales. Pinnae are borne laterally and lack a midrib. The detached leaves are included in the morphotaxon Yelchophyllum (FIG. 17.36) (Hermsen et al., 2007a). Pollen cones of Delemaya spinulosa occur in the same silicified peat and are interpreted as belonging to Antarcticycas. The cones are 3.0 cm long and consist of helically arranged microsporophylls characterized by distal projections (Klavins et al., 2003). Each microsporophyll contains two clusters of partially fused pollen sacs on the abaxial surface and pollen is 20 μm long and monosulcate. The Antarcticycas plant is hypothesized as growing from a subterranean stem (FIG. 17.37), based on the narrow stem diameter, presence of numerous adventitious roots, and polar light regime in which it grew, with dark winters.

Figure 17.36 Cross section of Yelchophyllum leaf (Triassic).

Bar  100 μm.

Figure 17.37 Suggested reconstruction of Antarcticycas schopfii. (From Hermsen et al., 2008.)

CHAPTER 17

Figure 17.38 Suggested reconstruction of Leptocycas gracilis. (From Delevoryas and Hope, 1971.)

Yelchophyllum leaves appear to be fusinized, whereas Antarcticycas stems do not, suggesting that fire was a component of the ecosystem in which it grew and providing support for the hypothesis of an underground stem. Antarcticycas schopfii was initially thought to be closely related to Bowenia (Smoot et al., 1985); however, a more recent phylogenetic analysis suggests that it may not be in the lineage that gave rise to modern cycads (Hermsen et al., 2006b). Leptocycas gracilis is a tree-size cycad that has been reconstructed based on fossils from the Upper Triassic of North Carolina (Delevoryas and Hope, 1971). The reconstruction is based on the attachment of stems, leaves, and possible pollenproducing cones (FIG. 17.38). It is unlike the modern cycads because the stem was rather smooth and probably no more than 5.0 cm in maximum diameter. Based on the size of the leaves and the diameter of the stem, L. gracilis is thought to have been 1.5 m tall. Along the stem, leaf bases were widely separated and at the apex there was a loose crown of pinnately compound leaves of the Pseudoctenis type associated with cataphylls. Individual fronds were petiolate and

Figure 17.39

CYCADOPHYTES

717

Hugo Rühle von Lilienstern.

exceeded 20.0 cm in length; cuticle fragments indicate that the stomata are characteristically cycadalean with haplocheilic guard cell ontogeny. Attached to one stem is a slender strobilus that morphologically resembles a pollen cone. Although L. gracilis lacks cellular preservation, the morphology of the plant has been especially important in defining the habit of early cycads. Bjuvia simplex has been reconstructed as a tree bearing leaves of Taeniopteris gigantea (Florin, 1933a). This Late Triassic plant is associated with megasporophylls interpreted as having ovules borne along the leaf midrib (Florin, 1933a). Because of the association of the two organs and the similar epidermal pattern, the sporophylls and leaves were used to reconstruct a plant with a stout, heavily armored trunk with persistent leaf bases and a terminal crown of leaflike megasporophylls that superficially resemble those of Cycas. The fragmentary nature of the remains, absence of seeds, and the simple organization of the leaves, however, are features that have made some skeptical of Florin’s original interpretation (Axsmith et al., 2003). Based on material from southern Germany, Hugo Rühle von Lilienstern (FIG. 17.39), an amateur collector from Thuringia,

718

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 17.40 Dioonitocarpidium liliensternii showing detail of

ovule attachment (arrow) (Triassic). Bar  5.0 mm. (Courtesy C. Pott.)

provided a tentative reconstruction of Dioonitocarpidium (Dioonites) pennaeformis, an enigmatic Late Triassic plant with possible affinities in the Cycadales (Rühle von Lilienstern, 1928). The plant is suggested to have large, entire-margined leaves, which were originally described as Daeniopsis (Taeniopteris) angustifolia. Megasporophylls of Dioonites–Dioonitocarpidium exhibited a proximal, entiremargined region with a strong midrib to which are attached several seeds (FIG. 17.40), and a distal portion with a highly dissected lamina (Schenk, 1865–1867). The reconstruction, however, has variously been questioned based on the fact that foliage and megasporophylls have not been found in organic connection (Kelber and Hansch, 1995). Kustatscher et al. (2004) correlate Dioonitocarpidium with foliage assignable to Bjuvia based on co-occurrences in Ladinian (upper Middle Triassic) rocks of northern Italy. Zamioidea macrozamioides (FIG. 17.41) is a morphotaxon used for isolated Jurassic megasporophylls that are structurally similar to megasporophylls of Zamia.

17.41 Zamioidea macrozamioides, isolated cycad megasporophyll showing site of seed attachment (arrows) (Jurassic). Bar  1.0 cm. (Courtesy BSPG.)

Figure

JURASSIC CYCADS

Androstrobus is a morphogenus for small, cylindrical pollen cones consisting of helically arranged microsporophylls with abaxial pollen sacs (FIGS. 17.42, 17.43A, B); some species have been assigned to the Cycadales, although others do not belong to this group. Bernettia (FIG. 17.44) has also been used for microsporangiate cones that are now included in Androstrobus (van Konijnenburg-van Cittert, 1993). Androstrobus was originally described from the Jurassic of Etrochey in France (Schimper, 1869–1874) but is best known from a number of species in the famous Middle Jurassic floras from Yorkshire, England (Harris, 1941b; Thomas and Harris, 1960; Van Konijnenburg-Van Cittert, 1968). Androstrobus has also been reported from

Figure 17.42 Suggested reconstruction of Androstrobus manis

cone (Jurassic). (From Delevoryas, 1962.)

CHAPTER 17

CYCADOPHYTES

719

Microsporophyll

(B)

Androstrobus nathorstii

(A)

Zamiostrobus luciae (D)

Figure 17.45 Diagrammatic reconstruction of Androstrobus piceoides microsporophylls and sporangia. (From Schweitzer et al., 2000.)

Megasporophyll (C)

Figure 17.43 A. Suggested reconstruction of Androstrobus nath-

orstii. B. Isolated microsporophyll. C. Seed cone of Zamiostrobus luciae D. Isolated megasporophyll. (Courtesy J. Watson.)

Figure 17.44 Impression of Bernettia ( Androstrobus) cone surface showing microsporophylls (Jurassic). Bar  2.5 cm.

several other Triassic to Cretaceous sites in Argentina (Archangelsky and Villar de Seoane, 2004), China (Hu et al., 1999) Greenland (Van Konijnenburg-van Cittert, 1993), Iran (FIG. 17.45) (Schweitzer et al., 2000), and South Africa (Anderson and Anderson, 2003), among other countries. In

A. wonnacottii from the Jurassic of Yorkshire, the cone is 7.5 cm long and the distal ends of the microsporophylls are upturned (C. Harris, 1964). Androstrobus balmei is a slightly larger cone with 50 microsporangia arranged in clusters on the abaxial surface of the microsporophyll (C. Hill, 1990). The pollen grains are ovoid, monosulcate, and up to 30 μm long. This pollen morphotype shares some features with the modern family Cycadaceae (e.g., surface sculpture), whereas the fine structure of the exine is more similar to that found in the Zamiaceae. Ovulate cones of the plant that bore Androstrobus are called Beania (FIG. 17.46). They consist of an axis bearing widely separated megasporophylls, each with a distally expanded, peltate portion which resembles the megasporophylls of certain extant cycads (Carruthers, 1869; Harris, 1964; Schweitzer et al., 2000). Two sessile, orthotropous seeds are attached to the inner surface of each megasporophyll. The nucellus and integument are fused (adnate) in the lower (chalazal) two-thirds of the ovule. Because the genus is known only from compressed specimens, features of the vascular system and histology of the ovules have not been determined. Specimens of B. mamayi are known in which the axis is at least 18.0 cm long (Thomas and Harris, 1960). Associated with the megasporophylls are strap-shaped leaves up to 80.0 cm long with a prominent midvein and entire margin. Stomata with slightly sunken guard cells are scattered on the lower surface of the leaves between the veins; trichomes are also common along the veins. Foliage

720

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 17.47 Suggested reconstruction of the ovulate cone of Microzamia gibba. (From J. Kvacˇ ek, 1997.) Figure 17.46 Suggested reconstruction of Beania gracilis.

(From Taylor and Taylor, 1993.)

has been called Nilssonia tenuinervis, but as noted by Cleal et al. (2006), these leaves from the Middle Jurassic of Yorkshire require the naming of a new species, as they do not fit with the original species concept. The entire plant was reconstructed by Harris based on co-occurrence, including A. wonnacottii pollen cones, B. mamayi ovulate cones, and N. tenuinervis foliage, but it does not look much like a modern cycad (Harris, 1961b). Rather, the leaves and strobili were borne at the ends of slender, branched axes, although Harris (1961b) noted that there

is no evidence for the stem morphology of this plant. It is not known whether the pollen and ovule-bearing cones were borne on the same plant or on different plants, as in modern cycads. Based on the biovulate sporophylls and the cell structure of the integument (see Harris, 1941b), Beania closely resembles Microzamia gibba (FIG. 17.47), a compact ovulate cycad cone with cataphylls on the peduncle from the Cenomanian (Upper Cretaceous) of Bohemia, Czech Republic (J. Kvacˇek, 1997). Lioxylon is a stem from the Middle Jurassic of China which has been attributed to the cycads (W. Zhang et al.,

CHAPTER 17

2006). This permineralized stem shares a number of features in common with several fossil and living cycads.

Paleozoic

CYCADOPHYTES

Mesozoic

721

Cenozoic

POLLINATION BIOLOGY

The presence of glandular bodies (FIG. 17.18) associated with seeds of Phasmatocycas suggested to Mamay (1976a) that they may have functioned in some form of biotic pollination syndrome, perhaps involving insects, although at the time it was thought that all extant cycads were wind pollinated (Kono and Tabe, 2007). This interesting hypothesis has since found support from studies of extant cycads, which indicate that many, if not all, modern species are pollinated by insects (mostly beetles) (Norstog and Nichols, 1997; Hall et al., 2004). Support for Mamay’s hypothesis has also come from the fossil record. The Triassic pollen cone Delemaya spinulosa was described as containing coprolites composed exclusively of Delemaya pollen—the oldest evidence of cycad pollinivory in the fossil record (Klavins et al., 2005). Although not direct evidence of pollination, the presence of these coprolites may signal an early stage in the evolution of the plant–pollinator syndrome found in modern cycads. DISCUSSION: CYCAD EVOLUTION

Mesozoic cycads were historically thought to have been plants with slender stems and widely separated leaves that abscised, similar to Leptocycas gracilis (Delevoryas and Hope, 1971; Delevoryas, 1975), that is, morphologically unlike the short, squat trunks of modern cycads. Additional research, however, has shown that some Mesozoic cycads, such as Charmorgia dijolii, had short trunks like extant taxa (Ash, 1985a), and that both growth forms existed at the same time. The reconstruction of Antarcticycas schopfii with a subterranean stem has further expanded our knowledge of habit diversity in Mesozoic cycads. Permineralized cycads indicate that monostelic (e.g., Antarcticycas) and polystelic (e.g., Worsdellia) cycads existed at the same time and that by the Mesozoic, the group was exploiting a variety of growth habits and habitats. In his review of fossil cycads, Mamay (1976a) suggested that the earliest cycads had simple rather than pinnate leaves and that there was a progressive dissection of the simple leaf blade in the course of evolution of the cycad frond. This evolutionary series would have begun in the Paleozoic with a simple, taeniopterid leaf and culminated in a pinnate leaf type in the Cenozoic (FIG. 17.48). The starting point for this transformational series was no doubt influenced by the taeniopterid leaves associated with Spermopteris/Phasmatocycas. The discovery of pinnate leaves in the Lower Permian of China in association with the megasporophyll Crossozamia, however, suggests an

(B)

(A) (C)

Suggested evolution of cycad leaves beginning with entire margined, A. ancestral taeniopterid type, B. extending to forms with more incised margins leading to cycad, and C. zamioid types. (From Mamay, 1976a.) Figure 17.48

alternative interpretation for cycad leaf evolution. It is important to point out, however, that neither of the two leaf types at that locality, Tianbaolinia and Yuania, has been found attached to Crossozamia (FIG. 17.49). Still another interpretation is that by the Early Permian several types of cycads had already evolved, each with a distinctive leaf type (Gao and Thomas, 1989). This hypothesis fits with the suggestion, based on a phylogeny of fossil and extant cycads which included character and minimum age mapping, that major extant lineages of cycads had diverged by the Permian–Triassic and have remained relatively unchanged (FIG. 17.50) to the present (Hermsen et al., 2006). It has long been believed that the Cycadales have their origin among the Paleozoic seed ferns, the Medullosales being most often suggested as ancestors. This hypothesis is supported by most morphologic phylogenetic analyses (Crane, 1985a; Nixon et al., 1994; Doyle, 2006), which view the medullosans as sister to all other seed plants except the Lyginopteridales. Adding further support to this hypothesis is a developmental analysis of male sporangiophores of extant Zamia amblyphyllidia, which suggests that cycads evolved from a pinnate, pteridospermous ancestor with radial synangial groups (Mundry and Stützel, 2003). Irrespective of their origin, the cycads reached their maximum development during the Mesozoic, both in terms of geographic distribution and in numbers of taxa, and have steadily

722

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

(E) (B)

A

(D) (F) (G)

B

(A)

(I) (C)

(H)

Suggested evolution of the cycad megasporophyll beginning with the Crossozamia type (a). One lineage leading to the Cycas form (i) and the second leading to the Zamia form (f, g). Intermediate forms include Crossozamia minor (b), Beania gracilis (d), Beania ovule (e), Crossozamia chinensis (c) and Cycas revoluta (h). (From Gao and Thomas, 1989.)

Figure 17.49

declined during the intervening 240 million years. Phylogenies that include extant and fossil members of the Cycadales suggest that some of the features seen in extant members, including girdling traces and omega-shaped petiole traces, had evolved by the Permian (Hermsen at al., 2006). In spite of these character analyses, however, there remain major vegetative and reproductive differences between cycads and seed ferns, including the presence of dioecy in modern cycads. The data are not conclusive for all medullosans, but at least some Carboniferous forms produced both ovules and synangia on the same plant. Although we believe that the medullosans continue to represent the best ancestral group in the evolution of the Cycadales, perhaps if only by the process of elimination, the fact that they co-existed with cycads underscores the fact that these relationships are far from settled (Delevoryas, 1982; Crane, 1988).

BENNETTITALES The plants included in the Bennettitales, or Cycadeoidales of some authors, extend from the Triassic to the Cretaceous and occur in both hemispheres. The leaves in this clade are morphologically identical with those of many cycadaleans,

except for the presence of syndetocheilic stomata in the Bennettitales. Bennettitalean foliage is very similar morphologically to that of cycads, and the leaves are often distinguishable only by epidermal features, that is, the presence of syndetocheilic stomata, as noted earlier (see Chapter 16 for details). Moreover, it has recently been documented that immature Pterophyllum filicoides (FIG. 17.51) leaves from the Carnian (Upper Triassic) of Lower Austria display circinate vernation (FIG. 17.52) similar to that seen in modern cycads (Pott and Krings, 2007). This suggests that bennettitalean foliage may resemble cycad leaves not only in macromorphology but also in development. There are several foliage genera that are attributed to the bennettitales based on epidermal anatomy and the morphology of the leaf segments (also termed leaflets or pinnules) where they attach to the rachis (FIG. 17.53) (see Chapter 16). Several of the more common morphogenera are Nilssoniopteris (FIG. 17.53), Pseudocycas, Dictyozamites, Ptilophyllum, Pterophyllum (FIG. 17.54; 17.55), Otozamites, and Zamites (Watson and Sincock, 1992), all of which are described in more detail in Chapter 16. Bennettitaleans also produce cataphylls; two names used for these scale leaves are Deltolepis and Cycadolepis (Chapter 16). The vast majority

CHAPTER 17

CYCADOPHYTES

723

17.51 Leaf of Pterophyllum filicoides (Triassic). Bar  4.0 cm. (Courtesy C. Pott.)

Figure

Figure 17.50 Leaflet of Eostangeria pseudopteris (Paleogene).

Bar  1.0 cm. (From Z. Kvacˇek and Manchester, 1999.)

of bennettitalean leaves are once pinnate; however, there are also several accounts of fossil bipinnate leaves referred to the Bennettitales, for example, Banatozamites chlamydostomus from the Liassic (Hettangium, Lower Jurassic) of Romania (Czier, 1996), Coreanophyllum variisegmentum from the Upper Triassic of Korea (Kimura and Kim, 1982, 1988) (FIG. 17.56), Nipponoptilophyllum bipinnatum from the Upper Jurassic of Japan (Kimura and Tsujii, 1984), and Pterophyllum bavieri (FIG. 17.57) from the Rhaetian (Upper Triassic) of Iran (Schweitzer and Kirchner, 2003). Eoginkgoites is an unfortunate name for another cycadeoid leaf type. The morphogenus is restricted to the Upper Triassic of North America and includes pinnate leaves with up to five lateral pinna pairs (Axsmith et al., 1995). Subsidiary cells of abaxial stomata sometimes have papillae overarching the

Figure 17.52 Coiled leaf segment of Pterophyllum filicoides show-

ing circinate vernation (Triassic). Bar  3.0 mm. (Courtesy C. Pott.)

stomatal pit. Eoginkgoites has been suggested as an example of a bennettitalean leaf in which lateral pinnae have been reduced (Ash, 1976). Although the vast majority of bennettitalean leaves are preserved as impression–compressions, there

724

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

A–N usually with stomata confined to abaxial surface (hypostomatic)

A B C

D A–D Zamites Equally contracted basal angles

F G

E, F Otozamites Acroscopic basal angle forming auricle

E

H

M G, H Ptilophyllum Basiscopic basal angle decurrent

I

K

K Anomozamites Basal angles usually expanded, pinna about as wide as long

I,J Pterophyllum Basal angles usually expanded, pinna longer than wide L

L Dictyozamites Reticulate venation

J Stomatal groove

O

N

O Pseudocycas Stomata confined to abaxial central groove

M Nilssoniopteris Leaf entire

N Sphenozamites Pinna rhomboid, attached laterally

Figure 17.53 Pinna morphology used to define bennettitalean foliage genera. (From Watson and Sincock, 1992.)

are a few that are anatomically preserved (see Chapter 16). One of these is Otozamites mortonii from the Middle Jurassic of Scotland (Dower et al., 2004). In this pinnate leaf the vascular bundles in the rachis form an omega-shaped strand that surrounds a smaller, cylindrical bundle. A sclerenchyma sheath surrounds each vascular bundle and the abaxial surface contains numerous epidermal papillae, a character that has been interpreted as being linked to the environment of growth. The habit of the plants varies from short and squat to slender and highly branched. Cycads and bennettitaleans are quite

similar anatomically except for the absence of girdling leaf traces in the latter. The two groups differ greatly, however in the anatomy and morphology of their reproductive organs, which, in the bennettitaleans, consist of relatively small to medium-sized but complex mono- or bisporangiate structures. Historically, the Bennettitales has been divided into two families that differ in their geologic age, habit, organization, and features (e.g., seed integument) of the reproductive organs (Rothwell and Stockey, 2002). We have followed the familial designations outlined by Watson and Sincock (1992) (FIG. 17.58).

CHAPTER 17

CYCADOPHYTES

725

Figure 17.55 Suggested reconstruction of Williamsonia with

Nilssoniopteris leaves. (From Watson and Sincock, 1992.)

Figure 17.56 Tatsuaki Kimura.

Figure 17.54 Pterophyllum lyellianum leaf (Cretaceous).

Bar  4.0 cm.

CYCADEOIDACEAE

STEM ANATOMY The best-known member of this family is Cycadeoidea, a genus erected by Buckland in 1828 and known principally from silicified trunks (FIG. 17.59) with embedded reproductive structures (see a review and history of the genus in

Watson and Lydon, 2004). Numerous specimens have been collected from throughout the world, and the holotype of C. etrusca was even discovered in an Etruscan grave, in which it was apparently placed as a burial object (Capellini and Solms-Laubach, 1890, 1892). Some of the earliest studies are those of Buckland (1828), Carruthers (1868), and Ward (1898), all of whom initially suggested that the fossils had similarities to modern cycads. These investigations were followed by the extensive two-volume study of George R. Wieland (FIG. 17.60), American Fossil Cycads, based on numerous silicified trunks collected from the Lower

726

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 17.57 Pterophyllum bavieri. (From Schweitzer and

Kirchner, 2003.)

Cretaceous Lakota Formation of the Black Hills in South Dakota (Wieland, 1906, 1916). The petrified trunks of Cycadeoidea are morphologically similar in general organization to many modern cycads in being cylindrical to columnar (sometimes globose), and usually 1 m tall. Most trunks are unbranched, although there are reports of some that apparently branched profusely. Covering the surface are helically arranged, persistent leaf bases (FIG. 17.61) and multicellular, scalelike hairs that formed a ramentum around the bases. Reconstructions of Cycadeoidea show a crown of helically arranged, pinnate leaves extending from the distal end of the trunk (FIG. 17.62), although we are unaware of any mature leaves being found attached to the trunks. Wieland did describe tightly coiled immature leaves associated with one of the trunks. Bisporangiate cones are embedded in the trunk among the leaf bases (Delevoryas, 1963). In transverse section the stem of Cycadeoidea may be up to a meter in diameter and consists of a large pith containing

Figure 17.58 Joan Watson.

secretory canals surrounded by an endarch eustele and secondary xylem and phloem. Secondary xylem is composed of scalariform tracheids and uni- and biseriate vascular rays. An interesting anatomical feature of some Cycadeoidea stems is the nearly equal distribution of xylem and phloem (Wieland, 1906), and the secondary phloem contains tangential bands of fibers alternating with sieve cells (FIG. 7.11) and parenchyma (Ryberg et al., 2007). Surrounding the vascular cylinder is a broad cortex of thin-walled parenchyma and secretory cells. Leaf traces are C-shaped as they pass through the cortex and some contain secondary xylem (Delevoryas, 1960). The vascular system of the cone is

CHAPTER 17

CYCADOPHYTES

727

Figure 17.61 Top of Cycadeoidea sp. trunk showing heli-

cal arrangement of leaves (Cretaceous). Bar  1.5 cm. (Courtesy J. Watson.) Figure 17.59 Silicified Cycadeoidea sp. trunk (Cretaceous).

Bar  6.0 cm.

Figure 17.62 Reconstruction of Cycadeoidea plant bearing

leaves. (From Delevoryas, 1971.)

Figure 17.60 George R. Wieland.

(Courtesy T. Delevoryas)

derived from segments of several leaf traces that form a cylinder in the outer cortex. At a higher level, one segment of the cylinder separates to vascularize a subtending leaf; the remaining vascular tissue extends into the cone peduncle.

In the closely related genus Monanthesia, every leaf contains a cone in its axil (FIG. 17.63) (Delevoryas, 1959). In M. magnifica, from the Late Cretaceous (Campanian) Mesa Verde Group of New Mexico, USA, the vascular system of the cone is derived from the fusion of two cortical bundles that develop from two leaf traces, neither of which supplies the subtending leaf. Trunks initially described as Cycadites saxbyanus from the Isle of Wight are morphologically identical

728

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 17.63 Surface of Monanthesia sp. trunk showing cones in axils of leaf bases (Cretaceous). Bar  5.0 cm. (Courtesy J. Watson.)

to Monanthesia magnifica (Watson and Lydon, 2004). Monanthesia is also known from England and may parallel the stratigraphic range in North America (Watson and Lydon, 2004). In spite of vegetative similarities, there are numerous features of the reproductive structures, for example, the presence of a cone in the axil of every leaf, long stalks on the ovules, a small-domed receptacle, and apparently synchronous development of the cones that distinguish this genus from Cycadeoidea. REPRODUCTIVE STRUCTURES The cones of all members of the Cycadeoidaceae are bisexual and borne on short lateral branches or peduncles (FIG. 17.64). It is thought that they did not extend beyond the level of the trunk surface (Delevoryas, 1968). Some cycadeoids produced cones that were nonsynchronous in development, whereas others, such as Monanthesia, appear to have produced cones that all matured at about the same time. It should be pointed out, however, that because cycadeoids are silicified and thus require the preparation of thin sections to study, not a large number of cones have actually been sectioned. Despite the fact that initial work on Cycadeoidea cones was done as early as 1870 by Carruthers, details on the structure and organization were first known in the 1960s and 1970s. In general Cycadeoidea cones consist of an outer whorl of microsporangiate structures that surround a central, dome-shaped receptacle (FIG. 17.65). Arising from the receptacle are orthotropous ovules surrounded by hairlike interseminal scales (FIGS. 17.66, 17.67). Small cones of Cycadeoidea are known that only contain ovules, but subsequent studies have shown that these were also bisporangiate.

Figure 17.64 Cycadeoidea trunk with several cones (arrows)

(Cretaceous). Bar  10.0 cm.

R

Figure 17.65 Longitudinal section of Cycadeoidea cone

showing receptacle (R) and microsporophylls and pollen sacs (arrows) (Cretaceous). Bar  1.0 cm.

CHAPTER 17

CYCADOPHYTES

729

Figure 17.66 Longitudinal section of Cycadeoidea cone

Figure 17.67 Longitudinal section of Cycadeoidea cone with

receptacle (Cretaceous). Bar  8.0 mm. (From Crepet, 1974.)

seeds along edge separated by interseminal scales (Cretaceous). Bar  1.0 cm.

The microsporophylls in these cases had shed their pollen and subsequently disintegrated (Crepet, 1974). In some cones there is a structure termed the corona that extends out from receptacle and is made up of modified interseminal scales. This structure has also been reported in some cones of the Williamsoniaceae, but it is not known whether these structures are homologous (Nixon et al., 1994). Vascular tissue appears to be present only in the nucellus of the radially symmetrical seeds. Below the microsporangia is a whorl of sterile bracts. As a result of detailed studies by Wieland (1906), the bisporangiate cones of Cycadeoidea became the focal point of cycadeoids as angiosperm ancestors. Wieland suggested that the microsporophylls surrounding the receptacle and ovules opened at maturity, much like a fern crozier, although subsequent work showed that this was incorrect. The concept of the open cycadeoid “flower” of Wieland made its way into many museum and textbook reconstructions and, as a result, remained well entrenched for many decades; the open reconstruction is even used by some today to describe these reproductive structures.

The next stage in understanding the structure and organization of the cycadeoid cone was provided by Delevoryas (1968). He interpreted the pollen-bearing microsporophylls as shaped like the segments of a citrus fruit in which trabeculae extend between the dorsal and ventral surfaces, and to which were attached synangia containing pollen sacs. In contrast to Wieland’s reconstruction, Delevoryas’ interpretation viewed the entire cone as remaining closed at maturity. Subsequent work by Crepet (1974), utilizing cones at various stages of development, provided the basis for a modification of this earlier interpretation. Crepet suggested that each microsporophyll was homologous with a pinnate frond that is curved in on itself so that the distal tip is adjacent to the microsporophyll base (FIG. 17.68), and each pinna is folded toward the receptacle. In this interpretation the microsporophylls are attached to the rachis only at their base and are appressed or ontogenetically fused at the tip. Crepet noted that there is no convincing evidence to suggest that the structure ever opened or expanded, as was originally hypothesized by Wieland. Fusion of the distal

730

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

MS

Figure 17.68 Diagrammatic reconstruction of Cycadeoidea

cone showing microsporophylls in various planes of section. (From Taylor and Taylor, 1993.)

tips of the microsporophylls, adhesion between the proximal and distal portions of the pinnae, and the presence of a mass of sterile tissue at the distal end of the cone, at the point of maximum curvature of the microsporophylls, all suggest that the cones were incapable of opening (Crepet, 1974). The pollen-bearing structures in Cycadeoidea consist of thick-walled, bean-shaped synangia, each containing 8–20 tubular sporangia arranged around the periphery (FIGS. 17.69, 17.70) (Delevoryas, 1965). Pollen grains of C. dacotensis, a common species in North America, are broadly bilaterally symmetrical and 25 μm long (FIG. 17.71). On the distal surface is an elongated sulcus, and ornamentation is psilate to punctate (Osborn and Taylor, 1995). The sporoderm consists of two zones, the outer one composed of granules. What were originally interpreted as cells of a microgametophyte (Wieland, 1906) are actually folds of the thin exine (T. Taylor, 1973).

Figure 17.69 Longitudinal section of Cycadeoidea cone showing massive sterile tissue at apex (arrow) and microsporophylls (MS) (Cretaceous). Bar  1.0 cm. (From Crepet, 1974.)

DEVELOPMENT One of the special aspects of this group of seed plants is that each trunk generally contains several cones, and many can be preserved in exquisite detail, so that cone development can be studied even in a single trunk. Development stages that have been described range from cones with a mass of undifferentiated tissue representing the immature receptacle, to others which contain hundreds of seeds, some with well-developed embryos. Immature ovules are stalked and surrounded by five to six interseminal scales each; megaspores develop in a linear tetrad (FIG. 17.72) (Crepet and Delevoryas, 1972). As ovule development and megasporogenesis continue, the ovule stalks elongate, and the ovulate receptacle becomes more club shaped. Mature seeds are small (2.0 mm long), slightly flattened, and characterized by an elongate extension of the distal end of the integument that forms a micropylar tube. The integument is three parted

CHAPTER 17

CYCADOPHYTES

731

s

Figure 17.70 Section of several Cycadeoidea dacotensis synangia (S) bracket (Cretaceous). Bar  1.0 mm. (From Osborn, 1991.)

Figure 17.72 Longitudinal section of immature Cycadeoidea ovule after megasporogenesis. Arrow indicates tetrad. (From Crepet and Delevoryas, 1972.)

Figure 17.71 Cycadeoidea dacotensis pollen grain (Cretaceous). Bar  5.0 μm. (From Taylor and Taylor, 1993.)

(endotesta, sclerotesta, and sarcotesta), with the nucellus and seed coat fused from the base to near the micropylar region. Instead of a distinct pollen chamber, the distal end of the nucellus in cycadeoid seeds is represented by a nucellar plug. Embryos have been found in some seeds, and these possess two cotyledons and a protuberance below the epicotyl interpreted as a feeder, a structure which also occurs in the gnetophytes (Chapter 19). The presence of an abscission layer at the base of the receptacle suggests that the cone may have been shed as a unit (Watson and Sincock, 1992). In Monanthesia all the cones on a single trunk appear to be at

732

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

17.74 Cross section of Cycadeoidea maccafferyi cone showing stellate seed in cross section surrounded by interseminal scales (Cretaceous). Bar  1.0 mm. (Courtesy R. A. Stockey.)

Figure

Figure 17.73 Permineralized cone of Cycadeoidea maccafferyi (Cretaceous). Bar  1.5 cm. (Courtesy R. A. Stockey.)

the same level of maturity, suggesting that these plants were monocarpic. It has also been hypothesized that fertilized seeds may have been dispersed by animals, perhaps large herbivores (Watson and Sincock, 1992). Crepet’s discovery that all Cycadeoidea cones were bisporangiate reduced some of the confusion over the method of pollination in the genus (Crepet, 1972). Structural features of the microsporophylls suggest that the units did not open at maturity but rather disintegrated, thus suggesting that the ovules were perhaps self-pollinated. Many Cycadeoidea cones, however, show evidence of insect borings, both in the receptacle and in the area of the pollen sacs. This implies that some arthropod group, perhaps beetles, may have acted as secondary pollinators in this predominantly self-pollinating system (Crepet, 1972; reviewed in Labandeira et al., 2007b). Crepet (1974) has suggested that both pollination systems may have been a factor in the rapid decline of the cycadeoids during the Cretaceous. Self-pollination would have contributed to a homozygous population that would not be well adapted to environmental changes. It has also been hypothesized that the selective pressure on primary pollinators may have caused them to shift from cycadeoids to the angiosperms that were rapidly radiating during the Early Cretaceous. It is interesting to speculate as to whether the pollination syndrome seen in the Cycadeoidea is universal in all cycadeoids. Cycadeoidea maccafferyi is a small (7.8 cm long), extraordinarily well-preserved ovulate cone (FIG. 17.73) from the

Campanian (Upper Cretaceous) Haslam Formation of British Columbia (Rothwell and Stockey, 2002). At the apex, the seeds are stellate in cross section and interseminal scales are conspicuous (FIG. 17.74). An interesting feature in this seed is the vascular bundle, which is convoluted and anatomically mimics the vascular architecture seen in contractile roots. This may suggest that the ovules in C. maccafferyi were exerted during pollination and were subsequently retracted below the level of the interseminal scales. Some seeds possess embryos (FIG. 17.75). WILLIAMSONIACEAE

Members of the Williamsoniaceae include plants with slender, branching stems and stalked reproductive organs that were borne among the leaf bases on the trunks. The plants are believed to have been 2.0 m tall with leaves widely separated along the stems. Some reproductive organs assigned to this group contained only ovules or pollen sacs; others were apparently bisporangiate like Cycadeoidea, but differed in that they opened at maturity. As with a number of fossils, reconstructions of entire plants in this family are based on fragmentary evidence. Members of the Williamsoniaceae are thought to be slightly older than the Cycadeoidaceae. One of the most frequently illustrated members of the Williamsoniaceae is Williamsonia sewardiana (FIG. 17.76), a permineralized plant from the Jurassic of India (Sahni, 1932b). This was a small tree 2.0 m tall with persistent, helically arranged leaf bases. At the distal end of the trunk was a crown of pinnate leaves of the Ptilophyllum type. On the trunk between the scars were smaller scars that marked the former position of scale leaves. Ovulate cones were

CHAPTER 17

CYCADOPHYTES

733

N E

Figure 17.75 Longitudinal section through the apex of a

Cycadeoidea maccafferyi seed showing the tip of the nucellus (N) and radicle of embryo (E) (Cretaceous). Bar  1.0 mm. (Courtesy R. A. Stockey.)

attached to short lateral branches that also produced leaves; each cone was borne on a long peduncle, which was covered with hairy bracts. Ovules were elongated, with the distal end of the integument attenuated to form an elongated micropylar tube. The vascular system of the cone is a narrow ring of scalariform tracheids surrounding a massive pith. Williamsonia gigas has also been reconstructed as an unbranched, erect plant with a stem covered by rhombic leaf scars (Williamson, 1870). Leaves were of the Zamites type and produced in a group at the crown. The species was based on such fragmentary material that the actual appearance of the plant continues to be controversial. Bucklandia (FIG. 17.77) includes Mesozoic stems bearing relatively large leaf scars. The name encompasses various preservational modes, including compressions, casts, and petrifactions (Sharma, 1967). Harris (1969) reconstructed a portion of a plant with Bucklandia pustulosa stems, cones of Williamsonia leckenbyi, and foliage of Ptilophyllum pecten. The axes were 14.0 cm in diameter and produced laterals

Figure 17.76 Suggested reconstruction of Williamsonia sewardiana. (Courtesy T. Delevoryas.)

that bore the reproductive organs and leaves. Scattered over the surface of the stems were leaf scars and what are interpreted as lenticels. Ischnophyton is a Late Triassic bennettitalean with a slender stem (1.0 cm in diameter) and pinnate fronds from the Pekin Formation of North Carolina, USA (Delevoryas and Hope, 1976). The surface of the stem is slightly wrinkled, but there are no persistent leaf bases present. Leaves are 20.0 cm long, with pinnae attached to the upper surface of the rachis. When compared with other bennettitalean foliage types, the leaves of I. iconicum appear to share features with forms assigned to both Otozamites and Zamites. Nothing is known about the reproductive parts of the plant.

734

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 17.78 Williamsonia sp. williamsonian reproductive

organ (Early Cretaceous). Bar  1.0 cm. (Courtesy BSPG.)

Figure 17.77 Suggested reconstruction showing Bucklandia

stem, Williamsonia cones, and Ptilophyllum leaves. (Modified from Watson and Sincock, 1992.)

OVULATE STRUCTURES The genus Williamsonia (Delvoryas, 1991) is also used for monosporangiate cones preserved as either impression– compression (FIG. 17.78) or structurally preserved forms (Sharma, 1969a; Rothwell and Stockey, 2002). The cone is generally cup shaped and in some species, such as W. margotiana, may be up to 15.0 cm in diameter (Watson and Sincock, 1992). In the center of a Williamsonia cone is a dome-shaped receptacle bearing numerous small seeds and tightly packed interseminal scales (FIG. 17.79). The elongated micropylar tubes of the seeds extend above the level of the scales (FIG. 17.80). Subtending the seed-bearing receptacle is a whorl of bracts that can give the structure a somewhat flowerlike appearance. In W. bryonyae the bracts are 7.0 cm long, whereas the seed-bearing receptacle is just a few millimeters in diameter (Watson and Sincock, 1992). Compact cones of W. netzahualcoyotlii from the Middle Jurassic of Oaxaca, Mexico, range from 2.0 to 3.0 cm in diameter and exhibit a small peduncle scar at the base

17.79 Diagrammatic cut away reconstruction of Williamsonia margotiana. (From Watson and Sincock, 1992.)

Figure

(FIG. 17.81) (Delevoryas and Gould, 1973). The number of ovules present on the receptacle is estimated to have been between 25 and 50, and apparently not all reached maturity at the same time. At the distal end of the cone is a ring of fused scales that form the corona. The development of Williamsonia ovules, including the formation of the pollen chamber, appears somewhat similar to Cycadeoidea (Sharma, 1974). Williamsonia diquiyui is a permineralized specimen from the Jurassic of Mexico that contains dicotyledonous embryos in the seeds (Delevoryas and Gould, 1973). A permineralized ovulate cone (FIG. 17.82) from the Upper Cretaceous of British Columbia, Canada, provides further details about the development

CHAPTER 17

CYCADOPHYTES

735

Interseminal scale Stalked ovule Stalked ovule

Surface of gynoecium

Micropylar plate Micropyle Interseminal scale

Micropyle Scale heads

Figure 17.80 Arrangement of stalked ovules and interseminal scales in Williamsonia. (From Watson and Sincock, 1992.)

Figure 17.81 Reconstruction of Williamsonia netzahualcoyotlii. (From Delevoryas and Gould, 1973.)

Figure 17.82 Longitudinal section of Williamsonia bockii cone

of seeds and embryos in the williamsonians. Interseminal scales and ovules extend from the receptacle in W. bockii in an approximate 300° arc (Stockey and Rothwell, 2003). Ovules are tightly compacted (FIG. 17.82), ellipsoidal, and 8 mm long (FIG. 17.83). The presence of numerous pollen tubes (FIG. 17.84) between the nucellus and integument in W. bockii lends support to the interpretation that williamsoniaceous plants possessed an anemophilous pollination syndrome. The genus Williamsoniella is a bisporangiate cone of Jurassic age (Thomas, 1915; Harris, 1944; Cridland, 1957). It consists of a whorl of wedge-shaped, pinnate microsporophylls surrounding a reduced receptacle bearing interseminal scales and seeds (FIG. 17.85). In W. coronata the cone is 2.0 cm long, with microsporophylls extending out 1.5 cm. Pollen sacs are synangiate like those of Cycadeoidea, and

the pollen is monosulcate and smooth walled (Harris, 1969). Based on their common occurrence, Harris (1969) suggested that W. coronata was produced by a plant that bore leaves of the Nilssonia type. Cones are thought to have been borne in the axils of leaves and in the angle between two branches. Pollen from some Williamsoniella cones, such as W. lignieri, has been compared with dispersed grains of Exesipollenites (Harris, 1974). These grains are 25 μm in diameter and monoporate. Cones produced in the angle of two branches appear to have been the habit in Wielandiella as well. This Late Triassic genus produced leaves of the Anomozamites type beneath the cone (FIG. 17.86). Another ovulate reproductive structure known from the Upper Triassic of East Greenland is Vardekloeftia (Harris,

showing seeds and interseminal scales (Cretaceous). Bar  8.0 mm. (Courtesy R. A. Stockey.)

736

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 17.84 Detail of Williamsonia bockii seed apex show-

ing pollen tube (arrow) (Cretaceous). Bar  200 μm. (Courtesy R. A. Stockey.)

Figure 17.83 Longitudinal section of Williamsonia bockii seed

and interseminal scales (Cretaceous). Bar  1.0 mm. (Courtesy R. A. Stockey.)

1932a; Pedersen et al., 1989a). Cones are preserved as coaly compressions in the form of spheres, each up to 3.0 cm in diameter. On the surface are interseminal scales and 20 ovoid seeds (FIGS. 17.87, 17.88). The seed is surrounded by another cutinized layer that has been interpreted as a cupule which terminates just below the seed apex (Pedersen et al., 1989a). Within the micropyles of the seeds are monosulcate pollen grains. Leaves of Ptilophyllum kochii occur in the same rocks and this foliage is thought to have been produced by Vardekloeftia. Bennetticarpus is a morphogenus used for reproductive structures (FIGS. 17.89, 17.90) showing distinct bennettitalean characters, but which are insufficiently preserved to be assignable to more completely understood genera (Watson and Sincock, 1992). Fredlindia fontifructus is an enigmatic, presumably ovulate cone from the Upper Triassic Molteno Formation in South Africa (Anderson and Anderson, 2003). The structure

Figure 17.85 Longitudinal section of Williamsoniella reproduc-

tive structure showing arrangement of microsporophylls and seedbearing receptacle. (From Harris, 1969, in Watson and Sincock, 1992.)

CHAPTER 17

CYCADOPHYTES

737

Figure 17.88 Details of Vardekloeftia sulcata seeds (Triassic). Bar  5.0 mm. (From Pedersen et al., 1989a.)

Figure 17.86 Suggested reconstruction of a williamsoniacean

plant with bisexual cones and leaves of the Anomozamites type. (From Nathorst, 1902, in Watson and Sincock, 1992.)

Figure 17.89 Suggested reconstruction of Bennetticarpus

antoinetteae. (From Watson and Sincock, 1992.)

Figure 17.87 Vardekloeftia sulcata reproductive structure showing hexagonal impressions of interseminal scales and two seeds (Triassic). Bar  1.0 cm. (From Pedersen et al., 1989a.)

consists of an elongate axis 10 cm long that bears bilaterally symmetrical, apparently fleshy structures (described as “gynoecia”) in a series of whorls (FIG. 17.91). Each of these flattened structures consists of a tongue-shaped

(FIG. 17.92), thickened lamina, 2.5 cm long by 10 mm wide, which bears a honeycomb aggregate of columnar, presumably ovuliferous units on its abaxial surface. In surface view, these ovuliferous units are similar to the ovules and interseminal scales of some bennettitaleans, although an order of magnitude smaller; no ovule has been recovered to date. If each of these flattened units is homologous with the receptacle of the Bennettitales, they differ markedly in being

738

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 17.90 Bennetticarpus wettsteinii receptacle with seeds

(arrows) (Triassic). Bar  2.0 cm. (Courtesy C. Pott.)

bilaterally symmetrical and in being attached in whorls to an axis. As a result, Anderson and Anderson (2003) erect a new order and family for Fredlindia but include it in the class Bennettitopsida. POLLEN ORGANS Weltrichia (FIG. 17.93) encompasses pollen cones assignable to the Williamsoniaceae which were formerly called Williamsonia (Braun, 1849; Schuster, 1911; Harris, 1969). Earlier interpretations that ovulate Williamsonia cones contained microsporophylls like those of Cycadeoidea are now known to have been in error. The pollen cones appear to exhibit more morphologic variability than their ovule-bearing counterparts, including the position and shape of the pollen sacs and number and size of the sterile bracts (Li et al., 2004a). Weltrichia sol (Williamsonia gigas) from the Jurassic of Yorkshire is a large pollen cone that consists of an open urn- or cup-shaped receptacle 10.0 cm in diameter. Extending from the rim of the cup are 30 microsporophylls, each 6.0 cm long. These are distally tapered and bear numerous pollen sacs on the inner surface. Pollen grains are thin walled, monosulcate, and 46 μm long. On the inner surface of the cup are numerous semicircular structures that look like synangia but that lack pollen. These have been referred to as resinous sacs (Harris, 1969) and may have functioned as some form of pollination attractant. The uniformity of the stomatal pattern and consistent occurrence of W. sol cones with bracts of Williamsonia gigas and

Figure 17.91 Fredlindia fontifructus, ovulate strobilus with

bilaterally symmetrical “gynoecia” borne in a series of whorls (Triassic). (From Anderson and Anderson, 2003.)

leaves of Zamites gigas have been used to suggest that these parts represent organs of the same plant. In another species, Weltrichia pecten, the basal cup is smaller (4.0 cm wide) and the edge is dissected into only 10–12 microsporophylls. Synangia are borne in two rows on the inner surface of the microsporophyll. These pollen cones are continuously associated with Ptilophyllum pecten leaves. Weltrichia hirsuta from the upper Liassic (Lower Jurassic) of Iran is one of only a few species in the genus Weltrichia in

CHAPTER 17

CYCADOPHYTES

739

Figure 17.94 Westersheimia pramelreuthensis (Triassic). Bar  5.0 mm. (Courtesy C. Pott.)

Figure 17.92 Fredlindia fontifructus, bilaterally symmetri-

cal “gynoecium” composed of a tongue-shaped lamina and abaxial honeycomb aggregate of ovuliferous units (Triassic). (From Anderson and Anderson, 2003.)

17.93 Weltrichia microdigitata showing radiating microsporophylls (Jurassic). Bar  6.0 mm.

Figure

which the sporangia are not attached to the inner side of the microsporophyll, but rather occur on elongate, forked sporangiophores developed from the apical lobes of the hirsute sporophylls (Schweitzer and Kirchner, 2003). Another species

showing this arrangement of sporangia is W. harrisiana from the Middle to Upper Jurassic of Kachchh in India (Bose and Banerji, 1984). There are several less well-known bennettitalean pollen organs described from Upper Triassic rocks. One is Bennettistemon amblum, a lanceolate-shaped microsporophyll with syndetocheilic stomata (Harris, 1932a). On the adaxial surface are a series of flanges, each bearing 20–30 sporangia. In B. ovatum, the entire surface of the sporophyll is covered with pollen sacs. Both species produced monosulcate pollen. The presence of syndetocheilic stomata on Haitingeria (Kräusel, 1949a), Leuthardtia (Kräusel and Schaarschmidt, 1966), and Leguminanthus (Kräusel and Schaarschmidt, 1966) suggest that these organs from the Carnian (Late Triassic) floras of Neuewelt near Basel, Switzerland and Lunz in Lower Austria may also represent detached microsporophylls of bennettitaleans (Crane, 1986). The affinities of other male reproductive organs from these famous floras, such as Westersheimia (FIG. 17.94), Lunzia (FIG. 17.95), and Pramelreuthia (Krasser, 1916, 1917), are still inconclusive. Pramelreuthia is a planar pinnate structure composed of a naked axis bearing stalked synangia (FIG. 17.96), each containing up to 20 pollen sacs, in subopposite or opposite pairs. The taxon has also been recorded from the Upper Triassic Chinle Formation of the southwestern United States (Ash and Litwin, 1996). Based primarily on the occurrence of bisaccate pollen grains in the sporangia, Ash and Litwin (1996) questioned the affinities of this taxon within the cycadophytes. DISCUSSION: BENNETTITALES

The Bennettitales have received considerable attention during the last several decades as they have figured prominently

740

PALEOBOTANY: THE BIOLOGY AND EVOLUTION OF FOSSIL PLANTS

Figure 17.96 Pramelreuthia haberfelneri (Triassic). Bar  5.0 mm. (Courtesy C. Pott.) Figure 17.95 Lunzia austriaca showing pollen sacs (arrows)

(Triassic). Bar  2.0 mm. (Courtesy C. Pott.)

in discussions dealing with relationships among seed plants and the origin of the angiosperms. Phylogenetic analyses based on morphological and anatomical features have placed them within the anthophyte clade, a group that also includes the flowering plants (Nixon et al., 1994; Doyle, 2006). This inclusion, however, may have more to do with the fact that many specimens of bennettitaleans include both vegetative and reproductive structures, and that they are anatomically preserved, thus providing many more characters to consider. To some degree, this situation makes it difficult to determine whether or not a plant like Cycadeoidea with bisporangiate reproductive organs is, in fact, a derived form (Crane, 1988). One often-cited difference between flowering plants and bennettitaleans is the bitegmic ovules of the former. Crane (1986) suggested that perhaps bennettitalean ovules did possess a double integument, the outer integument being homologous with what has been termed the cupule in bennettitaleans such as Vardekloeftia (Pedersen et al., 1989a). In this hypothesis, both the bitegmic ovule, consisting of an

ovule and its cupule, and the interseminal scales are considered to represent highly reduced megasporophylls. An alternative hypothesis was presented by Rothwell and Stockey (2002), who view the cupule in Vardekloeftia as originating from the seed integument. It has also been suggested that the interseminal scales are homologous with the angiosperm carpel through fusion of the scales around each ovule. The beautifully preserved seeds and scales in Cycadeoidea maccafferyi (Rothwell and Stockey, 2002), however, show that most ovules share scales with neighboring ovules, so production of additional scales would be necessary for this scenario. The pollen organs of the Bennettitales have been homologized with those of certain Paleozoic seed ferns. For example, there are numerous Paleozoic pollen organs that are synangiate, and some of these are known to have been borne on a pinnate, leaflike organ that may be homologous with the microsporophylls of Cycadeoidea. Delevoryas (1968) first suggested that the bisporangiate cone of Cycadeoidea represents a portion of a foliar system, containing seeds and pollen organs, which became phylogenetically reduced to the position it occupies interspersed among the leaf bases on the

CHAPTER 17

Figure 17.97 Suggested reconstruction of Monanthesia sp.

(background) and Cycadeoidea dacotensis (foreground). (Courtesy J. Watson.)

stem. This transformation of the reproductive organs may be concomitant with the reduction of overall plant habit from narrow, elongated stems to the shorter, squat trunks seen in Cycadeoidea. Fossil evidence indicates that the earliest members of the Cycadales also had elongated, small-diameter stems (Delevoryas, 1982). The reproductive organs of the Bennettitales include both monosporangiate and bisporangiate forms. As noted earlier, all of the reproductive structures in the Cycadeoidaceae, such as Cycadeoidea and Monanthesia (FIG 17.97), are bisexual (bisporangiate). In the Williamsoniaceae, with separate ovulate and pollen organs, these structures can either be borne on the same plant (monoecious) or on two different plants (dioecious). Unfortunately, the distribution of reproductive

CYCADOPHYTES

741

organs on individual plants in the Williamsoniaceae, as is the case with most fossil seed plants, is not known. One interesting avenue of research about the Bennettitales focuses on how the reproductive organs functioned at the time of pollination and ovule maturation. There is increasing evidence that many fossil seed plants had developed more than chance associations with insects, which may have served as secondary pollinators as early as the Pennsylvanian. For example, Harris (1969) noted that the resin-containing “synangia” on the bracts of the pollen cone Weltrichia may represent some form of attractant. Others have described coprolites in cones of several species of Cycadeoidea (Crepet, 1974; Stockey and Rothwell, 2003), further suggesting that insects were associated with at least some cycadeoids. Although some members of Cycadeoidea appear to have been self-pollinated, the open morphology of the cones in the Williamsoniaceae suggests that outcrossing may have been the norm in those plants. There are, however, still a number of unanswered questions about williamsonian reproductive structures. For example, did the microsporophylls of Williamsonia and Weltrichia expand at maturity so that pollen could be easily transported from pollen sacs to micropyles, perhaps by wind? There is a high frequency of detached Williamsonia cones in which the seeds are all at the same stage of development. This sort of evidence in other fossils has been used to suggest that the reproductive organs abscised before seed development was complete or even before pollination, but these cones could also have been detached from the parent plant in a storm. The beautiful preservation of Williamsonia bockii shows an intact reproductive organ, including megagametophytes and embryos in some of the seeds, providing evidence that development was completed while the cone was still intact, at least in this plant (Stockey and Rothwell, 2003). Taphonomic evidence would be helpful in trying to explain dispersed cones with immature ovules. In addition to the open morphology of the cones, outcrossing could also have been achieved by nonsynchronous development of cones within a population. Although the bennettitaleans are some of the earliest seed plants to be studied, the complexities of the group are only now becoming apparent. It appears that not only do they demonstrate a wide range of morphological and structural features, but the fossils also suggest interesting patterns in their reproductive biology. The questions that are posed earlier are but a few of the fascinating aspects of these seed plants, and the ones that promise exciting contributions to the biology and phylogenetic history of this group in the future.

This page intentionally left blank

18 Ginkgophytes Paleozoic record................................................................ 744

Umaltolepidiaceae.............................................................................752

Ginkgophyte wood ............................................................747

Yimaiaceae ........................................................................................752 Schmeissneriaceae ............................................................................753

Ginkgophyte foliage ........................................................747 Taxa with uncertain affinities ...............................754 Pollen-producing structures ................................ 750 Conclusions ...........................................................................755 Ginkgophyte plants ......................................................... 750 Ginkgoaceae ......................................................................................750 Karkeniaceae .....................................................................................752

The trees are coming into leaf Like something almost being said “The Trees” by Philip Larkin

The ginkgophytes are an ancient group of gymnosperms that is believed to have originated during the late Paleozoic, although Paleozoic fossils are rare and often difficult to interpret. For example, large fan-shaped leaves from the Carboniferous, reminiscent of Ginkgo leaves, have been described as Ginkgophytopsis (Chapter 16), and variously interpreted as fern aphlebia, seed fern or cordaite foliage, or as leaves of an early ginkgophyte or ginkgophyte progenitor. The ginkgophytes had their highest diversity and widest distribution in the Mesozoic, after which the group drastically declined during the Cenozoic (Z. Zhou and Wu, 2006). Today they are represented by a single species, the maidenhair tree Ginkgo biloba, which is planted as an ornamental in temperate areas of the world, both because of its beauty (FIG. 18.1) and its resistance to air pollution and attacks by fungi and insects. Ginkgo is sometimes referred to as a living fossil because it had a worldwide distribution beginning in the Mesozoic (Tralau, 1968; Z. Zhou, 1997). The megagametophyte of the Ginkgo seed has been used as a food source in China and Japan, and extracts of the leaves have been used in Chinese medicine for several thousand years.

Although it is an extremely slow grower, G. biloba is an attractive, stately tree that may reach a height of almost 30 m. The primary vascular system consists of a eustele and is surrounded by a bifacial vascular cambium. Shoot dimorphism is a characteristic feature of the genus, with spur or short shoots produced in the axil of leaves borne on the long shoots. Both long and spur shoots are developmentally interchangeable. Secondary xylem is composed of thin-walled tracheids and narrow vascular rays. Tracheal pitting is circular bordered and restricted to the radial walls. On some of the old cultivated specimens, woody structures similar to burls, which are termed chi chi, are produced on the lower side of large branches. Some become quite large (30 cm in diameter), and grow downward; if they contact the ground they may produce roots and leaves. One distinctive morphologic feature of the genus is its fan-shaped leaf with open dichotomous venation pattern, although some leaves do show vein anastomoses. There is considerable variability in the leaf margin, ranging from entire to deeply notched, and this can be related to the position of the leaf or series of leaves on the shoot system.

743

744

PaleoBOtany: the biology and evolution of fossil plants

flagellated, free-swimming male gametes (spermatozoids); this process usually occurs on the tree. Ginkgo biloba and cycads (Chapter 17) are the only extant seed plants that produce flagellated male gametes (Norstog et al., 2004; Vaughn and Renzaglia, 2006); for details about gametophytegamete formation and the fertilization process, as well as information on the history of discovery of spermatozoids in G. biloba, the reader is referred to Lee (1955) and Ogura (1967), and references cited therein. After fertilization, the ovule develops into a yellowish, plumlike “fruit” up to 2.5 cm long that consists of a single large seed with a fleshy outer layer. The mature seed is anatomically similar to that of a cycad, except that in Ginkgo the vascular system consists of two bundles that extend through the innermost region of the integument. Ginkgophytes appear to have persisted relatively unchanged in their leaf morphology since the Mesozoic, and there are reports of ginkgophyte foliage from many stratigraphic levels and geographic regions in both the Northern and Southern Hemispheres. Figure 18.1 Decorative lapel pin showing Ginkgo leaves.

HigheR TAxa in THIS chapter:

Colonization of G. biloba roots by endomycorrhizal fungi of Glomus sp. and subsequent establishment of VAmycorrhizae is a common phenomenon (Fontana, 1985; Aoki, 1997). An interesting report by Trémouillaux-Guiller et al. (2002) documents a second, much more strange association in G. biloba, in which Coccomyxa-related green algae and their immature precursor forms live as intracellular endophytes in the leaves. This latter type of algal association has not been documented for any other group of plants but does occur in certain animals (e.g., cnidarians). Ginkgo biloba produces separate ovulate and pollenbearing structures and it is dioecious, that is the structures are borne on different plants. Pollen organs resemble lax, angiospermous catkins and are borne in the axils of bud scales or leaves on the spur shoots (Gifford and Foster, 1989). Attached to the strobiloid axis are numerous fertile appendages that have been termed sporangiophores or sporophylls, each bearing two to four pollen sacs at the tip. Pollen is prolate and germinates from the distal surface. The ovules are borne on stalks that arise in the axil of a leaf on a short shoot. Each stalk generally bears two ovules on either side of the stalk (Douglas et al., 2007). Each ovule is surrounded at its base by a rimlike collar, which has been variously interpreted as the homologue of a megasporophyll, a second integument, or an aril. Developmental studies, however, suggest that it is none of these and is a structure unique to Ginkgo (Douglas et al., 2007). The ovule is fertilized by

Ginkgoales Ginkgoaceae Karkeniaceae Umaltolepidiaceae Yimaiaceae Schmeissneriaceae

Paleozoic record Fossils are known from the Permian that have been variously interpreted as either ginkgophytes or ginkgophyte ancestors. One of the oldest genera regarded as having affinities with the ginkgophytes, perhaps as the sister group to all Mesozoic ginkgoaleans, is Trichopitys heteromorpha (FIG. 18.2) from the Lower Permian of Lodève in southern France (Florin, 1949; Z. Zhou, 1991). Compressed specimens consist of vegetative shoots with helically arranged, nonlaminar leaves. The leaves are slightly decurrent at the base and up to 10 cm long; each is dissected into four to eight segments. Along some of the vegetative axes are zones of fertile structures. The fertile structures each consist of a poorly defined axis borne in the axil of a leaf (FIG. 18.3); the axis bears two to six helically arranged branches, each terminated by a single recurved ovule. Ovules are slightly flattened and about 6 mm long. The ovulebearing branches of T. heteromorpha are morphologically

CHAPTER 18

Ginkgophytes

745

Figure 18.2 Fertile shoot of Trichopitys heteromorpha. (From

Andrews, 1961.)

Figure 18.4 Suggested reconstruction of Polyspermophyllum sergii branch. (From Taylor and Taylor, 1993.)

Figure 18.3 Trichopitys heteromorpha fertile shoot (arrow)

(Permian). Bar  1 cm. (Courtesy H. Kerp.)

similar to some aberrant ovulate structures known from G. biloba that may have up to 15 stalked ovules per shoot. Florin suggested that Trichopitys was a primitive member of the ginkgophyte line based on the organization of the ovulatebearing structures and the vegetative leaves. Another putative Permian ginkgophyte known from both vegetative and fertile specimens is Polyspermophyllum (Archangelsky and Cúneo, 1990). The specimens come from Chubut Province, Argentina, and include branches that bear both fertile and sterile appendages (FIG. 18.4). The helically arranged, linear leaves of P. sergii dichotomize several times and are up to 10 cm long. Each leaf segment has a single vein and on the abaxial surface are two furrows. Morphologically, they appear similar to the leaves of Dicranophyllum (Barthel, 1977), a genus used for certain enigmatic Late Pennsylvanian and Early Permian gymnosperms that are characterized by unbranched stems up to about 2 m tall (FIG. 18.5), to which are attached linear, forked leaves (FIG. 18.6). Reproductive

746

PaleoBOtany: the biology and evolution of fossil plants

Figure 18.5 Dicranophyllum hallei, leafy stem (Permian). Bar  3 cm. (Courtesy R. Noll and H. Kerp.)

structures in these dioecious plants occur in the form of male and female cones clustered into distinct fertile zones along the stem (FIG. 18.7) (Barthel et al., 1998; Barthel and Noll, 1999; Wagner, 2005). However, ovule-bearing organs of Polyspermophyllum consist of a dichotomous branching system, with the ultimate segments recurved and each terminating in an ovule (FIG. 18.8). Isolated seeds thought to belong to P. sergii are 4 mm long and ornamented by longitudinal striations. At the apex is a bifid extension of the integument. Archangelsky and Cúneo (1990) suggested a more formal

Figure 18.6 Dicranophyllum hallei, leafy stem (Permian).

Bar  4 cm. (Courtesy BSPG.)

classification that would include Polyspermophyllum, together with Trichopitys and Dicranophyllum, in the Dicranophyllales. This order includes woody plants with dichotomous, elongate, linear leaves and compound ovulate structures. Others suggest Trichopitys is related to the peltasperms (Meyen, 1987) and that Dicranophyllum represents an early coniferophyte (Rothwell et al., 2005).

CHAPTER 18

S

P

Ginkgophytes

747

that Ginkgo wood is particularly susceptible to cell wall degradation in contrast to the greater chemical resistance of many gymnospermous and angiospermous woods (Scott et al., 1962). However, fossil ginkgophyte wood may simply be difficult to recognize or to discriminate from conifer wood in some instances (Zheng and Wang, 2000). Permineralized wood displaying ginkgoalean anatomy is usually assigned to the morphogenus Ginkgoxylon, but other taxa, such as Ginkgomyeloxylon and Protoginkgoxylon (Giraud and Hankel, 1986; Zheng and Zhang, 2000), are also used (Süss, 2003). Charcoalified remains of ginkgophytalean short shoots from the Upper Cretaceous Peruc-Korycany Formation in the Czech Republic have been described as Pecinovicladus kvacekii by Falcon-Lang (2004). Subsequent studies suggested that these short shoots belong to the whole-plant concept of Nehvizdyella bipartita (Kvacek et al., 2005) (discussed below).

Ginkgophyte foliage

Figure 18.7 Dicranophyllum hallei, stem with sterile leaves

and laterally positioned reproductive structures containing pollen (P) and seeds (S) (Permian). Bar  1 cm. (Courtesy M. Barthel.)

Figure 18.8 Portion of a branching system of Polyspermophyllum sergii with terminal ovules (arrows) (Permian). Bar  2.5 cm. (From Archangelsky and Cúneo, 1990.)

Ginkgophyte wood Wood with characteristic Ginkgo-like anatomy is rare in the fossil record. Some believe this may be due in part to the fact

There are numerous reports of fossil leaves that are indistinguishable from the leaves of modern Ginkgo biloba (FIG. 18.9). Some have been referred to the extant taxon Ginkgo, whereas others have been designated as Ginkgoites (FIG. 18.10) (Florin, 1936b, c; Cúneo, 1987). Harris and Millington (1974) proposed that the name Ginkgoites be abandoned because the genus is poorly defined and cannot be applied with consistency to all the morphologic forms known to have existed within a single genus. Watson et al. (1999), however, argued in favor of retaining Ginkgoites as a morphogenus for foliage lacking associated reproductive organs (FIGS. 18.11, 18.12). In spite of taxonomic ambiguities, the name continues to be routinely used for Ginkgo-like leaves that cannot be attributed to a natural genus (Z. Zhou, 1997). It is interesting to note that the cuticle fine structure of extant Ginkgo and Ginkgoites tigrensis is quite similar (Del Fueyo et al., 2006; Villar de Seoane, 1997). Others have used Ginkgo for various leaf morphologies that are interpreted as being related to the ginkgophytes (L. Zhao et al., 1993). Ginkgo digitata is a relatively common Jurassic species that has a wedge-shaped lamina with the distal margins dissected by shallow indentations. The veins are numerous and small resin bodies are common in the intercostal areas. Stomata are crowded on the lower surface in the intercostal areas and trichomes are present along the veins. Ginkgo huttonii is similar to G. digitata, but characterized by more deeply divided leaves and a wedge-shaped lamina with crowded veins. Seeds, bract scales, and pollen

748

PaleoBOtany: the biology and evolution of fossil plants

Figure 18.9 Ginkgo adiantoides (Miocene). Bar  1 cm.

(Courtesy BSPG.)

Figure 18.11 Ginkgoites tigrensis showing attachment of several leaflets (Cretaceous). Bar  1 cm.

Figure 18.12 Ginkgo sp. leaf (Cretaceous). Bar  1 cm. (Courtesy J. Watson.)

Figure 18.10 Ginkgoites tigrensis showing basal venation

(Cretaceous). Bar  1 cm.

cones from the Jurassic of Yorkshire are also referred to G. huttonii (FIG. 18.13) because of the consistent association of this foliage species in the same beds. The outer cuticle of the integument of the seeds is preserved and shows impressions

of epidermal cell patterns. Another species from the Lower Cretaceous of Montana is G. pluripartita (Brown, 1975b). The leaves of this form are relatively small; the largest is about 2 cm long. The lamina is typically divided into four or six segments, with the central sinus most prominent. Leaves are hypostomatic, and stomata confined to the intercostal areas. Subsidiary cells have prominent, overarching papillae. An example of the diversity of Ginkgo-like leaves extending from the Jurassic to the Lower Cretaceous in northeast China is provided by Z. Zhao et al. (1993).

CHAPTER 18

Figure 18.13 Suggested reconstruction of Ginkgo huttonii pollen cone. (From Taylor, 1988b.)

A widespread foliage taxon with possible affinities in the ginkgophytes is Sphenobaiera (FIG. 18.14). This morphogenus, which extends from the Early Permian well into the Cretaceous, is used for highly dissected, wedgeshaped leaves that have dichotomous venation but lack a distinct petiole. The absence of a petiole distinctly offset from the lamina can be used to separate Sphenobaiera from the morphogenera Ginkgo, including Ginkgoites, and Baiera (FIG. 18.15) (Florin, 1936b). One of the betterknown species of Sphenobaiera is S. longifolia (Harris and Millington, 1974) from the Middle Jurassic of Yorkshire. Leaves are at least 13 cm long and divided four times into narrow ultimate segments. Leaves are amphistomatic, and the cuticle is thickest on the adaxial surface. Stomata on the upper surface are irregularly oriented in bands, with the guard cells surrounded by six papillate subsidiary cells. In S. ophioglossum, some leaves are slightly less than 20 cm long and have from two to four broad lobes. Another foliage type from the Lower Permian of Argentina, Chiropteris, can be distinguished by a larger number of vein anastomoses (Archangelsky, 1960). Oblanceolate leaves with an entire

Ginkgophytes

749

Figure 18.14 Sphenobaiera digitata (Permian). Bar  1 cm. (Courtesy M. Barthel.)

Figure 18.15 Baiera muensteriana (Triassic–Jurassic). Bar 

1 cm. (Courtesy BSPG.)

750

PaleoBOtany: the biology and evolution of fossil plants

margin and closely spaced, dichotomizing veins have been placed in the genus Eretmophyllum (Hluštík, 1986), whereas highly dissected leaves with a single vascular trace include Baierella, Euryspatha, Baieraphyllites, and Kirjamkenia. Other putative ginkgophyte foliage types are listed in Z. Zhou (1997) and Z. Zhou and Wu (2006), and although some show similarities in morphological features to one another and to modern Ginkgo leaves, they differ in venation patterns, for example Kerpia (FIG. 18.18) (Naugolnykh, 1995). Fossil Ginkgo leaves have become increasingly important in reconstructions of Mesozoic and Cenozoic environments based on their stomatal characters and isotopic composition (Chen et al., 2001; Royer et al., 2001; Beerling and Royer, 2002; Sun et al., 2003).

Pollen-producing structures Pollen organs of fossil ginkgophytes are not well known. One putative pollen organ has been reported from the Yorkshire Jurassic (Harris and Millington, 1974). It consists of loosely arranged, once-dichotomized appendages, each terminated by what is believed to be a pollen sac. Pollen grains adhering to the sac are elongate (29–42 μm long) and ornamented with a sulcus on the presumed distal surface. Ginkgo liaoningensis, from the Lower Cretaceous of China, consists of a cluster of highly reduced microsporophylls arising from a central axis. Each sporophyll bears two to four pendulous pollen sacs with longitudinal dehiscence and monocolpate pollen (Liu et al., 2006a). At the tip of the sporophyll is a triangular extension thought to contain mucilage. Pollen sacs together with seeds and leaves have also been reported from the Upper Cretaceous of Canada (Serbet, 1996, 1997). Another putative Cretaceous pollen organ that may have affinities with either the ginkgophytes or some Mesozoic pteridosperm group is Brenneria (Pedersen et al., 1993), described from Drewry’s Bluff in Zone I of the Potomac Group, Virginia, USA. This interesting microsporangiate organ produces pollen of Decussosporites, a bisaccate grain with a sulcus that is perpendicular to the long axis of the grain on the proximal surface. Identical pollen is found in seeds of Brennerispermum from the same site.

or have reassembled stems, foliage, and reproductive structures based on correspondences in macromorphology, epidermal anatomy, and/or internal anatomy. Z. Zhou (1991) distinguishes five lineages of Mesozoic ginkgophytes that are known as whole plants: (1) Ginkgo, in the Ginkgoaceae, which also includes Grenana (Samylina, 1990) and Nehvizdyella (Kvaˇcek et al., 2005); (2) Karkenia in the Karkeniaceae (Archangelsky, 1965); (3) Toretzia and Umaltolepis in the Umaltolepidiaceae (Krassilov, 1969, 1972b; Stanislavsky, 1973); (4) Yimaia in the Yimaiaceae (Zhou and Zhang, 1992); and (5) Schmeissneria in the Schmeissneriaceae (Kirchner and Van Konijnenburg-Van Cittert, 1994). We will discuss the examples of the characteristic genera in each of these lineages later in this chapter. Ginkgoaceae

Ginkgo yimaensis (FIG. 18.16) is a Middle Jurassic ginkgophyte from Henan Province, China (Z. Zhou and Zhang, 1989a)

Ginkgophyte plants During the last several decades there have been a number of reports that have documented ginkgophyte leafy twigs with various types of reproductive structures in organic connection,

Figure 18.16 Suggested reconstruction of Ginkgo yimaensis.

(From Z. Zhou and Zhang, 1989a.)

CHAPTER 18

(FIG. 18.17). Leaves have a long petiole and a lamina divided into four to eight segments, with resin bodies present between the veins. Attached to short shoots are elongate pedicles, each terminating in an ovule with a mucronate tip. Although the leaves are different, the organization of the short shoot and ovulate structure is similar to that in G. biloba, especially some aberrant forms. Details of the megaspore membrane in the fossil are also similar to those of G. biloba (Z. Zhou, 1993). Based on ovulate reproductive organs and leaves it is hypothesized that there has been relatively little evolutionary change in Ginkgo since the Jurassic (Z. Zhou and Zheng, 2003), although the number of ovules and pollen sacs appears to have been reduced in the clade leading to extant G. biloba (X.-C. Liu et al., 2006a; Naugolnykh, 2005b, 2007) (FIG. 18.18). Nehvizdyella bipartita is a compound ovuliferous reproductive structure from the Upper Cretaceous (Cenomanian) Peruc-Korycany Formation, Czech Republic, that consists of a main axis and two short secondary axes, each terminated by large cupule-like structure, probably homologous  to the collar of the recent Ginkgo (J. Kvacek et al., 2005). Cupules enclose orthotropous ovules. In early developmental stages, the entire ovule and young seed are confined within the cupule, with the exception of the micropylar area.

Ginkgophytes

751

Monosulcate pollen grains of Cycadopites type have been found adhering to the seeds. Associated with N. bipartita is foliage assignable to Eretmophyllum obtusum (formerly Nehvizdya obtusa; see Hluštík, 1977, 1986; J. Kvacek, 1999; Gomez et al., 2000), short shoots described as Pecinovicladus kvacekii (Falcon-Lang, 2004), and trunk wood identified as Ginkgoxylon cf. G. gruettii. The co-occurrence of these fossils in monodominant taphocoenoses at four geographi cally distant localities has been used by J. Kvacek et al. (2005) to suggest that these remains all belong to the same plant (FIG. 18.19). Moreover, facies analysis indicates that this plant thrived in a coastal salt-marsh environment, and thus represents the first evidence for halophytes among the ginkgophytes.

Figure 18.18 Suggested transitional stages leading to modern

Figure 18.17

Zhiyan Zhou.

Ginkgo biloba. 1. Kerpia macroloba and Karkenia. 2. Karkenia spp. 3. Yimaia hallei. 4. Ginkgo yimaensis. 5. Ginkgo biloba. (From Naugolnykh, 2005b.)

752

PaleoBOtany: the biology and evolution of fossil plants

Figure 18.19 Suggested reconstruction of Nehvizdyella bipartita showing Eretmophyllum obtusum foliage and ovulate structures. (From J. Kvacˇek et al., 2005.)

Grenana is a morphogenus used for isolated leaves from the Middle Jurassic of Angren (Middle Asia) (Samylina, 1990). The leaves were initially interpreted as belonging to a seed fern, but later reinterpreted as a ginkgophyte (Z. Zhou, 1997). Associated with the leaves are ovuliferous reproductive structures that resemble Nehvizdyella bipartita in having seeds enveloped in a large cupule. Karkeniaceae

Karkenia is an ovulate reproductive organ that was originally described from the rich, Early Cretaceous Ticó flora of Santa Cruz province, Argentina (Archangelsky, 1965), but is known today from additional areas, including Germany (Kirchner and Van Konijnenburg-Van Cittert, 1994), Iran (Schweitzer and Kirchner, 1995), China (Z. Zhou et al., 2002), and New Zealand (Retallack, 1981). The type species, Karkenia incurva, consists of a short axis to which are attached more than 100 anatropous ovules (Del Fueyo and Archangelsky, 2001). Each ovule is 3 mm long and attached to the axis by a slender stalk. Several cuticle layers and a megaspore were macerated from each compressed seed. Closely associated in the assemblage are small woody axes believed to be spur shoots. These rocks also contained Ginkgoites tigrensis leaves (FIG. 18.10), which are interpreted as the foliage of the plant that produced the Karkenia reproductive organs. Karkenia hauptmannii from the Liassic (Lower Jurassic) of Germany, however, has been found attached to short shoots bearing leaves assignable to Sphenobaiera spectabilis (Kirchner and Van Konijnenburg-Van Cittert, 1994). Other species of Karkenia have been found associated with Sphenobaiera foliage in

Late Jurassic–Early Cretaceous rocks of Siberia (Krassilov, 1969), Rhaeto–Liassic (Late Triassic–Early Jurassic) deposits of Iran (Schweitzer and Kirchner, 1995), and the Jurassic of China (Z. Zhou et al., 2002). In these specimens, of which K. cylindrica from Iran is the largest at 12 cm long, the ovules were produced in what have been termed cones, each containing 100 ovules. Helically arranged ovules morphologically similar to those of Karkenia have been described from the Cretaceous of Russia as Semionangyma (Krassilov and Bugdaeva, 1988). The associated morphotaxa, however, suggest affinities with the cycads instead of the ginkgophytes. These include Cladophlebidium foliage and pollen cones of Semionandra which contain monosulcate pollen. Umaltolepidiaceae

Umaltolepis and Toretzia are ovuliferous reproductive structures characterized by one or two ovules per axis which were originally described from the Lower Cretaceous of Siberia in the case of Umaltolepis (Krassilov, 1972a) and from the Triassic of the Ukraine for the morphotaxon Toretzia (Stanislavsky, 1973). Umaltolepis has since been found in Iran (Schweitzer and Kirchner, 1995). Both reproductive structures resemble Nehvizdyella bipartita, but Umaltolepis has bracts beneath the ovule and at the base of the seedbearing axis (Krassilov, 1972a), which are absent in N. bipartita. Moreover, Umaltolepis lacks a cupule. Toretzia also lacks a cupule and has anatropous ovules, whereas they are  orthotropous in N. bipartita (J. Kvacek et al., 2005). Both Umaltolepis and Toretzia are associated with linear, ribbonlike leaves assignable to Pseudotorellia (Barnard and Miller, 1976). Yimaiaceae

Leaves assignable to the morphogenus Baiera (FIG. 18.20) have been reported to be associated with the ovulate organ Yimaia from the Middle Jurassic of China (X. Wu et al., 2006). These specimens are preserved in a paper coal and exhibit monocyclic stomata with sunken guard cells. Ovules are produced in groups of three to five at the terminal end of a peduncle. Yimaia is also described with leaves of Ginkgoites still attached to short shoots (Z. Zhou et al., 2007). The ovules Y. capituliformis are believed to have been helically arranged in a cluster. Morphologically similar seeds have been described from Greenland and Sweden and placed in the morphogenus Allicospermum. Seeds of Allicospermum found attached to a bifurcating peduncle in a small, cup-shaped collar are referred to as A. ginkgoidae (X. Yang et al., 2008). The presence of Ginkgoites regnellii leaves at the same locality,

CHAPTER 18

Ginkgophytes

753

Figure 18.20 Leaf of Baiera furcata. (From Taylor and Taylor,

1993.)

together with certain cuticular features, suggests both organs represent parts of the same Jurassic ginkgophyte. Schmeissneriaceae

Schmeissneria microstachys is a Liassic (Early Jurassic) plant from Germany with long and short shoots, to which are attached slender leaves and ovule-bearing structures consisting of an elongate axis (FIG. 18.21) up to 8 cm long, which bears up to 45 drop-shaped, cupulate ovules, each up to 3.5 mm long; mature ovules are winged and up to 5 mm long (Kirchner and Van Konijnenburg-Van Cittert, 1994). Associated with S. microstachys is the male fructification Stachyopitys preslii, which probably belongs to the same plant. The genus Stachyopitys is used for pedunculate pollenbearing fructifications composed of a main axis to which are attached unbranched or branched lateral secondary axes at irregular intervals. Each secondary axis produces a terminal cluster of radiating pollen sacs (Anderson and Anderson, 2003). Earlier interpretations suggested that Schmeissneria was allied with the ginkgophytes, but the plant also has features that might suggest affinities with the gnetophytes. A specimen of S. sinensis from the Haifanggou Formation (Middle Jurassic) in western Liaoning, northeast China, has been used to suggest affinities with the angiosperms (X. Wang et al., 2007). Here, the female organs (cupulate ovules) are borne in pairs on short peduncles that are arranged along an axis. Each of the presumed ovules has a central unit that is surrounded by an envelope with characteristic longitudinal

Figure 18.21 Ovule-bearing axis of Schmeissneria microstachys

(Jurassic). Bar  1 cm. (Courtesy BSPG.)

ribs. The apex of what is interpreted as a distal central unit is completely closed by the wall and each central unit has two locules (biloculate structure) that are completely separated by a vertical septum.

754

PaleoBOtany: the biology and evolution of fossil plants

Taxa with uncertain affinities There are numerous fossil plants for which the systematic affinities remain highly conjectural. A foliage morphogenus tentatively included with the ginkgophytes based on epidermal anatomy is Glossophyllum (Arberophyllum; see Doweld, 2000). It forms an isolated taxon that differs in various morphological traits from other members of the Mesozoic ginkgophytes, including Ginkgoites, Baiera, and Sphenobaiera (Tralau, 1968; Dobruskina, 1998). The most characteristic features of Glossophyllum are tongue-shaped leaves (FIG. 18.22), which are 10 cm long and lack petioles or subepidermal secretory cavities. Specimens of G. florinii, a Triassic form with a rounded apex from the Carnian (Upper Triassic) of the Northern Calcareous Alps of central Europe (Kräusel, 1943), have stomatiferous costal and non-stomatiferous intercostal fields on the abaxial surface. On the surface of the cuticle are multiple, longitudinally oriented striations that are hypothesized to be some sort of mechanical defense against phylloplane herbivores or

Figure 18.22 Glossophyllum florinii (Triassic). Bar  1 cm.

(Courtesy C. Pott.)

microorganisms (Pott et al., 2007e). Similar striations have also been recorded for the cuticles of a number of other fossil ginkgophytes. For example, Florin (1936) describes surface striations on ginkgophyte cuticles from the Wealden (Lower Cretaceous) of Franz Josef Land that are similar to the striae seen in G. florinii, and Denk and Velitzelos (2002) reported cuticular striae from Cenozoic representatives of the Ginkgoales. Since the reproductive structures of Glossophyllum remain unknown, the possibility exists that there are other groups of plants that possessed this foliage type, such as conifers or early gnetophytes. From the Upper Triassic Molteno Formation (Karoo Basin) of South Africa, Anderson and Anderson (2003) described a number of putative ginkgophyte fossils, including Avatia bifurcata, an ovuliferous reproductive structure composed of a once-forked axis bearing a pair of megasporophylls. Each megasporophyll consists of a single multiovulate head. The authors pointed out that these ovuliferous structures resemble anomalous fructifications sometimes produced by extant Ginkgo biloba, in which leafy expansions develop around the ovules (Krassilov, 1972a; Hara, 1997). Seeds of A. bifurcata are winged, however, in contrast to the fleshy seeds produced by G. biloba. Associated with Avatia are various

Figure 18.23 Hamshawvia baccata megasporophyll with seeds (arrows) (Triassic) Bar  5 mm. (Courtesy J. M. Anderson.)

CHAPTER 18

types of ginkgophyte leaves and microsporangiate strobili named Eosteria, which consist of a main axis bearing helically arranged, cupulate microsporophylls. Another ovuliferous structure from the Molteno Formation, which is also known from Australia (Holmes and Anderson, 2007) is Hamshawvia (Anderson and Anderson, 2003). This fossil is up to 4 cm long and composed of a once-forked axis bearing a pair of megasporophylls with ovules (FIG. 18.23). The attachment of H. longipedunculata ovulate structures and Sphenobaiera schenckii leaves (FIG. 18.24) indicates affinities in the ginkgophytes, given that Sphenobaiera was produced by ginkgophytes.

Conclusions Historically G. biloba was thought to be closely related to members of the Cycadales based on certain reproductive features such as flagellated male gametes. The hypothesis has also been advanced that the ginkgophytes diverged from some Carboniferous pteridosperm group and have evolved separately since then (Meyen, 1982b 1984) (FIG. 18.25). Ginkgophytes have also been related to conifers and cordaites (Doyle and Donoghue, 1986), suggested as a sister group to the cycadophytes (Chaw et al., 1997), and to a clade including conifers and gnetophytes (Bowe et al.,

Ginkgophytes

755

2000). A recent comparative morphological and morphogenetic study of male sporangiophores (Mundry and Stützel, 2004a) alludes to a relationship of G. biloba with Coniferales, Gnetales, and Cordaitales, and rejects a close relationship with groups that have pinnately organized sporophylls, such as Cycadales, as assumed by some molecular studies. Another interpretation of the evolution within the Ginkgoales has been proposed by Z. Zhou (1991). Based on a cladistic analysis using cordaites and early conifers as the outgroup, the author suggested that the Ginkgoales are a monophyletic group that included Trichopitys, Karkenia, Toretzia, Umaltolepis, Yimaia, and Ginkgo. According to this phylogeny, the ovule-bearing axes of Trichopitys are primitive, with a decrease in the number of ovules and shoot length in the short shoots occurring in geologically later and more derived taxa. Florin (1949) speculated that the ovule-bearing shoots of Ginkgo were homologous with the fertile axillary branch system in Trichopitys. Others have pointed to the presence of several ovules on an axillary shoot system in some aberrant extant specimens of G. biloba as additional support for such a reduction (Rothwell, 1987b; Z. Zhou and Zhang, 1989a).

Figure 18.24 Hamshawvia longipedunculata ovulate organs

attached to a stem bearing Sphenobaiera schenckii leaves (Triassic). (From Anderson and Anderson, 2003.)

Figure 18.25

Sergei V. Meyen. (Courtesy M. Barthel).

756

PaleoBOtany: the biology and evolution of fossil plants

A key focus in discussing the evolution of the ginkgophyte reproductive organ has been the collar that surrounds the mature seed. This structure has been interpreted as a fused megasporophyll (Crane, 1985b), a bract (Krassilov, 1970), or a homologue of the seed-bearing head of a peltasperm (Meyen, 1987). Developmental studies of Ginkgo ovules suggest that the collar develops after the integument has formed and lacks vascular tissue at maturity (Pankow and Sothmann, 1967). This analysis suggests that Ginkgo ovules are more accurately interpreted as terminal at the ends of axillary branches. This hypothesis would appear to be consistent with Archangelsky and Cúneo’s interpretation of the Early Permian genus Polyspermophyllum, in which the cup-like structure at the base of the seed is thought to be homologous with the collar of G. biloba (Archangelsky and Cúneo, 1990). Despite uncertainty regarding the affinities of the ginkgophytes, foliar remains as early as the late Paleozoic underscore the ancient age of this clade of seed plants. The extensive diversity of Mesozoic ginkgophyte leaf types suggests the occurrence of several natural groups (Z. Zhou, 1991), with the one represented by G. yimaensis giving rise to the modern lineage represented by G. biloba. This hypothesis appears to be borne out by foliage which shows a rapid increase in diversity beginning in the Middle

Triassic, and fossil evidence supports the conclusion that by the Late Triassic all of the features encompassed by G. biloba were in place (Z. Zhou and Wu, 2006). Despite the limited number of reproductive organs known in sufficient detail, several evolutionary trends have been proposed. One of these is the overall reduction in vegetative shoots, ovules, and pedicles (Naugolnykh, 2007), which morphologically give the ovulate organs a cone-like appearance. In addition, ovules appear to have become larger and leaves more entire. An interesting question concerns what led to the decline in diversity to a single species today. Tiffney (1986a) suggested that the evolution of larger seeds like those in Ginkgo improved opportunities for successful seedling establishment in closed canopies because of the increase in stored food, but at the same time decreased dispersal range. Rothwell and Holt (1997) used this scenario to suggest that the demise of large vertebrates like dinosaurs who may have consumed and dispersed these seeds may have had a longterm effect in reducing the diversity of ginkgophytes. They reasoned that studies of extant G. biloba seeds indicate that germination is greatly improved once the sarcotesta is removed, and that this odoriferous seed coat layer may represent a relic attractant mechanism for ancient seed dispersers.

19 GYMNOSPERMS WITH OBSCURE AFFINITIES GIGANTOPTERIDALES ..............................................................758

HERMANOPHYTALES ................................................................773

Vegetative Remains ...........................................................................758

GNETALES .......................................................................................775

Reproductive Organs .........................................................................762

Extant Genera....................................................................................776

VOJNOVSKYALES .........................................................................763

Extant Reproductive Structures ........................................................777

CZEKANOWSKIALES...................................................................765

Fossil Gnetophyte Pollen...................................................................777 Gnetophyte Megafossils ....................................................................778

IRANIALES...................................................................................... 768

Putative Gnetophytes ........................................................................781

PENTOXYLALES ............................................................................ 768

DIRHOPALOSTACHYACEAE ....................................................785

It is better to know some of the questions than all of the answers. James Thurber

There are a number of enigmatic gymnospermous plants that deserve mention because of their interesting morphology and, in some instances, unusual internal structure (FIG. 19.1). Some of these are known from very few specimens or are restricted both geographically and stratigraphically, for example, the Hermanophytales. In other instances, the plants are known in some detail, but their affinities continue to remain elusive, for example, the Vojnovskyales. Some, such as the gnetophytes, have been included in phylogenetic analyses of seed plants. Obviously, as additional specimens of existing taxa are discovered, which are more completely preserved, together with the discovery of new genera and species, some of the enigmatic taxa today will perhaps be easier to classify in the future. This is why we have continued to emphasize the importance of searching for new localities, as well as visiting existing ones to collect new and more complete specimens, which will provide the opportunity to investigate a broad range of issues such as the reproductive biology, development, and ecology of these enigmatic plants. No doubt some of the plants considered in this chapter will form the basis of

new groups of gymnosperms, or perhaps will no longer be included with the gymnosperms. In the discussion that follows there is no intent to specifically identify any relationships among other orders or families, but only to offer a few of the suggested affinities that have been hypothesized. We have also included the Gnetales in this section since their fossil history is poorly known and their relationship with other groups, despite the availability of extant members, remains unsettled.

757

Higher taxa in this chapter:

Gigantopteridales (Permian–Triassic) Vojnovskyales (Upper Carboniferous–Lower Permian) Czekanowskiales (Jurassic–Cretaceous) Iraniales (Upper Triassic) Pentoxylales (Jurassic) Hermanophytales (Jurassic–Cretaceous) Gnetales (Lower Cretaceous–recent) Dirhopalostachyaceae (Jurassic–Lower Cretaceous)

758

paleobotany: the biology and evolution of fossil plants

GIGANTOPTERIDALES The Gigantopteridales (Gigantonomiales of Meyen, 1987) are a loosely defined group of primarily Permian plants that are known principally from leaf remains from China, Japan, Korea, southeastern and western Asia, and North America (K. Sun, 2006). Some forms assigned to this group are also known from the Permo–Triassic of Central China and the Late Triassic (Rhaetian) of Greenland. Nearly all the plants in this order are known from impressions and compressions, in part with well-preserved cuticles (Y. Guo et al., 1993; Z.-Q. Wang, 1999; Yao and Liu, 2004), although a few permineralized stems have been reported (H. Li et al., 1996; H. Li and Taylor, 1998, 1999). VEGETATIVE REMAINS

Figure 19.1 Wattia texana (Permian). Bar  1 cm. (Courtesy

D. S. Chaney and W. A. DiMichele.)

Gigantopterids encompass a large diversity of leaf forms, both inter- and intraspecific, and many bear a striking resemblance to the leaves of extant woody angiosperms with regard to leaf size, shape, margin, and venation. Other similarities between angiosperms and gigantopterids include the presence of vessels. These features have led some to suggest that late Paleozoic gigantopterids can be regarded as vegetative analogs for flowering plants (Glasspool et al., 2004a). Gigantopterid leaves are large and characterized by a venation pattern in which the veins dichotomize in some genera, whereas in others there is a distinct pinnate pattern. There are multiple orders of veins, and the higher order veins anastomose and fuse to form a reticulate pattern (Halle, 1927; Asama, 1984). Several specimens have been described that display pinnately organized leaf portions indicative of compound leaf architecture (Kon’no and Asama, 1956); in other species, the leaves are dichotomously forked (Z.-Q Yao and Liu, 2002). Despite the size of the leaves, however, it remains unknown in most forms as to whether the leaves were simple or represent parts of even larger, compound leaves or fronds (Glasspool et al., 2004a). Leaves of the type species, Gigantopteris (Megalopteris) nicotianaefolia, from the Permian of South China (Z. Yao, 1983) are up to 50 cm long (Halle, 1927; Glasspool et al., 2004c) and possess an entire margin. Each leaf is characterized by pinnate venation with at least four orders of veins that ultimately give rise to large, reticulate polygonal meshes with the finer veinlets anastomosing to form meshes and blind endings (Glasspool et al., 2004c). Another species, Gigantopteris americana, was described from the Permian flora of North America (White, 1912). Since the specimens from North America were distinctly different

chapter 19 gymnosperms with obscure affinities

759

Figure 19.2 Portion of Gigantopteridium americanum leaf showing venation (Permian). Bar  3 cm. (From Read and Mamay, 1964.)

Figure

from the Chinese material, however, they were later excluded from Gigantopteris and placed into a new genus, Gigantopteridium (FIG. 19.2) (Koidzumi, 1934, 1936). Gigantopteridium was historically regarded as typical for the North American Gigantopteris flora, but has more recently also been recorded from China (L.-J. Liu and Yao, 2002b). Gigantopteridium marginervum from the Kuhfeng Formation (Middle Permian) of Jiangsu Province, China, has amphistomatic leaves with cyclocytic stomata surrounded by papillate subsidiary cells on the abaxial leaf surface and heavily cutinized subsidiary cells lacking papillae on the adaxial surface (Z.-Q. Yao and Liu, 2004). Other gigantopterid leaves from North America (Read and Mamay, 1964) have been placed in the genera Cathaysiopteris and Zeilleropteris (Mamay, 1986). Cathaysiopteris is a form that includes both forked and pinnately compound leaves up to 20 cm long (FIG. 19.3).

The margin is sinuous, with pinnate venation and up to three orders of venation. Close spacing of the tertiary veins and the long, narrow meshes are characteristic of this leaf type (Mamay et al., 1986). Zeilleropteris leaves (FIG. 19.4) may exceed 30 cm in length and be up to 27 cm wide. This leaf is also dichotomously divided, but may be distinguished by a venation pattern of up to four orders (Koidzumi, 1936). Both Cathaysiopteris and Zeilleropteris have been recorded for Permian floras of Asia (L.-J. Liu and Yao, 2002b). Gigantonoclea leaves occur both in North America (Mamay, 1988) and Asia (Z.-Q. Wang, 1999). This leaf type is quite similar to Gigantopteris with many forms having a toothed margin. The principal difference between the two appears to be that the pattern of veins in Gigantonoclea is less complex (FIGS. 19.5, 23.44), often with no blind endings, and veins are less sinuous than those of Gigantopteris (Glasspool et al., 2004a). Several permineralized specimens of G. guizhouensis

19.3 Cathaysiopteris yochelsonii Bar  2.5 cm. (From Read and Mamay, 1964.)

(Permian).

760

paleobotany: the biology and evolution of fossil plants

Figure 19.4 Zeilleropteris wattii showing venation (Permian).

Bar  3 mm. (From Mamay, 1986.)

Figure 19.5 Venation pattern of Gigantonoclea sp. (From

Taylor and Taylor, 1993.)

have been reported from Late Permian coal balls of southeast China (H. Li and Tian, 1990). In these leaves the base of the petiole is slightly expanded, suggesting that perhaps the leaf was simple. Stomata are paracytic and occur on the abaxial surface together with uniseriate trichomes (H. Li et al., 1994). Numerous secretory cavities are reported within the mesophyll of the leaf. An interesting feature of these leaves is the presence of prickles, or outgrowths of epidermis and cortex without vascular tissue, along the midrib on the abaxial surface, at the point where secondary veins depart. Aculeovinea yunguiensis is a permineralized stem that has been reconstructed bearing G. guizhouensis leaves, based on permineralized and compressed leaves and stems (H. Li and Taylor, 1998). Stems of A. yunguiensis are 1 cm in diameter, vesselless, and contain secondary xylem that is nearly storied; alternating with every two to three files of tracheids are multiseriate rays. The cortex exhibits a sparganum organization. Stems of A. yunguiensis bear the same type of prickles as the leaves of G. guizhouensis from the same site. Li and Taylor (1998) also reported on the occurrence of attached compressed leaves and stems from a nearby locality. The epidermal outgrowths, along with the presence of narrow diameter stems and large leaves, have been used to suggest a lianescent habit for Gigantonoclea (H. Li et al., 1994; H. Li and Taylor, 1998). Z.Q. Wang (1999), however, suggested that Gigantonoclea was an aquatic plant based on frond dimorphism that can be compared with the variation seen in submerged versus emergent leaves in extant aquatic plants. Vasovinea tianii is another permineralized axis described from Upper Permian coal balls of China (H. Li and Taylor, 1999). The axes are 1 cm in diameter, eustelic, and consist of five to nine wedge-shaped segments of mesarch primary xylem, each with a limited amount of secondary xylem. Secondary xylem tracheids have scalariform to bordered pits. One of the features that makes this plant so interesting is the presence of vessels with foraminate perforation plates in the secondary xylem (H. Li et al., 1996). Compound hooks, trichomes, and prickles occur along the stems of V. tianii, suggesting structures of a climbing plant. Vasovinea is reconstructed as bearing Gigantopteris-type leaves (H. Li and Taylor, 1999). Delnortea abbottiae is a partially permineralized gigantopterid leaf type from the Lower Permian of North America (Mamay et al., 1988). The oblong–elliptical leaves are simple, petiolate, and have margins that range from entire to crenulate (FIG. 19.6). Sclerenchymatous tissue occurs along the enrolled leaf margin, suggesting that this tissue served some mechanical function. Most specimens of D. abbottiae are 3–5 cm long, but some may have been up to 30 cm; all

chapter 19 gymnosperms with obscure affinities

761

Figure 19.6 Delnortea abbottiae leaf (Permian). Bar  1.5 cm. (From Mamay et al., 1988.)

leaves may have had clasping bases. Venation is pinnate, with four orders of veins (FIG. 19.7). The vascular system of the leaf midrib consists of 8–10 collateral bundles with scalariform–bordered pitted tracheids within a ground tissue containing abundant sclerotic cells. Delnortea is unique among gigantopterid foliage in that the secondary veins terminate in the sinuses of the crenulate margin, and the secondary veins extend all the way to the lamina margin where they fuse with the marginal indurated border. A thickened border is not known in other gigantopterids (DiMichele et al., 2000). Although Mamay et al. (1988) suggested that

Figure 19.7 Impression of Delnortea abbottiae leaf (Permian).

Bar  1.5 cm. (From Mamay et al., 1988.)

762

paleobotany: the biology and evolution of fossil plants

D. abbottiae may represent a late Paleozoic pteridosperm, they also do not rule out affinities with the gnetophytes. The geographic distribution of Delnortea is largely restricted to North America, but the genus has also been reported from Venezuela by Ricardi et al. (1999). These authors suggested that Delnortea belongs to a flora that was adapted to dry equatorial climatic conditions and was quite possibly widespread in the central western region of Pangea during the Artinskian–Kungurian (Early Permian). Another gigantopterid leaf that shares some features with Delnortea is Evolsonia. The specimens occur as finegrained impressions and are known from several different Lower Permian sites in north central Texas (Mamay, 1989). Evolsonia texana is a simple, petiolate leaf with a margin that may be either entire, sinuate, or with small teeth (FIG. 19.8). Like all gigantopterid leaves, this form is large (nearly 30 cm long). Venation can be subdivided into four orders; if marginal leaf teeth are present, the secondary veins end in the teeth. The four orders of venation are herringbone in form throughout (Mamay, 1989; DiMichele et al., 2004). REPRODUCTIVE ORGANS

There is little information about the reproductive organs of the Gigantopteridales. X. Li and Yao (1983) described leaves of Gigantonoclea from the Permian of South China that are up to 13 cm long with what are interpreted as seeds on either side of the midvein (FIG. 19.9). They proposed the name Gigantonomia for these ellipsoidal seeds, which were believed to be attached on the abaxial surface with the micropyle exposed on the upper surface of the leaf. Gigantotheca is the name used for synangiate structures borne on the lower surface of Gigantonoclea foliage from the same Permian rocks from South China. The presumed pollen sacs are arranged in linear rows, each 4 cm long, that extend over a secondary vein. Preservation was too poor to extract any pollen from the pollen sacs. In Linophyllum xuanweiense, a putative male reproductive structure associated with Gigantopteris dictyophylloides, Gigantonoclea guizhouensis, and G. hallei foliage, two rows of rounded bodies, up to 0.5 mm in diameter and interpreted as sporangia, are arranged on both sides of the midvein near the leaf margin (Zhao et al., 1980; Z. Yao and Liu, 2004). Another putative pollen organ is Jiaochengia (Z.-Q. Wang, 1999) which is found closely associated with Gigantonoclea foliage in Upper Permian rocks from Shanxi Province, North China. The specimen consists of three pairs of what are termed elongate microsporophylls bearing synangia, each constructed of two to eight elongate sporangia. Nothing is known precisely about where the synangia were borne or what type of pollen was produced.

Figure 19.8 Evolsonia texana (Permian). (Courtesy D. S. Chaney and W. A. DiMichele.)

Bar  5 cm.

The absence of detailed information about the reproductive organs of the gigantopterids makes it difficult to relate them to other groups of plants. Early reports suggested they were ferns, but more recently they have been included with the pteridosperms, and the possibility of foliar-borne seeds and pollen organs could strengthen this hypothesis (X. Li and Yao, 1983). Nevertheless, the precise nature of these

chapter 19 gymnosperms with obscure affinities

763

Figure 19.9 Suggested reconstruction of Gigantonomia fukienensis leaf with marginal seeds. (From Taylor and Taylor, 1993.)

structures remains questionable. The presence of sparganum-type cortex in Aculeovinea also suggests pteridosperm anatomy (Chapter 14). Meyen (1984) suggested that the gigantopterids were closely related to the Callistophytales, Caytoniales, Ginkgoales, and several other Mesozoic seed plant groups. Venation and vessels in some members have been used to suggest affinities with the angiosperms, although these features could also be explained as functional, based on the large leaf size and small stem diameters. Obviously the nature of the reproductive organs will be critical in defining the systematic position of these unusual plants, although in some instances, even when the reproductive organs are anatomically preserved, as in the case of the Pentoxylales (discussed below), assignment remains questionable.

VOJNOVSKYALES This order of gymnosperms was instituted for an unusual plant that was initially described from the Lower Permian of the Pechora Basin in the former USSR (Neuburg, 1955). Since that time, specimens have been described from other continents, including North America, Africa, and Argentina.

The original specimen of Vojnovskya paradoxa consists of an axis about 13 cm long and 2 cm in diameter; on the surface of this compression are large scars that mark the former position of the fan-shaped leaves or cataphylls of the Nephropsis type. Venation in the leaves is parallel. The reproductive organs consist of fertile branches or cones that have been termed “polysperms;” they are borne scattered on the vegetative branches (FIG. 19.10). In V. paradoxa the cones consist of an axis to which are attached reflexed seed stalks with widened apices. The distal portion of the cone was densely covered by linear scales with interspersed seeds. Mature seeds were apparently shed from the cones, and thus not normally found in organic connection. Seeds associated with V. paradoxa are of the Samaropsis type. At one time the cones were interpreted as being bisporangiate, but this is apparently not the case (Meyen, 1987; Rothwell et al., 1996). Additional information about the structure of ovulate cones assigned to the Vojnovskyales has been reported by Rothwell et al. (1996) from permineralized axis segments bearing ovulate cones discovered in marine shales of Late Pennsylvanian age. The axes are eustelic with ovulate cones borne in an axillary position. Cones of Sergeia neuburgii are simple with helically arranged scale leaves below

764

paleobotany: the biology and evolution of fossil plants

Figure 19.11 Diagrammatic reconstruction of Sergeia neuburgii showing axis and seed cones. (From Rothwell et al., 1996.) Figure 19.10 Suggested reconstruction of Vojnovskya para-

doxa. (Redrawn from Meyen, 1982.)

(FIG. 19.11), and sporophylls above, each bearing a single ovule. Ovules are flattened, with vascular tissue extending into the integumentary wings. The pollen chamber is simple with a small nucellar beak. Nothing is known about the pollen cones or pollen. Paravojnovskya (Gaussia; Doweld, 2004) from the Early Permian is another genus that has been suggested as closely related to Vojnovskya. The reproductive organs of Paravojnovskya are interpreted as a receptacle 2 cm in diameter (FIG. 19.12). Extending from the surface are elongate structures, each with a swollen base and expanded tip (FIG. 19.12). Krassilov and Burago (1981) suggested that this structure is not a strobilus, but rather a uniovulate carpel (swollen base) with an elongate style. What was initially suggested to be a pollen sac at the tip of the unit, is considered by these authors to be a hollow pistil. Still another interpretation views the seeds of Paravojnovskya as being borne at the tips of elongate stalks like those in certain cordaites (Meyen, 1984). Scirostrobus (Pholidophyllum) pterocerum is a vojnovskyalean fructification from the Lower Permian of the Middle

Cis-Urals (Naugolnykh, 1998, 2001b; Doweld and Naugolnykh, 2002). It is a stalked, bilaterally symmetrical, umbrella-like structure formed by sterile scales and short seed stalks (FIG. 19.13), which are fused at the base. An interesting structure that is morphologically similar to Neuburg’s original material of Vojnovskya was described from the Lower Permian of the southwestern United States (Mamay, 1976b). The conelike unit is 5 cm long, with the apical portion containing an aggregation of about 30 slender appendages possibly representing microsporophylls. Near the base are several flattened seeds, each with a bifid apex. In an earlier paper, Early Permian flabelliform leaves from the same locality were described as Sandrewia texana (Mamay, 1975). They closely resemble the leaves called Nephropsis from the Permian of Russia (Zalessky, 1912). Kerp (1988) re-interpreted Sandrewia, suggesting it is a peltasperm, and that the leaves are in fact ovuliferous organs (megasporophylls) of Autunia conferta (Chapter 15). Fan-shaped leaves with parallel venation similar to that in Nephropsis have also been reported from the Pennsylvanian of Argentina (Archangelsky and Leguizamón, 1971). Associated with the leaves is a portion of a strobilus that contains what are interpreted as microsporophylls.

chapter 19 gymnosperms with obscure affinities

765

The affinities of the Vojnovskyales have generated considerable speculation. Meyen (1984) placed all the genera within two families of Angaran Cordaitales, whereas Krassilov and Burago (1981) suggested the group has distant affinities with the angiosperms. The Vojnovskyales have also been associated with the Noeggerathiales (Chapter 12) (Neuburg, 1955), a group whose affinities are also problematic. It is interesting to note that Cardiocarpus samaratus, the ovule associated with the Early Permian cordaite Shanxioxylon sinense from northern China (Chapter 20; Wang et al., 2003b), closely resembles the ovules found inside the vojnovskyalean cone Sergeia neuburgii from North America (Rothwell et al., 1996). This appears to support the Vojnovskyales as perhaps most closely aligned with late Paleozoic coniferophytic plants (Naugolnykh, 2001b). For now, Mamay’s (1976b) assessment of the Vojnovskyales as a “bizarre, short-lived group of late Paleozoic gymnosperms” is perhaps most accurate.

CZEKANOWSKIALES

Figure 19.12 Suggested reconstruction of Paravojnovskya (Gaussia). (From Krassilov and Burago, 1981.)

Figure 19.13 Scirostrobus (Pholidophyllum) pterocerum. (From Naugolnykh, 1998.)

The Czekanowskiales include plants with persistent leaves borne on caducous, that is, deciduous, spur shoots, subtended by scalelike leaves (Pant, 1957; Krassilov, 1968, 1972a,b). The group extends from the Late Triassic into the Cretaceous (Ash, 1994). The name Czekanowskia is used for elongated, highly dissected leaves (FIG. 19.14), with a single vein that enters the base and dichotomizes several times before reaching the leaf margin. This feature has been useful in separating Czekanowskia leaves from those assigned to Baiera and Sphenobaiera, in which two vascular strands enter the base. In some species of Czekanowskia, laminar dissection is so pronounced that the leaves appear to be aggregated into fascicles (FIG. 19.15). Cuticle preparations indicate that there are both hypostomatic and amphistomatic leaf types in the genus, and that the stomata are arranged in short longitudinal files or bands (Samylina and Kiritchkova, 1973, 1993; Kostina, 1999; Watson et al., 2001). Samylina and Kiritchkova (1993) distinguished three subgenera within the morphogenus Czekanowskia based on stomatal distribution and arrangement: subgenus Czekanowskia for amphistomatic leaves with stomata arranged in files, subgenus Harrisiella for amphistomatic leaves with stomata arranged in bands, at least on the abaxial side, and subgenus Vachrameevia for hypostomatic leaves with stomata arranged in files or bands. Moreover, these authors document that within the subgenus Czekanowskia two speciation events took place in the Middle Jurassic and Early Cretaceous, which they relate to the paleoclimate

766

paleobotany: the biology and evolution of fossil plants

Figure 19.14 Leaf fascicles of Czekanowskia sp. (Jurassic).

Bar  2 cm. (Courtesy BSPG.)

during these time periods. One paleoecological interpretation suggests that the leaves were deposited under a humid, temperate to tropical climate based on the presence of leaves in Eurasian coal-bearing sequences (Ash, 1994). There are other morphogenera of leaves assigned to the Czekanowskiales. Leaves of Solenites are arranged in distinct bundles of 11–16 that are borne on short shoots. Rows of stomata occur on both surfaces. Leaves of Sphenarion are narrowly wedge shaped, with no distinction between the petiole and lamina (Harris and Miller, 1974). They are distinguished from Czekanowskia and Solenites by being wider than 1 mm. Leaves of Phoenicopsis are produced in fascicles on short caducous shoots that also bear small persistent leaves (Heer, 1876 (FIG. 19.16); Watson et al., 2001). They are narrowly wedge shaped but lack a petiole. Florin (1936b) subdivided Phoenicopsis into three genera, sometimes also referred to as subgenera, based on stomatal distribution: Phoenicopsis s.str. for hypostomatic leaves, Culgoweria for amphistomatic leaves with stomata scattered over the entire leaf surface, and Windwardia for amphistomatic leaves with stomata arranged in longitudinal rows. Leafy shoots believed to represent young, long shoots of Phoenicopsis s.l. from the Middle Jurassic Yima Formation in central China have been described under the name

Figure 19.15 Czekanowskia microphylla. (From Taylor and

Taylor, 1993.)

Tianshia patens (Z. Zhou and Zhang, 1998). Another morphogenus, Arctobaiera, is characterized by leaves attached to either long or short shoots. Stomata are arranged in rows and the veins superficially appear parallel, but have numerous anastomoses (Z. Zhou and Zhang, 1996). Based on material from the Yima Formation, Z. Zhou and Guignard (1998) studied the cuticle ultrastructure of Phoenicopsis euthyphylla and Arctobaiera renbaoi and found that, except for minor differences in thickness of the layers, the cuticle ultrastructure of these taxa is quite similar in both the upper and lower cuticles. Their upper cuticles are thicker and consist of three layers, whereas the abaxial cuticles typically have two layers. Species of the ovulate cone Leptostrobus (for a review of the genus, see X.-Q. Liu et al., 2006b) have been associated with foliage of Czekanowskia, Sphenarion, Phoenicopsis s.l., and Solenites (Clifford and Camilleri, 1998). This

chapter 19 gymnosperms with obscure affinities

767

Figure 19.16 Oswald Heer. Figure 19.17 Suggested reconstruction of Leptostrobus cancer. (From Taylor and Taylor, 1993.)

association has been based on their common occurrence in the same fossil beds, and the similarity in epidermal anatomy. Although incomplete, some specimens of L. cancer are up to 15 cm in length (Harris, 1951b; X.-Q. Liu et al., 2006b). Helically attached to the axis at regular intervals of 5 mm are flattened, globose, bivalved structures termed capsules (FIG. 19.17). Each is 5 mm long and consists of two valves; the outer surface in some species is crenulate (Krassilov, 1968). The outer margin of the valve is partially enrolled and forms a broad flange that has been interpreted as enclosing the seeds (FIG. 19.18). Each capsule contains up to five seeds, with each seed attached to the rim of the capsule valve so that the micropyles are directed toward the open portion of the capsule. Nothing is known about the tissues of the seeds, but megaspore membranes macerated from the seeds suggest that the ovules were of a basic gymnospermous type. Although numerous pollen types have been identified in association with the seed micropyles, none occurs with sufficient frequency to be regarded as the pollen of Leptostrobus. Some grains appear to be inaperturate, whereas others are distinctly bisaccate (Harris and Miller, 1974).

Ixostrobus is the name of a pollen cone thought to be the strobilus of Czekanowskia. Another point of view is offered by Harris and Miller (1974), who believe that Ixostrobus was the pollen cone of a plant that produced the ribbon-shaped leaves of the Desmiophyllum type. Specimens of I. whitbiensis (Middle Jurassic of Yorkshire) are at least 10 cm long and contain helically arranged microsporophylls with pollen sacs. Each microsporophyll is flared where it is attached to the axis and pointed at the distal end. In I. longicalcaratus from northern Iran, four ellipsoidal pollen sacs are attached slightly back from the tip (Schweitzer and Kirchner, 1995). Specimens of I. siemiradzkii from the Lower Jurassic (Liassic) of Hungary have microsporophyll tips that are recurved (WcisłovLuraniec and Barbacka, 2000). Pollen in I. whitbiensis is circular–oval, with a conspicuous sulcus on the distal surface. Grains range from 35 to 40 μm in diameter and, based on the description, may have possessed a tectate sporoderm. The affinities of the Czekanowskiales continue to remain problematic. They have been included with the ginkgophytes based on association of highly dissected foliage (Harris, 1935; Schweitzer and Kirchner, 1995), whereas others have suggested the group may be distantly related to the seed

768

paleobotany: the biology and evolution of fossil plants

Figure 19.19 Suggested reconstruction of Irania hermaphro-

ditica. (From Schweitzer, 1977; courtesy U. Schweitzer and R. Gossmann.)

Figure 19.18 Capsule of Leptostrobus stigmatoides show-

ing marginal flange (arrow) (Jurassic). Bar  8 mm. (Courtesy V. A. Krassilov.)

ferns, or as noted by Meyen (1987), they are simply a group of poorly defined gymnosperms.

IRANIALES This order was established in 1977 by Schweitzer for hermaphroditic reproductive structures of Late Triassic age from northern Iran. The reproductive structures were termed “flowers” and consisted of an axis bearing helically arranged clusters of aggregated microsporangia (FIG. 19.19). Nothing is known about the pollen. Arising from an extension of the main reproductive axis are narrow, dichotomizing, secondary axes, each terminating in a flattened, heart-shaped structure

termed a capsule. Extending around the edge of the capsule is a ridge of delicate tissue. Another interpretation is that the ridge of tissue may represent the integumentary wing of a flattened conifer seed. If this interpretation is accurate, then the pointed tip represents the micropylar end of the seed. The presence of Desmiophyllum type foliage in the same rocks was used to suggest affinities with the Czekanowskiales. Moreover, some have suggested that the capsules of the czekanowskian female cone Leptostrobus (discussed earlier) are structurally similar to the capsules seen in Irania (X.-Q. Liu et al., 2006b).

PENTOXYLALES The Pentoxylales constitute a small group of Jurassic and Cretaceous gymnosperms, many of which are known from structurally preserved specimens. The first report of these interesting plants was made by the distinguished Indian

chapter 19 gymnosperms with obscure affinities

769

Figure 19.21 Cross section of Pentoxylon stem showing five

sympodia of secondary xylem. Note some periodicity in wood segments (arrow) (Jurassic). Bar  5 mm. Figure 19.20 Birbal Sahni.

paleobotanist Birbal Sahni (1948) (FIG. 19.20) from plant remains collected from the Rajmahal Hills of northeastern India. Since the initial report, numerous additional specimens have been collected from the Rajmahal Hills (Sharma, 1969b, 1973b,c, 1979, 1980), and the group has also been reported from Jurassic and Lower Cretaceous localities in New Zealand (Harris, 1962), Australia (White, 1981b), and most recently Antarctica (Césari et al., 1998; Howe and Cantrill, 2001), although the affinities of these latter specimens have been challenged (Sharma, 2003). The group was initially based on petrified specimens, but compression– impression fossils have also been attributed to the order. Although the exact habit of these plants remains unknown, some consider them to have been small trees that possessed both long and short shoots or spur shoots (Bose et al., 1985; Pole, 1988). The abundance of leaves, short shoots, and ovulate cones from Alexander Island, Antarctica, has been used to suggest that the plants were more shrublike and lived in monodominant communities in river floodplains (Howe and Cantrill, 2001). Pentoxylon includes stems that in transverse section consist of five or six wedge-shaped vascular segments embedded in a thin-walled ground tissue (FIG. 19.21). In some descriptions, these segments are referred to as steles of a polystelic stem, but as has been noted for Medullosa (Chapter 14), the stele of Pentoxylon is a eustele. Each xylem wedge consists

of a mesarch primary xylem strand or sympodium that is completely surrounded by dense secondary xylem, with growth rings indicating some level of periodicity. The most common number of rings in the wood is five or six, although up to 16 have been reported. In Pentoxylon sahnii, the most common species, the vascular cambium apparently functioned in a way so that in older stems there was strong endocentric development (FIG. 19.22) of the secondary xylem, that is, more wood was produced toward the pith region of the stem. The wood of Pentoxylon is pycnoxylic, with unior biseriate circular-bordered pits on the radial walls of the tracheids, and uniseriate medullary rays up to seven cells high. Wood parenchyma, ray tracheids, and resin canals are not a component of this simple wood. Pith and cortex consist of thin-walled parenchyma with scattered nests of sclerotic cells. In a few stems, the cortex contains radial rows of cells suggestive of a periderm. The secondary phloem is composed of sieve elements and phloem parenchyma; fibers are lacking (Sharma and Bohra, 1977). Spur shoots of Pentoxylon, each less than a centimeter in diameter, have only primary vascular bundles in the pith and cortex (Sharma, 1973b, c, 1979). Each bundle produces a strand from the inner margin that repeatedly branches to form the leaf traces. The outer surface of the spur shoots is often ornamented by helically arranged, rhomboidal leaf cushions, each with seven to nine vascular bundle scars.

770

paleobotany: the biology and evolution of fossil plants

Figure 19.22 Cross section of Pentoxylon stem showing extraxylary tissues (Jurassic). Bar  5 mm.

Figure 19.24 Taeniopteris type leaf on surface of silicified peat containing Pentoxylon stems (Jurassic). Bar  1 cm.

Figure 19.23 Suggested reconstruction of Pentoxylon sah-

nii with Taeniopteris leaves attached to short shoot. (From Sahni, 1948.)

Scars of two different sizes have been described as occurring in zones on the stem, with the smaller ones thought to have been scale leaves. At the end of spur shoots in specimens from Antarctica is a terminal bud (Howe and Cantrill, 2001). Leaves of Pentoxylon are strap shaped and may be up to 20 cm long (FIG. 19.23). They exhibit a prominent midrib

of several bundles, and in general are identical with those of the morphogenus Taeniopteris (FIG. 19.24). Because some pentoxylalean leaves are structurally preserved, the generic name Nipaniophyllum was instituted for these. Leaves of N. raoi are 1 cm wide and characterized by a short petiole. The midrib is broad and consists of several parallel veins that terminate at the rounded apex; lateral veins are borne at right angles, occasionally branching and fusing near the leaf margin. The petiole has five to nine traces, each surrounded by a sclerenchymatous sheath. Stomata are sunken and confined to the abaxial surface. In some species stomata appear to be arranged in rows. Stomatal development in some leaves is syndetocheilic, but in others, each pair of guard cells is surrounded by a ring of subsidiary cells, that is, the anomocytic pattern (Bose et al., 1985). Insect

chapter 19 gymnosperms with obscure affinities

771

19.26 Sahnia laxiphora microsporangiate organ (Cretaceous). Bar  1 cm. (From Osborn et al., 1991.)

Figure

Figure 19.25 Sahnia nipaniensis showing microsporophylls produced on a cylindrical receptacle and positioned terminally on a dwarf shoot (Jurassic). (From Suthar and Sharma, 1988.)

galls occurring on a Nipaniophyllum leaf from the Upper Cretaceous of the Rajmahal Basin, India, suggest that pentoxylalean leaves were used by developing insect larvae (Banerji, 2004). The pollen organ of Pentoxylon is called Sahnia and believed to have been produced terminally on parenchymatous short shoots (FIG. 19.25) (Vishnu-Mittre, 1953). In S. nipaniensis there are numerous elongate microsporophylls (microsporangiophores of Bose et al., 1985) that are arranged in close spirals around (according to Vishnu-Mittre,

1953) or on a dome-shaped or cylindrical receptacle (according to Suthar and Sharma, 1988). Each sporophyll bears several to many pear-shaped, sub-sessile or stalked sporangia (FIGS. 19.25, 19.26) (Suthar and Sharma, 1988). The proximal portion of the pollen organ is surrounded by several modified leaves (bracts), up to 6 mm long, that are curved distally (FIG. 19.25). In S. laxiphora they terminate in four to seven unilocular pollen sacs (FIG. 19.26) (Drinnan and Chambers, 1985). Pollen is monosulcate and 25 μm long. The sporoderm in this species is two parted, consisting of an inner granular zone and outer homogeneous region (Osborn et al., 1991). Spore mother cells containing structures interpreted as chromosomes in distinct meiotic figures have been reported from a permineralized Sahnia microsporangium preserved in a Lower Cretaceous chert matrix from Nipania, Rajmahal Hills (Bonde et al., 2004). The seed-bearing organs in this group have been termed infructescences, female flowers, seed-bearing fruits, or seed-bearing cones. Specimens of Carnoconites have a central peduncle or axis that branches, with each branch terminating in an ovulate head (FIG. 19.27) (Bose et al., 1983). Each head or megasporophyll of C. compactus (Bose et al., 1985) is 2.5 cm long and contains 20 tightly packed orthotropous, helically arranged seeds (FIG. 19.28). Seeds are platyspermic and apparently sessile (FIG. 19.29). The nucellus and integument are attached only at the base, and vascular tissue is reported as extending through the

772

paleobotany: the biology and evolution of fossil plants

Figure 19.27 Suggested reconstruction of the seed-bearing organ Carnoconites compactus. (From Crane, 1988.)

sclerotesta. The outer integument consists of a thick, fleshy sarcotesta that surrounds the fibrous sclerotestal layer of the integument. Putative polyembryony has been reported in a seed of C. compactus (Sharma, 1989). Additional specimens from the Albian of Antarctica include ovulate organs attached to short shoots by elongate pedicles (Howe and Cantrill, 2001). Each cone of C. cranwellii is 7 mm long and contains 35 ovules (Harris, 1962). An extraordinary specimen from the Lower Cretaceous of the South Shetlands Island (Antarctica) includes seven cones that appear to be helically attached to an axis (FIG. 19.30). This specimen also bears a Taeniopteris leaf (Césari et al., 1998). The ovules of C. llambiasii are about 1 mm long. The Pentoxylales represent a unique group of seed plants that have been associated with various groups of gymnosperms, depending on which characters were analyzed. For example, the platyspermic ovules, thick sclerotesta, and ring of pollen-bearing structures suggest affinities with the Bennettitales (Rao, 1974). Monosulcate pollen is found in bennettitaleans as well as in Ginkgo, cycads, and certain pteridosperms. Anomocytic stomata are found in all groups except the bennettitaleans, Gnetum, and Welwitschia, but this character is considered unreliable in phylogenetic analyses. Several phylogenetic analyses place Pentoxylon together with

Figure 19.28 Longitudinal section of Carnoconites compac-

tus seed-bearing axis with numerous attached seeds (Jurassic). Bar  5 mm.

the bennettitaleans as the sister group of a clade that also includes the flowering plants and Gnetales (Crane, 1985a,b; Doyle, 2006; Doyle and Donoghue, 1986). In an analysis of seed plants by Hilton and Bateman (2006), Pentoxylon and the glossopterids are closely related, but in Nixon et al.

chapter 19 gymnosperms with obscure affinities

773

Figure 19.29 Cross section of Carnoconites compactus with

attached seeds (Jurassic). Bar  1 cm.

(1994) Pentoxylon is part of a large polytomy that includes several seed plant groups. Many of the characters used in these and other cladistic analyses require rather large steps in homologizing some of the features that are ultimately coded. For example, Crane (1985b) interpreted the seeds as possessing a double integument of which part of the outer integument is described as a cupule homologue. Although the Pentoxylales may ultimately turn out to be related to gnetophytes, bennettitaleans, and the flowering plants, the evidence to link them more strongly with one particular group is not convincing, or is biased by missing values (Nixon and Davis, 1991). In this context, we view Pentoxylon as a Mesozoic gymnosperm whose affinities continue to remain obscure.

Figure 19.30 Axis with multiple Carnoconites llambiasii reproductive structures (arrows) (Cretaceous). Bar  5 mm. (Courtesy S. N. Césari.)

HERMANOPHYTALES The Hermanophytales are perhaps the most enigmatic group among the gymnosperms discussed in this chapter. The order is known only from permineralized stems, and consists of a single genus, Hermanophyton, with four species described to date (Tidwell and Ash, 1990). Specimens come from the southwestern United States, where they have been

discovered in Upper Jurassic rocks of the Morrison Formation. The genus also exists in the Santonian (Upper Cretaceous) of Europe, based on a single specimen from the Aachen region in northwestern Germany. Hermanophyton stems are typically unbranched and only a single branched specimen is known (Tidwell, 2002)

774

paleobotany: the biology and evolution of fossil plants

Figure 19.31 William D. Tidwell.

(FIG. 19.31). They range from 3 to 40 cm in diameter and are characterized by a complex internal structure (Tidwell and Ash, 1990; Tidwell, 2002). Up to 15 prominent, wedgeshaped segments of primary and secondary xylem are radially arranged around a narrow parenchymatous pith. The individual segments of xylem are separated from one another by wide primary rays (FIG. 19.32); anastomoses of segments may be present. The primary xylem of each segment is separated into strands, whereas the secondary xylem is compact and shows irregular growth increments. Secondary xylem consists of radially aligned tracheids with uniseriate pitting; apertures are generally circular, but sometimes elliptical. Numerous uni- to biseriate xylem rays, some up to 32 cells high, occur in each of the wedges. The primary rays consist of parenchymatous cells and irregular strands of thick-walled fibers. Bundles of meandering tracheids from the xylem wedges flank either side of the ray. In addition, numerous vascular traces originate from the primary xylem on each side of a primary pith ray (FIG. 19.33), and extend through the primary ray. The stele is surrounded by a twolayered parenchymatous cortex, which, in turn, is surrounded by a ramentum composed of densely spaced, club-shaped appendages and long trichomes. Passing through the cortex are numerous leaf traces and cortical bundles. Located between cortex and ramentum is what has been termed a

19.32 Petrified stem of Hermanophyton taylorii (Jurassic). Bar  8 cm. (Courtesy BSPG.)

Figure

modified meristematic region. Integrated in the ramentum are numerous small, rounded leaf bases that are vascularized by strands originating from the fusion of bundles from different leaf traces and cortical bundles (FIG. 19.33). Nothing is known about the systematic affinities of Hermanophyton due primarily to the lack of information on the other organs, especially reproductive structures of this plant. Arnold (1962) compared Hermanophyton with Rhexoxylon, a permineralized stem genus from Gondwana today referred to the seed fern order Corystospermales (Chapter 15). Rhexoxylon stems, like Hermanophyton, are characterized by wedges of secondary xylem separated by parenchymatous tissue that includes numerous bundles. However, the secondary xylem in Rhexoxylon develops on both sides of the primary xylem (i.e., centrifugally and centripetally), whereas it is only formed on the outside in Hermanophyton. Arnold (1962) speculated that Hermanophyton might eventually prove to be a cycadophyte, but other plant groups variously compared

chapter 19 gymnosperms with obscure affinities

775

which strongly suggests a non-self-supporting growth habit for this plant.

GNETALES

Figure 19.33 Diagrammatic reconstruction of the anatomy of Hermanophyton glismannii. (From Tidwell and Ash, 1990.)

with Hermanophyton include the medullosan seed ferns, Pentoxylales, and Bennettitales (Tidwell and Ash, 1990). Almost nothing is known about the growth habit of the Hermanophyton plant. Arnold (1962) and Tidwell and Ash (1990) suggested that it was a short-to medium-sized, narrow-stemmed tree crowned with small leaves, based on the size of the leaf traces, which were successively shed as the plant grew, leaving numerous leaf bases on the lower parts of the stem. Tidwell (2002) suggested that the Hermanophyton plant may have grown to 18 m high. Another interpretation is that Hermanophyton was a large liana based on the multi-stranded cable-like organization of the cauline vascular tissue. The longest Hermanophyton stem ever found was 10 m long, and had a relatively consistent diameter of 12 cm,

Like the other groups discussed in this chapter, the plants included in the Gnetales have been variously treated taxonomically, although the group includes living representatives. Historically, they were placed in an intermediate position between the gymnosperms and angiosperms in a subdivision termed the Chlamydospermae (Pulle, 1938). In other systems, each of the three extant genera, Gnetum, Ephedra, and Welwitschia, occupied its own family, whereas still other treatments each has been elevated to the level of an order. With the increasing use of molecular phylogenetics, the gnetophytes became the focus of what has become known as the anthophyte hypothesis (Doyle and Donoghue, 1992). The anthophyte clade includes not only the gnetophytes and angiosperms but also the Bennettitales and Pentoxylales. Although the view that the gnetophytes are a sister group to the flowering plants has been supported in some seed plant analyses (Rothwell and Serbet, 1994), this relationship is now controversial since there is an apparent conflict between molecular and morphological data (Donoghue and Doyle, 2000; Burleigh and Mathews, 2004). Another interpretation is the so-called gnepine hypothesis, in which there is a sister relationship inferred between the gnetophytes and pines (Hajibabaei et al., 2006), but other data sets reject the conifer–gnetophyte relationship (Rydin et al., 2002). In the seed plant analysis of Nixon et al. (1994), the gnetophytes are regarded as paraphyletic with the angiosperms nested within them, whereas Doyle (1996) suggested that the ancestor of the gnetophytes and angiosperms may be as old as the Permian. We have included Gnetum, Welwitschia, and Ephedra in a single order for convenience only, despite the fact that the group is considered to be monophyletic (Price, 1996). The principal features that have been used to unite these unique gymnosperms are the compound reproductive organs, the topographic position of the reproductive parts in which the ovule-producing parts are more apical than the pollenproducing parts of the bisexual organ complexes (Martens, 1971; Endress, 1996), and the presence of foraminate vessels and their development in the xylem. Although a type of double fertilization has been reported in Ephedra and Gnetum, and used to support a relationship with the angiosperms, in the gnetophytes studied in detail to date, this type of double fertilization does not produce an endosperm like that in angiosperms (Friedman and Carmichael, 1996).

776

paleobotany: the biology and evolution of fossil plants

EXTANT GENERA

EPHEDRA Ephedra is a profusely branched shrub (FIG. 19.34) confined to cool, desert regions today; a few of the 35–45 extant species are small trees. The primary vascular system is a eustele (Marsden and Steeves, 1955). Leaves vary from decussate to whorled, and range from scalelike to needles up to 3 cm long. The pollen of Ephedra is ellipsoidal and ornamented with a series of striations that extend from one end of the grain to the other (FIG. 19.35). In at least one species the pollen wall appears tectate (Van Campo and Lugardon, 1973). Although there is an increasing number of fossils that morphologically appear identical to modern Ephedra, and thus indicate that the crown group may be Mesozoic (Rydin et al., 2004), rbcL sequence data suggest that extant Ephedra diverged only 8–32 Ma ago (Huang and Price, 2003; Huang et al., 2005). GNETUM Almost all the species of Gnetum (FIG. 19.36) consist of lianas that inhabit tropical rainforests in Asia, South America, and Africa (Won and Renner, 2006). One interesting feature of this genus is the pinnate reticulate venation pattern in the leaves, which superficially resembles that found in the leaves of many dicotyledonous angiosperms (FIG. 19.36). Pollen grains are relatively small (16–20 μm) and typically spherical; some are ornamented with delicate spines.

Figure 19.35 Ephedra pollen grain (Extant). Bar  10 μm.

WELWITSCHIA The third extant genus of this group is the monotypic Welwitschia, which is confined to arid areas of southwest Africa. Morphologically, the plant is unusual in that it

Figure 19.36 Gnetum gnemon. Bar  5 cm. (Courtesy Günter

Fuchs.)

Figure 19.34 Ephedra torreyana branch with cones.

consists of a short, cone-shaped axis, two strap-shaped leaves that arise from its rim (FIG. 19.37), and a deep-seated tap root. Seeds have conspicuous lateral wings formed by the fusion of a pair of bracts. Pollen of W. mirabilis is like that of Ephedra, but appears to possess a permanent leptoma

chapter 19 gymnosperms with obscure affinities

777

Figure 19.38 Selected examples of ornamentation patterns in

Figure 19.37 Welwitschia mirabilis. (Courtesy J. Silander.)

(Bharadwaj, 1963), or shed the exine at germination (Rydin and Friis, 2005).

Cretaceous ephedroid pollen grains from the Santana Formation in northeastern Brazil. (Redrawn from de Lima, 1980.)

secrete a pollination droplet that may serve to attract insects, whereas in other forms there is evidence of wind pollination.

EXTANT REPRODUCTIVE STRUCTURES

FOSSIL GNETOPHYTE POLLEN

There has been considerable discussion and debate about the nature, organization, and homology of the reproductive parts of the gnetophytes (e.g., Frohlich and Meyerowitz, 1997). In extant gnetophytes, they occur as cone-like structures sometimes with several cones within a branching system. It is important to point out that although the terms flower and inflorescence are sometimes used to describe these units, we will avoid these terms here, as they imply homologies with angiosperms which, as noted earlier, are tenuous. Most species are dioecious and cones consist of an axis bearing bracts at each node. In the axil of the bracts are the microsporangiate (male) reproductive units, each of which consists of a so-called perianth of bracts that surround a pair of microsporophylls, also termed antherophores, that often branch at the tips (Hufford, 1996; Mundry and Stützel, 2004b). At the tips are synangia, with the microsporangia dehiscing longitudinally. The megasporangiate reproductive structures in the group are more apical than the microsporangiate structures (Endress, 1996). They are compound and consist of one or two pairs of decussate bracteoles that surround the ovule. The seed integument is elongated into a micropylar tube which extends beyond the bracteoles. Ovules may be ensheathed by one or two additional bracts, and it is these structures that have been discussed relative to their homologies as an integument, cupule, perianth, carpel, megasporophyll, sterile leafy organ, or a bract (Takaso and Bouman, 1986). Some ovules

The fossil record of the gnetophytes is limited based on megafossils; however, there have been numerous polyplicate (or ephedroid) pollen grains reported for this group. The earliest putative gnetophyte pollen comes from the Permian (Azéma and Boltenhagen, 1974; de Lima, 1980; Pocock and Vasanthy, 1988). In the Cretaceous, pollen with gnetophyte characteristics (FIG. 19.38) is abundant and has been recorded from numerous sites (Stover, 1964; Srivastava, 1968; Takahashi et al., 1995; Rydin et al., 2004). Pollen resembling gnetophyte grains is also produced by a number of other plants, for example, the polyplicate pollen types in Arales, Laurales, and Zingiberales (Hesse et al., 2000; Hesse and Zetter, 2007), and thus it is likely that not all late Mesozoic and Cenozoic fossil ephedroid-like pollen was actually produced by gnetophytes. Many of the Permian putative gnetophyte pollen grains are morphologically similar to Ephedra grains, and this pollen type is also a common component in northern Gondwana and southern Laurasian Cretaceous floras (Muller, 1984). Despite the gnetalean affinities, this pollen type is unfortunately called Equisetosporites. Finestructural details of Equisetosporites indicate that the wall is two parted with a lamellate nexine and homogeneous tectum (Osborn et al., 1993), features which are found in both Ephedra and Welwitschia. At a few localities, the pollen co-occurs with gnetophyte foliage remains (de Lima, 1978; Pons et al., 1992).

778

paleobotany: the biology and evolution of fossil plants

Another polyplicate pollen grain often included with the gnetophytes is Ephedripites. Ephedripites has gnetalean affinities based on ultrastructural features (Trevisan, 1980), whereas Equisetosporites is suggested as more closely related to the angiosperms (Pocock and Vasanthy, 1988; Zavada, 1990a). A different point of view was offered by Osborn (1991) who suggested that some polyplicate grains from northern Gondwana referred to as Equisetosporites (de Lima, 1980) are, in fact, more closely related to gnetophytes. It is interesting that there is a highly diverse ephedroid and angiosperm palynoflora during the mid-Cretaceous, but the ephedroid pollen declines toward the Cenomanian and represents only a small percentage of geologically younger pollen floras (Crane, 1996). GNETOPHYTE MEGAFOSSILS

The absence of a well-documented megafossil record of gnetophytes was initially interpreted as indicating that these plants lived predominantly in arid environments as they do today, and thus would not be easily fossilized. In addition, it was suggested that perhaps some leaves thought to have been produced by Gnetum were in fact those of various dicots since the venation of the two is superficially similar. More recently, however, there have been additional gnetophyte megafossils discovered (Cornet, 1996), and information about the fossil record of this group continues to expand. Pollen of the Equisetosporites type has been described in a Masculostrobus cone from the Late Triassic Chinle Formation of Arizona (Ash, 1972). Associated with these cones, but not organically attached, are leaves of Dechellyia gormanii (FIG. 19.39). In this genus the leaves are arranged in opposite and decussate pairs (FIG. 19.39) (Ash, 1972). Associated with D. gormanii foliage are seeds characterized by a single large, flattened wing extending from the wall. Although there is some superficial resemblance between D. gormanii and the leaves of the conifers Taxodium and Metasequoia, the association with Equisetosporites pollen in the rocks has largely shaped opinion that this fossil is a gnetophyte. Piroconites kuespertii, a microsporophyll from the Lower Jurassic of Germany, has also been suggested to be a gnetophyte based on the occurrence of polyplicate pollen (Van Konijnenburg-Van Cittert, 1992). This fossil bears synangia containing Ephedripites pollen on the adaxial surface. It is hypothesized that scales called Bernettia represent the megasporophylls of the same plant. These were subtended by bracts called Chlamydolepis, and long linear leaves with parallel venation (Desmiophyllum gothanii) are also thought to be parts of the same plant that produced Piroconites (Crane, 1996).

Figure 19.39 Suggested reconstruction of Dechellyia gormanii showing shoot system and fertile branches. (From Ash, 1972.)

The first unequivocal megafossil evidence of a gnetalean is the description of Drewria by Crane and Upchurch (1987). Stems of Drewria potomacensis from the Lower Cretaceous Potomac Group of Virginia, USA, are slender, with opposite and decussate simple leaves with entire margins; each is up to 2 cm long (FIGS. 19.40, 19.41). Primary venation consists of two to three pairs of parallel veins that extend to near the apex of the lamina. Second-order veins extend between the primaries at nearly right angles and are arched apically. The reproductive organs of D. potomacensis are dichasia organized in three spikes (FIG. 19.42). The lateral spikes are larger and produce flattened seeds (FIG. 19.43), each 2 mm long, and subtended by a pair of bracts. It is hypothesized that the central spike may have produced pollen sacs, although no in situ pollen was obtained; however, polyplicate pollen grains occur in the same rocks and these are ellipsoidal and range from 18–32 μm long. They appear most similar to modern pollen of Welwitschia mirabilis (FIG. 19.44) and to other gnetophyte grains described from the Lower Cretaceous (de Lima, 1978; Trevisan, 1980). Sedimentological data, together with the floral composition

chapter 19 gymnosperms with obscure affinities

779

Figure 19.41 Suggested reconstruction of the vegetative parts of Drewria potomacensis. (From Crane and Upchurch, 1987.)

19.40 Drewria potomacensis showing vegetative leaves (arrow) and axillary reproductive shoots (Cretaceous). Bar  3 mm. (Courtesy G. R. Upchurch.)

Figure

of the rocks containing D. potomacensis, support the hypothesis that the gnetophytes and early angiosperms had similar ecological tolerances (Doyle et al., 1982), living in mesic environments, rather than being initially adapted to hot and dry climates (Brenner, 1976). The more recent discovery of Cratonia cotyledon from the Lower Cretaceous of Brazil is an important fossil because not only is it a gnetophyte but it also represents a seedling, a rare fossil plant structure (Rydin et al., 2003). The compressed specimen consists of two cotyledons, a lateral feeder, and a root (FIG. 19.45). The cotyledons have numerous parallel veins with secondary veins fused to form arches that are similar to those described in Drewria potomacensis (Crane and Upchurch, 1987). Another possible Early Cretaceous gnetophyte is Liaoxia from northeastern China (Rydin et al., 2006a). The type species, Liaoxia chenii (FIG. 19.46), was originally described

Figure 19.42 Suggested reconstruction of Drewria potomacensis dichasium. (From Crane and Upchurch, 1987.)

as a monocotyledonous angiosperm (Cyperaceae) by Cao et al. (1998), later re-interpreted as Ephedrites chenii (S. Guo and Wu, 2000), and eventually transferred to Liaoxia (Rydin et al., 2006a). The fossils consist of stems up to 3 mm in diameter with swollen nodes and internodes that range from 8 to 40 mm long. Leaves are 5–20 mm long and are produced in an opposite-decussate phyllotaxis. Cones up to 1 cm long consist of two to six pairs of bracts, and seeds occur in the axils of the cone bracts. Liaoxia chenii is probably a close relative of the extant Ephedra (Liu et al., 2008) but diagnostic reproductive details that might confirm this relationship remain unknown.

780

paleobotany: the biology and evolution of fossil plants

s

Figure 19.45 Cratonia cotyledon, a gnetophyte seedling (Cretaceous). Bar  2 cm. (Courtesy B.A.R. Mohr.)

19.43 Detail of Drewria potomacensis showing seeds (S) and associated bracts from a dichasium (Cretaceous). Bar 1.5 mm. (Courtesy G. R. Upchurch.)

Figure

Figure 19.44 Welwitschia-type pollen grain associated with

Drewria (Cretaceous). Bar  15 μm. (Courtesy G. R. Upchurch.)

Figure 19.46 Suggested reconstruction of Liaoxia chenii. (From Rydin et al., 2006a.)

Striate stems of Ephedra archaeorhytidosperma have also been reported from the same rocks that contain L. chenii (Y. Yang et al., 2005). Seeds assigned to E. portugallica from the Early Cretaceous of Portugal and E. drewriensis from the

Early Cretaceous of North America are up to 1.4 mm long and possess the same complement of structures, including outer sclerenchymatous envelope and elongate micropylar tube, as modern Ephedra seeds (Rydin et al., 2006b).

chapter 19 gymnosperms with obscure affinities

781

Figure 19.48 Cotyledons of Priscowelwitschia austroamericana (Cretaceous). Bar  5 mm. (Courtesy D. L. Dilcher and T. A. Lott.)

Priscowelwitschia (Welwitschiella) austroamericana for young shoots with paired cotyledons (FIG. 19.48). PUTATIVE GNETOPHYTES

Figure 19.47 Several cones of Welwitschiostrobus murilii

(Cretaceous). Bar  5 mm. (Courtesy D. L. Dilcher and T. A. Lott.)

Inside the micropylar tube of some seeds are the exines of polyplicate pollen grains interpreted as being discarded during germination. Seeds lacking certain epidermal structures and pollen inside the micropylar tube are included in Ephedrispermum (Rydin et al., 2006b). Leaves and cones that morphologically resemble Welwitschia have been reported from the Lower Cretaceous of Brazil (Dilcher et al., 2005). Welwitschiostrobus murilii is used for dichasial terminal or axillary cones formed of decussate bracts (FIG. 19.47), whereas Welwitschiophyllum brasiliense is used for isolated, generally elongated leaves,

A single specimen interpreted as a gnetophyte has been reported from the Upper Permian of China (Z.-Q. Wang, Z. 2004). Palaeognetaleana auspicia is described as bisexual (FIG. 19.49), but because the specimen is compressed, critical details that would corroborate affinities with the gnetophytes are not obvious. It should be noted, however, that masses of Ephedripites pollen 40–50 μm long were extracted from the specimen. Dinophyton spinosus (FIG. 19.50) is an interesting fossil first described from vegetative remains of Late Triassic age (Chinle Formation) (Ash, 1970). Additional specimens composed of four bract-like appendages in a cruciate arrangement, so-called pinwheels, have been described with uniovulate cupules and sporangiophores with pollen sacs containing grains of the Alisporites or Pteruchipollenites type (Krassilov and Ash, 1988). Although it is suggested that the bract-like appendages subtending the cupules provide evidence of affinities with the gnetophytes, the compressed specimens could just easily represent a member of a seed fern group. Another Late Triassic fossil that has been referred to the gnetophytes is the enigmatic reproductive structure Nataligma dutoitii from the Molteno Formation of South Africa (Anderson and Anderson, 2003). This compound

782

paleobotany: the biology and evolution of fossil plants

MP SC

IC OS MM BR

TP

MS

CH

A

B

Figure 19.49 Palaeognetaleana auspicia. A. Reconstruction of the cone. B. Axillary unit, showing structure of ovular integument;

MM: megaspore membrane; TP: tapetal tissue; IC: cuticle of inner envelope; SC: sclerotic envelope; OS: outer scales or fibers; MS: a presumed pollen sac; BR: bract; CH: chalazal end. (From Z.-Q. Wang, 2004.)

Figure 19.50 Portion of Dinophyton spinosus leaf. (From Ash,

1970.)

cone consists of a series of whorls of small, compact ovulate strobili borne on elongated slender axes (FIGS. 19.51, 19.52). Each strobilus consists of up to six whorls of megasporophylls (FIG. 19.53). The distal portions of the

megasporophylls form a continuous enclosure around the strobilus. Anderson and Anderson (2003) suggested that foliage assigned to the morphogenus Gontriglossa may belong to N. dutoitii. Gontriglossa includes axes bearing oppositely positioned fascicles of narrow, elliptic leaves with frequently anastomosing open-mesh arching venation. Based on the distinctness of Nataligma, Anderson and Anderson (2003) placed the fossil in a separate order within the gnetophytes. Another megafossil that has been linked to the gnetophytes is Eoantha zherikhinii, a reproductive organ from the Lower Cretaceous of Transbaikalia (Krassilov, 1986; Krassilov and Bugdaeva, 2000). The specimens represent pedicellate, rosette-shaped, cupulate, flower-like structures 4 mm wide, which are composed of a whorl of four scales (FIG. 19.54). Each of the four scales has a median groove and is described as containing a single ovule. The entire unit is subtended by elongate bracts; the pedicel extends through the unit and terminates in a cluster of filiform appendages. The suggested affinities of this unusual reproductive organ are based on the presence of polyplicate pollen of the Ephedripites–Gnetaceaepollenites type in the pollen chambers of the ovules. Although this type of association has been used many times to link dispersed fossil plant organs, foreign pollen can become associated with seeds of various plant groups, and thus when used alone may be a less reliable character in determining affinities. Baisianthus ramosus and Vitimantha crypta are two names of pollen organs

chapter 19 gymnosperms with obscure affinities

Figure 19.51 Nataligma dutoitii compound reproductive structure showing lateral axes bearing ovulate cones (arrows) (Triassic). Bar  3 mm. (Courtesy J. M. Anderson.)

with possible affinities in the gnetophytes discovered at the same locality along the Vitim River as Eoantha (Krassilov and Bugdaeva, 2000). Baisianthus ramosus is a synangiate pollen organ organized in whorls, whereas V. crypta consists of a pedicellate, flower-like structure formed by four bractlike scales enclosing follicles that contain polyplicate pollen grains. Affinities with the gnetophytes have also been suggested for Gurvanella dictyoptera from the Lower Cretaceous of Mongolia and China. This fossil was initially interpreted as an angiospermous fruit (Krassilov, 1982), but later reinterpreted as a samaroid cupule composed of four decussate scales, each subtending a crescent-shaped seed (Akhmetiev and Krassilov, 2002). G. Sun et al. (1998) pointed out that the winged seeds, that is, the seed–bract

783

Strobilus

Lateral axis Primary axis

Figure 19.52 Nataligma dutoitii showing primary axis and

whorls of laterals bearing cones. (From Anderson and Anderson, 2003.)

784

paleobotany: the biology and evolution of fossil plants

Peduncle

(A)

(B)

Figure 19.55 Afrasita lejalnicoliae. A. Showing suggested reconstruction of fertile branch. B. Cone with one of the seeds partially exposed. (From Krassilov et al., 2004.)

Lateral axis

Figure 19.53 Nataligma dutoitii, compact ovoid strobilus with

megasporophylls arranged in whorls and showing axis and peduncle (Triassic). (From Anderson and Anderson, 2003.)

Figure 19.54 Eoantha zherikhinii. (From Crane, 1988.)

complexes, of G. dictyoptera are similar to those of Welwitschia marabilis, and suggested that Welwitschia-like plants appear to have extended to China and Russia during the Mesozoic. Exceptionally well-preserved specimens

of Gurvanella sp. have also been reported from the Early Cretaceous Jehol Group (Jehol Biota) in northeastern China (Z. Zhou et al., 2003). These fossils show both the vegetative and reproductive features of the plant, and document that Gurvanella is indeed reproductively similar to extant Welwitschia, but has vegetative features more similar to Ephedra. Akhmetiev and Krassilov (2002) suggested that, although morphologically different, the cupules of Eoantha and Gurvanella are organized on the same structural plan. Another reproductive structure that deserves mention here is Afrasita lejalnicoliae (FIG. 19.55) from the Lower Cretaceous of Egypt and Israel. This enigmatic fossil, originally described as Leguminocarpon abbubalense (Lejal-Nicol, 1981), consists of radially arranged, crescentshaped scales subtending large, axially borne seeds (FIG. 19.55) constructed of three envelopes (Krassilov et al., 2004). The systematic affinities of A. lejalnicoliae remain unresolved, but the surface of the scales resembles the surface ornamentation of coarse ridges present on the scales in Eoantha. The Gnetales represent an interesting albeit perplexing group of gymnospermous plants, not only in their relationship to each other but also in their evolutionary history. Recent phylogenetic analyses suggest the group is monophyletic and shares more features with bennettitaleans, Pentoxylon, and angiosperms than any other group of seed plants. The fossil evidence, although limited, does indicate that gnetophytes were relatively cosmopolitan during the mid-Cretaceous, becoming more restricted geographically up to the present. The suggestion that the gnetophytes occupied ecological niches similar to those of the early angiosperms raises the possibility that one reason for their decline may

chapter 19 gymnosperms with obscure affinities

785

Figure 19.57 Single capsule of Dirhopalostachys rostrata

with conspicuous keel (Cretaceous). Bar  1.25 mm. (Courtesy V.A. Krassilov.)

torus-bearing circular-bordered pits and perforation plates that suggest a closer similarity to the tracheids of conifers than those of angiosperms (Carlquist, 1989). Doyle’s (1996) suggestion that perhaps the gnetophytes and angiosperms are related to the glossopterids is an interesting hypothesis that will certainly be tested in morphological analyses that are developed as we continue to learn more about the fossil relatives of Ephedra, Gnetum, and Welwitschia.

Figure 19.56 Dirhopalostachys rostrata axis with capsules

DIRHOPALOSTACHYACEAE

(Cretaceous). Bar  2.5 mm. (Courtesy V. A. Krassilov.)

be linked to their anemophilous pollination syndrome or an inefficient entomophilous syndrome (Bino et al., 1984). The report by Carlquist (1988), which relates the distribution of vessels in Ephedra to climate, suggests that this character may be less important phylogenetically than was previously thought. Moreover, the tracheids of Ephedra possess large

Another interesting group of seed plants, which is based on limited material from the Upper Jurassic–Lower Cretaceous of Siberia, is the Dirhopalostachyaceae (Krassilov, 1975). In this group the ovule-bearing organs, which have been termed gynoclads, are interpreted as organized in raceme-like clusters (FIG. 19.56). Each pair of capsules (?megasporophylls) of Dirhopalostachys rostrata contains a single ovule (FIG. 19.57). The affinities are suggested as being with the cycad Beania (Krassilov, 1975).

This page intentionally left blank

20 CORDAITALES VEGETATIVE FEATURES ............................................................ 788

Seeds .................................................................................................798

Stems .................................................................................................788

ANGARAN CORDAITES ............................................................ 801

Foliage ...............................................................................................791 PHYLOGENETIC POSITION AND ORIGIN OF THE

Roots .................................................................................................794

CORDAITES ................................................................................... 803 REPRODUCTIVE FEATURES .................................................... 795 Reproductive Organs .........................................................................795

Time passes and the world changes. The remains of the past are shrouded in uncertainty. Basho

The Cordaitales are an extinct group of gymnosperms that can be traced from the Mississippian to the Permian. The most complete records of these plants come from central Europe (Grand’Eury, 1877), North America (Rothwell and Warner, 1984; Rothwell, 1993), and China (S.-J. Wang, 1998; S.-J. Wang et al., 2003b), where they represented a conspicuous portion of the late Paleozoic flora. Paleoecological studies indicate that many forms inhabited lowland peat mires where they either formed more or less monotypic stands or were part of mixed vegetation types, growing interspersed among calamites, tree ferns, and lycophytes (Raymond and Phillips, 1983; Trivett and Rothwell, 1985; Raymond, 1988; DiMichele and Phillips, 1994). Other forms lived in poorly to well-drained mineral soils, forming fire-prone communities on drier sites (Scott et al., 1997; Falcon-Lang, 2003a–c; Falcon-Lang et al., 2004). A few giant forms grew in upland settings (Falcon-Lang and Scott, 2000; Falcon-Lang and Bashforth, 2004, 2005). Some cordaitaleans are suggested as inhabiting marine-influenced coastal mires and thus represented the earliest known mangrove communities, the so-called cordaitalean mangrove hypothesis; however, this hypothesis has more recently been challenged. For example, Falcon-Lang (2005a) described cordaitalean trees from the Lower Pennsylvanian of Joggins, Nova Scotia, which grew

in a coastal setting that was flood prone but not permanently submerged. Because the forest layers do not show evidence of marine influence, and several key physiological adaptations to growth in brackish waters are absent in the trees, he concluded that the term “mangrove” for these fossil forests is probably inappropriate (see also Plaziat et al., 2001). Several cordaitaleans have been reconstructed as whole plants, including Pennsylvanioxylon birame (Costanza, 1985), Cordaixylon dumusum (FIG. 20.1) (P. nauertianum) (Rothwell and Warner, 1984; Costanza, 1985; Rothwell, 1993), Mesoxylon priapi (Trivett and Rothwell, 1985), C. iowensis (Trivett, 1992), and Shanxioxylon sinense (S.-J. Wang et al., 2003b). The names used to refer to the entire plants are typically the names originally given to the stems (as in Lepidodendron, Psaronius, and Medullosa), because this part of the plant usually provides the most distinctive characters. Arborescent Cordaitales consist of monopodial trunks with a distal crown that produces large, strap-shaped leaves. They represent the largest trees in the Pennsylvanian tropical zone, with some forms growing to more than 45 m tall (Falcon-Lang and Bashforth, 2005), but the group also includes long-lived, scrambling shrubs, such as C. dumusum, and small trees, such as S. sinense. Some of the small trees, such as P. birame, have aerial stilt roots that form at the base

787

788

PaleOBotany: the biology and evolution of fossil plants

Figure 20.1 Diagrammatic reconstruction of Cordaixylon dumusum. (From Rothwell and Warner, 1984.)

of the stem and probably looked much like the basal portions of some modern mangrove trees. Pollen and seed-producing organs of the cordaites consist of compound cones and are generally borne on distal leaf-bearing branches. Pollen is saccate and the seeds are typically bilaterally symmetrical; some have conspicuous lateral extensions of the integument in the form of wings.

VEGETATIVE FEATURES STEMS

Although the generic name Cordaites was originally instituted for foliage (Unger, 1850) (FIG. 20.2), it has also been used for structurally preserved stems and for the entire plant (Wilson and Johnston, 1940; Cohen and Delevoryas, 1959). Cordaixylon (Grand’Eury, 1877; Rothwell and Warner, 1984; Trivett, 1992) and Pennsylvanioxylon (Vogellehner, 1965; Costanza, 1985; Tian and Wang, 1987, 1988; S.-J. Wang, 1997, 1998) are two of the generic names specifically given to anatomically preserved stems that earlier were referred to as Cordaites.

Figure 20.2

Franz Unger.

Distinguishing Pennsylvanioxylon from Cordaixylon, however, appears to be difficult. Because of the apparent uncertainties with regard to the differences between the two taxa, S.-J. Wang et al. (2003b) refer to the stems in a more informal way

CHAPTER 20

CORDAITALES

789

M. Barthel.)

Figure 20.4 Oblique longitudinal section of cordaite stem showing pith septations and secondary xylem (Pennsylvanian). Bar  1 cm.

as Cordaixylon–Pennsylvanioxylon. Impression–compression fossils of cordaitalean branches and twigs with helically arranged leaf scars are usually assigned to the morphogenus Cordaicladus (FIG. 20.3) (Grand’Eury, 1877). Cordaixylon stems are eustelic and contain a rather large pith region surrounded by a small number of separate primary xylem strands (sympodia). The primary vascular architecture of these stems is like that of modern conifers. In other cordaite stem genera, the sympodia are interconnected like those of some pteridosperms (Trivett and Rothwell, 1988b). The pith is a conspicuous portion of the stem,

with some specimens possessing a pith 10 cm in diameter. The most conspicuous features of the pith are the evenly spaced horizontal septations or diaphragms (FIGS. 20.4, 20.5) that alternate with lacunae formed during the elongation of the stem. In some cordaites, septations are lacking, and the pith is wholly parenchymatous, for example, in Mesoxylon priapi. In Piracicaboxylon meloi, a cordaitalean stem from the Lower Permian of Uruguay, the pith is differentiated into central and peripheral zones that are separated by a parenchymatous ring (Crisafulli, 1998). Some of the cordaitalean

Figure 20.3 Cordaicladus sp. (Permian). Bar  1 cm. (Courtesy

790

PaleOBotany: the biology and evolution of fossil plants

Figure 20.5 Oblique cross section of cordaite stem showing

histology of pith septations (Pennsylvanian). Bar  5 mm.

pith cells contain amorphous contents, and have been suggested as being secretory. Because of the transverse septations in Cordaixylon, pith casts are relatively easy to distinguish. In the early 1800s the generic name Sternbergia was established by Edmund T. Artis for cordaitalean pith casts in honor of his friend and colleague, Kaspar Maria von Sternberg. At about the same time, Sternberg published a description of the pith casts in honor of Artis. Because the name Sternbergia was already in use as the generic name of certain extant flowering plants, that is, “autumn daffodils” in the Amaryllidaceae, cordaite pith casts are today referred to by the name Artisia (FIG. 20.6). The primary xylem of a cordaite stem is relatively inconspicuous and composed of just a few tracheids in the discrete sympodia at the ends of the narrow wedges of secondary xylem. Pitting of primary xylem tracheids ranges from annular to circular bordered with little distinction between primary and secondary xylem tracheids in pitting patterns. When structurally preserved stems were first described by Renault (1896a,b), the primary xylem was characterized as endarch. Later, cordaitalean axes were discovered in coal balls from Great Britain (Scott and Maslen, 1910) and the genus Mesoxylon was established for stems with mesarch stem and leaf traces. Today Mesoxylon is also distinguished by a discontinuous or non-sympodial vascular architecture, whereas Cordaixylon is delimited by a sympodial system with endarch traces (Trivett and Rothwell, 1988b). Secondary xylem in cordaites is extensive and simple (FIG. 20.7), consisting only of tracheids and rays that are variable in height (Costanza, 1985; Cichan, 1986b). Xylem parenchyma and resin canals are absent. Ray cells are thin walled and often preserved with amorphous cell contents. Secondary xylem tracheids exhibit closely spaced vertical rows of uni- or biseriate pits similar to those seen in some

Figure 20.6

Artisia sp. (Pennsylvanian). Bar  2 cm.

conifers (Rothwell and Warner, 1984); they have narrow, diagonal pit apertures, but lack tori (Schmid, 1967). Wood of the Cordaixylon type typically has short uniseriate rays. Other cordaite woods, such as Mesoxylon, possess biseriate rays and multiseriate pitting on larger tracheids. In some specimens, abrupt changes in tracheid diameter produce patterns in the secondary xylem that resemble growth rings

CHAPTER 20

Figure 20.7 Cross section of Mesoxylon sp. stem (Pennsylvanian). Bar  5 mm.

CORDAITALES

791

Secondary phloem in Cordaixylon may be up to 5 mm thick in some specimens (E. Taylor, 1988). It consists of alternating, tangential bands of sieve cells and axial parenchyma; fibers are not present. Sieve areas are oval to rectangular and located on the radial cell walls. Secondary phloem in cordaites supports what appears to be an overall conservative evolutionary history seen in gymnosperms, that is, an early development of the arrangement of cell types and patterns that has remained relatively unchanged to the present (E. Taylor, 1988). The cortex of cordaitaleans is characterized by alternating bands of sclerenchyma that form the Sparganum- or Dictyoxylon-type cortical patterns which are also present in certain groups of seed ferns (Chapter 14). In some species a well-defined periderm is produced. The genus Shanxioxylon includes cordaitalean stems from the Upper Permian of China (Tian and Wang, 1987; S.-J. Wang et al., 2003b) that display a parenchymatous and septate pith, sympodial primary vascular architecture, endarch primary xylem maturation of the cauline bundles, and mesarch maturation in the leaf traces. Leaf traces diverge from the pith margin as single bundles and leaves approach a helical phyllotaxis. Secondary xylem is of the Dadoxylon type. In contrast to Cordaixylon– Pennsylvanioxylon and Mesoxylon, branch traces are rarely observed in Shanxioxylon. Moreover, Shanxioxylon lacks the Sparganum- or Dictyoxylon-type cortical pattern of peripheral sclerenchymatous bands seen in other cordaitean stems. FOLIAGE

Figure 20.8 Cordaixylon tianii showing reaction wood (arrow)

(Pennsylvanian). Bar  2.5 mm. (Courtesy J. Hilton.)

(FIG. 20.8). Large, isolated segments of secondary xylem have been described under the morphogenus Dadoxylon (Vogellehner, 1964). Although such wood is believed to have been produced by cordaitaleans, it is not sufficiently preserved to show the positions of protoxylem strands or leaf traces.

Impression and compression fossils of leaves represent the most frequently encountered parts of the plants, and bedding planes containing monotypic assemblages of Cordaites leaves are known from various sites in Europe and North America (e.g., Martín-Closas and Galtier, 2005). Cordaitalean leaves are typically strap shaped (FIG. 20.9), with some reaching 1 m in length, although most are in the 10-to-20-cm size range; a few forms are known to have produced small, needle-like leaves. The tip of the leaf is usually bluntly rounded and the area of attachment is broad. Cordaitalean leaves lack a midrib and instead possess parallel veins that occasionally dichotomize. Subgeneric names, including Eucordaites, Dorycordaites, and Poacordaites, are based on external morphology and surface ribbing patterns and have been used to distinguish Euramerican impression and compression foliage specimens (Grand’Eury, 1877; Remy and Remy, 1977). Cordaitalean leaves usually are helically arranged, and many species probably resembled certain modern members of Podocarpus and Araucaria. Some

792

PaleOBotany: the biology and evolution of fossil plants

Figure 20.9 Cordaite branch tip with several leaves. (Pennsylvanian). Bar  2 cm.

species are believed to have been heterophyllous (Rothwell, 1982a). Cordaite leaf cuticles have been studied intensively with regard to micromorphology and epidermal anatomy (Barthel, 1962b, 1964; Glukhova and Menshikova, 1980; G. Sun, 1991; Z. Liu et al., 1998; Šimu˚nek, 2000, 2007; S. Li et al., 2003) but also with regard to chemical constituents (Stankiewicz et al., 1998; Zodrow et al., 2000a, 2003; Mösle et al., 2002; Auras et al., 2006a, b). The epidermal cell pattern is composed of isodiametric or elongate and usually more or less rectangular cells that are arranged in long rows oriented parallel to the leaf margins. Stomata typically occur in longitudinally oriented intercostal bands. The stomatal

apparatus is haplocheilic and relatively simple, consisting of two bean-shaped guard cells, two lateral subsidiary cells, and a terminal cell at either of the polar ends. The guard cells are slightly sunken below the subsidiary cells. Normal epidermal cells and subsidiary cells may bear cuticular papillae (Barthel, 1962b, 1964; Wartmann, 1969). Most Cordaites leaves have historically been regarded as hypostomatic. In the Middle Pennsylvanian permineralized form C. felicis and the Late Pennsylvanian compressed C. principalis, however, stomatiferous bands are present on both surfaces (Good and Taylor, 1970; Zodrow et al., 2000b), whereas in the Late Pennsylvanian compressed C. schatzlarensis, stomata occur in bands on the abaxial surface but are scattered or in short rows on the adaxial side (Šim˚unek and Libertín, 2006). This suggests that many species described from compression– impression specimens were actually amphistomatic. Šim˚unek (2000) reported that the number of cordaite cuticle morphotypes from the Carboniferous of Bohemia is considerably higher than the number of species described based on macromorphological features. This suggests that the diversity of cordaitaleans in this region was greater than generally appreciated. Another interpretation is that cuticle micromorphology displays intraspecific differences (Zodrow et al., 2000b). These differences might be related to the position on the plant, age of the leaf, heteroblastic development, or local differences in habitat conditions. More recently, Šim˚unek (2007) provided a new classification of the genus Cordaites from the Carboniferous and Permian of the Bohemian Massif based on macromorphology and cuticle micromorphology that demonstrates the value of cuticular and epidermal features in cordaitalean foliage taxonomy and systematics. Permineralized cordaitean leaves have been described from Europe (Stopes, 1903), North America (Reed and Sandoe, 1951; Harms and Leisman, 1961), and China (S.-J. Wang and Tian, 1995; S.-J. Wang et al., 2003b) and have been placed in different species based on differences in the distribution of cell types within the leaf; for example, Harms and Leisman (1961) distinguished 10 species. Anatomically, the leaves are rather similar, consisting of a well-differentiated palisade and spongy mesophyll. In some species, the distinctions between these two tissues are difficult to determine; in others, either tissue may be absent. The most conspicuous anatomic feature in some taxa is the presence of fibrous bands between the veins (FIG. 20.10) which may resemble structural I-beams in cross section. The presence and extent of these fibrous partitions have been useful in delimiting species. In C. felicis, the number of veins ranges from 10 to 25 per centimeter, with masses of hypodermal fibers above and

CHAPTER 20

20.10 Cross section of Cordaites (Pennsylvanian). Bar  1 mm. (Courtesy BSPG.)

Figure

sp.

CORDAITALES

793

leaf

below the veins (Good and Taylor, 1970) (see FIG. 20.27). The vascular elements are surrounded by a sheath of thinwalled cells four to seven cell layers thick and maturation of the bundles is mesarch. Noeggerathiopsis has been used for elongated, strapshaped, Cordaites-like leaves with sub-parallel venation from the Carboniferous and Permian of the Southern Hemisphere (Pant, 1982). Although some authors use the name Cordaites for foliage of this type (Meyen, 1982b), others believe the two genera are distinct. Among the characters used to distinguish Noeggerathiopsis are its smaller spathulate shape, absence of interstitial veins or fibers between the veins, and several other anatomical features (FIG. 20.11). Permineralized Noeggerathiopsis leaves from Permian rocks of the Prince Charles Mountains, East Antarctica (McLoughlin and Drinnan, 1996) have an upper epidermis, a mesophyll differentiated into an ill-defined, adaxial palisade layer and abaxial spongy layer (FIG. 20.12), and a lower epidermis. Leaves have prominent abaxial stomatiferous furrows or grooves, one between each vein, which are missing in the Euramerican Cordaites, but are present in the Angaran taxon Rufloria (see below). The veins in the Antarctic Noeggerathiopsis consist of vascular bundles composed of an adaxial wedge of thick-walled xylem tracheids and an abaxial cavity representing the former position of the phloem. Xylem is mesarch and composed of small, centrally located protoxylem cells and slightly larger, irregularly arranged centripetal metaxylem tracheids. Trichomes are present on the abaxial surface and especially abundant in the stomatiferous grooves (FIGS. 20.12, 20.13). Noeggerathiopsis leaves are hypo- or amphistomatic with the stomata haplocheilic, and the guard cells longitudinally oriented and surrounded by six subsidiary cells (McLoughlin and Drinnan, 1996). Additional cuticular and epidermal features of Noeggerathiopsis leaves

Figure 20.11 Noeggerathiopsis sp. (Permian). Bar  1 cm. (Courtesy S. McLoughlin and A. Drinnan, 1996.)

have been reported by Pant and Verma (1964), Lele and Maithy (1964), Srivastava and Tewari (2002), and Chauhan and Tiwari (2006). Some studies fail to document the typical abaxial grooves (Lele and Maithy, 1964; Pant and Verma, 1964), but this may be a preservational or preparational artifact (see discussion in McLoughlin and Drinnan, 1996). The status of Noeggerathiopsis continues to be controversial (surveyed in McLoughlin and Drinnan, 1996, and Singh et al., 2007). Maheshwari and Singh (1999) suggest that cordaitean

794

PaleOBotany: the biology and evolution of fossil plants

Figure 20.12 Section of Noeggerathiopsis sp. leaf show-

ing vascular bundles and stomata grooves (arrows) (Permian). Bar  200 μm. (Courtesy S. McLoughlin.) Figure 20.14 Cross section of Amyelon sp. with secondary xylem (center) and well-developed cortical tissues (Pennsylvanian). Bar  2 mm.

This is a small, simple ellipsoidal leaf with a sessile base, obtuse apex, and coriaceous and folded lamina (Singh et al., 2007). It differs from other Cordaites-like leaves in that the stomata are not arranged in longitudinally oriented intercostal bands but rather are irregularly oriented and distributed (Singh, 2000). The systematic affinities of Kawziophyllum remain unresolved. Figure 20.13 Section of Noeggerathiopsis sp. leaf show-

ing distinct trichomes in stomatal grooves (arrows) (Permian). Bar  200 μm. (Courtesy S. McLoughlin.)

leaves of this type from the Permian of India, which have traditionally been placed in Noeggerathiopsis, should be called Pantophyllum (Rigby, 1984b) if they are from Gondwana and have cuticle. Euryphyllum is a morphogenus used for Early Permian leaves from the Southern Hemisphere that are ovate-spathulate and often asymmetrical, with narrow tapering bases, obtuse apices, and dichotomous, sub-parallel venation (surveyed in Singh et al., 2007). Although some consider Euryphyllum to be a broader form of Noeggerathiopsis leaves, most regard the two taxa as distinct based primarily on the fact that the lateral veins arch in the former and are straight in the latter taxon (Seward and Sahni, 1920; Chandra and Singh, 1996; Singh et al., 2007). Another Gondwanan foliage taxon similar to Noeggerathiopsis is Kawziophyllum (Singh, 2000).

ROOTS

Structurally preserved cordaitalean roots are named Amyelon, which includes both protostelic and eustelic “species” (Barnard, 1962; Cridland, 1964; Z. Li, 1986a). Z. Li (1986b) suggested that some so-called species of Amyelon represent root ecotypes produced by a single, natural cordaite species in response to habitat conditions. The protostelic forms of Amyelon (FIG. 20.14) range from diarch to pentarch, and pitting ranges from spiral to multiseriate pitting on the primary xylem tracheids. The secondary xylem is composed of uniseriate rays and radially aligned tracheids with 1–5 rows of bordered pits on the radial walls. Secondary phloem contains sieve elements, phloem parenchyma, phloem fibers, and rays, with the outer zone loosely constructed and possibly aerenchymatous (E. Taylor, 1988). Surrounding the secondary phloem zone is a layer of phelloderm and a compact layer of phellem with lenticels. In the roots of the Cordaixylon dumusum plant, the periderm forms near the periphery of the primary cortex (Rothwell and Warner, 1984). Lateral roots are borne in large clusters from phellem-covered protuberances

CHAPTER 20

CORDAITALES

795

Figure 20.15 Cross section of Premnoxylon sp. (Pennsylvanian).

Bar  1 mm.

on the larger roots. They show no definite arrangement but appear to have been borne on only one side of the primary root. They possess the same complement of tissues as the main roots but have a well-developed endodermis. Endophytic fungi believed to be endomycorrhizal have been reported in Amyelon-type roots of some European cordaitalean specimens (Osborn, 1909; Halket, 1930), but it is difficult to determine whether the fungi were true symbionts, parasites, or saprotrophs that invaded the tissues after the death of the plant (Cridland, 1962). Another name used for some permineralized woody cordaitalean roots is Premnoxylon (FIG. 20.15) (Pierce and Hall, 1953). Cichan and Taylor (1982c) provided evidence for a complex system of anastomosing wood borings in Premnoxylon wood from the Middle Pennsylvanian of Kentucky. Since wood borings typically occur in above-ground plant parts, this adds support to the hypothesis advanced by Cridland (1964) that some cordaitalean roots were aerial.

REPRODUCTIVE FEATURES

C

Figure 20.16 Compressed cordaitean strobilus showing several secondary fertile shoots (C) along the primary cone axis. Note conspicuous bracts (arrows) (Pennsylvanian). Bar  5 mm.

REPRODUCTIVE ORGANS

The reproductive structures of cordaitaleans are considered to be compound, lax cones and are composed of primary shoots (primary cone axis) bearing bracts and secondary fertile shoots (secondary cone axis) in the axils of the bracts (FIG. 20.16). In some of the most extensive compression specimens, the primary cones are 30 cm long; permineralized specimens rarely exceed 10 cm in length. Cordaitalean reproductive structures are generally believed to be borne on distal branches. In the Cordaixylon dumusum plant

(FIG. 20.1), cones are attached at proximal levels of mature branches where leaves have been shed and a great deal of wood has been produced but are also borne apically among the leaves (Rothwell, 1988, 1993). In proximal regions, they extend outward from the cortex of the branch. Cones near the apex of C. dumusum are vascularized by a single bundle that arises by tangential division of a stem bundle. This implies

796

PaleOBotany: the biology and evolution of fossil plants

that the cones do not represent axillary branches, which are vascularized by two traces that diverge from the two stem bundles that flank the position of a leaf trace. In the P. birame plant, however, the cones have been interpreted as axillary branches (Trivett and Rothwell, 1988b). Many reconstructions show the position of the cones in the same organotactic spiral as the leaves. Interestingly, Rothwell (1993) found no evidence for a relationship between the positions of leaf divergence and cone divergence in C. dumusum. Cordaitalean cones are monosporangiate (FIG. 20.17), and it is uncertain whether the plants were monoecious or dioecious. Most of the reproductive structures have been assigned to Cordaianthus or Cordaitanthus (for nomenclatural issues, see Fry, 1955), regardless of whether they ultimately produced ovules or pollen. In cross section, the primary shoot axis is slightly flattened and contains a centrarch or endarch medullated stele with tracheids that have scalariform pitting on both the radial and the tangential walls. The primary axis bears bracts in distichous or tetrastichous arrangement, and a secondary fertile shoot (cone) occurs in the axil of each bract. The secondary axis is radial and produces helically arranged, scale-like leaves. Ovules or pollen sacs are attached to the most distal scales. In Cordaitanthus duquesnensis the secondary shoots are 5.5 mm long and contain 60–70 helically arranged scales (Rothwell, 1982b). Each of the distal scales bears a single, platyspermic ovule in which the nucellus and integument are attached only at the chalaza. Ovules begin development as a swollen tip on an immature fertile scale (Rothwell, 1982b). Integument ontogeny is initiated with a bilobed collar of tissue at the periphery of the swollen tip. Growth at the base of the collar results in the formation of an integument that forms a micropyle in the apical region. In some cordaites there was an elongation of the fertile scale during ovule development so that the ovule was suitably positioned for pollination. Rothwell (1982b) speculated that development in these ovules was continuous, with no seed dormancy. As a result, he suggested that many cordaites shed their ovules from the cone close to the time they reached full size and prior to germination. In Cordaitanthus concinnus (Middle Pennsylvanian) the secondary fertile shoots are 1.6 cm long but are not present at some nodes due to mechanical disruption or abortion of primordia (Rothwell, 1977b). Each secondary shoot axis contains axial bundles (sympodia) that increase in number distally similar to the eusteles of other gymnosperms. The largest number of sympodia in C. concinnus secondary axes is eight, with each giving rise to a scale trace. Surrounding the secondary axis is a variable number of helically arranged

B

B

20.17 Cordaitanthus sp. primary axis with tightly packed secondary fertile shoots in axils of bracts (B) (Pennsylvanian). Bar  1 cm.

Figure

modified leaves, that is, scales with all but the most distal ones sterile (Delevoryas, 1953). In C. schuleri from the Middle Pennsylvanian of Iowa, the number of sterile scales per secondary shoot is estimated to be 95 (Fry, 1956),

CHAPTER 20

whereas in C. concinnus and C. compactus, the number is closer to 30. In the pollen cones (FIG. 20.18), only the most distal five to ten scales contain pollen sacs. Each fertile scale contains 3–6 elongate pollen sacs, each 1 mm long. In C. concinnus the pollen sacs are arranged in pairs, with dehiscence taking place toward the center of the cluster. The pollen sac wall is a single cell layer thick and dehiscence is longitudinal. Each pollen sac is vascularized by a small trace that departs from the vascular bundle of the fertile scale. The ontogeny of the pollen sacs has been documented for male reproductive structures associated with the Cordaixylon dumusum plant (Rothwell, 1993). It begins with a simple swelling of the sporophyll apex, followed by division of the apex into five primordia, each of which subsequently develops into a lobe and eventually develops into an elongate pollen sac. Pollen sacs within a single secondary shoot probably matured in a sequential pattern based on the fact that less-developed, small pollen sacs with a solid interior co-occur with larger, collapsed pollen sacs with a hollow interior. Pollen grains extracted from Cordaitanthus pollen sacs are monosaccate, consisting of a central body or corpus surrounded by a saccus or air bladder attached to both the proximal and the distal surfaces (Millay and Taylor, 1974). In C. concinnus the grains measure 65 μm by 45 μm and exhibit a conspicuous ornamentation of endoreticulations on the inner surface of the saccus; the outer surface is relatively smooth (Delevoryas, 1953). Other species possess pollen grains that approach 100 μm in diameter. Pollen of Cordaitanthus is identical with the common Pennsylvanian sporae dispersae grain type Florinites (FIG. 20.19). Florin (1936a) described pollen grains from the pollen sacs of several silicified Pennsylvanian Cordaitanthus cones that contained what he believed were sterile jacket cells surrounding a row of lenticular cells inside the corpus. Subsequent studies have indicated that the sterile jacket cells were probably an artifact caused by folds in the corpus wall (Millay and Taylor, 1974). There can be little doubt, however, that the lenticular cells represent stages in the development of the microgametophyte in the form of axially aligned prothallial cells (Taylor and Millay, 1979). Although haptotypic markings are generally absent from the proximal surfaces of Florinites pollen, the ultrastructure and thin exine suggest that germination probably took place from the distal surface. It is not known whether pollen tubes were produced in this grain type. Cordaitalean cones characterized by bracts and secondary fertile shoots in helical arrangement are assigned to the genus Cathayanthus (S. Wang et al., 2003b). Cathayanthus sinensis, the ovulate cone of the Shanxioxylon sinense plant,

CORDAITALES

797

Figure 20.18 Longitudinal section of Cordaitanthus concin-

nus pollen cone. Arrows indicate pollen sacs (Pennsylvanian). Bar  2 mm.

Figure 20.19 Florinites sp. pollen grain (Pennsylvanian).

Bar  20 μm.

consists of a slender, subcircular, or subtriangular primary axis that bears bracts and secondary fertile shoots in a compact helical arrangement (S.-J. Wang and Tian, 1993). Secondary shoots are up to 15 mm long and consist of 105–115

798

PaleOBotany: the biology and evolution of fossil plants

sterile and fertile scales in approximately equal numbers; these are helically arranged around the axis. Fertile scales extend from the distal portion of the secondary axis and are helmet-like in transverse section. Ovules are borne terminally on extended tips of the fertile scales. The pollenbearing reproductive structure of the S. sinense plant has been named Cathayanthus rametrarus (S.-J. Wang and Tian, 1991b; S.-J. Wang et al., 2003b). Secondary fertile shoots are bilaterally symmetrical, bud-like in longitudinal section, up to 5 mm long and have up to 13 sterile scales and 3–5 fertile scales positioned distally on the secondary axis. Each of the fertile scales bears five pollen sacs containing Florinitestype pollen grains. Another pollen-bearing cone included within the Cordaitales is Gothania (FIG. 20.20). The genus was originally delimited by Hirmer (1933a) for cordaitalean fructifications believed to be associated with the European stem Mesoxylon multirame. Specimens have since been discovered in Middle Pennsylvanian strata in eastern Kentucky (Daghlian and Taylor, 1979). Morphologically, they are similar to specimens of Cordaitanthus but differ in the number of pollen sacs per scale (FIG. 20.21), their arrangement, and the type of pollen produced. Pollen grains isolated from Gothania lesliana correspond to the sporae dispersae genus Felixipollenites (Millay and Taylor, 1974). The monosaccate pollen of this type ranges up to 180 μm in diameter and differs from Florinites pollen by possessing a conspicuous, complex trilete suture on the proximal surface (FIG. 20.22). Saccus and corpus are initially attached at both poles in Felixipollenites, but separation occurs at the distal pole late in development. The inner surface of the saccus contains prominent endoreticulations. Sullisaccites is another cordaitalean pollen type that is similar to Felixipollenites (Millay and Taylor, 1974). These monosaccate grains are bilateral, large (80 μm long), and trilete. Sullisaccites pollen has been found in Gothania cones attached to Mesoxylon priapi stems (Trivett and Rothwell, 1985).

Figure 20.20 Diagrammatic reconstruction of Gothania les-

liana compound strobilus containing pollen cones. (From Daghlian and Taylor, 1979.)

SEEDS

Cordaitalean ovules/seeds are often heart-shaped or cordate in outline, platyspermic (i.e., dorsiventrally flattened), and display bilateral or 180° rotational symmetry; some forms are characterized by conspicuous lateral extensions of the integument in the form of wings. One of the more common Pennsylvanian seed taxa thought to have been produced by Cordaitanthus-like reproductive structures is Cardiocarpus (FIG. 20.23) (Brongniart, 1881; Hilton et al., 2003a and references therein). Structurally preserved seeds of this type are so common in some coal-ball localities in Iowa that

Figure 20.21 Section of Gothania lesliana pollen sac show-

ing monosaccate pollen grains of Felixipollenites (Pennsylvanian). Bar  0.5 mm.

CHAPTER 20

CORDAITALES

799

Figure 20.22 Felixipollenites grain extracted from Gothania pollen sac (Pennsylvanian). Bar  60 μm.

hundreds can be removed from weathering coal balls in a relatively short time. Isolated ovules/seeds of similar morphology that are preserved as casts, impressions, or compressions (FIG. 20.24) have variously been referred to the genus Cordaicarpus as suggested by Seward (1917). Other scholars, however, use the name Cardiocarpus for seeds with an acute apex, whereas specimens with rounded apices are assigned to Cordaicarpon or Cordaicarpus (Remy and Remy, 1977). Structurally preserved seeds of Cardiocarpus spinatus, the most common Pennsylvanian form, are 2 cm long, with the apex attenuated into a micropyle and the base rounded or cordate shaped (Roth, 1955). They are 1 cm wide in the primary plane and approximately half that size in the secondary plane. The integument is three parted, adnate to the nucellus only at the base, and there is a simple pollen chamber at the distal end of the nucellus. A single terete vascular strand enters the base of the seed and gives off two slightly flattened strands that extend distally through the sarcotesta in the primary plane. The main strand continues to the base of the nucellus and flares slightly to form a shallow disk of tracheids. No Euramerican cordaite seeds are known in which the nucellus is vascularized. In C. samaratus and C. tuberculatus from the Lower Permian of China, however, the main bundle continues distally toward the base of the nucellus where it forms a vascularized sheath that extends distally for about one-third of the length of the nucellus (S.-J. Wang et al., 2003a, b).

Longitudinal section of Cardiocarpus taiyuanensis (Permian). Bar  500 μm. (Courtesy J. Hilton.)

Figure 20.23

Figure 20.24 Compressed (Pennsylvanian). Bar  1 cm.

seed

of

Cordaicarpus

sp.

Several C. spinatus ovules have been described with exquisitely preserved megagametophytes within the megaspore membrane; some have archegonia, which can be identified by jacket cells (Andrews and Felix, 1952). One

800

PaleOBotany: the biology and evolution of fossil plants

Figure 20.26 Longitudinal section in primary plane of Mitrospermum compressum (Pennsylvanian). Bar  2 mm.

Figure 20.25

Baolin Tian (Courtesy J. M. Hilton.)

exceptional case of fossil preservation was demonstrated by Baxter (1964b), who described starch grains, each with a hilum or nucleus, in the megagametophyte cells of a C. spinatus ovule. Cardiocarpus affinis, C. oviformis, and C. magnicellularis are three additional species that differ slightly in size and the disposition of cells in the integument (Reed, 1946; Baxter and Roth, 1954; Leisman, 1961; Rothwell, 1993). One particularly interesting species of Cardiocarpus is C. samaratus (S.-J. Wang and Tian, 1991a) (FIG. 20.25), the ovule associated with the Early Permian Shanxioxylon sinense plant from northern China (S.-J. Wang et al., 2003b). Although C. samaratus is similar in overall morphology to other cardiocarpalean ovules, it most closely resembles the ovules found inside the putative vojnovskyalean cone Sergeia neuburgii from the Upper Pennsylvanian of North America (Rothwell et al., 1996; Chapter 19). The genus Samaropsis includes impression–compression fossils of ovules of cordaitaleans, walchian conifers, and

several groups of pteridosperms. Samaropsis ovules are usually relatively small (5–20 mm long), circular to oval, and have a rounded, pointed or cordate base and rounded, pointed, or, notched apex. Surrounding the seed is a prominent membranaceous wing (Göppert, 1864). Cuticle preparations have been obtained from a few compressed Samaropsis seeds; for example, Šim˚unek and Libertín (2006) described wellpreserved cuticles of S. newberryi from the Westphalian of the Czech Republic. This species is one of the largest specimens of Samaropsis recorded to date, that is, up to 45 mm long and 52 mm wide. It has stomata arranged in distinct stomatiferous bands on both sides of the seed; however, one side is distinctly cutinized and has fewer stomata than the other side. In the Žacléˇr area of the Czech Republic, S. newberryi is associated with Cordaites schatzlarensis leaves (Šim˚unek and Libertín, 2006). Mitrospermum is another genus of structurally preserved ovules thought to have been produced by cordaitalean plants and is known from both European (Arber, 1910) and North American Carboniferous rocks (FIG. 20.26) (Taylor and

CHAPTER 20

CORDAITALES

801

Figure 20.27 Cross section of Mitrospermum compressum

showing expanded integumentary wings. Beneath seed is a section of the cordaite leaf (arrow) Cordaites felicis showing alternating fiber bands and vascular bundles (Pennsylvanian). Bar  1 mm.

Stewart, 1964; Grove and Rothwell, 1980). Specimens are 7.5 mm long and about the same width in the primary plane, where lateral extensions of the sarcotesta form conspicuous wings (FIG. 20.27) (Taylor and Stewart, 1964). In M. bulbosum, the integument is homogeneous with no differentiation of layers (Long, 1977c). The vascular system consists of a small, terete strand that enters the base of the seed and extends to the level of the nucellus; at this point, the bundle divides and a single strand enters each of the wings. Hilton et al. (2003a) suggested that M. bulbosum may be similar to Lyrasperma (Long, 1977c) and therefore may belong to a hydrasperman-type pteridosperm rather than the cordaites. In other species of Mitrospermum, such as M. compressum and M. vinculum, however, the consistent occurrence of Felixipollenites pollen in the micropyle suggests that these seeds may have been borne on cordaitalean plants that produced Gothania-type pollen cones. Kamaraspermum is a seed differentiated by the presence of an attenuated micropylar tube and a chamber in the base of the seed (Kern and Andrews, 1946). Baxter (1971b) suggested that these features are preservational and assigns these ovules to Mitrospermum. As is often the case in paleobotany, seeds of the Mitrospermum type have been found associated with different pollen cones, including those producing Florinites, Sullisaccites, and Felixipollenites pollen. Florinites pollen

Figure 20.28 Mature pollen cone of Mesoxylon priapi. (From Taylor and Taylor, 1993.)

was also produced in Cordaitanthus pollen cones, which have been found associated with seeds of the Cardiocarpus type. Trivett and Rothwell (1985) have shown that a number of vegetative and reproductive organs are variously associated with the stems Pennsylvanioxylon birame, Mesoxylon priapi (FIG. 20.28), and Cordaixylon dumusum, thus demonstrating the diversity in cordaitalean plants (FIG. 20.29). Their studies indicate that association alone cannot always be effectively used in understanding the biology of certain fossil plants. Based on the number of different types of detached Pennsylvanian ovules, these authors suggested that there are at least ten distinct cordaitalean plants from this time period.

ANGARAN CORDAITES A number of genera from Angara, principally from northeastern Asia, have been considered to be members of the Cordaitales (Meyen, 1984). These range from the Late Mississippian or Early Pennsylvanian to the Late Permian and locally represent significant, perhaps even predominant, constituents of the flora (Durante, 1995; DiMichele et al., 2001a). Meyen (1987) includes the families Rufloriaceae and Vojnovskyaceae in the Cordaitales

802

PaleOBotany: the biology and evolution of fossil plants

Cordaixylon dumusum

Mesoxylon birame

Mesoxylon priapi

Non-axillary cones

Cordaitanthus-type cones

Axillary cones

Gothania-type cones

Endarch leaf traces

Sympodial vascular architecture

Mesarch leaf traces

Nonsympodial vascular architecture

Cardiocarpus-type ovules

Mitrospermum-type ovules

Florinites-type pollen

Sullisaccites-type pre-pollen

Reparatory strands

Figure 20.29 Relationships among various cordaite genera based on common features and occurrence. (From Trivett and Rothwell,

1988.)

(Cordaitanthales of Meyen). Although there is some similarity in leaf morphology between these two groups and the cordaites, and ovule morphology of at least one cordaitalean species (i.e., the Shanxioxylon sinense plant) has been shown to parallel that in the vojnovskyalean cone Sergeia neuburgii, the reproductive biology of the Rufloriaceae and Vojnovskyaceae generally remains poorly understood. We have chosen to treat members of the Vojnovskyales as a group of enigmatic gymnosperms (Chapter 19) and will briefly consider Rufloria and a few associated organs here. Rufloria is a strap-shaped leaf with parallel veins that is one of the common Angaran foliage types and ranges from the Pennsylvanian to the Permian (Meyen, 1963). Like Noeggerathiopsis, Rufloria is characterized by abaxial furrows containing stomata and the apparent absence of sclerenchyma between the veins. Along the margin of the stomatiferous bands are epidermal cells with papillae (Archangelsky and Leguizamón, 1980). Glukhova (1976) established several subgenera of Rufloria based on cuticular and epidermal features. Cordaites is also known from the Permian of Angara and is defined by Meyen (1982a) to include leaves lacking abaxial furrows. Another presumed Angaran cordaitalean leaf type is Zamiopteris. This leaf is lanceolate, with the veins crowded near the base like those of Glossopteris. Stomata are present on the abaxial surface in wide bands. Zamiopteris has also been recorded for several Late Permian mixed Cathaysian–Angaran floras from China (reviewed in K. Sun, 2006). LePage et al. (2003) reported Rufloria and Zamiopteris leaves and associated putative cordaitalean reproductive structures, including Pechorostrobus, Bardocarpus, and

Sylvella from the late Early Permian of the Canadian Arctic. These have been used to suggest that a strong phytogeographic connection existed between this part of North America and the Angaran floral realm. This hypothesis, however, has been disputed by Ignatiev (2003) who regards the identifications as questionable. Strap-shaped, parallel-veined, amphistomatic leaves with stomata arranged in longitudinal intercostal bands have also been recorded for the Middle Permian of western Texas (DiMichele et al., 2004). Although these leaves resemble Rufloria, the exact affinities remain obscure. A perplexing aspect of dealing with the Angaran cordaites is the lack of detailed information about the reproductive organs. The reproductive structure Gaussia (Paravojnovskya, see Doweld, 2004) that has been associated with the Vojnovskyales was discussed in Chapter 19. Meyen (1984), however, considered this taxon to represent an ovulate organ of the Cordaitales. At least two distinct types of pollen cones are present among the Angaran cordaites, neither of which is morphologically similar to the pollen cones of Cordaitanthus or Gothania. Pechorostrobus from the Late Permian of the Pechora Basin in Russia consists of an axis containing ellipsoidal microsporophylls (microsporoclads of Meyen, 1982c), with pollen sacs borne in clusters along the sporophyll axis (Meyen, 1982c), suggestive of the aneurophytes (Ignatiev and Meyen, 1989). Pollen is monosaccate and a monolete-dilete mark occurs on the proximal surface. Cladostrobus is the other late Paleozoic Angaran pollen cone that is included in the Cordaitales, principally based on its association with Rufloria leaves (Maheshwari and Meyen, 1975). This taxon consists of

CHAPTER 20

helically arranged microsporophylls with rhomboidal distal laminae. Clusters of sporangia are attached to the pedicel of the sporophyll and appear to have been borne in pairs. The attachment of pollen sacs is similar to some of the cones included in the Voltziales. Pollen grains are oval and 30 μm in diameter. A weakly developed saccus covers the central body, except for a small region on the distal face. The pollen has been assigned to the sporae dispersae genus Cladaitina. Suchoviella synensis is a Late Permian cordaitalean seed cone from Angara (Ignatiev and Meyen, 1989). It consists of an axis with helically arranged seeds on short stalks (FIG. 20.30). The seeds, which are included in the genus Samaropsis, are flattened, with the integumentary wings most pronounced near the chalaza. Other Angaran Samaropsistype seeds have been described from the Permian (Gzhelian) of the Donets Basin (Boyarina, 2004). Ignatiev and Meyen (1989) suggested that Suchoviella synensis leaves assignable to Rufloria synensis, and microsporangiate reproductive structures of the Pechorostrobus type are parts of the same plant. In the same paper, these authors also introduced an alternative classification system for cordaitalean reproductive structures based on the morphology of sporangiophores and microsporangia, and the structure of the seed stalks and axillary complexes. This classification system has never become widely accepted and used. Other Angaran seeds with possible affinities in the Cordaitales include Bardocarpus and Sylvella. Bardocarpus seeds are winged, cordate, and sessile; they are helically arranged on an unbranched axis and characterized by a large, semicircular proximal notch and attenuated, “two-horned” apex (Meyen, 1982c, 1988b). Sylvella has a prominent micropylar projection and asymmetrical wings on both sides of the projection (Meyen, 1982a).

PHYLOGENETIC POSITION AND ORIGIN OF THE CORDAITES There are several opinions regarding the phylogenetic position of the Cordaitales. Although most consider the cordaites and conifers to be closely related, there is no universal agreement on the phylogenetic position of the cordaites. Based principally on Florin’s (1951) studies of cordaitalean and conifer reproductive structures, many scholars initially believed that the conifers evolved from the Cordaitales. The primary objection to this hypothetical relationship is that it is difficult to homologize the compound pollen cones of the cordaites with the simple pollen cones of the conifers (Grauvogel-Stamm and Galtier, 1998). The discovery of compound pollen cones in Pennsylvanian walchian conifers

CORDAITALES

803

Figure 20.30 Diagrammatic reconstruction of Suchoviella

synensis. (From Ignatiev and Meyen, 1989.)

from North America (Hernández-Castillo et al., 2001a, 2003) demonstrated, however, the existence of a basic structural equivalent between ovulate and pollen cones in at least some ancient conifers and thus added strong support to Florin’s hypothesis of a close relationship between conifers and cordaites. Both cordaites and conifers have pollen and seed cones in the form of compound shoot systems with fertile, secondary dwarf shoots in the axils of bracts. A modification of Florin’s idea views conifers and cordaites as evolving from a common ancestor, which some believe is most likely to be found within the archaeopteridalean progymnosperms (Beck, 1981). This hypothesis is supported by similarities in vegetative organization in both groups and wood anatomy (Elkins and Wieland, 1914; Schmid, 1967). An alternative hypothesis views the cordaites and conifers as evolving from a Pennsylvanian seed fern group such as the Callistophytales (Rothwell, 1982a). This idea focuses on the similarities in the reproductive biology of both groups. Rothwell suggested that the morphological and structural disparity present

804

PaleOBotany: the biology and evolution of fossil plants

between the fronds of seed ferns and the simple leaves of cordaites and conifers may result from heterochrony—morphological and structural changes resulting from modification in developmental patterns and their timing. Viewing the seed ferns as the progenitors of the cordaites also requires an evolutionary modification of the reproductive organs, however. This is interpreted as a result of a reduction in the size of pteridosperm leaves and a concomitant aggregation of the leaves to form a strobilus (Rothwell, 1982a). Although this hypothesis presents another alternative regarding the evolution of cordaites and conifers, there is currently little evidence to support this theory. Although the origin and phylogenetic position of the Cordaitales continue to remain problematic, research during the last decade has graphically demonstrated that this group is far more diverse than previously understood. This diversity has provided an expanded range of characters that may prove to be useful in more accurately evaluating the

evolutionary history of these late Paleozoic seed plants. Especially interesting in this context are several plants from the Cathaysian realm that clearly belong to the Cordaitales, but which share some features with certain fossil conifers, and others that, although they clearly belong to the conifers, share features with the Cordaitales (Hilton et al., 1999, 2003a). For example, Shanxioxylon sinense is undoubtedly cordaitean but shares with conifers the helical arrangement of the secondary fertile shoots (Cathayanthus) and small needle-like leaves (S.-J. Wang et al., 2003b). These plants are likely to play a significant role in understanding conifer and cordaite phylogeny in the future. A recent phylogenetic analysis by Hilton and Bateman (2006) resolves the cordaite Shanxioxylon as sister to a clade consisting of Cordaixylon– Pennsylvanioxylon plus Mesoxylon, which in turn is sister to a conifer group including the Paleozoic, Mesozoic, and several modern conifers.

21 CONIFERS EARLY CONIFERS ........................................................................ 806

Summary Discussion: Cheirolepidiaceae .........................................837

VOLTZIALES ................................................................................... 807

Podocarpaceae...................................................................................838 Summary Discussion: Podocarpaceae ..............................................843

Utrechtiaceae.....................................................................................808

Araucariaceae ....................................................................................843

Thucydiaceae ....................................................................................814

Summary Discussion: Araucariaceae................................................848

Emporiaceae ......................................................................................815

Cupressaceae .....................................................................................849

Majonicaceae ....................................................................................816

Summary Discussion: Cupressaceae ................................................859

Ullmanniaceae...................................................................................819

Sciadopityaceae.................................................................................860

Bartheliaceae .....................................................................................820

Pararaucariaceae................................................................................861

Other Voltzialeans .............................................................................820

Pinaceae ............................................................................................861

Ferugliocladaceae..............................................................................823

Summary Discussion: Pinaceae ........................................................868

Buriadiaceae ......................................................................................826

Cephalotaxaceae ...............................................................................868

Pollen Cones .....................................................................................826

Taxaceae ............................................................................................869

Summary Discussion: Voltziales.......................................................828

Summary Discussion: Cephalotaxaceae and Taxaceae .....................869

CONIFERALES .............................................................................. 830 CONCLUSIONS ........................................................................... 870 Palissyaceae ......................................................................................830 Cheirolepidiaceae ..............................................................................831

Between every two pines is a doorway to a new world. John Muir

Many of the plants included within the Coniferales are woody trees of enormous size (FIG. 21.1); others are small shrubs. The oldest living plants are found among the conifers— a bristlecone pine (Pinus longaeva) in eastern Nevada is estimated to be 5000 years old. Extant conifers are typically placed in seven families consisting of 60 genera and 650 species. With the exception of a few deciduous taxa, that is, Larix, Taxodium, Metasequoia, Glyptostrobus, and Pseudolarix, all are evergreen, and their foliage functions for more than a single growing season. In some high latitude populations of Pinus longaeva, the foliage may remain on the plant up to 40 years. Leaves are needles or scales with one to two veins, although in some genera the leaves

are broad and contain many vascular bundles (e.g., Agathis). Wood is dense (pycnoxylic) and characterized by circularbordered pits; the primary vascular system is a eustele with endarch xylem maturation. Reproductive organs consist of simple pollen cones with helically arranged microsporophylls and pollen sacs borne on the abaxial surface. Most conifers have ovulate cones; these are compound and consist of helically arranged ovuliferous scales, each subtended by a bract. Ovules are borne on the adaxial (upper) surface of the scales, and the entire unit is termed the bract–scale complex. The compound ovulate cone is considered to have evolved from a cordaitean-type reproductive unit by reduction of the dwarf shoot to form the modern bract. Cones may be borne

805

806

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.1 Life history of Pinus. (From Taylor and Taylor, 1993.)

on the same (monoecious) or, in some instances, different (dioecious) plants; all are monosporangiate. Numerous conifers, but not all, produce saccate pollen, and various types of polyembryony, that is, the production of more than one embryo per seed, are common in many taxa.

EARLY CONIFERS Perhaps, the oldest indisputable conifer is Swillingtonia denticulata from the Westphalian B (Pennsylvanian) of Yorkshire,

denticulata England (Scott and Chaloner, 1983). The taxon consists of fusinized twigs bearing leaves, which exhibit paracytic stomata arranged in two rows. Specimens of S. denticulata share some similarities with Walchia (FIG. 21.2), a Permian leafy conifer. Similar material has been reported from Pennsylvanian sites in North America and often referred to the walchian conifers (Lyons and Darrah, 1989). A stem with secondary xylem from the Upper Pennsylvanian (Westphalian C) may represent the oldest conifer remains based on anatomy (Galtier et al., 1992). Work on fossil conifers can be traced back to the detailed cuticular studies of

CHAPTER 21

CONIFERS

807

Higher taxa in this chapter:

Voltziales: Utrechtiaceae (Pennsylvanian–Lower Permian) Thucydiaceae (Pennsylvanian) Emporiaceae (Pennsylvanian) Majonicaceae (Upper Permian) Ullmanniaceae (Upper Permian) Bartheliaceae (Pennsylvanian) Ferugliocladaceae (Lower Permian) Buriadiaceae (Lower Permian) Coniferales: Palissyaceae (Triassic–Jurassic) Cheirolepidiaceae (Triassic–Cretaceous) Podocarpaceae (Triassic–recent) Araucariaceae (Triassic–recent) Cupressaceae (Triassic–recent) Sciadopityaceae (Cretaceous–recent) Pararaucariaceae (Jurassic) Pinaceae (Triassic–recent) Cephalotaxaceae (Jurassic–recent) Taxaceae (Middle Jurassic–recent)

Figure 21.3

Rudolph Florin.

Florin (FIG. 21.3) (1938, 1939a, b, c, d, 1940a, b, 1944a, b, 1945), who was instrumental in focusing attention on the evolution of the modern conifer bract–scale complex. Although some of Florin’s initial ideas have been challenged (Clement-Westerhof, 1988), his studies provided the impetus to critically examine both the morphology and the evolution of this group of fossil plants. The discovery of new and better-preserved specimens of several late Paleozoic conifers has greatly increased our understanding of the diversity of the group (Clement-Westerhof, 1988; Mapes and Rothwell, 1990; Rothwell et al., 1997). These studies have resulted in the addition of several new fossil families and the reorganization and redefinition of others. They have led to a more accurate characterization of vegetative shoots and a better understanding of features of the ovulate cones, which are important in defining families. The late Paleozoic and some Mesozoic conifer families are often included within the Voltziales, whereas the modern families and fossil members are included in the Coniferales. The reader is referred to the papers by Clement-Westerhof (1984, 1988), Visscher et al. (1986), Mapes and Rothwell (1990), and Rothwell et al. (1997, 2005) for a more in-depth treatment of late Paleozoic conifers.

VOLTZIALES

Figure 21.2 Walchia sp. (Permian). Bar  5 cm. (Courtesy W. A. DiMichele and D. Chaney.)

Historically, the voltzialean conifers have orthotropic branching in which leaves are broad, dwarf shoots bear flattened, partially fused scales (FIG. 21.4), stomatal complexes are scattered, and pollen is bisaccate. The so-called walchian conifers are characterized by plagiotropic branching with

808

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.5 Walchia sp. shoot (Pennsylvanian). Bar  2 cm.

Figure 21.4 Walchia sp. bract-scale complex showing fused

scales (Permian). Bar  5 mm.

needlelike leaves, dwarf shoots with radially arranged scales, stomata in bands, and pollen that is monosaccate (Looy, 2007). Phylogenetic analysis suggests that the walchians are nested within the voltzialeans (Rothwell et al., 2005). UTRECHTIACEAE

The genera included in the Utrechtiaceae, that is, Lebachia, Ernestiodendron, Ortiseia, Moyliostrobus, and Otovicia, were formerly included in the Lebachiaceae (Florin, 1938, 1939a, b, c, d, 1940a, b, 1944a, b, 1945) and the Walchiaceae (Clement-Westerhof, 1988). Some authors use the Walchiaceae in the place of Utrechtiaceae (Broutin and Kerp, 1994; Poort et al., 1997). Members of the Utrechtiaceae are forest trees with orthotropic (upright) stems and plagiotropic (horizontal) leafy shoots (Mapes and Rothwell, 1990; Rothwell and Mapes, 2003). Florin (1944a) illustrated lateral shoots of Utrechtia floriniformis as being borne in whorls. Leaves are scalelike, helically arranged, and only a few millimeters long; each is vascularized by a single vein.

Stomata are present on both surfaces of the leaf and tend to be organized into two elongated rows. In some species, heterophylly may occur in which leaves on the main axis and penultimate branches bifurcate, whereas those on the ultimate branches have an acute tip (Clement-Westerhof, 1984). In general, the morphology of the leaves closely resembles the living conifer Araucaria excelsa (the Norfolk Island pine). The stems were eustelic, and resin canals were absent. Many authors use the morphogenus Walchia (FIG. 21.5) for poorly preserved vegetative shoots in which cuticular features cannot be determined. The preservation of many early conifer leaves is variable, depending on the depositional environment, and may be due to the chemical composition of the cuticles (Rothwell et al., 1997). Walchia is also used for reproductive material in which only one ovule is produced per dwarf shoot (Clement-Westerhof, 1984). There is considerable disagreement regarding the taxonomy and systematics of many early conifers, in part based on the type of preservation and the need to develop concepts of whole plants. The reader is referred to Hernández-Castillo et al. (2001b) for a detailed discussion of many of these conflicting points of view. If found with well-preserved cuticle in which stomatal features can be examined, some authors place these plants in Lebachia or Ernestiodendron. Culmitzschia is

CHAPTER 21

CONIFERS

809

another generic name used for vegetative shoots with helically arranged leaves and a cuticular anatomy like that found in other early conifer genera. The relationship between vegetative and reproductive parts in a fossil assemblage is a basic requisite in order to define an entire plant, and the reproductive organs of the Pennsylvanian–Permian conifers consist of strobiloid structures borne at the ends of leafy branches. Like the cordaites, these are monosporangiate, with pollen and seedproducing cones perhaps borne on separate shoot systems of the same plant. In at least one species, Otovicia hypnoides, the reproductive structures were terminally positioned on lateral axes of the penultimate or ultimate order of branches (see the following sections). UTRECHTIA The organization of the ovulate cone is of particular significance in defining Paleozoic conifer families (ClementWesterhof, 1988). In Utrechtia, the compound ovulate cone consists of a central axis bearing helically arranged, bifid bracts. Arising from the axil of each bract is a short shoot consisting of an axis bearing numerous sterile scales and one or more fertile scales. Each fertile scale produced a single inverted ovule (with the micropyle facing the axis) attached to the lateral surface (Mapes and Rothwell, 1990). Utrechtia floriniformis is an ovulate cone 7 cm long with each short shoot ~8 mm long (Mapes and Rothwell, 1990). The fertile scales are obovate with cordate bases.

Figure 21.6 Ernestiodendron filiciforme (Permian). Bar 

2 cm.

ERNESTIODENDRON Leaves of Ernestiodendron (FIG. 21.6) appear to be borne at nearly right angles to the stems (FIG. 21.7), and the stomatal complexes are less well organized than in Lebachia. Each stoma is surrounded by four to eight papillate subsidiary cells. The primary axis of the Ernestiodendron ovulate strobilus (FIG. 21.8) contains helically arranged, bifid bracts. The cone (secondary) axes bear 30 scales, with only the distal four to six bearing ovules. The cones of E. filiciforme are up to 10 cm long and bear bracts that are entire. ORTISEIA In Ortiseia (Florin, 1964), leaves are helically arranged (FIG. 21.9) and possess dicyclic stomata. In O. leonardii, each ovule is 1.3 mm long and has two hornlike processes at the micropylar end (Clement-Westerhof, 1984). The genus is rare outside the Southern Alps (Poort et al., 1997). Pollen produced by Ortiseia is of the Nuskoisporites dulhuntyi type. These large, trilete monosaccate grains range up to 300 μm in diameter and possess an unusual shape—the

Figure 21.7 Ernestiodendron filiciforme, macerated twig segment (Permian). Bar  1 mm. (Courtesy M. Barthel.)

810

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.9 Ortiseia jonkeri Bar  2 mm. (Courtesy H. Kerp.)

leafy

axis

(Permian).

Figure 21.8 Ernestiodendron filiciforme, seed cone (Walchiostrobus gothanii) showing bifid bract (arrow) (Permian). Bar  1 cm. (Courtesy M. Barthel.)

sexine forms a flattened or distally inclined expansion giving the grain a bell-shaped configuration in lateral view (Poort et al., 1997). Internally this expanded region is similar to the protosaccate condition described in other conifer grains (e.g., Potonieisporites), and Nuskoisporites is interpreted as being protosaccate. Based on these unusual pollen grains, Ortiseia is suggested to be endemic to the European portion of the Euramerican floral province (Poort et al., 1997). OTOVICIA This taxon includes ovulate cones that morphologically resemble Utrechtia, but which differ in the shape of the fertile scales (FIG. 21.10) and the number of ovules per short shoot

Figure 21.10 Otovicia hypnoides ovuliferous scale (Permian). Bar  4 mm. (Courtesy H. Kerp.)

CHAPTER 21

CONIFERS

811

Figure 21.11 Diagrammatic reconstruction of Otovichia hyp-

noides ovuliferous scale. (From Kerp et al., 1990.) Otovicia hypnoides (Lebachia hypnoides), sterile twig (Permian). Bar  5 mm. (Courtesy M. Barthel.)

Figure 21.12

(FIG. 21.11) (Kerp et al., 1990). In O. hypnoides (FIGS. 21.12, 21.13), ovules are up to 9 mm long and have pollen of Potonieisporites inside the elongated, funnel-like nucellar beak (FIG. 21.14). The large saccus on these early conifer pollen grains suggests that anemophily was the principal pollination syndrome in these plants. MOYLIOSTROBUS Another conifer cone known from permineralized specimens of the Lower Permian of Texas, USA, is Moyliostrobus (Miller and Brown, 1973b). Moyliostrobus texanum appears to share a number of features with ovulate cones of members of the Utrechtiaceae. The compound cone is 2 cm in diameter and bears ovate bracts, each with a single vascular bundle. Short shoots are flattened and have 20–50 sterile bracts; each ovule is attached to the central surface of the cone scale. Ovules are dorsiventrally flattened and recurved toward the cone axis (Mapes, 1987). The integument is three parted and attached to the nucellus only in the base. A few seeds contain an elliptical zone of parenchyma with a few tracheids that are interpreted as representing embryos (Miller and Brown, 1973a). OTHER TAXA The morphogenus Walchiostrobus (FIG. 21.15) is used for ovulate cones that have a number of sterile scales and three fertile scales per short shoot; a single ovule is attached

Figure 21.13 Branch of Otovicia hypnoides with pollen cones (arrows) (Permian). Bar  3 cm.

812

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.14 Ovule of Otovicia hypnoides showing bifid apex

(Permian). Bar  2.5 mm.

on the abaxial surface of each scale. Walchianthus (FIG. 21.16) is used for isolated pollen cones with the same epidermal features as those found in the vegetative members of the Utrechtiaceae. Thuringiostrobus was separated from Walchiostrobus (FIG. 21.17) when it was discovered that it bore a different number of fertile and sterile scales. In Thuringiostrobus meyenii (FIG. 21.18), there are four fertile scales per short shoot and none to many sterile scales (Kerp and Clement-Westerhof, 1991). The pollen cones of this family, like those of many fossil conifers (FIG. 21.19), are relatively modern in appearance (Visscher et al., 1986). Unlike the seed cones, they are borne terminally on leafy branches. The cone consists of a central axis surrounded by numerous helically arranged, flattened microsporophylls. Each sporophyll has an upturned distal segment that overlaps the sporophyll above and a small basal projection that partially covers the pollen sacs. On the abaxial surface of the microsporophyll are two elongated

Figure 21.15 Walchiostrobus gothanii (Permian). Bar  5 mm. (Courtesy BSPG.)

microsporangia (pollen sacs) that dehisce longitudinally. Pollen grains of the Utrechtiaceae are monosaccate, with some 300 μm in diameter. On the proximal surface of some grains is a trilete mark. Because the grains are typically compressed,

CHAPTER 21

CONIFERS

813

it is difficult to determine whether they are eusaccate or protosaccate. Detailed ultrastructural examination has been completed on the pollen of several of these cones indicating they are protosaccate (Taylor and Grauvogel-Stamm, 1995).

Figure 21.16 Walchianthus sp. cone (Pennsylvanian). Bar 

4 mm.

Figure 21.17 Walchiostrobus gothanii showing short shoots (arrows) (Pennsylvanian). Bar  1.5 cm. (Courtesy BSPG.)

(A)

(B)

(C)

Figure 21.18 Adaxial view of ovuliferous short shoots of: A. Walchiostrobus gothanii, B. Thuringostrobus florinii, and

C. Thuringiostrobus meyenii. Dark ovals represent attachment points of ovules. (From Kerp and Clement-Westerhof, 1991.)

814

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.20 Cycadocarpidium sp. (Triassic). Bar  5 mm. (Courtesy L. Grauvogel-Stamm.)

THUCYDIACEAE

Figure 21.19 Otovicia hypnoides pollen cone (Permian). Bar 

4 mm.

When dispersed, the pollen extracted from cones of this group is included in the sporae dispersae genus Potonieisporites. Cycadocarpidium (FIG. 21.20) is used for Triassic sporophylls with parallel venation of the Podozamites type that were initially interpreted as cycads, based on the occurrence of structures thought to be as seeds at the base of the sporophyll (Nathorst, 1886). Subsequent studies suggested that the fossil is more closely associated with conifers (Harris, 1935; Florin, 1953) or with forms such as Aethophyllum and Swedenborgia (Grauvogel-Stamm, 1978). Kimura and Ohana (2000) identified Cycadocarpidium bracts with two parallel veins; each bract is attached to an ovuliferous scale, which bears a single seed.

This family is known from numerous specimens from a Late Pennsylvanian (Stephanian) shale in Ohio, USA (HernándezCastillo et al., 2001b). Thucydia mahoningensis includes three orders of vegetative and fertile shoots that demonstrate different levels of plagiotropy (FIG. 21.21). Ovulate fertile zones are present between vegetative regions (FIGS. 21.22, 21.23) on the stems and consist of compact bracts, each subtending an axillary dwarf shoot, which consists of 10–15 sterile scales and 3–4 fertile ones. The fertile scales (sporophylls) have a recurved apex and bear a single, inverted ovule. One of the unusual features of the pollen cone is its compound organization (Hernández-Castillo, 2001a). The pollen cones consist of small helical, leaflike bracts (FIG. 21.24). In the axil of each bract is a dwarf shoot, consisting of sterile and fertile scales, the latter bearing pollen sacs (Hernández-Castillo et al., 2001b). Pollen is of the Potonieisporites type. This organization for pollen cones is unlike any known from the earliest fossil conifers described to date and can be compared to both cordaitean pollen cones (Chapter 20) and

CHAPTER 21

CONIFERS

815

V

O

Figure 21.21 Suggested reconstruction of Thucydia mahonin-

gensis. (From Hernández-Castillo et al., 2001.)

modern gnetaleans (Chapter 19). Another unusual feature of T. mahoningensis is the difference in stomatal distribution between fertile and sterile leaves. Bracts and sterile scales have evenly distributed stomata, whereas stomata on the vegetative leaves occur in bands on the adaxial surface.

EMPORIACEAE

Permineralized conifer remains from the Upper Pennsylvanian of Kansas are included in the family Emporiaceae (Mapes and Rothwell, 1990). These conifers have penultimate vegetative shoots with simple and/or bifid leaves. Papillae and trichomes are present on the epidermis of both the stems and the leaves. The ovulate cones of Emporia lockardii (Lebachia lockardii of Mapes and Rothwell, 1984) are up to 5.5 cm long and consist of an endarch, eustelic primary axis with helically arranged, bifid bracts. In the axil of each bract is a flattened fertile shoot (Rothwell, 1982a), consisting of 20–30 sterile scales and 1–5 fertile scales. Each fertile scale bears a single, inverted platyspermic ovule attached at the tip (Mapes

Figure 21.22 Thucydia mahoningensis showing vegetative axis (V), ovulate fertile zone (O), and isolated seeds (arrows) (Pennsylvanian). Bar  1 cm. (Courtesy G. Hernández-Castillo.)

and Rothwell, 1990). Pollen grains of the Potonieisporites type are preserved in the pollen chambers of several ovules (FIG. 21.25), including those that are described as immature. The larger size of the pollen grains when compared to the diameter of the micropyle suggests that a pollination droplet may have been a component of the reproductive biology of these plants (Mapes and Rothwell, 1984). Polycotyledonous embryos (FIG. 21.26) occur in several seeds (Mapes and Rothwell, 1988). Hanskerpia is a Late Pennsylvanian conifer from North America that combines characters of several existing groups, including the Emporiaceae (Rothwell et al., 2005). The branches are plagiotropic and produce simple, amphistomatic leaves up to 10 cm long. The stems are eustelic with a zone

816

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.23 Ultimate leaves of vegetative shoot of Thucydia mahoningensis (Pennsylvanian). Bar  1 mm. (Courtesy G. Hernández-Castillo.)

of dense wood and narrow band of periderm. Seed cones are compound and produced on penultimate shoots; ovules are inverted and produced singly or in pairs on each sporophyll. Although pollen cones are not known for H. hamiltonensis, pollen in the pollen chamber of some seeds is comparable to Potonieisporites. Based on a cladistic analysis of this walchian conifer and 21 other Paleozoic conifer taxa, Hanskerpia is nested within the walchian Voltziales or walchian conifers of Mapes and Rothwell (1991), a paraphyletic group, according to this analysis. The analysis also shows a strong provinciality in these extrabasinal conifer taxa from Europe and North America, which contrasts with contemporaneous wetland floras. Rothwell et al. (2005) suggested that the familial arrangement of the Carboniferous–Late Permian conifers may require reinterpretation. MAJONICACEAE

The vegetative shoots of the Majonicaceae bear helically arranged, flattened leaves that are amphistomatic. In some taxa, leaves are up to 3 cm long and, like many of the other early conifers, heterophylly is common. The stomata in Majonica alpina occur in more or less irregular rows. The primary feature used to characterize this family is the lateral-

Figure 21.24 Thucydia mahoningensis pollen cone (Pennsylvanian). Bar  1 cm. (Courtesy G. Hernández-Castillo.)

to-adaxial position of the ovule on the fertile scale (FIGS. 21.27, 21.28), a feature that results in ovules that are not in a single plane (Clement-Westerhof, 1987, 1988). The short shoots in M. alpina may be partially fused to the subtending bract and are somewhat flattened. Lebowskia is a Late

CHAPTER 21

CONIFERS

817

B SS

FS O Figure 21.25 Potonieisporites pollen grain (arrow) in pol-

len chamber of Emporia lockardii (Pennsylvanian). Bar  70 μm. (Courtesy G. Mapes.)

Figure 21.27 Diagrammatic reconstruction of bract-ovuliferous

scale Majonica alpina. O  ovule, B  bract, FS  fertile scale, SS  sterile scale. (From Taylor and Taylor, 1993.)

E

Figure 21.26 Emporia lockardii megaspore containing embryo

(E) (Pennsylvanian). Bar  225 μm. (Courtesy G. Mapes.)

Permian member of this family characterized by leaves that are more than a centimeter long with an obtuse apex (Looy, 2007). Stomata are scattered on both leaf surfaces, and a single large median ovuliferous scale is flanked by two smaller ones (FIG. 21.29). Lebowskia grandifolia (FIG. 21.30) suggests a lineage of Permian conifers in North America that is different from the voltzialeans in Europe. Pseudovoltzia is another member of this family that has vegetative organs and cuticle (FIGS. 21.31, 21.32) similar to those of many early conifers (Florin, 1927; Uhl, 2004). The ovuliferous cones of Pseudovoltzia liebeana occur as permineralized specimens from the Upper Permian of Europe (Schweitzer, 1963). In this genus, the bract and the dwarf shoot are partially fused, and the ovules are borne directly

Figure 21.28 Ovuiferous scale of Majonica alpina (Permian). Bar  1 mm. (Courtesy H. Kerp.)

818

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.29 Ovuliferous cone and dwarf shoots (arrows) of Lebowskia grandifolia (Permian). Bar  5 mm. (Courtesy C. V. Looy.)

on fertile scales (FIG. 21.33), not on stalks as was initially interpreted by Florin (1951). The dwarf shoots are partially flattened and include two sterile scales and three fertile ones (FIG. 21.34). Each fertile scale has a single ovule, but the ovules are not all in the same plane. Schweitzer (1996) used P. hexagona as the starting point in a transformational series ending in a cone–scale complex like that seen in extant seed cones of Cryptomeria japonica.

Figure 21.30 Diagrammatic reconstruction of Lebowskia grandifolia. (Courtesy C. V. Looy.)

Dolomitia cittertiae from Italy has ovulate cones 2 cm wide that contain 13 triangular, sterile scales (ClementWesterhof, 1987). There are three fertile scales, with the two lateral ones slightly recurved (FIG. 21.35); ovules were attached to the abaxial surface. Another conifer that has a unique combination of characters is Cassinisia orobica, a three-dimensionally preserved vegetative branch from the Permian of northern Italy (Kerp et al., 1996). It has thick, overlapping leaves with falcate tips. Although the affinities of this vegetative material remain unknown, the presence of

CHAPTER 21

CONIFERS

819

Figure 21.32 Cuticle of Pseudovoltzia liebeana showing sto-

matal complexes (Permian). Bar  10 μm. (Courtesy H. Kerp.)

Figure 21.31 Pseudovoltzia liebeana leafy branch (Permian). Bar  1 cm. (Courtesy H. Kerp.)

sunken stomata on the abaxial leaf surface suggests the plant may have grown in a meso-xerophytic ecosystem. ULLMANNIACEAE

The vegetative branches of the Ullmanniaceae have lanceolate-to-ovate decurrent leaves that are up to 8 cm long.

Figure 21.33 Lateral view of Pseudovoltzia liebeana showing

ovule attached to fertile scale and bract. (From Taylor and Taylor, 1993.)

Monocyclic stomata occur on both surfaces and are arranged in irregular rows (Florin, 1944a); papillae are present on some epidermal cells. The ovulate cones of Ullmannia are up to 6 cm long and include bracts with an acute apex. In this family, the short shoots are flattened, with the individual

820

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.34 Adaxial view of Pseudovoltzia liebeana showing position of ovules (dark ovals) on fertile scales and sterile scales between. (From Taylor and Taylor, 1993.)

Figure 21.36 Morphology of Ullmannia sp. fertile scale show-

ing position of ovule and tip of bract. (From Taylor and Taylor, 1993.)

Figure 21.35 Adaxial view of Dolomitia cittertiae showing three fertile scales (dark ovals) and sterile scales between. (From Taylor and Taylor, 1993.)

scales fused together, much like the ovuliferous scale in extant conifers. The flattened short shoot, however, is still free from the subtending bract (FIG. 21.36). A single ovule is present on the adaxial surface of the fertile scale in U. bronnii (Florin, 1944b). Pollen cones are simple, with abaxially borne, elongated pollen sacs. Although the pollen in this family has been described as including both monosaccate and bisaccate grains (Florin, 1951), it appears that all the pollen is monosaccate with endoreticulations inside the eusaccus. Germination probably occurred through the aperture on the distal surface.

(Rothwell and Mapes, 2001). Leaves of Barthelia furcata (Pennsylvanian, Hamilton Quarry, Kansas) have epidermal cells that are papillate, and trichomes are present near the base of the leaf. Stomatal complexes are arranged in bands, with each complex consisting of four to seven subsidiary cells that typically have thickened walls. Pollen cones are small, up to 5 mm long, and consist of helically arranged microsporophylls, some with a bifid tip. The number of pollen sacs per microsporophyll is not known; pollen is monosaccate with endoreticulations extending inwardly a short distance from the inner surface of the saccus (eusaccate). They compare with grains of the Potonieisporites type. Seed cones of B. furcata are compound and occur in fertile regions along the axis. Each cone is radial and includes numerous sterile scales with ovules orthrotropus on narrow sporophylls. Barthelia furcata shares a number of features with Emporia lockardii, which also occurs at the same site but differs in having longer bifid leaves with stomatiferous bands confined to the adaxial surface. Transverse pith septations with sclerotic nests also serve to distinguish Barthelia from other coniferophytes. OTHER VOLTZIALEANS

BARTHELIACEAE

This family includes plants with irregularly branched shoots with leaves up to 5 cm long that may be forked or simple

There are a number of additional Paleozoic and Mesozoic conifers that are not known in sufficient detail to be included within well-defined families. In earlier treatments, taxa now included

CHAPTER 21

CONIFERS

821

Figure 21.37 Glyptolepis hungarica showing two recurved

ovules and five sterile scales. (From Miller, 1977.)

Figure 21.38 Adaxial surface of Voltziopsis africana. (From

Miller, 1977.)

in the Utrechtiaceae, Majonicaceae, and Ullmanniaceae were classified in the Lebachiaceae, with the remaining morphotaxa placed in the Voltziaceae (Miller, 1977). We have chosen to treat some additional forms as morphogenera, but understand that even their assignment to families within the Voltziales is only provisional. Glyptolepis is a morphogenus known from the Permian and the Triassic. It was originally described as having bract– scale complexes with long cone–scale stalks. The cone scales are distinctly flattened and have been reconstructed with two recurved seeds (FIG. 21.37). A reanalysis of the genus suggests that Glyptolepis as presently interpreted includes a number of different forms that do not conform to the type species, G. keuperiana (Axsmith and Taylor, 1997). A species from the Upper Triassic of Germany, G. richteri, does conform to the original description of the genus. Glyptolepis richteri includes bract–scale complexes with long cone–scale stalks; the distal end of the stalk is expanded into eight, equal-sized lobes. The subtending bract is fused to the cone– scale stalk and curves upward, completely covering the cone scale on the surface of the cone. Conewagia is another Triassic ovulate cone that is associated with helically arranged leaves on vegetative shoots (Axsmith et al., 1998a). The flattened bract–scale complexes of C. longiloba are oriented in a single plane and composed of 11 foliar components. Isolated seeds occur at the same collecting site, but none have been found attached. The seeds

are elliptical and characterized by a structure that may represent a winglike extension of the integument. Voltziopsis (Permian–Late Triassic) is a form known only from the Southern Hemisphere and includes both vegetative remains and seed cones (Townrow, 1967a). Branches have dimorphic leaves—the smaller scalelike ones are triangular and 3 mm long, whereas the larger leaves (3 cm long) extend out from the stems in two rows and are slightly twisted. Ovulate cones occur at the ends of short branches. Each strobilus consists of 25 bract–scale units; the scales are five lobed, with each lobe bearing a single ovule (FIG. 21.38). In Voltzia heterophylla, leaves are helically arranged and needle like (FIG. 21.39). Quadrocladus is used for leafy shoots with irregular branching and short, subtriangular leaves with conspicuous papillae surrounding the stomata (FIG. 21.40) (Meyen, 1997; Uhl and Kerp, 2005). Florinostrobus andrewsii includes seed cones from an Upper Triassic site (Pekin Formation) (Delevoryas and Hope, 1987). The ovuliferous scales in this cone were attached to the axis at nearly right angles, with the distal portion five lobed and bent upward. On the adaxial surface of the scale are three scars thought to represent the position of ovules. The subtending bract in F. andrewsii is believed to have been fused with the scale and only free at the tip (Delevoryas and Hope, 1987). Depending on the degree of fusion of the fertile scale, the presence of three ovules in F. andrewsii might suggest affinities with the Majonicaceae.

822

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.41 Adaxial view showing three recurved ovules and

extended lobes of Tricranolepis frischmannii. (From Miller, 1977.)

Figure 21.39 Branch and elongated leaves of Voltzia hetero-

phylla (Triassic). Bar  1 cm. (Courtesy L. Grauvogel-Stamm.)

Figure 21.40 Stomatal complex of Quadrocladus orobiformis (Permian). Bar  15 μm. (Courtesy H. Kerp.)

Tricranolepis is another isolated seed cone (Late Triassic– Early Jurassic) (Roselt, 1958). There is a single scale with three large lobes, each fused to the stalk of a recurved ovule (FIG. 21.41). A similar scale morphology is seen in

Figure 21.42 Swedenborgia sp. seed cone (Triassic). Bar 

1 cm. (Courtesy K.-P. Kelber.)

Swedenborgia (FIG. 21.42), except that in this genus there are five scale lobes and an equal number of recurved ovules (FIG. 21.43); the bract is fused with the fertile scale (Harris, 1935).

CHAPTER 21

CONIFERS

823

Figure 21.43 Swedenborgia cryptomerioides seed cone scale (Jurassic–Triassic). Bar  5 mm. (Courtesy BSPG.)

Aethophyllum is an interesting conifer from the Lower Triassic Voltzia sandstones (Grès à Voltzia) of France that has historically been affiliated with the angiosperms and sphenophytes (Grauvogel-Stamm, 1978). Each leaf is linear and decurrent on the stem. Mature plants of A. stipulare are small plants that rarely exceeded 2 m (FIG. 21.44). Stems were up to 2 cm in diameter and contained an endarch eustele with parenchymatous pith and numerous air spaces (Rothwell et al., 2000). Ovulate cones are lax with acuminate bracts (FIGS. 21.45, 21.46). The fertile scale is characterized by five hornlike projections (FIG. 21.47), each associated with a recurved seed (Grauvogel-Stamm and Grauvogel, 1975) (FIG. 21.48). Pollen is bisaccate (FIG. 21.49). Discovery of immature plants, some as small as 4 cm tall, suggestive of the seedling stage, and fertile specimens up to 30 cm tall provide convincing evidence that growth in A. stipulare was rapid, similar to that in modern, herbaceous ruderals.

Figure 21.44 Suggested reconstruction of Aethophyllum stip-

ulare (Triassic). (From Grauvogel-Stamm, 1978.)

FERUGLIOCLADACEAE

The Ferugliocladaceae is a family of Early Permian conifers from Gondwana, which includes both vegetative branches and reproductive organs (Archangelsky and Cúneo, 1987). Specimens of Ferugliocladus from Argentina include several

orders of branches bearing small (1 cm long), linear leaves, each with a single vein (FIG. 21.50). Ovulate cones (FIG. 21.51) are borne terminally and consist of helically arranged bracts and axillary, orthotropous, platyspermic ovules, each

824

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.45 Aethophyllum stipulare branch with immature

pollen cone (Triassic). Bar  2 cm. (Courtesy L. Grauvogel-Stamm.)

Figure 21.47 Fertile scale of Aethophyllum stipulare characterized by hornlike projections with each bearing a single recurved ovule. (From Grauvogel-Stamm and Grauvogel, 1975.)

Figure 21.46 Aethophyllum stipulare pollen cone (Triassic). Bar  1 cm. (Courtesy L. Grauvogel-Stamm.)

7 mm long (FIG. 21.52). In F. riojanum, the seeds have a bifid apex. Pollen cones are terminal and 2 cm long. The position of the pollen sacs is not clear, but the pollen grains are monosaccate and of the Cannanoropollis type. Ugartecladus has the same general organization as ovulate cones of Ferugliocladus, except that the seeds lack a bifid micropyle (Archangelsky and Cúneo, 1987). Two of the most unusual features of this family are the apparent absence of subtending sterile scales and the orthotropous position of the ovules. Archangelsky and Cúneo (1987) suggested that, perhaps, the sterile and fertile scales have been fused into the stalk of the ovule and

CHAPTER 21

CONIFERS

825

Figure 21.50 Leafy branch of Ferugliocladus patagonicus

(Permian). Bar  1 cm. (From Archangelsky and Cúneo, 1987.)

Figure 21.48 Léa Grauvogel-Stamm.

Figure 21.49 Aethophyllum stipulare pollen grain (Triassic). Bar  30 μm. (Courtesy L. Grauvogel-Stamm.)

therefore cannot be distinguished in the compressed fossils. This hypothesis would mean an ancestor like that proposed for other early conifers in which the sterile and fertile scales have become flattened, and phylogenetic analyses place the Ferugliocladaceae in a basal position to other late Paleozoic forms (see below). Another theory is that the axillary ovule in the Ferugliocladaceae has evolved from an ancestor with axillary, orthotropous ovules. Genoites demonstrates features that could be interpreted as ancestral to the ovulate cone organization in the Ferugliocladaceae (Cúneo, 1985). This plant has branches bearing helically arranged, bifid leaves. In the axils of some leaves

Figure 21.51 Seed cone attached to vegetative shoot of Ferugliocladus patagonicus (Permian). Bar  1 cm. (Courtesy I. Escapa.)

826

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.52 Seed with bifid apex assigned to Ferugliocladus patagonicus (Permian). Bar  1 cm. (Courtesy I. Escapa.)

are stalked, orthotropous ovules (FIG. 21.53). A reduction of the lax, ovule-bearing branch of G. patagonica could result in an ovulate cone much like Ugartecladus or Ferugliocladus. The discovery of these morphologically distinct, ovulate reproductive organs in Gondwana rocks suggests that the diversity of early conifers is higher than presently postulated. In addition, paleoecological studies suggest that the members of the Ferugliocladaceae were ecologically more diverse and perhaps reproductively more advanced than their equatorial counterparts. BURIADIACEAE

The Permian Buriadiaceae are a Gondwanan family that includes plants with woody branches and helically arranged polymorphic leaves that are decurrent at the base (FIG. 21.54). Leaf tips may range from simple to many times forked. Stomata are absent from the abaxial leaf surface. Some leaves of Buriadia heterophylla are nearly 3 cm long and have papillae along the margin (Florin, 1940b). The secondary xylem of the branches is of pycnoxylic type with uniseriate- to multiseriate-bordered pits on the radial walls of the tracheids. Although ovules were described attached along the stems (FIG. 21.55) (Pant and Nautiyal, 1967b), a reexamination of the original material indicates that none are actually attached (Singh et al., 2003). Similarly, it remains unknown whether

Figure 21.53 Fertile shoot of Genoites patagonica show-

ing position of stalked ovule (arrow) in the axil of the helically arranged bifid leaves (Permian). Bar  1 cm. (Courtesy I. Escapa.)

the ovules were aggregated into any type of strobilus. The seeds are platyspermic with a prominent, curved micropylar beak. Seed coat and nucellus are fused only at the base of the seed. Monocolpate pollen grains up to 100 μm long have been identified at the tip of the nucellus in some specimens. In Coricladus from the Lower Permian of Brazil, which may be related to Buriadia, the distal ends of vegetative branches produce four ovuliferous scales, each with a bifid tip and two seeds (Jasper et al., 2005). POLLEN CONES

Included with the early conifers are a number of interesting pollen cones in which the attachment of the pollen sacs differs from modern and younger conifers (FIG. 21.56). One of these from the Triassic is Masculostrobus acuminatus (Grauvogel-Stamm and Grauvogel, 1973; Grauvogel-Stamm

CHAPTER 21

CONIFERS

827

Figure 21.54 Branched shoot of Buriadia heterophylla. (From

Taylor and Taylor, 1993.) Figure 21.55 Suggested reconstruction of Buriadia hetero-

and Schaarschmidt, 1979). In this genus, each microsporophyll bears approximately six elongated pollen sacs attached to the heel of the microsporophyll and parallel to the stalk of the microsporophyll. Pollen is bisaccate, 70–130 μm long, and similar to modern conifers in appearance. There is a proximal cap and distal furrow. Another cone that is morphologically similar to Masculostrobus is Voltziostrobus (Early Triassic) (Grauvogel-Stamm, 1969). Specimens of V. schimperi may be up to 10 cm long and are organized as a series of helically arranged, peltate microsporophylls. Up to 20 elongate pollen sacs per sporophyll are attached by dichotomizing stalks to the pedicel of the microsporophyll. Pollen sacs are attached in a similar manner in Sertostrobus (Grauvogel-Stamm, 1969), but this cone is more lax and there are a smaller number of pollen sacs on each sporophyll pedicel. Pollen is bisaccate and in the 50- μm size range. In

phylla shoot with lateral seed. (From Taylor and Taylor, 1993.)

Ruehleostachys and Willsiostrobus (FIG. 21.57), the microsporangia extend toward the cone axis from the lower portion of the peltate microsporangiophore (FIG. 21.58), whereas in Hercynostrobus, pollen sacs are club shaped and arise in a digitate pattern from the surface of the sporophyll (bract) (Arndt, 2002). In Leastrobus, a permineralized form from the Triassic, the pollen sacs are attached to the inner surface of the laminar portion of the microsporophyll (Hermsen et al., 2007b). In many of these cones, the pollen is bisaccate (FIG. 21.59). In characterizing the diversity of pollen cones attributed to the late Paleozoic and Mesozoic conifers, Meyen (1997) proposed three basic types: those with sporangia attached to a shield-like microsporophyll, those with sporangia adnate

828

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.57 Willsiostrobus sp., pollen cone (Triassic). Bar  1 cm. (Courtesy K.-P. Kelber.)

Figure 21.56 Longitudinal section of the permineralized pol-

len cone Masculostrobus sp. (Jurassic). Bar  3.5 cm. (Courtesy BSPG.)

to the sporophyll stalk, and those attached to the stalk of the microsporophyll by delicate structures he termed sporangiophores (e.g., Darneya) (FIG. 21.60). Others have suggested that the latter type should be regarded as a compound structure since the delicate stalks are attached at regular intervals along the stalk of the microsporophyll and are therefore suggestive of a fertile shoot (Grauvogel-Stamm and Galtier, 1998). Figure 21.58 Pedicle (arrow) and elongated pollen sacs

SUMMARY DISCUSSION: VOLTZIALEANS

The Voltziales have historically been referred to as transition conifers, that is, intermediate between Carboniferous Cordaitales and modern families within the Coniferales. The most complete phylogenetic analyses of the late Paleozoic conifers to date are those of Hernández-Castillo (2005) (FIG. 21.61) and Rothwell et al. (2005), which included 20 and 18

of Willsiostrobus sp. L. Grauvogel-Stamm.)

(Triassic).

Bar  2.5 mm.

(Courtesy

conifer taxa, respectively, along with a vojnovskyalean and two cordaites. Both analyses resolve the Cordaitales and Vojnovskyales as basal to the Voltziales, which form a clade of all the early conifers. Within the voltzialean clade are three

CHAPTER 21

21.59 Willsiostrobus rhomboidalis pollen (Triassic). Bar  35 μm. (Courtesy L. Grauvogel-Stamm.)

Figure

grain

CONIFERS

Figure 21.60 Darneya peltata microsporophyll with peltate head (arrow) and numerous pollen sacs (Triassic). Bar  5 mm. (Courtesy L. Grauvogel-Stamm.)

Cordaiteans

 Callistophyton poroxyloides Cordaixylon dumusum

59 54

Mesoxylon priapii

100

Vojnov. Plant

66 60

84 79

Ferugliocladus spp. Genoites patagonica

Angara

Timanostrobus muravievii 83 83

Gondwana

Dicranophyllum hallei 53

829

Concholepis harrisii Kungurodendron sharovii Thucydia mahoningensis Ernestiodendron filiciforme

Aethophyllum stipulare 61 60

Dolomitia cittertiae

69 63

Voltziales

Voltzia hexagona 54 56

Voltzian

Ortiseia spp.

Majonica alpina

Utrechtia floriniformis

Barthelia furcata Hanskerpia hamiltonensis Emporia lockardii 66 56

91 79

Lebachioid

Otovicia hypnoides

Emporia royalii Emporia cryptica

Figure 21.61 Suggested phylogeneteic relationships among voltzialean conifers. (From Hernández-Castillo et al., 2003.)

830

PaleoBOtany: the biology and evolution of fossil plants

additional clades: (1) the Gondwanan voltzialeans, which includes the Ferugliocladaceae; (2) the voltzian Voltziales (Majonica, Dolomitia, Aethophyllum, Voltzia hexagona, and Ortiseia); and (3) the lebachioid Voltziales (Utrechtia, Otovicia, Emporia, Barthelia, and Hanskerpia). The walchian Voltziales resolve as a paraphyletic group, which includes the lebachioid Voltziales plus Ernestiodendron and Thucydia. The relationships of these late Paleozoic taxa with modern conifer taxa are still unclear. Some analyses suggest that the Cordaitales and modern conifers are separate clades (Doyle, 2006) or simply constitute a polyphyletic basal grade of conifers (Hilton and Bateman, 2006), although the former analysis only included Emporia and the later only Emporia and Thucydia among the Paleozoic conifers. The discovery of additional voltzialean fossils leading to the reconstruction of entire plants, perhaps based on detailed cuticular analysis techniques, may be the principal method available to better understand the phylogenetic history of many of these unusual plants.

CONIFERALES PALISSYACEAE

The Palissyaceae are a small family of Triassic and Jurassic conifers that have distinctive bract–scale complexes (Florin, 1958). Palissya (FIG. 21.62) is believed to have been a woody plant with alternately arranged, vegetative branches and persistent, helically arranged, linear leaves. The stomata are confined to narrow bands on the lower surfaces of the leaves, one band on each side of the midrib. In P. sphenolepis, stomata are aligned longitudinally and the subsidiary cells contain overarching papillae. Ovulate cones in Palissya occur singly at the terminal end of lateral branches (FIG. 21.63). Each cone is up to 10 cm long and borne on a peduncle with a few vegetative leaves at the base. In P. elegans, the axis contains helically arranged, overlapping, stalked ovule-bearing structures (megasporophylls or bract–scale complexes). On the adaxial surface of each are several pairs of cup-shaped structures (FIG. 21.64) that are interpreted as the sites of seed attachment (Parris et al., 1995). Seeds are thought to be orthotropous, and the cuplike collar was originally thought to be an aril (Florin, 1951). Seeds of Palissya are ovate and ~2.5 mm long. Nothing is known about the pollen-producing organs. Another genus included in this family is Stachyotaxus, a Late Triassic plant known from Greenland, Switzerland, and southern Sweden. Leaves of S. elegans are morphologically similar to those of Palissya but appear to be borne in two

Figure 21.62 Palissya sp. (Jurassic). Bar  2 cm. (Courtesy S.

McLoughlin.)

ranks (Nathorst, 1908). Stomata are scattered on the abaxial surface and lack papillae on the subsidiary cells. The principal difference between the two genera is in the ovulate cones. In Stachyotaxus, the bract–scale complexes are simpler,

CHAPTER 21

CONIFERS

831

Figure 21.64 Diagrammatic reconstruction of Palissya ele-

gans. (From Parris et al., 1995.)

Figure 21.63 Palissya aptera, branch with immature cones

(Triassic–Jurassic). Bar  2 cm. (Courtesy BSPG.)

consisting of a pair of fused ovuliferous scales that terminate in a single ovule. Like Palissya, there is a cuplike structure at the site of ovule attachment in Stachyotaxus. Macerated S. elegans ovules suggest that the nucellus and integument are fused only at the base. Pollen cones recovered from the same rocks are also thought to belong to the plant that bore Stachyotaxus. Morphologically, they are small aggregations of helically arranged microsporophylls, each containing two elongated pollen sacs on the abaxial surface. Dehiscence is apparently longitudinal, and the spherical pollen grains lack sacci (Harris, 1935).

Metridiostrobus palissyaeoides is an interesting ovulebearing organ known from the Upper Triassic of North Carolina (Delevoryas and Hope, 1981). Helically arranged appendages are loosely attached to the axis, each consisting of a bract subtending an ovule-bearing unit. On either side of the ovuliferous appendage is a row of 8–10 seeds, with the micropyles directed toward the lamina margin. Ovules are small (2 mm long) and show no evidence of the cuplike point of attachment between ovule and sporophyll. The morphology of the bract–scale complexes in Palissya and Stachyotaxus is similar to the ovulate structures in the modern conifers Cephalotaxus and Dacrydium, prompting some authors to suggest affinities with these taxa. Florin (1958), however, suggested that the two genera were probably distinct from any living conifers, possibly evolving directly from the voltzialean (Utrechtiaceae) Ernestiodendron. Schweitzer (1963), however, interpreted the taxa as representing a reduction series leading to extant Cephalotaxus. Delevoryas and Hope (1981) called attention to the morphological similarity of Palissya, Stachyotaxus, and Metridiostrobus with the putative ginkgophyte Trichopitys (Chapter 18) and suggested that the ovule-bearing structures in the Palissyaceae may represent an independent origin for conifer ovulate cones. CHEIROLEPIDIACEAE

The Cheirolepidiaceae is a large family of Mesozoic conifers that most certainly represent several different types of plants

832

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.66 Cuticle of Pseudofrenelopsis parceramosa showing sunken stomata with papillae extending over stoma (Cretaceous). Bar  50 μm. (Courtesy B. J. Axsmith.)

Figure 21.65 Several Classopollis pollen grains. Note stria-

tions that make up the subequatorial rimula (arrow) (Cretaceous). Bar  20 μm. (Courtesy B. J. Axsmith.)

based on habit and ecology. Some were trees many meters tall, whereas others have been interpreted as small herbs or shrubs. As noted by Watson (1988) in her comprehensive review of the family, perhaps the only unifying character of the group is the unique type of pollen grain—Classopollis (FIG. 21.65). Several cheirolepidiaceous plants have been reconstructed either from multiple parts found in attachment or through cuticular similarities. Others consist of vegetative and reproductive structures that occur together at a single locality within a monotypic stand. One of the best-known members of this family is Pseudofrenelopsis parceramosa, originally described from the Wealden (Lower Cretaceous) of the Isle of Wight (Watson, 1977; Alvin et al., 1981). The plant is reconstructed as a large tree, based on associated logs up to 70 cm in diameter with wood of the Protopodocarpoxylon type (Alvin, 1983). Some specimens show uneven growth rings with a small amount of latewood. Tracheid pitting is of the protopinaceous and cupressoid types (Alvin, 1982). Leaves of Pseudofrenelopsis parceramosa are borne one at a node, with the base of the leaf forming a cylindrical sheath (Watson, 1977). The triangular tip of the leaf is 2 mm long, and along the margin of the leaf are hollow, unicellular hairs or delicate teeth (Watson, 1977). Based on similar cuticular anatomy (FIG. 21.66), pollen cones of Classostrobus

comptonensis associated with the leafy twigs are also reconstructed as part of the same species (Alvin et al., 1994). These are 12–14 mm in diameter and consist of a central axis with helically arranged, peltate microsporophylls. The number of pollen sacs is more than two, and pollen is of the Classopollis type. Classopollis-containing pollen cones associated with P. dalatzensis have been reported from the Lower Cretaceous of China (X.-J. Yang, 2008). Cheirolepidiaceous fossils consisting of cones attached to vegetative remains are known from the Lower Cretaceous of Arkansas, North America (Axsmith et al., 2004a). At the end of vegetative shoots of P. parceramosa are pollen cones that consist of a central axis with helically arranged, peltate microsporophylls. The cones are round or ovoid, up to 14 mm wide by 20 mm long, and contain Classopollis pollen. These cones are described as Classostrobus arkansensis (FIG. 21.67), as they differ from the cones known from the Wealden. Although the foliage is very similar, more detailed work suggests that this plant is not the same as the one reconstructed from the Isle of Wight (Axsmith, 2006). The vegetative branches are helical rather than whorled, and there are differences in leaf and wood anatomy. Ovulate cones occur at the same site (Axsmith and Creen, 2005) but have not yet been described in detail. This research suggests that P. parceramosa foliage is a morphospecies and should not be regarded as belonging to the same plant in the various habitats in which it occurs (Axsmith, 2006). Tomaxellia (FIG. 21.68) includes leafy shoots, pollen cones, and ovulate cones from the Lower Cretaceous Baqueró Formation (now Anfiteatro de Ticó Formation, Baqueró Group;

CHAPTER 21

21.67 Classostrobus arkansensis Bar  5 mm. (Courtesy B. J. Axsmith.)

Figure

CONIFERS

833

(Cretaceous).

Cladera et al., 2002) of Argentina (Archangelsky, 1963). The leaves of T. biforme are sharply tapered, helically arranged, decurrent, and up to 1.3 cm long. Stomata are amphistomatic, with the guard cells sunken and surrounded by four to five subsidiary cells. Although the morphology of Tomaxellia leaves suggests affinities with several genera, including Sequoia, Athrotaxis, and Taxodium, the fine structure of the cuticle shows significant differences from these taxa (Villar de Seoane, 1998). Pollen cones are 3 mm long and 1 mm in diameter and attached both laterally and at the tip of branches. Microsporophyll arrangement and the number of pollen sacs are not known. Pollen extracted from the cones conforms to the genus Classopollis (Srivastava, 1976b). Ovule-bearing cones associated with the foliage are 3 cm long and consist of bract–scale complexes arranged helically on an axis. Bracts are ovate, decurrent at the base, and partially surround the axis. The shorter scale is partially fused to the bract and exhibits a lobed margin. Two circular scars near the scale base on the adaxial surface are thought to represent the positions of the ovules. Isolated ovuliferous scales of Tomaxellia have also been found, suggesting that this cone, like several other cheirolepidians, shed its scales at maturity. Cupressinocladus valdensis (Protocupressinoxylon) is a Late Jurassic member of the family that was the dominant tree in southern England during the Wealden (Francis, 1983). In situ trees are associated with Cupressinocladus

Figure 21.68 Tomaxellia biforme leafy shoot (Cretaceous).

Bar  1 cm. (Courtesy B. A. R. Mohr.)

foliage and cones containing Classopollis pollen. Uniseriate, circular-bordered pits occur on the radial walls of the tracheids, and cross-field pits have slitlike apertures. No cheirolepidiaceous woods possess resin canals. Cheirolepidiaceous foliage has historically been included in a number of conifer families, most commonly, the Cupressaceae, Taxodiaceae, and Araucariaceae. The vegetative shoots in the family have helically arranged leaves (e.g., Brachyphyllum) or foliar members arranged in whorls (e.g., Frenelopsis). In Brachyphyllum (FIG. 21.69), the leaves have a basal cushion and a short, free tip. In B. mamillare,

834

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.70 Pagiophyllum diffusum (Triassic). Bar  1 cm. Figure 21.69 Brachyphyllum sp. (Cretaceous). Bar  5 mm.

a common form from the Yorkshire Jurassic, the vegetative shoots are pinnately branched and the leaves 1.5 cm long (Harris, 1979). Morphologically, they have a rounded apex and appear succulent. Amphicyclic stomata are present on all surfaces and are slightly sunken. Pagiophyllum is another foliage morphogenus with helically arranged leaves (FIG. 21.70) and stomata in rows on all surfaces except near the margin or leaf tip. In P. maculosum, the leaves are hypostomatic, and stomata are found on the branches between leaves, suggesting that perhaps the shoots of these plants remained green (Van Konijnenburg-Van Cittert, 1987). In the other basic type of cheirolepidiaceous arrangement, sometimes called the frenelopsid type, leaves are attached in whorls on segmented shoots. The most common taxon with this arrangement is Frenelopsis. In this type, the leaf bases of each segment form a smooth cylinder around the stem and lack sutures (a line marking the edge of a leaf within a whorl) (Watson, 1977). This genus is common in the

Cretaceous and has been reported from the Upper Jurassic of Spain. Frenelopsis alata is known from specimens that exhibit at least three orders of branching. Triangular leaves, each 0.6 mm long, are produced in whorls of three (Pons, 1979). Alvin and Hluštík (1979) have documented that a modified form of axillary branching is present in specimens of F. alata. This type of branching is most similar to that seen in some species of the Cupressaceae and is not known to occur in other conifers. In F. teixeirae from the Lower Cretaceous of Portugal, the leaves exhibit an opposite, decussate arrangement (Alvin and Pais, 1978), whereas in F. silfloana, the number of leaves at a node varies from two to three (Watson, 1983). Stomata are arranged in rows, and each subsidiary cell is ornamented by a papilla overarching the stomatal pit. Ovulate and pollen cones, along with wood associated with foliage of F. ramosissima, suggest that this Lower Cretaceous cheirolepidiaceous plant may have been at least 20 m tall (Axsmith and Jacobs, 2005). Pseudofrenelopsis foliage is also included in this group (Alvin, 1983). These leaves are similar to those of Frenelopsis, but only one leaf is borne at each node. Usually

CHAPTER 21

CONIFERS

835

Figure 21.72 Phylloclade of Androvettia statenensis bear-

ing short radially symmetrical lateral shoots (Cretaceous). Bar  1.5 mm. (Courtesy F. M. Hueber.)

Figure 21.71 Narrow phylloclade of Androvettia statenensis

showing several orders of branching (Cretaceous). Bar  5 mm. (From Hueber and Watson, 1988.)

the base of the leaf forms a cylindrical sheath, but in some cases, the cylinder is not complete, resulting in a longitudinal suture (Watson, 1977); internodes are typically short. The cuticle of most frenelopsids is generally quite thick (Alvin, 1982). Androvettia is a Late Cretaceous conifer with suggested affinities in the Cheirolepidiaceae (Hueber and Watson, 1988). The foliage is fernlike and consists of flattened and fused branches (FIG. 21.71) that assume the function of leaves (phylloclades); individual leaves are scalelike (FIG. 21.72) and only a few millimeters long with stomata typically sunken and surrounded by papillae. Androvettia shows some resemblance to the extant conifer Phyllocladus.

Cupressinocladus is a foliage genus established by Seward (1919) for decussately arranged leaves like those in some extant Cupressaceae. Leaves occur in decussate pairs or in alternate whorls. They are small and scalelike and extend downward into broad decurrent bases that are separated by narrow sutures. Stomata occur in well-defined rows, but the apertures are randomly oriented. Another probable foliage member of the Cheirolepidiaceae is Glenrosa (Watson and Fisher, 1984), described from the Lower Cretaceous Glen Rose Formation of Texas, USA. The leaves are small and stomata occur in clusters, with each cluster sunken in a common stomatal pit. Extending over the pit are several elongated papillae, a configuration often seen in plants from xeric environment. Daghlian and Person (1977), in their description of Frenelopsis varians from the Glen Rose Formation, noted that the geological evidence supports a coastal salt marsh environment with high evaporation rates and high salinity. Based on epidermal characters, Tarphyderma, from the Lower Cretaceous of Argentina (Archangelsky and Taylor, 1986), is another cheirolepidiaceous taxon thought to have grown in a desiccating environment. The hypostomatic cuticles of T. glabra are thick and characterized by sunken suprastomatal chambers up to 250 μm deep. Guard cells are located at the bottom of the chamber, which is formed by a ring of tubelike cells just beneath the surface of the cuticle (FIG. 21.73). The depositional environment of Tarphyderma suggests that the structure of the stomatal complexes may have been an adaptation to a habitat with frequent ash falls. There is not a great deal of information about the seed cones of the Cheirolepidiaceae. Specimens of Hirmeriella muensteri (a name that has also been applied to pollen cones) include ovuliferous scales (FIG. 21.74), each bearing what has been interpreted as a single seed partially covered

836

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.75 Walter W. Jung.

Figure 21.73 Cuticle of inner surface of Tarphyderma glabra leaf showing elongate suprastomatal chambers and guard cells (Cretaceous). Bar  50 μm.

Figure 21.74 Cheirolepis muensteri, isolated scale (Triassic–

Jurassic). Bar  5 mm. (Courtesy BSPG.)

by a flap of tissue (Jung, 1968) (FIG. 21.75). In H. kendalliae from the Jurassic of Yorkshire, the scales contain two abortive seeds on the adaxial surface (Harris, 1979). Harris suggested that the seeds were perhaps borne in some type of cutinized sac, based on the large number of cuticles recovered from macerations of the scales. Pseudohirmerella is known from isolated cone scales, but cones are thought to be 3 cm long with broadly attached scales, each showing two impressions on the adaxial surface, which are thought to represent the sites of seed attachment. Pseudohirmerella delawarensis is based on specimens from the Upper Triassic Passaic Formation in the Newark Basin of North America. The assemblage of associated remains includes ovulate cones, possible pollen cones, poorly preserved charcoalified wood, and leafy shoots assigned to Brachyphyllum–Pagiophyllum (Axsmith et al., 2004b). As the cone scales are still attached to the cone axis in three specimens, it is possible to show that the adaxial surface of the scale is concave with two depressions separated by a ridge. In addition, these specimens clearly show that the five lobes at the distal end of each scale are within a single plane. As noted earlier, several species of pollen cones have been described under the name Classostrobus (FIG. 21.67) (Alvin et al., 1978; Axsmith et al., 2004a). The cones are small, spherical, and possess a central axis with helically arranged, often peltate microsporophylls. Pollen sacs are abaxial and may number from two per microsporophyll in Hirmeriella muensteri to eight in C. pseudoexpansum.

CHAPTER 21

CONIFERS

837

Figure 21.76 Ultrathin section of Classopollis pollen grain

wall showing inner lamellate layer and outer spinate ornamentation (Cretaceous). Bar  1.5 μm. (From Taylor and Alvin, 1984.)

Pollen extracted from cheirolepidiaceous cones is included in the pollen morphotaxon Classopollis (FIG. 21.65) (Pflug, 1953). In some literature, Corollina or Circulina is still used for these grains, but the name Classopollis was officially conserved at the International Botanical Congress in 2006 (Traverse, 2004, 2007). Pollen of this type extends from the Upper Triassic to the Upper Cretaceous (Maastrichtian). Grains macerated from Classostrobus comptonensis cones are spherical, with a furrow (rimula) that encircles the grain just below the equator (Taylor and Alvin, 1984). Just above the rimula is a region where the pollen wall is thicker. On the proximal surface is a triradiate mark, and the distal pole is ornamented by a cryptopore. The mature pollen wall is divided into an inner, lamellate nexine and an outer sexine constructed of radially arranged rodlike units interwoven with transverse subunits (Rowley and Srivastava, 1986). The organization of the sexine gives this region a tectate type of organization in which there are spaces between the sporopollenin (FIG. 21.76) units of the wall (Zavialova, 2003). On the surface of the grains are numerous, closely spaced spinules (FIG. 21.76). It has been possible to determine the developmental pattern of the pollen wall in some Classopollis grains (Taylor and Alvin, 1984) (FIG. 21.77). The organization of the pollen wall and the presence of delicate channels through the wall in the region of the proximal pole suggest that the pollination biology of this group may have been more complex than that of other gymnosperms. Other pollen grains thought to be related to Classopollis but not found in situ have a different wall organization. For example, in Duplicisporites, the exine is more homogeneous and lacks a distinctive external ornament (Zavialova and Roghi, 2005).

Figure 21.77 Kenneth L. Alvin.

SUMMARY DISCUSSION: CHEIROLEPIDIACEAE

The affinities of the Cheirolepidiaceae continue to remain speculative, with earlier ideas, based principally on foliage, aligning them with the Taxodiaceae, the Cupressaceae, and the Araucariaceae. Hilton and Bateman (2006) included the Cheirolepidiaceae as a composite terminal in their phylogenetic analysis of seed plants, and the family is nested within the extant conifers. Based on the distribution of Classopollis pollen, cheirolepidians dominated certain coastal environments in the Mesozoic and were widespread in warm habitats at low paleolatitudes, especially during the Cretaceous. Alvin (1982) noted that they had particularly massive cuticles, sometimes up to 100 μm thick. In the Yorkshire Jurassic flora, members of the Cheirolepidiaceae have cuticles that range from three to ten times the thickness of other conifers in that flora. Depositional environments of macrofossil remains show that they grew not only as monodominant stands in xeric, hypersaline environments (Francis, 1983) but also in fluvial environments, where they were elements in more diverse floras (Upchurch and Doyle, 1981). The unique morphology of the pollen is an unusual component of the reproductive biology of the group that did not occur in any other conifer groups. Clement-Westerhof and Van Konijnenburg-Van Cittert

838

PaleoBOtany: the biology and evolution of fossil plants

(1991) agreed with earlier ideas postulating that the ovules were covered by an epimatium or some other fleshy structure. Alvin (1982) suggested that pollination biology in this group was more complex than in most conifers. He cited evidence of pollen on the cone scale, so perhaps a pollination droplet was involved in bringing the pollen to the micropyle. As more macrofossils are discovered, and especially now that there are some reconstructions of entire plants, the diversity of the family is becoming clearer. What is needed now is additional information about the organization and the reproductive biology of the ovulate cones, and how these similarities and differences, when placed in a phylogenetic context, can be used to more accurately characterize this important group of Mesozoic gymnosperms. PODOCARPACEAE

The Podocarpaceae includes evergreen shrubs and trees with helically arranged, sometimes opposite, leaves that range from scalelike to linear or broad. They are the second largest extant conifer family in terms of number of genera and exhibit the greatest amount of morphological diversity. All genera have sunken stomata in discontinuous rows (Stockey and Ko, 1988). Pollen cones contain two sporangia per sporophyll, and pollen grains have sacci reduced in size. Seed cones are variable in morphology but uniform in their axillary attachment to the primary axis. Stoffberg (1991) indicated that the ovule-bearing cone is a modified shoot based on its axillary position as determined from ontogenetic studies. Each cone consists of an axis with a small number of sterile bracts and one or two ovules. In many species, the bracts develop into a fleshy structure (epimatium) that surrounds the ovule. Today, the family is restricted to the Southern Hemisphere and includes 17 or 18 genera. Fossils attributed to the Podocarpaceae are known from the Lower Triassic and extend throughout the Mesozoic and Cenozoic. Possibly the oldest podocarpaceous remains are compressed leafy twigs and detached pollen and seed cones from the Triassic of South Africa, Australia, Chile, Argentina, Antarctica, and Madagascar (Hill and Brodribb, 1999; Troncoso et al., 2000). Rissikia media bears spur shoots, each ~6 cm long, that are thought to have been deciduous based on the presence of abscission scars on the axes (Townrow, 1967b). At the base of each spur shoot are several small, scalelike leaves; the remainder of the shoot contains 30 flattened, helically arranged leaves, each ~1 cm long. Stomata are arranged in four rows on the abaxial surface, with slightly sunken guard cells. Arching over the stomatal pit are cuticular projections or papillae that originate from the subsidiary cells. The pollen cones of R. media are 1 cm

long and contain 25 microsporophylls, each with two elongate, abaxially oriented pollen sacs. The sacs are thick walled and partially protected by the upturned distal end of the sporophyll. Pollen grains possess distally inclined sacci. The grains are 50 μm long and contain widely spaced endoreticulations on the inner surface of the saccus wall. On the distal surface is a slightly thinner region or leptoma; proximally the cappa exhibits numerous parallel striations. Similar sporae dispersae grains are included in the genus Taeniasporites. Seed cones range up to 6 cm long, each with a variable number of bract–scale complexes identified by a trifid bract that subtends the cone scale. Extending from the lobe of each cone scale are two elongated ovules. Stalagma is an interesting conifer from the Upper Triassic of China (Z. Zhou, 1983b). The fertile shoot bears scale leaves, each 4 mm long, and an ovulate cone at the end of the shoot; cone scales are loosely arranged and decurrent at the base. On the adaxial surface of the scale is a single inverted, platyspermic seed. Arising from the adaxial surface of the cone scale is a flap of tissue interpreted as an epimatium that partially covers the seed. If this is a podocarp, then it is especially interesting since it is reported to contain monosulcate pollen (50–60 μm long) in the seed micropyles and also in fragments of closely associated cones. Associated with the cones are simple, elongated leaves ( 15 cm long) with parallel veins, assignable to the genus Desmiophyllum (FIG. 21.78). Nothodacrium is a podocarp from the Jurassic of Antarctica (Townrow, 1967c). In this woody conifer, there is no division of the vegetative parts into long and short shoots. Superficially, the axes appear pinnately arranged, although they were probably produced in a helical pattern. The leaves are unusual in that they diverge in several directions from the axis. Each is 3 mm long and rhomboidal in transverse section. Ovule-bearing cones occur at the ends of branches, each cone consisting of 10–15 bract–scale complexes. The ovuliferous scale is trilobed and bears a single inverted seed. Pollen cones assignable to the morphogenus Masculostrobus are known from the same rocks but have been found attached to leaf-bearing branches. Pollen grains are trisaccate and 110 μm in diameter. Similar Jurassic pollen is included in the sporae dispersae taxon Tsugaepollenites. Structurally preserved pollen cones from the Jurassic of India that were formerly assigned to Masculostrobus are now included in Podostrobus (Rao and Bose, 1971). Pollen of this podocarpaceous cone is reported as being both biand trisaccate. Another woody podocarp known from the Jurassic of Australia and New Zealand is Mataia. The leaves of this plant are helically arranged, but twisted at the base to form two rows (FIG. 21.79). Each is 1.5 cm long with rows

CHAPTER 21

of monocyclic stomata on the lower surface. Seed cones of M. podocarpoides are 3 cm long and contain basal foliage leaves that subtend a zone of 8–12 helically arranged, bract– scale complexes (Townrow, 1967b). The bract is triangular in

Figure 21.78 Branch fragment with two opposite Desmiophyllum

type leaves (Triassic). Bar  1 cm. (Courtesy B. Meyer-Berthaud.)

CONIFERS

839

outline, whereas the tip of the ovuliferous scale is recurved and produces two stalked ovules from the adaxial surface. Podocarpaceous pollen cones, seed cones, and leaves from the Lower Cretaceous of Argentina are included in Trisacocladus (FIG. 21.80). The helically arranged leaves

Figure 21.80 Trisacocladus tigrensis pollen (Cretaceous). Bar  5 mm. (Courtesy S. Archangelsky.)

Figure 21.79 Portion of a leafy shoot of Mataia podocarpoides (left) and detail of leaf attachment. (From Taylor and Taylor, 1993.)

cone

840

PaleoBOtany: the biology and evolution of fossil plants

are 6 mm long and leave a rhomboidal scar when detached from the twigs (Archangelsky, 1966). The largest ovule-bearing cone (2 cm long) consists of a fleshy axis bearing ovules; none of the specimens have bracts or scales. Ovules of T. tigrensis are orthotropous and 3 mm long. Pollen cones produce helically arranged microsporophylls with probably two pollen sacs on each abaxial surface. Pollen grains are small (25–35 μm in diameter) with three sacci extending from the distal hemisphere of the grain. Ultrastructural features of the fossil grains are similar to those of the extant genera Dacrydium, Podocarpus, and Dacrycarpus (Baldoni and Taylor, 1982). Another Early Cretaceous member of the family from the Baqueró Formation, Argentina, is Squamastrobus (Archangelsky and Del Fueyo, 1989). Both pollen and ovule-bearing cones are associated with foliage of the Brachyphyllum type. The ovulate cones are terminal and bear helically arranged bract–scale complexes, each with a single seed (FIG. 21.81). Pollen cones are simple, small (1.8 cm long), and contain bisaccate pollen of the Podocarpidites type. Pollen cones associated with Elatocladus foliage from the Baqueró Formation are described as Morenoa (Del Fueyo et al., 1990). Affinities with the podocarps are based on the organization of the cones and the pollen ultrastructure. Two foliage morphogenera from the Lower Cretaceous of Brazil are Podozamites and Lindleycladus (Kunzmann et al., 2004). In Lindleycladus, the leaves may be several centimeters long and helically arranged (FIG. 21.82). On the surface are longitudinally oriented stomata arranged in narrow bands, whereas in Podozamites, the stomata are transversely oriented. Notophytum krauselii includes permineralized woody stems and leaves from the Middle Triassic of East Antarctica (MeyerBerthaud and Taylor, 1991). The eustelic stems range from a few millimeters to more than 20 cm in diameter and produced helical branches at wide angles. The secondary xylem has growth rings (FIG. 21.83) of variable thickness and uniseriate rays; tracheids are polygonal in transverse section. The multiveined leaves are elongated, apetiolate, and 3 cm in diameter. Roots contain a phi layer (FIG. 21.84), suggesting that the plants grew in an environment with fluctuating water levels (Millay et al., 1987). The presence of transfusion tissue in two patches adjacent to the vascular bundles, resin canals below the vascular bundles, and sclereids in the mesophyll are features that are used to compare Notophytum with members of the Podocarpaceae, especially the extant genus Nageia (Axsmith et al., 1998b). Permineralized wood from an in situ Middle Triassic forest in Antarctica was described as Jeffersonioxylon (Del Fueyo et al., 1995). Although it shows some features of podocarpaceous wood, this genus is believed to belong

Figure 21.81 Pollen cone of Squamastrobus tigrensis (Cretaceous). Bar  1.5 mm. (From Archangelsky and Del Fueyo, 1989.)

to the corystospermalean seed ferns (Chapter 15), based on Dicroidium leaf litter surrounding the silicified stumps. Permineralized leaves of Notophytum are similar to those of the impression–compression form Heidiphyllum (FIG. 21.85), a common foliage genus in the Triassic of Gondwana. These leaves are apetiolate, slightly obovate, and have 8–12 parallel veins (FIG. 21.86). Each pair of guard cells

CHAPTER 21

CONIFERS

841

Figure 21.84 Cross section of root of Notophytum krauselii

showing phi layer (arrows) (Triassic). Bar  50 μm.

Figure 21.82 Lindleycladus sp. (Cretaceous). Bar  5 cm. (Courtesy B. A. R. Mohr.)

Figure 21.83 Cross section of Notophytum krauselii showing

growth rings (Triassic). Bar  5 mm.

is surrounded by five subsidiary cells, and the abaxial surface of the epidermis is papillate. The morphology and epidermal features of N. krauselii and compressed H. elongatum (Late Triassic of Antarctica) leaves are very similar (Axsmith et al.,

1998b). Telemachus elongatus (FIG. 21.87) is thought to represent the cone of H. elongatum as both taxa co-occur at multiple localities in South Africa (Anderson, 1978). Telemachus also occurs in the Triassic of Antarctica (X. Yao et al., 1993) along with Heidiphyllum and Notophytum (Axsmith et al., 1998b). Telemachus elongatus cone scales are five lobed, with each scale producing two to three recurved ovules. An interesting find of podocarpaceous wood is an in situ Cretaceous forest reported from 140 m below sea level 32 km off the coast of South Africa (Bamford and Stevenson, 2002). None of the specimens recovered show evidence of growth rings. Miocene representatives of the family occur in Australia in the form of pollen cones and foliage assignable to the extant genus Dacrycarpus (R. Hill and Whang, 2000). Microsporophylls are helically arranged and pollen grains, each with three sacci, are believed to have been produced by these cones. Sigmaphyllum and Falcatifolium are two foliage genera from the Paleogene of Australia that are closely related to modern podocarps (R. Hill and Scriven, 1999). The use of micromorphological cuticular analysis of Falcatifolium was especially useful in demonstrating that during the Eocene this plant lived at very high paleolatitudes, whereas today it is restricted to lower latitudes. Prumnopitys is one of several Paleocene podocarps preserved in carbonaceous mudstones from New Zealand (Pole, 1998). Specimens include shoots with helically arranged, distichously flattened leaves. An analysis of the type and distribution of stomata were used to suggest the affinities of the specimens. The Northern Hemisphere record of Mesozoic podocarps is relatively meager, including only a few pollen grains and some megafossils from the Eocene of Tennessee (Dilcher, 1969). These specimens consist of stout stems with helically arranged, needle-shaped, amphistomatic leaves that show thickened rings of cutin surrounding the stomata. These

842

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.86 Detail of Heidiphyllum sp. venation (Triassic).

Bar  1 cm. Figure 21.85 Leaf of Heidiphyllum sp. (Triassic). Bar  2 cm.

cuticular features clearly indicate the affinities of the leafy shoots with the Podocarpaceae. There are a number of generic names applied to isolated pieces of petrified or permineralized wood interpreted

as being podocarpaceous. These include Podocarpoxylon, Paraphyllocladoxylon, Phyllocladoxylon, and Metapodocarpoxylon (Bamford et al., 2002). The reader is referred to the authoritative account of the fossil Podocarpaceae, Cupressaceae, and Araucariaceae from the Southern Hemisphere by Hill and Brodribb (1999).

CHAPTER 21

Figure

21.88 Agathis

sp.

CONIFERS

leaves

(Extant).

843

(Courtesy

K. Nixon.)

Figure 21.87 Seed cone of Telemachus elongatus (Triassic).

Bar  2 cm.

The present restriction of the family to the Southern Hemisphere, Central America, and Mexico has suggested that evolution within the family was linked to the breakup of Gondwana (Woltz, 1986). Although the podocarps were once believed to have been restricted to Gondwana throughout their geologic history, megafossil evidence, together with scattered Northern Hemisphere reports of pollen, suggest that some of the members of this group may have been more cosmopolitan in their distribution.

SUMMARY DISCUSSION: PODOCARPACEAE

ARAUCARIACEAE

Most extant podocarps have highly reduced ovulate reproductive organs. This has suggested to some that the possible ancestors of the group should be sought in families in which the reproductive organs are loosely arranged, possess simple seed–scale complexes, or have cauline-borne orthotropous ovules subtended by a bract. Both Ferugliocladaceae and Buriadiaceae, two families that are confined to the Southern Hemisphere, share these features with the podocarps. Others have suggested a relationship with the Paleozoic conifer families that have well-defined seed–scale complexes. The seed cones of Rissikia and Mataia suggest that the cone scales of some extant podocarps may be morphologically equivalent to one of the scale lobes in these fossil members. According to this idea, the epimatium represents a single, bivascularized unit rather than two structures that have fused together. In Stalagma, there is some evidence of an epimatium occurring as early as the Late Triassic. Since it is difficult to identify such a feature in most fossil bract–scale complexes, it is not known whether the epimatium was present in all podocarpaceous taxa or has persisted only in the remaining modern lineages.

The Araucariaceae is a family of ancient conifers that today includes three genera, Araucaria, Agathis (FIG. 21.88), and Wollemia (FIG. 21.89), with approximately 40 species. All genera are monophyletic with Wollemia interpreted as the most primitive living genus (Setoguchi et al., 1998). All extant species are evergreen and lack shoot dimorphism. Leaves may be helically arranged or opposite, linear or broad, and cones are typically large. The number of pollen sacs within the family varies from 5 to 20, with pollen grains typically nonsaccate. Seed cones take 2 years to mature, and the ovuliferous scale is highly reduced and fused to the bract, usually with one ovule produced per complex. The three genera are confined today to South America and Australasia, but, like the Podocarpaceae, they were present to a limited extent in the Northern Hemisphere during the Mesozoic. There are a few reports of araucarian fossils in the Paleozoic and some specimens from the Upper Triassic, but the greatest diversity and widest distribution of araucarians apparently occurred during the Jurassic. Fossil evidence suggests that from the Cretaceous to the present, the family has gradually declined in numbers of taxa and geographic distribution (Stockey, 1982; Kunzmann, 2007a, b).

844

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.90 Agathoxylon (Araucarioxylon) sp., segment of a permineralized stem (Jurassic). Bar  5 cm. (Courtesy BSPG.)

Figure

21.89 Wollemia

nobilis

(Extant).

(Courtesy

M. A. Gandolfo.)

All living araucarians are trees, some that exceed 60 m in height, with the leaf-bearing branches forming a dense crown. Because of the enormous size of some of these trees and the large amount of wood produced, it is no wonder that one of the most common fossils assignable to the family is petrified or permineralized wood. This wood was referred to as Araucarioxylon for many years, but many different, unrelated taxa were placed in the morphogenus over the years. Most of the araucarian woods should now be placed in Agathoxylon (FIG. 21.90) (Bamford and Philippe, 2001). The nomenclatural history of Araucarioxylon is complex, and the reader is referred to Vogellehner (1964), Philippe (1993), and Philippe and Bamford (2008) for further information. Two of the best-known sources for araucarian wood are the Petrified Forest in Arizona, USA (Late Triassic, Chinle Formation)

and the Cerro Cuadrado Petrified Forest in Patagonia, Argentina (Middle Jurassic, La Matilde Formation), where enormous silicified logs are found. Some of the wood from Argentina has definitely been assigned to Agathoxylon (Zamuner and Falaschi, 2005). Agathoxylon (Araucarioxylon) includes secondary xylem that contains both uniseriate- and multiseriate-bordered pits on the radial walls of tracheids. There is no xylem parenchyma nor have resin canals been described in wood of this type. Vascular rays are uniseriate. In some instances, wood with these characteristics is difficult to distinguish from specimens of Dadoxylon (Mussa, 1986) (FIG. 21.91), a morphogenus that no doubt contains specimens of cordaitean secondary xylem (Meyer-Berthaud and Taylor, 1992). Philippe et al. (2004) provided a detailed summary of the stratigraphic and geographic distribution of araucarian woods in Gondwana. Leafy twigs of the morphogenus Brachyphyllum (FIG. 21.92) are sometimes found associated with reproductive organs believed to be araucarian. One of these

CHAPTER 21

Figure 21.91

Diana Mussa.

Figure 21.92 Brachyphyllum hondurense (Jurassic). Bar 

2 cm.

is B. mamillare from the Jurassic of Yorkshire in which cuticle structure demonstrates affinities with ovulate cone scales of Araucarites phillipsii (Harris, 1979). Specimens of B. patens from the Upper Cretaceous of Europe, however,

CONIFERS

845

show stomata, suggesting that it may have affinities with the Cheirolepidiaceae (Van der Ham et al., 2003), underscoring the fact that Brachyphyllum is a morphogenus and has been found with several different types of conifer remains. Permineralized specimens of B. vulgare from Japan provide important details about the histology of the stem and fleshy leaves, but these specimens offer nothing conclusive about the systematic affinities of the genus (Ohana and Kimura, 1993). Araucarioides is a foliage type described from the Eocene of Tasmania that was regarded as an intermediate between Agathis and Araucaria based on epidermal characters (Bigwood and Hill, 1985). Since the discovery of the living Wollemia in Australia in 1994, however, others have suggested that the genus may represent fossil specimens of that taxon (Chambers et al., 1998). The validity of distinguishing foliar features in fossil araucarian leaves rests on wellreferenced modern analogs (Stockey and Taylor, 1978a). These have been assembled for Agathis and Araucaria (Stockey and Ko, 1986). Fossil specimens of Araucaria grandifolia are known from the early Albian of Patagonia, Argentina (Del Fueyo and Archangelsky, 2002). Woody shoots have large multiveined leaves that have an acute apex and decurrent base. Leaves are amphistomatic, closely spaced on the stem, and helically arranged. The genus Araucarites was initially used for cones, isolated ovuliferous scales, and sterile twigs thought to have araucarian affinities (Philippe, 1993), but today it is used only for cones and cone parts. Araucarites cutchensis from the Upper Jurassic–Lower Cretaceous of India is based on detached, wedge-shaped seed scales up to 3.5 cm long (Bose and Maheshwari, 1973). Partially embedded in the upper scale surface is a single ovoid seed. In the cone A. rudicula from the Petrified Forest of Arizona, there is a single small seed located toward the distal end of the bract–scale complex (Axsmith and Ash, 2006). A characteristic feature of the extant araucarian bract–scale complex is the ligule, which is an extension of the ovuliferous scale that is not fused to the bract. Many of the presumed seed scales of Araucarites do not exhibit a ligule. Doliostrobus (FIG. 21.93) is known from vegetative and reproductive remains from the Cretaceous and Cenozoic that morphologically resembles those of Agathis, although the Cenozoic forms are sometimes placed in their own family (K. Kvacˇek, 2002a). Foliage remains and dispersed cuticle attributed to Agathis have been described from the Eocene of Australia (Scriven, 1993; Pole, 1995). Some of the most spectacular and beautifully preserved araucarian fossils are silicified seed cones found in the Cerro Cuadrado Petrified Forest of Patagonia, Argentina. The forest is within the La Matilde Formation and is considered to be

846

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.94 Longitudinal section of silicified Araucaria mirabilis cone showing seeds (Jurassic). Bar  1.5 cm. (Courtesy BSPG.)

Figure 21.93 Doliostrobus taxiformis, twig with needle leaves (Oligocene). Bar  2 cm. (Courtesy BSPG.)

Middle Jurassic. Cones of Araucaria mirabilis (FIG. 21.94) are completely silicified by α-quartz and range from spherical to ellipsoidal in shape (Stockey, 1978). The largest specimens are almost 10 cm in diameter and consist of a central axis with helically arranged bract–scale complexes. Each unit is constructed of an ovuliferous scale subtended by a woody, winged bract. The axis of the cone contains a parenchymatous pith surrounded by a ring of fused bundles that become separate at higher levels. Each bract–scale complex is vascularized

by a double set of bundles that provide traces to the bract and ovuliferous scale. Resin canals are associated with each vascular strand. Anatomic features of the Patagonian cones suggest affinities with modern Araucaria bidwillii (FIG. 21.95), which is known only from Queensland, Australia, today. Each ovuliferous scale bears one seed partially embedded in the upper surface. Mature seeds of A. mirabilis are 1.3 cm long and about half as wide. The integument of the seed is composed of three distinct tissue systems: parenchymatous sarcotesta, fibrous sclerotesta, and partially preserved endotesta. The seed coat and nucellus are adnate only at the base. Megagametophyte tissue is common, and in a few specimens, elliptical cavities suggestive of archegonia have been identified. An interesting feature of the fossil araucarian seeds is the occurrence of dicotyledonous embryos within the megagametophyte tissue (Stockey, 1975). They are typically in the telostage period of development and, in longitudinal section, clearly exhibit shoot apex, cotyledons, root meristem, vascular system, columella, and calyptroperiblem (FIG. 21.96). Cellular preservation within these embryos is so good that nuclei appear to be present in some of the cells. The remnants of a tightly coiled suspensor are also visible in some of the sections. Rarely does the fossil record provide any easily identifiable evidence of postfertilization stages or embryo and seedling

CHAPTER 21

21.95 Araucaria bidwillii Bar  5 cm. (Courtesy R. Stockey.)

Figure

seed

cone

CONIFERS

847

(Extant).

Figure 21.97 Leafy twig of an araucarian with terminally posi-

tioned lignotuber (Jurassic). Bar  4.5 cm. (Courtesy BSPG.)

Figure 21.96 Longitudinal section of Araucaria mirabilis

seed containing embryo. Arrow indicates shoot apex (Jurassic). Bar  0.5 mm. (Courtesy R. Stockey.)

development. Several specimens that have been compared with extant araucarian seedlings have been found among the Cerro Cuadrado fossil remains. They are top-shaped to turbinate, with the largest ~3.5 cm long. Some of these fossil specimens, such as the seedlings of living araucarians, have a swollen hypocotyl composed of a parenchymatous cortex with patches of resin canals (Stockey and Taylor, 1978b). In

the central portion of the axis is a pith surrounded by secondary xylem. Additional research has demonstrated, however, that these unusual fossils represent three different categories of structures. Some are true woody seedlings, as described by Stockey and Taylor (1978b), whereas others represent decorticated shoots. The turbinate structures are now believed to be aerial lignotubers (FIG. 21.97) (Stockey, 2002) that may have formed in the axils of leaves. Despite the numerous ovulate cones, twigs, and wood discovered in the Cerro Cuadrado deposits, no pollen cones have been discovered to date. Similarly, no pollen grains or evidence of pollen tubes have been found within the seeds. This is especially unusual because pollen tubes in the Araucariaceae are quite erosive and typically leave conspicuous channels in the seed tissues. The absence of pollen in the deposits is believed by some to indicate that the forest was covered by volcanic ash before

848

PaleoBOtany: the biology and evolution of fossil plants

the ontogenetic development of the pollen cones; the presence of embryos suggests that the cones may have developed parthenogenetically. Interestingly, pollen cones within the Araucariaceae are quite rare in the fossil record, but one has been described from the Aptian of Patagonia as Alkastrobus peltatus (Del Fueyo and Archangelsky, 2005). Pollen sacs in this species are attached to the abaxial surface of a peltate microsporophyll much like that in a number of voltzialean pollen cones, and pollen is of the Cyclusphaera type. These grains are different from Balmeiopsis and Araucariacites, two dispersed pollen types known to have been produced by araucarians. Leafy twigs of Brachyphyllum occur in the same deposits as A. peltatus. Wairarapaia is a seed cone described from the Cretaceous of New Zealand that is thought to be closely related to cones of extant Wollemia (Cantrill and Raine, 2006). Specimens of W. mildenhallii are 3 cm in diameter and have wedgeshaped cone–scale complexes with the micropyle of the ovule directed toward the cone axis. Seeds are attached to a narrow pad of tissue similar to that in extant Agathis and Wollemia. Some seeds possess dicotyledonous embryos. Specimens of the extant W. nobilis can be compared with published examples of araucarian cone scales, pollen cones, and juvenile and adult foliage from the Cretaceous of Australia (Chambers et al., 1998). Although perhaps not the same species, the Cretaceous fossils appear more comparable to Wollemia than Araucaria. Further supporting the hypothesis that Wollemia (FIGS. 21.98–21.100) was present during the Cretaceous are the similarities between the dispersed fossil pollen type Dilwynites and that of W. nobilis (Chambers et al., 1998). Araucaria vulgaris, from the Upper Cretaceous of Hokkaido, consists of permineralized seed cones of A. nihongii attached to leafy shoots of Yezonia vulgaris, a genus for permineralized foliage that is probably equivalent to Brachyphyllum (Ohsawa et al., 1995). These cones bear a single seed per cone–scale complex and represent the first evidence from the Northern Hemisphere of the section Eutacta of the genus Araucaria.

Figure 21.98 Mature leaves of araucarian that are comparable to extant Wollemia (Cretaceous). Bar  5 mm. (Courtesy S. McLoughlin.)

SUMMARY DISCUSSION: ARAUCARIACEAE

The fossil record of the Araucariaceae (FIG. 21.101) suggests that the oldest and geographically most widespread members of the family belong to Araucaria and that at least some species in the family occurred in the Northern Hemisphere (R. Hill and Brodribb, 1999). Because of the relatively recent discovery of Wollemia, it will be interesting to see if additional araucarian fossils are related to this genus and if so to what extent the geographic distribution of the genus is extended in

Figure 21.99 Juvenile foliage of araucarian that is compa-

rable to extant Wollemia (Cretaceous). Bar  1 cm. (Courtesy S. McLoughlin.)

CHAPTER 21

CONIFERS

849

Figure 21.101 Compressed araucarian cone (Cretaceous).

Bar  2 cm. (Courtesy B. A. R. Mohr.) Figure 21.100 Araucarian pollen cone that is compara-

ble to extant Wollemia (Cretaceous). Bar  2 cm. (Courtesy S. McLoughlin.)

the geologic record. One interesting anatomical character in members of the Araucariaceae is the presence of undifferentiated exogenous axillary meristems that are buried beneath the periderm in the leaf axils (Burrows, 1999). This situation is unlike that in some conifers that lack buds in the leaf axils and may provide a useful character in distinguishing permineralized fossil araucarian stems. Extant species of Araucaria are divided into four sections Columbea, Bunya, Eutacta, and

Intermedia. Del Fueyo and Archangelsky (2002) suggested that Patagonia could have been a center of Mesozoic distribution for the genus, as three of the four sections are represented there, one in the Jurassic (Bunya) and two by the Cretaceous (Colmbea and Eutacta). The presence of a single ovule per ovuliferous scale in the Araucariaceae has been used to suggest affinities with certain voltzialeans. Like the other modern conifer families, however, a consensus on the origin and phylogenetic position of the Araucariaceae is not yet available. CUPRESSACEAE

The Cupressaceae constitute the largest and most widely distributed of the conifer families in terms of numbers of

850

PaleoBOtany: the biology and evolution of fossil plants

extant genera (30), with many cultivated today as ornamentals. Among the living members are found the largest single organisms on earth, the giant sequoias, Sequoiadendon giganteum. We will use the more recent classifications that include taxa previously in the family Taxodiaceae (Eckenwalder, 1976; Farjon, 2005). This classification, with the exception of Sciadopitys, is well resolved by certain morphological features (Farjon and Ortiz Garcia, 2003) and molecular phylogenetics (Gadek et al., 2000). All species are evergreen and characterized by small, scalelike linear leaves that are borne oppositely or in whorls. Pollen cones contain whorled microsporophylls, each with three to six pollen sacs. Ovulate cones are typically small, with the bract generally larger than the cone scale to which it is fused. Ovule number is typically two to three per cone scale, but in some species of Cupressus, the number may vary from 6 to 20. Pollen is spheroidal, monoporate, and nontectate. At one time, the family was thought to have occurred in the early Mesozoic based on certain fossils, for example, Cupressinocladus, Cupressinostrobus, and Paleocyparis, but today many of those fossils are regarded as members of the Cheirolepidiaceae (K. Kvacˇek et al., 2000). Two vegetative characters that help distinguish cupressaceous and cheirolepidiaceous taxa are the presence of papillae arising from epidermal cells and cuticular thickenings, termed Florin rings, around the stomatal pore in cupressoid taxa. The reader is referred to the authoritative accounts of fossil members of the family found in R. Hill and Brodribb (1999) and Stockey et al. (2005). We have followed the classification of Stockey et al. (2005) in using subfamilies to organize the discussion of the fossil record of the Cupressaceae. CUNNINGHAMIOIDEAE Cunninghamiostrobus is an ovulate cone from the Cretaceous associated with needlelike leaves, which closely resemble extant specimens of Cunninghamia (Miller, 1975). The leaves are helically arranged (FIG. 21.102) and contain conspicuous resin canals that are arranged in a ring around the vascular tissue. Leaves contain a palisade mesophyll surrounded by a fibrous hypodermis. Ovule-bearing cones of C. hueberi are attached both laterally and terminally to structurally preserved stems. They are ellipsoidal and consist of a central axis 2.5 cm long with helically arranged bract–scale complexes. The presence of a partial growth ring in the cone axis suggests that cone development required more than a single growing season to complete, much like some modern species of the family. In C. hueberi, the bract is conspicuous, with the ovuliferous scale reduced to a pad of tissue about one-third the size of the bract. On the upper surface of each scale are three flattened seeds (FIG. 21.103).

Figure 21.102 Cunninghamia marquettii leafy axis (Oligocene). Bar  1 cm. (From Axelrod, 1998a.)

In C. goedertii from the Oligocene, the bract is conspicuous, whereas the ovuliferous scale is much reduced (Miller and Crabtree, 1989). Permineralized specimens from Japan show distinct differences in the vascular architecture of the cone scales compared to modern Cunninghamia (Ohana and Kimura, 1995). The fossil cones possess a combination of

CHAPTER 21

CONIFERS

851

21.103 Reconstruction of ovuliferous scale of Cunninghamiostrobus yubariensis. (From Ohana and Kimura, 1995.)

Figure

features found in cones of modern Athrotaxis, Taiwania, and Cunninghamia. It is suggested that these cones represent a group of conifers that were geographically widespread during the Late Cretaceous to the Oligocene but which are now extinct (Miller and Crabtree, 1989). Elatides is a taxon used principally for ovulate cones, although some authors have extended the taxonomic designation to also include vegetative remains and pollen cones. Specimens are known from a number of geographic localities ranging from the Middle Jurassic to the Cretaceous (MacLeod and Hills, 1991). In E. bommeri from the Lower Cretaceous (Wealden) of Belgium, the vegetative axes are irregularly branched, with 3 mm long leaves borne in a tight helix (Harris, 1953). In section view, the leaves are rhomboidal and characterized by a single, large resin canal within the fibrous hypodermis. Stomata are arranged in rows with the individual guard cells slightly sunken; papillae are absent on the subsidiary cells. Ovule-bearing cones are borne terminally on vegetative branches and have stalked cone scales that terminate in a conspicuous distal spine. The seeds are 3 mm long and lack distinct wings. Elatides williamsonii is a pollen cone from the Jurassic of Greenland in which each microsporophyll contains three pollen sacs on the abaxial surface (Harris, 1943); pollen is oval with a distal aperture. The wood associated with these fossils is included in Cupressinoxylon (Bamford et al., 2002). A cone from the Upper Cretaceous Magothy Formation of New Jersey, USA, Rhombostrobus cliffwoodensis, resembles modern cones of Cunninghamia in the distribution of resin canals and vasculature of the bract– scale complex (LaPasha and Miller, 1981).

Figure 21.104 Athrotaxites lycopodioides, (Jurassic). Bar  1 cm. (Courtesy BSPG.)

leafy

twig

TAIWANIOIDEAE This subfamily is very rare in the fossil record. Two genera of permineralized cones from the Cretaceous of Japan are considered to belong to the Taiwanioideae, Parataiwania (Nishida et al., 1992) and Mikasastrobus (Saiki and Kimura, 1993). In P. nihongii, there are four-winged seeds per cone scale; in M. hokkaidoensis, the vascular supply to the ovuliferous scale is like that in extant Taiwania. ATHROTAXOIDEAE One of the better known fossil members of the Cupressaceae is Athrotaxites (FIG. 21.104) from the Cretaceous (Aptian) of Montana and coeval deposits in Canada (Miller and LaPasha, 1983). The species is based on impressions and compressions,

852

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.105 Metasequoia sp. (Paleocene). Bar  1 cm.

and the plant has been reconstructed with small, scalelike leaves, each ~2 mm long. Stomata are confined to the abaxial surface and sunken. Pollen cones are simple and contain helically arranged microsporophylls, each with two abaxial pollen sacs. Seed cones are borne in a terminal position on a lateral branch. The bract and ovuliferous scale are fused, with each probably producing a single seed per scale. Some species of Athrotaxis, such as a number of fossil cupressoids, share several features with modern taxa, suggesting that extensive radiation was taking place within the family during the Early Cretaceous (Miller, 1988). Other species of Athrotaxis show morphological variation that is not evident in modern forms (R. Hill et al., 1993). SEQUOIOIDEAE Metasequoia (FIG. 21.105), the dawn redwood, is a member of this subfamily and is often described as a living fossil. It was initially described from fossil remains, only later found to be living in central China (H.-L. Li, 1964). The history of this genus provides an interesting account of how information about living plants can be used in interpreting fossils. The genus Metasequoia (FIG. 21.106) was described by Miki (1941) (FIG. 21.107) from vegetative remains and cones collected in Pliocene clays and lignite beds. The distinguishing feature of M. glyptostroboides is the deciduous leafy shoots borne in opposite pairs along the branches. Needles are twisted and opposite; stomata are arranged in parallel rows on either side of the leaf midrib. Leaves are petiolate and diverge at an acute angle from the axis. Possibly, the best taxonomic characteristic of the genus is the decussate arrangement of the pollen and seed-cone scales. In Sequoia and Taxodium,

Figure 21.106 Metasequoia occidentalis leafy twig (Eocene).

Bar  2 cm. (Courtesy BSPG.)

the two genera most often confused with Metasequoia, the cone scales are arranged helically (Chaney, 1951). Some years earlier, Endo (1928) had realized that fossil cones described from Oligocene rocks as Sequoia langsdorfii were improperly designated because the cones possessed decussate

CHAPTER 21

CONIFERS

853

Figure 21.109 James E. Canright. (Courtesy T. Delevoryas.)

Figure 21.107 Shigeru Miki. (Courtesy T. Delevoryas.)

Figure 21.108 Metasequoia sp. cones (Eocene). Bar  1 cm. (Courtesy J. F. Basinger.)

Figure 21.110 Ralph Works Chaney. (Courtesy H. N. Andrews.)

cone scales. He also noted the blunt apices of the needles and the short peduncle of the cones, the two characteristics that are used today to distinguish Metasequoia. In 1943, T. Wang, a forester, collected some specimens of a tree in central China that he could not identify. When these specimens were analyzed and compared with the published descriptions of the fossil Metasequoia (FIG. 21.108), it became obvious that the living plant was the same genus.

Many of the fossil specimens originally described as species of Sequoia are, in fact, specimens of Metasequoia, and the genus was rather widespread in North America and Asia during the Cretaceous and the Cenozoic (Canright, 1972) (FIG. 21.109). One final note on this interesting plant is worth mentioning. Beginning in about 1948, seeds of Metasequoia were obtained from Asia through the efforts of E. D. Merrill and Ralph W. Chaney (FIG. 21.110) and were

854

PaleoBOtany: the biology and evolution of fossil plants

like those of extant Sequoiadendron but lack the tricyclic stomata characteristic of the genus. The seed cones also lack a transverse median groove, but all other features are comparable to those of Sequoia and Sequoiadendron. Austrosequoia is an anatomically preserved cone that shares features with Sequoia (Peters and Christophel, 1978). Some information is known about the anatomy of taxodiaceous seed cones from the Upper Cretaceous (LaPasha and Miller, 1981). In Nephrostrobus, the axis contains a ring of resin canals in the cortex. The ovuliferous bract–scale complex in this fossil is morphologically similar to that in several extant genera but differs significantly from M. glyptostroboides.

Figure 21.111 Metasequoia glyptostroboides branch with

cones (Extant). (Courtesy K. Nixon.)

widely planted throughout North America and Europe. As a result, today the “living fossil” may be appreciated by all. Perhaps the best-known foliage type is M. occidentalis (FIG. 21.106), a morphogenus based on impression–compression specimens, which is morphologically similar to M. glyptostroboides (FIG. 21.111). Many of the fossil species of Metasequoia have been placed in synonymy (Y. Liu et al., 1999). Metasequoia foxii is a Paleocene plant from Canada based on more than 10,000 compression specimens of vegetative and reproductive structures (Stockey et al., 2001b). Leaves are opposite/decussate and up to 4 cm long. Permineralized wood from in situ stems shows growth rings, and tracheids with circular-bordered pits are interspersed with uniseriate rays. Pollen cones occur in pairs with microsporophylls producing three pollen sacs on the abaxial surface. The number of seeds per scale is not known, but isolated seeds and seedlings are common on the bedding plane. Seedlings possess two cotyledons. Metasequoia milleri from the Eocene of British Columbia, Canada, was initially used for pollen cones, but the diagnosis has been expanded to accommodate large trees that produced seed cones, foliage, and even roots with endomycorrhizae (Rothwell and Basinger, 1979; Basinger, 1981, 1984). The cones are small and constructed of 30 helically to irregularly arranged microsporophylls, each bearing three pollen sacs on the abaxial surface. Pollen grains are subspheroidal and characterized by a distinctive, erect papilla. In many features, the Eocene cones appear remarkably similar to pollen cones of modern M. glyptostroboides, differing only in minor histologic characteristics and pollen morphology. Leafy twigs with attached seed cones from the Upper Cretaceous of Sweden have been described as Quasisequoia (Srinivasan and Friis, 1989). Scale leaves are morphologically

TAXODIOIDEAE Silicified remains of Taxodium wallisii come from the Late Cretaceous of Alberta, Canada (Aulenback and LePage, 1998). The three-dimensionally preserved branches bear dimorphic leaves that are both taxodioid, that is, triangular in cross section with a decurrent base and an abaxial keel, and cupressoid, that is, with a sharply recurved, acuminate apex. Both seed and pollen cones are attached to leafy shoots (FIG. 21.112). The ovuliferous scale in T. wallisii is conspicuously lobed and bears two seeds. Pollen cones are small and arranged in terminal panicles (FIG. 21.113); they have peltate microsporophylls with five to nine pendulous pollen sacs arranged in two rows. Pollen is small (20 μm) with a papilla that is 2 μm long. When all characters are analyzed, these fossils not only appear closely related to extant Taxodium but also show similarities with Glyptostrobus (FIG. 21.114) and Cryptomeria. Taxodium balticum is based on Paleogene compression specimens from Russia (Vickulin et al., 2003). Microanatomical characters associated with the stomata suggest that the fossil shares characters with two extant species, T. mucronatum (Mexico) and T. distichum, which are widespread in North America. Glyptostrobus is known from vegetative (FIG. 21.115) and reproductive (FIG. 21.116) specimens and is the principal biomass component in the formation of Neogene coals in central and eastern Europe; it was also widespread in Eocene swamp habitats in North America (LePage, 2007). Parataxodium is an interesting fossil from the Cretaceous of Alaska that was instituted for pollen and seed cones, as well as leaf-bearing twigs (Arnold and Lowther, 1955). Both long and short shoots occur on P. wigginsii, with entire leaf-bearing shoots apparently shed annually like those in Taxodium (FIG. 21.117) and Metasequoia. The leaves are attached by a short stalk and appear to be alternate in arrangement. The leaf tip is generally blunt. Detached seed cones

CHAPTER 21

CONIFERS

855

Figure 21.113 Reconstruction of Taxodium wallisii. (From Aulenback and LePage, 1998.)

Figure 21.112 Branch of Taxodium wallisii bearing panicles of

pollen cones (Cretaceous). Bar  1 cm. (Courtesy R. Serbet.)

thought to belong to Parataxodium are 1.3 cm long and bear helically arranged, flattened scales; the number of seeds produced per scale is not known. Compressed pollen cones, 2 mm long, are present in the matrix and some contain pollen (Aulenback and LePage, 1998). Drumhellera kurmanniae, from the Upper Cretaceous Horseshoe Canyon Formation

of Alberta, Canada, includes permineralized leafy branches with pollen cones attached in the axils of leaves (Serbet and Stockey, 1991). Each microsporophyll has a single resin canal and two pollen sacs. Pollen is tiny (12–16 μm), and the exit papilla is short. The axillary arrangement of cones is similar to Taxodium, but the number of pollen sacs compares to those in Sequoia and Sequoiadendron. Like Parataxodium, the cones of D. kurmanniae are most similar overall to pollen cones of Taxodium and Metasequoia, suggesting that perhaps these extant genera had a common origin.

856

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.114 Glyptostrobus sp. axis (Eocene). Bar  1 mm.

(Courtesy B. LePage.)

Fokienia is a monotypic genus in the Cupressaceae that is endemic to southeastern Asia. Both vegetative and reproductive parts have been described from the Paleocene of Canada (McIver and Basinger, 1990) as F. ravenscragensis. The fossils consist of pinnate branches bearing four-ranked, persistent, scalelike leaves. Seed cones are 6 mm long, borne in opposite pairs, and consist of 8–10 woody, peltate cone scales. Although two seeds are borne on each cone scale in extant specimens, the number of seeds per scale in the fossil remains unknown. The fossil and extant cones of Fokienia are quite similar, but the vegetative morphology in the two plants is quite different. McIver and Basinger (1990) provided two theories to explain this discrepancy. One hypothesis is that vegetative morphology in this taxon was much more diverse in the past, but only one form has survived today. The other possibility is that the vegetative parts of the plant continued to evolve through the Cenozoic, whereas the reproductive organs remained relatively static.

Figure 21.115 Leafy branch of Glyptostrobus oregonensis

(Miocene). Bar  5 mm. (Courtesy D. Erwin.)

Mesocyparis is another Paleocene foliage type from Saskatchewan, Canada, with scalelike decussate leaves (FIG. 21.118) (McIver and Basinger, 1987). The taxon is known from leafy twigs and associated cones and seeds (FIG. 21.119). Branches are pinnate and the leaves of M. borealis share the greatest number of similarities with leaves of the extant genus Thuja. Interestingly, the seed cones share a number of similarities with those of Chamaecyparis (FIG. 21.120), in which there are three to six decussate pairs of cone scales, with the upper scale containing ovules. On the adaxial surface is a prominent pointed umbo, and the seeds are platyspermic with conspicuous wings. Pollen cones of M. borealis are like those of extant members of the family, with two to three pollen sacs per microsporophyll (FIGS. 21.119, 21.121). McIver and Basinger (1987) suggested that

CHAPTER 21

CONIFERS

857

21.118 Leafy branch of Mesocyparis umbonata (Cretaceous). Bar  1 mm. (Courtesy R. Serbet.)

Figure

Figure 21.116 Seed cones of Glyptostrobus oregonensis

(Miocene). Bar  5 mm. (Courtesy D. Erwin.)

Figure 21.117 Taxodium dubium leafy twigs (Miocene). Bar  1 cm. (Courtesy BSPG.)

Mesocyparis existed with other cupressoid taxa during the Cenozoic and that this group has its closest affinities with the Southern Hemisphere cupressoid genera Libocedrus, Papuacedrus, and Austrocedrus. Sewardiodendron is a Jurassic conifer that is placed in the Taxodioideae (Florin, 1958). It is a woody plant with helically arranged leaves that appear to be two ranked, because the leaves twist at their base and spread in two dimensions. Stomata are confined to the lower surface in distinct bands, with subsidiary cells possessing overarching papillae.

In the same assemblage are helically arranged scale leaves at the bases of lateral shoots. Ovulate cones of Sewardiodendron laxum attached to leafy shoots have been described from the Middle Jurassic of China (X. Yao et al., 1989). The upper portion of the ovuliferous scale is lobed on the distal margin and bears six inverted ovules on the adaxial surface. Pollen cones were also found attached to the same shoots. These were small (4 mm long) with three elongated pollen sacs on the abaxial surface. Specimens of S. laxum from central China are suggested to have their closest affinities with the modern genus Cunninghamia. The plant is reconstructed as a shrub or small tree, perhaps with deciduous branches, based on the presence of a thin cuticle and the position of bud scales (X. Yao et al., 1998). Austrohamia minuta is a Jurassic conifer from Argentina that is known from leafy twigs, branches, and pollen and seed cones (Escapa et al., 2008). This interesting conifer also possesses a combination of characters suggesting placement within the basal Cupressaceae. There are a number of taxodioid leaf morphotaxa. Cryptomeriopsis is used for Cretaceous amphistomatic, falcate leaves with adaxial stomatal bands and Elatidopsis for similar leaves that are lanceolate and epistomatic (Van der Ham et al., 2001). CUPRESSOIDEAE Cretaceous and Cenozoic leaves morphologically similar to Widdringtonia (Widdringtonites) have been reported from numerous sites (McIver, 2001). Leaves of W. americana are

858

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.120 Chamaecyparis eureka ovulate cone. Bar  2 mm. (Courtesy J. F. Basinger.)

Figure 21.119 Suggested reconstruction of Mesocyparis borea-

lis pollen cone. (From McIver and Basinger, 1987.)

heterophyllous, with scalelike leaves borne on long shoots and fleshy leaves on short shoots (FIG. 21.122). Seed cones are globose and consist of four cone scales with an irregular outer surface. Leaves of Fitzroya acutifolius are known from Oligocene–Miocene rocks of Tasmania (R. Hill and Paull, 2003). The fossils indicate that the genus was much more widespread in Gondwana than its restricted habitat today in the Valdivian forests of southern Chile and Argentina, where the single extant species, F. cupressoides, is found.

Tetraclinis is another extant genus that is monotypic. Fossil specimens of Tetraclinis occur in the Oligocene of North America as seeds, seed cones, and branches bearing leaves (Z. Kvacˇek et al., 2000). Branches are flattened and leaves are dimorphic, with tightly appressed, four-ranked scale leaves borne in pseudowhorls. Leave segments have round to blunt apices and are amphistomatic with monocyclic stomata. Cone scales were produced in opposing pairs, and each seed has two broad, cordate-shaped wings. Fossils of this type are also common in the Cenozoic of Europe and called Tetraclinis salicornioides (formerly called Hellia or Libocedrites) because of their similarity to the extant genus Tetraclinis. Wood associated with Tetraclinis foliage in Europe is sometimes included in Tetraclinoxylon (Sakala, 2003). Specimens from the middle Miocene of Denmark include well-preserved branchlets with leaves that could be easily embedded in paraffin and sectioned with a rotary microtome (Friis, 1977b). Vegetative remains of Thuja are also known from the fossil record (FIG. 21.123). A single branch with seed cones of the Thuja type was discovered in the Late Cretaceous of Alaska (LePage, 2003c). The fossil seed cones of T. smileya are morphologically identical to those of modern Thuja and are interpreted as abortive because none of the cone scales are reflexed. In T. polaris from the Paleocene of the Arctic, branches bore scalelike decussate leaves, and cones had eight to nine pairs of thin cone scales (McIver and Basinger, 1989b).

CHAPTER 21

CONIFERS

859

Figure 21.122 Reconstruction of Widdringtonia americana leafy shoot with cones. (From McIver, 2001.) Figure 21.121 Mesocyparis borealis fertile branch showing opposite arrangement of seed cones (arrows). Bar  1 cm. (Courtesy J. F. Basinger.)

CUPRESSACEOUS WOOD Numerous coniferous woods believed to be cupressoid have been reported from a number of Cretaceous and Cenozoic localities around the world. In Taxodioxylon taxodii (Upper Cretaceous of western Canada), the tracheids are large, with characteristic cross-field pit patterns (Ramanujam and Stewart, 1969a). Xylem rays are uni- to biseriate, and the rays are up to 35 cells high. Sieve elements have been described in taxodiaceous bark from the Upper Cretaceous of Alberta under the binomial Taxodioxylon gypsaceum (Ramanujam and Stewart, 1969b). Margeriella cretacea includes silicified wood and leaves collected in Upper Cretaceous rocks in central California and is believed to be closely related to the Taxodium group (Page, 1973). The leaves are helically arranged, amphistomatic, and 4 cm long. They are apetiolate and have a triangular point at the

tips. The wood consists of tracheids with a single row of bordered pits on the radial walls and low, uniseriate rays. In general, the histology of the wood is similar to the form genus Cupressinoxylon, a type that has also been attributed to the Podocarpaceae (Kräusel, 1949b). An in situ forest of permineralized stumps of Glyptostroboxylon is known from the Paleocene–Eocene boundary in Belgium. The wood lacks resin canals and has cross-field pits that vary from glyptostroboid to taxodioid, making assignment to a modern genus difficult (FaironDemaret et al., 2003). The trees are believed to have existed in an environment in which there was an oscillating water table, much like that in modern Taxodium distichum (bald cypress) swamps. Other commonly described wood types often placed in this family include Widdringtonioxylon, Thujoxylon, Libocedroxylon, and Juniperoxylon. SUMMARY DISCUSSION: CUPRESSACEAE

The oldest generally accepted fossil assignable to the Cupressaceae is Parasciadopitys aequata, a permineralized

860

PaleoBOtany: the biology and evolution of fossil plants

cupressoid forms, were competing for niche space. R. Hill and Brodribb (1999) hypothesized that the ability of certain species to flatten their leaves and the species longevity exhibited by a number of taxa may represent strategies that increased the ability of the conifer to compete for space with angiosperms. Two environmental parameters that are believed to have greatly affected the distribution of cupressoids include the inability to rapidly regenerate after fire and the high sensitivity to changes in rainfall. The similarity in certain foliage types as seen in the fossil record has also been used to suggest a response to environmental pressure. In this instance, however, the morphological uniformity of certain foliage types known to represent widely divergent conifer groups has made interpreting the past biodiversity of taxa especially difficult. Among modern plants, the general slow growth and long-lived nature of these plants has made them unattractive experimental organisms; however, they do provide interesting models of life history strategies in relationship to ecology and adaptive physiology. SCIADOPITYACEAE

Figure 21.123 Impression of Thuja dimorpha (Miocene).

Bar  5 mm. (Courtesy D. Erwin.)

cone from the Middle Triassic of Antarctica (X. Yao et al., 1997). This taxon has bract–scale complexes that are helically arranged on the cone axis. The bract and scale are fused only at the base; ovuliferous scales have five distal lobes and an equal number of recurved seeds. The fossil cone combines characters of the Cupressaceae and Sciadopitys, supporting some phylogenies that suggest a common origin for these plants (Gadek et al., 2000). There is good fossil evidence of the Cupressaceae back to the Jurassic and some woods and leaves with cupressaceous characters in the Late Triassic. By the Cretaceous, the record of fossil Cupressaceae is worldwide, suggesting that certain features were already present before that time. Both vegetative and reproductive features of the fossils clearly indicate that the modern genera represent a small part of what must have been a conspicuous component of the Mesozoic conifer flora. The nature of the cone–scale complex and certain vegetative features have been used to link the Cupressaceae with members of the Voltziales (Miller, 1982), including some members of the Majonicaceae. Although the Cretaceous was a time of extensive radiation of flowering plants, it was also a time during which certain conifers, including some

The extant conifer Sciadopitys verticillata (umbrella pine) is included in its own family (Farjon, 2005); in other treatments, it is included within the Cupressaceae or Taxodiaceae. This evergreen tree is endemic to Japan and characterized by green cladodes (photosynthetic branches) that function as leaves. Sciadopitys has several features in common with members of the Cupressaceae sensu lato. These include seed wings that are formed from the seed integument and nonsaccate pollen. Unique features of Sciadopitys include wood lacking axial parenchyma, unique vasculature of the bract–scale complex, and the double-needle nature of the leaf (Ohsawa, 1994). Sciadopitophyllum includes compressed shoots bearing whorls of leaves that are morphologically identical with modern Sciadopitys (Christophel, 1973). The genus is known from the Upper Cretaceous to the Paleocene. Specimens of S. canadense consist of foliar shoots bearing 8–12 lanceolate leaves, each up to 16 cm long, but not all the fossil leaves have the apical notch at the tip that is characteristic of Sciadopitys. It is important to note, however, that this feature apparently becomes obscured even in extant specimens as development continues. Takaso and Tomlinson (1991) examined cone and ovule development in S. verticillata and indicated that the development of this bract–scale complex may be similar to that in certain voltzialeans such as Pseudovoltzia. Sciadopityostrobus kerae is an ovulate cone from the Upper Cretaceous of Japan characterized by five-lobed cone scales

CHAPTER 21

Figure 21.124 Sciadopityoides microphylla leaf and epider-

mis. (From Taylor and Taylor, 1993.)

with recurved ovules (Saiki, 1992). Because the cones are permineralized, it is possible to show that the origin of the vascular trace to the cone scales is different from that in the modern species. Another cone, Sciadopitys yezo-koshizakae, also from the Upper Cretaceous of Hokkaido, has nine seeds per scale complex like the modern species (Ohsawa et al., 1991a). Based on palynology and maceral analysis of Miocene lignite deposits in Germany, it is suggested that Sciadopitys grew in drier sites in the swamps, whereas Sequoia grew in moister areas (von der Brelie and Wolf, 1981). Sciadopityoides leaves possess an expanded leaf base with a prominent abscission scar (FIG. 21.124) (Sveshnikova, 1981). The margin of the stomatal groove forms a lip that contains papillae. Sciadopityoides microphylla can be up to 2 cm long and contains closely spaced, randomly distributed stomata. In Sciadopityoides variabilis, the stomatal band is five to eight stomata wide with cuticular papillae prominent along the flanks of the stomatal groove (Bose, 1955). The guard cells are greatly thickened and not sunken beneath the surface. Pollen grains obtained from the same macerations are spherical, up to 48 μm in diameter, and almost identical to the pollen of modern Sciadopitys verticillata. Many leaves lack a notched tip (Manum, 1987). Oswaldheeria is a genus of leaves from the Cretaceous of Siberia and the Arctic which are similar to those of Sciadopitys (Bose and Manum, 1990). They differ in the nature of the median stomatal furrow and features of the

CONIFERS

861

epidermis. None of the fossil leaves possesses the same whorled arrangement as in S. verticillata, nor are they dimorphic in venation and stomatal distribution as is the case in the extant species. Although nothing is known about the reproductive organs of these plants, they did represent a conspicuous component of the flora between 55° and 65° N in Spitsbergen, Greenland, and Baffin Island during the Lower Cretaceous (Nosova, 2001). Bose and Manum (1990) erected the family Miroviaceae for fossil leaves with Sciadopitys morphology and placed Oswaldheeria in this family. The family, however, is based only on vegetative remains as no cones have been found. Some have placed Oswaldheeria in the ginkgophytes (Reymanówna, 1985), but Gordenko (2007) noted epidermal similarities between O. eximia and some podocarps. Wood of Podocarpoxylon occurs at the same site as the leaves. Oswaldheeria eximia is described as having two vascular bundles and three resin canals, with transfusion tissue located below the bundles. PARARAUCARIACEAE

Pararaucaria patagonica is a small cone from the Middle Jurassic Cerro Cuadrado petrified forest in Argentina. Despite the generic name, this taxon includes features of both the Cupressaceae and the Pinaceae (Stockey, 1977). The cones are large and woody, with fused ovuliferous scales and bracts (FIG. 21.125). Each scale contains a single, cordate-shaped seed ~6 mm long, many with embryos. Extending from the surface of the seed are two wings composed of anastomosing rows of glandular hairs that are continuous with the upper surface of the ovuliferous scale. Because of the intermediate status of this cone, it is included in its own family (Stockey et al., 2005). PINACEAE

The Pinaceae today are principally a Northern Hemisphere, temperate family and are the largest modern conifer family in terms of the number of species, with 10 or 11 genera and 250 species. The family includes shrubs and trees, some up to 100 m tall, and several widespread taxa, such as Abies (firs), Picea (spruce), Tsuga (hemlock), and Pinus (pines). Leaves in the Pinaceae are linear; wood is pycnoxylic, resinous, and commercially valuable. Pollen and seed cones are borne on the same plant (monoecious) with the ovule-bearing cones large and consisting of helically arranged ovuliferous scales, each bearing two seeds on the upper surfaces. Each scale is subtended by a bract and free for most of its length. Pollen cones are small and composed of sporophylls arranged helically on an axis; each microsporophyll bears two elongated pollen sacs on the abaxial surface. In some members, pollen

ˇ

862

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.125 Longitudinal section of Araucaria mirabilis

showing embryos within seeds (Jurassic). Bar  12 mm. (Courtesy R. A. Stockey.)

Figure

21.127 Permineralized

pine

cone

(Cretaceous).

Bar  2 cm. Figure 21.126 Pinus strobipites (distal surface) (Miocene). Bar  10 μm. (Courtesy M. S. Zavada.)

grains are bisaccate (FIG. 21.126), and the saccus wall is internally ornamented with endoreticulations. It has not been possible to determine exactly when the family evolved, but judging from the diversity of the seed cones (FIGS. 21.127,

21.128) present during the Cretaceous (Miller and Robison, 1975; X.-Q. Ohsawa et al., 1991b), the group must have been well established early in the Mesozoic. Four subfamilies based on cone and seed characters are used here to classify the fossil taxa (Farjon, 1990), and there is some support for this classification scheme based on a

CHAPTER 21

CONIFERS

863

Compressed pine cone (Miocene). Bar  1.5 cm. (Courtesy BSPG.)

Figure 21.128

Figure 21.130 Pinus crossii cone (Oligocene). Bar  1 cm. (Courtesy G. R. Upchurch.)

three-genome data set (X.-Q. Wang et al., 2000). Although Wang et al.’s analysis divided the family into two large clades, as LePage (2003a) has pointed out, each of these large clades contains two smaller clades, which correspond to the four subfamilies that are based on morphological characters (FIGS. 21.129, 21.130).

Figure 21.129 Five needle fascicle of Pinus crossii leaves (Oligocene). Bar  1 cm. (Courtesy G. R. Upchurch.)

PINOIDEAE One of the oldest fossils that may be a member of the Pinaceae is Compsostrobus (FIG. 21.131), an ovulate cone from the Upper Triassic of North Carolina (Delevoryas and Hope, 1973, 1987). Specimens are preserved as compressions, but sufficient details are present so that characteristics of the family can be identified. The largest cone is 13 cm long and contains loosely arranged bracts and ovuliferous scales. The scales are spatulate in outline with two flattened seeds on the upper surface. Extending from the distal end of each seed is an elongated micropylar tube, suggesting that pollination droplets were produced. Associated with the

864

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.132 Pinus leptophylla twig with needle leaves (Oligocene). Bar  4 cm. (Courtesy BSPG.)

Figure 21.131 Suggested reconstruction of a portion of a Compsostrobus neotericus cone (left). Adaxial surface of ovuliferous scale with two seeds (upper right), and lateral view of microsporophyll with pollen sacs (lower right). (From Delevoryas and Hope, 1975.)

ovulate cones are shoots bearing linear leaves. It has been suggested that the leaves persisted on the stems for some time since they occur on axes of varying diameter. Prepinus was established for dwarf shoots and attached, needlelike leaf bases of Cretaceous age that were thought to represent a transitional stage between the Paleozoic cordaites and modern members of the Pinaceae. Anatomical studies indicate, however, that these leaves are distinct from those of Pinus and thus cannot be considered an early stage in the evolution of a Pinus leaf (Robison, 1977). In addition, leaves assignable to Pinus are known from several Cretaceous sites. GENUS PINUS Extant species of Pinus are generally divided into two monophyletic subgenera, subgenus Strobus (haploxylon or soft pines), with one vascular bundle per leaf, and subgenus Pinus (diploxylon or hard pines), with two vascular bundles per leaf, and these morphological distinctions are confirmed based on plastid DNA sequence data (Geada-López et al.,

2002; Gernandt et al., 2005). These subgenera are further subdivided into sections and subsections. Based on the fossil leaf characters, it is noteworthy that some subsections of Pinus as they are recognized today were not in existence during the Late Cretaceous. Foliage characters do suggest, however, that the subgenus Pinus may be the oldest within the genus (Stockey and Nishida, 1986). Although there are numerous reports of the genus Pinus in the literature (FIG. 21.132), many of these specimens are not known in sufficient detail to provide information about the evolutionary history of the taxon. There are numerous reports of pinaceous ovulate cones (FIG. 21.133) that have provided important information about the diversity of the family in general and seed cones in particular (Ohsawa, 1997). A large number of these structurally preserved cones thought to have been members of the Pinaceae (Miller, 1976) have been separated into two broad groups based on whether they can be assigned to modern genera or fall outside the circumscription of extant taxa. Perhaps the oldest cone assignable to Pinus is P. belgica, a lignitic cone from the Lower Cretaceous of Belgium (Alvin, 1960a). The cones are 4.5 cm long and include a small triangular bract that subtends a large (2 cm long) ovuliferous scale. The seeds possess a thin membranous wing, and one seed per cone scale is most common. Histologically, the cone scale is similar to the modern cone of P. sylvestris, a form with abundant sclerenchyma and resin canals. Another pinaceous cone from the Late Cretaceous is Pinus cliffwoodensis, an ovulate strobilus nearly 5 cm long from the Magothy

CHAPTER 21

CONIFERS

865

Figure 21.134 Cross section of Pinus arnoldii (Eocene). Bar  1 cm. (Courtesy R. A. Stockey.) Figure 21.133 Pinus sp., twig with two seed cones preserved

within a barite concretion (Oligocene). Bar  1 cm. (Courtesy BSPG.)

Formation in New Jersey, USA (Miller and Malinky, 1986). In the pith of the cone axis are nests of sclereids and a ring of secondary xylem with resin canals. Pitting is uniseriate circular bordered. Bract and ovuliferous scale vasculature begins as a single trace that divides in the outer cortex. Vasculature in the scale occurs as a series of traces, each with a resin canal on the abaxial surface. These anatomical features are most like those of Pinus subgenus Pinus section Pinus. Silicified remains of pines are also known from the Miocene of North America and include needles, ovulate and pollen cones, and short shoots (Miller, 1992). Character comparison suggests a close relationship with closed-cone pines. By the Pliocene, ovulate cones of Pinus possess most of the characters distributed in modern subsections of the genus (McKown et al., 2002). Of the remaining structurally preserved cones, most can be included in the morphogenus Pityostrobus if they cannot be related to a modern genus, but possess various combinations of anatomical characters (Miller, 1985); Obirastrobus has been used for Cretaceous cones from Japan that are similar to Pityostrobus (Ohsawa et al., 1992). Most of the anatomical and morphologic characteristics in Pityostrobus also appear in the genus Pinus. For example, Miller (1976) noted that most species of Pityostrobus have at least one feature in common with modern Pinus cones, whereas the majority have at least two. Thus, at present, it remains an impossible task to determine whether the pityostroboid cones represent a plexus from which modern Pinus cones evolved or are merely

divergent lines of seed cone evolution that include combined features of the ancestral cone types (Smith and Stockey, 2001). For example, in Pityostrobus hallii (Late Cretaceous of Maryland), the ovuliferous scales are very thin, as in modern species of Picea, whereas the vascular trace pattern is similar to that of Pinus (Miller, 1974). Bract and scale fusion, however, are more similar to the modern genera Abies, Larix, and Tsuga. Pityostrobus palmeri has seed scales like those of Pinus but a vascular system more similar to ovulate cones of Cedrus (Miller, 1972). Partially enclosed seeds represent a unique feature in Pityostrobus beardii (Smith and Stockey, 2002). Other seed cones are placed in the morphogenus Pseudoaraucaria (Alvin, 1957). Each cone scale has two seeds that are separated by a ridge of tissue, giving the impression that the seeds are embedded like those of Araucaria. In P. major, the seeds are, in fact, partially embedded and believed to have been dispersed by a separation of the upper part of the scale (Alvin, 1988). By the Paleogene, cone characteristics appear to be more consistent with extant subgenera. Pinus arnoldii is a silicified cone from the middle Eocene Princeton chert, Allenby Formation, British Columbia (FIG. 21.134) (Miller, 1973). This species is thought to be closely related to certain extant species in the subgenus Pinus, based on the abaxial position of resin canals in the ovulate scale, the inflated scale apices, and the strong curvature of the scale vascular strands. It is similar to two other Paleogene species in the subgenus Pinus, P. princetonensis, also from the Princeton chert (Stockey, 1984), and P. baileyi, which is known from the Eocene Thunder Mountain and Oligocene Haynes Creek floras of

866

PaleoBOtany: the biology and evolution of fossil plants

Figure 21.135 Cross section of Pinus similkameenensis (Eocene). Bar  2 mm. (Courtesy R. A. Stockey.)

Figure 21.136 Cross section of fascicle of Pinus similkameen-

In addition to P. similkameenensis, the Princeton chert contains two other species of leaves assignable to Pinus. The leaves of P. allisonii are amphistomatic and occur in fascicles of two; each is vascularized by two bundles separated by a band of fibers. These leaves occur with a three-leaved pine named P. andersonii (Stockey, 1984). This species has an expanded band of fibers on the adaxial side of the leaf. There are several reports of permineralized Pinus leaves in the Cretaceous (Stockey and Ueda, 1986). One of these is P. haboroensis from the Upper Cretaceous of Hokkaido, Japan (Stockey and Nishida, 1986). Leaves are borne in fascicles of three to four, and each leaf contains a double vascular strand associated with six to eight medial resin canals. Stockey et al. (1990) have described several different types of seedling-like structures from the same Upper Cretaceous rocks. Some may represent stages in abnormal seedling growth, whereas others may be decorticated shoots. Pinus and Prepinus vegetative remains are also known from the Upper Cretaceous in North America (Gandolfo et al., 2001). These fusinized fossils occur with an isolated pollen cone described as Amboystrobus cretacicum. The cone is 2.5 cm long, has two abaxial pollen sacs per microsporophyll, and eusaccate, monosulcate pollen. PINUS WOOD Miocene petrified wood from the Succor Creek region of eastern Oregon is called Pinuxylon woolardii (Tidwell et al., 1986). The wood is most similar to subgenus Strobus and has distinct growth rings and numerous heterogeneous rays 8–21 cells high. Based on floral associations at several of the Succor Creek sites (Taggart and Cross, 1980; Cross and Taggart, 1982), it is suggested that P. woolardii lived under xeric conditions in a relatively cool climate. This is supported by anatomical comparisons between the fossil and several living species that inhabit dry sites at relatively high elevations, for example, Pinus monophylla and P. balfouriana.

ensis leaves (Eocene). Bar  0.5 mm. (Courtesy R. A. Stockey.)

Idaho, USA (Erwin and Schorn, 2006). Occurring in the same rocks with P. arnoldii cones are stems (FIG. 21.135), dwarf shoots, and needles (FIG. 21.136) that have been assigned to P. similkameenensis (Miller, 1973). Each spur shoot bears a fascicle of five needles that are anatomically identical with leaves of extant white pines (e.g., P. strobus). In cross section, they have the shape of an equilateral triangle. Stomata are slightly sunken in two or three rows on the ventral (adaxial) faces. Stems possess well-developed growth rings and abundant resin canals.

LARIX The earliest record of the genus Larix is from the middle Eocene of Axel Heiberg Island in the Canadian Arctic (LePage and Basinger, 1991). Larix altoborealis includes mummified vegetative and reproductive specimens (FIG. 21.137). The branching systems are dimorphic with the brachioblasts (reduced branches) giving rise to 70 narrow leaves in each fascicle. Seed cones are borne singly and are believed to have persisted once seeds were shed. Cone scales, each with two winged seeds, are subtended by a delicate bract. Modern species can be segregated into two morphological categories based on bract length; long bract forms are geographically

CHAPTER 21

CONIFERS

867

in the identification of species and perhaps useful in phylogenetic analyses (LePage, 2001). Permineralized leaves of Picea from the Miocene of Japan are quadrangular in cross section with stomata on all faces (Matsumoto et al., 1994). Differences in histology have been used to distinguish the fossil leaves from modern forms.

Figure 21.137 Mummified branch and cone of Larix altobo-

realis (Eocene). Bar  1 cm. (Courtesy B. LePage.)

restricted today, whereas those with shorter bracts are widespread and include all the fossils described to date (LePage and Basinger, 1995a,b). Structurally preserved seed cones of Larix and Picea have also been reported from the Pliocene of Alaska (Miller and Ping, 1994). PICEOIDEAE Members of this subfamily are evergreen and characterized by well-developed pulvini, small swellings at the base of the leaves, on the shoots. Leaves are individually attached, flattened, and helically arranged. Various fossils attributed to Picea have been reported from the Cenozoic of the Northern Hemisphere (LePage, 2001). Three species of vegetative and reproductive remains occur in the middle Eocene Buchanan Lake Formation of Axel Heiberg Island. A comparative study of several fossil species together with extant forms suggests that bract morphology may be an important character

ABIETOIDEAE The genus Abiocaulis is used for structurally preserved stems that exhibit anatomic features similar to those of modern Abies. Specimens of Abiocaulis verticillatus (Cretaceous of Belgium) consist of long shoots with helically arranged, decurrent leaf bases; no leaves were found attached to the stems (Alvin, 1960a). In cross section, the stems consist of a distinct zone of periderm that follows the outline of the leaf bases. Nests of irregular sclereids are present in both the pith and the outer portion of the cortex. Features of the wood, including spiral checking of the tracheid walls and the pattern of rays, suggest a wood of the Cedroxylon type. Similar wood is produced by extant species of Keteleeria and Abies. The presence of persistent bud scales on the fossil specimens further supports affinities within the Pinaceae. Associated with the stems were numerous leaves of the Elatocladus type (Miller and LaPasha, 1985), although none were found attached. The number of vertical resin canals in the wood of extant species of Keteleeria is useful systematically (Linet al., 2000) and may represent an important characteristic in defining certain types of fossil conifers (Blokhina et al., 2006). Cedrus-like silicified wood from the Lower Cretaceous of Alaska has been given the binomial C. alaskensis. The wood contains widely spaced growth rings, with earlywood tracheids that are hexagonal in cross section, and resin canals in the ray system, that is, horizontally oriented. Pitting on tracheids is opposite and the pit torus is large; distinct crassulae are present between the pits. No ray tracheids are present, but vascular rays are narrow and have simple pits on the horizontal cell walls. Tsuga or hemlock is known from North America and Asia today, and fossils of all plant parts are known to extend from the Late Cretaceous onwards (LePage, 2003b). The seed cone axis in T. swedaea consists of secondary xylem surrounding a pith. Cone scale morphology and anatomy are hypothesized as useful in species identification, and the decrease in cone size during the Miocene is interpreted as a response to global climate warming. Permineralized leaves from the Miocene of Japan are anatomically similar to those of the extant species T. heterophylla (Matsumoto et al., 1995). The leaves of T. shimokawaensis possess a stomatal band on either side of the abaxial surface and a resin canal on the

868

PaleoBOtany: the biology and evolution of fossil plants

abaxial side of the single vascular strand; an endodermis partially encloses the vascular trace. SUMMARY DISCUSSION: PINACEAE

Like the other conifer families, the origins of the Pinaceae are difficult to resolve. What is known about the subsequent evolution of the family suggests that considerable diversity was already present by the Cretaceous and that the genus Pinus was a conspicuous element early in the evolution of the family, based on both foliage (FIG. 21.138) and cones (FIG. 21.139). By the mid-Cretaceous, pines were present in North America and can be recognized as an eastern and western lineage, based on a combination of both modern and fossil species. Axelrod (1986) suggested that subsections of the genus Pinus were formed during the Late Cretaceous–

Paleogene in association with the formation of new habitats that were initially dry and edaphically poor. He also suggested that modern pines represent an excellent example of a group that has flourished during the past several hundred years because of human disruption of the ecosystem. Of particular importance in understanding the fossil history of the family are the limitations imposed by using isolated organs of various conifers collected from various geographic and stratigraphic sites to erect species. Erwin and Schorn (2005) have discussed this problem as it applied to isolated winged seeds of Pinus. They underscore the necessity of accurate identifications based on a large sample size and a suite of characters that can be related to extant taxa in tracing biogeographic and phylogenetic trends within conifer families. An interesting ancillary hypothesis that has been suggested interprets the radiation of the pines as being directly associated with their ability to form symbiotic associations with ectotrophic mycorrhizae. The discovery and recognition of these biotic interactions both in the fossil record and in the modern taxa represents an important but relatively unstudied dimension in tracing the evolution of certain plant groups such as the Pinaceae. CEPHALOTAXACEAE

Figure 21.138 Cross section of permineralized pine leaf

(Eocene) Bar  750 μm (Courtesy R. A. Stockey.)

Figure 21.139 Section of Pinus sp. microsporophyll showing pollen sac with pollen (Eocene). Bar  200 μm. (Courtesy R. A. Stockey.)

The Cephalotaxaceae today consists of a single genus, Cephalotaxus, which includes about six species of small trees and shrubs. Leaves are opposite and needlelike, and most species appear to be dioecious. The small ovulate cones produce one or two erect ovules that morphologically appear similar to those of some taxads. Pollen cones have three to eight pollen sacs per microsporophyll. Pollen grains are small and characterized by a small distal leptoma. Thomasiocladus is the name used for Jurassic impression– compression specimens from England (Florin, 1958). Only vegetative remains are known, and these consist of branches with helically arranged, two-ranked linear leaves. Stomata are haplocheilic and arranged in rows on the abaxial surface. Each subsidiary cell contains a small papilla; guard cells are sunken. Miller (1977) noted that similar vegetative remains have also been placed in the extant genus Cephalotaxus, with specimens known from numerous geographic regions in both Jurassic and Cretaceous. Reproductive structures that have been assigned to this family include Cephalotaxospermum and Cephalotaxites. Cephalotaxospermum includes Cretaceous, drupe-like structures 1.3 cm long that are constructed of an outer fleshy zone surrounding a more sclerotic, inner integument (Berry, 1910). The affinity of these structures as podocarpaceous has also been suggested. Based on vegetative remains, the

CHAPTER 21

Cephalotaxaceae can be traced from the Middle Jurassic onward. The small number of fossil reports, however, makes it impossible to determine whether the family was more widespread during the past than it is today. Molecular analyses using chloroplast matK genes and nuclear ITS sequences suggest that Cephalotaxaceae and taxad genera are monophyletic and that Cephalotaxus is basal (Cheng et al., 2000). TAXACEAE

Extant members of the Taxaceae are evergreen shrubs or small trees with helically arranged leaves. Pollen-bearing cones have microsporophylls with three to nine pollen sacs. Members of this group differ from other conifers in bearing a single ovule (occasionally two) terminally on a modified shoot. Below the ovule are scales, but there is no evidence of a bract–scale complex like that in other conifers. In the genus Taxus, the ovule is partially surrounded by a fleshy envelope termed an aril. Pollen in the group is nonsaccate. Most authorities recognize four to five extant genera and 20 species. Possibly, the best-known fossil member of the family is Palaeotaxus from the Lower Jurassic (Florin, 1951). Both vegetative and ovulate parts are known. The foliage consists of linear needles twisted near the base so that they lie in a single plane, much like the leaves of modern Taxus (yew). Leaves are hypostomatic, and the haplocheilic stomata are arranged in two rows, one on either side of the midrib. Subsidiary cells lack papillae and trichomes are absent. The fine, undulating, anticlinal cell walls with irregular thickenings have been used to separate the fossils from extant taxa. The ovule-bearing shoots are radially symmetrical in cross section and borne in the axils of vegetative leaves. The cone axis is covered with helically arranged acuminate scales and terminates in an ellipsoid, orthotropous ovule. The compressed nature of the specimens makes it difficult to determine whether an aril is present or not (Florin, 1963). Two-ranked leaves are included in the Cretaceous genus Tomharrisia (Krassilov, 1967). The genus is also known from the Jurassic and has hypostomatic leaves with monocyclic stomata surrounded by four to six papillae. Marskea is another Jurassic taxad that has leaves almost identical with those of living Taxus (Florin, 1958). The ovulate organ consists of a terminal ovule 7 mm long and 3 mm in diameter, with a faint outline of an aril. The axis of the ovule is smooth, except near the base of the seed, where there is a whorl of scales. The pollen-bearing organ of Marskea is thought to have consisted of sporophylls with upturned tips and three pollen sacs attached to the abaxial surface in a row (Harris, 1979). Pollen grains are round and

CONIFERS

869

similar to grains of extant Torreya in the presence of a thin, distal germination site. Poteridion is another ovuliferous shoot that morphologically resembles Taxus (Harris, 1979). This Jurassic specimen also has a terminal ovule that is slightly enclosed by scale leaves. Vesquia includes fossil seeds thought to be related to Torreya (Alvin, 1960b). Specimens of V. tournaisii from the Early Cretaceous of Belgium are 1 cm in diameter; the integument is composed of angular stone cells with numerous pits in the wall. The distal end of the seed is attenuated into a micropylar canal. In the lower third of the seed are two grooves (one on either side of the integument) through which vascular tissue may have passed. The megaspore membrane is smooth, unlike the granulose texture present in most gymnosperms, but this may simply reflect the stage of development before fossilization. Foliage of Amentotaxus can be traced from the Cenozoic and includes linear leaves with abaxial bands of stomata. A ring of subsidiary cells surrounds each deeply sunken stoma. During the Cenozoic, the genus was a common element in many mesophytic forests (Ferguson et al., 1978). Several genera of Mesozoic fossil woods have been included in the Taxaceae (Greguss, 1967). Taxaceoxylon has distinct growth rings, and uniseriate-bordered pits are present on the radial walls of the tracheids. The rays are generally homogeneous; xylem parenchyma and resin canals are not present. Several of these features are common in other wood types, indicating the difficulty in determining the affinities of coniferous petrified wood. SUMMARY DISCUSSION: CEPHALOTAXACEAE AND TAXACEAE

The earliest fossils assignable to the Taxaceae exhibit foliage and reproductive structures almost identical with those of extant forms. As a result, it is difficult to trace the origins of the group. Molecular phylogenetic analyses (Cheng et al., 2000) suggest that the two families form a clade, with the Cephalotaxaceae basal to the taxads. Several ideas have been offered to explain the unusual nature of the ovulate organs in the taxads. Florin (1954) suggested that the terminal ovule in the Taxaceae is not homologous with the ovules in other conifers but rather evolved from an ancestor that had terminal ovules. However, Harris (1976b) suggested that the taxads might be related to a Carboniferous ancestor in which the dwarf shoots were not borne on a primary axis but rather were distributed on leafy shoots. According to this interpretation, modifications of the cone apex resulted in ovules being in a terminal position. Another potential ancestor might have demonstrated an organization like that found in the Buriadiaceae or Ferugliocladaceae, where a morphological reduction of

870

PaleoBOtany: the biology and evolution of fossil plants

a simple strobilus could lead to terminally borne ovules. At the present time, the most widely accepted hypothesis is that the ovule in the taxads is a remnant of a fertile short shoot that has become modified from its original axillary position (Miller, 1988). The origin of the aril is difficult to interpret since the nature of the structure in fossils remains equivocal. Accordingly, the fleshy aril that surrounds the terminal ovule in this group may represent a modification of the sterile scales at the base of the short shoot or a portion of the axis.

CONCLUSIONS All the families of modern conifers appear approximately at the same time in the Mesozoic. The only exceptions are Cephalotaxaceae and Taxaceae, the two families that have a poor fossil history. The remaining families all have their earliest representatives appearing at least by the Late Triassic. In attempting to trace the origins of the modern conifer families, the emphasis has been on the ovulate bract–scale complex and the suggested homologies between this structure and the compound cones of the Paleozoic conifers (Miller, 1982). Florin’s initial idea was that the compound cones of many Paleozoic conifers could be traced directly to the morphologically similar reproductive organs of the Cordaitales. He regarded the ovuliferous cone scale in modern conifers as homologous with a reduced dwarf shoot that had become flattened and was associated with the fusion of the fertile and sterile scales. Accompanying this change in symmetry was a reduction in the number of stalked ovules, an inversion of the ovules so that the micropyle was directed toward the cone axis, and a fusion of the seeds to the newly evolved ovuliferous scale. This idea was subsequently modified by Clement-Westerhof (1988), who suggested that the early conifers had ovules that were abaxially attached to the fertile scales. She suggested that some of the early voltzialeans from the Pennsylvanian represent the ancestral condition, with radially symmetrical dwarf shoots, each with several fertile scales, whereas the younger members of the Utrechtiaceae (Walchiaceae of Clement-Westerhof) are derived. Still another point of view was expressed by Mapes and Rothwell (1990), who consider the large number of sterile and fertile scales in Emporia as representing a condition not far removed from the cordaitalean Cordaitanthus (Chapter 20). They regarded the basal or sublateral point of ovule attachment in Emporia as evidence of an ancestral condition. Irrespective of the derivation of the conifer cone scale, a considerable gap remains between the organization of these early reproductive organs and the forms present in the younger conifers (Jurassic–recent).

Today, most consider the conifer cone scale to have evolved from a reduced dwarf shoot in a compound strobilus. There is perhaps less agreement regarding whether the ovules were attached to stalks or were abaxial (in relationship to the dwarf shoot axis) on fertile scales. In the Buriadiaceae and the Ferugliocladaceae, two families that possess a coniferophytic organization of vegetative parts, the ovules are orthotropously borne on stalks that arise in the axil of a bract. There are no features to suggest that the reproductive organs in these families are compound. In the Thucydiaceae, the ovule-bearing parts are produced in fertile zones to form vegetative regions; thus, there are no terminal cones. The evolution of pollen organs in the conifers is an equally perplexing situation. A number of pollen cones in the Triassic, for example, Masculostrobus, Sertostrobus, and Voltziostrobus, are simple and constructed like those of extant conifers, with a central axis bearing helically arranged sporophylls. Pollen sacs in some forms are attached to the stalk of the microsporophyll or to the lower portion of an otherwise peltate microsporophyll. A minor change in the orientation of the laminate portion of the microsporophyll would result in the pollen sacs being abaxial, like those in modern families. The ancestry of these pollen cones continues to be problematic, although, as in the ovulate organs, the cordaites have traditionally been suggested as a morphological starting point. Such a transformation would require viewing a cluster of axillary pollen cones in Pinus as homologous with an ancestral compound strobilus such as that found in Cordaitanthus. The homologies of the compound pollen cones in the Thucydiaceae are even more problematic. As is so often the case, the more we learn about fossil plants, the more difficult it becomes to use a particular set of characters to define a taxonomic group or explain the origin of certain morphological features. This is particularly true of the conifers, which were at one time considered the direct descendants of the cordaites, although it has been suggested that the Podocarpaceae, the Cephalotaxaceae, and the Taxaceae are related to certain Mesozoic seed ferns based on the number of integuments and the nature of a cupule (Semikhov et al., 2001). To further confound classification schemes, many of the fossil taxa discussed earlier combine characters from multiple families, a situation which is not unexpected, especially in the older fossils. The inability to classify a number of late Paleozoic coniferophyte plants and the absence of transitional forms leading to modern families led Miller (1988) to suggest that perhaps the conifers constitute a polyphyletic group. This hypothesis has been supported by several cladistic analyses (Rothwell and Serbet, 1994), whereas other analyses have supported monophyly for

CHAPTER 21

the conifers (Rydin et al., 2002) or were unresolved (Nixon et al., 1994). All the analyses that have included fossil taxa only serve to confirm that we have much to unravel about conifer systematics and evolution, as has been suggested by Rothwell and Mapes (2001). Finally, although paleobotanists provide species descriptions that include variation where possible, preservational biases and developmental data have not been widely used in placing fossils into families. For example, in some extant conifers, the ovule changes from orthotropous to reflexed during development. This is also the case for the size, shape, and degree of separation of the bract and cone scale, characters that have been historically used to define plesiomorphic states. To deal with these variables, it will become increasingly necessary to examine the development of structures in modern taxa and to correlate these, where possible, with a sufficient

CONIFERS

871

number of fossil specimens. This is being done in a number of studies, either directly (Bobrov et al., 2004) or based on the primary literature. But, there is another dimension that is often overlooked in discussions of conifer seed cone evolution, which involves pollination biology. Tomlinson and Takaso (2002) provided a detailed description of seed cone development in modern conifer families that demonstrates the timing of specific developmental patterns relating to pollen capture, seed protection, and seed dispersal. Cresswell et al. (2007) studied the aerodynamics of various modern conifer cones and demonstrated that the cone morphology has little to do with how pollen is captured. Such functional and life history parameters need to at least be kept in mind when discussing fossil conifer seed-cone evolution. Such an approach is necessary to test hypotheses about the adaptive value of structural and morphological features observed in the fossil record.

This page intentionally left blank

22 Flowering plants Angiosperm origins .......................................................... 876

Monocotyledons .................................................................917

Origin of the Flower ..........................................................................877

Alismatales........................................................................................917

Habit..................................................................................................879

Asparagales .......................................................................................921

Ecological Considerations.................................................................879

Dioscoreales ......................................................................................922

Site of Origin.....................................................................................880

Liliales...............................................................................................922

Pre-cretaceous fossil evidence .............................. 880

Pandanales.........................................................................................923

Sanmiguelia.......................................................................................881

Commelinids .............................................................................923

Furcula ..............................................................................................882

Arecales.............................................................................................923

Problematospermum .........................................................................883

Commelinales ...................................................................................925

Pre-Cretaceous Pollen .......................................................................883

Poales ................................................................................................925

Dispersed Pollen ...............................................................................884

Zingiberales.......................................................................................928

Early angiosperm evidence ..........................................885

Eudicots ........................................................................................929

Pollen ................................................................................................885

Proteales ............................................................................................933

Pollen Evolution ................................................................................889

Ranunculales .....................................................................................940

Evidence from Leaves .......................................................................889 Core eudicots ..........................................................................941 Angiosperm ancestors .....................................................893

Gunnerales ........................................................................................941

Caytoniales ........................................................................................894

Caryophyllales ..................................................................................941

Czekanowskiales ...............................................................................895

Saxifragales .......................................................................................942

Glossopteridales ................................................................................895 Bennettitales......................................................................................895

Rosids ..............................................................................................946

Pentoxylales ......................................................................................895

Myrtales ............................................................................................948

Gigantopteridales ..............................................................................895

Eurosids I (Fabids) .................................................................950

Phylogenetic analyses and

Fabales...............................................................................................950

angiosperm origins ............................................................895

Fagales...............................................................................................953

Selected angiosperm families.....................................897

Malpighiales......................................................................................967 Oxalidales .........................................................................................971

Basal angiosperms ................................................................898

Rosales ..............................................................................................971

Nymphaeales .....................................................................................901 Austrobaileyales ................................................................................902

Eurosids II (Malvids) ...........................................................976

Ceratophyllales..................................................................................904

Brassicales.........................................................................................976 Malvales ............................................................................................976

Magnoliids .................................................................................904

Sapindales .........................................................................................977

Canellales ..........................................................................................904 Laurales .............................................................................................906

Asterids..........................................................................................981

Magnoliales .......................................................................................908

Cornales ............................................................................................981

Piperales ............................................................................................915

Ericales..............................................................................................985

(Continued)

873

874

paleobotany: the biology and evolution of fossil plants

Euasterids I (Lamiids).........................................................986

Asterales...........................................................................................990

Garryales ..........................................................................................987

Dipsacales ........................................................................................ 991

Gentianales.......................................................................................987

Cenozoic floras.................................................................... 991

Lamiales ...........................................................................................988 Solanales ..........................................................................................988

Conclusions ............................................................................996

Euasterids II (Campanulids) .........................................988 Apiales .............................................................................................989 Aquifoliales ......................................................................................989

People from a planet without flowers would think we must be mad with joy the whole time to have such things about us. Iris Murdoch The angiosperms or flowering plants (sometimes designated the Magnoliophyta or Anthophyta) represent the dominant plants in most regions of the world today and include some 400–500 families and perhaps as many as 400,000 species. With the possible exception of bacteria and pathogenic fungi, angiosperms are the organisms that most directly affect human existence on Earth today and contain at least 95% of all living vascular plant species. The group includes all crop plants, for example wheat, rice, and corn, that form the basic food supply to the world. Flowering plants are necessary to human survival in the form of wood for building, medicinal uses, fibers for paper, and textiles. They demonstrate tremendous variation in habit, ranging from small stemless, free-floating plants such as duckweed to huge trees such as oaks and beeches. Angiosperms can be found in an extraordinary range of habitats, extending from saltwater to freshwater and from the tropics to polar regions. Structurally, they are adapted to a terrestrial environment, although some have become secondarily aquatic; other forms, such as the cacti, are able to tolerate exceptionally dry desert environments. Some are colorless parasites on other flowering plants. Within the angiosperms is an extensive array of chemicals, called secondary metabolites, that are effective in both communication and as defense agents against phytopathogenic microorganisms and herbivores. Flowering plants are interpreted as a monophyletic group and have historically been subdivided into two major lineages, the monocotyledons and the dicotyledons. In the classification system of Cronquist (1988), these groups were called the Liliopsida and Magnoliopsida. The number of cotyledons represents only one of the structural features that distinguishes these plants. Monocots and dicots (FIG. 22.1) also differ in the arrangement of the primary vascular tissue in their stems, scattered in monocots in an atactostele versus

Figure

22.1 Sassafras

cretaceum

leaf

(Cretaceous).

Bar  2 cm.

forming a cylinder in dicots. If they are woody, dicots have a vascular cambium and monocots do not. Leaf venation in monocots is parallel whereas it is reticulate or net-veined in dicots; monocot sieve tube plastids are of a particular type (subtype 2) containing cuneate protein, whereas dicots exhibit different types (Behnke, 1971). Other characters that may be more difficult to use in distinguishing these groups include: (1) the number of floral parts, three or multiples of three in monocots; two, five, seven, or many multiples thereof

chapter 22

in dicots (FIG. 22.2); (2) pollen with a single elongate furrow in monocots and three furrows or pores in dicots; and (3) adventitious roots in monocots and primary roots developing from a radicle in dicots. Aspects of embryo development have also been used to suggest a separation between monocots and dicots (Juguet, 1989). Today, however, the classification of major lineages of angiosperms is more difficult, even with the availability of molecular and morphological data sets and the application of phylogenetic analysis. Although more than 60,000 species of monocots are interpreted as being monophyletic, the dicots are considered a paraphyletic group that is no longer formally recognized. To put this in the context of geologic time, one estimate, based on analysis of extant taxa, suggests that the divergence between the monocots and dicots occurred during the Late Jurassic–Early Cretaceous (Chaw et al., 2004).

Higher taxa in this chapter (for families, see Appendix 1):

Basal Angiosperms Orders: Nymphaeales, Austrobaileyales, Ceratophyllales Magnoliids Orders: Canellales, Laurales, Magnoliales, Piperales Monocotyledons Orders: Alismatales, Asparagales, Dioscoreales, Liliales, Pandanales Commelinids Orders: Arecales, Commelinales, Poales, Zingiberales Eudicots Orders: Proteales, Ranunculales Core eudicots Orders: Gunnerales, Caryophyllales, Saxifragales Rosids Order: Myrtales Eurosids I (Fabids) Orders: Fabales, Fagales, Malpighiales, Oxalidales, Rosales Eurosids II (Malvids) Orders: Brassicales, Malvales Asterids Order: Cornales Euasterids I (Lamiids) Orders: Garryales, Gentianales, Lamiales, Solanales Euasterids II (Campanulids) Orders: Apiales, Aquifoliales, Asterales, Dipsacales

flowering plants

875

Fossil flower showing perianth parts (Cretaceous). Bar  2 mm. (Courtesy D. L. Dilcher.)

Figure 22.2

There are several synapomorphies of the flowering plants, but many of their characteristics are shared to some degree with members of other groups of plants. Characteristics that are often used to separate angiosperms from other seed plants, especially when considered together, include the following: (1) the presence of flowers (FIG. 22.3), sometimes with associated accessory parts; (2) seeds enclosed by a carpel; (3) specialized conducting elements in the form of vessels in the xylem and sieve tube members in the phloem, although vessels are present in other groups as well; (4) double fertilization, which is also shared with some gymnosperms; (5) bitegmic (two integuments) ovules (possibly shared with some gymnosperms); (6) the presence of endosperm; (7) a reduced, three-nucleate microgametophyte; (8) a reduced female gametophyte (when compared to those in gymnosperms); and (9) the presence of tectate-columellate pollen and presence of pollenkitt (Hesse, 1980); (10) syringyl lignin (but see Weng et al., 2008); (11) and rapid pollen tube growth (Williams, 2008). Tectate-columellate pollen is also present in several fossil gymnosperms. Because of the extraordinary diversity within the group, there are exceptions to all these features, including double fertilization and endosperm formation. For example, a form of double fertilization has been reported in several members of the Gnetales, as well as a few conifers (Friedman, 1990a, b, 1998; Carmichael and Friedman, 1996). As a result, some have suggested that the common ancestor of angiosperms and gymnosperms possessed double fertilization (Friedman, 1998). The product of the second fusion event in the gymnosperms is not endosperm, but rather a transient embryo (Raghavan, 2003). Features such as flowers, stamens with

876

paleobotany: the biology and evolution of fossil plants

Meiosis

Sporophyte Ovary

Flower

Four magaspores Three megaspores degenerate

Megaspore mother cell undergoes meiosis, producing

Three successive divisions of the remaining megaspore nucleus take place

Ovule

Anther Microspore mother cells undergo meiosis, each producing

Young sporophyte

Quartet of microspores

2n

n

Pollen grain Generative nucleus

Seed coat

Germination

Embroyo

Seed

The germinated pollen grain with its tube and two sperma constitutes the Zygote mature male develops gametophyte into embryo

Fertilization

Figure 22.3

3n endosperm tissue which develops from a cell formed when the two polar nuclei and one sperm unite

Pollen tube Sperm

Antipodals Polar nuclei Egg

Ovary Fruit Ovary develops into fruit; ovule develops into seed

Tube nucleus

Pollination Synergids Tube nucleus

Three eight nuclei produced by the three successive divisions of the megaspore nucleus become rearranged in what is now called the embryo sac. It constitutes the female gametophyte

Integuments Micropyle

Idealized angiosperm life history. (From Taylor and Taylor, 1993.)

four microsporangia, ovules with two integuments, and carpels are characters that have high preservation potential. Others, such as double fertilization (Faure and Dumas, 2001), the formation of endosperm (Baroux et al., 2002) and features of the female gametophyte (Battaglia, 1989) are beyond identification in the fossil record at the present time. In spite of this problem, groups regarded as basal angiosperms based on molecular phylogenetics have provided important information about these character states (Friedman et al., 2003; Williams and Friedman, 2004).

Angiosperm origins The origin of the flowering plants continues to be one of the major unsolved problems in paleobiology, just as Darwin recognized in 1859 when he called the sudden appearance of angiosperms an “abominable mystery” (Davies et al.,

2004; Crepet, 2000). Much has been written and debated about angiosperm ancestors (Takhtajan, 1969, 1987), migration routes, evolution of features, and the subsequent evolution of modern taxa. Although molecular phylogenetic analyses have proposed extant groups that are sister to all other angiosperms, precisely when the divergence between angiosperms and gymnosperms took place continues to remain unresolved. The inclusion of reliable fossil evidence in phylogenetic hypotheses to assess timing in major angiosperm clades is a critical component of determining the age of angiosperm groups (Crepet et al., 2004), and it is becoming increasingly clear that the methodology and choice of fossils can greatly alter the results (Anderson et al., 2005). Using molecular clock assumptions, angiosperms predate the first fossil evidence of the group in the Early Cretaceous. The continued discovery of new fossils that further expands available information about the diversity of types in the Early Cretaceous (e.g. Friis et al., 2000a, b)

chapter 22

also points to an earlier origin for the group. Fossil evidence, especially of floral structures (Takahashi et al., 1999) and other reproductive structures, indicates that during the Cretaceous there were several phases of angiosperm radiation that can be recognized based on floral morphology, pollen diversity, and the composition of the floras (Friis et al., 2006a). In addition, age estimations based on molecular sequence data combined with biogeographic data represent an important tool that can be used to suggest dispersal patterns (Renner, 2005a). Despite all of these tools there is still no unambiguous fossil that can be used to illustrate a pre-Cretaceous angiosperm. Although paleobotanists have struggled to identify the gymnosperm clade or clades from which the flowering plants arose, it might be argued that, when it comes to identifying the group or groups of angiosperm precursors, little or no progress has been made. Others take a perhaps less pessimistic viewpoint, suggesting that some Mesozoic and even Paleozoic plants represent possible flowering plant ancestors (Doyle, 2006). In recent years there have been several phylogenetic analyses of seed plants that have incorporated increasingly larger sets of combined molecular and morphological data. In some instances fossils have been used to calibrate these phylogenies (Bremer et al., 2004), whereas in others they have not. Discussions continue as newer techniques and approaches are utilized, and new fossil discoveries are incorporated into the phylogenetic analyses. Others have focused on the evolution of specialized angiosperm features, such as the closed carpel or bitegumented ovule, and have developed hypotheses that rely on fossils to suggest transitional morphological stages in the evolution of these structures (Doyle, 1994; Frohlich, 2006). Another approach is to use flowering plant structures from an assemblage of presumed gymnospermous ancestors in the reconstruction of an early flowering plant progenitor (Retallack and Dilcher, 1981a). The reader is referred to the paper by Bateman et al. (2006) that discusses a rational approach to understanding some of the questions relating to the origin of angiosperms. Origin of the Flower

Central to any discussion of the evolution of the flowering plants is a definition of precisely what constitutes an angiosperm and what features define the group. As noted earlier, both questions remain difficult problems that do not have simple solutions. They require determining homologies among organs in different seed plant groups while attempting to establish a rational basis for classifying some characters as synapomorphies. Several of the traditional angiosperm characteristics have already been noted. Many of these, at least in

flowering plants

877

part, overlap and are found in other groups of vascular plants. For example, vessels are known in some gymnosperms and vascular cryptogams, whereas several gymnosperms have superficially enclosed ovules at certain stages of development. Other angiosperm features, such as endosperm formation, a character once thought to represent the only unambiguous feature of the group, by its very nature may be impossible to distinguish in the fossil record. This feature is even more difficult to resolve in the fossil record since in some angiosperms the endosperm is cellular, and thus in a fossil seed, might look quite similar to a cellular megagametophyte of a gymnosperm. The character becomes more problematic since double fertilization is absent in certain angiosperms, such as the seasonal aquatics (Podostemaceae) (Raghavan, 2003). Moreover, in certain basal groups of extant angiosperms, nothing is known about double fertilization or endosperm formation (Friedman and Floyd, 2001). Historically there have been numerous theories advanced regarding the origin of the angiosperms in general and more specifically, the origin of the flower (Meyen, 1988a). Some of these obviously pre-date research on certain groups of gymnosperms and the ever-changing ideas about homologies between structures. Many of these are summarized in Angiosperm Origins: Morphological and Ecological Aspects (Krassilov, 1997). A few are discussed below. PSEUDANTHIAL THEORY It is perhaps not surprising that the earliest theories about angiosperm origins focused on the origin of the flower (Dilcher, 2001). One such idea, termed the pseudanthial theory, regards the flower as homologous with the unisexual reproductive organ of a gymnosperm, a concept that was initially proposed by Eichler (1876). According to the proponents of this theory, primitive flowers were small, windpollinated, unisexual, simple, and derived from branching systems that became clustered together. These flowers are often aggregated on elongate axes and possess a single carpel that encloses an ovule with a single integument. Continued coalescence would result in the evolution of terminal carpels subtended by stamens, and eventually in the modification of leaves to form sepals and petals. Flowers like those in the Fagales, Juglandales, and Myricales have been used as examples of primitive angiosperm flowers according to the pseudanthial theory. A modification of this theory, termed the anthophyte hypothesis, suggested that the flowering plants were most closely related to the Gnetales, together with the fossil orders Bennettitales and Pentoxylales (Crane, 1985a; Crane et al., 1995), and later the Caytoniales (Doyle and Donoghue,

878

paleobotany: the biology and evolution of fossil plants

1992). In another cladistic analysis, the angiosperms were found to be nested within the Gnetales (Nixon et al., 1994). As additional genes have been incorporated into molecular phylogenetic analyses of seed plants, several analyses have the Gnetales nested within the conifers, and not the closest living relative to the angiosperms (e.g., Donoghue and Doyle, 2000; Burleigh and Mathews, 2004). EUANTHIAL THEORY The second theory, and the one that has historically received the greatest amount of attention, is termed the euanthial theory. It was also proposed in the late 1800s (see Endress, 1993) and has subsequently been modified as a result of the addition of fossils (Arber and Parkin, 1907; Endress, 1993). According to this idea, the angiosperm flower is interpreted as being homologous with a bisexual strobilus containing numerous, helically arranged carpels and fleshy pollen-bearing parts. Further, the perianth parts in the primitive flower type are conspicuous, and the pollination syndrome is thought to have involved beetles. One group of fossils that was used to support this theory was the Bennettitales, because early ideas suggested that the cones opened once the ovules were developed (Chapter 17). Proponents pointed to Magnolia and a number of related genera as examples of this flower type. Today we know that the flower is a highly plastic structure that has changed multiple times (Endress, 1987, 1990, 2001), even within the same taxonomic group. This means that the symmetry of some fossil flowers may represent reversals and may demonstrate innovations that we do not yet understand within the context of modern analogs (Citerne et al., 2006). Moreover, the genetics involved in flower symmetry and the control of petal and stamen development have also underscored that certain structures evolved multiple times (Kramer and Irish, 2000; Hernández-Hernández et al., 2007). The euanthial theory suggested that the primitive angiosperm had vesselless wood, leaflike stamens, monosulcate pollen, partially closed carpels, and numerous, free, helically arranged floral parts. Accordingly, the most primitive angiosperms belonged to the magnoliids. Historically there have been two major ideas about the origin of the carpel. In the phyllosporous or megasporophyll theory, ovules are borne on a leaf that ultimately becomes folded to enclose the seeds. This differs from the stachyosporous concept, which views the carpel as evolving from a bract that subtends a branch bearing ovules. Using information from both fossils and extant plants, the ancestral carpel is interpreted as ascidiate with a marginal stigmatic surface and reproductive axes containing many flowers, each with a few carpels and

ovules, represent the primitive condition (Taylor and Kirchner, 1996). The evolution of the carpel has also been sought in the Paleozoic, based on the cupule structure found in a number of seed ferns. One of these is Anasperma, an anatropously borne ovule arising on the inner surface of a cupule (Long, 1966). MICROSPORANGIAL THEORIES One of the theories that used information from developmental genetics to explain the evolution of the flower is termed the mostly male theory (Frohlich and Parker, 2000). According to this hypothesis, the gymnosperm ancestor possessed a pollen organ of helically arranged microsporophylls. In this evolutionary scenario the carpel wall is derived from a microsporophyll—not a megasporophyll—with the ovule(s) arising ectopically. Some fossil evidence has been used to support this theory (Frohlich, 2006). Other ideas based on developmental genetics suggest that the origin of a bisexual flower is the result of certain genes that cause the tips of pollen cones to become female, and as a result of certain cumulative gene effects, later give rise to perianth parts (Baum and Hileman, 2006). Subsequent research postulated that the diversity of floral structures seen in extant flowering plants is the result of changing the expression domains of floral homeotic genes that alter the identity of whorls of floral parts (Cronk, 2001; Theissen and Melzer, 2007). These developmental studies are fundamental to answering several questions about the origin of the flower (Tucker, 1999) because they critically examine organ homologies (e.g., carpel) and even floral symmetry, and can be used to compare these phenotypic expressions with those seen in fossils. It will be interesting to see how further developments in molecular genetics relate to the origin of angiosperms. Potentially important topics include the role of MADS-box genes in floral development (Wu et al., 2007) and how such changes are evaluated within systematically defined groups of flowering plants and gymnosperms, as well as whether or not these genes are in fact homologous in both groups (Sundström et al., 1999). This will be a critical area of investigation because several studies suggest that floral parts can easily be modified. When these changes are considered in the context of millions of years of seed plant evolution, it makes defining a flower a very difficult – perhaps impossible – problem to solve (Zanis et al., 2003). TRANSITIONAL–COMBINATIONAL THEORY The most recent theory for the origin of angiosperms is called the transitional–combinational theory (Stuessy, 2004). This hypothesis views various floral parts as evolving at different rates, initially carpels, then double fertilization, and finally the flower. It is only when these components are in place that

chapter 22

flowering plants

879

angiospermy is achieved, according to this theory. Krassilov’s (1997) cyclic angiospermization concept shares a number of features in common with the transitional–combinational theory. Regardless of when features like carpels and double fertilization appear, however, they all must evolve. Although this theory delays the definition of an angiosperm until all selected components are in place, it still requires the transformation of new organs from structures present in some progenitor gymnospermous group or groups. It should be obvious that the definition of a flower, like the concept of angiospermy, is still far from settled (Frohlich and Chase, 2007), and the homology of organs will require far more detailed attention than either molecules or morphology alone can currently provide (Bateman et al., 2006). As will be seen later in this chapter, there are multiple sources of extant and fossil data that can be used to support most of these theories, whether they deal with the origin of the flower, the carpel, or any other feature of the angiosperms, and currently there is no single line of evidence that specifically supports one theory over the other. Habit

In addition to theories relating to the origin of the flower, there are several ideas that speculate on the habit of the earliest flowering plants (Meeuse, 1967). One idea suggests that they had small flowers with a reduced number of stamens and carpels (D. Taylor and Hickey, 1996); this theory, termed the herbaceous origin hypothesis was based on an interesting fossil angiosperm from the Aptian Korumburra Group of Australia. The specimen consists of two alternately arranged leaves attached to a small axis with an inflorescence in the axil of each leaf (FIG. 22.4). The leaf is 3 cm long, slightly asymmetrical, and unlobed. Venation is weakly pinnate, with random higher-order veins. Each inflorescence consists of two axial bracteoles subtending a small ovary (FIG. 22.5). Vegetative and reproductive characters suggest affinities among several members of the magnoliids, as well as within the basal monocots (D. Taylor and Hickey, 1990). If this early angiosperm were in fact herbaceous, it would help to explain why angiospermous woods have not been reported from deposits older than the Early Cretaceous. A slightly different view sees the earliest angiosperms as small weedy shrubs that gave rise to both herbaceous and arborescent forms. Others have variously championed the idea that they were xeric shrubs and most recently, aquatic herbs (G. Sun et al., 2002). Ecological Considerations

Another aspect of the earliest angiosperms concerns their hypothetical physiology and ecology (Feild and Arens, 2005, 2007).

Figure 22.4 Cretaceous angiosperm with two attached leaves,

each subtending an inflorescence. Bar  5 mm. (Courtesy D. W. Taylor.)

In this context, Feild et al. (2003) examined the extant basal angiosperm Austrobaileya and, based on certain physiological parameters, suggested that the earliest angiosperms were woody plants that inhabited wet and dark (low light), disturbed habitats (Feild et al., 2004). The concept of early angiosperms inhabiting disturbed habitats, such as riverine environments, is an old one, (Retallack and Dilcher, 1986) and is supported by paleoecological studies of fossil angiosperm floras (Wing and Boucher, 1998). Some of the ecological and life history strategies of angiosperms, such as a shorter reproductive cycle and the ability to move water and solutes more rapidly in specialized cells (i.e., vessels and sieve tubes), would have represented adaptive advantages in disturbed habitats. These factors have been hypothesized as one of the driving forces that gave flowering plants a selective advantage over ferns and gymnosperms (Stebbins, 1965; Lupia et al., 2000). Increased CO2 levels have also been suggested as a driver in the origin

880

paleobotany: the biology and evolution of fossil plants

types of environments (Coiffard et al., 2007). Certain types of evolutionary adaptations, when coupled with changing major global climatic regimes during the Mesozoic, are believed to have been important evolutionary factors leading to the rapid angiosperm diversification that is recorded in the Early Cretaceous fossil record (Heimhofer et al., 2005). Site of Origin

Figure 22.5 Suggested reconstruction of angiosperm from Aptian sediments (FIG. 22.4) consisting of two alternately arranged leaves and an inflorescence in the axil of each leaf. (Courtesy L. J. Hickey.)

of angiosperms (Barrett and Willis, 2001). Another interrelated factor in the early evolution of angiosperms concerns the potential competition between gymnosperm and angiosperm seedlings in various habitat settings. Termed the slow seedling hypothesis, Bond (1989) hypothesized that gymnosperm seedlings would not have been able to match the faster growth rates of flowering plants. Although limited, there are some quantitative data that support Bond’s hypothesis (Becker, 2000). Using parsimony analysis Coiffard et al. (2006) suggested that during the Albian–Cenomanian in Europe, gymnosperms dominated brackish-water environments, whereas angiosperms were most common in freshwater settings. From this point onward flowering plants continued to expand their ecological range and by the Cenomanian were present in all

Another interesting problem in angiosperm paleobotany has been concerned with defining the geographic region(s) where the angiosperms are believed to have originated (D. Taylor, 1990). Although at one time the angiosperms were believed to have originated in the Arctic (boreal origin), the more widely held view today is that they arose in the tropics and spread poleward (Axelrod, 1959). This view depends on the assumption that the site of greatest modern diversity represents the site of origin of the group. This theory has been supported by the high percentage of so-called primitive modern taxa in the tropics of the southwestern Pacific and southeast Asia, and by a number of fossil records. For example, using pollen data, the first appearance of tricolpate pollen occurs in continually younger rocks as one proceeds from the tropics to higher latitudes on both sides of the equator (Brenner, 1976; Lupia et al., 1999). The fossil leaf record also supports a tropical origin and subsequent poleward radiation (Crane and Lidgard, 1989; Lidgard and Crane, 1990). Among the more important lines of evidence is the diversity of angiosperm monosulcate pollen. Based on these data, it appears that the Atlantic–South American–African rift zone represent the most probable sites for the initial radiation of the group (Doyle, 1984; Zavada, 2007). Palynofloras from this region includes forms indicative of semiarid conditions and thus provide support for Stebbins’ (1974) suggestion that the earliest angiosperms originated and diversified in semiarid conditions. Bisaccate grain types are characteristically absent from this province, with gymnosperms represented by various araucarians and members of the Podocarpaceae. A high percentage of the palynomorphs from this province include monosulcate cycadophyte-type grains; this province also has provided the oldest tricolpates. Other pollen records and the suggestion that early angiosperm pollen grains may have been inaperturate, however, compromise the use of pollen in determining the site of origin (Brenner, 1996).

Pre-cretaceous fossil evidence The rapid radiation of the angiosperms in the Early Cretaceous and the apparently meager fossil record of the group prior to that time have been used as the basis for

chapter 22

suggesting that the flowering plants evolved in the Early Cretaceous (Burger, 1990). Assuming that the angiosperms underwent explosive radiation, this suggestion is supported by an ever-increasing database of considerable morphological diversity in fossil flowers from the Lower Cretaceous, many of which appear to represent evidence of apparently derived angiosperm lineages. Others have argued that angiosperms have an earlier origin, perhaps in the late Paleozoic or early Mesozoic. DNA sequence data, together with molecular clock assumptions, have been used to suggest that the origin of the angiosperm crown group is Early–Middle Jurassic (Wikström et al., 2001). The presence of certain types of pollen in the Triassic also appears to support an earlier origin (Cornet and Habib, 1992; Hochuli and Feist-Burkhardt, 2004), as does some phytochemical biomarker evidence (D. Taylor et al., 2006). This type of evidence has served as the principal impetus to search older rocks for fossils that might provide links between gymnosperm ancestors and the flowering plants. Throughout the history of paleobotany there have been a number of early Mesozoic plants believed to represent preCretaceous flowering plants. Some of these fossils have been transferred to gymnospermous groups, whereas others have remained equivocal due to vagaries of preservation, interpretation, or the absence of a suitable number of specimens. Still others have been discounted because they were found in rocks lacking precise stratigraphic resolution. Finally, there are other fossils that have simply not been accepted as angiosperms because they occur in rocks that are too old, preservation is poor, or the interpretation is deemed too radical. In this section we will examine some of these pre-Cretaceous plant fossils that have historically been linked with the angiosperms.

flowering plants

881

Figure 22.6 Portion of a leaf of Sanmiguelia lewisii (Triassic).

Bar  1 cm. (From Brown, 1956.)

Sanmiguelia

One of the most widely cited fossils is Sanmiguelia (FIG. 22.6) a genus established for palmlike leaves collected from the Triassic of Colorado by Roland Brown (1956) (FIG. 22.7). Since the initial description, additional specimens have been discovered in situ which suggest that the plants are 60 cm tall and consist of helically arranged leaves attached to a conical stem (Tidwell et al., 1977) (FIG. 22.8). Leaves are broadly elliptical with an acute apex and clasping base. The lamina is plicate, with four orders of delicate parallel veins extending to the apex; occasional cross veins are present (Cornet, 1986). Stomata are abundant and actinocytic. The discovery of additional specimens of S. lewisii from the Upper Triassic (upper Carnian) of western Texas provides a more complete picture of this interesting Triassic plant (Cornet, 1989a). The stems are woody and contain a zone of secondary xylem made up of tracheids. Of

Figure 22.7 Roland W. Brown. (Courtesy H. N. Andrews.)

particular importance is the discovery of reproductive organs in the same beds which are believed to belong to Sanmiguelia (Cornet, 1986). The pollen organs, Synangispadixis, are preserved as compressions and consist of an axis that bears

882

paleobotany: the biology and evolution of fossil plants

Figure 22.8 Suggested reconstruction of Sanmiguelia lewisii. (From Taylor and Taylor, 1993.)

numerous, helically arranged, sessile microsporophylls (FIG. 22.9). Each highly reduced microsporophyll consists of a pair of elongate, biloculate pollen sacs that dehisced longitudinally. The in situ pollen grains are elongate, with a single sulcus and a granular sporoderm (Cornet, 1989a). The ovulate structures, called Axelrodia, are produced at the end of secondary branches and consist of an axis bearing numerous structures interpreted as carpels; the entire unit is subtended by several bracts (FIG. 22.10). The putative carpels are variously shaped and possess a stigmalike apex and ventral suture (Cornet, 1989a). Anatropous ovules were produced in pairs and are thought to have had a double integument. Pollination is believed to be biotic via insects. Nemececkigone is the name applied to isolated seeds (1.4 cm long) that occur in the same rocks as Sanmiguelia (Cornet, 1986). One inherent difficulty in working with fossil plants involves the interpretation of various organs, especially in instances where preservation and morphology make it difficult to understand homologies. Prior to the discovery of associated reproductive organs, Sanmiguelia was thought to have affinities with monocots, in particular Veratrum (Melanthiaceae) (Tidwell et al., 1977). Others have challenged this interpretation based on certain features of the leaves and suggested that Sanmiguelia is a cycadophyte leaf (Read and Hickey, 1972). Based on his studies, Cornet (1989a) considered Sanmiguelia to be a primitive angiosperm that combines features of both monocots and dicots. Does this interesting suite of fossils represent evidence of a Triassic plant with certain angiosperm features? Perhaps.

Figure 22.9 Suggested reconstruction of Synangispadixis tidwellii. (From Taylor and Taylor, 1993.)

The collection of additional specimens and perhaps the utilization of new techniques, however, will help to decipher the affinities of this interesting Triassic plant. Furcula

Another pre-Cretaceous plant that once captured the imagination of paleobotanists as a possible early angiosperm is Furcula (FIG. 22.11). The genus is known from the Triassic (Rhaetian) of Greenland and consists of lanceolate-shaped leaves that usually dichotomize in the middle (Harris, 1932b). Some leaves are up to 15 cm long and have entire margins. Venation consists of a prominent midrib from which are produced lateral veins about every 2 mm; between these veins are smaller veinlets that anastomose to form a reticulate pattern similar to that seen in modern dicots (FIG. 22.12). Cuticle preparations of F. granulifera show a rather simple epidermal

chapter 22

flowering plants

883

Figure 22.10 Suggested reconstruction of the ovulate flower Axelrodia burgeri. (From Taylor and Taylor, 1993.)

pattern of slightly sunken stomata surrounded by a ring of six subsidiary cells. Furcula possesses syndetocheilic stomata, which are typical of a number of plant groups, including angiosperms. Although Furcula has been regarded as a preCretaceous angiosperm by some, others believe that features of the venation pattern more strongly indicate affinities with some seed fern groups (Scott et al., 1960). Another dicot-like leaf showing some similarity to Furcula is Pannaulika, a fossil recovered from the Upper Triassic of the Newark Supergroup in eastern North America (Cornet, 1993). The small leaf has an entire margin with pinnate venation and four orders of veins. Interestingly, the leaf occurs in rocks that also contain small, reticulate, angiosperm-like pollen. Leaves with reticulate venation, however, are relatively common in the fossil record and do not necessarily indicate angiosperm affinities. The reader is referred to the summary of these by Trivett and Pigg (1996). Problematospermum

Problematospermum is a Jurassic fossil that was initially described as a gymnosperm. Specimens of P. ovale are small (2.5 cm long), ovoid, and have a pappus-like tuft at one end (Turutanova-Ketova, 1930). A few pollen grains have been found on the pappus tube. Morphologically, these seedlike specimens resemble the achene and pappus of certain Asteraceae (Krassilov, 1973b); however, they also have been compared to certain Paleozoic pteridosperm seeds that are known to have possessed integumentary or nucellar-borne processes like those in Gnetopsis (FIG. 14.74).

Figure 22.11 Leaf of Furcula granulifera. (From Taylor and

Taylor, 1993.)

Pre-Cretaceous Pollen

In addition to megafossils, pollen grains have been used to suggest the existence of angiosperms before the Cretaceous. One pollen type that was believed to have been produced by a pre-Cretaceous angiosperm is Eucommiidites, a grain originally described from Jurassic rocks in Sweden (Erdtman, 1948). Subsequent reports have extended the taxon from the Triassic well into the Cretaceous (Scheuring, 1970). The smooth-walled grains of E. troedssonii are boat shaped and range from 30–40 μm long. On one surface is a prominent furrow or colpus. What was interpreted as a tricolpate organization in Eucommiidites was used to suggest affinities with the modern dicot genus Eucommia, a grain type in which the colpi are of slightly unequal size. Subsequent studies of Eucommiidites have substantiated that the grains

884

paleobotany: the biology and evolution of fossil plants

Figure 22.12 Portion of a Furcula granulifera leaf showing venation. Bar  250 μm. (From Harris, 1932b.)

possess a single colpus, with the remaining “colpi” formed by a proximally positioned zonisulculus that surrounds the bilaterally symmetrical grain. In addition, there are now a number of reports (Brenner, 1967; Reymanówna, 1968; Friis and Pedersen, 1996; Kvacˇek and Pacltová, 2001) that have documented the presence of Eucommiidites-type pollen grains in the micropyle and pollen chamber of several gymnospermous ovules (e.g., Spermatites, Allicospermum, Eucommiitheca). Pollen of Eucommiidites has also been found in the Jurassic cycadalean cone Hastystrobus (Van Konijnenburg-Van Cittert, 1971). Pedersen et al. (1989b) described a Cretaceous fructification containing pollen of the Eucommiidites type. The ultrastructure of Eucommiidites pollen also supports the gymnospermous affinities of this pollen type (Doyle et al., 1975). The presence of Eucommiidites pollen in several different types of seeds suggests the existence of a diverse group of plants, all producing pollen of the same basic morphological type. For example, seeds of Erdtmanispermum from the Lower Cretaceous of Denmark are included in Erdtmanithecaceae (Erdtmanithecales). The seeds are ellipsoid and slightly triangular in cross section, with an elongate micropylar tube (Pedersen et al., 1989b) (FIG. 22.13). The megaspore is 3 μm thick and surrounded by several cutinized layers

Figure 22.13 Erdtmanispermum balticum seed (Cretaceous).

Bar  250 μm. (From Pedersen et al., 1989b.)

interpreted as the nucellus. Recently, another species, E. juncalense, has been reported from the Cretaceous of Portugal further expanding the geographic distribution of the Eucommiidites pollen-producing group of plants (Mendes et al., 2008). Several Eucommiidites pollen-containing structures, together with other reproductive organs, are included in the Erdtmanithecaceae, a family characterized by radial, peltate microsporangia aggregated in spikes, seeds possessing a long micropylar tube like that in certain gnetophytes, and pollen of the Eucommiidites type (Friis and Pedersen, 1996). More recent studies, however, suggest that some of these seeds may be more closely related to the Gnetales and Bennettitales (Friis et al., 2007). Bayeritheca (FIG. 22.14) is a Cretaceous pollen organ that also produced Eucommiidites grains (Kvacˇek and Pacltová, 2001). Dispersed Pollen

The fossil record of dispersed pollen types that have been linked to the flowering plants is even older than that of in situ pollen.

chapter 22

flowering plants

Figure 22.15

Figure 22.14 Bayeritheca hughesii (Cretaceous). (Courtesy

J. Kvacˇ ek.)

For example, there is a rich assemblage of Late Triassic (early– middle Carnian) palynomorphs from the Richmond Basin of Virginia that includes several angiosperm-like grain types with a reticulate–columellate exine (Cornet, 1989b). Several of these are included in the Crinopolles group, an assemblage of grains with two or more sulci located on the distal and equatorial sides of the grains. These grains are dimorphically sculptured and lack apertures on the proximal surface. Several Mesozoic polyplicate pollen types, for example, Welwitschiapollenites and Ephedripites, have been regarded as being produced by gnetophytes. Others, however, suggested that they may represent another lineage of angiosperm pollen that parallels the monosulcate types (Cornet, 1989b), whereas Hughes (1994) (FIG. 22.15) believed that grains of this type may represent a group of unknown origin. Hesse and Zetter (2007) indicated that some Ephedripites grains are similar to pollen of Spathiphyllum, an extant pantropical member of the Araceae.

885

Norman F. Hughes.

Two additional Triassic (Carnian) pollen types also demonstrate features that could be regarded as angiospermous. One of these is Equisetosporites, an ellipsoidal grain ornamented by diagonally disposed, thickened bands that separate harmomegathic furrows (Pocock and Vasanthy, 1988). A slightly different sporoderm construction is present in Cornetipollis (FIG. 22.16) (Pocock and Vasanthy, 1988). The bands in C. reticulata are not rotated and the ultrastructure of the pollen wall shows the presence of distinctly larger columellae. Both pollen types possess some angiospermous features, suggesting that perhaps the flowering plants or the group(s) from which they evolved may be as old as the Triassic (Vasanthy et al., 2004). To date the oldest Triassic pollen suggesting affinities with the flowering plants comes from Middle Triassic rocks of the Norwegian Arctic (Hochuli and Feist-Burkhardt, 2004). The grains are similar to many Early Cretaceous forms and include monosulcates and perhaps triaperturates; interestingly all possess a reticulate exine ornamentation. The grains were recovered from shallow wells and thus there are no megafossils with which the grain types can be equated. In spite of this, the diversity of angiosperm features on these grains is striking.

Early angiosperm evidence Pollen

What are interpreted as the earliest Cretaceous angiosperm pollen grains are reported from the Hauterivian (Lower

886

paleobotany: the biology and evolution of fossil plants

Figure 22.17 Retimonocolpites pollen grain (Cretaceous). Figure 22.16 Cornetipollis reticulata pollen grain (Triassic).

Bar  10 μm. (From Pocock and Vasanthy, 1988.)

Cretaceous) of England (Hughes et al., 1991). Other small (20–35 μm), round pollen grains from the Hauterivian of Israel are also considered to represent some of the earliest unequivocal angiosperm pollen grains (Brenner and Crepet, 1987). The pollen wall is tectate–semitectate, and there is a single, poorly defined sulcus. Another pollen type with angiospermous features is Retimonocolpites, a small (14–25 μm) monosulcate grain from the Aptian (FIG. 22.17). The surface is coarsely reticulate, a feature that was initially responsible in part for the initial consideration of this grain as a monolete fern spore. Retimonocolpites is described as being semitectate with the reticulum directly attached to the foot layer. In Stellatopollis, another early Cretaceous monosulcate pollen type, the exine is tectate to semitectate (Doyle, 1975) (FIG. 22.18). Supratectal projections occur on the reticulum (FIG. 22.19) in some species (Cornet and Habib, 1992). The occurrence of a columellate exine has been used to suggest that grains of this type may have angiospermous affinities (FIG. 22.20).

Bar  5 μm. (Courtesy. M. Zavada.)

22.18 Stellatopollis pollen Bar  5 μm. (Courtesy M. Zavada.)

Figure

grain

(Cretaceous).

chapter 22

flowering plants

887

Figure 22.19 Crotonoid sculpture of Stellatopollis exine

(Cretaceous). Bar  3 μm. (Courtesy B. Cornet.)

Figure 22.21 Pollen grain of Afropollis jardinus (Cretaceous). Bar  15 μm. (From Doyle et al., 1982.)

Figure 22.20 Ultrathin section of Stellatopollis pollen grain wall (Cretaceous). Bar  1 μm. (Courtesy M. Zavada.)

Liliacidites (Barremian–early Aptian) is a grain type suggested as representing an early monocotyledonous pollen type (Zavada, 1984a; Brenner, 1996). These grains are described as both monosulcate and inaperturate. They average 36 μm long and possess a reticulate exine, with the ornament more delicate at the opposite ends of the grain (Walker and Walker, 1984). Dichastopollenites is an Albian operculate grain type that has been related to extant pollen found in the Nymphaeaceae (Schrank and Mahmoud, 2000). Specimens of Clavatipollenites have been reported from the Lower Cretaceous (Barremian–Cenomanian) of several geographic areas including England, Argentina, the United States, Africa, Portugal, and Israel and, because of their morphologic variability (Walker and Walker, 1984), are interpreted as representing the pollen of several different taxa. They are a common component of many early angiosperm assemblages (Heimhofer et al., 2007), are generally 25 μm long, and contain a single sulcus or are nearly inaperturate. In a

few morphotypes, a Y-shaped aperture termed a trichotomosulcus replaces the single furrow. The structure of the exine in Clavatipollenites is variable, but in some forms the radially disposed rods of sporopollenin are fused at their summits to form a tectate-columellate exine (Brenner, 1996). Although several non-angiospermous pollen types morphologically resemble Clavatipollenites, such as those found in Ginkgo, Cycadeoidea, and certain seed ferns, the grains also have several characters that are similar to those found in modern chloranthaceous pollen (Pedersen et al., 1991). Ultrastructural studies (Chlonova and Surova, 1988) show similarities between Clavatipollenites and Ascarina (Chloranthaceae). Another reticulate pollen type is Afropollis (Doyle et al., 1982). The genus ranges from the Barremian to the Cenomanian and includes spherical, inaperturate grains with a zonisulculus that extends around the equator or may be slightly displaced toward a pole (Hughes and McDougall, 1990). The sculpture is an open network of ridges (muri) that are separated from the inner portion of the exine over much of the grain surface (FIGS. 22.21, 22.22). Fine structure of the pollen wall includes a thick nexine and fine lamellae in what is interpreted as the endexine (Doyle et al., 1990a). The coarse sculpture on many of these early pollen types has been used to suggest these grains were adapted to insect pollination. An alternative hypothesis (Zavada, 1984b; Zavada and Taylor, 1986b) suggests that the reticulat-perforate exine of many early angiosperm pollen types demonstrates the existence of a sporophytic self-incompatibility system, since

888

paleobotany: the biology and evolution of fossil plants

Figure 22.22 Ornamentation of Afropollis wall showing

coarse muri (Cretaceous). Bar  5 μm. (Courtesy N. F. Hughes.)

Figure 22.24 Tricolpites pollen grain found in the flower Platananthus hueberi (Cretaceous). Bar  2 μm. (From Friis et al., 1988.)

22.23 Several Tricolpites minutus pollen grains (Cretaceous). Bar  15 μm. (From Doyle et al., 1975.)

Figure

there is a high degree of correlation between this system and similar exine structure in a number of modern groups. This hypothesis has been challenged by Gibbs and Ferguson (1987), but is supported by additional data reported by Zavada (1990b). The first distinctly tricolpate pollen, which is diagnostic of the eudicots, appears near the Barremian–Aptian boundary (Brenner, 1976). Grains of this type have been described by the generic names Tricolpites (FIG. 22.23) and Tricolpopollenites. They are small, some 15 μm long, and possess three equidistant colpi. Like many of the monosulcate

types, the exine is semitectate to tectate. A number of pollen grains of this type have been found in small staminate flowers such as Platananthus (FIG. 22.24) (Friis et al., 1988). Ultrastructural features of the pollen walls have been used to distinguish early angiosperm pollen from gymnospermous grains. Some studies have relied on ultrastructural characters as the basis for interpreting pollen-wall features in the exine of gymnosperms and angiosperms (Zavada, 1984a; Vasanthy et al., 1990; Osborn, 1991). For example, Doyle et al. (1975) postulated that the endexine in angiosperm pollen is nonlaminated, whereas in gymnosperms the nexine has distinct lamellae. In Clavatipollenites, however, the nexine is nonlaminated but occurs only beneath the aperture, whereas in some extant angiosperms, for example in the Asteraceae, Campanulaceae, and Lamiaceae, the endexine is lamellate. The columellate exine of angiosperms was once believed to represent a significant difference between angiosperm and gymnosperm pollen, but this character is now known to be unreliable, since a number of both angiosperms and gymnosperms possess what has been termed a granular exine (Walker, 1976). Currently, early angiosperm pollen is now interpreted as being columellate with a thin endexine (Doyle, 2005). Studies of fossil pollen ultrastructure also indicate, however, that the degree of post-depositional compression and other diagenetic factors, including thermal alteration, can greatly affect the wall structure (Zavada, 2007).

chapter 22

Pollen Evolution

Brenner (1996) suggested that there are several evolutionary stages that can be recognized from early angiosperm pollen morphology. The earliest pollen is believed to have been small, circular, and inaperturate with recognition proteins located in the tectate-columellate exine. At the next evolutionary stage, the intine and the endexine become thicker, followed by the development of the sulcus; recognition proteins now occur in the intine. Rapid diversification from the basic monosulcate grain type gives rise to the tricolpate grain morphology seen in the eudicots. Thus, the earliest angiosperm grains were inaperturate and not monosulcate. Especially noteworthy is Brenner’s (1996) suggestion that the tectate-columellate exine was the original exine condition, and therefore not necessarily linked to insect pollination syndromes. Supporting this proposed evolutionary sequence of pollen types is the presence of small, round grain types from the Lower Cretaceous (Valanginian–Hauterivian) Helez Formation of Israel that are inaperturate to weakly monosulcate (Brenner and Bickoff, 1992). Based simply on dispersed pollen, this hypothesis is supported by the reticulate–columellate grains that are increasingly being reported from Triassic rocks. This means that the gymnosperm–angiosperm relationship should not necessarily be sought in plant groups that possess pollen of the monosulcate type. Brenner’s (1976) studies indicate that the oldest tricolpates occur in tropical latitudes and become more common at higher latitudes in progressively younger rocks. Several ideas have been presented to explain the evolution of tricolpate pollen from the more primitive monosulcate condition. One hypothesis suggests that the tricolpate condition evolved from a tricotomosulcate grain (Doyle, 1973). Another point of view is expressed by Brenner (1976), who views the tricolpate condition as evolving from a zonisulculate type (e.g., Eucommiidites) via a change in tetrad geometry. The origin and early diversification of angiosperms based on molecular phylogenetics suggest that the first angiosperms had globose monosulcate rather than boat-shaped pollen and that the exine was columellate with a thin endexine rather than possessing a granular infratectum and lacking an endexine as in many Magnoliales (Doyle, 2005). As a result, it has now been suggested that the alveolate exine structure present in some late Paleozoic and Mesozoic gymnosperms, such as the glossopterids, Caytonia, and some Triassic grain types, may represent the early angiosperm pollen prototype. Moreover, examination of pollen features framed within a functional context (i.e., self-incompatibility) argues that angiosperm origins took place in the Triassic (Zavada, 2007, but see Friis et al., 2006).

flowering plants

889

Although there is still considerable uncertainty regarding the ancestors of the flowering plants, the Early Cretaceous angiosperm pollen types indicate a great deal of taxonomic diversity (Lupia et al., 1999). Such diversity is also seen in a number of dispersed Triassic pollen types that possess angiospermous features (Cornet, 1989b; Brenner, 1996; Zavada, 2007). These include size, shape, exine structure, nature and position of aperture, and ornamentation type. Many of these features have made it possible to recognize major lineages of flowering plants early in the Cretaceous, and the discovery of in situ pollen in charcoalified flowers has allowed assignment of certain grain types with a greater degree of confidence. Thus, although the origin of angiosperms from gymnospermous ancestors continues to remain obscure, the first occurrence of major taxonomic groups can be recognized based on distinctive features of their pollen (Muller, 1984; Lupia et al., 1999). Evidence from Leaves

Among the more common Cretaceous angiosperm fossils are impression and compression leaf remains. During the late nineteenth and early twentieth centuries, many paleobotanists attempted to make direct, taxonomic comparisons between fossil angiosperm foliage types and the leaves of modern flowering plants. Most of these studies used such features as leaf size, shape, and margin and, to a lesser degree, the pattern of venation. Little attempt was made to place the fossil leaves in any developmental pattern that reflected their stage of maturity at the time they were fossilized, or to consider the plasticity in leaf morphology within a taxon and the influence of habitat and even leaf position on the plant. As a consequence, many of these early studies represent “picture matching” attempts to relate the fossil leaves to modern counterparts. When modern analogs were not immediately apparent, a proliferation of new genera and species resulted. The application of systematic principles, together with techniques such as cuticular anatomy and details of venation patterns, indicates that a number of the earlier identifications with modern taxa are incorrect (Hickey, 1973; Dilcher, 1974). This is not to suggest that all the early angiosperm leaf studies were completed without trying to utilize taxonomically useful and reliable characters to describe the taxa. For example, one needs only to examine the contributions of von Ettingshausen (1854a, b, 1856, 1858a, b, 1865) to realize that he was aware of the importance of venation as a systematic character. Other paleobotanists, such as Berry (1916a, b, 1924, 1930) (FIG. 22.25) and Lesquereux (1883, 1891), incorporated some of the terminology developed by von Ettingshausen, but in general, von Ettingshausen’s use

890

paleobotany: the biology and evolution of fossil plants

Figure 22.26 Cleared angiosperm leaf showing multiple

orders of venation (Extant). Bar  5 mm.

Figure 22.25 Edward W. Berry. (Courtesy H. N. Andrews.)

of multiple characters did not gain universal acceptance. It is also important to note that many of the early collections of angiosperm fossil leaves were made under circumstances in which the precise stratigraphic position of the fossils was either not recorded or not known to the collector. As a result, in the absence of detailed stratigraphic control, many of these collections became useless as subsequent investigators attempted to trace angiosperm taxa through time. The consequences of some of these early problems and misidentifications continue today as non-paleobotanists attempt to polarize characters or use various fossil taxa as benchmarks in phylogenetic analyses. During the last four decades there has been unparalleled activity in angiosperm paleobotany, in part reflecting the utilization of new instrumentation, that is, various imaging systems, methodologies (e.g., phylogenetic analyses), and especially the discovery of important new specimens (e.g., charcoalified flowers; Tiffney, 1977). These paleobotanical approaches have been complemented by various molecularbased phylogenies and new statistical techniques. In addition many paleobotanists have extensively incorporated ecological, sedimentological, and biological approaches into the study of fossil flowering plants. Many of these studies have relied heavily on structural and morphological investigations of modern

plants, and a new body of information about the genetic, developmental, and biochemical mechanisms in angiosperms has also contributed to our increased understanding of fossil representatives. One approach involves the definition, development, and classification of leaf characteristics on the basis of leaf architecture (FIG. 22.26). In this method emphasis is placed on features such as shape, venation pattern, configuration of the margin, and presence or absence of structures such as glands or trichomes. As a basis for the development of this classification scheme, Hickey (1973) examined more than 1000 genera representing 135 families of dicots. Particular emphasis was directed at what are termed lower-order (basic) features of the leaf, for example venation pattern, margin, and leaf shape, since these would have an increased probability of preservation in fossilized leaves. Using these characteristics as base-level data, it was possible to carefully examine Early Cretaceous angiosperm leaves and to identify the characteristics that may be regarded as primitive (Hickey and Wolfe, 1975). This approach also attempted for the first time to separate developmental features from those that have systematic significance. Because some leaf features have apparently evolved in parallel, this approach has made it easier for researchers to identify taxa more reliably and to consider their evolution based on fragmentary specimens (see also Leaf Architecture Working Group, 1999). One of the most important and useful features used in these studies has been the analysis of venation patterns. Four distinct levels of venation have been recognized. In leaf

chapter 22

forms of the first rank, secondary and higher-order veins are irregular, and the spaces between the veins (intercostal areas) are irregular in size and shape. In leaves of the next order, secondary veins are more regular, but the tertiary veins and intercostal areas remain irregular. Most modern dicot leaves are represented by third-rank leaves, in which the tertiary veins are regular and the areoles (areas between the tertiary veins) are irregular in size and orientation. In fourth-rank leaves, the areoles are of uniform size and shape. In classifying leaf types according to the leaf architecture scheme, it is important to understand that not all foliar features represent stages in the evolution of leaves. Some represent morphological and anatomical expressions of the ecological niche in which the plant was living or a particular developmental stage leading to an adult form. In recent years the cuticular structure and epidermal anatomy of fossil leaves, including that revealed by scanning electron microscopy and other types of imaging systems, has become a routine and powerful research tool in the identification and analysis of fossil angiosperm leaves. The cuticle expresses a number of important characteristics that are useful in determining relationships at the specific, generic, or familial levels (Upchurch, 1984). These include the thickness and chemical composition of the cuticle, morphology of the epidermal cells, presence of trichomes, glands, and scales, and the number of stomata and organization of the stomatal complex. A precise terminology has been developed to refer to the arrangement of the epidermal cells that surround the guard cells of the stomatal complex (Dilcher, 1974; Upchurch, 1995). Upchurch and Dilcher (1990) listed several major principles utilized in their classification of Cenomanian angiosperms from southeastern Nebraska. The classification was based on foliar architecture and cuticular anatomy and included the use of multiple characters that have systematic significance in extant angiosperms. New taxa were only described based on specimens containing three or four orders of venation. In utilizing cuticular features, it is important to underscore that the total complement of characteristics available in the fossil need to be utilized. Ultrastructural features of fossil cuticle have been examined for various gymnosperms (Guignard et al., 1998; del Fueyo et al., 2006), but to date have not been widely used to study fossil angiosperms. The oldest leaves from the Southern Hemisphere that show distinct angiosperm characters come from the Aptian Baqueró Formation (now Baqueró Group; see Cladera et al., 2002) of Patagonia, Argentina (Romero and Archangelsky, 1986) (FIG. 22.27). The simple leaves are pinnately lobed and 14 cm long. The margin is serrate, with teeth separated

flowering plants

891

Figure 22.27 Sergio Archangelsky.

Figure 22.28 Angiosperm leaf from the Bajo Tigre site,

Patagonia, Argentina (Cretaceous). Bar  2 cm.

by angular sinuses. Venation is of the simple craspedodromous type with random tertiary veins (FIG. 22.28). The majority of foliar features suggest affinities within the eudicots.

892

paleobotany: the biology and evolution of fossil plants

by asymmetrical bases. A high percentage of these leaves exhibit entire margins. In the zones above (IIA, IIB), there are many more tricolpates and monocolpates, both with reticulate exine patterns. This is also the first appearance of tricolporoidate pollen and grains that are more triangular in outline. Leaves from this zone are more variable in shape than those in zone I (FIG. 22.33), and the first palmately veined forms appear here. In many of the zone IIB leaves, primary veins dichotomize to form symmetrical loops along the margin. Both pinnately and palmately lobed forms make their first appearance here. Other leaves suggest the initial appearance of various types of glands. In zones IIC and III, the pollen flora includes greater numbers of tricolporoidate types (FIG. 22.34). Leaves of the platanoid type dominate at this level and the tertiary veins are better developed (FIG. 22.35). In addition, the leaves in these zones often have expanded petiole bases that are suggestive of the abscission zones in deciduous leaves. Studies of this type provide an excellent basis upon which to test hypotheses about the evolution of both pollen and leaf characters during a major radiation of the flowering plants. Within such a framework it becomes possible to compare various features along both a stratigraphic and geographic gradient, and to

IV “Raritan”

III

Albian Patapsco

Cenomanian

In the Northern Hemisphere, the study by Hickey and Doyle (1977) on fossils from the Potomac Group of eastern North America represents a classic study on early angiosperm remains. What makes this paper important is that the authors combined two different data sets, leaves and pollen (Doyle and Hickey, 1976), to examine an extensive sequence of Cretaceous rocks in order to provide greater resolution of the origin and subsequent evolution of flowering plants (FIG. 22.29). Using the biostratigraphic framework (based on palynological zones) established by Brenner (1963), they were able to document a continuum of changes in leaf architecture and venation from the Aptian into the Cenomanian, as well as changes in the palynoflora for this sequence. They divided these assemblages into a series of six zones, I, IIA, IIB, IIC, III, and IV. Angiosperm remains are rather rare in zone I, now interpreted as Albian (Hochuli et al., 2006), with most of the pollen types consisting of monosulcate forms; only in the upper part of the zone were a few tricolpates present. Leaves of this zone are characterized by a lack of differentiation between the lamina and petiole and a disorganized pattern of venation (FIGS. 22.30, 22.31, 22.32). They are generally elliptical in outline with later-appearing forms characterized

IIC

IIB

Aption Patuxent Arundel

IIA

I

Figure 22.29 Suggested trends in the evolution of pollen and leaves from the Potomac Group (Cretaceous) of North America. (From Hickey and Doyle, 1977.)

chapter 22

Figure 22.30 Rogersia angustifolia (zone I). (From Hickey and

flowering plants

893

Figure 22.31 Ficophyllum crassinerve (zone I). Bar  3 cm. (Courtesy L. J. Hickey.)

Doyle, 1977.)

access a mosaic of characters that may reflect more subtle ecological adaptations.

Angiosperm ancestors Almost every group of vascular plants has at one time or the other been implicated as the progenitor to the flowering plants, including ferns (Ophioglossales; Kato, 1990). As additional fossil and molecular evidence has been assembled, some of these suggestions have lost support, whereas others have gained proponents. Historically, as is the case today, the group that has received the greatest amount of attention is the Mesozoic seed ferns, with the Caytoniales often suggested as a likely ancestral group based on characters of the cupules (see Chapter 15). The following is a brief summary of some of the salient points that have been considered when discussing the gymnosperm ancestry of the flowering plants. The

Figure 22.32 Vitaphyllum multifidum (zone I). (From Hickey

and Doyle, 1977.)

894

paleobotany: the biology and evolution of fossil plants

Figure 22.35 Araliopsoides cretacea (zones IIC and III).

(From Hickey and Doyle, 1977.)

Figure 22.33 Sapindopsis (zone IIB). (From Hickey and

Doyle, 1977.)

reader is referred to specific chapters in the book for a more in-depth treatment of each of these gymnosperm groups. Caytoniales

Figure 22.34 Tricolporopollenites kruschii pollen grain (Miocene). Bar  10 μm. (Courtesy M. S. Zavada.)

These plants, which were initially described by H. H. Thomas (1925) as possible Jurassic angiosperms, have been extensively studied and are now regarded as seed ferns by paleobotanists. The intriguing feature that has drawn many to this group is the nature of the seed-bearing cupules, structures which contain several ovules enclosed by a thinly cutinized inner sac. Proponents of the Caytoniales as angiosperm ancestors point to the almost-sealed cupule as an indication of one way in which the carpel may have evolved. This hypothesis does not explain the origin of the second integument of angiosperms, however, and it has been suggested that the second integument is homologous with the cupule of Caytonia, and that the carpel has its origin from the rachis or a portion of the rachis that bears the cupules in Caytonia (Doyle, 1978).

chapter 22

There are some other interesting functional aspects of Caytonia that may have a bearing on the evolution of angiospermy. The outer edge or lip of the cupule was connected to each ovule by a cutinized tube subdivided into a number of channels, corresponding to the number of ovules. It has been suggested that pollen grains were moved to the micropyles via these channels, perhaps with the aid of a pollination droplet (Chapter 15). In the transition from gymnosperm to angiosperm, these tubes might have functioned as “style” mechanisms prior to the evolution of rapidly growing pollen tubes. In this scenario the group would have had some type of compatibility system, perhaps controlled by the gametophyte. Other characters that have been used to suggest an angiospermous level of evolution within the Caytoniales is the reticulate pattern of venation seen in the leaves of Sagenopteris and the loculate microsporangia of Caytonanthus that superficially resemble anthers. Czekanowskiales

This group of Jurassic seed plants has linear leaves attached to short shoots. The Czekanowskiales, like the Caytoniales, possess cupules, sometimes termed capsules, with a papillate flange that superficially resembles a stigmatic surface. Some specimens suggest that pollen grains landed directly on the micropyle of the ovule, indicating that the capsule was not entirely closed. At least with reference to carpel closure, these Jurassic gymnosperms parallel the condition seen in several angiosperms in which the carpel is not completely closed, for example Drimys, Degeneria, Bubbia, or Exospermum.

flowering plants

895

stratigraphic gap between their last appearance (Late Permian) and the first angiosperm-like fossils. Bennettitales

The Bennettitales are another group that has been repeatedly suggested as containing possible angiosperm ancestors. This assumption was initially based on the organization of the reproductive structures, which were first thought to have opened like flowers (Wieland, 1906), and was later supported by numerous phylogenetic analyses (Doyle, 2006 and references therein). Bisexual bennettitalean reproductive structures consist of a basal receptacle to which are attached numerous stalked ovules (Chapter 17). Surrounding the ovules is a whorl of microsporophylls. Between the ovules were large interseminal scales, hypothesized as fusing together to form carpel-like structures. The seeds of the bennettitaleans have been interpreted by some as having a double integument (Crane, 1986) and, with just a few exceptions, members of the group have reproductive organs subtended by helically arranged bracts. This arrangement is unknown in any other group of gymnosperms. Pentoxylales

The Pentoxylales have received attention because of their interesting reproductive organs which place them close to the angiosperms based on phylogenetic analyses (Crane, 1985b). The ovulate structures Carnoconites contain numerous seeds, each with a double integument and attached to what is interpreted as a fleshy receptacle (Chapter 19). The microsporophylls of Sahnia are arranged around a sterile determinate axis and this structure has been described as flowerlike.

Glossopteridales

The glossopterids have periodically been suggested as angiosperm progenitors (e.g., Melville, 1969). The discovery of seeds partially enveloped by a megasporophyll (Gould and Delevoryas, 1977), presence of ovules attached to the adaxial surface of the megasporophyll (E. Taylor and Taylor, 1992; Nishida et al., 2007), and ovules produced within a cupule (E. Taylor et al., 2007) have perhaps provided support for this hypothesis (Chapter 14). Again this theory focuses on the evolution of the carpel. Retallack and Dilcher (1981a) presented a scenario for the glossopterids giving rise to angiosperms in which the ovules are borne on the abaxial surface of the megasporophyll and several assumptions are made about the relationship of the seed-bearing unit and the vegetative leaf. Recent cladistic analyses suggest that the glossopterids, together with the Pentoxylales, Bennettitales, and Caytoniales, are sister to the angiosperms (Doyle, 2006; Hilton and Bateman, 2006). If the glossopterids are considered as possible ancestors, however, there is a large

Gigantopteridales

This is a group of late Paleozoic putative seed ferns (Chapter 19). They possess large simple leaves with a reticulate-type venation pattern consisting of multiple orders of veins. Permineralized stems believed to be those of gigantopterids have vessels with perforation plates. Unfortunately nothing is known about the reproductive organs of this group. The presence of a biochemical signature in the form of oleanane, a compound found in angiosperms, has been reported in the Gigantopteridales as well as the Bennettitales (D. Taylor, et al., 2006).

Phylogenetic analyses and angiosperm origins In general, candidacy as a flowering plant progenitor focuses on the nature and organization of the reproductive parts, in

896

paleobotany: the biology and evolution of fossil plants

particular the morphological changes necessary to enclose seeds in a carpel, and the presence of seeds with two integuments. Most seed plants that are known in sufficient detail during the Mesozoic appear to have evolved in the direction of enclosing their seeds. Historically this has been explained as a response to predation pressure. An equally viable interpretation, however, suggests that the placement of sterile tissue around seeds in the form of a carpel may also represent an early stage in the origin of a self-incompatibility system (Zavada and Taylor, 1986a). Paleobotanists have also employed the principles of cladistic analysis or phylogenetic systematics in an attempt to determine the relationships of all seed plants. This has resulted in an increasingly larger data set of characters that can be coded and used in the analysis. In this approach, shared derived characters, or synapomorphies, are used to reconstruct the branches of the phylogenetic tree, each of which represents a clade or monophyletic group. Such a system relies on determining ancestral (plesiomorphic) and derived (apomorphic) characters based on outgroup comparisons of taxa in which the characters of an outgroup help determine the polarity of the tree (Nixon and Carpenter, 1993). One of the major problems in cladistic methodology is how to deal with character state evolution within a lineage (Stuessy and König, 2008). Sometimes, however, patterns of evolution can result in resemblance between taxa that is not due to direct inheritance from a common ancestor, but rather from convergence or parallel evolution; this similarity due to convergence is termed homoplasy. One method used to resolve such conflicting hypotheses employs the principle of parsimony, a method used to develop a phylogenetic tree with the fewest character state changes. Since many characters cannot be determined in fossils, for example, molecular, genetic, cytological, embryological, and sometimes anatomical characters, it is necessary to evaluate the various relationships that are presented and to seek the hypothesis that is supported by the largest number of known characters. More recently, Bayesian methods (a likelihood approach) have been employed; these are based on a model of morphological evolution. The reader is referred to a textbook on cladistics or plant systematics for more information (e.g., Schuh, 2000; Judd et al., 2007). One of the biggest problems in incorporating fossil data on phylogenetic trees is the disposition of stratigraphic information. Although some strict cladists do not support including stratigraphic information in phylogenies, most paleobotanists believe that it is important in understanding plant evolution. Minimum age mapping has proven useful in this regard. In this method, characters present in fossils

are mapped onto the most parsimonious tree, thus providing minimum ages for the nodes of the tree (Crepet et al., 2004; Hermsen and Hendricks, 2007). This approach may be useful in testing hypotheses relating to the evolution of particular floral features. The application of phylogenetic systematics to seed plants and the origin of angiosperms were first used by Hill and Crane (1982) utilizing only extant seed plant groups. Since that time there have been numerous phylogenetic analyses of seed plants and the relationships of gymnosperms and angiosperms, some using fossils and others not. In some instances these analyses point to different groups as representing relationships with angiosperms. Subsequently, a number of studies have been done that incorporate both fossil and extant groups in cladistic analyses (Crane, 1985a, b, 1988; Doyle and Donoghue, 1986, 1987a, b; Nixon et al., 1994; Rothwell and Serbet, 1994; Doyle, 2006, 2008; Hilton and Bateman, 2006). Each of these studies has, in turn, utilized a greater number of character states in the analysis and has provided detailed rationales for the decision to use or reject various characters. There are also a number of studies that combine both molecular and morphological results of extant angiosperms, and some of these also include fossils. Several of these studies, especially the earlier ones, suggest that the angiosperms, gnetophytes, Pentoxylon, and the bennettitaleans form a clade that includes the Mesozoic seed fern Caytonia, or perhaps the glossopterids, the so-called anthophyte hypothesis (discussed earlier). Among the characters of the common ancestor of this clade are granular, nonsaccate pollen, syndetocheilic stomata, simple leaves, and aggregated sporophylls (Doyle and Donoghue, 1986). Although these ideas are not necessarily new, they do suggest what gymnospermous groups should be more critically evaluated in the future. It is important to note, however, that preservation plays a very significant role in such analyses. The bennettitaleans and Pentoxylon are known from permineralized specimens, which thus make it possible to include additional characters and determine homologies that are not always evident in other preservational modes. For example, it is possible to identify a double integument in Carnoconites (Pentoxylales) and possibly in the seeds of the bennettitaleans because they are permineralized and not simply compressions or impressions. Finally, a variety of other parameters directly influenced the origin and subsequent diversity of the angiosperms. Such factors as pollination biology, habitat, compatibility systems, species diversity, propagule dissemination, changes in life history, and competition are but a few of the biotic and abiotic ingredients that were responsible for the origin and

chapter 22

subsequent evolution and diversification of the flowering plants. Ultimately, it may never be possible to single out the sister group or groups directly involved in the ancestry of the flowering plants. Nevertheless, it is important to utilize all available methods, and the application of phylogenetic systematics, coupled with the utilization of a variety of neontological techniques, including a detailed analysis of embryology, contribute to our ability to pose specific hypotheses about angiosperm origins that can be duplicated and tested as additional fossil data are discovered and interpreted. If previous interpretations of fossil and molecular evidence are accurate, then why is there so little evidence of flowering plants prior to the Early Cretaceous? The upland origin hypothesis was one of the earliest ideas used to explain the absence of an angiosperm fossil record prior to the Cretaceous. According to this idea, angiosperms initially evolved in rolling, hilly tracts and lower to middle slopes of mountains in the ancient tropics (Axelrod, 1952). Because these sites were far from the depositional basins, the proponents argued that the opportunity for plants to become fossilized was extremely low. Others countered that, in spite of the absence of a megafossil record, some pollen grains produced by these upland plants should be preserved in the record (Scott et al., 1960), and current palynological evidence supports this argument. Supporters countered that the probable delicate nature of the exine decreased the preservation potential of the pollen or that the grains were impossible to distinguish from those of non-angiosperm types. Other reasons that have been historically noted include the initial small population size resulting from a probable allopatric speciation event, the elimination of the original sites owing to various physical processes, and, in the case of pollen, the very small amount produced by the presumably insect-pollinated plants. Another often-cited explanation for lack of a pre-Cretaceous angiosperm record is related to the difficulty in interpreting precisely what an angiosperm is and when a sufficient suite of characters is present to achieve the level of “angiospermy.” To a large degree, this concept is embraced in the transitional-combinational theory discussed earlier (Stuessy, 2004), which requires a specific combination of characters to achieve angiospermy. If one adheres tenaciously to certain characters that are probably not going to be found in the fossil record, such as double fertilization, development of an endosperm, and a shortened life history as evidence of angiospermy, then it may never be possible to identify the earliest flowering plants. As we have learned more about early angiosperms, there has been an important shift in emphasis from searching for ancestor–descendant relationships to examining levels of

flowering plants

897

evolution within particular clades of seed plants. It is likely that the earliest flowering plants will not have the full complement of angiosperm features, since evolution proceeds at different rates in different plant organs and even within the same group (Sanderson, 2002). Certainly the earliest flowering plants probably bear little or no resemblance to those we are most familiar with now, even if one includes the so-called basal angiosperms. In this context it is important to continue to be open to new ideas regarding the timing and interpretation of organs and groups of organisms that may provide clues about the earliest flowering plants, and the willingness to examine multiple sources of data will be especially important in the future (Bateman et al., 2006). One need only to recall that Wieland’s ideas about the relationship of Cycadeoidea cones and angiosperm flowers was not widely supported until the cladistic analyses in the mid-1980s reestablished this putative evolutionary relationship. Despite all of the advances that include fossils, techniques, molecules, and even philosophy, the issue of the origin of the flowering plants remains firmly unresolved.

Selected angiosperm families It is beyond the scope of this book to try to document, even in a cursory way, the fossil record of the perhaps more than 400 families of flowering plants that are known from the Cretaceous onward. Rather, we have provided some examples in this section from more than 95 families so as to demonstrate not only the diversity of the angiosperm fossil record but also the types of evidence that have been used to establish the existence of these groups in the fossil record. Some of the examples given will also illustrate grades of angiosperm evolution that have been attained at particular points in geologic time. To place the discussion of the fossil angiosperms within a phylogenetic context, we have adopted the classification system used in Flowering Plant Families of the World (Heywood et al., 2007) and Angiosperm Phylogeny Group II (2003), with the monocots nested within the dicots, and present the families in this order. Excellent coverage of selected extant angiosperm families can be found in Plant Systematics (Simpson, 2006), Plant Systematics: A Phylogenetic Approach (Judd et al., 2007), and the online publication, The Families of Flowering Plants: Descriptions, Illustrations, Identification, and Information Retrieval (L. Watson and Dallwitz, 1992 onward). The reader is also referred to the comprehensive treatment of the flowering plants by Thorne (2007) for up-to-date information on authorship and year of publication.

898

paleobotany: the biology and evolution of fossil plants

Basal angiosperms The oldest diverging flowering plants are included in a group sometimes termed the basal angiosperms. They represent a few groups that diverged from other flowering plants before the appearance of the eudicots. Flowering plants at this grade of evolution typically have combinations of features that include numerous tepals, laminar stamens with wide filaments, numerous individual carpels, alternate, spirally arranged leaves, and monoaperturate pollen. Included in this group are the Amborellaceae, Chloranthaceae, Nymphaeaceae, Austrobaileyales, Ceratophyllaceae, and magnoliids. Combinations of these families and orders based on DNA sequences are also referred to as the ANITA grade and include Amborellaceae, Nymphaeales, Illiciales, Trimeniaceae, and Austrobaileyaceae (Qiu et al., 1999). Floral features of the ANITA grade include small flowers with few to moderate numbers of spirally arranged organs, free styles, and ascidiate carpels with one to few ovules; modern members are insect pollinated (Endress, 2001). Epidermal features within the group show that ethereal oil cells and hydropotes are synapomorphies of the Austrobaileyales and Nymphaeales, respectively (Carpenter, 2006). The important discovery of several suites of fossils of Early Cretaceous age from western Portugal suggests that angiosperms were perhaps more diverse during the Barremian– Aptian than previously anticipated (Friis et al., 1999, 2000a, b, 2005). It is quite probable that as angiosperm phylogenies incorporating fossils become more robust, the placement of some of these groups may shift. Finally, it is becoming increasingly clear that future research needed to demonstrate relationships among these groups will require developmental evidence from a variety of sources, including the nature of the megagametophyte (Friedman et al., 2003). AMBORELLACEAE Phylogenetic analyses resolve the monotypic genus Amborella as sister to all other angiosperms (Jansen et al., 2007), or as occupying this position with the Nymphaeales (water lilies) (Zanis et al., 2002; Soltis et al., 2007). An analysis of 81 plastid genes, the largest number to date, places Amborella at the base of the angiosperm tree, with the Nymphaeales and Austrobaileyales on the two nodes above (Jansen et al., 2007). This basal position has been challenged by some, who suggest that not only is Amborella not a basal angiosperm, but it is not even basal within the dicots (Goremykin et al., 2003). Amborella has unisexual flowers with broad laminar stamens and lacks vessels. Research dealing with anatomy and leaf photosynthetic physiology

indicates that A. trichopoda is shade adapted and may indicate that the earliest flowering plants possessed limited structural and physiological flexibility to light (Feild et al., 2001). To date there is no fossil record of the group. HYDATELLACEAE The Hydatellaceae is another family that has been interpreted as close to the base of the flowering plant tree (Saarela et al., 2007). This is a family of dwarf aquatics that were once thought to be monocots perhaps related to Poaceae. They are characterized by unisexual flowers with a single carpel or stamen and cellular endosperm. Pollen is round and anasulcate; the tectum is perforate (Remizowa et al., 2008). There is no fossil record to date, and it is certain that the position of the family will change as additional molecular data are assembled. At the present time it is regarded as sister to the Nymphaeales (Friis and Crane, 2007). ARCHAEFRUCTACEAE This family includes three fossil species of Archaefructus (FIG. 22.36) from China, A. liaoningensis, A. sinensis (G. Sun et al., 2002), and A. eoflora (Ji et al., 2004), which are interpreted as herbaceous aquatic plants. The specimens were initially reported as being Jurassic (G. Sun et al., 1998) but are now considered to be 124–125 Ma (Barremian–Aptian boundary), based on 40Ar/39Ar (Swisher et al., 1999) and zircon dating (W. Yang et al., 2007). The specimens consist of reproductive axes borne in the axil of highly dissected leaves (FIG. 22.37). The axes bear up to 60 helically arranged conduplicate carpels borne on short pedicles (FIG. 22.38); each carpel produced from two to four ovules. Short pegs on the reproductive axis are interpreted as evidence of deciduous stamens or carpels (G. Sun et al., 1998). Stamens consist of a short filament and four pollen sacs and the in situ monosulcate pollen is 17–36 μm long. Leaves are small and pinnately dissected. In A. sinensis carpels are helical, opposite, or whorled with a swollen petiole base. There is no evidence of bracts or other perianth parts. In A. eoflora (FIG. 22.39) the reproductive axes are described as representing a bisexual flower (Ji et al., 2004). Another interpretation suggests that the reproductive axes in Archaefructus represent inflorescences and that the plant was completely submerged (Friis et al., 2003a) rather than having its reproductive organs above the water level. Despite the fact that Archaefructus is known from almost entire plants (FIG. 22.39) (Ji et al., 2004), the evolutionary position remains equivocal. There are several features, however, that suggest some similarities to members of the Nymphaeales (Schneider et al., 2003).

chapter 22

flowering plants

899

Figure 22.37 Archaefructus liaoningensis with highly dissected leaves (Cretaceous). Bar  3.5 cm. (Courtesy D. L. Dilcher.)

Figure 22.36 Reconstruction of Archaefructus sinensis plant.

(Courtesy D. L. Dilcher.)

At the present time its early geologic age and the presence of enclosed seeds may represent a level of angiospermy, in which at least some components of the reproductive system are defined (G. Sun et al., 2002). Finally, recall the discussion of other Mesozoic seed plants, namely the seed ferns, in which seeds are enclosed. Could the features of Archaefructus, including the aquatic habitat, be those of a gymnosperm, or do are our current ideas about gymnosperms exclude this interesting possibility? CHLORANTHACEAE In this monophyletic family are aromatic plants with small simple flowers and opposite, serrate leaves with sheathing stipules. Today this is a taxonomically isolated group in which the phylogenetic position remains equivocal; however,

there is agreement that Hedyosmum is sister to the remaining genera (L.-B. Zhang and Renner, 2003). The family includes 4 genera and 75 species that range from shrubs and trees to herbs. One of the unique features found in the family is the three-parted stamen in Chloranthus, a character that can potentially be used to establish the divergence of the extant genus, and to perhaps elucidate the origin of this structure (Herendeen et al., 1993). Chloranthus was initially postulated as an example of an early angiosperm flower, based on the simple organization and extensive fossil record, but molecular phylogenetics support the Chloranthaceae as diverging slightly above the ANITA grade (Doyle et al., 2003). The oldest record of the group is based on the pollen type Asteropollis from the Upper Cretaceous (Doyle, 1999). Grains of this type have been attributed to the living genus Hedyosmum (Friis et al., 2005). The fossil pollen morphotype Clavatipollenites is similar to grains of extant Ascarina, but this type appears to be plesiomorphic for the family (Eklund et al., 2004). Stephanocolpites has also been suggested as

900

paleobotany: the biology and evolution of fossil plants

Figure 22.38 Portion of an axis of Archaefructus liaoningensis

bearing helically arranged carpels (Cretaceous). Bar  1 cm. (Courtesy D. L. Dilcher.)

Figure 22.39 Complete Archaefructus eoflora plant includ-

ing rhizome D. W. Taylor.)

having chloranthaceous affinities (Walker and Walker, 1984). Although chloranthaceous pollen has been known for some time from the Early Cretaceous, the discovery of flowers in rocks from the Lower Cretaceous was an important find (Friis et al., 1986). Chloranthistemon is the name assigned to Late Cretaceous, bilaterally symmetrical, three-lobed androecia from Sweden (Crane et al., 1989; Eklund et al., 1997). The bisexual, zygomorphic flowers lack a perianth, are 1.5 mm long, and closely resemble those of extant Chloranthus (Chloranthaceae). They consist of a tripartite and broadened androecium (FIG. 22.40) borne in an abaxial to lateral position on the monocarpellate ovary. The flowers are arranged in the axils of decussate bracts and pollen sacs are borne laterally with dehiscence of the valvate type. Endress and Hufford (1989) reported that this dehiscence pattern is common in primitive extant angiosperms.

(arrow)

(Cretaceous).

Bar  5 cm.

(Courtesy

Pollen is spheroidal, 12–14 μm in diameter, and semitectate. Chloranthus stamens with in situ pollen from the Upper Cretaceous are referred to as C. crossmanensis (Herendeen et al., 1993). Flowers from the Lower Cretaceous Potomac Group consist of three-lobed androecia, each bearing two pairs of small pollen sacs. Those from the Upper Cretaceous have a single pair of pollen sacs on the lateral, androecial lobes. Features of the stamens suggest affinities with Chloranthus or Sarcandra, two modern taxa that possess an entomophilous pollination syndrome (Friis et al., 1986). As a result of a comprehensive morphological analysis using 6 fossil and 38 extant species, Eklund et al. (2004) hypothesized that the ancestors to the family were either trees or shrubs that had unisexual or bisexual flowers with monosulcate

chapter 22

flowering plants

901

Figure 22.41 Rhizome of a water lily or lotus (Oligocene).

Bar  4 cm. (Courtesy BSPG.)

Figure 22.40 Portion of a flower of Chloranthistemon endres-

sii showing three-parted androecia (Cretaceous). Bar  140 μm. (From Eklund et al., 1997.)

pollen. The results of this study demonstrate how morphological data, including those from fossil taxa, can be of value in reconstructing early morphotypes within the family. It has been suggested that the small flowers of the Chloranthaceae and anemophily in two genera may be ecologically linked to sun tolerance and to less reliable pollinators as the group spread to cooler regions (Doyle et al., 2003). Nymphaeales

NYMPHAEACEAE This order is considered to be one of the most basal groups of flowering plants and includes two well-defined families, the Nymphaeaceae and the Cabombaceae (Les et al., 1999). Extant members are thought to have diversified during the Eocene (Yoo et al., 2005). Modern members of this family include water lilies that grow partially submerged in freshwater lakes, ponds, and streams. The group includes 8 genera and 70 species of aquatic plants that are worldwide in distribution. The fossil record of the family includes large rhizomes (FIG. 22.41), leaves, and flowers (e.g., Anoectomeria), in addition to seeds. Some that occur in the Lower Cretaceous (Albian) include Brasenites (FIG. 22.42), Aquatifolia, and Scutifolium, three leaf morphotaxa that are interpreted as belonging to the Nymphaeaceae and Cabombaceae respectively (H.-S. Wang and Dilcher, 2006; D. Taylor et al., 2008).

Figure 22.42 Leaf of Brasenites kansense (Cretaceous).

Bar  2 cm. (From H.-S. Wang and Dilcher, 2006; Courtesy S. R. Manchester.)

Leaves ascribed to Nymphaea have been reported from the Miocene of Argentina (Anzótegui, 2004), and there are several reports of the family based on vegetative remains, such as leaves and rhizomes, as well as dispersed pollen and seeds (Collinson, 1980b; Borsch, 2000; I. Chen et al., 2004; Hesse and Zetter, 2005; Löhne, 2007). Recent discoveries now include floral remains in the form of charcoalified mesofossils of flowers from the Early Cretaceous sediments

902

paleobotany: the biology and evolution of fossil plants

PC

Figure 22.43 Lateral view of Microvictoria svitkoana with

receptacle cup, paracarpels (PC), and central sterile column (Cretaceous). Bar  450 μm. (Courtesy K. Nixon.)

in the western basin of Portugal (Friis et al., 2001, 2006a). The small (3 mm long) flower is radially symmetrical, perigynous, and contains both stamens and carpels. Pollen associated with the floral parts is monocolpate, but unlike other members of the order, the pollen wall is reticulate and tectate. The arrangement of carpels is suggested as being similar to that found in the Nymphaeales, but family placement is uncertain. Microvictoria svitkoana is another fossil flower reported from the lower Upper Cretaceous (Turonian) Raritan Formation of New Jersey (Gandolfo et al., 2004). Currently, it represents the oldest fossil flower that can unequivocally be assigned to the Nymphaeaceae. Specimens are up to 4 mm long and consist of numerous helically arranged floral parts (FIG. 22.43). A combined morphological and molecular analysis suggests that Microvictoria is most closely related to the modern Victoria–Euryale clade. Pollination biology of the fossil is interpreted as being similar to that in extant Victoria, where there is a sophisticated trap-and-release mechanism for the beetle pollinators, further substantiating the coevolution of plant–insect interactions in the Late Cretaceous. The earliest seeds referable to the group are from the Maastrichtian and assigned to the genus Barclayopsis; they

also appear closely related to the Cabombaceae (Knobloch and Mai, 1984). There are several other extinct genera based on seed morphology and anatomy (Takahashi et al., 2007), for example, Symphaenale futabensis, a form reported from the Santonian (Upper Cretaceous) of Honshu, Japan. Specimens of S. futabensis are anatropous with an inconspicuous raphe (Takahashi et al., 2007). Cells of the exotesta are columnar with lobed outer walls. Although there are some similarities between Symphaenale seeds and those of modern Brasenia, other combinations of features are found in other extant and fossil seeds. Permineralized seeds have also been described from the early Eocene of China (I. Chen et al., 2004). Seeds of Nuphar wutuensis are 5 mm long with a raphe ridge on one side; surface cells are pentagonal to polygonal in outline, depending on which layers of the integument are exposed. Structurally preserved seeds have also been reported from the middle Eocene Allenby Formation of British Columbia (Cevallos-Ferriz and Stockey, 1989). Specimens of Allenbya include carpels enclosing at least four ovules. The seeds are about ~7 mm long and characterized by an operculum at the micropylar end and columnar cells in the seed coat (FIG. 22.44). Although A. collinsonae seeds share the greater number of similarities with the fossils Dusembaya and Palaeonymphaea, they may be related to the extant genus Victoria. Fossil seeds from the late Paleocene and Pleistocene share some features with those of the extant genus Euryale (Miki, 1960; W. Taylor et al., 2006). The abundance of seeds of this order in the fossil record suggests that the group was far more diverse in the past than now and that the families Nymphaeaceae and Cabombaceae were well defined by at least the middle Eocene (Cevallos-Ferriz and Stockey, 1989). Austrobaileyales

AUSTROBAILEYACEAE This monotypic family includes a single species, Austrobaileya scandens, a woody vine endemic to tropical Queensland, Australia, where it grows in low-light forest understorey habitats (Feild et al., 2003). To date, no fossil representatives are known. ILLICIACEAE This family includes evergreen trees and shrubs with simple leaves and small, bisexual flowers composed of numerous tepals. The number of stamens and carpels are few to many, with each carpel containing a single seed; floral features are generally highly variable (Oh et al., 2003). Extant species have aromatic oil cells. In Illicium the female gametophyte is four celled/four nucleate, which is considered to be a plesiomorphic character in members of the Austrobaileyales and Nymphaeales (Williams and Friedman, 2004).

chapter 22

flowering plants

Figure 22.45

903

Dieter H. Mai.

has been reported from several localities, including the Upper Cretaceous and Paleogene of Antarctica (Poole et al., 2000).

Figure 22.44 Longitudinal section of Allenbya collinsonae

seed showing distal operculum (Eocene). Bar  1.5 mm. (Courtesy R. A. Stockey.)

Anacostia is known from the Lower Cretaceous of North America and Europe and includes carpels with an indistinct stigma, each containing a single seed (Friis et al., 1997a). The affinities of these fossil fruits may lie with magnoliids or monocots. Several fossil leaves described as Illicium have been reported from the Eocene of Germany based on cuticular anatomy including the presence of ethereal oil cells (Jähnichen, 1976). The foliage type Longstrethia varidentata from the Upper Cretaceous (Cenomanian) of the Dakota Formation in Nebraska, USA, is also placed within the family by some authors (Upchurch and Dilcher, 1990). Illiciospermum includes isolated seeds from the Cretaceous that share several features, such as epidermal anatomy, size, and integument thickness, with modern Illicium seeds (Frumin and Friis, 1999). Seeds and a few fruits attributed to the family are known from the Miocene of Europe (Mai, 1970) (FIG. 22.45) and the lower Miocene Brandon lignite of North America (Tiffney and Barghoorn, 1979). Wood assigned to Illicioxylon

SCHISANDRACEAE The 50 extant species in two genera in this family are represented by woody vines and lianas with small unisexual flowers, the majority confined to Japan, China, and southeastern Asia. Based on sequence (Zhong et al., 2006) and morphological data it remains unclear whether the two extant genera, Schisandra and Kadsura, are monophyletic (G. Hao et al., 2001). Molecular phylogenies delimit two clades within the family—one including only Schisandra species and the other with species of both genera (Z. Liu et al., 2006). The close relationship of the two genera is also supported by cuticular analysis (Z.-R. Yang and Lin, 2005) and the presence of a four-celled/four-nucleate embryo sac in both taxa (Tobe et al., 2007). Leaves assignable to the family have been reported from the Upper Cretaceous of China (S. Guo, 1984) and Miocene of Germany (Ferguson, 1971). The fact that extant leaf genera can be distinguished based on epidermal anatomy holds promise for future identifications of members of the family (Denk and Oh, 2006). Tricolpate, heteropolar grains from the Upper Cretaceous of California have also been attributed to the family (Chmura, 1973), and seeds have also been reported from the Eocene Clarno Formation (Manchester, 1994a) that are similar to the extant species S. glabra. Other seeds have been reported from several Miocene sites in Europe.

904

paleobotany: the biology and evolution of fossil plants

Figure 22.46 Ceratophyllum muricatum fruit (Paleocene). Bar  4 mm. (Courtesy P. S. Herendeen.)

Ceratophyllales

Historically, the herbaceous aquatics placed in this order have been included in the Nymphaeales (Cronquist, 1988). Studies by Les (1988), however, have supported the recognition of a separate order. CERATOPHYLLACEAE The family is monotypic and consists of a single extant genus Ceratophyllum. Fossil fruits of Ceratophyllum are known from the Paleocene (FIG. 22.46). Each is 4 mm long, ornamented by 8–11 elongate spines (FIG. 22.47), and is included in the extant species C. muricatum (Herendeen et al., 1990). Additional remains of Ceratophyllum are known from the Eocene Green River Formation of western North America. These include an elongate axis with swollen nodes bearing whorls of finely dissected leaves associated with ceratophyllacean fruits (Herendeen et al., 1990). The occurrence of extant species in the Paleogene is unusual and is interpreted as demonstrating the very small amount of evolutionary change in Ceratophyllum. The family demonstrates no close affinities with any other extant group of angiosperms, and the presence of a number of features, for example orthotropous, unitegmic ovules, and simultaneous

Figure 22.47 Ceratophyllum muricatum Bar  1 mm. (Courtesy P. S. Herendeen.)

fruit

(Extant).

microsporogenesis, suggests that the family may be a vestige of an ancient line of angiosperms (Les, 1988).

Magnoliids Canellales

WINTERACEAE This family includes four to nine genera of woody plants that occur principally in the South Pacific, including Australia and New Guinea. One genus, Takhtajania, is monotypic and occurs in Madagascar. The leathery leaves of these evergreen trees and shrubs are simple and entire, with pinnate primary venation and are alternately arranged on the stem. Flowers are bisexual with few to numerous stamens with short broad filaments, and carpels with ill-defined stigma and style. The wood lacks vessels, but this has been suggested to represent a secondary acquisition (Young, 1981). Most species shed pollen in permanent tetrads; individual grains are coarsely reticulate and characterized by a circular sulcus. Two flowers from

chapter 22

Figure 22.48 Flower of Cronquistiflora (Cretaceous). Bar  1.2 mm. (Courtesy K. Nixon.)

flowering plants

905

sayrevillensis

Figure 22.50 Lateral view of Detrusandra mystagoga flower

showing laminar stamens (Cretaceous). Bar  575 μm. (Courtesy K. Nixon.)

Figure 22.49 Two carpels of Cronquistiflora sayrevillensis showing peltate stigmas (Cretaceous). Bar  525 μm. (Courtesy K. Nixon.)

the Late Cretaceous, Cronquistiflora (FIGS. 22.48, 22.49) and Detrusandra (FIG. 22.50), possess laminar stamens and monosulcate pollen (Crepet and Nixon, 1998a). They share several characters with cupulate magnoliids and also have features found in the Winteraceae.

Figure 22.51 Tetrahedral tetrad of Walkeripollis gabonensis (Cretaceous). Bar  5 μm. (From Doyle et al., 1990a.)

Pollen of the Winteraceae has been reported from the Lower Cretaceous (Aptian–Albian) of Israel (J. W. Walker et al., 1983). Grains are produced in tetrahedral tetrads and have been placed in the genus Walkeripollis (FIG. 22.51) (Doyle et al., 1990a). Individual grains of W. gabonensis range from

906

paleobotany: the biology and evolution of fossil plants

25–35 μm and have an elliptical aperture. Ultrastructural studies of the pollen wall help to confirm the affinities of Walkeripollis within the Winteraceae. Cladistic analyses of several early pollen types, including Walkeripollis, imply that the early members of the Winteraceae were far more widespread geographically in the Early Cretaceous than previously thought (Doyle et al., 1990b). For example, a tetrad of Walkeripollis grains has recently been reported from the mid-Cretaceous of Patagonia, Argentina (Barreda and Archangelsky, 2006). Wood with affinities to the Winteraceae has been described from the Upper Cretaceous of California (Page, 1979). The tracheids are small and thick walled; vessels are absent. Rays are narrow (four to five cells) with minute pits on the radial cell walls. Another fossil wood placed in the Winteraceae is Tetracentronites panochetris (Upper Cretaceous) (Page, 1968, 1981). Wood assigned to the Winteraceae has also been described from the Upper Cretaceous of Antarctica as Winteroxylon (Poole and Francis, 2000). Anatomical features suggest close affinities with extant Bubbia. Laurales

This order includes 7 families and 3300 species, with 5 families known from the fossil record (Renner, 2005b). All are insect pollinated and most have seeds distributed by animals. Diversity in the group is suggested to be the result of habitat and geology rather than pollinators and dispersers. CALYCANTHACEAE This is a small family of shrubs and small evergreen trees with simple leaves and bisexual, actinomorphic, solitary flowers. Today the group is restricted to North America, eastern Asia, and northern Queensland, Australia. Jerseyanthus calycanthoides is a Cretaceous flower characterized by cupulate receptacles with recurved tepals near the margin of the cup (FIG. 22.52) (Crepet et al., 2005). Stamens are fleshy and produced in pairs; pollen is disulculate. On the inside of the cupule are elongate trichomes. Carpels are free and contain a single seed. Although the fossil flowers share a number of characters with extant members of the family, the closest correspondence with living members appears to be with Calycanthus. LAURACEAE There are nearly 3000 tropical and subtropical species in this family, with most trees and shrubs bearing simple, often leathery leaves. Many species possess aromatic oils. The flowers are regular and bisexual, with flower parts usually in threes. Stamens are organized in four whorls of three each,

Figure 22.52 Diagrammatic reconstruction of Jerseyanthus calycanthoides. (Courtesy W. L. Crepet.)

but in some instances they are reduced to staminodes. The fruit is typically a berry. The family has an extensive fossil record that includes leaves, wood, pollen, fruits, seeds, and flowers from the Cenozoic, with a few reports as early as the Cretaceous (Upchurch and Dilcher, 1990; Herendeen et al., 1994). Crassidenticulum is an oblong leaf type from the Upper Cretaceous (Cenomanian) of southeastern Nebraska (Upchurch and Dilcher, 1990). The margin is finely serrate with chloranthoid teeth; the midvein consists of several strands with craspedodromous secondary venation. Other leaves from the same site that are believed to be lauraceous include Densinervum, Landonia, Pabiania (FIG. 22.53), and Pandemophyllum. Leaves and flowers attached to the same branch have also been reported from the Lower Cretaceous of Brazil (Mohr and Eklund, 2003). Leaves of Araripia florifera are three lobed with pinnate venation (FIG. 22.54), although shape may be highly variable (FIG. 22.55). The flowers possess helically arranged tepals. Wood assignable to the Lauraceae is relatively easy to recognize based on the presence of variable vessel-ray pitting (cross-field pitting), heterocellular rays one to three cells wide, oil cells (FIG. 22.56), vasicentric paratracheal parenchyma, and small to medium pores (Scott and Wheeler, 1982). Ulminium is a common lauraceous wood type reported from numerous Cretaceous and Cenozoic localities, including the Lower Cretaceous of central California (Page, 1967) (FIG. 22.280) and Eocene Clarno Formation of Oregon (Scott and Wheeler, 1982). Specimens of U. scalariforme from Oregon have scalariform perforation plates in association with oil cells which are located on the margins of the rays and interspersed among the fibers. Paraphyllanthoxylon marylandense is another type of lauraceous wood described from the

chapter 22

flowering plants

907

Figure 22.53 Pabiania variloba (Lauraceae) (Cretaceous).

Bar  1.5 cm. (Courtesy G. R. Upchurch.)

Figure 22.54 Twig bearing leaves and fruits of Araripia

Figure 22.56 Heterocellular rays accompanied by oil cells (arrow) in the lauraceous wood Beilschmiedioxylon africanum. (From Dupéron-Laudoueneix and Dupéron, 2005.)

florifera (Cretaceous). Bar  3 cm. (Courtesy B. A. R. Mohr.)

Figure 22.55 Different morphologies of Araripia florifera leaves. (From Mohr and Eklund, 2003.)

mid-Cretaceous Potomac Group of eastern North America (Herendeen, 1991). Other species of Paraphyllanthoxylon (Wheeler and Lehman, 2000) (FIGS. 22.57, 22.58), however, appear to be more difficult to assign to a particular family and have variously been placed into a number of plant families, including Lauraceae, Elaeocarpaceae, Flacourtiaceae, Anacardiaceae, Burseraceae, Euphorbiaceae, and Sapindaceae (see Cahoon, 1972 for P. alabamense). Additional information on fossil lauraceous woods and a survey of genera of silicified woods referred to the Lauraceae can be found in Süss (1958) and Dupéron-Laudoueneix and Dupéron (2005).

908

paleobotany: the biology and evolution of fossil plants

Figure 22.57 Tangential section of Paraphyllanthoxylon ari-

Figure 22.58 Tangential section of Paraphyllanthoxylon uta-

zonense showing multiseriate rays (Cretaceous). Bar  200 μm. (Courtesy E. A. Wheeler.)

hense showing multiseriate rays (Cretaceous). Bar  100 μm. (Courtesy E. A. Wheeler.)

Mauldinia is used for small (3.6 mm long), bisexual flowers from the mid-Cretaceous (Cenomanian) of North America (FIG. 22.59) (Drinnan et al., 1990b), Upper Cretaceous of central Asia (Frumin et al., 2004) and middle Cretaceous of Germany (Viehofen et al., 2008). The flowers are triangular in cross section, possess a perianth of triangular tepals arranged in two whorls (FIG. 22.60), and are borne in lateral inflorescences. The anthers demonstrate both introrse and extrorse dehiscence depending on their position in the flower. The ovary is unicarpellate with a triangular-shaped stigma. The fusinized preservation of some specimens shows the tissue organization of the endocarp and cellular nature of the endosperm. The presence of M. mirabilis and M. hirsuta in the Cenomanian represents early evidence of the family and further underscores the emerging picture of exceptional flowering plant diversity among the magnoliid dicotyledons early in the geologic history of the group. Perhaps the earliest evidence of the Lauraceae is the charcoalified flower Potomacanthus lobatus (von Balthazar et al., 2007). This early–middle Albian flower is bisexual and trimerous and is characterized by a single carpel with one seed. The presence of several derived characters in P. lobatus provides support for the hypothesis that diversification of the Laurales was well underway in the Early Cretaceous.

Magnoliales

The Magnoliales are woody angiosperms that inhabit tropical to warm-temperate climates. At one time this group was thought to represent the most ancient of the flowering plants based on the characters of simple leaves with pinnate venation, laminar stamens producing granular, monosulcate pollen, conduplicate carpels, and absence of vessels in some genera. More recently, however, some of these traits are now considered as synapomorphies that were independently derived in other magnoliid angiosperms (Sauquet et al., 2003). Molecular phylogenetic analyses regard the Magnoliales as a monophyletic group of six families (Zanis et al., 2002). All lines of evidence suggest the Magnoliales as being more derived, but still in a basal position relative to the eudicots and monocots (Friis et al., 1997b). ANNONACEAE Included in this family are 2400 species of trees, shrubs, and woody vines with simple leaves and solitary inflorescences. Wood is known from the Paleocene and Eocene (Crawley, 2001; Wheeler and Manchester, 2002), pollen from several different sites, and floral evidence from the Late Cretaceous of Japan (Takahashi et al., 2008). Eocene seeds are described

chapter 22

flowering plants

909

Figure 22.60 Mauldinia mirabilis flower showing separated inner tepals (Cretaceous). Bar  0.5 mm. (From Drinnan et al., 1990b.)

Figure 22.59 Mauldinia mirabilis Bar  1 mm. (From Drinnan et al., 1990b.)

flower

(Cretaceous).

from Africa (Chesters, 1955). Some have suggested that the family originated in South America and Africa in the Late Cretaceous (Doyle et al., 2004), whereas molecular dating suggests that some members of the group may have originated in South America during the Miocene (Pirie et al., 2006). MAGNOLIACEAE Several fossil woods possess anatomical characters like those of extant members of the Magnoliaceae. In Magnoliaceoxylon from the Eocene of western North America, the wood cannot be related to a modern genus (Wheeler et al., 1977). Liriodendroxylon princetonensis is another magnolian wood type from the Eocene of British Columbia, containing a chambered pith surrounded by diffuse-porous wood (Cevallos-Ferriz and Stockey, 1990a). This wood type, like so many other anatomically preserved angiosperm woods, shares a number of features with multiple extant genera.

In recent years there has been a major effort, building on the work of Carlquist (1975, 2001) to compile a list of fossil wood features (FIGS. 22.61, 22.62) believed to be of ecologic versus phylogenetic significance, and to use these with other proxy records, such as leaf physiognomy, to infer paleoclimate of fossil angiosperms (Wiemann et al., 1998, 1999, 2001). One of the centers of this activity is the laboratory of Elisabeth Wheeler and this work has resulted in a database that includes 1600 entries (Wheeler et al., 1986; LaPasha and Wheeler, 1987; Wheeler and Baas, 1993), which is now maintained by the North Carolina State University Library digital collection. Associated with the fossil images is a modern wood database, InsideWood, which includes an interactive key, 5500 descriptions, and 32,000 images (Wheeler et al., 2007; http://insidewood.lib.ncsu.edu/search/). Such a system provides a basis for characterizing primitive angiosperm wood features through time, such as scalariform perforation plates, scalariform and opposite intervessel pitting, long vessel elements, fibers with bordered pits, and heterocellular rays, and measuring these features against climatic parameters. Added to this are studies of chemical signatures in fossil woods that can be useful as proxy records of past climates (Poole and Van Bergen, 2006) and the development of models that utilize wood structure and ecological parameters in climate predictions.

910

paleobotany: the biology and evolution of fossil plants

Figure 22.61 Cross section of Metcalfeoxylon kirtlandense

Figure 22.62 Tangential section of Baasoxylon parenchy-

showing solitary vessels (Cretaceous). Bar  200 μm. (Courtesy E. A. Wheeler.)

matosum showing rays with two sizes of ray cells (Cretaceous). Bar  100 μm. (Courtesy E. A. Wheeler.)

Endressinia brasiliana is a magnolian fossil from the Lower Cretaceous of Brazil (Mohr and Bernardes-deOliveira, 2004). It consists of a branching axis with ovate leaves and several small, bisexual terminal flowers consisting of several rows of staminodes with glands borne on a flattened receptacle (FIG. 22.63). It is suggested that E. brasiliana may be related to Eupomatiaceae. Lesqueria is a fruiting axis from the mid-Cretaceous of Kansas that has obvious affinities within the Magnoliidae (Crane and Dilcher, 1984). The axis consists of a receptacle bearing numerous (175–250) helically arranged follicles (carpels) (FIG. 22.64). Each follicle is characterized by an adaxial suture; the distal end of the carpel has a bifid tip. Follicles contain 10–20 seeds borne in two rows. Below the fruits are helically arranged, laminar flaps (FIG. 22.64) that leave diamond-shaped scars after abscission. These three-dimensional sandstone molds possess a large number of features that are most similar to polycarpic taxa within the Magnoliidae. Another Cretaceous (upper Albian–mid-Cenomanian) magnoliid multifollicular angiosperm fruit is Archaeanthus

Figure 22.63 Axis of Endressinia brasiliana showing ter-

minal flowers (arrows) (Cretaceous). Bar  1 cm. (Courtesy B. A. R. Mohr.)

chapter 22

flowering plants

22.65 Archaeanthus linnenbergeri Bar  35 mm. (Courtesy D. L. Dilcher.)

Figure

911

(Cretaceous).

Figure 22.64 Suggested reconstruction of Lesqueria elocata

showing follicles. (From Crane and Dilcher, 1984.)

(Dilcher and Crane, 1984). The axis of A. linnenbergeri consists of an elongate axis bearing a cluster (100–300) of tightly packed follicles (FIGS. 22.65, 22.66). The carpels have an adaxial crest bearing unicellular trichomes. In the immature follicles are nearly 100 elliptical seeds, each 2 mm long. Beneath the follicles were several zones of different scars believed to represent the former position of stamens, perianth parts, and bud scales (FIG. 22.65). When found isolated, the bud scales are referred to the genus Kalymmanthus, the perianth parts to Archaepetala, and the leaves to Liriophyllum (Dilcher and Crane, 1984). Leaves of L. kansense are petiolate

Figure 22.66 Suggested reconstruction of Archaeanthus lin-

nenbergeri multifollicular fruit with some carpels revealing seeds. (Courtesy D. L. Dilcher.)

912

paleobotany: the biology and evolution of fossil plants

Figure 22.67 Leaf of Liriophyllum kansense (Cretaceous).

Bar  3 cm. (Courtesy D. L. Dilcher.)

Figure 22.69 Protomonimia kasai-nakajhongii (Cretaceous). Bar  7 mm. (Courtesy H. Nishida.)

Figure 22.68 Suggested reconstruction of Archaeanthus linnenbergeri attached to stem with leaves of the Liriophyllum kansense type. (From Dilcher and Crane, 1984.)

and deeply lobed (FIG. 22.67), with some specimens exceeding 18 cm long. Venation is pinnate with the secondary veins camptodromous. In addition to the close association of the disarticulated plant parts, the reconstruction of Archaeanthus (FIG. 22.68) is based on the presence of amber-colored structures (secretory?) on all organs.

Protomonimia is a permineralized, magnoliaceous reproductive organ (FIG. 22.69) from the mid-Cretaceous of Japan that morphologically appears similar to Archaeanthus and Lesqueria (H. Nishida and Nishida, 1988) (FIG. 22.70). The apocarpous, apistillate reproductive organ consists of a receptacle bearing 55 helically arranged carpels (FIG. 22.71). Along the adaxial surface of each carpel is a crest bearing numerous trichomes; each unit contains 12–15 bitegmic, anatropous ovules. A Late Cretaceous multifolliculate fruit from British Columbia includes helically arranged follicles on an elongate

chapter 22

Figure 22.70

Makoto Nishida. (Courtesy H. Nishida.)

receptacle (Delevoryas and Mickle, 1995). Litocarpon is permineralized and includes winged seeds; dehiscence occurs along the dorsal suture. Litocarpon beardii is interpreted as having more features in common with modern magnoliids, including fewer carpels and dorsal dehiscence, than with Archaeanthus and Lesqueria. The early Cenomanian deposits of the Dakota Formation in the central Great Plains (USA) have produced a number of interesting vegetative (Vaez-Javadi and Dilcher, 1990) and reproductive organs (Dilcher, 1979). One of these is Prisca (FIG. 22.73), a raceme of multiple elongate, apetalous, unisexual follicles (Retallack and Dilcher, 1981b). The helically arranged, elliptical follicles may number 90 per receptacle (FIG. 22.73), each containing two to six orthotropous, bitegmic seeds. Because of their association at the locality, leaves of the Magnoliaephyllum-type are used in the reconstruction of this plant (Retallack and Dilcher, 1981b). Several lines of evidence, including abscission scar, thick cuticle, and woody multifollicular axis, suggest that Prisca was a woody shrub or tree, with an anemophilous pollination syndrome; it probably also dispersed seeds and fruits by wind.

flowering plants

913

Figure 22.71 Longitudinal section of Protomonimia kasainakajhongii showing several follicles, each containing seeds (Cretaceous). Bar  7 mm. (Courtesy H. Nishida.)

Figure 22.72

Harald Walther.

914

paleobotany: the biology and evolution of fossil plants

Figure 22.74 Restoration of Archaeanthus linnenbergeri flower. (Courtesy D. L. Dilcher.)

Figure 22.73 Suggested reconstruction of Prisca reynoldsii multifollicle axes. (From Taylor and Taylor, 1993.)

Flowers like Archaeanthus (FIG. 22.74) and Lesqueria are excellent examples of magnoliid reproductive organs that contain many of the features traditionally considered to represent the archetype of the primitive flowering plant (Cronquist, 1988). They occur, however, in rocks that are younger than those containing the chloranthaceous and platanaceous flower types reported from the Lower Cretaceous (Dilcher, 1979). Thus, when one considers only the fossil evidence of flower types, it appears that two rather different floral morphologies occurred at about the same time, one being the polycarpellate type with a large number of

perianth parts, that is, Archaeanthus and Lesqueria, and the other the small, chloranthaceous flowers with few parts that occur in slightly older rocks (see Chloranthaceae discussed below). This might be explained by viewing the flowering plants as polyphyletic, a suggestion that has been voiced by some researchers (Krassilov, 1977b). Or the large Magnoliatype polycarpellate flowers may represent a unit that has evolved from the aggregation of many smaller flowers. Finally, Endress (1987, 1990), based on extant members of the Magnoliidae, suggested that these different early flower types merely reflect the diversity in floral morphology within certain groups of angiosperms. MYRISTICACEAE Today this family is pantropical and includes trees and a few evergreen shrubs with alternate, entire leaves that lack stipules. Flowers are small, unisexual, and borne in terminal inflorescences; staminate flowers have stamens fused to a synandrium and pistillate flowers with a single large seed. Phylogenetic analyses using morphology and several plastid genes indicate that the family originated in AfricaMadagascar and that the Asian taxa are derived (see Doyle

chapter 22

flowering plants

915

B

S C

Figure 22.75 Pollen tetrad of Lactoripollenites africanus

B

(Cretaceous). Bar  10 μm. (From Zavada and Benson, 1987.)

et al., 2004). The fossil record of the family is sparse and based on a few leaves (Wolfe, 1977) and wood (Boureau, 1950; LaPasha and Wheeler, 1987). Piperales

LACTORIDACEAE Although the taxonomic position of this family has been debated (Takhtajan, 1980), several lines of evidence suggest affinities within the Magnoliales, although a recent study of the wood suggests affinities that are closer to the Piperales (Carlquist, 1990). This monotypic family is geographically isolated today, being confined to the Juan Fernández Islands off the west coast of Chile in wet montane forests. Pollen of Lactoris is shed in tetrads, with the individual grains anasulcate and scabrate (Zavada and Taylor, 1986b). The grains differ from other angiosperm pollen in the possession of sacci, a feature found in a large number of gymnosperms. Pollen grains have been recovered that possess characteristics of the family from Turonian–Campanian (Upper Cretaceous) rocks off the coast of southwest Africa. Lactoripollenites occurs in tetrads (FIG. 22.75) up to 36 μm in diameter and the pollen has the same ornament and saccus organization as that in extant grains (Zavada and Benson, 1987). The discovery of fossil lactoridacean pollen supports the hypothesis (based on floral and vegetative features) that the family diverged from its magnolialean ancestor in the Late Cretaceous (Lammers et al., 1986). The discovery of this pollen near Africa indicates that the family was more widespread in the Southern Hemisphere during the Cretaceous.

Figure 22.76 Cross section of Saururus tuckerae flower showing bracts (B), stamens (S), and four carpels (C) (Eocene). Bar  100 μm. (Courtesy R. A. Stockey.)

SAURURACEAE These plants are small perennial herbs with cordate leaves and a spikelike inflorescence of bisexual flowers that lack a perianth. Today Saururaceae occur in eastern Asia and North America and are represented by four genera. Fossil pollen has been reported from the middle Eocene and includes small (10 μm) monosulcate grains that are tectate-columellate (Smith and Stockey, 2007a). Fruits and seeds are known from the Eocene to the Pliocene (Mai and Walther, 1978 (FIG. 22.72); Mai, 1999). Permineralized Eocene inflorescences are included in the extant genus Saururus. Inflorescences are 3 mm long and contain delicate flowers 1 mm in diameter (FIG. 22.76) (Smith and Stockey, 2007b); flowers lack a perianth and have four carpels each (FIG. 22.77). Pollen ranges from 6–11 μm long. One of the ambiguities of the fossil record is the fact that not all specimens, regardless of how beautifully they may be preserved, can be placed within modern families of flowering plants. As a result, new families are created to represent the biodiversity at a particular point in geologic time. One potential candidate for inclusion in such a family is the Eocene

916

paleobotany: the biology and evolution of fossil plants

L

L

Figure 22.79 Cross section of Eorhiza arnoldii showing leaves

(L) (Eocene). Bar  475 μm (Courtesy K. B. Pigg.)

Figure 22.77 Cross section of gynoecium of Saururus tuck-

erae showing four connate carpels (Eocene). Bar  100 μm. (Courtesy R. A. Stockey.)

22.80 Cross section of Princetonia allenbyensis fruit with enclosed seeds (Eocene). Bar  2 mm. (Courtesy R. A. Stockey.) Figure

Figure 22.78 Cross section of Eorhiza arnoldii rhizome

with branch trace (arrow) (Eocene). Bar  2 mm. (Courtesy K. B. Pigg.)

plant Eorhiza arnoldii (Robinson and Pearson, 1973) from the Princeton chert in the Allenby Formation of British Columbia. The plant consists of an extensive rhizome (FIG. 22.78) system that gives rise to upright axes, some of which may be 6 cm in diameter. The central region of the rhizome consists of a massive parenchymatous pith (FIG. 22.78) surrounded by a narrow ring of vascular bundles (Stockey and Pigg, 1994). Leaves are scalelike (FIG. 22.79) and characterized by considerable aerenchyma. An interesting feature of this plant is

the presence of secondary xylem with very narrow-diameter vessels. The presence of ensheathing, scalelike leaves is suggestive of monocots. Fruits and seeds of Princetonia allenbyensis occur in the same chert deposit as Eorhiza (Stockey and Pigg, 1991). The inflorescences are racemes with helically arranged flowers, each with three to five carpels. Fruits are loculate capsules with 40 seeds per locule (FIG. 22.80). Seeds exhibit an outer integument of thick-walled elongate sclereids and an inner integument of rectangular, thin-walled cells (FIG. 22.81). These fossils also have some characters found in aquatic or semiaquatic plants, and it has been proposed that they may have been produced by the Eorhiza plant (FIG. 22.82) (see Stockey and Pigg, 1994). Because the fossils possess characters found in multiple modern families of angiosperms, it is suggested that they represent a family of extinct aquatic plants within the magnoliids.

chapter 22

flowering plants

917

Figure 22.83 Eocene Princeton chert locality in British Figure 22.81 Several seeds of Princetonia allenbyensis

showing micropyle (arrow) (Eocene). Bar  1.5 mm. (Courtesy K. B. Pigg.)

Columbia, Canada. (Courtesy R. A. Stockey.)

Cyclanthaceae, S. Smith et al., 2008). The absence of a large number of fossil monocots compared to the number of dicot fossils probably reflects the higher percentage of total dicot genera, as is the case among extant plants today. Alismatales

Figure 22.82 Diagrammatic reconstruction of the basal parts of Eorhiza arnoldii. (From Stockey and Pigg, 1994.)

ALISMATACEAE The plants in this group live in swamps or marshes and have branched inflorescences. Flowers may be uni- or bisexual and are hypogynous. The presence of laticifers is believed to have some form of protective function. Some genera possess leaves with reticulate venation. Heleophyton is a small (1.5 mm in diameter) petiole from the Princeton chert (middle Eocene) that represents some of the oldest evidence of the family to date (Erwin and Stockey, 1989). Within the ground tissue are 36 oval collateral vascular bundles that are arranged in several series. Like other plants from the Princeton chert (FIG. 22.83), Heleophyton shows anatomical adaptations for life in aquatic habitats in the form of aerenchyma in the vegetative tissues (Cevallos-Ferriz et al., 1991). In addition to documenting the existence of the family by middle Eocene time, the discovery of H. helobiaeoides also demonstrates that the Princeton chert site was an aquatic ecosystem.

Monocotyledons Like the dicots, the monocots are well represented in the fossil record, with pollen and leaves occurring in the Lower Cretaceous (Herendeen and Crane, 1995). Daghlian (1981) indicated that good fossil evidence of members of the Pandanaceae, Arecaceae (Palmae), Potamogetonaceae, Araceae, and several other possible families occurs in the Cretaceous, with the remaining monocots first found in the Cenozoic (e.g.,

ARACEAE A number of extant araceous taxa are found in marshes, but most occur as herbs on the forest floor, although some are lianes (vines). The flowers lack bracts and are arranged in a spikelike inflorescence termed a spadix that is surrounded by a leaflike and sometimes colored spathe. The stamens are united and fewer than six; the ovary is generally a single carpel. This family (Mayo et al., 1997) includes the duckweeds (Bogner and

918

paleobotany: the biology and evolution of fossil plants

Figure 22.85 Portion of a Nitophyllites (Philodendron) limnestis leaf (Eocene). Bar  1.5 cm. (Courtesy D. L. Dilcher.)

Figure 22.84 Suggested reconstruction of Limnobiophyllum scutatum. (Courtesy R. A. Stockey.)

Petersen, 2007), a group sometimes placed in its own family, Lemnaceae. Limnobiophyllum is a fossil that was interpreted as bridging the gap between Lemnaceae and Araceae (uppermost Cretaceous–mid-Cenozoic; see Bogner et al., 2005, 2007). It is a free-floating, stoloniferous water plant (FIG. 22.84) with one or two sessile, suborbicular to reniform leaves of different size, and numerous simple and one or two longer branched roots on a reduced shoot axis (Z. Kvacˇek, 1995a, b (FIG. 22.87); Stockey et al., 1997). Characterization of leaf venation and foliar morphology in extant Araceae has provided a framework to deal with fossil leaves of this type (Wilde et al., 2005). Based on these features four fossil leaf morphogenera are delimited: Araceophyllum, Araciphyllites, Caladiosoma, and Nitophyllites. Large ( 75 cm wide) leaves with a sagittate base and entire margin are included in Nitophyllites (Philodendron) limnestis (FIG. 22.85) (Dilcher and Daghlian, 1977). The secondary and tertiary veins fuse at the margin of the leaf to form submarginal veins (FIG. 22.86); stomata are present on both surfaces. This Eocene aroid is thought to have been a small shrub that inhabited floodplains. Many of the other fossil genera are believed to represent helophytes in swampy sites. Structurally preserved seeds of Keratosperma allenbyense from the middle Eocene are up to 3.2 mm long and possess

22.86 Reconstruction of leaf of Nitophyllites (Philodendron) limnestis leaf showing venation. (From Dilcher and Daghlian, 1977.)

Figure

idioblasts in some areas of the integument (Cevallos-Ferriz and Stockey, 1988). Specimens share the largest number of features with members in the subfamily Lasioideae (Smith and Stockey, 2003). Other monocots from the Princeton chert include Uhlia (Arecaceae) (Erwin and Stockey, 1991a, 1994), Helophyton helobiaeoides (Alismataceae) (Erwin and Stockey, 1989), Soleredera rhizomorpha (Liliales) (Erwin

chapter 22

Figure 22.87

flowering plants

919

Zlatko Kvacˇek.

and Stockey, 1991b), and Ethela sardantiana (Juncaceae– Cyperaceae) (Erwin and Stockey, 1992). Spadix-like fossils of Acorites heeri were described from the middle Eocene Claiborne Formation of Henry County, Tennessee (Crepet, 1978). They lack a spathe, but bear numerous perfect florets. Each flower has bilocular anthers and trilocular ovaries, as does the extant genus Acorus. These compression specimens suggest that each floret has a perianth. Epidermal anatomy of the inflorescence is similar to that in modern aroid genera. Araceites is another genus used for fossil spadices thought to belong to the Araceae (Fritel, 1910). Numerous aroid seeds have been described from the upper Eocene, but the largest number are known from the Oligocene (Madison and Tiffney, 1976). An exceptional suite of floating aquatic monocots have been reported from the Upper Cretaceous (Campanian) Dinosaur Park Formation of southern Alberta, Canada. Cobbania corrugata consists of interconnected rosettes of leaves and roots attached to short stems. They are interconnected by stolons and it is suggested that they grew as floating mats in calm water in an oxbow lake or pond (Stockey et al., 2007). Leaves are up to 7.5 cm long and characterized by a hollow base and trichomes on both surfaces, which would have increased the buoyancy of the plant. Although the fossils show some similarity to modern Pistia (Kvacˇek and Bogner, 2008), detailed examination of the vegetative features serves to distinguish the extant and fossil forms and these authors placed leaves originally described by Lesquereux (1876) into synonymy with Cobbania. Because the morphology of submerged leaves can differ from emergent leaves (Gerber and Les, 1994), this factor can be especially important in paleoecological studies. Leaves of Araciphyllites have been reported from the Upper Cretaceous

Figure 22.88 Mayoa pollen grains showing polyplicate organ-

ization (Cretaceous). Bar  6 μm. (Courtesy E. M. Friis.)

(Campanian) of Austria, which have eucamptodromous venation (J. Kvacˇek and Herman, 2004). Floral structures discovered in the Early Cretaceous (Barremian or Aptian) of Portugal were described as Pennistemon for staminate flowers and Pennicarpus for pistillate flowers (Friis et al., 2000a); Heimhofer et al. (2007) dated the rocks as slightly younger, indicating a major angiosperm radiation during the early Albian. These specimens, together with the in situ pollen Pennipollis, suggest alismatalean affinities, possibly within the Araceae (Friis et al., 2000a). Others place these genera with the Chloranthaceae (Wilde et al., 2005). A Late Cretaceous aroid infructescence has been described from the late Campanian of Canada (Bogner et al., 2005). Albertarum pueri is a bisexual flower with a trilocular ovary and is most similar to extant specimens of Symplocarpus. Rhodospathodendron tomlinsonii is a late Maastrichtian permineralized stem from India that is considered to be a lianescent aroid (Bonde, 2000). The dispersed pollen record of the Araceae is scanty; it begins in the late Early Cretaceous, and peaks in the Paleocene–Eocene. It includes three distinct pollen types (Hesse and Zetter, 2007): a zona-aperturate pollen of the Monstera or Gonatopus type, which is very similar to Proxapertites operculatus; an ulcerate-spiny type typical for Limnobiophyllum; and a polyplicate, omniaperturate pollen type, Mayoa portugallica (FIG. 22.88) (an ephedroid pollen morphology with non-gnetalean affinities), which was reported from late Lower Cretaceous deposits in Portugal (Friis et al., 2004). The fossil Mayoa grains are most similar to extant pollen of Holochlamys, a modern member of the family that is

920

paleobotany: the biology and evolution of fossil plants

Figure 22.90 Several rosettes of Limnobiophyllum (Spirodela)

scutatum showing overall morphology of the plant (Paleocene). Bar  1 cm. (Courtesy R. A. Stockey.) Figure 22.89 Two rosettes of Limnobiophyllum (Spirodela)

scutatum with radiating roots (Paleocene). Bar  5 mm. (Courtesy R. A. Stockey.)

today restricted to New Guinea. Specimens from the Paleocene have been included in Spirodela (McIver and Basinger, 1993) (FIG. 22.89), an extant floating aquatic plant with bilaterally symmetrical lamina (FIG. 22.90) and leaves that range from 1–3 cm long with campylodromous venation. Some of these, however, are now included in Limnobiophyllum. What are interpreted as leaf buds extend from a budding pouch at the base of each leaf. HYDROCHARITACEAE The plants in this family are marine and freshwater aquatics with simple leaves and generally unisexual flowers. Hydrochariphyllum is an entire-margined leaf from the early Miocene of the Czech Republic (Z. Kvacˇek, 1995a). Although several other leaf genera have been described for the family, many of these are no longer considered reliable (Z. Kvacˇ ek, 1995a). Some fossil leaf and seed remains have been assigned to the extant genus Stratiotes. The leaves of S. schaarschmidtii are broadly obovate and may be up to 12 cm long with parallel venation that is acrodromous (Kvacˇek, 2003). Seeds are oblong with a hooked base (Holy and B˚uzek,

1965; Mai, 1985). On the dorsal edge is a keel and on the ventral side a collar through which the micropyle passes (Sille et al., 2006). Morphometric analyses of Stratiotes seeds suggest that there has been an increase in size of the keel accompanied by a change in shape, with the elongate forms the most recently evolved (Sille et al., 2006). This study is an excellent example of how careful analysis is necessary prior to suggesting structure–function relationships in fossil plants, and yet sometimes the results are still difficult to interpret. ZOSTERACEAE (SEAGRASSES) Modern seagrasses are interpreted as polyphyletic and organized into 13 genera within 5 families (Les et al., 1997). These are plants that grow submerged in coastal and estuarine environments, are attached to perennial rhizomes, and have small flowers that are pollinated underwater. Leaves are often ribbonlike or oval. Despite the enormous modern communities and their involvement in shallow marine ecosystems, fossil seagrasses are relatively rare. It has long been believed that the Zosteraceae are of relatively recent origin (Kuo et al., 1989; Aioi, 2000). Phylogenetic analyses by Kato et al. (2003), however, revealed divergence times between the Zosteraceae and its closest relative, Potamogetonaceae, as 100 Ma, which suggests that the Zosteraceae emerged

chapter 22

somewhere in the mid-Cretaceous. This seems to concur with the fossil record. Among the Cretaceous fossils compared to extant seagrasses are Posidonia cretacea and Thalassocharis westfalica from the upper Campanian of Westphalia, Germany, two taxa introduced for stems with attached fibrous bases (P. cretacea) and for stems with attached leaves (T. westfalica) by Hosius and von der Mark (1880). Archaeozostera longifolia from the Campanian or Maastrichtian of Japan has also been compared with modern seagrasses (Koriba and Miki, 1960). The affinities of A. longifolia, however, remain equivocal (Kuo et al., 1989). Specimens of Thalassotaenia debeyi, another Cretaceous– Paleocene (Maastrichtian–Danian) seagrass fossil, consist of strap-shaped leaves, each with 9–15 parallel veins; some leaves exhibit cross veins alternating with bands of fibers (Van der Ham et al., 2007). Seagrasses from the Eocene of Florida are given the generic names Thalassodendron and Cymodocea. Blades of T. auricula-leporis are 2 cm wide, up to 14 cm long, and contain up to 35 parallel veins (Ivany et al., 1990). In C. floridana leaves are attached to branching rhizomes. What is most striking about this fossil ecosystem is the co-occurrence of juvenile ophiuroids (brittle and basket stars), true star fish, and other echinoderms, suggesting that this seagrass community served as a nursery community for these organisms similar to modern seagrass communities. Also present are epiphytic algae. The fossil seagrasses are interpreted as living in a shallow, nearshore warm-water environment that was rapidly buried by siltation caused by an offshore storm (Ivany et al., 1990). Although fossil seagrasses have been related to modern genera based on leaf features, to date there are no fossil flowers known that might make affinities with extant taxa more reliable. An exceptional Pliocene seagrass community, including numerous specimens of Posidonia oceanica, has been described from the island of Rhodes, Greece (Moissette et al., 2007). Included are leaves, in situ rhizomes, and a diverse community of skeletal biotopes, some specific to seagrass leaves, that are comparable to those that exist today in the Mediterranean. There are numerous fossils that may demonstrate some monocot characters, but nevertheless have affinities that remain equivocal. One of these is Klitzschophyllites, an Early Cretaceous angiosperm known from North America, southern Europe, and East Asia as well as parts of the Southern Hemisphere (Mohr et al., 2006). Specimens possess a trifurcate axis, each bearing a single, apetiolate leaf with a serrate margin (FIG. 22.91). Venation is reported as acrodromous– parallelodromous. On the surface of each leaf are round bodies interpreted as glands. The plant is interpreted as living near rivers, deltas, or estuaries (Mohr et al., 2006).

flowering plants

921

Figure 22.91 Axis with leaves of Klitzschophyllites (Cretaceous). Bar  3 cm. (Courtesy B. A. R. Mohr.)

Asparagales

AGAPANTHACEAE Protoyucca shadishii is a monocot with secondary growth from the Miocene of northwestern Nevada (Tidwell and Parker, 1990). Protoyucca has an atactostele with anastomosing collateral bundles (FIG. 22.92) and a secondary thickening meristem. The permineralized fossil specimens have the largest number of features in common with Yucca brevifolia, the Joshua tree. HEMEROCALLIDACEAE This is a small family of herbs with trimerous flowers and nectaries on their sepals. Dianellophyllum is a partial leaf specimen from the Eocene of Australia (Conran et al., 2003). The leaf is compared with the extant taxon Daniella, but differs in the lack of costal differentiation in the epidermal cells. ORCHIDACEAE The Orchidaceae is perhaps the largest family of flowering plants today and includes perennial herbs, with many forms containing aerial roots. The reader is referred to the paper by

922

paleobotany: the biology and evolution of fossil plants

22.92 Cross section of Protoyucca (Miocene). Bar  5 mm. (Courtesy W. D. Tidwell.)

Figure

shadishii

Freudenstein and Rasmussen (1999) for a detailed analysis of specific morphological features used in orchid systematics. The most recent classification delimits five subfamilies (Chase et al., 2003; Freudenstein et al., 2004). Flowers are bisexual and zygomorphic, and one tepal is usually modified to form a labellum which attracts pollinators and serves as a landing platform for them. Pollen, usually in tetrads, occurs in clumps termed pollinia. In a very real sense, orchids have one of the most highly evolved and specialized life history strategies on Earth, and one that perhaps evolved by mimicking other angiosperm flower types without providing a reward to the pollinator (Chase, 2001). The fossil record of the family, however, is scanty and includes only a few specimens of flowers preserved as impression–compressions, including Protorchis, Palaeorchis, and Eoorchis (Schmid and Schmid, 1977). Eoorchis miocaenica, a Miocene compression, has been described as the oldest orchid, although there is only poorly preserved specimen. It consists of a flower 1.3 cm long (FIG. 22.93) (Mehl, 1984). The absence of fossil orchid pollen in the rock record is attributed to the lack of preservational potential and perhaps methods of recovery (Wolter and Schill, 1985). Molecular clock assumptions suggest that the family is at least 65 Ma old (Chase, 2005). Paleontological evidence suggests that the orchids extend at least to the Neogene based on the report of orchid-pollinating bees in amber (Engel, 1999). The report by Ramírez et al. (2007) of pollinia on a stingless bee preserved in Miocene amber is not only an extraordinary example of a specific plant–animal interaction, but also provides unequivocal evidence of the Orchidaceae in the Neogene. Morphological and molecular phylogenies suggest the Orchidaceae evolved perhaps by the mid-Cretaceous (Chase, 2001).

Figure 22.93 Compressed flower of Eoorchis miocaenica (Miocene). Bar  1 cm. (From Mehl, 1984.)

Dioscoreales

DIOSCOREACEAE This is a small family of tropical vines and lianas often with broad reticulate-veined leaves that make them look like dicots. Inflorescences are typically racemes, and flowers are small with three fused carpels. The fruit is a winged capsule. The largest genus is Dioscorea (600 species) which has unisexual flowers with one to two ovules per locule. The group contains plants that are important to humans and includes the true yams (i.e., Dioscorea). Historically there have been a number of fossil leaves assigned to Dioscorea (Berry, 1929; Becker, 1961), but some of these have been questioned (Daghlian, 1981; Conran et al., 1994). Well-preserved leaves have been reported from the Oligocene of Ethiopia (FIG. 22.94). Molecular phylogenetic analyses suggest that Dioscorea represents an early diverging monocot (Hansen et al., 2007). Liliales

PETERMANNIACEAE This is a small family of monocots with net venation rather than parallel veins. Today the group occurs in coastal rain forests in eastern Australia. Petermanniopsis is an Eocene fossil leaf that has parallelodromous venation with random tertiary veins and stomata confined to the abaxial surface (Conran et al., 1994). The leaf is ovate with an acuminate drip tip (Conran and Christophel, 1999). Petermanniopsis

chapter 22

flowering plants

923

Figure 22.95 Lateral view of Mabelia archaia showing val-

vate tepals (Cretaceous). Bar  450 μm. (Courtesy K. Nixon.)

Figure 22.94 Dioscorea sp. leaf (Oligocene). Bar  2 cm. (Courtesy A. D. Pan.)

is an excellent example of a fossil that, while sharing some characters with those seen in a modern family, can only be tentatively assigned to that family. Pandanales

PANDANACEAE Strap-shaped leaves with parallel veins and marginal spines are placed in Pandanites (J. Kvacˇek and Herman, 2004). They are M-shaped in transverse section and have tetracytic stomata arranged in two bands. They appear to have been a common component of coal forming wetlands during the Cretaceous and may be used as evidence to support the hypothesis that monocots originated in wetland habitats (Les and Schneider, 1995). TRIURIDACEAE The nine genera of this family are achlorophyllous herbs with scalelike leaves that today occupy tropical and subtropical habitats. The plants form symbiotic relationships with fungi. Flowers are unisexual and borne in racemose inflorescences, each flower characterized by elongate tepal-like structures that are thought to mimic fungal structures that attract insects (Leake, 1994; Rudall, 2003). Mabelia is a small, unisexual flower with six tepals (FIG. 22.95) from the Upper Cretaceous (Turonian) of New Jersey (USA) (Gandolfo et al., 2002). Anthers are dithecal with in situ pollen that is prolate and monosulcate, with a reticulate tectate structure. Nuhliantha is another triurid flower from the same deposit (FIG. 22.96), which also has

Figure 22.96 Top view of Nuhliantha nyanzaiana flower showing remains of perianth surrounding three stamens and central pistillode (Cretaceous). Bar  400 μm. (Courtesy K. Nixon.)

six tepals and monosulcate pollen. Trimerous, staminate flowers have been reported from the Upper Cretaceous (Santonian) that may also be included in the family (Herendeen et al., 1999).

Commelinids Arecales

ARECACEAE (PALMAE) The palm family includes shrubs, vines, and trees that are primarily found in tropical and subtropical climates. The leaves are large, with the base of the petiole sheathing the stem. The inflorescence is a spadix with the flowers unisexual and regular; fruit is a berry or a drupe. The reader is

924

paleobotany: the biology and evolution of fossil plants

Figure 22.97 Cross section of Palmoxylon wood showing fiber caps (opaque areas) (Cretaceous). Bar  2 cm.

referred to Daghlian (1978b), Erwin and Stockey (1994), and Harley (2006) for authoritative accounts of fossils in this family, and Pan et al. (2006) for a discussion of the occurrences of fossils in Africa. Fossil fruits and pollen of Nypa are know back to the Cretaceous (Chitaley and Nambudiri, 1995; Gee, 2001). One of the most common fossil members of the palm family is Palmoxylon (FIG. 22.97). The genus includes silicified axes that contain evenly distributed vascular bundles (atactostele); individual bundles possess caps of fibers (FIG. 22.98). Two interesting species, P. simperi and P. pristina, believed to come from the Middle Jurassic Arapien Shale Formation near Redmond, Utah (USA), were once believed to represent pre-Cretaceous angiosperms (Tidwell et al., 1970). Subsequent studies, however, revealed that these fossils were actually mid-Cenozoic in age (Scott et al., 1972). Sabal dortchii is a middle Eocene palm with large costapalmate leaves that exhibit axially elongate, abaxial stomata (FIG. 22.99). Each stomatal complex has six subsidiary cells (Daghlian, 1978b). Costapalma, Palustrapalma (FIG. 22.100), and Palmacites are additional genera that have been described from the Gulf Coastal Plain that support the diversity of the group during the Eocene. Daghlian (1978b) postulated that differences in these palms may reflect the differing substrates in which they grew. Palms (FIG. 22.101) appear as early as the middle of the Late Cretaceous (Senonian), with their major radiation apparently taking place during the Paleogene.

Figure 22.98 Vascular bundle of Palmoxylon sp. showing bun-

dle cap of fibers (Cretaceous). Bar  0.5 mm.

Figure 22.99 Lower epidermis of Sabal dortchii show-

ing stomatal complexes (Eocene). Bar  60 μm. (Courtesy C. P. Daghlian.)

Extraordinarily well-preserved palm flowers have been reported from Oligocene–Miocene Dominican, Baltic, and Mexican amber (Poinar, 2002a, b). One of these, Socratea brownii, a staminate flower with up to 50 stamens, represents

chapter 22

flowering plants

925

Figure 22.101 Costapalmate palm frond Sabalites inquirenda

(Paleocene). Bar  5 cm. (Courtesy G. R. Upchurch.)

22.100 Portion of Palustrapalma agathae leaf (Eocene). Bar  2.5 cm. (Courtesy C. P. Daghlian.)

Figure

the first fossil representative of the genus; in Trithrinax dominicana, the stamens are bent inward. Commelinales

COMMELINACEAE The Commelinaceae today are mostly a tropical and subtropical family of herbs and are characterized by succulent stems and sheathing leaves. Often the bisexual flowers contain a blue corolla, and the stamens are arranged in two whorls of three. The fruit is a capsule. The family includes 50 extant

genera of which slightly less than half occur in Africa. The fossil record of the subclass Commelinidae extends from the Paleocene onward. Pollia tugenensis includes Miocene leaves and fruits collected from a tuff deposit in Kenya (Jacobs and Kabuye, 1989). The spirally arranged leaves are lanceolate and 12 cm long; numerous trichomes cover the upper surface. Attached to some stems are clusters of spherical, indehiscent fruits. Features of the fossil leaves suggest affinities with the extant Asian species P. zollingeri, whereas the fruits are more similar to modern species from both Asia and Africa. Poales

CYPERACEAE This family inlcudes grasslike herbs that can be found today in especially wet environments. The culms are often triangular and bear three-ranked, sheathing leaves. Flowers may be unisexual or bisexual and are borne in the axil of a glume (bract). The gynoecium consists of two or three united carpels; stamens number one to three and pollination is anemophilous.

926

paleobotany: the biology and evolution of fossil plants

Figure 22.103 Achene of Cyperocarpus pulcherrima (Miocene). Bar  500 μm. (From Thomasson, 1983.)

Figure 22.102 Achene of Carex graceii (Miocene). Bar  600 μm. (From Thomasson, 1983.)

Cyperacites is a Cenozoic morphogenus used for elongate floral spikes with helically arranged, imbricate scales that terminate naked stems (Hickey, 1977). Carex graceii consists of trigonous achenes that include a perigynium (modified leaflike structure that surrounds the ovary). It measures 3 mm long and exhibits hexagonal–octagonal epidermal cells (FIG. 22.102) (Thomasson, 1983). Achenes lacking a perigynium are placed in Cyperocarpus (FIG. 22.103). Rhizocaulon amatitlani, a structurally preserved rhizome tentatively assigned to the Cyperaceae, has been described from Pleistocene sediments of Lago Amatitlán in Guatemala (Davila Arroyo et al., 2007). The rhizome is elongate and characterized by prominent, upwardly oriented points of shoot insertion and numerous small root stigmata. Nodal diaphragms are porous and show round to oval bundle canals. Other Rhizocaulon-type rhizome fossils, however, have been referred to the Zingiberaceae (Worobiec and Lesiak, 1998; Kovar-Eder, 2004). POACEAE (GRAMINEAE) From an economic standpoint, this is probably the most important family of flowering plants today, since human existence is so dependent on cereal and forage crops. There are more than

600 genera and 8000 species in the family. Long, narrow leaves typically alternate in two rows on opposite sides of the stem and the small, bisexual flowers are organized into inflorescences. The gynoecium consists of three fused carpels, each with a single ovule; stamens are generally in threes. Micromorphological features have provided an excellent method to classify modern and fossil grasses (Thomasson, 1978) and to evaluate a number of adaptive features of the androecium (Thomasson, 1985). The presence of certain types of rust and smut fungi on living members of the Poaceae and Cyperaceae has also been used to infer relationships within these families (Savile, 1990). One of the most abundant and widespread types of fossil evidence for the presence of grasses in the Cenozoic are phytoliths, siliceous infillings of epidermal cells that may be preserved even in the absence of macrofossils. Grass phytoliths, which often have a group or species-specific morphology (FIGS. 22.104, 22.105), represent useful proxy indicators for the reconstruction of Cenozoic grass vegetation and paleoclimatological and paleoecological parameters (Jones, 1964; Strömberg, 2005; Retallack, 2007; Strömberg et al., 2007). Many of the modern grass families are believed to have appeared during the early Eocene (Savile, 1987), although the fossil record is not well demonstrated before the Eocene. Crepet and Feldman (1991), however, have reported grass spikelets (FIG. 22.106) with pollen from the Paleocene– Eocene of western Tennessee that suggest the existence of the Poaceae at least by the Late Cretaceous (FIG. 22.107).

chapter 22

flowering plants

927

Figure 22.104 Phytolith of the Cusquea-type short cell

from a bambusoid grass (Eocene). Bar  22 μm. (Courtesy C. A. E. Strömberg.)

Figure 22.106 Grass spikelet with two florets (Paleocene– Eocene). Bar  2 mm. (Courtesy W. L. Crepet.)

Figure 22.105 Phytolith of the inverted bilobate-type from a grass (Miocene). Bar  20 μm. (Courtesy C. A. E. Strömberg.)

One of the best-known fossil grasses is Tomlinsonia, a permineralized, herbaceous rhizome (FIG. 22.108) 3 mm in diameter that was initially reported from the upper Miocene Ricardo Formation of California (Nambudiri et al., 1978). The vascular tissue consists of alternate rings of bundles within a parenchymatous ground tissue. Attached to the culms are sheathing leaf bases that are two ranked (Tidwell

Figure 22.107 Grass rhizome with leaf sheaths (Paleocene– Eocene). Bar  3 mm. (Courtesy W. L. Crepet.)

928

paleobotany: the biology and evolution of fossil plants

Figure 22.108 Cross section of Tomlinsonia thomassonii rhizome (Miocene). Bar  7.5 mm. (Courtesy W. D. Tidwell.)

Figure 22.109 Fractured leaf surface of chloridoid grass showing prominent vascular bundle sheath (arrow) (Miocene). Bar  40 μm. (From Thomasson et al., 1986.)

and Nambudiri, 1989, 1990). One interesting feature of this plant is the presence of Kranz anatomy in the leaves. Kranz anatomy occurs in plants that exhibit C4 photosynthesis and is characterized by the presence of bundle sheaths surrounding leaf veins, a low ratio of mesophyll to bundle-sheath area, and vascular bundles that are separated from each other by few mesophyll cells. Well-preserved leaf fragments assigned to the subfamily Chloridoideae that also demonstrate Kranz anatomy have been reported from the Miocene Ogallala Formation in Kansas (FIG. 22.109) (Thomasson et al., 1986). These studies are particularly important in that they provide a means of examining some physiological parameters in fossil plants. In this instance, the cellular preservation of the fossils makes it possible to infer a physiological pathway,

C4 photosynthesis, that is efficient at high temperatures and light intensities. Not only does the presence of Kranz anatomy provide some information about the physiology of the plant but it also indirectly provides paleoclimatic data, since C4 plants are most often found in warm to hot tropical and subtropical areas. Since C4 grasses are adapted to low CO2 conditions and water-stressed environments, models linked to global climate have been developed to explain the expansion of C4 grasses during the Late Oligocene (Lunt et al., 2007). Analysis of habitats based on phytoliths suggests two possible scenarios for the rise of grassland-dominated habitat. Using a more general method, Strömberg (2004) found phytolith evidence from northwestern Nebraska that suggested a late Oligocene–early Miocene increase in grasslands. With the use of an index to assess tree cover based on phytoliths, however, the author found that grasslands were present by the late Eocene–early Oligocene, although these were dominated by festucoid, C3 grasses (Strömberg, 2002). Although C4 photosynthesis is not widely distributed among monocots and dicots today, its presence by at least the Miocene indicates that this system is not of recent origin (Monson, 1989; Cerling et al., 1993). Other grasses, also from the late Miocene, have the C3 photosynthetic pathway, as inferred from the anatomy of fossil specimens (Thomasson, 1984). Miocene grass androecia from the Ash Hollow Formation of Nebraska are included in the morphogenus Archaeoleersia (Thomasson, 1980a). The silicified specimens are 3.4 mm long and consist of a subequal lemma and palea bearing delicate prickle hairs. The fossils are assigned to the tribe Oryzeae, which includes wild rice and domesticated rice. Their occurrence in areas where rice does not grow today suggests that this group of grasses migrated southward during the climatic deterioration of the late Cenozoic– Quaternary. Guadua zuloagae is a petrified bamboo culm from the Pliocene of Argentina that is characterized by distinct nodes and internodes (Brea and Zucol, 2007). Each vascular bundle has two large metaxylem vessels and short cells containing silica are present in the epidermis, structures that are also found in some grasses (FIG. 22.110). Other bamboo fossils include leaves and pollen from the Cenozoic of Europe and Japan (Worobiec and Worobiec, 2005). Zingiberales

MUSACEAE The plants in this small family are large herbs with stems formed of overlapping leaf sheaths; inflorescences are large and bracteate. Banana (Musa) is the most widely recognized member of this family.

chapter 22

flowering plants

929

Figure 22.111 Spirematospermum wetzleri fruit with seeds (Miocene). Bar  4 mm. (Courtesy T. Fischer.)

unresolved (Boyd, 1992). Another morphogenus used for Cenozoic Musa-like leaves is Musophyllum (Göppert, 1854).

Figure 22.110 Grass epidermis (Miocene). Bar  250 μm. (Courtesy C. A. E. Strömberg.)

A slightly curved fruit 4 cm long from the Eocene Clarno Formation in central Oregon is included in Ensete, a genus native to the Old World tropics today (Manchester, 1995). Although the specimens are impressions, two rows of seeds can be distinguished. Isolated permineralized seeds have also been reported from the Clarno nut beds (Manchester and Kress, 1993; Manchester, 1994a). Since today Ensete is not known in the Americas, the discovery of seeds and fruits indicates that the genus was far more widespread geographically and adds support to the idea that today southeast Asia represents a refugium for genera affected by climatic cooling during the Cenozoic (Manchester, 1995). The genus Musopsis is used for certain Cenozoic leaf fossils from Greenland that are morphologically similar to leaves seen in modern genera in Heliconiaceae, Musaceae, and Streliziaceae; since the leaves lack distinctive morphological features, the exact systematic affinities remain

ZINGIBERACEAE The Zingiberaceae, or ginger family, consists of perennial herbs with creeping horizontal or tuberous rhizomes. It is comprised of 52 genera and more than 1300 species that are distributed throughout tropical Africa, Asia, and the Americas. Many species are economically important as ornamental plants, spices, or are used in folk medicine. A fossil fruit and seed type variously referred to the Zingiberaceae is Spirematospermum, which occurs from the Late Cretaceous to the Pliocene in Europe and North America (Kirchheimer, 1936; Koch and Friedrich, 1971; Friis, 1988). The fruits are highly variable in size, for example up to 10 cm long and 1.5–2 cm wide in S. wetzleri (FIG. 22.111), and remotely resemble small bananas; ovaries are inferior and placentation is parietal. Seeds are 5–10 mm long, irregularly elongate, and spirally striate on the surface. In the hilar area they are truncate with a slight depression, and around the chalaza rounded or slightly pointed. Zingiberoideophyllum is a morphogenus used for fossil leaves that resemble the leaves of extant Zingiberaceae (Worobiec and Lesiak, 1998). The leaves are hypostomatic and characterized by parallelodromous venation, paratetracytic stomata, and the presence of a hypodermis and secretory cells. In several European localities, Zingiberoideophyllum liblarense leaves co-occur with Rhizocaulon zingiberoides rhizomes and Spirematospermum wetzleri fruits and isolated seeds. As a result, Worobiec and Lesiak (1998) hypothesized that all three fossil species may belong to a single plant.

Eudicots The eudicots are an informal name used for several monophyletic orders of dicotyledons that are estimated to include

930

paleobotany: the biology and evolution of fossil plants

Figure 22.114 Detail of Striatopollis grains extracted from

anther of Spanomera mauldinensis (Cretaceous). Bar  5 μm. (From Drinnan et al., 1991.)

sometimes termed the paleodicots, and the core eudicot clade, which includes two major groups: rosids and asterids. The monophyly of the core eudicots is strongly supported by molecular phylogenetics (Soltis et al., 2003). Figure 22.112 Staminate flower of Spanomera mauldinensis arising in the axil of a bract pair (Cretaceous). Bar  300 μm. (From Drinnan et al., 1991.)

Figure 22.113 Papillae on stigmatic surface of Spanomera mauldinensis showing several pollen grains (arrows) (Cretaceous). Bar  20 μm. (From Drinnan et al., 1991.)

75% of extant angiosperm species. The clade is supported by molecular sequence data, and is also characterized by pollen grains with three germination sites, that is, triaperturates (Doyle and Hotton, 1991), so this monophyletic assemblage has also been called the tricolpates (Judd and Olmstead, 2004). Fossil tricolpate pollen first appears around the Barremian–Aptian boundary. Included in the eudicot clade are a basal grade of orders, including the Ranunculales and Proteales, together with several families, such as Buxaceae, Trochodendraceae, and Tetracentraceae, which are

BUXACEAE Buxaceae belong to a grade of families near the base of the eudicot tree (von Balthazar and Endress, 2002). This family includes four extant genera of tropical and temperate shrubs and trees with typically unisexual flowers that are organized into opposite and decussate inflorescences. The superior ovary is constructed of three carpels, each with two ovules; endosperm is fleshy. The number of stamens ranges from four to six. A review of the fossil record of the genus Buxus can be found in Z. Kvacˇek et al. (1982) and Köhler and Brückner (1989). Pollen that is morphologically similar to some species of the Buxaceae is known from the Cretaceous (Boltenhagen, 1963), with additional reports from the Eocene (Bessedik, 1983) and lower Miocene (Muller, 1981). Grains of Erdtmanipollis from the Campanian (Late Cretaceous) are polyporate and range from 25–35 μm in diameter; the exine is crotonoid-reticulate (Srivastava, 1972). Flowers assignable to the Buxaceae have been described from the Potomac Group of eastern North America (Drinnan et al., 1991). Spanomera mauldinensis is used for small unisexual flowers with four to five opposite and decussate tepals. Staminate specimens (FIG. 22.112) from the lower Cenomanian (Upper Cretaceous) of Maryland have tricolpate (FIG. 22.113), semitectate pollen in the 20 μm size range that conforms to the sporae dispersae genus Striatopollis (FIG. 22.114). Pistillate flowers are bicarpellate with an elongate suture. The affinities of Spanomera within the Buxaceae confirm the hypothesis that this family is closely associated to the lower Hamamelididae (Drinnan et al., 1991). Other tricolpate-striate pollen of this type has been found in staminate

chapter 22

flowering plants

931

flowers of Lusistemon, and attached to the stigmatic surface of pistillate flowers named Lusicarpus (Pedersen et al., 2007). Although these Early Cretaceous flowers from Portugal share a number of characters with early diverging lineages of eudicots, they are perhaps most similar to those of the Buxaceae. Leaves and infructescences of Sinocarpus decussatus from a Lower Cretaceous site in China are thought to have some relationship to the Buxaceae, but characters in the fossils are also seen in members of the Ranunculaceae and Myrothamnaceae (Leng and Friis, 2006). TROCHODENDRACEAE Plants included in this family are unique in the absence of vessels in the wood and the occurrence of idioblasts in the leaves. Some authorities include Cercidiphyllum in the order. Both Trochodendron (FIG. 22.115) and Tetracentron are monotypic and restricted to southeast Asia today, but were more widely distributed early in the Cretaceous based on fossils (Manchester et al., 1991). In Trochodendron the leaves are evergreen and elliptical to obovate, whereas in Tetracentron leaves are linear and deciduous. The fossil record of the group includes characteristic leaves that first appear in the Cretaceous (Hickey and Doyle, 1977) and wood (Vozenin-Serra and Pons, 1990). Crane et al. (1991) described Nordenskioeldia (sometimes spelled Nordenskioldia) borealis, a trochodendralean infructescence from the Paleocene of western North America (Montana, Wyoming, and North Dakota). Associated with the fruits are leaves of Zizyphoides flabella. The plant is reconstructed as bearing simple petiolate leaves with rounded apices and entire margins. Venation is actinodromous. Some leaves of this type have been included in the genera Zizyphoides, Cercidiphyllum, and Joffrea (Pigg et al., 2007). Associated shoots possess tracheids in the secondary xylem, but lack vessels. Reproductive organs consist of woody, schizocarpic fruits arranged in a whorl on an infructescence axis. Fruits of Nordenskioeldia are infructescences (FIG. 22.116) with multiple fruitlets, each containing a single seed (FIG. 22.117). Although several features of the plant remain to be determined, such as pollen and floral structure, Nordenskioeldia (FIG. 22.118) expands our knowledge of the diversity of Cenozoic members of the Trochodendraceae. The genus has also been recorded from the Paleocene of China (G. Zhang and Guo, 2006) and the middle Eocene flora of Republic, Washington (USA) (Pigg et al., 2001). Palmately veined Trochodendron nastae leaves have been reported from this same flora, as well as Nordenskioeldia sp. infructescences and Zizyphoides leaves (Pigg et al., 2001). It is suggested that Nordenskioeldia may have been a floodplain colonizer (Crane et al., 1991).

Figure 22.115 Infructescence of Trochodendron sp. (Eocene).

Bar  1 cm.

Lhassoxylon is a wood type from the upper Aptian of Lhasa, Tibet, that is reported to have both gymnospermous and angiospermous characters. Tracheids are arranged in irregular clusters (FIG. 22.119) like those in the extant genus Tetracentron (Vozenin-Serra and Pons, 1990). Secretory canals, however, are similar to those described in Araucariopitys, a stem type found in association with leaves of the Czekanowskiales. Suzuki et al. (1991) described

932

paleobotany: the biology and evolution of fossil plants

Figure 22.117 Section of Nordenskioeldia borealis infruct-

escence with numerous fruitlets, each containing a single seed (Paleocene). Bar  3 mm. (Courtesy P. R. Crane.)

22.116 Infructescences (Paleocene). Bar  2 cm.

Figure

of

Nordenskioeldia

sp.

Miocene wood from central Japan under the binomial Tetracentron japonoxylum. Two types of tracheids are produced; one type (usual) has typical scalariform-bordered pits on the radial walls. The wood also contains sporadic radial files of broad tracheids with alternate bordered pits. On the radial walls of these tracheids are half-bordered pits to ray cells, but there are no pits to the “usual” tracheids. These unusual tracheids are found only in the extant genus Tetracentron. Wood has also been described from several Eocene and Oligocene localities, and pollen from Miocene Tetracentron fruits (Manchester and Chen, 2006; Grímsson et al., 2008). Infructescences consisting of branched

Figure 22.118 Nordenskioeldia sp. leaf (Paleocene). Bar  1 cm.

peduncles bearing single fruits with six to many fused carpels from the Eocene of Canada are called Trochodendron drachukii (Pigg et al., 2007). Leaves of Tetracentron hopkinsii from the Eocene Allenby Formation of British Columbia are 9.5 cm long with a cordate base and actinodromous venation. Other fossil occurrences of Tetracentron leaves are reviewed by Ozaki (1987).

chapter 22

flowering plants

933

Figure 22.119 Cross section of Lhassoxylon aptianum show-

ing parenchyma in the wood (Cretaceous). Bar  4 mm. (From Vozenin-Serra and Pons, 1990.)

Proteales

NELUMBONACEAE This group consists of aquatic, perennial herbs with rhizomatous stems that bear emergent, concave to peltate leaves; stems contain laticifers. Flowers are large and the fruit is an aggregate of nuts sunken in an accrescent receptacle, that is, one that increases in size after flowering. In early systematic treatments Nelumbo was included within the Nymphaeaceae; however, molecular studies now place the group within a grade of eudicots that contains Platanus (Hoot et al., 1999). This placement is also supported by the polysymmetric floral development, anatomy, and carpel closure in Nelumbo, which is similar to that in several other basal dicots (Hayes et al., 2000). Moreover, extant Nelumbo pollen development is different from that seen in both the Nymphaeaceae and lower eudicots (Kreunen and Osborn, 1999). Leaves of Nelumbo have been reported from the Albian of Portugal that measure up to 30 cm in diameter, but these may represent members of the Nymphaeaceae. Nelumbites is used for orbicular leaves with actinodromous primary venation from the Potomac Group (Lower Cretaceous) (Upchurch et al., 1994). Isolated tepals and structures that may be receptacles containing nuts occur in the same deposits. Other Late Cretaceous and Cenozoic leaves are given the name Nelumbium (FIGS. 22.120, 22.121), suggesting affinities with Nelumbo (e.g., Matsuo, 1962). Reproductive structures

Figure 22.120 Nelumbium buchii, leaf of a waterlily (Miocene). Bar  10 cm. (Courtesy BSPG.)

and leaves of Nelumbo puertae have been reported from the Upper Cretaceous (Campanian–Maastrichtian) of Patagonia (FIGS. 22.122, 22.123) (Gandolfo and Cúneo, 2005). The leaves are orbiculate with the petiole centrally inserted.

934

paleobotany: the biology and evolution of fossil plants

22.123 Nelumbo puertae leaf showing orbicular, symmetrical lamina (Cretaceous). Bar  2 cm. (Courtesy M. A. Gandolfo.) Figure

Figure 22.121 Portion of a Nelumbium buchii leaf (Oligocene). Bar  6 cm. (Courtesy BSPG.)

R

Figure 22.124 Nelumbo puertae showing primary veins form-

ing brochidodromous arches (Cretaceous). Bar  5 mm. (Courtesy M. A. Gandolfo.)

Figure 22.122 Top view of Nelumbo-like fossil receptacle (R)

showingseveral fruits (Cretaceous). Bar  5 mm. (Courtesy M. A. Gandolfo.)

Twelve or more primary veins dichotomize at least once to form brochidodromous arches; areoles are four to five sided and well developed (FIG. 22.124). Present in the same rocks are portions of what are interpreted as receptacles with

chapter 22

flowering plants

935

Small plant with leaves and roots suggesting affinities with the Nymphaeales (Cretaceous). Bar  3 cm. (Courtesy B. A. R. Mohr.)

Figure 22.125

casts of nuts (FIG. 22.122). Although the fossil record of the group is considered to be extensive, the Patagonian fossil represents the first report of the genus from the Southern Hemisphere. Small plants (FIG. 22.125) with possible nymphaealean affinities have also been reported from the Crato Formation of Brazil (Mohr and Friis, 2000). PROTEACEAE This family, with 75 genera of shrubs and trees, is one of the most diverse in the Southern Hemisphere. The plants are characterized by exstipulate leaves, four-merous flowers in the form of a raceme, with paired, unisexual or bisexual flowers, each with a single carpel and fewer than eight stamens. Petals are typically absent. Mycorrhizae are believed to be absent from the roots. The family is restricted to the Southern Hemisphere today and is believed to have originated in Australia as trees in mesothermic, closed forests during the mid-Cretaceous (Johnson and Briggs, 1975; Sauquet et al., 2007). Although a diverse assemblage of follicular fruits from the Late Cretaceous of Sweden have some features in common with extant members of the Proteaceae, they also possess characters seen in members of the Cercidiphyllaceae (Leng et al., 2005), underscoring the difficulty of assigning fossil plants to modern families, especially fossils from the Cretaceous. Lomatia includes proteaceous leaves described from the upper Eocene–Oligocene that are morphologically identical to the living genus, which is a Tasmanian endemic (Carpenter and Hill, 1988). The leaves of the fossil are bipinnate with conspicuous lobes and brochidodromous venation. A thick cuticle and reduced leaves imply xeromorphic conditions, but are believed to have reflected edaphic conditions,

Figure 22.126 Palmately lobed leaf, Parafatsia subpeltata

(Eocene). Bar  5 cm. (Courtesy R. J. Carpenter.)

that is, available soil moisture, since the climate is interpreted as cool, wet, and humid (Beadle, 1968). Parafatsia subpeltata is a palmately lobed leaf (FIG. 22.126) from the middle Eocene of Australia characterized by actinodromous primary venation (Carpenter et al., 2006). Although the leaf morphology is unique when compared to extant members of the family, the epidermal anatomy, including annular trichome bases and unique ornamentation of the epidermis, align Parafatsia with the Proteaceae. Leaves described from the Eocene of Argentina include Lomatia, Embothrium, Orites, and Roupala (Gonzalez et al., 2007). Serlin (1982) described proteaceous leaves from the Albian of northwest Texas in which the petiole and blade are not well differentiated. On the lower surface of Tenuiloba canalis are two grooves containing paracytic stomata (FIG. 22.127). Wood from the same locality is called Aplectotremas and consists of radial groups of vessels separated by files of fibers. The rays are narrow and paratracheal parenchyma is present in small amounts. Features of the ring-porous wood of A. halistichum are unlike those of extant angiosperms, but similar to what has been hypothesized as proto-proteaceous wood (Johnson and Briggs, 1975). Silicified wood from the Eocene of Santa Cruz Province, Argentina, is believed to represent a proteaceous root and is assigned to the genus Lomatia (Ancibor, 1989). The wood has scarce paratracheal parenchyma and vessels in tangential rows.

936

paleobotany: the biology and evolution of fossil plants

Figure 22.127 Stomatal patterns: A. anomocytic; B. anisocytic; C. paracytic; D. diacytic; and E. actinocytic.

22.128 Beaupreaidites verrucosus pollen grain (Miocene). Bar  20 μm. (From Pocknall and Crosbie, 1988.)

Figure 22.129 Beauprea montana pollen grain (Extant). Bar  20 μm. (From Pocknall and Crosbie, 1988.)

Beaupreaidites (FIG. 22.128) is used for proteaceous pollen that is morphologically similar to the isopolar, tricolpoid pollen of extant Beauprea (FIG. 22.129) (Pocknall and Crosbie, 1988). The fossil grains are found as early as the Late Cretaceous in New Zealand and Australia; the extant genus is endemic to New Caledonia.

Musgraveinanthus is a Cenozoic inflorescence axis from Australia bearing 30 pairs of perfect flowers, each subtended by three bracts (Christophel, 1984). Flowers possess three elongated hypogynous glands, a superior ovary, and pollen that is described as being diporate. These Eocene flowers are most similar to extant genera in the tribe Banksiae, subtribe Musgraveinae.

Figure

chapter 22

flowering plants

937

Figure 22.130 Platanus wyomingensis (Eocene). Bar  2 cm. (Courtesy Denver Museum of Nature.)

PLATANACEAE Modern members of this family include temperate to tropical monoecious trees consisting of 9–10 species in the genus Platanus, the plane tree or sycamore. Leaves in the family are palmately lobed with encircling stipules and unisexual flowers; fruits are achenes. The pollen record for the family is extensive and can be traced from the Early Cretaceous, where there are a number of distinct types that are not known in modern Platanus (FIG. 22.130) (Denk and Tekleva, 2006). Palmately lobed leaves in the mid–late Aptian have also been used to suggest the Early Cretaceous evolution of this group. One of the common leaf types in the Paleocene of Scotland is Platanites (FIG. 22.131) (Crane et al., 1988). The leaves are compound with a trilobed, terminal leaflet and a pair of asymmetric lateral leaflets. Manchester (1986) has been able to reconstruct one fossil member of the family, Macginitiea, from the middle–upper Eocene of Oregon. The leaves are palmately lobed, usually with five to seven lobes, and characterized by a pattern of inverted V’s or chevrons formed by secondary veins (Wolfe and Wehr, 1987) (FIG. 22.132). The leaves are consistently associated with pistillate and staminate inflorescences of Macginicarpa and Platananthus, respectively (discussed below). Silicified petioles show numerous collateral vascular bundles arranged in a ring like those in Platanus occidentalis. Stipules are absent in the fossil. Plataninium is a relatively common fossil wood type at many Upper Cretaceous and Cenozoic localities in Europe (Meijer, 2000; Poole et al., 2002) and North America (Page, 1968; Wheeler et al., 1995). Anatomically, the fossil wood is almost identical to the modern Platanus, but lacks simple

Figure 22.131 Platanites hebridicus (Paleocene). (From Crane

et al., 1988.)

Figure 22.132 Wesley C. Wehr. (Courtesy K. B. Pigg.)

vessel perforations and has wider rays (FIG. 22.134) (Wheeler et al., 1977; Wheeler, 1991). Other permineralized woods similar to the modern Platanus have been assigned to the genus Platanoxylon (FIG. 22.133) (Süss and Müller-Stoll, 1977; Süss, 1980a; Selmeier, 1996) (FIG. 22.135). The morphogenus Spiroplatanoxylon is used for Platanus-like woods that are characterized by vessels with spiral thickenings (Süss, 2007). Since spiral thickenings are unknown in Platanus, the

938

paleobotany: the biology and evolution of fossil plants

Cross section of Platanoxylon sp. (Paleocene). Bar  400 μm. (Courtesy E. A. Wheeler.)

Figure 22.133

Figure 22.134 Cross section of secondary xylem of Plataninium haydenii showing large (center) and small (left) rays (Eocene). Bar  35 μm. (Courtesy E. A. Wheeler.)

Figure 22.135 Alfred Selmeier.

relationship between Platanus and Spiroplatanoxylon remains uncertain. Miocene Platanus wood has been described from Hungary that shows characteristic medullary spots or socalled pith flecks (Markflecken), which are due to the mining of insect larvae of Palaeophytobia platani (Agromyzidae, Diptera) in the cambium (Süss and Müller-Stoll, 1975). The Aquia plant is an excellent example of uniting detached parts of fossil plants to develop a concept of the entire organism. This reconstruction of an early platanoid from the early–middle Albian Patapsco Formation (Potomac Group) of northern Virginia (USA) is based on consistent co-occurrence at the locality, cuticular features, and pollen on the carpels (Crane et al., 1993). Inflorescences of Platanocarpus brookensis consist of elongate axes that bear pistillate flowers with a perianth of tepals. Staminate flowers of A. brookensis have long filaments bearing anthers with valvate dehiscence. In situ pollen is columellate and tricolpate and is also found on the Platanacarpus carpels. Sapindopsis (FIG. 22.136) leaves were related to the flowers on cuticular features. The similarity in cuticle structure, however, may be equivocal since certain features are found in many angiosperm genera (Carpenter et al., 2005). Another member of the Platanaceae for which isolated parts have been united is Platanus neptuni from the

chapter 22

flowering plants

939

Figure 22.137 Cross section of a floret of Macginicarpa

glabra showing five carpels (Eocene). Bar  2 mm. (Courtesy S. R. Manchester.)

Figure 22.136 Pinnately compound Sapindopsis (Cretaceous). Bar  3 cm. (Courtesy G. R. Upchurch.)

leaf

upper Eocene to upper Miocene of Europe (Z. Kvacˇek and Manchester, 2004). The P. neptuni plant includes globose infructescences, staminate and pistillate flowers, and simple to quinquefoliate leaves. Infructescences and leaves bear large, peltate, glandular trichomes. Hamatia elkneckensis is used for staminate inflorescences from the late Albian Potomac Group of North America (Pedersen et al., 1994), and Sarbaya for staminate heads from the mid-Cretaceous of Kazakhstan (Krassilov and Shilin, 1995). A staminate inflorescence from the Upper Cretaceous of the Amur region of Russia combines features of the Platanaceae and Hamamelidaceae (Maslova et al., 2007). Bogutchanthus has tetramerous flowers with staminodes, crescent-shaped pollen sacs, and pantocolpate pollen. The flowers of B. laxus are found associated with leaves of the Platanus type. Macginicarpa (FIG. 22.137) is used for pistillate inflorescences that were borne in helically arranged clusters on an axis (Manchester, 1986). Each floret is made up of five free, conduplicate carpels subtended by a perianth. Silicified specimens from the Eocene Clarno nut beds reveal features about the carpel wall. The mature fruits (FIG. 22.138) are achenes, but lack dispersal hairs common in living Platanus fruits. The staminate heads are called Platananthus and are constructed of numerous florets of five anthers each. The pollen is tricolpate, ranges from 11–16 μm in diameter, and is tectate-columellate. Dispersed stamens are called Macginistemon. A comparison of vegetative and reproductive features in the family suggests that the fossil was less

22.138 Infructescence of Macginicarpa (Eocene). Bar  5 mm. (From Manchester, 1986.)

Figure

glabra

specialized for wind dispersal of fruits and pollen than modern taxa (Manchester, 1986; Friis et al., 1988). Another infructescence from the Cenomanian Dakota Formation of Kansas is Caloda (Dilcher and Kovach, 1986). This taxon consists of an axis bearing smaller, helically arranged branches that terminate in receptacles (FIG. 22.139). Each receptacle bears 35–50 free, indehiscent conduplicate carpels 2 mm long. When compared with extant flowers, Caloda shares the greatest number of features with members of the Platanaceae. Geologically older platanaceous flowers have been recovered from the Lower and Upper Cretaceous of North America and Sweden (Crane et al., 1986). They include both staminate and pistillate flowers (Friis et al., 1988;

940

paleobotany: the biology and evolution of fossil plants

22.140 Leaf of Mahonia simplex (Oligocene). Bar  1 cm. (From Axelrod, 1998b; courtesy S. R. Manchester.)

Figure

Figure 22.139 Restoration of Caloda delevoryana axis bear-

BERBERIDACEAE The plants in the Berberidaceae include woody shrubs, some with evergreen, persistent leaves. Some shoots have elongate spines representing modified leaves. Flowers are bisexual and the fruit is edible. Fossil leaves of Mahonia (FIG. 22.140) (Meyer and Manchester, 1997; Z. Kvacˇek and Hurník, 2000; Ramírez and Cevallos-Ferriz, 2000a) and Berberis have been described from the Miocene (Z. Kvacˇek and Erdei, 2001) with specimens of B. kymeana occurring as lanceolate leaves that have widely spaced teeth along the margin. Some specimens of Lomatites that were initially included in the Proteaceae are believed to represent members of the Berberidaceae (Z. Kvacˇ ek and Erdei, 2001).

ing several carpels. (Courtesy D. L. Dilcher.)

Magallón-Puebla et al., 1997). Specimens of Platananthus potomacensis include staminate flowers surrounded by elongate tepals. The pollen grains are very small (9–13 μm) and similar to the sporae dispersae taxon Tricolpites minutus. Platanocarpus is the name used for pistillate inflorescences and infructescences that lack trichomes on the carpel and have a poorly developed style (Friis et al., 1988). Ranunculales

The Ranunculales are interpreted to be sister to the eudicots (Soltis et al., 2005). Within the families in this order are plants with wide multiseriate rays in the wood, sieve-tube plastids of the S-type, and young stems with separate bundles.

RANUNCULACEAE Bisexual flowers from the middle Albian of Kazakhstan are thought to be related to the members of the Ranunculaceae and Paeoniaceae (Krassilov et al., 1983). Specimens of Hyrcantha consist of bracteate inflorescences bearing three to five ascidial carpels, each 7 mm long. Stamens are about half as long as the carpels, but are poorly preserved. Another Cretaceous flower with possible affinities in this order is Teixeiraea lusitanica (von Balthazar et al., 2005). The staminate flower is subtended by bracts and includes stamens of two sizes; sepal-like and petallike tepals constitute the perianth. Pollen grains are small (15– 20 μm) and tricolpate, with a tectate-columellate wall. Although the characters present in these unisexual flowers can be found today in several families of the Ranunculales, the character of stamens of different sizes is not found in this order. The authors

chapter 22

Figure 22.141 Fractured surface of fruit of Paleoactaea nagelii showing two rows of seeds (Paleocene). Bar  5 mm. (Courtesy K. B. Pigg.)

flowering plants

941

Figure 22.142 Tricolpites reticulatus pollen grain polar view (Cretaceous). Bar  5 μm. (From Jarzen and Dettmann, 1989.)

suggested that T. lusitanica may represent an extinct phylogenetic lineage within the crown group of the order. Permineralized Paleocene fruits were described as Paleoactaea (Pigg and DeVore, 2005b). The bilaterally symmetrical specimens are 13 mm long with the anatropous seeds borne in two rows (FIG. 22.141). Morphologically and anatomically the fruits are nearly identical to those produced by the extant plant Actaea, a herbaceous wildflower from northern temperate regions (Pigg and DeVore, 2005b). A similar seed is known from the early Eocene London Clay flora.

Core eudicots Gunnerales

The Gunnerales has been suggested to be the sister to the remaining other core eudicots, and includes two families, Gunneraceae and Myrothamnaceae, with a total of 42 species (Soltis et al., 2003). GUNNERACEAE This is a monotypic family, included by some authors in the Saxifragales (Fuller and Hickey, 2005), with 40 species in the single genus Gunnera. The plants are perennial herbs and are mainly restricted today to the Southern Hemisphere. There is considerable morphological diversity in Gunnera, ranging from small, stoloniferous herbs to gigantic, rhizomatous herbs with massive leaves; symbiotic cyanobacterial associations make it possible for the plants to invade habitats poor in nutrients (Fuller and Hickey, 2005). The pollen of Gunnera is distinct and can be traced back to the Late

Figure 22.143 Gunnera macrophylla pollen grain polar view

(Extant). Bar  5 μm. (From Jarzen and Dettmann, 1989.)

Cretaceous (Wanntorp et al., 2004). Tricolpites reticulatus is the name given to fossil Gunnera pollen (Jarzen, 1980). This fossil pollen (FIG. 22.142) is indistinguishable from that of modern Gunnera (FIG. 22.143), and suggests that the family was far more cosmopolitan beginning in the Late Cretaceous (Jarzen and Dettmann, 1989). Caryophyllales

PHYTOLACCACEAE Plants included in this family have alternate leaves with small, typically bisexual flowers. The perianth occurs as a

942

paleobotany: the biology and evolution of fossil plants

Figure 22.144 Permineralized infructescence of Coahuilacarpon phytolaccoides (Cretaceous). Bar  4 . (Courtesy S. R. S. Cevallos-Ferriz.)

single whorl of four to five free segments. The majority of the plants in this family are native to tropical America and the West Indies. Coahuilacarpon is used for ovoid, multiple infructescences (FIG. 22.144) from the Upper Cretaceous of Mexico (Cevallos-Ferriz et al., 2008). Specimens are up to 4.3 cm long and may include 30 berries. The fossils are permineralized and thus provide the opportunity to detail tissue systems in the seeds. This discovery represents the oldest record of the family and demonstrates the existence of the group in the Late Cretaceous. Saxifragales

Using both extant and fossil taxa in a combined morphological and molecular cladistic analysis, Hermsen et al. (2006a) suggested that the order Saxifragales is not supported by any morphological synapomorphies that might be recognized in fossils. Although this may imply that the major lineages were already well diverged by the Late Cretaceous, additional fossils and the increased resolution they bring to morphological character placement may alter this hypothesis. CERCIDIPHYLLACEAE Today Cercidiphyllum is a dicotyledonous tree restricted to eastern Asia, but it was a common floral component at middle and high latitudes in the Northern Hemisphere during

Figure 22.145 Suggested reconstruction of Joffrea speirsii

showing pistillate inflorescences (upper figure) and short shoots bearing mature infructescences. (From Crane and Stockey, 1985.)

the Late Cretaceous (Mai and Walther, 1983). Zavada and Dilcher (1986) used pollen characters in a cladistic analysis of 20 families in the Hamamelidae, and they suggested that the Cercidiphyllaceae and Platanaceae are two of the more primitive families in the group. Cercidiphyllaceous leaves are petiolate and generally elliptical, with the base ranging from acute to rounded. The margin is crenate and often glandular, with actinodromous venation (Chandrasekharam, 1974). Abundant specimens from the upper Paleocene Paskapoo Formation near Red Deer, Alberta, Canada, form the basis of a reconstruction of the Cercidiphyllum-like plant Joffrea (Crane and Stockey, 1985). Joffrea speirsii (FIG. 22.145), a plant named after the discoverer of the site and avid amateur collector, Betty Speirs (FIG. 22.146), is reconstructed from six different disarticulated plant parts and indicates the presence of both long and short shoots (FIG. 22.147). Attached to the short shoots are typical Cercidiphyllum-type leaves (FIG. 22.148), as well as pistillate and staminate inflorescences containing 40 ellipsoidal carpels (FIG. 22.149). One of the more spectacular aspects of this report is the presence of numerous dispersed winged seeds 1 cm long and in situ seedlings (FIG. 22.150). The seedlings are typically preserved in growth position and at varying developmental stages (Stockey and Crane, 1983). Germination

chapter 22

flowering plants

Figure 22.148 Leaf of Joffrea speirsii Bar  2 cm. (From Stockey and Crane, 1983.)

943

(Paleocene).

Figure 22.146 Betty Speirs. (Courtesy R. A. Stockey.)

of Joffrea seedlings was epigeal. Based on the large number of seedlings at the collecting site, it is suggested that germination was rapid and synchronous, a reproductive strategy sometimes termed r-selection, which is typically found in colonizer species that are adapted to living in disturbed habitats. Nyssidium is another Cercidiphyllum-like plant from the Cenozoic of England (Crane, 1984). The leaves are called Trochodendroides (FIG. 22.151). Inflorescences were apparently produced on the long shoots, unlike the condition in Joffrea. It is speculated that Nyssidium was wind pollinated and that the seeds were dispersed by water or wind. The condensed inflorescences in the modern Cercidiphyllum are hypothesized to have evolved from the elongate type seen in Joffrea, accompanied by a change in suture orientation (Crane and Stockey, 1986). The earliest fossil woods assignable to the genus Cercidiphyllum with confidence come from the Oligocene of northwestern Bohemia (Czech Republic) and have been described as Cercidiphylloxylon kadanense (Sakala and Privé-Gill, 2004).

Figure 22.147 Immature staminate inflorescence of Joffrea

speirsii (Paleocene). Bar  10 mm. (Courtesy R. A. Stockey.)

HALORAGACEAE This is a small family that includes 120 species in 8 genera of mostly aquatics; a few are small trees. The family has been interpreted as having a minimum age of 65 Ma, currently the Cretaceous–Paleogene boundary (Fishbein et al., 2001), based on the following fossils. Tarahumara sophiae is known from permineralized specimens collected

944

paleobotany: the biology and evolution of fossil plants

Figure 22.150 Seedling of Joffrea speirsii showing a pair

of cotyledons (arrows) and opposite first leaves (Paleocene). Bar  1 mm. (From Crane and Stockey, 1985.)

Figure 22.151 Leaf of Trochodendroides sp. (Paleocene).

Bar  1 cm.

Figure 22.149 Infructescence of Joffrea speirsii (Paleocene).

Bar  2 cm. (From Crane and Stockey, 1985.)

from the Upper Cretaceous (Campanian–Maastrichtian) Tarahumara Formation of Mexico (Hernández-Castillo and Cevallos-Ferriz, 1999). Flowers are small and unisexual (FIG. 22.152) with the gynoecium composed of four carpels that are united at the base. Fruits contain a single, pyriform ovule. Obispocaulis is used for stems from the same locality that are believed to be parts of the same plant. The plant is

chapter 22

flowering plants

945

Figure 22.153 Infructescence of Microaltingia apocarpela

showing several broken carpels containing seeds (Cretaceous). Bar  450 μm. (From Z. Zhou et al., 2001.)

Figure 22.152 Longitudinal section of Tarahumara sophiae flower showing central carpel with seed and two lateral carpels (Cretaceous). Bar 0.5 mm. (Courtesy S. R. S. Cevallos-Ferriz.)

interpreted as a nearshore herb in which the aerial parts floated on the water surface. The occurrence of these permineralized vegetative and reproductive remains in the latest Cretaceous suggests that diversification within the family is much older than previously thought. HAMAMELIDACEAE This family consists of 30 genera and 120 species of trees and shrubs with leaves that are typically alternate and range from simple to palmately lobed. The inflorescence is compact with bisexual or unisexual flowers, and is known from fossil material from the Late Cretaceous onward (Crane and Blackmore, 1989a, b; Magallón et al., 2001). The sweet gum tree Liquidambar is in this family, subfamily Altingioideae. Leaf morphology ranges from palmately lobed in Liquidambar to evergreen with pinnate venation in Altingia, which is sometimes included in its own single family, Altingiaceae (Ickert-Bond et al., 2005, 2007). Fruits are bicarpellate, lack a perianth, and dehisce to release the seeds. Fossil inflorescences of Liquidambar are known from the Upper Cretaceous of

North America. Specimens of Microaltingia consist of pistillate inflorescences and fruits (Z. Zhou et al., 2001). Florets are 1.2 mm in diameter with the gynoecium surrounded by two to three whorls of sterile phyllomes. Each carpel contains several seeds, and tricolpate pollen is found on the stigmas. Pollen is 10 μm in diameter and tricolpate. Evacarpa (late Paleocene–early Eocene of Russia) consists of globose, pistillate heads with tightly packed florets (Maslova and Krassilov, 1997). Carpels are paired and contain numerous seeds. Both Evacarpa and Microaltingia (FIG. 22.153) share features with Liquidambar and Altingia, as well as inflorescences of platanaceous affinity. Because fossils like these demonstrate a mosaic of characters (Ickert-Bond et al., 2007), they provide an important source of information useful in tracing the phylogenetic position of early angiosperms. Liquidambar changii is a Miocene taxon used for anatomically preserved infructescences (FIG. 22.154) (Pigg et al., 2004). Interestingly the characters found in these specimens are more similar to the extant taxon L. acalycina from eastern Asia than to forms living in North America, suggesting a Beringian biogeographic corridor during the Miocene. Silicified Liquidambarlike wood, assigned to Liquidambaroxylon, is known from the Oligocene of the Czech Republic (Sakala and Privé-Gill, 2004) and the Miocene of southern Germany (Selmeier, 2002c), among other localities. ITEACEAE The two modern genera in this family have small simple leaves with stipules. Branches have a chambered pith and

946

paleobotany: the biology and evolution of fossil plants

Figure 22.154 Section of Liquidambar changii infructescence

(Miocene). Bar  5 mm. (Courtesy K. B. Pigg.) Figure 22.156 Lateral view of Divisestylus brevistamineus

showing styles (Cretaceous). Bar  350 μm. (Courtesy K. Nixon.)

by a cladistic analysis that uses both morphological and molecular data (Hermsen et al., 2003).

Figure 22.155 Lateral view of Divisestylus brevistamineus

showing papillae on petals and adpressed styles. Sepal tips are broken off. (Cretaceous). Bar  100 μm. (Courtesy K. Nixon.)

flowers are pentamerous with a bicarpellate gynoecium. A Late Cretaceous (Turonian) charcoalified flower from the Raritan Formation of New Jersey (USA) is tentatively placed in the family (Hermsen et al., 2003). Divisestylus brevistamineus (FIGS. 22.155, 22.156) has styles that do not extend above the tips of the sepal lobes, but are inserted opposite the lobes; blunt trichomes ornament the styles. The presence of both fused carpels and stigmas with free styles suggests affinities with the Iteaceae. This placement is also supported

SAXIFRAGACEAE Many of the plants in this family occur in north temperate regions as herbs, shrubs, and a few small trees. The leaves are simple and exstipulate. Flowers are bisexual, regular and borne in racemes. Sepals and petals are free and borne in fours or fives. Stamens are usually produced in two whorls. Although there are relatively few fossils known from this family, there is a report of two species from the Upper Cretaceous of southern Sweden (Friis and Skarby, 1982). Scandianthus includes two species, S. costatus and S. major. The genus includes small (2.2 mm long), bisexual flowers with sepals and petals in fives. The gynoecium is formed by two fused carpels. Insect pollination is suggested based on the presence of a nectary disk. Similar floral morphology is found in Vahlia of the Vahliaceae, a group now included in the Euasterids I.

Rosids The rosids demonstrate extensive floral diversity and represent the largest clade included in the eudicots (Schonenberger and von Balthazar, 2006). It is estimated that the rosids include more than a third of all angiosperms (Soltis et al., 2005) and 68,000 species (Magallón et al., 1999). Synapomorphies remain to be well identified for the group, but in general

chapter 22

flowering plants

947

Figure 22.157 Vitis teutonica, leaf (Cenozoic). Bar  1 cm.

(Courtesy BSPG.)

features such as two-to-several whorls of stamens, reticulate pollen, and nuclear endosperm are shared by most members. To date the oldest fossils assignable to the group are reported as Turonian to late Santonian (Crepet et al., 2004). Many of the rosid lineages appear to have diverged during the Cretaceous. This is based on molecular sequence data and to some degree fossils, especially floral remains. The lack of a more robust fossil record has been hypothesized as a result of taphonomic, ecological, and distributional factors (Crepet et al., 2004). VITACEAE This family appears closely related to the large, well-supported rosid clade (Soltis et al., 2000; Ravi et al., 2007), and includes 15 genera and 700 species. Members are mostly tropical and subtropical, although Vitis (grapevine) is worldwide. Plants in this family are mainly lianas and possess adaptations for climbing. The leaves are alternate and may be simple or compound (FIG. 22.157). Flowers are regular with four to five petals. The earliest unequivocal evidence of the family is the late Paleocene (Chen and Manchester, 2007). Compressed seeds (FIG. 22.158) attributed to this family are known as early as the Paleocene (Chen and Manchester, 2007), with most included within the modern genus Vitis (Tiffney and Barghoorn, 1976). Cevallos-Ferriz and Stockey (1990b) have described several anatomically preserved seeds, for example Ampelocissus similkameenensis, that they include within the family from the middle Eocene Princeton chert locality in British Columbia, Canada. The ruminate seeds appear W-shaped in cross section and have a

Figure 22.158 Vitis seed (Miocene). Bar  2 mm. (Courtesy

M. S. Zavada.)

Figure 22.159 Cross section of Ampelocissus similkameenensis

seed (Eocene). Bar  750 μm. (From Cevallos-Ferriz and Stockey, 1990b.)

five-zoned integument (FIG. 22.159). Additional seed morphotypes of Ampelocissus (FIG. 22.160) are reported in Chen and Manchester (2007) who indicate a far wider geographic range of the genus during the warmer Eocene and Miocene. Silicified vitaceous wood has been described from the lower Eocene of England (Poole and Wilkinson, 2000).

948

paleobotany: the biology and evolution of fossil plants

Figure 22.160 Various views of fossil (Eocene) Ampelocissus bravoi seeds in upper panel compared with extant Ampelocissus cavicaulis seeds. Bar  2 mm. (From Chen and Manchester, 2007; courtesy S. R. Manchester.)

Figure 22.161 Cross section of Decodon allenbyensis fruit showing seeds (Eocene). Bar  1 mm. (Courtesy K. B. Pigg.)

Myrtales

LYTHRACEAE Many of the plants in this family (30 genera) live in mesic to wet habitats and are generally characterized by radially symmetrical flowers with brightly colored petals and twice as many stamens as sepals. Fruits, seeds, roots, and stems from the Eocene Princeton chert of British Columbia are included in Decodon allenbyensis (Cevallos-Ferriz and Stockey, 1988; Little and Stockey, 2003). Permineralized leaves from the same site are similar to extant leaves of Duabanga (Little et al., 2004), but are considered to be the leaves of Decodon allenbyensis (Little and Stockey, 2003). They have a C-shaped midvein surrounded by a sheath of sclerenchyma. Other lythraceous fossils include fruits (FIG. 22.161) and seeds (Tiffney, 1994; Pigg and Devore, 2005a; Estrada-Ruiz and Cevallos-Ferriz, 2007), wood (Little and Stockey, 2006), extraxylary tissues (Little and Stockey, 2003), and pollen (Graham and Graham, 1971). Permineralized fossil fruits consisting of capsules with five to seven locules are given the name Shirleya grahamae (Pigg and DeVore, 2005a). In the locules (FIG. 22.162) of these middle Miocene fossils are numerous distally winged seeds. Based on detailed anatomy, the fossil is most similar to extant fruits of Lagerstroemia. TRAPACEAE This family, which is sometimes included in the Lythraceae, contains only a single extant genus, Trapa (Watson and Dallwitz, 1992 onward). Members of this genus are floating plants with triangular, toothed leaves. The floating leaves are connected to an inflated petiole that increases buoyancy. Flowers are small and four petalled; fruits are nutlike and characterized by four sharp and barbed spines. Although

Figure 22.162 Section of Shirleya grahamae fruit showing packing of seeds (Miocene). Bar  5 mm. (Courtesy K. B. Pigg.)

vegetative remains of Trapaceae are rare as fossils, for example, the Miocene leaf Mikia pellendorfensis from Austria (Kovar-Eder et al., 2002), Trapa fruits (FIG. 22.163) have been recorded from numerous Cenozoic localities (Wójcicki et al., 1999; Wójcicki and Zastawniak, 2002; Graham et al., 2005). A closely related form has been given the generic name Hemitrapa (FIG. 22.164) (Gregor, 1982; Wójcicki and Kvacˇ ek, 2002; Kovar-Eder et al., 2005). MYRTACEAE This family (100 extant genera and 3000 species) of mostly trees and shrubs is found today principally in Australia and tropical America. Flowers are bisexual and often have two

chapter 22

flowering plants

949

Figure 22.165 Dispersed cuticle assigned to the Myrtaceae

(Cretaceous). Bar  20 μm. (Courtesy G. Upchurch.) alabamensis, Bar  4 mm. (Courtesy B. J. Axsmith.)

Figure

22.163 Trapa

nut

(Pliocene).

R

Figure 22.164 Hemitrapa Bar  2 cm. (Courtesy BSPG.)

heissigii,

fruit

(Miocene).

bracts at the base. Calyx and corolla consist of four to five free or united parts, and the stamens are numerous. Leaves are generally opposite with entire margins and are characteristically glandular. The fruits may be in the form of a nut, berry, drupe, or capsule. The family is known from the Cretaceous (FIG. 22.165) onward and pollen has been described under the name Myrtaceidites. The grains are small, 12–23 μm in diameter, and triangular in outline (Krutzsch, 1969). Myrtaciphyllum is a simple late Eocene leaf 3 cm long with brochidodromous venation and anomocytic stomata on the abaxial surface (Christophel and Lys, 1986). These leaves are mummified and thus provide a large number of characters that can be used in their identification. In M. undulatum, hairs are present on both surfaces. Using 19 foliar characters in a numerical analysis, Christophel and Lys (1986) determined

Figure 22.166 Cross section of Paleomyrtinaea princetonensis seed showing integument and raphe (R) (Eocene). Bar  650 μm. (Courtesy K. B. Pigg.)

that the Australian fossils are not closely allied with any extant species. Another name used for myrtaceous leaf fossils is Myrciophyllum (Zastawniak, 1994). In Syzygioides, a morphotaxon that includes attached leaves, twigs, flowers, and fruits from the Eocene of North America, the leaves are decussate and opposite (Manchester et al., 1998). Similar foliar features also occur in Eucalyptus. Because of the wide variation in leaf architecture in the Myrtaceae, these authors suggested that records of the family should also, where possible, include flowers and remains of fruits. Flowers and fruits have also been reported from the Eocene of British Columbia as Paleomyrtinaea (FIG. 22.166) (Pigg et al., 1993) and South Australia as Tristaniandra (Greenwood et al., 2007). The flowers are pentamerous with a tricarpellate ovary, with sepals, petals, and stamens inserted along the rim of a hypanthium,

950

paleobotany: the biology and evolution of fossil plants

Figure 22.167 Diporites aspis pollen grain showing viscin threads (Miocene). Bar  20 μm. (Courtesy D. Pocknall.)

which remain in mature fruits. Myceugenelloxylon is used for silicified, diffuse-porous woods with heterocellular rays that are referred to the Myrtaceae (Poole et al., 2001). Other myrtaceous fossil woods have been described by Gottwald (1966), Kramer (1974), and Ragonese (1980), among others.

22.168 Corsinipollenites Bar  10 μm. (Courtesy M. S. Zavada.)

Figure

warrenii

(Miocene).

ONAGRACEAE This family includes 650 species of mostly perennial herbs with four-parted, bisexual, generally perfect, and often protandrous flowers. The fruit is a capsule and the pollen is characterized by protruding apertures and viscin threads (FIG. 22.167) (Skvarla et al., 1978). The fossil record of the family extends back to the Late Cretaceous (Maastrichtian), with numerous reports of pollen in Oligocene and younger rocks (Martin, 2003). One of these is Corsinipollenites (FIG. 22.168), a triporate grain that is 42 μm in diameter. Specimens of C. epilobioides from the Oligocene of New Zealand demonstrate a close morphological and ultrastructural similarity with pollen of the extant genus Epilobium (Daghlian et al., 1984).

Eurosids I (Fabids) Fabales

FABACEAE (LEGUMINOSAE) In comparison to the large size of the legume family (14,000 extant species), the fossil record is somewhat poorly documented. The group is believed to have evolved during the Late Cretaceous based on the well-defined subfamilies Caesalpinioideae, Mimosoideae, and Papilionoideae, that were in existence by the Cretaceous (FIGS. 22.169, 22.170)

Figure 22.169 Protomimosoidea buchananensis flowers with 10 stamens (Paleocene–Eocene). Bar  3 mm. (Courtesy W. L. Crepet.)

chapter 22

flowering plants

951

C P

O

Figure 22.170 Barnebyanthus buchananensis showing petal (P),

calyx lobe (C), and an ovary (O) (Paleocene–Eocene). Bar  1 cm. (From Crepet and Herendeen, 1992; courtesy W. L. Crepet.)

(Raven and Polhill, 1981; Calvillo-Canadell and CevallosFerriz, 2005). The fossil record of wood, leaves and leaflets, flowers, fruits, and pollen (Senesse and Cavagnetto, 1990) is extensive and dates back to the Late Cretaceous (Giraud and Lejal-Nicol, 1989; Cenozoic fossils reviewed in Herendeen et al., 1992). Exquisitely preserved specimens (FIG. 22.171) come from the Eocene Claiborne Formation of North America (Herendeen and Dilcher, 1990a). Leaves of extant Cercis have been used to examine the preservation potential of the upper pulvinus, a character that has been the basis for demonstrating the presence of the genus in the fossil record (Owens et al., 1998). These studies have identified other foliar characters which now make it possible to identify Cercis from the middle Miocene. Legume fruits and leaves assigned to the extant genus Wisteria have also been reported from the Miocene of China (Q. Wang et al., 2006). Bauhcis (FIG. 22.172) is used for leaves of Oligocene age that show a similar pattern of venation to that of extant species of Bauhinia and Cercis (Calvillo-Canadell and Cevallos-Ferriz, 2002). Duckeophyllum is a bipinnate leaf with at least three pairs of opposite pinnae, each producing 14 pairs of coriaceous, opposite leaflets (Herendeen and Dilcher, 1990a). Secondary veins are brochidodromous, and paracytic stomata are abaxial. Detached leaves with brochidodromous venation are placed in the genus Parvileguminophyllum (Call and Dilcher, 1994). Fruits of Eliasofructus are in the form of pods and may be up to 15 cm long. They contain numerous seeds

Figure 22.171 Mimosoid legume flower (Paleocene–Eocene).

Bar  2 mm. (Courtesy W. L. Crepet.)

Figure 22.172 Bauhcis moranii (Oligocene). Bar  5 mm.

(Courtesy S. R. S. Cevallos-Ferriz.)

based on the presence of transverse seed chambers. The wall of the pod is believed to have been relatively thin. Caesalpinia subgenus Mezoneuron is an extant group of 35 species of lianas and shrubs that occur in lowland rain forests and along river banks in tropical Africa, Asia,

952

paleobotany: the biology and evolution of fossil plants

Figure 22.173 Frank H. Knowlton. (Courtesy H. N. Andrews.)

and Australia, as well as on several islands. Caesalpinia claibornensis is an Eocene fruit up to 8 cm long with a vascularized wing along the placental suture (Herendeen and Dilcher, 1991). Similar fruits have been recorded from other Eocene and Miocene localities (Knowlton, 1926) (FIG. 22.173). Although Caesalpinia subgenus Mezoneuron was present in North America during the Eocene and Miocene, the taxon does not occur in either North or South America today. Several leaves assigned to the mimosoids and papilionoids, however, have been reported from the upper Miocene of Argentina (Anzótegui et al., 2007). A highly diverse flora of legumes from an Eocene and an Oligocene site in Mexico contains 10 genera of the Mimosoideae, 3 Caesalpinioideae, and 2 Papilionoideae. Genera include Stryphnodendron (FIG. 22.174), Inga (FIGS. 22.175, 22.176), and Acacia (Calvillo-Canadell and Cevallos-Ferriz, 2005), among others. The occurrence of this flora in Mexico underscores that legumes were an important element in low latitudes of North America during the Cenozoic, and that at least some forms are evolutionarily linked to the development of the boreotropical flora. Mimosoid flowers are perfect and attached to a spicate inflorescence axis (Crepet and Dilcher, 1977). Specimens of Eomimosoidea plumosa range from the Eocene into

Figure 22.174 Stryphnodendron emarginatum (Oligocene). Bar  5 mm. (Courtesy S. R. S. Cevallos-Ferriz.)

the Oligocene (Daghlian et al., 1980). The calyx is fourlobed, and the androecium consists of about eight exerted stamens with bilocular anthers. The tricolporate pollen is shed in permanent tetrads and the gynoecium is a single carpel. Eomimosoidea shares some features with flowers of the extant genera Dinizia and Fillaeopsis, subfamily Mimosoideae. Other zygomorphic mimosoid flowers believed to belong to the subfamily Papilionoideae suggest bee pollination (Crepet and Taylor, 1985, 1986). Barnebyanthus is a pedicellate flower 1.5 mm long that represents the earliest evidence of papilionoid legumes (Crepet and Herendeen, 1992) (FIG. 22.170). This Eocene flower shares several features with the extant genus Sophora. Dichrostachoxylon zirkelii, permineralized wood assignable to the Mimosoideae, has been reported from Miocene rocks at Küçük Çekmece in Turkey (Selmeier, 1990). Another specimen of a permineralized Dichrostachys-type wood comes from the middle Miocene of Bavaria, southern Germany (Selmeier, 1986). Crudia (subfamily Caesalpinioideae) is known from the middle Eocene of North America based on leaflets and fruits (Herendeen and Dilcher, 1990b). Leaflets are asymmetrical and ovate, each 2.7 cm long. Secondary venation is brochidodromous in C. brevifolia. Fruits, each containing two

chapter 22

flowering plants

22.177 Ostrya oregoniana fruit Bar  5 mm. (From Meyer and Manchester, 1997.)

Figure

953

(Oligocene).

Fagales

This order includes eight families, including the Fagaceae, Betulaceae, Nothofagaceae, and Balanopaceae, that are primarily amentiferous with simple, alternate leaves. Figure 22.175 Inga poblana leaf showing two terminal sessile leaflets (Oligocene). Bar  1 cm. (Courtesy S. R. S. Cevallos-Ferriz.)

Figure 22.176 Inga popensis (Courtesy S. R. S. Cevallos-Ferriz.)

(Eocene).

Bar  5 mm.

ovules, measure up to 11.5 cm long and are placed in C. grahamiana. Pollen of Crudia has been reported from the Eocene of Panama (Graham, 1989), and permineralized Cenozoic (Miocene or Pliocene) Crudia woods have been described from Toluviejo–Corozal in Colombia (Pons, 1980) and Río Paranaiba in Brazil (Selmeier, 2004) and the United States (Page, 1993). Additional references to the fossil record of fabaceous woods can be found in Müller-Stoll and Mädel (1967).

BETULACEAE Members of the Betulaceae are well represented in the Cenozoic of the Northern Hemisphere (Crane and Stockey, 1987) (FIG. 22.177). Paracarpinus is an elliptic leaf with a finely toothed margin that is similar to leaves of the extant genus Carpinus. The areoles of P. chaneyi are regularly arranged and four to five sided (Manchester and Crane, 1987). Winged fruits (nutlets) up to 1.3 cm long are assigned to Asterocarpinus (FIG. 22.178). Leaves of Carpinus and samaras (FIG. 22.179) have also been reported from the Eocene of North America. Infructescences (FIG. 22.180), leaves, and wood of Alnus has been described from the Eocene of North America. It is diffuse porous with scalariform perforation plates and homocellular, uniseriate rays (Wheeler et al., 1977). Alnoxylon, Betuloxylon, and Eucarpinoxylon represent silicified betulaceous woods from the Cenozoic (Lakowitz, 1890; Müller-Stoll and Mädel, 1959). Palaeocarpinus (FIGS. 22.181, 22.182) is a relatively common genus that occurs from the Paleocene to late Eocene and combines features seen in extant Carpinus (FIG. 22.183) and Corylus (FIG. 22.184). Specimens of P. dakotensis

954

paleobotany: the biology and evolution of fossil plants

Figure 22.178 Asterocarpinus perplexans nutlet (Oligocene). Bar  2 cm. (From Manchester and Crane, 1987.) 22.180 Alnus sp. infructescences Bar  1 cm. (Courtesy National Museum, Prague.)

Figure

(Eocene).

Figure 22.179 Winged samaras of Carpinus perryae (Eocene). Bar  3 mm. (Courtesy K. B. Pigg.)

infructescences from the late Paleocene of North Dakota are associated with staminate inflorescences (catkins), pollen, and leaves (Manchester et al., 2004). The permineralized infructescences are 11 cm long and contain helically arranged cymules, each containing paired fruits. Morphologically the nuts (FIG. 22.185) are smaller than those in P. joffrensis (F. Sun and Stockey, 1992) and P. orientalis (Manchester and Guo, 1996). Elongate Pliocene infructescences with helically arranged involucres, some with bifurcate tips, and elliptical nuts are placed in the genus Cranea (Manchester and Chen, 1998). These Pliocene

Figure 22.181 Paracarpinus chaneyi (Oligocene). Bar  2 cm.

(Courtesy S. R. Manchester.)

chapter 22

flowering plants

Figure 22.183 Carpinus (Courtesy B. J. Axsmith.)

bract

(Pliocene).

955

Bar  2 cm.

Figure 22.182 Infructescence of Palaeocarpinus dakotensis (Paleocene). Bar  2 cm. (From Manchester et al., 2004.)

deposits from Wyoming also include staminate inflorescences, pollen, and leaves that may represent parts of the same plant. The fossils conform most closely to the extant genus Ostryopsis. CASUARINACEAE This family, with 4 extant genera and 90 species, consists of trees with scalelike leaves and is restricted today to Australia, southeast Asia, and some islands in the southwest Pacific Ocean. The flowers are borne in catkins, with staminate flowers consisting of one stamen and subtending scales,

Figure 22.184 Corylus johnsonii infructescence on long peduncle (Eocene). Bar  5 mm. (Courtesy K. B. Pigg.)

whereas pistillate flowers have two carpels, with the entire structure cone-like in organization and woody at maturity. Pollen is triporate. Dilcher et al. (1990) have provided a detailed analysis of vegetative and reproductive features of

956

paleobotany: the biology and evolution of fossil plants

Figure 22.185 Section of Palaeocarpinus dakotensis nut

showing sclerenchymatous endocarp (Paleocene). Bar  1 mm. (Courtesy S. R. Manchester.)

the extant genera in the family as a basis for interpreting the fossil members more accurately. Casuarina is known from the Eocene of Australia and includes both vegetative and reproductive specimens (Christophel, 1980). The fossils show the greatest similarity to species of Gymnostoma, believed to be the most primitive member of the family. Features useful in defining the genus include stomata in open furrows, male flowers in a compound arrangement and subtended by two bracts, and articles (vegetative branchlets) bearing small leaves that are sometimes termed teeth (Scriven and Hill, 1995). Fossil branches bearing whorls of four leaves have been reported from the upper Paleocene of Australia as G. antiquum (Scriven and Hill, 1995). Small staminate and pistillate flowers together with vegetative remains from the lower Oligocene of Tasmania and Oligocene of Australia were described as G. tasmanianum (Guerin and Hill, 2003, 2006). There are several reports of casuarinaceous pollen from the Cenozoic (Macphail et al., 1994). Permineralized cladodes have been reported from the Oligocene of Australia that compare favorably with those of Casuarina (Guerin, 2004; Arena, 2008). FAGACEAE Evidence from pollen suggests that the Fagaceae originated prior to the Santonian (mid-Late Cretaceous) (Wolfe, 1973). The family includes two subfamilies, the Castaneoideae and the Fagoideae. Castaneoid inflorescences have been described from the middle Eocene of Tennessee (Crepet and Daghlian, 1980). The staminate catkins of Castaneoidea are 9 cm long, with helically arranged dichasia (FIG. 22.186), each containing three florets. Each floret has up to 10 stamens and the anthers contain tricolporate pollen

Figure 22.186 Inflorescence axis with several dichasia of Castaneoidea puryearensis (Eocene). Bar  1.25 mm. (From Crepet and Daghlian, 1980.)

(FIG. 22.187). The family is believed to have undergone a major radiation during the Paleocene (Mindell et al., 2007). Pollen of a unique morphological type with protruding apertures has been described from the Late Cretaceous onward as the Normapolles group (FIG. 22.188), an extinct lineage of possible wind-pollinated taxa (Schönenberger et al., 2001a). These grains are triporate with the germinal openings located on the equator of the grain. In many forms the aperture is complex and protruding (Batten and Christopher, 1981) and the grains are triangular. Normapolles grains are first encountered in the Cenomanian, diversifying during the remainder of the Cretaceous, and becoming extinct by the Oligocene. Orders with which the Normapolles have been related include the Juglandales (Wolfe, 1973), Casuarinales, Haloragales, Myricales, Myrtales, Sapindales, Santalales, and Urticales (Batten and Morrison, 1987; Batten, 1989). As a result of a number of important discoveries of floral structures with Normapolles grains in situ (e.g., Friis et al., 2006b), the Normapolles complex is today regarded as a highly diverse group of wind-pollinated fagalean plants that occupied a variety of ecological settings during the Cretaceous. Charcoalified and lignified pistillate and staminate flowers from the Upper Cretaceous of Portugal are borne in

chapter 22

flowering plants

957

Figure 22.188 In situ pollen grain from Endressianthus miraensis staminate flower (Cretaceous). Bar  6 μm. (Courtesy E. M. Friis.)

22.187 Castaneoidea puryearensis pollen grain (Eocene). Bar  4 μm. (From Crepet and Daghlian, 1980.)

Figure

compound inflorescences and assigned to Endressianthus (FIG. 22.189). The genus includes unisexual flowers in dichasia subtended by a large bract and characterized by unicellular trichomes covering the axes and inflorescence parts (Friis et al., 2003b). One principal difference between the fossil flowers and modern members of the family is the porate pollen type Interporopollenites found in Endressianthus. Campanian staminate flowers, fruits, and cupules from central Georgia (USA) are included in Protofagacea allonensis (Herendeen et al., 1995). The dichasia include up to seven staminate flowers subtended by several bracts; each flower is constructed of imbricate tepals and twelve stamens in two cycles of six; the stamens contain tricolporate pollen. The combination of characters that occur in Protofagacea is not known in extant taxa but rather can be found today in both the Fagaceae and the Nothofagaceae. In this regard the discovery of these fossils not only extends the group back to the Late Cretaceous but also demonstrates that certain features existed in the group, for example pollen type, which are unknown in extant genera. Castanopsis is a tropical–subtropical extant timber tree of Asia. The fossil record of the genus is known principally from fruits up to 3 cm long that can be traced back to the

Figure 22.189 Endressianthus miraensis staminate flower (Cretaceous). Bar  150 μm. (Courtesy E. M. Friis.)

958

paleobotany: the biology and evolution of fossil plants

Figure 22.190 Branch with staminate inflorescence of Fagopsis sp. (Oligocene). Bar  1 cm. (Courtesy S. R. Manchester.)

Oligocene (Mai, 1989) and silicified wood (Selmeier, 1972, 1992a). Some authors use Castanoxylon for the wood of this genus (Navale, 1964). Fagopsis longifolia is an Oligocene plant that consists of alternately arranged, simple leaves (FIG. 22.190) with prominent teeth and craspedodromous venation (Manchester and Crane, 1983). Both pistillate inflorescences and infructescences are preserved in attachment to the vegetative shoots. The genus has a number of features that are unlike any living taxa in the family, but in total indicate affinities within the Fagaceae. Of particular interest is the suggestion that the fruits may have been wind dispersed, whereas those in the modern members of the family are dispersed by animals. Fruits, catkins, and leaves from the Oligocene Catahoula Formation of Huntsville, Texas, demonstrate a variety of characters that can be found in the two subfamilies (Crepet and Nixon, 1989a). Contracuparius is used for deeply lobed, decussate cupules that enclose three to seven winged fruits (FIG. 22.191), whereas Amentoplexipollenites refers to staminate catkins bearing dichasia of small florets. The pollen is tricolporate and transitional between modern pollen species in its ornamentation. Lax catkins assigned to Paleojulacea from the Paleocene–Eocene have tricolporate,

Figure 22.191 Infructescence of Contracuparius huntsvil-

lensis showing lobed cupules (Oligocene). Bar  1 cm. (Courtesy W. L. Crepet.)

tectate-columellate pollen. This taxon has both derived (lax catkin) and ancestral (pollen) features. Castanea (Castaneoideae) includes 12 species that are restricted today to the temperate regions of the Northern Hemisphere (Jones, 1986). Castaneoid fossils (FIG. 22.192) are first reported as pollen from the Santonian, with wood common in the Eocene and Miocene (Selmeier, 1991). Castaneophyllum is a simple, narrow lanceolate leaf up to 28 cm long (Jones and Dilcher, 1988). Leaves exhibit pinnate,

chapter 22

flowering plants

959

Figure 22.193 Quercus sp. cupules (Miocene). Bar  2 mm.

(Courtesy M. S. Zavada.)

Figure 22.192 Castanea sp. (Miocene). Bar  2 cm. (Courtesy

BSPG.)

simple craspedodromous venation. The species in this genus were formerly included in the morphogenus Dryophyllum. Berryophyllum is another Eocene leaf type. Jones and Dilcher (1988) suggested that the foliar characteristics of this leaf type may be the primitive type in the family. Permineralized fruits of Cascadiacarpa spinosa from the Eocene Appian Way locality of Vancouver Island, Canada, are similar to those of Castanea in having an enclosing cupule and branched spines (Mindell et al., 2007). Specimens have a bilocular ovary and spines are randomly distributed on the cupule. Modern species of Quercus range from the tropics to temperate areas and have simple leaves with pinnate venation and cupulate fruits (FIG. 22.193). The unisexual inflorescences have staminate florets with a lobed perianth and four

22.194 Quercoidites pollen Bar  10 μm. (Courtesy M. S. Zavada.)

Figure

grain

(Miocene).

to nine stamens. Pollen is tricolporate (FIGS. 22.194, 22.195) and the fruit is an acorn. The earliest fossils of Quercus are found in the Oligocene (Crepet, 1989). Staminate inflorescences of Q. oligocenensis from the Oligocene Catahoula

960

paleobotany: the biology and evolution of fossil plants

22.195 Quercipollenites pollen grain (Miocene). Bar  10 μm. (Courtesy M. S. Zavada.)

Figure

Formation consist of sessile florets, each with a slightly lobed perianth and six stamens (FIG. 22.196) (Daghlian and Crepet, 1983). Leaves are 3 cm long with two to four lobes on each. Permineralized acorns are known from the Miocene (FIG. 22.197). Paraquercinium is a Cretaceous wood that shares many anatomical features with extant Quercus and Lithocarpus (Wheeler et al., 1987). The wood has rays of two sizes and indistinct growth rings. Pores on the vessels are solitary and perforation plates are simple. Quercinium (FIG. 22.198) is another fagaceous wood reported from the Eocene of Yellowstone National Park (Wheeler et al., 1978). The wood is semi-ring porous. In Quercinium lamarense, growth rings are 2–10 mm wide and apotracheal parenchyma occurs in distinct bands one cell wide (Wheeler et al., 1978). Permineralized Cenozoic wood that is structurally similar, or even identical, to modern Quercus wood has been assigned to the genus Quercoxylon (FIG. 22.199) (Selmeier, 1992b). Based on pollen micromorphology, Crepet (1989) suggested that castaneoid pollen approximates the ancestral condition in the family. Both modern and fossil members imply that the extinction of the Fagaceae in the Cenozoic was more pronounced in the Americas than in the Old World (Crepet and Nixon, 1989b). For example, Z. Kvacˇek and Walther (1992) described vegetative remains from the Cenozoic, and Denk and Meller (2001) have plotted cupule morphology of Fagus from the Cenozoic to the recent, and

Figure 22.196 Portion of a Quercus oligocenensis catkin

(Oligocene). Bar  2 mm. (Courtesy W. L. Crepet.)

Figure 22.197 Section of Quercus hiholensis acorn (Miocene). Bar  3 mm. (Courtesy K. B. Pigg.)

chapter 22

flowering plants

961

22.198 Cross section of Quercinium centenoae (Cretaceous). Bar  450 μm. (Courtesy S. R. S. Cevallos-Ferriz.)

Figure

Figure 22.200 Cleared bracts of Engelhardia roxburghiana

(Extant). Bar  8 mm. (From Dilcher et al., 1976.)

Figure 22.199 Quercoxylon sp., segment of a permineralized

oak stem (Miocene). Bar  11.5 cm. (Courtesy BSPG.)

they suggested that the absence of some features in the fossils was the result of transport prior to fossilization. For detailed information about the fossil history and evolution of the Fagaceae in Europe, see Palamarev and Mai (1998).

JUGLANDACEAE This family, which can be traced from the upper Paleocene, includes 8 extant genera and 60 species that are mainly restricted to the Northern Hemisphere. Juglans extends into South America and Engelhardia (sometimes misspelled Engelhardtia) (FIG. 22.200) to New Guinea, Sumatra, and Java (Manchester, 1987). Caryapollenites (FIG. 22.201) and Momipites are fossil juglandaceous pollen morphotypes (Nichols and Ott, 2006). Dryophyllum is a morphogenus for leaves that were initially believed to be fagaceous, but are now included within the Juglandaceae (FIG. 22.202) (Jones et al., 1988). The narrow elliptical leaflets have a length to width ratio of 2.5:4 (Jones et al., 1988). The margin is sometimes serrate with non-glandular teeth. Venation is pinnate camptodromous to semi-craspedodromous. Carya is used for Paleocene leaflets, catkins (FIG. 22.203), and fruits (FIGS. 22.204, 22.205) that are elliptical and up to 16 cm long. The secondary veins of the leaves are craspedodromous, whereas the tertiary are weakly developed. At least some of these leaves are thought to represent the genus Aesculus (Manchester, 2001a). Two leaflike wings of unequal size that subtend a Carya-like nutlet are given the generic name Casholdia (Crane and Manchester, 1982).

962

paleobotany: the biology and evolution of fossil plants

22.201 Caryapollenites Bar  10 μm. (Courtesy M. S. Zavada.)

Figure

inelegans

(Miocene).

Abundant fossil specimens have been described that are assignable to the tribe Engelhardieae. Characteristic features of this group include the even pinnate organization of the venation and the presence of peltate scales on the abaxial leaf surface. Leaves of the Eocene species Oreoroa claibornensis are linear and 20 cm long (Dilcher and Manchester, 1986). The margin is sometimes serrate, with the teeth confined to the upper third of the leaf. On the lower surface are stalked, circular peltate trichomes. Palaeocarya is the name now used for fossil fruits subtended by a trilobed wing that were earlier placed in Paraoreomunnea. The wings are up to 7 cm long and have a venation that ranges from pinnate to triveined. The nutlet is divided into two or more compartments. In Paleooreomunnea (middle Eocene) the wing of the nut is also trilobed, but here the sinuses are very shallow (Dilcher et al., 1976), and in this genus the nut is large (2 cm). Cyclocarya is the only genus that can be traced back to the late Paleocene based on the unusual attachment of the wing (Manchester and Dilcher, 1982; Manchester, 1987). Permineralized catkins and attached leaves have been found near Almont, North Dakota (USA), in the same upper Paleocene deposits as Cyclocarya (W. Taylor, 2007). In this genus the nutlet is surrounded at the equator by a large circular wing which is at right angles to the long axis of the fruit. In Pterocarya, two large wings are attached to each fruit, whereas in Juglans (walnuts and butternuts) the fruits are larger and lack wings. Fruit dispersal was accomplished both by wind and animals during

Figure 22.202 Dryophyllum subcretaceum Bar  1.5 cm. (From Jones et al., 1988.)

(Paleogene).

the early radiation of the family (Manchester, 1989a). Other morphogenera used for juglandaceous fruits (samaras) from the Cenozoic of Europe and North America include Hooleya and Cruciptera (FIG. 22.205) (Manchester et al., 1994). Amurcarya lobata is the name given to a juglandaceous fruit morphotype from the Paleocene of the Amur Province in

chapter 22

flowering plants

963

Figure 22.204 Carya sp. nut (Miocene). Bar  3 mm. (Courtesy M. S. Zavada.)

Figure 22.203 Carya sp. catkin (Pliocene). Bar  2 cm.

(Courtesy B. J. Axsmith.)

the Russian Far East (Kodrul and Krassilov, 2005). Tiffney (1986b) suggested that fruit and seed dispersal within the Juglandaceae shifted from an abiotic to a biotic syndrome around the Cretaceous–Paleogene boundary. A staminate catkin from the Eocene Claiborne Formation of Tennessee (USA) is also placed in this family (Crepet

Figure 22.205 Winged fruit of Cruciptera sp. (Eocene). Bar  1 cm. (From Manchester, 1991.)

964

paleobotany: the biology and evolution of fossil plants

Figure 22.206 Staminate florets of Eokachyra aeolia showing two bracts (arrows) with floral parts beneath (Eocene). Bar  2 mm. (From Crepet et al., 1980.)

et al., 1980). Eokachyra aeolia specimens consist of an axis 6 cm long, with widely spaced, helically arranged flowers. The flowers are bilaterally symmetrical, each constructed of a three-lobed bract and a three-parted floral envelope (FIG. 22.206). Cuticular preparations indicate that the surfaces of the perianth parts contain large peltate scales (FIG. 22.207). Extending down from the elongate floral receptacle are 10– 15 stamens (FIG. 22.208). The pollen of E. aeolia is triporate and 20 μm in diameter through the equator. Based on a comparison with extant flowers, this taxon shares characters with Engelhardia, Oreomunnea, and Alfaroa. The morphologic arrangement of the parts, coupled with the type of pollen, suggests that, like the related extant forms, the fossil specimens were probably wind pollinated. Polyptera manningii infructescences contain 30 fruits; each nut is pyramidal (FIG. 22.209) and bears a wing with 8–12 lobes (Manchester and Dilcher, 1997). Juglandiphyllites is a pinnately compound leaf of five to seven serrate leaflets found associated with the fruits (FIG. 22.210). Associated staminate catkins have florets with 10–15 anthers with triporate pollen. Juglandaceous flowers from the Upper Cretaceous (Senonian) of southern Sweden are placed in the genus Manningia (Friis, 1983). The flowers are small (2 mm long) and perfect, with five tepals and five stamens. The stigma is trilobed and the gynoecium unilocular with a single seed. In Caryanthus (FIG. 22.211), another flower from the same site, the perianth is epigynous and the number of stamens ranges from six to eight. Pollen extracted from C. knoblochii is referred to the Normapolles genus Plicapollis (FIG. 22.212). Another flower included in the wind-pollinated Normapolles complex is Dahlgrenianthus

Figure 22.207 Peltate scale of Eokachyra aeolia (Eocene).

Bar  50 μm. (From Crepet et al., 1975.)

Figure 22.208 Suggested reconstruction of several staminate

flowers of Eokachyra aeolia showing three-lobed bract and floral parts. (From Crepet et al., 1975.)

chapter 22

22.209 Polyptera manningii fruit Bar  6 mm. (From Manchester and Dilcher, 1997.)

Figure

flowering plants

965

(Paleocene).

Figure 22.211 Caryanthus knoblochii flower (Cretaceous).

Bar  300 μm. (Courtesy E. M. Friis.)

Figure 22.212 Pollen grain from Caryanthus knoblochii flower

(Cretaceous). Bar  6 μm. (Courtesy E. M. Friis.)

Figure 22.210 Juglandiphylloides glabra branch with leaves

(Paleocene). Bar  8 cm. (From Manchester and Dilcher, 1997.)

(FIG. 22.213) (Friis et al., 2006b). This Late Cretaceous bisexual flower includes a perianth of five tepals that are fused to form a floral cup. Pollen is triaperturate and triangular in outline (FIG. 22.214).

966

paleobotany: the biology and evolution of fossil plants

and differs from the others in the presence of diagonal pores on the tracheids (Wheeler et al., 1978). These woods are typically ring porous with thick-walled vessels and fibers. Engelhardioxylon is a morphogenus of middle Eocene woods from the Clarno nut beds of Oregon that shows the greatest affinities with the tribe Engelhardieae (Manchester, 1983). It ranges from diffuse porous to semi-diffuse porous and contains vessel elements with scalariform perforation plates. MYRICACEAE This family is most closely related to the Juglandales and includes plants with simple leaves and compound, unilocular ovaries. The pollen is similar to Late Cretaceous Normapolles group pollen. Pollen similar to Myrica has been reported from the Upper Cretaceous (Santonian) of North America (Doyle and Robbins, 1977). Comptonia columbiana is a simple leaf ~ 12 cm long from the middle Eocene of northeastern Washington (Wolfe and Wehr, 1987). The margin is doubly serrate, a feature that appears to be unique in the family. Figure 22.213 Floral structure of Dahlgrenianthus sueci-

cus showing portion of floral cup (Cretaceous). Bar  150 μm. (Courtesy E. M. Friis.)

Figure 22.214 Detail of stigma of Dahlgrenianthus suecicus showing adhering pollen grains (Cretaceous). Bar  15 μm. (Courtesy E. M. Friis.)

Caryojuglandoxylon, Eucaryoxylon, and Pterocaryoxylon are wood morphogenera that resemble Carya and Juglans (Müller-Stoll and Mädel, 1960; Selmeier, 1995). The middle Eocene species P. knowltonii is believed to be the oldest

NOTHOFAGACEAE Some authorities include Nothofagus, the Southern Hemisphere beech, within the Fagaceae, whereas others have placed the genus in a separate family based on pollen features (Kuprianova, 1962), or on the basis of phylogenetic analyses of numerous characters (Crepet and Nixon, 1989b). R. Hill and Read (1991) suggested that the combination of vegetative, cupule (FIG. 22.215) and pollen characters in both fossil and living taxa provide a more accurate grouping of modern species, and one that may provide valuable insights into the evolutionary history of the family. The modern genus includes 34 species that inhabit South America, Australia, New Zealand, New Caledonia, and New Guinea. Fossil leaves of Nothofagus (FIG. 22.216) often have composite teeth and first appear in the Eocene (FIG. 22.217) (Romero, 1986a). The pinnate venation is craspedodromous in N. elongata (Romero and Dibbern, 1985). In N. muelleri the symmetrical leaves are up to 7 cm long and have distinctly serrate margins (FIG. 22.218) (R. Hill, 1988). Two types of trichomes are present on the lower surface of these Eocene leaves. Specimens of Nothofagus plicata from the Eocene of Australia have plicate vernation and a single set of teeth per secondary vein (Scriven et al., 1995). Wood of Nothofagoxylon is first found in the Upper Cretaceous (Torres and Rallo, 1981), and also occurs in the Eocene of Antarctica (Poole et al., 2001). Some Cenozoic woods have been placed in Nothofagus (Ohsawa and Nishida, 1990). These woods are characterized by scalariform

chapter 22

Figure 22.215 Cupule of Nothofagus bulbosa (Oligocene).

Bar  1 mm. (Courtesy R. S. Hill.)

perforation plates, helical thickenings of the vessels, and circular, alternate intervessel pitting. Pollen of Nothofagus is referred to the genus Nothofagites and first appears in the Santonian (Dettmann and Playford, 1969). The grains are small (25 μm) and stephanocolpate; species from Patagonia are larger and ornamented by spinules (Zamaloa and Varreda, 1992). Nothofagites grains are sometimes placed in separate groups based on aperture morphology (Dettmann et al., 1990). Available fossil evidence suggests that Nothofagus and related families evolved in west Gondwana and that by the Santonian the genus was well established in the Southern Hemisphere (Romero, 1986a). R. Hill (1991) has described several new species of Nothofagus from West Antarctica and Tasmania based on foliage and cupules. He suggested that the climatic shift during the Cenozoic (temperature decline and more seasonal rainfall) resulted in relatively minor changes within the subgenera of Nothofagus. The size, shape, and margin of the fossil leaves of the genus

flowering plants

967

Figure 22.216 Nothofagus plicata leaf (Eocene). Bar  1 cm. (Courtesy S. McLoughlin.)

have been used to suggest information about the habitat and biogeography of various species through time (R. Hill, 1991). Jones (1986), in a review of the vegetative features of the Fagaceae, suggested that the Nothofagaceae and Betulaceae are closely related, perhaps having evolved from the same ancestor. Malpighiales

CLUSIACEAE Minute fossil flowers from the Turonian (Upper Cretaceous) of North America suggest that the family was morphologically diversified by that time (Crepet and Nixon, 1998b). Paleoclusia chevalieri is pentamerous, with some type of resinous material in many flower parts. Pollen on the stigmas is prolate to tricolpate and reticulate (FIG. 22.219). There are some characters in the fossils that may be interpreted as evidence of bee pollination, a condition that has been documented in extant members of the family (Crepet and Nixon, 1998b).

968

paleobotany: the biology and evolution of fossil plants

Figure 22.218 Three leaves of Nothofagus muelleri (Eocene).

Bar  2 cm. (From R. Hill, 1988.)

Figure 22.217 Nothofagus lobata (Oligocene). Bar  1 cm. (Courtesy R. S. Hill.)

EUPHORBIACEAE This is a very large and diverse group of flowering plants that contains nearly 8000 extant species. The pollen record suggests that the group is as old as the Paleocene (Muller, 1984). Hippomaneoidea is an Eocene inflorescence that bears cymules, each of which is composed of three staminate florets (FIG. 22.220) (Crepet and Daghlian, 1982). The pollen is tricolporate, tectate-columellate, and ornamented by a reticulum (FIG. 22.221). Crepetocarpon is a middle Eocene capsulelike fruit 4 cm in diameter (FIG. 22.222) (Dilcher and Manchester, 1988). At the apex is a short protrusion; at the base, a peduncle

Figure 22.219 Pollen grain of Paleoclusia chevalieri showing reticulate sculpture (Cretaceous). Bar  5 μm. (From Crepet and Nixon, 1998b.)

scar. Placentation is axile with radially arranged locules. The Eocene occurrence of Crepetocarpon is interesting because these fossils show the greatest similarities with fruits in the Hippomaneae, considered to be one of the highly evolved tribes in this family.

chapter 22

flowering plants

969

Figure 22.221 Ultrathin section of pollen grain of Hippomaneoidea warmanensis (Eocene). Bar  3 μm. (From Crepet and Daghlian, 1982.)

Figure 22.222 Fruits of Crepetocarpon pekinsii (Eocene). Bar  1 cm. (Courtesy D. Dilcher.) Figure 22.220 Inflorescence of Hippomaneoidea warmanensis (Eocene). Bar  4 mm. (From Crepet and Daghlian, 1982.)

Paraphyllanthoxylon is a silicified Cretaceous–Paleogene wood that shares a number of characters with the extant euphorbiaceous genus Bridelia (Thayn and Tidwell, 1984). The fossil wood is diffuse porous with simple to scalariform perforation plates. Rays are heterocellular and may range from uni- to multiseriate (FIGS. 22.57, 22.58). Paraphyllanthoxylon abbottii from the Big Bend National Park in Texas was the first Paleocene dicotyledonous wood described from North America (Wheeler, 1991). The species shares a number of features

with several dicot families, but is believed to be most similar to genera included within the Burseraceae. Specimens of Paraphyllanthoxylon from the mid-Cretaceous Potomac Group have been suggested as belonging to the Lauraceae (Herendeen, 1991). Other species have been suggested as belonging to the Elaeocarpaceae, Flacourtiaceae, and Anacardiaceae (Herendeen, 1991). Euphorbioxylon ortenburgense is the name given to euphorbiaceous wood discovered in Cenozoic Molasse sediments of southern Germany (Selmeier, 1998), and E. bussonii has been reported from the Upper Cretaceous of Africa (Koeniguer, 1967). Another fossil wood referred to the

970

paleobotany: the biology and evolution of fossil plants

Euphorbiaceae, Bischofia javanoxyla, comes from the lower Miocene of northern Taiwan (C. Li et al., 2003). SALICACEAE There are two living genera, Salix and Populus (FIG. 22.223), in this family. The wood is characterized by vessels with simple perforations and uniseriate rays. Populus wilmattae (FIG. 22.224) is a species known from the Parachute Creek Member of the Green River Formation (middle Eocene) (Manchester et al., 1986). The ovate leaves are up to 10 cm in length and have a crenate margin with regularly spaced teeth, each associated with a gland. The pattern of venation is not distinctly pinnate, as it is in extant Populus (Ramírez and Cevallos-Ferriz, 2000b). The fruiting axis is a raceme with helically arranged, pedicellate fruits (FIG. 22.224); capsules are elliptical. Another fossil poplar from the Parachute Creek flora has been named Populus tidwellii (Manchester et al., 2006). This species differs from extant Populus in that the inflorescences are terminal, whereas they are axillary and borne on the previous season’s wood in modern taxa. Although the origins of the Salicaceae are difficult to trace, modern foliage and fruit characters appear as early as the middle Eocene (Manchester et al., 1986). These are trees, shrubs, and herbs with typically alternate, simple or compound, stipulate leaves. The flowers are usually bisexual and the receptacle hollow. Sepals and petals are pentamerous. Wood assignable to Salix is known from the upper Miocene of Colorado (Wheeler and Matten, 1977). Pseudosalix handleyi is an extinct species in the Salicaceae from the Eocene Green River Formation of Utah and Colorado that is based on twigs with attached flowers, fruits, and foliage (Boucher et al., 2003). MALPIGHIACEAE In this family are tropical climbers, herbs, shrubs, and trees. Many species possess two-armed multicellular, stellate hairs, called Malpighian hairs. Many have conspicuous flowers. The record of the family dates from the Eocene and includes both megafossil (D. Taylor and Crepet, 1987) and pollen evidence (Martin, 2002). The presence of pairs of glands on the sepals of fossil flowers suggests that entomophily was in existence in this group by the Eocene (D. Taylor and Crepet, 1987). Oxalidales

CUNONIACEAE This family includes evergreen trees and shrubs with leathery leaves from the Southern Hemisphere. Flowers almost

Figure 22.223 Populus dentiacuminata Bar  4 mm. (Courtesy S. R. S. Cevallos-Ferriz.)

(Oligocene).

always have numerous stamens and the fruit is a woody capsule. Pollen assigned to Concolpites (Romero, 1986a; Macphail et al., 1994), wood (Poole et al., 2000, 2001), and leaves, flowers, and fruits (Barnes and Jordan, 2000) have been reported from the Cretaceous and Cenozoic. Small bisexual, charcoalified flowers from the Upper Cretaceous

chapter 22

flowering plants

971

(Delevoryas, 1964c). In transverse section each drupe is hexagonal with the individual fruits fused together; between the fruits are what are interpreted as secretory ducts. Silicified wood similar to that seen in the extant mulberry tree (Morus) has been described from the Miocene of southern Germany and referred to the morphogenus Moroxylon (Selmeier, 1993).

Figure 22.224 Fruiting raceme (arrow) attached to branch

bearing leaves of Populus wilmattae (Eocene). Bar  6 cm. (From Manchester et al., 1986.)

of Sweden are tetramerous with valvate sepals and eight stamens in two rows (Schönenberger et al., 2001b). Specimens of Platydiscus have a broad nectary ring and numerous anatropous ovules. Pollen is small and perhaps tricolpate. The discovery of these fossil rosids, like that of several other families, indicates that a number of modern lineages that are now found principally in the Southern Hemisphere were far more cosmopolitan early in their fossil history. ELAEOCARPACEAE In this family are trees and shrubs that consist of small populations ranging from the Pacific area to Australia. Leaves are simple and flowers are regular with three to five valvate sepals and usually four to five petals. Stamens are free and pollen is released through a pore. Leaves of Sloanea from the Oligocene of Italy are simple with irregularly spaced, delicate teeth and craspedodromous venation (Hably et al., 2007). The genus is interpreted as a proxy record for a tropical–subtropical climate during the Oligocene in certain parts of Europe. Rosales

MORACEAE Although a few species inhabit temperate regions, for example Morus, this family occurs in tropical and subtropical regions in the form of trees or shrubs. The leaves are simple and flowers regular and unisexual. Fruits are in the form of an achene or drupe and often multiple. Such plants as breadfruit and mulberry are members of this family. Arthmiocarpus hesperus is a Late Cretaceous fruiting axis that bears helically arranged drupes, each 1.5 cm long

RHAMNACEAE The earliest fossil record of this family is from the Upper Cretaceous (Calvillo-Canadell and Cevallos-Ferriz, 2007) with the earliest pollen record in the Oligocene and leaves in the Eocene. Berhamniphyllum is a simple narrow leaf with eucamptodromous venation from the Eocene of North America (Jones and Dilcher, 1980). Other leaf impressions similar to extant Pomanderris are reported from the Miocene of New Zealand (Campbell, 2002) and the Oligocene of Mexico (CalvilloCanadell and Cevallos-Ferriz, 2007). Coahuilanthus belinade has fused perianth parts that form a floral cup (Calvillo-Canadell and Cevallos-Ferriz, 2007). This Late Cretaceous flower known from impression remains has clawed petals and shares features with Rhamnus. Nahinda is a bisexual flower from the Oligocene with opposite stamens and an ovary that develops into a drupe (FIGS. 22.225, 22.226) (Calvillo-Canadell and Cevallos-Ferriz, 2007). Impressions of what are interpreted as rhamnaceous fruits have also been reported from the Middle Jurassic of north China (Pan, 1990). Paliurus is a cup-shaped fruit 4 mm long, whereas Zizyphus liaoxijujuba is fusiform and up to 10 mm long. If, in fact, these fossils are the remains of angiospermous fruits, then it goes without saying that their Jurassic age greatly predates demonstrable megafossil evidence of the flowering plants. ROSACEAE These are trees, shrubs, and herbs that typically have alternate, simple, or compound stipulate leaves. Flowers are usually bisexual and the receptacle hollow. Sepals and petals are pentamerous. The family consists of 100 genera and 3400 species with a worldwide distribution. Molecular phylogenetic studies suggest that the family is monophyletic (Potter et al., 2002). Ancestral states are interpreted as including shrubs with alternately arranged, simple stipulate leaves, stamens and pistils one to five, and a single ovule per locule (Potter et al., 2007). DeVore and Pigg (2007) provided an excellent summary of the fossil members of the family with particular emphasis on the Eocene of northwestern North America. Paleorosa is a permineralized, bisexual, actinomorphic flower from the middle Eocene (FIG. 22.227) (Basinger,

972

paleobotany: the biology and evolution of fossil plants

Figure 22.225 Nahinda axamilpensis flower with central

carpel and two anthers. (Oligocene). Bar  1 mm. (Courtesy L. Calvillo-Canadell and S. R. S. Cevallos-Ferriz.)

1976). Sepals and petals are in fives and borne in an alternate manner. The gynoecium consists of five free, pubescent carpels, each with two ovules. Stamens are numerous (13–19), and pollen morphology is not known. Another flower with rosaceous features has been described from the mid-Cretaceous Dakota Formation (Basinger and Dilcher, 1984). Flower parts are pentamerous, and sepals and petals are free. The interesting combination of primitive and derived features in this flower suggests affinities with members of the Rosales (Saxifragaceae) and Rhamnales. Süss et al. (1999) reported on a permineralized rosaceous wood, Rosoxylon hassiacum, from the Miocene of Lauterbach in Hesse, Germany. Foliage described as Cercocarpus (FIG. 22.228) has been reported from the Oligocene of Mexico (Velasco de León and Cevallos-Ferriz, 2000) and Stockeya creedensis (FIG. 22.229) is interpreted as an important component of a mesic woodland community during the late Oligocene (Wolfe and Schorn, 1989). The family has also been reported from the middle Eocene of British Columbia (Cevallos-Ferriz and Stockey, 1990c). Twigs of Prunus allenbyensis have a number of features that suggest the fossil is a member of the Rosaceae, including a heterocellular pith with perimedullary zone, wood that

Figure 22.226 Nahinda axamilpensis fruit with persistent perianth (Oligocene). Bar  3 mm. (Courtesy L. Calvillo-Canadell and S. R. S. Cevallos-Ferriz.)

Figure 22.227 Longitudinal section through gynoecium of Paleorosa similkameenensis (Eocene). Bar  400 μm. (Courtesy J. F. Basinger.)

chapter 22

22.228 Cercocarpus mixteca Bar  3 mm. (Courtesy S. R. S. Cevallos-Ferriz.)

Figure

flowering plants

973

(Oligocene).

is semi-ring porous (FIG. 22.230), fibers with circular-bordered pits, and scarce paratracheal parenchyma. The anatomy of the fossils, together with sedimentologic and taphonomic evidence, suggests that the plants grew in a seasonal climate with wet summers and dry winters. Prunus-like Cenozoic wood has been assigned to the morphogenus Prunium (Süss and Müller-Stoll, 1982). Specimens of Prunium gummosum from the Yellowstone National Park show evidence in the form of medullary spots or pith flecks suggesting interactions with the parasitic, cambium-mining fly Palaeophytobia prunorum (Agromyzidae, Diptera) (Süss, 1980b; Süss and Müller-Stoll, 1980). ULMACEAE One of the families included in the Rosales is the Ulmaceae, which includes 25–45 extant genera and a fossil

Figure 22.229 Stockeya creedensis (Rosaceae) (Oligocene). Bar  5 mm. (Courtesy G. R. Upchurch.)

pollen record that extends back to the Upper Cretaceous (Maastrichtian) (Muller, 1981). Extant species were once divided into two subfamilies, the Ulmoideae and the Celtidoideae, but the latter is now a family, Celtidaceae. The fossil record of the Celtidaceae includes Eoceltis dilcheri, a well-preserved flower from the Eocene of North America displaying stamens and petals (FIG. 22.231) (Zavada and

974

paleobotany: the biology and evolution of fossil plants

Figure 22.231 Two compressed flowers of Eoceltis dilcheri showing stamens (arrows) and petals (Eocene). Bar  4 mm. (Courtesy M. Zavada.)

Figure 22.230 Cross section of Prunus allenbyensis show-

ing pith, semi-ring porous wood, and extraxylary tissues (Eocene). Bar  3 mm. (Courtesy S. R. S. Cevallos-Ferriz.)

Crepet, 1981). Members of the Ulmaceae are trees with caducous stipules, no latex, simple distichous leaves, and two styles (Manchester, 1989b). Burnham (1986) has described a number of leaf morphotypes from the Paleogene of North America using discriminant analysis. The author’s studies indicate that Ulmus first appeared in the middle Eocene. Silicified wood (Selmeier, 1989) and winged fruits assigned to the genus Cedrelospermum (Manchester, 1989c) are common elements in some Cenozoic floras. The presence of one or two wings on the samara has been used to distinguish species. Cedrelospermum nervosum is used for branches with attached leaves (FIG. 22.232) and reproductive organs (FIG. 22.233) from the Eocene Green River Formation of Utah and Colorado (Manchester, 1989c). Leaves are narrow and lanceolate with an acute apex and entire to serrate margin. Flowers are unisexual and borne in fascicles of three to five, with the staminate flowers producing porate pollen. Leaves of the Tremophyllum

Figure 22.232 Twig with attached leaves and winged fruits

of Cedrelospermum nervosum (Eocene). Bar  1 cm. (From Manchester, 1989c.)

type are believed to have been produced by C. nervosum. Pollination and seed dispersal were accomplished via wind. The presence of a large number of Cedrelospermum remains in lake deposits near areas of volcanism suggests the plant was an early colonizer of open habitats (Manchester, 1989c). In his excellent review of the group, Manchester (1989b) suggested that the pattern of evolution of fruit dispersal in the Ulmaceae appears to parallel that in the Juglandaceae and Betulaceae.

chapter 22

flowering plants

Figure 22.234 Ulmus fruit (Oligocene). (Courtesy C. T. Gee and G. Oleschinski.)

975

Bar  2 cm.

Figure 22.233 Cedrelospermum nervosum fruit (Eocene).

Bar  3 mm. (From Manchester, 1989c.)

Leaves and fruits (FIG. 22.234) of Ulmus have also been reported from the Eocene of northwestern North America (Denk and Dillhoff, 2005). In U. okanaganensis the leaves are oblong-elliptical with the margin simple to doubly serrate; flower fascicles occur in the axils of leaves (FIG. 22.235). Fruits are samaras with reduced wings. This discovery is especially interesting because of the information provided about the life history of the plant. In modern Ulmus, inflorescences occur on the previous year’s shoots or in the axils of leaves on current year shoots. In the fossils, flower fascicles are formed in the leaf axils of the current year shoots and are found together with the leaves, a condition

Figure 22.235 Suggested reconstruction of Ulmus okanagan-

ensis showing sucker shoot (A) and fertile branchlet (B) (Eocene). (From Denk and Dillhoff, 2005.)

that is uncommon in modern Ulmus (Denk and Dillhoff, 2005). The authors suggested that the relationship between where and when flowers appear in Ulmus may be related to increased climate seasonality beginning in the Oligocene.

976

paleobotany: the biology and evolution of fossil plants

Eurosids II (Malvids) Brassicales

CAPPARACEAE This is a tropical and subtropical family of trees, shrubs, and lianas with alternate, simple leaves and flowers that are mostly perfect and tetramerous. Silicified Capparaceae wood is rare in the fossil record (see Selmeier, 2005), and includes Forchhammerioxylon scleroticum from the Eocene of Wyoming (Kruse, 1954), Capparidoxylon geinitzii from the Oligocene of Egypt (Schenk, 1883), and Capparidoxylon holleisii from the Miocene of Germany (Selmeier, 2005). Pollen is reported from the Pliocene and flowers from the Turonian of North America (Gandolfo et al., 1998a). The small bisexual flowers of Dressiantha bicarpellata are slightly zygomorphic (FIGS. 22.236, 22.237) and characterized by two whorls of stamens that suggest introrse dehiscence. Pollen is tricolporate. These Late Cretaceous flowers constitute the oldest record of the family. Malvales

TILIACEAE Tilia is a common Northern Hemisphere deciduous tree that is known from the Eocene onward, from pollen as well as leaves (Mai, 1961; Wolfe and Wehr, 1987). The presence of a leaflike bract attached to the inflorescence peduncle is a distinct character of the genus. Three types of bract morphologies (FIG. 22.238) have been reported in the fossil record beginning in the Eocene (Manchester, 1994b), suggesting that sessile bracts with pinnate venation and a peduncle that is adnate to the adaxial surface are apomorphic characters in the genus. Silicified tiliaceous wood assigned to extant Grewia has been described from various Miocene localities, including the Rauscheröd in southern Germany (Selmeier, 1985, 2000a; Gottwald, 1997), the island of Honshu in Japan (Watari, 1952), and the Chindwin basin in Myanmar (Burma) (Gottwald, 1994), as well as from the Pliocene of Vietnam (Vozenin-Serra, 1981). Fossil Grewia wood is usually referred to the morphogenus Grewioxylon, and characterized by heterocellular ray tissue that contains tile cells of the Pterospermum type (see Selmeier, 2000a).

Figure 22.236 Lateral view of Dressiantha bicarpellata flower showing part of the corolla (right) (Cretaceous). Bar  0.5 mm. (Courtesy M. Gandolfo.)

Sapindales

The extant Sapindales includes 9 families and 450 genera that are trees, shrubs, and woody vines. Many members produce noxious secondary compounds and have specialized herbivores that feed on them. Others produce gums and resins; flowers are often imperfect.

Figure 22.237 Lateral view of Dressiantha bicarpellata

showing five staminodes fused in a ring Bar  200 μm. (From Gandolfo et al., 1998b.)

(Cretaceous).

chapter 22

flowering plants

977

Figure 22.238 Bract of Tilia pedunculata (Miocene). Bar  1 cm. (From Manchester, 1994b.)

ANACARDIACEAE Today this family includes tropical and subtropical trees and lianas, including the cashew nut tree (Anacardium). Many members of this family produce chemicals that are irritants to humans, for example Toxicodendron. Diffuseporous woods with heterocellular rays have been described from the Miocene of Chile (Schöning and Bandel, 2004), and Glutoxylon, a fossil wood resembling the modern genus Gluta, has been reported from the Miocene–Pliocene of Bangladesh (Poole and Davies, 2001). Silicified Pistacia wood is known from upper Miocene rocks in southern Germany (Selmeier, 2000b). Compressed fruits 2 cm long have been described from numerous sites including the Eocene of the Messel oil shales (Manchester et al., 2007a), as well as flowers (FIGS. 22.239, 22.240) and leaves (Ramírez and Cevallos-Ferriz, 2002) (FIG. 22.241). Specimens of Anacardium germanicum have the same type

Figure 22.239 Staminate inflorescence of Pistacia sep-

timontana (Oligocene). Bar  3 mm. (Courtesy C. Gee and G. Oleschinski.)

of inflated pedicel that is involved in biotic dispersal of extant cashew nuts. Specimens from Germany suggest the group may have spread via the North American land bridge when the climate was warm (Manchester et al., 2007a).

978

paleobotany: the biology and evolution of fossil plants

Figure 22.240 Staminate inflorescence of Pistacia septimontana (Oligocene). Bar  1 mm. (Courtesy C. T. Gee and G. Oleschinski.)

MELIACEAE Included are trees that are pantropical in distribution today. Leaves are typically pinnate and flowers functionally unisexual with a perianth. Some of the mahogany wood genera are classified in this family. Phylogenetic studies suggest that the Meliaceae is most closely related to the Rutaceae and Simaroubaceae (Savolainen et al., 2000). The fossil record extends from the Eocene to the recent (Meyer, 2003); a fossil flower, Swietenia, has been reported in late Oligocene–early Miocene amber from Mexico (Castandea-Posadas and Cevallos-Ferriz, 2007). The family is believed to have originated in Africa and dispersed across Eurasia, with populations in North America, Europe, and East Asia becoming extinct with the loss of tropical climates in these areas (Muellner et al., 2006). References to fossil seeds, leaves, and wood including their global distribution are listed in Muellner et al. (2006 and references therein). RUTACEAE Fossil fruits with a persistent corolla earlier placed in Porana are now included in Chaneya (FIG. 22.242) (Y. Wang and Manchester, 2000). Additional specimens from the Miocene, including floral parts with resin bodies on the petals, have been used to further circumscribe the genus and more accurately define the systematic position of these fossils (Teodoridis and Kvacˇek, 2005; Manchester and Zastawniak, 2007). Ptelea (FIG. 22.243) is used for samaras with a central fruit and two

Figure Figure

22.241 Cotinus fraterna, leaf of Anacardiaceae

(Eocene). Bar  1 cm. (Courtesy G. R. Upchurch.)

22.242 Calyx

ies of Chaneya S. R. Manchester.)

tenuis

with two enlarged fruit bod(Eocene). Bar  1 cm. (Courtesy

chapter 22

to three wings, each with a coarse venation pattern (Call and Dilcher, 1995). The geologic range of the genus extends from the Miocene to the Quaternary. SAPINDACEAE Included in this family are 150 genera of tropical trees characterized by actinomorphic or zygomorphic flowers. The number of sepals is either four or five, with three to five petals, two to six carpels, and four to ten stamens. Pollen is morphologically variable, and seeds often possess an aril or fleshy sarcotesta; endosperm is lacking. The fossil record of pollen assignable to the Sapindaceae includes several tribes and is summarized by Erwin and Stockey (1990). Staminate flowers of sapindaceous affinity, including several in various stages of development, have been described from the Eocene of British Columbia as Wehrwolfea striata (FIG. 22.244) (Erwin and Stockey, 1990). The flowers are small (1 mm long) and characterized by free perianth parts and an intrastaminal nectary disk. Pollen is prolate and

Figure 22.243 Ptelea enervosa (Miocene). Bar  1 mm.

(Courtesy D. L. Dilcher.)

flowering plants

979

tricolporate. Ornamentation consists of striae, with perforations in the lumina between the striae. Wehrwolfea shares the largest number of features with members of the tribe Dodonaeeae. Cenozoic foliage believed to belong to members of the Sapindaceae has been reported by several authors (MacGinitie, 1953; Wolfe, 1977). Leaves of Aesculus hickeyi from the Paleocene of North America consist of three to five leaflets with finely serrate margins; associated fruits have three locules (Manchester, 2001a). Fruits have been described from the lower Eocene London Clay by Reid and Chandler (1933), Chandler (1961) (Paleoallophylus and Paleoalectryon), and Collinson (1983) (Cupanoides), and from the Paleocene of western North America (Wolfe and Wehr, 1987). Fossil sapindaceous wood of Sapindoxylon is known from the Eocene of southeast England (Wilkinson, 1988; Poole and Wilkinson, 1992) and the Miocene of southern Sumatra (Kräusel, 1922a). Acer consists of 140 extant species that occur mainly in temperate regions and on some tropical mountains. They

Figure 22.244 Diagrammatic reconstruction of Wehrwolfea striata flower showing stamens with filaments of two lengths. (From Erwin and Stockey, 1990.)

980

paleobotany: the biology and evolution of fossil plants

22.245 Acer sp. winged fruit (Oligocene). Bar  2 mm. (Courtesy C. T. Gee and G. Oleschinski.)

Figure

Figure 22.247 Acerites multiformis-type leaf (Cretaceous). Bar  2 cm. (Courtesy D. L. Dilcher.)

Figure 22.246

Michael C. Boulter.

are represented generally by monoecious trees and shrubs with opposite leaves and regular flowers. Sepals and petals number four to five and stamens, four to ten. The fruit is a two-winged schizocarp, with each half being a samara (FIG. 22.245). The fossil record of the genus is extensive and is reviewed in Boulter et al. (1996 and references therein) (FIG. 22.246). Pollen has been reported from the Oligocene (Piel, 1971), and fruits, seeds, leaves, and wood have been found at numerous sites throughout the Cenozoic. The foliage morphotype Acerites has also been reported from the Cretaceous (FIG. 22.247). The most exhaustive treatments of Cenozoic Acer fossils are those of Walther (1972) and Wolfe and Tanai (1987). While the former study provides a detailed analysis of central European fossils, the latter covers the systematics

and evolution of the genus throughout North America, and suggests that many radiations of lineages took place during the Eocene and extended from North America into Eurasia via Beringia. An interesting interaction between the gallinducing arthropod Artacris macrorhynchus (maple beadgall mite) and Acer pyrenaicum has been reported from the upper Miocene of the La Cerdana basin in Spain (Diéguez et al., 1996). Dipteronia, the sister genus of Acer, is also well represented in the fossil record. Today Dipteronia is endemic to eastern Asia, but during the Cenozoic was widespread in western North America. Fossil fruits (FIG. 22.248) assigned to D. brownii are smaller and include tricarpellate forms; modern fruits are bicarpellate (McClain and Manchester, 2001). Landeenia includes flowers, fruits (FIG. 22.249), seeds, and pollen from the Eocene of Wyoming (Manchester and Hermsen, 2000). The small flowers ( 9 mm) are bisexual with a pentamerous calyx; pollen is tricolpate with longitudinal striations on the surface. Seeds possess a small wing. The fruits share some similarities with Nordenskioeldia in the Trochodendraceae. The taxon is tentatively placed in Sapindales, incertae sedis because the combination of characters recovered from the fossils, sometimes even including flowers (FIG. 22.250), do not allow placement in any modern family.

chapter 22

flowering plants

981

Figure 22.248 Incompletely detached mericarps of Dipteronia brownii (Eocene). Bar  5 mm. (From McClain and Manchester, 2001.)

Figure 22.250 Fossil flower with funnelform corolla (Paleocene–Eocene). Bar  5 mm. (Courtesy W. L. Crepet.)

Cornales

Figure 22.249 Cyme of two incomplete fruits of Landeenia aralioides. Arrow indicates protruding style (Eocene). Bar  5 mm. (Courtesy S. R. Manchester.)

Asterids This is a large clade with up to 80,000 species that includes flowers with fused petals, cellular endosperm, iridoid compounds, and unitegmic ovules. Although there are no unequivocal fossil asterids in the Early Cretaceous, molecular phylogenetic dating of extant asterids suggests that the group and the subgroups euasterids, campanulids, and lamiids diversified during the Early Cretaceous (Bremer et al., 2004).

CORNACEAE Members of the dogwood family include 120 species of trees and shrubs that today inhabit both northern and southern temperate regions and tropical mountains. Leaves are typically simple and entire, and the inflorescences are organized into corymbs or umbels. The family includes both unisexual and bisexual flowers, with sepals, petals, and stamens in fours or fives. The ovary is formed of two fused carpels and the fruit is either a drupe or a berry. The fossil record of the family is extensive (FIG. 22.251), but is largely confined to the Cenozoic; one of the rare pre-Cenozoic fossils is Cornoxylon maderitschii, permineralized Cornus-like wood known from the Santonian (Upper Cretaceous) south of Aachen, Germany (Gottwald, 2000), and from Mexico. Another pre-Cenozoic fossil with possible affinities in the Cornaceae is Hironoia fusiformis, a cornalean fruit from the lower Coniacian (Upper Cretaceous) of northeastern Japan (Takahashi et al., 2002). The fruit is described as developed from an epigynous ovary with three to four locules. Each locule bears one seed and has a distinctive dorsal germination valve. This fossil provides the earliest record of this group in the fossil record, and establishes a minimum age for the early divergence of

982

paleobotany: the biology and evolution of fossil plants

Figure 22.251 Cross section of Cornaceae wood (Cretaceous). Bar  700 μm. (Courtesy S. R. S. Cevallos-Ferriz.)

the asterid clade. Eyde (1988) suggested that the dogwoods evolved along two principal lines, which can be distinguished by the type of ovule, presence or absence of bracts below the flower cluster, fruit color, pollen, and, to a lesser degree, fruit shape. Swida is an extant genus that is known from the Miocene of Denmark based on fossil fruit morphology (Friis, 1985a). The fossil is a biloculate, ovate-shaped endocarp with a slightly tapered apex. The outer surface is smooth. Swida discimontana is a fruit from the middle Miocene of Vogelsberg, Germany (Mai and Gregor, 1982). Leaves of Cornus from the Eocene of western North Dakota are ovate and symmetrical, and have acrodromous venation (Hickey, 1977). Although Nyssa has been included in its own family (Nyssaceae) by some authors, Eyde (1988) included the genus within the Cornaceae based on several lines of evidence ranging from chromosome number to nodal anatomy. The fossil record of the genus Nyssa can be traced back to the Cretaceous of North America, the Eocene of Europe, and the Oligocene of Asia; it consists of fruits (FIG. 22.252), wood, and leaves (reviewed in Eyde, 1997). The name Nyssoxylon is used for Cenozoic wood that resembles wood of extant Nyssa (Mädel, 1959; Gottwald, 2000; Meijer, 2000). The genus Nyssa became widely distributed in the Northern Hemisphere during the Cenozoic, then suffered wide extinctions, probably due to climatic changes in the late

Figure 22.252 Section of Nyssa sp. fruit (Cretaceous). Bar  1 mm. (Courtesy R. Serbet.)

Figure 22.253 Leaf impression of Davidia (Paleocene). Bar  2 cm. (Courtesy S. R. Manchester.)

antiqua

Cenozoic and Quaternary. The genus survived to date in refugia in eastern and northeastern Asia and North and Central America (Wen and Stuessy, 1993). Davidia (FIGS. 22.253, 22.254), an extant genus in China, is known from fruit fossils from the Paleocene of North America (Manchester, 2002) and the Pliocene of Japan

chapter 22

flowering plants

983

Figure 22.255 Leaf of Browniea serrata (Paleocene). Bar  1 cm. (Courtesy S. R. Manchester.)

Figure 22.254 Fruit attached laterally to Davidia antiqua

(Paleocene). Bar  1.5 cm. (Courtesy S. R. Manchester.)

(Kokawa, 1965; Tsukagoshi et al., 1997). Fruits are borne singly and have a longitudinally ribbed stone (endocarp). Numerous autapomorphies of the fossils, for example, large number of carpels, one fruit per inflorescence and so on, suggest they were in place by the Paleocene, perhaps signifying

a rapid evolution of the group in the Cretaceous (Manchester, 2002). Amersinia is used for Paleocene infructescences in which each carpel contains a single seed (Manchester et al., 1999). They are similar to living forms in Camptotheca. In the same rocks are leaves assigned to Beringiaphyllum that compare favorably with extant leaves of Davidia, some of which had earlier been included in Viburnum. Both reproductive and vegetative organs (FIG. 22.255) from the Cretaceous to Paleocene have been used to reconstruct Browniea (Manchester and Hickey, 2007). The reproductive parts (FIGS. 22.256, 22.257) of the plant include fruits, infructescences, and flowers. Pollen is small and tricolpate and about half the size of pollen of Camptotheca, an extant member of the Cornales with which it shares a number of characters. The leaves of B. serrata have pinnate venation and evenly spaced teeth. In spite of the extensive Cenozoic record of the Cornaceae, the absence of a well-documented fossil record in the Cretaceous, with the exception of Hironoia fruits, continues to obscure the affinities and patterns of radiation of the dogwoods. Xiang et al. (2006), using a total evidence tree, suggested that Cornus originated in Europe and had multiple migrations to North America and a few colonizations in the Southern Hemisphere in the early Cenozoic.

984

paleobotany: the biology and evolution of fossil plants

CURTISIACEAE Today this family is represented by a single species of evergreen tree found in montane forests on the eastern coast of South Africa. Leaves are opposite and regular, with flowers tetramerous. The ovary consists of four locules, each with a single seed. Tetralocular endocarps have been reported from the early Eocene of the London Clay and from southern England (Manchester et al., 2007b). In Curtisia quadrilocularis the structurally preserved specimens show details of the endocarp wall, including the presence of a central vascular bundle, which is diagnostic of fruits in the Curtisiaceae. The discovery of Curtisia in Europe during the Eocene suggests that perhaps this genus had a Laurasian origin, which would be consistent with the fossil record of other members of the order. HYDRANGEACEAE Small pentamerous, bisexual flowers from the Late Cretaceous (Turonian) are called Tylerianthus crossmanensis (Gandolfo et al., 1998b). Stamens occur in a single whorl, and in situ pollen is small and tricolporate. Each carpel has a recurved style that may be glandular (FIG. 22.258). These coalified flowers were suggested as demonstrating biotic pollination and show a suite of characters that are found in both the Saxifragaceae and the Hydrangeaceae today. Additional Hydrangea fossils, including calyces, have been described from several Cenozoic localities in Europe and North America (Hollick, 1925; Givulescu, 1967; Łan´cuckaSrodoniowa, 1975; Mai, 1985; Z. Kvacˇek, 2002b).

Figure 22.256 Infructescence of Browniea serrata with long peduncle (Paleocene). Bar  1 cm. (Courtesy S. R. Manchester.)

22.258 Lateral view of Tylerianthus crossmanensis showing two carpels with curved styles (Cretaceous). Bar  300 μm. (From Gandolfo et al., 1998a.) Figure

Figure 22.257 Infructescence and dispersed fruits of Browniea

serrata (Paleocene). Bar  1 cm. (Courtesy S. R. Manchester.)

chapter 22

flowering plants

985

Ericales

EBENACEAE About 450 species of woody, tropical plants are included in this family. The flowers are unisexual with stamens in pairs; pollen is tricolporate. Ebenaceous flowers and leaves have been described from the upper Eocene of Australia (Christophel and Basinger, 1982). Leaves of Austrodiospyros demonstrate a great range in size from 4–110 mm long, and a range in the shape of the apex (Basinger and Christophel, 1985). Venation is pinnate and brochidodromous, with randomly distributed actinocytic stomata on the abaxial surface. Only staminate, mummified, actinomorphic flowers of A. cryptostoma have been found. The perianth consists of four sepals and the same number of petals. Pollen is tricolporate and 28–40 μm in diameter at the equator. Seeds and fruits assignable to the Ebenaceae have been described from the Cenozoic of Europe (Vaudois-Miéja, 1982). It is believed that the family originated in the Southern Hemisphere and diversified widely during the late Cenozoic (Basinger and Christophel, 1985). ERICACEAE Many of these plants are evergreen shrubs with leathery leaves and stamens that dehisce by an apical pore. In the extant genus Monotoca, pollen tetrads possess one large functional cell and three small aborted cells. Tetrads with this geometry are known from the Miocene (Martin, 1993). Flowers with fluted, syncarpous ovary and inverted, U-shaped anthers with pseudoterminal awns as well as fruits of Paleoenkianthus sayrevillensis (FIG. 22.259) have been described from the Turonian (Upper Cretaceous) of New Jersey (Nixon and Crepet, 1993). Monadinous pollen (released as a single grain) was present on a stigma, and this, along with flower morphology, suggests ericalean affinities, probably near the extant genus Enkianthus in the basal Ericaceae. Permineralized, diffuse-porous wood with indistinct growth rings has been reported from the Paleocene of Belgium as Agaristoxylon and is compared with the modern small evergreen shrub Agarista (Gerrienne et al., 1999a). THEACEAE Included in this family are nearly 600 species of mostly tropical and subtropical trees and shrubs, as well as a few temperate species. Flowers are typically solitary with numerous stamens and 2–10 fused carpels; the fruit is a capsule or dry drupe. It is within this family that the economically important tea plant is found. The fossil record of the family is widespread in the Northern Hemisphere and dates back to the Late Cretaceous,

Figure 22.259 Lateral view of mature fruit of Paleoenkianthus

sayrevillensis (Cretaceous). Bar  2 mm. (From Nixon and Crepet, 1993.)

where leaves, wood, pollen, fruits, and seeds are known (Grote and Dilcher, 1989). Andrewsiocarpon is a middle Eocene genus that is based on fruit and seed remains (Grote and Dilcher, 1989). The capsules are five parted and subtended by the same number of imbricate sepals; bracteole scars are present on the fruit pedicel. Each of the five locules contains two seeds. Specimens of A. henryense share the largest number of features with fruits of the modern tribe Gordonieae and especially with the monotypic genus Franklinia. Gordoniopsis is another five-valved, dehiscent capsule from the Eocene that contains 6–10 anatropous seeds per locule; each seed is 7 mm long (Grote and Dilcher, 1992). In general these Eocene fruits are similar to those of the extant genus Gordonia, except that Gordoniopsis seeds lack wings. The presence of fruits of Gordoniopsis and Gordonia from stratigraphically equivalent, closely associated Eocene rocks suggests that perhaps these plants demonstrate two different propagule strategies. In Gordonia the winged seeds may have been dispersed by wind, whereas in Gordoniopsis the lack of wings may indicate either a more passive, abiotic system or perhaps a biotic syndrome. Hartia quinqueangularis is used for Cenozoic fruits and seeds that demonstrate the greatest similarity to those within the subfamily Camellioideae (Mai, 1975).

986

paleobotany: the biology and evolution of fossil plants

Leaves of Ternstroemites, a common Cretaceous and Eocene form, are ovate and variable in size. The apex is acuminate and possesses glandular serrations along the margin (Hickey, 1977). Other fossil foliage types assigned to this family include Ternstroemia, Thea, and Gordonia (Huzioka and Takahasi, 1970; Tanai, 1970). Theaceous woods from the Upper Cretaceous have been named Schimoxylon (Kramer, 1974) or Sladenioxylon (Giraud et al., 1992), and woods from the lower Miocene Yanagida Formation in central Japan and the Kungkuan Tuff in northern Taiwan have been described as Camellia japonoxyla (Suzuki and Terada, 1996) and C. kueishanensis (C. Li et al., 2003). Pollen of Pelliceria, which today is a small tree, has been reported from the Cenozoic of the Caribbean region (Graham, 1977).

endocarps up to 2 cm long. Fruits of Palaeophytocrene have been reported from the Eocene of British Columbia (Rankin et al., 2008). Endocarps (FIG. 22.261) are 2 cm long and ornamented on the outer surface by conical tubercles. The presence of Icacinaceae, a family that today is principally subtropical to tropical, provides additional evidence from the fossil record of a warmer climate in North America during the Eocene (Pigg et al., 2008).

Euasterids I (Lamiids) ICACINACEAE The plants in this family have racemose inflorescences of small flowers with fruits in the form of drupes. Leaves are simple with entire margins. The fossil record is summarized in Tanai (1990) and Kvacˇek and Bu˚zek (1995). Several Cenozoic fossils from Europe have been included in the family, as have Paleocene permineralized fruits from North America (FIG. 22.260). Icaciniphyllum is used for leaves with blunt teeth on the margin and semi-craspedodromous venation, whereas Palaeohosiea is used for flattened

22.260 Section of permineralized fruit of the Icacinaceae (Paleocene). Bar  2.5 mm. (Courtesy K. B. Pigg.)

Figure

22.261 Cross section of Palaeophytocrene cf. pseudopersica fruit (Eocene). Bar  1.5 mm. (Courtesy R. A. Stockey.)

Figure

chapter 22

flowering plants

987

Figure 22.262 Eucommia constans (Neogene). Bar  5 mm. (Courtesy D. L. Dilcher.) Figure 22.263 Eucommia eocenica (Eocene). Bar  4 mm.

Garryales

EUCOMMIACEAE The single species Eucommia ulmoides is included in this family that today it is restricted to forests in China. Flowers are unisexual and the fruit is a flat samara. Fossil fruits are common and can be distinguished based on the morphology and venation (FIGS. 22.262, 22.263) (Manchester, 1999). In Europe there is an excellent record beginning in the Oligocene (Mai, 1995). Some of the fossil fruits are about half the size of their modern counterparts (Call and Dilcher, 1997). Gentianales

GENTIANACEAE Most of the members of this family have opposite, entire leaves with gamopetalous flowers. They are cosmopolitan and

(Courtesy D. L. Dilcher.)

include annual and perennial herbs, shrubs, lianes, and trees. Small (22 mm in diameter) sympetalous flowers with a sevenlobed corolla have been reported from the lower Eocene (Crepet and Daghlian, 1981). Pollen in the anthers is triporate with gemmate ornamentation. Dispersed pollen of this type is included in Pistillipollenites. Fossil pollen from the Eocene Gatuncillo Formation in Panama has been compared to that produced by the extant genus Lisianthus (Graham, 1984). RUBIACEAE In this large family, plants are tropical to subtropical in distribution and they have simple, entire leaves with stipules. Simple perfect flowers are typically four- to five-merous,

988

paleobotany: the biology and evolution of fossil plants

and the ovule is generally inferior. Included in this family is the coffee tree, Coffea. Paleorubiaceophyllum is a narrow Eocene leaf up to 8.5 cm long; venation is brochidodromous (Roth and Dilcher, 1979). Wood of Grangeonixylon danguense from the Ypresian (Eocene) of France has tentatively been assigned to the Rubiaceae (Sakala et al., 1999; Sakala, 2000). Pollen of the extant Central and northern South American genus Sabicea has been discovered from lower Miocene rocks of the Culebra Formation in Panama (Graham, 1987). For details on the fossil pollen record of Rubiaceae, refer to Dessein et al. (2005). Lamiales

AVICENNIACEAE This family includes tree mangroves with stilt roots; fruit is a capsule with a beaked apex. Possible fruits have been described from the Miocene of New Zealand (Campbell, 2002). BYBLIDACEAE There are six to seven species in the single genus Byblis and these include herbs to shrubs in Australia and New Guinea. The family is known from at least the Eocene in the form of a mummified seed from South Australia (Conran and Christophel, 2004). On the surface are longitudinal ridges that form a honey-combed pattern. LENTIBULARIACEAE This is a family of carnivorous herbs in which the fossil record is scanty, but does include a few seeds (Mai, 1985; Collinson, 1989), turions (Jung, 1976), and pollen (Tsuji, 1979). OLEACEAE This family includes 600 species distributed in 25 genera that are worldwide in distribution. Fossil leaves are included in Olenites, a genus used for simple petiolate leaves with peltate hairs on both surfaces. Dispersed pollen together with leaf remains from the Miocene of Italy have been used to more accurately characterize the family in other geologic horizons (Sachse, 2001). Cenozoic silicified woods structurally similar to that of the extant genus Fraxinus have been described as Ornoxylon (Süss, 2005). The oldest macrofossil evidence of the family (Eocene) occurs in the form of bilaterally symmetrical samaras with a single seed derived from a superior ovary (Call and Dilcher, 1992). Solanales

SOLANACEAE The plants in this family include numerous species of economic importance such as tomato, peppers, and potato; another plant in this family, tobacco, also has economic

Figure 22.264 Marjorie E. J. Chandler. (Courtesy H. N.

Andrews.)

importance, but at the same time is a major health problem worldwide. Flowers are actinomorphic, pentamerous, and typically bicarpellate with numerous seeds. Berries, one of the fruit types in the family, represent a synapomorphy for a large clade in the family (Knapp, 2002). The fossil record is not extensive (Collinson et al., 1993). Fruits have been reported from the lower Eocene London Clay Flora (Reid and Chandler, 1933) (FIG. 22.264) and seeds from the Miocene of central Europe (Van der Burgh, 1987).

Euasterids II (Campanulids) BRUNIACEAE This is a small family of shrubs confined to the Western Cape region of South Africa. Leaves are coriaceous and the inflorescence may contain 400 bisexual, five-merous flowers (Hall, 1987). A possible fossil member of this family is Actinocalyx bohrii, from the Upper Cretaceous of Sweden. The flowers are perfect with a calyx formed of five free sepals and five stamens containing tricolporate pollen (Friis, 1985b). QUINTINIACEAE This is a small, unplaced family that is monotypic and today restricted to the southwestern Pacific and Australia. Flowers are bisexual with an inferior ovary and five stamens. A flower that may be included in this family from the Late Cretaceous of Scania (southern Sweden) is Silvianthemum (Friis, 1990).

chapter 22

Figure 22.265 Flower bud of Silvianthemum suecicum showing small size of sepals (Cretaceous). Bar  1 mm. (Courtesy E. M. Friis.)

flowering plants

989

Figure 22.267 Aralia wellingtoniana (Cretaceous). Bar  2 cm.

Apiales

ARALIACEAE In this family are trees or shrubs with lobed, compound leaves (FIG. 22.267) and umbellate inflorescences (FIG. 22.268). The fruit is a drupe. Flowers have been described from the upper Eocene–lower Oligocene Florissant flora of Colorado (USA), as well as fruits, for example Paleopanax, from the Eocene Clarno nut beds of Oregon (Manchester, 1994a). Aralia sp. is a five-lobed leaf from the Oligocene (FIG. 22.269); other leaves from the Eocene of North America are included in Dendropanax based on leaf shape, venation, and epidermal anatomy (Dilcher and Dolph, 1970). Aquifoliales

Figure 22.266 Apical view of Silvianthemum suecicum

showing nature of corolla (Cretaceous). Bar  0.5 mm. (Courtesy E. M. Friis.)

These perfect flowers are small and have pentamerous floral parts (FIGS. 22.265, 22.266). Pollen is small (10 μm) and tricolpate. The affinities of S. suecicum are suggested to be with species of the extant genus Quintinia.

AQUIFOLIACEAE This family is regarded as basal to many members of the Asterales. Ilex includes 400 species that are distributed worldwide. Flowers are unisexual, small, and have valvate sepals. The fossil record of the genus Ilex is primarily composed of carpological material and leaf fossils (e.g., Walther and Kvac˘ ek, 2008). Fossil tricolpate pollen grains with poorly defined pores have been included in the genus Ilexpollenites, and have been recorded as early as the Eocene

990

paleobotany: the biology and evolution of fossil plants

Figure 22.269 Aralia sp. leaf (Oligocene). Bar  6 cm.

(Courtesy BSPG.)

Figure 22.268 Araliaceous infructescence of globose heads

(Eocene). Bar  1 cm. (Courtesy S. R. Manchester.)

(Martin, 1977). Ilicoxylon austriacum is a permineralized Ilex wood described from the Ottnangian (early Miocene) of Gallspach in Upper Austria (Selmeier, 1970). Asterales

Cronquist (1988) included 60,000 species in this subclass, with nearly one-third of the species included in the largest family of dicotyledons, the Asteraceae or Compositae (old name). The majority of the plants are herbs with sympetalous flowers; iridoid compounds are common in many species. ASTERACEAE (COMPOSITAE) The aster or sunflower family is one of the most successful families of flowering plants today in terms of both species number and habitat. They possess alternate, opposite,

or whorled leaves, and a well-developed secretory system. Flowers are arranged in heads that may be grouped into larger inflorescences; heads are subtended by involucral bracts. Individual flowers are termed florets and have united petals; on a single head the flowers may be all alike and bisexual, or arranged in a ring with the marginal florets irregular (ray florets) and the central ones regular. Flower parts are in fives with the ovary inferior and bicarpellate; the fruit is an achene or cypsela. Although a few Asteraceae pollen grains have been described in older rocks, there are several records from the Eocene (Scott et al., 2006) and Oligocene (Muller, 1981), as well as the Miocene (FIGS. 22.270, 22.271). One of the oldest fossil pollen grains attributed to the family is the long-spined Echitricolporites from the Eocene of Brazil (Graham, 1996). There are relatively few reports of megafossils of the Asteraceae, with Viquiera cronquistii one of the most frequently cited reports (Becker, 1969) (FIG. 22.272). The species is based on a flattened head attached to a short stalk. The surface of the head is covered by numerous helically arranged bracts, but florets, cuticle, and pollen data are lacking from these late Oligocene specimens. After reexamining the type specimen, Crepet and Stuessy (1978)

chapter 22

flowering plants

991

22.270 Pollen grain of Asteraceae (Miocene). Bar  10 μm. (Courtesy M. S. Zavada.)

Figure

Figure 22.272 Herman F. Becker. (Courtesy D. L. Dilcher.)

and the epidermis may have various patterns produced by the epidermal cells. Although there are reports as early as the Paleocene, Kokawa (1961) suggested that reports of Meyanthes earlier than the Oligocene are problematic. Seeds are especially common in Holocene deposits (Watts, 1971). Dipsacales

Figure 22.271 Pollen grain of Asteraceae Bar  10 μm. (Courtesy M. S. Zavada.)

(Miocene).

concluded that the fossils may just as likely represent any one of a number of reproductive organs of other angiosperms or gymnosperms. Therefore, it is their opinion that V. cronquistii should not be used as evidence of Asteraceae. MENYANTHACEAE The members of this family are clonal aquatic plants with heart-shaped leaves covered by trichomes; they inhabit shallow bogs and river margins. Flowers are heterostylous and dimorphic. Fruits are 1 cm long; seeds are circular

CAPRIFOLIACEAE Modern members include shrubs and a few woody lianes with opposite leaves, sometimes in pairs. The fruit is a capsule. Dipelta is known from samaras and has a single fruit per dispersal unit (Reid and Chandler, 1926). Diplodipelta has been described from several sites spanning the Eocene to Miocene as bilaterally symmetrical, samara-like structures, each with two achene-like fruits (Manchester and Donoghue, 1995). Pollen is known from numerous Cenozoic sites (Böhnke-Gütlein and Weberling, 1981).

Cenozoic floras The distribution and biogeography of angiosperm taxa through time has historically been a focal point of angiosperm paleobotany (Axelrod, 1970). Numerous floras of Cenozoic (Axelrod, 1998a, b) age have been described worldwide—far too many to enumerate here. Rather, we will discuss a few selected examples that demonstrate the wealth of information that can be obtained, not only about the plants in a particular

992

paleobotany: the biology and evolution of fossil plants

Figure 22.273 Aureal T. Cross.

flora and their biogeography, but also about the environment in which they lived, including their use as paleoclimatic indicators (e.g., Traiser et al., 2007; Uhl et al., 2007b, c). One of the most extensively studied paleofloras of the Columbia Plateau region in the northwestern United States is the Succor Creek flora, located near the Idaho-Oregon border in southeastern Oregon (Fields, 1996). This important middle Miocene flora actually represents a number of florules collected within the same geographic area from the Sucker Creek Formation. Taggart and Cross (1980) (FIG. 22.273) listed more than a dozen collecting sites and a combined megafossil flora and palynoflora that includes representatives of almost 50 families in the cryptogams (algae, fungi, lycopods, sphenophytes, and ferns), gymnosperms (conifers, ginkgophytes, and gnetophytes), and angiosperms (50 genera of monocots and dicots). The most commonly occurring taxa are species of Quercus, Cedrela, Acer, Woodwardia, Platanus, Ulmus, and Picea (Graham, 1965; Fields, 1996). Morphologic analysis of the fossil specimens indicates that the most closely related extant taxa occur principally in eastern North America and the deciduous hardwood forests of eastern Asia. Based on the climatic tolerance limits of living Cedrela, that is, the nearest living relative method, NLR (Chapter 1), the climate during the late Miocene is reconstructed with winter temperatures that

did not fall below freezing for any extended period of time (Graham, 1965). Based on the proportion of montane conifer pollen, Taggart and Cross (1980) suggested that the climate was cool but equable. Based on the sedimentation cycle, regional geology, and taxa present at the Succor Creek sites, Graham (1965) has postulated that the flora was deposited in a freshwater lake that existed at an elevation of 2000 feet. An analysis of the composition of the Succor Creek megafossil flora and palynological remains suggests the presence of several identifiable plant communities in the area. Surrounding the lake in a narrow band was herbaceous marsh vegetation that included species of Equisetum, Typha, and Potamogeton. One of the more common elements of the lowland community was Glyptostrobus, an Asiatic genus of plants that grows in the evergreen, broad-leaved forests of China. The upland community of Succor Creek was dominated by species of oak, together with Acer, Alnus, Betula, Fagus, and several other taxa. The palynology from higher elevations shows high concentrations of Abies and Picea pollen. The presence of some grass pollen together with the remains of grazing animals led Graham (1965) (FIG. 22.274) to speculate that open grasslands were also present in the forest community. This brief description of one of the well-documented Columbia Plateau Miocene floral assemblages serves as an example of the interface of biological and geological data that are necessary for a complete analysis of fossil plant communities and the important role such studies play in tracing vegetational history. For further information, the interested reader is encouraged to examine some of the original papers describing the floras of the Columbia Plateau region (Chaney, 1959, 1967; Chaney and Axelrod, 1959; Cranwell, 1964; Graham, 1965; Taggart, 1973; Taggart and Cross, 1980). The Eocene floras of western Kentucky and Tennessee are also some of the better-known Cenozoic floras. The work on these floras began with E. W. Berry in the early 1900s and commenced again in the 1960s during the rebirth of angiosperm paleobotany in North America, led to a large degree by David Dilcher (1963). In addition to exceptionally well-preserved megafossil floras from a number of localities in this region, the stratigraphy has also been well defined based on palynomorphs (Frederiksen, 1988). A sufficient number and variety of specimens have been described from sites in the Claiborne Formation of Tennessee so that an analysis of the community provides data for understanding paleoclimates. These Eocene communities are an excellent example of why caution and careful evaluation of the fossils, however, must precede generalizations about past

chapter 22

flowering plants

993

Figure 22.275 Ficus affinis (Cretaceous). Bar  2.5 cm. (CourFigure 22.274 Alan Graham.

climates based on the plants alone. For example, there are several genera, such as Sabal, Philodendron, Ficus (FIG. 22.275), and Ocotea, that suggest a lowland tropical environment. Associated with these fossils are other floral elements that are common in montane areas, for example, pollen of Podocarpus and Pinus. The juxtaposition of these plants, which today do not coexist, suggests that the climatic tolerances of some of these taxa may have also changed since the Eocene. For these reasons, the use of the NLR method to reconstruct paleoclimate for these floras may not be very accurate. For angiosperm floras, the analysis of leaf margin and size, or foliar physiognomy, provides an estimate of climate (paleotemperature and paleoprecipitation) that is basically independent of taxonomy (Wolfe, 1979; Uhl et al., 2007b) (Chapter 1). For example, the Eocene sites in the Mississippi embayment have been regarded as representative of a tropical rain forest based principally on the identification of certain taxa. An analysis of the floras from this area based on foliar physiognomy suggests, to the contrary, that during the Eocene the area was a seasonally dry to slightly moist, warm temperate to cool subtropical regime (Dilcher, 1973). Wolfe and Upchurch (1987) and Wolfe (1990) (FIG. 22.276) utilized foliar physiognomy and cuticular

tesy G. R. Upchurch.)

Figure 22.276 Jack A. Wolfe. (Courtesy S. R. Manchester.)

994

paleobotany: the biology and evolution of fossil plants

Phase 5 York Canyon

formation

Phase 4 Angiosperm recovery

m 100

0

Raton

Phase 2 Fern spike

cm 10

Phase 1 Lancian

Vermejo formation

K–T

Phase 3 Angiosperm recolonization

0

Figure 22.277 Changes in leaf physiognomy from the latest Cretaceous to the early Paleocene in the Raton Basin. Phase 1: High diversity flora suggestive of warm subhumid conditions. Phase 2: Vegetation dominated by ferns following extinction at the K/P boundary. Phase 3: Angiosperms become dominant, but not diverse; suggest warm wet conditions. Phases 4, 5: Increasing angiosperm diversity. Leaves and cuticles suggestive of warm rain forest environments. In Phase 5 the angiosperm diversity is not as great as in Phase 1. (Modified from Wolfe and Upchurch, 1987.)

analysis of leaf floras from the Raton Basin, western United States, to analyze floral and paleoclimate changes across the Cretaceous–Paleogene boundary. They report a marked temperature and precipitation increase across the boundary. They also plot species diversity across the boundary (FIG. 22.277) and conclude that diversity was highest during the latest Cretaceous (FIG. 22.277, Phase 1). Following the K–P extinction event, the recovery of angiosperms (FIG. 22.277, Phases 3–5) meant that diversity was still low compared to Phase 1. This detailed study also identifies an increase in fern spores, the so-called fern spike, immediately following the extinction event.

Similar studies using leaf physiognomic features have been reported by Uhl et al. (2006) for the middle Miocene of Germany. As these studies underscore, it is a combination of data from different sources that provides the best estimate of paleoclimate (Uhl et al., 2003). This approach is frequently utilized in modern floral analyses and today is a necessary component of many paleobotanical studies. For example, the paleoclimate of high-latitude Cretaceous floras from Alaska has been examined utilizing measures of floral diversity (Spicer and Parrish, 1986), including amount of deciduousness (Wolfe, 1987). These data have been combined with information from foliar physiognomy (Spicer et al., 1987), tree rings (Parrish and

chapter 22

Spicer, 1988), and detailed sedimentological work (Ahlbrandt, 1979) to provide an overall picture of the environment in which these plants were growing. A different approach was applied by Böhme et al. (2007) in their reconstruction of the early and middle Miocene paleoclimate in southern Germany. These authors studied some 1000 silicified wood samples from the North Alpine Foreland Basin and inferred climate estimates based on the Coexistence Approach. This is a refinement of the NLR method, in which suites of taxa are compared in order to determine paleoclimate. Examining the ecological requirements of the NLRs of fossil taxa has also been useful in arranging the major vegetation units of the Eocene within a global context and subsequently correlating these data with independent proxy records of paleoclimate (Utescher and Mosbrugger, 2007). Another measure of Cenozoic floral diversity from a different type of flora is the extensive treatment of fruits and seeds in the London Clay Flora (Reid and Chandler, 1933; Chandler, 1961; Collinson, 1983). These fossils, long collected by amateurs from numerous localities in southeastern England, are early Eocene in age. The flora is dominated by angiosperms, with nearly 250 species identified. Only about a third of the genera in the flora, however, are referable to extant forms. Many of the taxa are tropical in distribution today, and the most closely comparable modern flora is considered to be that of the forests of southeast Asia. Other floras that consist almost exclusively of fruit and seed remains are the middle Eocene Nut Beds of the Clarno Formation of Oregon (Manchester, 1994a), the early Miocene Brandon Lignite flora, including palynology (Traverse, 1994), from the northeastern United States (Barghoorn, 1950; Tiffney, 1981, 1993), and an early Miocene flora from New Zealand (Pole, 1993). Space does not permit a full description of the numerous Cenozoic floras known from localities around the world that have enriched our knowledge of fossil angiosperm distribution and evolution. These include both megafossils as well as floras based on palynofloras (Palazzesi and Barreda, 2007). Discussions of some of the more extensive and wellknown floras can be found in the following examples. For Europe, see the classic works of von Ettingshausen (1853, 1858, 1867–1869, 1888); the summary of Mai (1995), as well as his numerous other works; Szafer (1946); Koch (1963); Velichkevich and Zastawniak (2003); and the summaries and references for Neogene floras in Kovar-Eder et al. (2008). For China, see Tao (2000); in North America, for the Neogene, see Brown (1935); MacGinitie (1962) (FIG. 22.278); Axelrod (1956); Wolfe (1964); and Manchester (1999). For the Paleogene, see MacGinitie (1953); Becker (1961, 1973); Brown (1962); Axelrod (1966); Wolfe (1977); Wolfe and Wehr (1987); and Wing et al. (2005).

flowering plants

995

Figure 22.278 Harry D. MacGinitie. (Courtesy H. N. Andrews.)

Figure 22.279 Charles J. Smiley.

The Miocene Clarkia flora from northern Idaho (Smiley and Rember, 1979) (FIG. 22.279) is important for the exquisite preservation of fossil angiosperm leaves from the site, some of which are still green in color. From these leaves have been described ultrastructural details such as

996

paleobotany: the biology and evolution of fossil plants

chloroplasts (Niklas et al., 1978), as well as fossil DNA (Chapter 1). This locality also contains an extensive record of epiphyllous fungi (Phipps and Rember, 2004; Phipps, 2007). The preservation of these fungi is comparable to those from the Eocene of Tennessee (Dilcher, 1965; Sheffy and Dilcher, 1971). There is an increasing awareness that the presence of certain types of fungi on fossil leaves may be a useful proxy record for some paleoclimatic constraints. The floras from the Deccan Intertrappean beds of India have been studied for many years and are noteworthy because the plants are permineralized in chert. The floras have been variously considered to be Late Cretaceous to Eocene, but more recent work suggests a latest Cretaceous (Maastrichtian) age (Chenet et al., 2007). These deposits include numerous angiosperm remains, including wood, stems, flowers (Verma, 1956), and fruits (Chitaley and Nambudiri, 1973). Two other important Cenozoic floras are the mid-Eocene Messel flora from Germany (Schaarschmidt, 1988) and the Geiseltal flora (Germany) from which green angiosperm leaves have been reported (Dilcher, 1967). The lacustrine deposits of the Messel area have been extensively studied for both their fauna and flora (Schaal and Ziegler, 1988), as well as for the economic importance of the Messel Oil Shale (Goth et al., 1988). The biota from this site has also been used to reconstruct the food web from this Eocene Lagerstätte (Ferguson, 2003). North American Eocene floras that are also known for their diversity and excellent preservation include the flora of the Green River Formation (MacGinitie, 1969), Florissant Formation (Manchester, 2001b; Meyer, 2003), and the middle Eocene Clarno nut beds flora (Manchester, 1994a). A more recently described flora comes from the Laguna del Hunco site in Patagonia, Argentina, and is early Eocene (Wilf et al., 2003, 2005a). This flora contains 186 species of plant organs and 152 species of leaves. This level of diversity exceeds any other Eocene flora on any other continent. These authors noted that although the flora occurs at 47°S latitude, it was living during the Paleocene–Eocene climatic optimum, when tropical biotas extended to much higher latitudes than today. The diversity of the plants from Laguna del Hunco was reflected in the diversity of the insects that fed on them, as Wilf et al. (2005b) found a greater diversity of damage types and insect feeding groups than in floras from North America of similar age.

Conclusions Several characteristics of angiosperms are believed to have played a role in their rapid diversification and ability to

Figure 22.280 Virginia M. Page

outcompete other groups. One of the most often cited is a pollination syndrome intimately associated with insects. The reticulate sculpture of many Early Cretaceous grains suggests that this association was well established by the Barremian, although some fossil grains with uncertain affinities also possess reticulate sculpture. In addition to entomophily, stigmatic compatibility has also been suggested as a contributing factor to the increased speciation rates among the early flowering plants. Another selective advantage often cited in early angiosperms is the shorter reproductive cycle and rapid seedling growth, which certainly provided the opportunity to outcompete many gymnosperms and to exploit new habitats. High humidity and warm temperatures in the tropics require a more efficient conducting system than might be expected if the flowering plants originated in a seasonally arid environment, and new research directions are not only examining this parameter in fossil plants but also examining these constraints in extant angiosperms that are thought to be basal. The evolution of vessels would have provided increased water conduction efficiency throughout the plant, and represents still another adaptation important to the exploitation of new niches. Detailed studies of fossil woods from all geologic periods will be especially important in tracking evolutionary adaptations and ecological parameters in fossil plants (Page, 1979) (FIG. 22.280) (Süss, 2007) (FIG. 22.281). The evolution of chemical defenses to combat herbivory, in the form of secondary compounds, may have contributed to the success of angiosperms as well (Tiffney, 1981). It has also been hypothesized that high chromosome numbers in some presumably primitive angiosperm families suggest that polyploidy may have played an important role in the rapid speciation of certain groups (Soltis and Soltis, 1990). As we continue to learn more about the living flowering plants that are so much a part of our lives today, it is no wonder

chapter 22

flowering plants

997

that biologists are motivated to continue the search for answers to their questions about the origin and subsequent evolution of this group. Will the discovery of additional specimens in older (and younger) rocks, together with the development and utilization of newer techniques, alter our current ideas about the evolution of the angiosperms? Most certainly. There are still fossil plants being discovered and reanalyzed for which the affinities within extant groups remain elusive (Dilcher et al., 2007). Angiosperm paleobotany, as in all scientific disciplines, will make its most significant advances as new, unconventional questions are asked and as additional answers are sought from new directions in the biology and evolution of fossil plants. One of these might be—did angiospermy evolve only once? Although the incompleteness of the fossil record is often cited as a reason for a lack of knowledge in many areas of paleobiology, the fact is that we have seen an explosion of new fossil finds and new knowledge of fossil angiosperms in recent years, and there is no reason to think that this rate of discovery will slow in the years ahead.

Figure 22.281

Herbert Süss

This page intentionally left blank

23 Interactions between plants and animals Early terrestrial ecosystem associations ...1001

Herbivory ........................................................................................1016

Animals on Land .............................................................................1001

Dispersal .........................................................................................1018

Early Plant–Animal Associations....................................................1001

Plants as habitat ................................................................1019

Herbivory ................................................................................. 1003

Other plant–animal interactions .....................1021

Defenses Against Herbivory ...........................................................1004

Mimicry ..........................................................................................1021

Fossil Evidence of Herbivory..........................................................1007

Pollination .......................................................................................1022

Interactions with vertebrates .............................1016

Conclusions ......................................................................... 1024

This is the fascination of plant and animal relationships: the richness of species, the spectacular variety and complexity of interactions, the beguiling loveliness of evolutionary creations, dynamic evolutionary processes and the quest for understanding. P. W. Price, 2002, p. 4 In this book we have examined microorganisms, fungi, and plants through geologic time, with the principal theme of demonstrating the diversity within various major groups, many of which are extinct, and where possible, tracing the evolutionary history of these organisms. This chronology has focused on the concept of evolutionary adaptation and how various plants have changed in response to the selective pressures in the ecosystem in which they once lived. In paleobotany such changes can be examined at the structural and morphological levels; however, these differences in reality reflect a mosaic of cellular, biochemical, and molecular changes that were, in turn, regulated by the genetic system within the organism. Such interactions between the environment and the genome of the organism represent the processes that drive evolutionary change. This volume, like many of the fossils in the rock record, is a chronology of groups of species that did not adapt to change and are now extinct. Biodiversity through time is the manifestation of species interactions that range from parasitism to mutualism, and the resources that plants, as primary producers, provide to the structure and dynamics of ecosystems.

Despite the extraordinary advances that have been made in all areas of paleontology during the last several decades, paleobotany is a long way from obtaining and analyzing unambiguous genomic data from fossil plants. Nevertheless, paleobotanists are increasingly examining various biotic interactions that existed millions of years ago; these can be used to more accurately evaluate and test hypotheses that focus on interactions in fossil ecosystems. The purpose of this chapter is to take a broad approach at depicting some of the dynamics that existed between plants and animals in the fossil record (Scott et al., 1985), and to compare these interactions with those that exist today (Herrera and Pellmyr, 2002 and papers therein). We have excluded from this survey plant–animal interactions that occur in marine paleoecosystems, such as microscopic algae living as symbionts with animals such as sponges, corals, or molluscs, or calcifying algae that contribute to the formation of reefs. This in no way diminishes the importance of these associations in the geologic history of the Earth; however, the majority of this volume addresses organisms living in terrestrial ecosystems. We have included a few non-terrestrial interactions in Chapter 4. The examples we

999

1000

Paleobotany: the biology and evolution of fossil plants

Figure 23.1 Circular hole feeding on angiosperm leaf (Cretaceous). Bar  1 cm. (Courtesy C. C. Labandeira and K. R. Johnson.)

present here are not exhaustive, but rather serve to illustrate some of the interpretations that can be examined and some of the hypotheses that can be developed from the fossil record. There are a number of types of evidence that can be cited as examples of plant–animal interactions in the fossil record. Incomplete leaf margins, holes in leaves of various types (FIGS. 23.1, 23.2), missing tissue in permineralized specimens (FIG. 23.3), and coprolites (FIG. 23.3) are but a few of the examples that can be used to assess levels of plant– animal interactions at a single collecting site (Labandeira and Allen, 2007), in different communities, or in ecosystems separated both temporally and spatially. Plant–animal interactions can also be examined as indicators of climate change (Wilf and Labandeira, 1999; Currano et al., 2008). Although certain types of evidence, such as coprolites, are relatively widespread in the fossil record, others, such as feeding or oviposition damages on leaves or other plant parts, are comparatively rare. The paucity of persuasive fossil data on leaf damage due to herbivory or oviposition is partially due to the collecting strategies used for many years for systematic studies of fossil plants. Taxonomic collections are almost always concerned with obtaining the most complete specimens to demonstrate the full complement of morphological features. Tattered or fragmented specimens

23.2 Ovoid skeletonization pattern (Paleocene). Bar  5 mm. (Courtesy C. C. Labandeira and K. R. Johnson.)

Figure

(FIG. 23.4), unless they obviously represent different or new taxa, are typically left behind rather than being packed, transported, cataloged, and eventually acquisitioned into a collection. Within the last 10–20 years, however, paleobotanists have begun to collect plants using more quantitative methods, to analyze not only plant diversity but also paleoecology and taphonomy (e.g., DiMichele and Wing, 1988; Gastaldo, 1989; McElwain et al., 2007). Use of these field methods has provided important data on the proportion of damaged leaves within a deposit.

CHAPTER 23 Interactions between plants and animals

1001

ancestor. Among the early arthropods were trilobites and crustaceans, both groups that can be traced back to the Early Cambrian, and the eurypterids in slightly younger Ordovician rocks. Diplopods (millipedes), myriapods (centipedes), and Collembola (springtails) are known from the Devonian, the same point in geologic time in which the presumed first insects (Hexapoda) are found (Grimaldi and Engel, 2005). Animals on Land

Figure 23.3 Tissue disruption and frass (arrow) in permineralized plant material (Triassic). Bar  450 μm.

Figure 23.4 Populus sp. with local necroses (Miocene). Bar  2 cm. (Courtesy BSPG.)

It is generally assumed that once plants became established on land, aquatic invertebrates rapidly followed, thus initiating the long-standing interactions between plants and animals (Jeram et al., 1990); however, there were microbial mats and possibly lichen-like symbioses present earlier that may have supported some of the first animals (Gray and Shear, 1992). During the transition to land, animals, much like plants, evolved various adaptations that allowed them to escape the physiological barriers to life on land, such as osmo- and thermoregulation, respiration, and water balance. Although there are no good fossil records of terrestrial arthropods as early as the Ordovician, several Late Ordovician paleosols are known to contain burrows (para-aggrotubules) that may have been made by animals (Retallack and Feakes, 1987). These authors suggested that these organisms consumed some form of plant material. There were also Silurian eurypterids that possessed several adaptations that are believed to have allowed them to exist for brief periods of time in a desiccating environment (Størmer, 1977). From slightly younger terrestrial rocks (Upper Silurian), Sherwood-Pike and Gray (1985) described pellets consisting of fungal hyphae thought to have been produced by a terrestrial fungivore. By the Carboniferous, the insect fauna was highly diverse, and this parallels the diversity of Carboniferous plants (Labandeira and Sepkoski, 1993). Moreover, the fossil record indicates that insects had not only evolved various adaptations needed to obtain food, such as wings and specialized mouthparts, but that they also were selective in the tissue systems that they consumed (Labandeira and Beall, 1990). Early Plant–Animal Associations

Early terrestrial ecosystem associations As noted in Chapter 6, there is compelling evidence that embryophytes became established on land earlier than the Late Silurian. Associated with the charophycean algae that are believed to have given rise to these land plants were various forms of aquatic arthropods. These are believed to have evolved from some form of segmented, soft-bodied wormlike

Once plants became established on the land, a number of plant– animal interactions rapidly evolved that may have included all the early land plants (Labandeira, 2006, 2007). One of the earliest known associations comes from the Lower Devonian Rhynie chert (Chapter 8). Fossilized along with the plants are springtails, mites (the smallest of the arthropods), insects, spiders, and spiderlike organisms termed trigonotarbid arachnids (FIG. 23.5) (Anderson et al., 2004; Dunlop et al., 2004). Also present are animals whose affinities remain equivocal (Dunlop et al., 2006).

1002

Paleobotany: the biology and evolution of fossil plants

Figure 23.5 Suggested reconstruction of the trigonotarbid Palaeocharinus rhyniensis from the Rhynie chert. (Courtesy N. Trewin.)

Figure 23.6 Molted exoskeleton of trigonotarbid inside a

dehisced sporangium in the Rhynie chert (Devonian). Bar 125 μm (Courtesy P. A. Selden.)

The extraordinary quality of the Rhynie chert preservation not only permits analysis of the animals that inhabited this ecosystem but also provides direct evidence for plant–animal and animal–animal interactions. For example, the mite Palaeotydeus devonicus lives inside empty sporangia of Aglaophyton major where it is attached to the inner sporangial wall and probably feeds by extracting nutritious sap from the degrading plant tissues. The trigonotarbid Palaeocharinus rhyniensis also occurs in many of the land plant sporangia (FIG. 23.6). The presence of trigonotarbids inside sporangia was initially believed to indicate

that these organisms were spore feeders (Kevan et al. 1975); others have suggested that the empty sporangia may have represented a site for molting of these animals (Rolfe, 1980). This latter suggestion would appear to have merit since the majority of the arthropods in the Rhynie chert appear to be predators and not herbivores (Shear et al., 1987). Still another idea is that trigonotarbids were predatory animals that visited the sporangia in order to feed on the mites living inside. The springtails at this site perhaps fed on spores, microbes, or decomposing plant litter. Also present in the cherts are numerous coprolites that can be classified into four basic organizational types, which suggests that the fauna in the Rhynie chert ecosystem was even more diverse in diet preference (Habgood et al., 2004). The exceptional preservation of the organisms in the chert makes it possible to examine these associations and to characterize the ecosystem in great detail (Anderson and Trewin, 2003). Trigonotarbids and pseudoscorpions (Shear et al., 1989) have also been found with plants at a Middle Devonian (Givetian) site near Gilboa, New York (Shear et al., 1987). The rocks at this site were originally studied for plant fossils, but when the shales were macerated with hydrofluoric acid a diverse assemblage of early arthropod cuticles was released (Shear and Bonamo, 1988), including remains of possible insects (Shear et al., 1984). Although the Rhynie and Gilboa sites are dominated by microarthropods, larger arthropods have also been described from Canada in the form of millipedes, scorpions, and arthropleurids (Shear et al., 1996). In addition to decomposing plant material along the shores of aquatic ecosystems, there were no doubt plant–animal associations in algal mats and stands of bryophytes that both provided food and appropriate microhabitats for various invertebrates (Gerson, 1973, 1976). It is hypothesized that herbivory among the early terrestrial arthropods evolved within groups that initially fed on detritus, perhaps along the shore (Little, 1983). Details on the morphology of fossil arthropods, including their mouthparts, genitalia, cuticular vestiture, and body shape, also supports the belief that the earliest forms were not herbivores, but rather were predatory (Shear and Kukalová-Peck, 1990). An important question then is why did these early arthropods move onto land if food was not the primary selective pressure? Possible reasons include refuge on land from other predators, expanding their own food niche, or perhaps as a part of their reproductive cycle. The discovery of coprolites (FIG. 23.7) containing land plant spores in the Late Silurian and Early Devonian provides indirect evidence that at least some early arthropods were herbivores (Edwards et al., 1995). Research based on comparative developmental biology and molecular phylogenetics and using molecular clock assumptions, suggests

CHAPTER 23 Interactions between plants and animals

1003

Figure 23.7 Coprolite composed of spores and cuticle-like material (Devonian). Bar  50 μm. (Courtesy D. Edwards and P. A. Selden.)

that insects arose near the Ordovician–Silurian boundary, a point in geologic time that is consistent with the discovery of the earliest putative land-plant fossils (Gaunt and Miles, 2002). The relationships of these earliest arthropods may be less certain, however, as some hypothesize that hexapods (insects and related groups) are not monophyletic (Nardi et al., 2003).

Herbivory Biological diversity today constitutes a complex mosaic of interrelationships among plants, animals, and microorganisms that spans multiple trophic levels. Herbivory, broadly defined, is the result of any organism that eats, bores into, or uses plant parts as food or shelter. Regardless of when animals initially became associated with plants, herbivorous relationships clearly span geologic time. Some of these interactions can be measured by morphologic and structural evidence within the plants (FIG. 23.8); in other cases the mouthparts of the animals provide evidence of feeding behavior (Labandeira, 1997, 2002). Still other examples include gut contents and coprolites, perhaps as a measure of preferred diet (Krassilov et al., 1997, 2003). These interactions reveal a continuum of evolutionary adaptations in which plants develop defense mechanisms to discourage herbivory, and the animals in turn evolve mechanisms to overcome the defenses of the plants. Despite the complexities of these interactions, fossils provide the only opportunity to recognize levels of association among organisms through geologic time (Wing and Tiffney, 1987). It is difficult to determine what types of associations occurred, however, when particular animal groups do not have living

23.8 Leaf of Mahonia with herbivore damage (Oligocene). Bar  1 cm. (Courtesy D. Erwin.)

Figure

representatives. Although the interplay among the various components in the ecosystem is a fascinating topic, we will focus primarily on the evidence of interactions as represented by fossil plants. It is generally assumed that foliar damage represents herbivory, but without distinctive evidence of a host response in the form of wound tissue or some form of reaction rim (FIG. 23.9), the tissue damage may be the result of detritivores, pathogenic fungi, or mechanical damage. Large collections of fossil insects, or collections made from a specific locality, like the famous Miocene Clarkia beds, also provide an excellent means of evaluating fossil plant–animal interactions (Smiley et al., 1975). The extraordinary preservation of some of these insects offers the opportunity to relate them to modern groups and thus establish

1004

Paleobotany: the biology and evolution of fossil plants

foregoing is an excellent example of how the study of fossil plants can contribute not only to reconstructing paleoclimate but also to understanding the interplay of biotic and abiotic factors that affect plants in their environment. One factor that may contribute to the relative scarcity of evidence for herbivore damage on fossil leaves is the fact that leaves generally become less palatable as they mature due to changes in leaf chemistry, such as an increase in substances like epicuticular waxes, tannins, and other storage products. Moreover, plants may also produce an increased number of fibers and additional fibrous tissue with age, which also makes them less palatable to herbivores. As a result, some fossil leaves that consistently show no damage may merely signal a level of selectivity by the herbivores based on leaf phenology, rather than a lack of specific herbivores or the use of particular defenses by the plant. Wilf et al. (2001) documented this effect in a study of insect damage on Paleocene and Eocene leaves from western North America. They found the greatest amount of herbivory on leaves with a short leaf life span. Defenses Against Herbivory Figure 23.9 Circular holes with a broad flange of reaction tis-

sue (arrow) (Eocene). Bar  1 cm. (Courtesy C. C. Labandeira.)

faunal assemblages that perhaps can be correlated with the flora from the same region. An approach that has been used to study herbivory in the fossil record measures the quantitative proportion of the area of leaf surface removed, presumably by herbivores, as measured against the total leaf surface, which is then evaluated in light of the paleoenvironment (Adami-Rodrigues et al., 2004b). The study of herbivory in geologic time has also been used as a means of understanding CO2 changes in the paleoatmosphere and past global warming. Currano et al. (2008) documented the combined effects of temperature and pCO2 on multiple forms of insect herbivory across the Paleocene–Eocene boundary in western North America. Their study of insect damage based on more than 5000 fossil leaves correlates well with rising and falling temperatures, and suggests that increased herbivory may occur under conditions of increasing pCO2 and global warming. This hypothesis, based on the fossil record of leaf damage, is supported by some modern studies of herbivory. For example, DeLucia et al. (2008) suggested that a reduction in nutritional quality of the plant tissue resulting from an increased carbon-to-nitrogen ratio may be the stimulus for herbivores to consume more plant tissue during times of high pCO2. The

Since plants are unable to move away from potential herbivores, they have evolved a variety of mechanisms to reduce the impacts of herbivory on reproduction and survival (Karban and Baldwin, 1997). Most of the strategies plants use against herbivory fall into two broad categories— mechanical and chemical defenses. Mechanical protection includes the production of thorns, spines, various forms of trichomes (FIG. 23.10), sclerotic tissues, and other morphological and architectural strategies including, in some cases, the placement of reproductive organs on the plant. Chemical defense (Woodhead and Chapman, 1986) involves the production of crystalline compounds, complex macromolecules, resins, and toxins that make plant material less palatable. Some chemical defense is indirect, as plants may respond to herbivory by releasing volatile organic compounds (VOCs) that can repel herbivores or attract predatory arthropods (Heil and Silva Bueno, 2007). These substances can also provide a chemical signal to nearby plants, which may cause them to produce additional chemical responses to repel herbivores. Within a single plant, VOCs can cause the initiation of other defense mechanisms, including increased production of nectar in extrafloral nectaries (EFN), which may serve to attract additional predatory arthropods. Even the ploidy level of the host may affect herbivory (Nuismer and Otto, 2004; Nuismer and Thompson, 2001). Current studies also suggest that tolerance to herbivory is heritable and can evolve in plant populations (Strauss and Agrawal, 1999).

CHAPTER 23 Interactions between plants and animals

1005

Figure 23.10 Epidermal papillae on the microsporophyll of

Lasiostrobus polysacci. Bar  35 μm. (From Taylor, 1970b.)

Figure 23.11 Lycopsid megaspore with hole gnawed in side

(Pennsylvanian). Bar  500 μm. (From Scott and Taylor, 1983.)

MECHANICAL PROTECTION Pubescent leaf and stem surfaces, that is those covered by trichomes, are generally regarded as adaptations to limit desiccation or mechanisms to reflect light and heat. Various types of erect or hooked, glandular or non-glandular (FIG. 23.12) trichomes may also function as defenses against herbivory (Levin, 1973; Duffey, 1986). Trichomes may be effective as physical barriers to insects, by making it more difficult for the animal to move around on the plant, or as structures that trap or injure herbivores. Some glandular trichomes are involved in either mechanical or chemical defense strategies or a combination of both (Jeffree, 1986; Simpson et al., 1999; Kellogg et al., 2002). Among the fossil plants discussed in this volume are countless examples of leaves with trichomes or papillae (FIG. 23.10). Some, like the multicellular capitate glands in Lyginopteris (Chapter 14), may have functioned to discourage herbivores both physically and chemically. Krings et al. (2002, 2004) suggested that certain types of non-glandular trichomes on the Carboniferous seed fern Blanzyopteris praedentata may have been effective as a physical barrier. Other trichomes are characterized by specialized secretory tips, equipped with a touch sensitive mechanism that could be activated by arthropods moving over the leaf surface. In this scenario the secretion would accumulate on the distal portions of the herbivore leg (i.e., tarsi and pretarsi; FIG. 23.12) and make it impossible for the organism to remain on the leaf (Chapter 16). Plants have evolved a variety of morphological defense mechanisms to try to prevent losses incurred through herbivory. One example is the organization and arrangement of

Figure 23.12 Distal portion (i.e., tarsus and pretarsi) of aphid leg clinging to trichome (Extant). Bar  20 μm.

reproductive organs designed to protect the spores, pollen grains, seeds, or fruits. For example, the upturned distal end of many types of sporophylls in compact cones (e.g., Lepidocarpon) may be interpreted as an obvious morphological defense to discourage predation. In the medullosan seed ferns, the outer surface of the pollen sacs and compound pollen

1006

Paleobotany: the biology and evolution of fossil plants

organs is thick walled and contains sclerenchymatous tissue. Moreover, when pollen is shed, all of the pollen sacs rupture along the thinner, inner-facing walls (e.g., FIG. 14.127), thus presumably making the mature pollen less accessible to predators. The presence of accessory structures surrounding seeds in the form of cupules (e.g., FIG. 15.51), sclerified seed integuments (e.g., FIG. 14.109), and fibrous interseminal scales (e.g., FIG. 17.83) are all examples of morphological innovations that may have initially evolved to discourage predation. In plants that disperse their spores to germinate on the ground, where litter-feeding arthropods are active, the presence of an extremely thick spore wall, sometimes highly ornamented by variously shaped projections, may constitute some form of selective advantage useful in discouraging certain herbivores. The evolution of lateral meristems in plants, that is, vascular and cork cambia, and the subsequent development of an arborescent growth habit (e.g., FIGS. 9.77, 12.2) occurred rapidly in several groups after their first appearance on the land. The selective advantage conferred by arborescence remains highly speculative, but such a habit did afford a number of advantages to the plants. Perhaps the most obvious of these is the ability to distribute reproductive propagules over a wider area, which in turn resulted in the potential to colonize new niches. Another is the location of reproductive organs in the distal parts of the plant, presumably out of reach of litter-dwelling herbivores. A coevolutionary response by animals to this plant innovation may have been the evolution of flight, which is well documented during the Carboniferous, but appears to have evolved by the Late Devonian not long after the first woody plants appear. It has been hypothesized that flight evolved as a dispersal mechanism, perhaps in response to temporary habitats caused by seasonal climatic fluctuations. Wings must have also served as a means to escape predators. Another equally plausible scenario is that flight provided terrestrial arthropods with the ability to forage on the more nutritious reproductive parts that were now located in the more distal regions of the plants. By the Carboniferous, nearly all the major groups of vascular plants (sphenophytes, lycopsids, pteridosperms, and some true ferns) had growth forms that placed the reproductive organs some distance from the ground. The notable exceptions were most of the homosporous ferns, but perhaps these clades of plants evolved chemical defenses early in their geologic history and therefore did not rely as much on mechanical protection (Cooper-Driver, 1978). CHEMICAL DEFENSES Plants use a diverse array of secondary metabolites and complex signaling pathways to respond to microbial attack and to discourage herbivory. In some instances, however,

herbivores have developed counter adaptations that allow them to ingest plant tissues containing repellent or toxic chemicals without negative effects (Wittstock et al., 2004). In many living plants, leaf surface chemicals have the ability to affect insect behavior, and insects possess the necessary sensory apparatus to detect chemicals by contact or through olfaction. Thus, leaf-borne waxes may function both as a physical and chemical barrier in plants. Since waxes are resistant to decay, they can be preserved in fossils and represent a potential source of information about this type of interaction. One of the most important defense mechanisms in living plants is the production of complex polymers that make the cell walls of plants indigestible to animals. Such materials as the complex polysaccharides cellulose and hemicellulose, and phenolic polymers such as lignin may represent types of plant defense mechanisms that have no doubt existed since plants and animals first interacted. Also included in this arsenal are inorganic crystals that presumably reduce digestibility in living plants such as the silica in the cell walls of Equisetum and grasses. Calcium oxalate crystals (CaC2O4) occur in a number of extant plants (Franceschi and Horner, 1980), including bark tissues in conifers (Hudgins et al., 2003) and dicots, where they serve as both a physical barrier, as the crystals are needlelike, and a chemical barrier to herbivory, as even a small quantity is poisonous to many mammals. Although it may be impossible to identify some types of crystals chemically in fossils, the presence of inorganic crystals may be inferred from a particular cell type or cell shape that contains comparable minerals in extant plants. Phytoliths are siliceous (opaline) bodies that are produced by the deposition of silica within or between the cell walls of certain plants, especially those in the Poaceae (grasses), although they also occur in other plants. They are important in the fossil record (Strömberg and Ferance, 2004), as phytoliths may be preserved in paleosols or sediments, even when the corresponding macrofossil is not present (Chapter 1). Since many are diagnostic for particular taxa, they can be used to identify past vegetation. Phytoliths (e.g., FIG. 1.62) have been reported from as early as the Devonian (Carter, 1999), and have been especially valuable in reconstructing grass-dominated ecosystems from the Cenozoic to the recent (Strömberg, 2004). Phytolith morphology has been used extensively in archaeological studies, and more recently, carbon isotope geochemistry has become an important tool that can be used to investigate shifts in C3 and C4 grasses during specific intervals of geologic time (Smith and White, 2004; Strömberg, 2004; Behrensmeyer et al., 2007). Is it possible that the abundance of phytoliths in some grasses prevents the plants from being eaten by large herbivore mammals?

CHAPTER 23 Interactions between plants and animals

Secondary compounds and toxins that function in living plants to discourage herbivory are more difficult to resolve in fossils. It may be that some of the biomolecules or secondary compounds already identified in fossil plants functioned as repellants or various forms of attractants, and perhaps in the future these can be traced in association with particular groups of herbivores through geologic time. A number of these compounds are important as molecular markers in the systematics and phylogeny of fossil representatives of some plant groups, for example the conifers (Otto et al., 2002, 2003; Yamamoto et al., 2006). Another form of chemical defense by plants is the production of secretions such as resins; amber is a fossilized form of resin. The earliest evidence of possible resins appears to be in the form of secretory canals or sacs found in some Devonian plants; by the Carboniferous, these materials appear to be widespread in seed ferns and cordaites. The production of large amounts of resin by some plants is evidence that these secretions play an important role in discouraging herbivory and invasion by microorganisms (see Langenheim, 2003 and references therein). The presence of foliar, endophytic fungi in some extant plants is important in providing a chemical defense against herbivory. This type of tritrophic interaction—fungi, plants, and herbivores—is relatively common in grasses (Vicari et al., 2002), and these microbial interactions may have operated in herbivore dynamics during the geologic past. Fossil Evidence of Herbivory

COPROLITES The well-preserved Euramerican Carboniferous floras represent an excellent opportunity to examine several parameters associated with phytophagous arthropods and plants. Cellulose is the principal component of most of the plant material ingested by terrestrial herbivores. Since all organisms except certain bacteria and fungi, however, lack the necessary enzymes to digest cellulose, terrestrial herbivores must have a symbiotic relationship with bacteria or fungi that live in their gut and function in digesting this polysaccharide. Associated with the various plant organs in coal balls are small (1–3 mm), circular–elongate structures made up of fragmented plant material that represent the coprolites (fossil fecal material) or frass of various phytophagous arthropods, especially oribatid mites (Labandeira et al., 1997). Similar structures have also been isolated from shales of various ages using acid maceration techniques (Hill, 1976; Scott, 1977). Some of the oldest coprolites (e.g., FIG. 23.13) have been described from Ordovician rocks (Richter and Richter, 1939), and by the Late Silurian Early Devonian there are numerous examples of fossil feces that are composed of spores,

1007

Figure 23.13 Coprolite associated with early land plants

(Devonian). Bar  50 μm. (Courtesy D. Edwards and P. A. Selden.)

Figure 23.14 Coprolite from the Rhynie chert (Devonian).

Bar  400 μm. (Courtesy H. Kerp.)

cuticle, and macerated plant material (FIG. 23.7) (Edwards et al., 1995). Well-preserved coprolites are a component of the Rhynie chert ecosystem and represent an important segment of the food web at this time (FIG. 23.14) (Habgood et al., 2004). Because the Rhynie chert coprolites appear as several distinct types that today cannot be related to the organisms that produced them with certainty, they are placed in three ichnotaxa: Lancifaex, Rotundafaex, and Bacillafaex. These authors suggested that the coprolites represent two feeding strategies—one that consists of indiscriminate feeding, perhaps by collembolans and myriapods, and a second more discriminate type, which may have involved some trigonotarbids and oribatid mites. Coprolites of the latter type are composed almost entirely of specific types of plant (FIG. 23.15) or fungal remains (FIG. 23.16). At least one study of modern oribatid mites suggests that of the mite species studied under laboratory conditions, a diet of bark algae was preferred over grass and herb litter (Hubert et al., 2001). This suggests that it may be possible to match coprolite composition with mite producers in the fossil record (FIG. 23.17), and to speculate on which enzyme systems may have been present in the animals. One of the early attempts to determine the systematic affinities of Carboniferous coprolites was by Baxendale

1008

Paleobotany: the biology and evolution of fossil plants

Figure 23.16 Rhynie chert coprolite composed of fungal

hyphae (Devonian). Bar  750 μm.

Figure 23.15 Rhynie chert coprolite composed of spores

(Devonian). Bar  750 μm.

(1979), who categorized a large number of coprolites in North American coal balls based on overall morphology and content from several different stratigraphic levels and geographic sites. A similar but stratigraphically more restricted approach was used by Scott and Taylor (1983), who analyzed three basic types of permineralized Carboniferous coprolites from coal balls collected at the Middle Pennsylvanian Lewis Creek, Kentucky locality (FIG. 23.18). These authors also documented the occurrence of the coprolites either in the matrix or in plant tissue. The most common form of coprolite in these coal balls was small (50 μm) and occurred in clusters; the plant material in these forms was impossible to identify. Another group included larger (150–900 μm), typically circular forms in which the periphery of the coprolite consisted of highly compacted plant material. Most of these coprolites occurred in the coal-ball matrix, and many were made up of spores and pollen grains. The largest forms (1–2 mm) were exclusively found in the matrix and typically included a variety of plant materials, although some were represented by large numbers of spores and sporangial fragments.

Frass (arrow) in tissue of seed fern stem (Pennsylvanian). Bar  500 μm. Figure 23.17

Using the sizes of fecal pellets produced by modern analogs, Scott and Taylor (1983) suggested that the most common peatinhabiting arthropods in this Pennsylvanian ecosystem were mites and collembolans, with millipedes and other phytophagous and saprophagous forms less common (FIG. 23.19). Labandeira et al. (1997) extensively sampled detritivory in plant tissues from several coal-swamp floras, using coal balls from four Upper Pennsylvanian sites in Illinois, USA.

CHAPTER 23 Interactions between plants and animals

Class 3

Class 2

This study substantiated mite boring in almost all plant groups during the Early and Middle Pennsylvanian, while insect herbivory becomes increasingly important during the Late Pennsylvanian. Permineralized coprolites of Permian and Triassic age from Antarctica indicate, however, that at least on this continent, oribatid mites were active components of the xylophagous fauna into the Mesozoic (Kellogg and Taylor, 2004). All of these studies demonstrate the great potential of analyzing interactions between plants and animals in geologic time. Similar studies need to be carried out for other permineralized floras from different sites and time periods, as Rex and Galtier (1986), for example, did with a Viséan (Late Mississippian) flora from France. Not only should it be possible to document a variety of ecological interactions, such as food preferences (FIG. 23.20) and types of tissues (FIG. 23.21) inhabited, but such analyses should also make it possible to detail microfaunal composition associated with particular habitats and floras. In some reports a single type of coprolite has been identified in one plant organ. For example, Rothwell and Scott (1983) described numerous, oval (450 μm in diameter)

Class 1

Number of examples

50 40 30 20 10 0 m m m 00 m 2.  00 2. m 5– m 1. 1.5 m 0– m 1. 0 ) 1. .5 5– (0 0. 500 m 0– μ 21 210 m 0– μ 18 180 m 0– 0 μ 15 15 0– μm 12 20 –1 m 90 0 μ –9 m 60 0 μ –6 30 μm 30

0–

Size

Mites Size ranges of fecal pellets of modern groups

Collembola Millipedes Cockroaches Grasshoppers and crickets

Figure 23.18 Histogram of coprolites from a single collecting

site measured against size of the fecal pellets of several groups of modern arthropods. (From Scott and Taylor, 1983.)

Phytophagous insects

Megasecoptera Palaeodictyoptera

Canopy Feces

Leaves cones seeds

(Xylophagous) Insect dispersal

Coleoptera Acarida Phytophagous Rotting insects stump Orthoptera Feces

Leaves twigs spores pollen seeds Litter

Litter Litter

Litter

Litter layer

Peat

Diplopoda Saprophagous Arthropods

1009

Collembola

Acari

Hollow trunk

Arthropleura

Feces Humified remains Coprolites, coprophagous Humivores

Feces

Feces

Figure 23.19 Diagram showing potential pathways of herbivory in a Carboniferous swamp. (From Scott and Taylor, 1983.)

1010

Paleobotany: the biology and evolution of fossil plants

Figure 23.21 Gallery in Premnoxylon sp. wood filled with

elliptical frass pellets (Pennsylvanian). Bar  600 μm. Figure 23.20 Several Lycospora spores within a coprolite showing perforation of the wall (Pennsylvanian). Bar  50 μm.

coprolites constructed of finely ground plant material in the ground tissue of a Pennsylvanian stem, whereas Lesnikowska (1990) reported coprolites and wound tissue in petioles of Psaronius chasei. Similar structures have also been noted both in the ground tissue of the Mesozoic fern Tempskya (Seward, 1924) and in the stem parenchyma of the Triassic tree fern Itopsidema (Ash, 2000). In these examples, the animals responsible for the coprolites chewed into the softer, and perhaps more nutritious, parenchymatous tissues of the stem, or possibly attacked these tissues because they were easier to burrow through. Information about the distribution of fossil insect families through time is another important component of studying plant–animal interactions, especially herbivory (Labandeira and Sepkoski, 1993; Labandeira et al., 1994). Examining the diversity of insects in the fossil record (e.g., FIG. 23.22), together with data about mouthpart morphology (Labandeira, 1997), represent important avenues of research that are useful in documenting feeding guilds and mouthpart evolution during the evolution of herbivory. For example, some tree ferns in the Late Pennsylvanian show evidence of piercing and sucking by insects, based on the wounds present in the plant tissues (Labandeira and Phillips, 2002). Taken in a broader context, the presence or absence of insect damage on a large number of fossil plant specimens can be used to examine herbivore changes and even extinctions (e.g., K–P boundary; Labandeira et al., 2002), and to track specialized

Figure 23.22 Bittacus egestionis from the Eocene Green River Formation shales. (From Carpenter, 1955.)

feeding associations, recovery rates, and the appearance of new coevolutionary associations between plants and insects. In addition, such studies can also identify new feeding strategies (Labandeira et al., 1994). In the case of Carboniferous coal-swamp communities, where the composition of the plant community is well known, hypotheses can be constructed that match feeding type and mouthpart structure with the plant groups present (Labandeira and Phillips, 1996a). This information can then be used to elucidate coevolutionary relationships between plants and insects. Wood boring is a common feeding mode in the modern flora, and the fossil record provides evidence of this

CHAPTER 23 Interactions between plants and animals

interaction as well. In the wood of Premnoxylon, a presumed cordaitean root (see Chapter 20) (Cichan and Taylor, 1982c), the wood was apparently digested to form a complex series of ramifying tunnels ranging from 0.3–0.6 mm in diameter that contain the fecal remains of the organism (FIG. 23.21). Each of these coprolites, which are attributed to mites, is elliptical and 75 μm in diameter. Present in one gallery is a cylindrical structure 200 μm in diameter that may represent part of a body fossil of the animal. With this example, as with others, it is difficult to determine whether the galleries were formed as a result of foraging or represented the habitat of the organism. Galleries have also been reported in the wood of Protocupressinoxylon cupressoides from the Jurassic of China (Z. Zhou and Zhang, 1989b). These galleries range up to 500 μm in diameter and some contain frass pellets of two different sizes (100–165 μm and 50–55 μm). The authors suggested the fecal remains were produced by Coleoptera, although Kellogg and Taylor (2004), using criteria from Labandeira et al. (1997), suggested these remains were produced by mites rather than beetles. In mummified wood of Larix altoborealis from the Eocene of the Canadian Arctic, however, the gallery and tunnel network can be attributed specifically to Dendroctonus (bark beetles) (Labandeira et al., 2001). In this case the coevolutionary relationship between conifer and bark beetle lineages can be traced back to the earliest representative of the genus Larix, which has important implications for the phylogeny of both the plants and the insects. An interesting example of tritrophic interactions was reported from the Upper Jurassic Morrison Formation of western North America (Hasiotis, 2004). Petrified conifer wood shows signs of fungal rot, especially in the heartwood area, and oval to elliptical cavities occur along with the irregular cavities formed by the action of the fungi. By comparison with modern associations, Hasiotis suggested that these traces could have been formed by insects, such as termites, beetles, or ants, that were feeding on the fungus. Coupling information about chemical deterrents in modern hosts with symptoms seen in certain fossil plants provides the opportunity to trace potential chemical interactions between host and herbivore over extended periods of geologic time, and perhaps to provide a means to determine when some of these interactions evolved. As new techniques become available that can be used to discriminate the chemical signatures in fossil plants, the opportunities to examine the interplay between host defense mechanisms and herbivores will open up new and exciting areas of evolutionary biology. Not all coprolites are formed by arthropods that feed on living plant material. Some feed on rotting wood,

1011

whereas others consume the waste products of other organisms. Some of these types may be difficult to distinguish in the fossil record, but others, for example those that feed on fungi, should be detectable (Sherwood-Pike and Gray, 1985). GUT CONTENTS One of the most accurate methods of determining precisely what organism is feeding on which plants is by analysis of the gut contents of fossil animals (Molnar and Clifford, 2000). For obvious reasons there are not a large number of reports of this type of interaction, but paleobiologists continue to search for examples of this type of exceptional preservation. In one study, Rolfe and Ingham (1967) reported the presence of Carboniferous lycophyte tracheids in the presumed gut of Arthropleura, and spores have been extracted from several Mazon Creek invertebrates (Carpenter and Richardson, 1976; Scott and Taylor, 1983). Pollen grains have been reported as the primary diet of some insects in the Permian (Rasnitsyn and Krassilov, 1996) and of sawflies (Xyelidae) from the Lower Cretaceous of the Transbaikalian region, Russia (Krassilov and Rasnitsyn, 1982, 1997, 1998). These reports suggest the presence of pollinivory in some groups. Such studies can be useful in tracing coevolutionary patterns and provide insight into fossil community structure. In another example, the pollen found inside each of five Cretaceous xyelid species is of a single type, suggesting feeding preferences, or that the plants produced pollen over a short period of time (Krassilov et al., 2003). Several different pollen morphotypes in a primitive Permian booklouse suggest that this species was not dietary specific (Krassilov et al., 1999). Fossil seeds in stomach contents can also be used as indirect evidence about the dispersal agent of a particular plant species (Sturm, 1978). For example, the presence of whole seeds might suggest that the vertebrate was acting as a dispersal agent, whereas fragments of seeds would suggest they merely served as a food source. Gnaw marks and perforations on seeds also provide evidence of animals seeking plants as a source of food (Collinson and Hooker, 1991; 2000) (FIGS. 23.23, 23.24). MARGINAL FEEDING Another more direct method of measuring the effects of herbivory is by directly observing leaves (FIG. 23.25) and other plant parts that have been damaged by phytophagous organisms. This type of damage was once considered to be a rare feature in fossil foliage (Berry, 1916b, 1930; Van Amerom and Boersma, 1971), but today it is well documented in fossils (Labandeira et al., 2007a; Currano et al., 2008)

1012

Paleobotany: the biology and evolution of fossil plants

23.23 Seed of Stratiotes showing gnaw marks (Paleogene). Bar  1 mm. (Courtesy M. Collinson.)

Figure

Figure 23.24 Fossil seed showing ellipsoidal perforations (arrow) (Pliocene). Bar  2 mm. (Courtesy C. C. Labandeira.)

Figure 23.25 Glossopteris sp. leaf showing damaged margin

especially in Cenozoic angiosperm leaves. Marginal feeding has also been noted on Carboniferous Neuropteris pinnules (Scott and Taylor, 1983) (FIGS. 23.26, 23.27), Mississippian Triphyllopteris pinnules (Iannuzzi and Labandeira, 2008), and Permian gigantopterid leaves (Beck and Labandeira, 1998). Glasspool et al. (2003) reported on marginal, non-marginal,

and apical leaf-feeding traces in two gigantopterid species from the Permian of northern China that indicate selective feeding strategies within Cathaysian gigantopterids and the preferential targeting of Gigantonoclea leaves. Plumstead (1963) illustrated Permian Glossopteris leaves that were extensively damaged by phytophagous activity. The author reported that only those

(Permian). Bar  1 cm. (Courtesy S. McLoughlin.)

CHAPTER 23 Interactions between plants and animals

Figure 23.26 Neuropteris sp. pinnule showing feeding traces along the margin (Pennsylvanian). Bar  2 cm.

1013

herbivory has also been reported on glossopterid bracts from the Permian of southern Brazil (Adami-Rodrigues et al., 2004a). The Mesozoic fossil record also contains numerous examples for plant/animal interactions in the form of damage on plants (FIG. 23.28) caused by phytophagous or pollinivorous insects (Kelber and Geyer, 1989; Grauvogel-Stamm and Kelber, 1996; Ash, 1996, 1999, 2005; Scott et al., 2004). In a few instances, morphologically distinct marks and patterns (FIG. 23.29) on leaves can be used to identify the insect responsible for the pattern, since certain groups of insects produce specific patterns as they consume leaf tissue. For example, Brooks (1955) described an angiosperm leaf from a clay pit near Puryear, Tennessee (Eocene), that contained four semicircular cuts 5–6 mm in diameter along one margin, which are suggestive of patterns produced by a female megachilid or leaf-cutting bee (FIG. 23.30). DEFOLIATION In modern ecosystems, one of the most dramatic examples of herbivory involves the defoliation of forest trees. By reducing photosynthetic surface area, leaf-feeding herbivores reduce tree growth by depleting carbohydrate reserves. A variety of interrelated biological and physical parameters influence the development of insect populations that are involved in defoliation. Although it has been hypothesized that such interactions have existed throughout the geologic history of plants and animals, no data from the fossil record are yet available to document this interaction. Since there are distinct changes in the size of tree rings in the wood of trees that have been defoliated by insects (FIG. 23.31) (Schweingruber, 1989), interactions of this type might be studied by examining growth-ring production in fossil wood specimens. Although growth-ring structure has been examined in a number of fossil woods (Creber and Chaloner, 1984b; Parrish and Spicer, 1988; Taylor and Ryberg, 2007), none of these reports were able to correlate ring irregularities with possible defoliation.

Figure 23.27 Neuropteris pinnules showing various patterns

of marginal bite marks. (From Scott and Taylor, 1983.)

leaves assumed to be very young show evidence of herbivory. Additionally, on the basis of occurrence of wound tissue on the lamina, it appears that the foliage was eaten while the leaves were still on the plants. Although the different Glossopteris species possess distinctive types of bite marks, it is not known whether these herbivores were monophagous, that is, specific to a particular taxon, or polyphagous. In addition to feeding damage on leaves of Glossopteris and Gangamopteris,

LEAF MINERS In addition to the general feeding of phytophagous insects, there are other, more specific types of herbivory on plants. For example, some larvae feed selectively on the mesophyll tissues of leaves. This activity results in the formation of feeding channels or mines (FIGS. 23.32, 23.33), as well as disklike excavations or holes that can be identified in the leaf. The herbivore removes the intervein tissues, leaving primary and secondary veins generally untouched. The presence of a reaction rim outlining the damaged area indicates that the leaf was alive when the mining took place. This feeding behavior is employed by particular species of Diptera, Coleoptera,

1014

Paleobotany: the biology and evolution of fossil plants

Figure 23.28 Cuticle of Nilssonia sp. showing holes caused by herbivore (Cretaceous). Bar  40 μm. (Courtesy J. Watson.)

Figure 23.29 Series of circular or sucking marks on Sterculia leaf

(Paleocene). Bar  1 mm. (Courtesy C. C. Labandeira and N. R. Cúneo.)

Hymenoptera, and Lepidoptera (Hering, 1951). The shape of the mines and pattern of the frass are features that can help identify the herbivore in the fossil record. Leaf mines are also sometimes referred to as window feeding or skeletonization, and have been described principally from angiosperm leaves (FIGS. 23.34, 23.35), although they can also occur in bryophytes, ferns, and gymnosperms. Leaf mines are sometimes associated with other examples of herbivory, such as chewed margins of leaves (FIG. 23.36) (Rozefelds, 1988a). Although most fossil mines have been reported from Cenozoic leaves, especially angiosperm leaves, there are a few reports from the Carboniferous (Müller, 1982), Permian (e.g., Russellites; Labandeira and Allen, 2007), and Triassic.

A Triassic example is Triassohyponomus, a leaf mine in the form of a series of tightly curved coils on Heidiphyllum leaves (Rozefelds and Sobbe, 1987). Unlike modern leaf mines, there is no frass present and no apparent increase in the size of the mine in the fossil leaves. In modern miners, the size change results from an increase in the larva as it grows. Mines have also been reported from the Mesozoic seed-fern foliage type Pachypteris (Rozefelds, 1988b). Many of the mines on this leaf closely follow the margin of the pinnae and are morphologically similar to those produced by lepidopterans in the Nepticulidae. Leaf mines have been described from a number of Cenozoic angiosperm leaves (Crane and Jarzembowski, 1980; Wilf et al., 2005b). These can be divided into three basic types. The first is an S-shaped mine in which the channel gradually widens distally. The second type is much smaller (0.44 mm wide) and extends from a secondary vein to the leaf margin and back to its source. The third type includes groups of two to four mines that originate at the leaf margin and often cross veins. Most of these mines are attributed to the Lepidoptera. A large number of morphological types of mines and galls, some comparable to modern analogs, have been identified on angiosperm leaves from the Upper Cretaceous of southern Negev, Israel (Krassilov, 2007), and the Miocene of Spain (Diéguez et al., 1996). Mines with a very specific morphology were found on monocot (ginger) leaves from the latest Cretaceous and Eocene of the western United States (Wilf et al., 2000). The same sort of mines are found on leaves of members of the Zingiberales today in the Neotropics. Interestingly, their presence in the

CHAPTER 23 Interactions between plants and animals

1015

Cross section of wood showing numerous galleries along ring boundary (Permian). Bar  5 mm (Courtesy S. McLoughlin.)

Figure 23.31

23.30 Angiosperm leaf showing scalloped incisions along margin perhaps caused by leaf-cutting bee (Eocene). Bar  12 mm. (From Brooks, 1955.)

Figure

latest Cretaceous extends the fossil record of the beetle family, Chrysomelidae, by at least 20 myr as compared to known insect body fossils. WOUND TISSUE In addition to arthropods with chewing mouthparts in the Devonian, there were others from the Lower Devonian (Emsian) that appear to have possessed piercing mouthparts

Figure 23.32 Serpentine leaf mine showing zigzag frass trail

(Eocene). Bar  5 mm. (Courtesy C. C. Labandeira and N. R. Cúneo.)

1016

Paleobotany: the biology and evolution of fossil plants

Figure 23.33 Elongate, narrow abrasion pattern on lauraceous leaf (Eocene). Bar  3 mm. (Courtesy P. Wilf and C. C. Labandeira.)

(Labandeira et al., 1988). One line of evidence for this type of feeding behavior is the presence in the Carboniferous and Permian of two orders of insects, the Palaeodictyoptera and Megasecoptera, which had mouthparts modified for piercing and sap sucking (Wootton, 1976). The second line of evidence involves the identification of wound responses in plants formed as a result of tissue penetration. Kidston and Lang (1921a) described dark areas they interpreted as necrotic sites in the axes of Rhynia gwynne-vaughanii that were later interpreted as damage caused by arthropods with some form of piercing or sucking mouthparts. The presence of wound tissue attributed to an arthropod with piecing mouthparts in the Carboniferous fern Etapteris is another example of this type of feeding group. The presence of wound tissue in general may also represent evidence of plant–animal interactions (Stopes, 1907). Banks (1981) reported the presence of a limited amount of periderm in a stem of Psilophyton dawsonii which was interpreted as evidence of wound repair after attack by a phytophagous arthropod. Such responses indicate that the biochemical pathways necessary to heal wounds in plants were in existence not long after plants became established on the land.

23.34 Skeletonization of interveinal tissue on Macginitiea wyomingensis (Eocene). Bar  1 cm. (Courtesy P. Wilf and C. C. Labandeira.)

Figure

Interactions with vertebrates Herbivory

The most obvious interaction between vertebrates and plants is the consumption of plants as food. Herbivory by terrestrial vertebrates dates back to the Early Permian (Weishampel and Norman, 1989), although most of this evidence is based on the functional morphology and dentition of the animals and not on plant evidence (Hotton, 1986), since vertebrate herbivory generally results in the complete destruction of the plant or the leaf. In instances where plants were selectively browsed, that is, only partially defoliated, it is still difficult to distinguish this type of activity in the fossil record since plants are almost always preserved in disarticulated assemblages.

CHAPTER 23 Interactions between plants and animals

1017

Figure 23.36 Large holes on cercidophyllaceous leaf (Paleocene). Bar  6 mm. (Courtesy C. C. Labandeira and K. R. Johnson.)

Figure 23.35 Skeletonization pattern across major veins of

dicot leaf (Cretaceous). Bar  8 mm. (Courtesy C. C. Labandeira and K. R. Johnson.)

Today most of the vertebrates that eat plants are mammals, with just a few herbivorous reptiles, for example land tortoises; in the Mesozoic, however, just the reverse appears to have been true. The dominant group of Mesozoic reptiles were the dinosaurs, some of which were successful plant consumers. The ornithopods or bird-hipped dinosaurs had special muscles and jaw mechanics that allowed them to masticate and grind plant materials. Many also apparently had gizzard stones that are thought to have aided in grinding plant materials, and stomach microbe systems that aided in digestion. Norman and Weishampel (1987) suggested that certain types of plant-eating dinosaurs were able to remain as the dominant vertebrates throughout the Mesozoic because they possessed jaws that allowed them to grind plant material successfully. Studies have also incorporated quantitative data into hypotheses about jaw mechanics and feeding behavior

(Barrett and Rayfield, 2006). Although dinosaurs have been suggested as a possible causative agent in the origin of angiosperms, data from variety of sources now indicate that this assumption is not valid (Barrett and Willis, 2001). Paleontologists have refined estimates of dinosaur size and weight based on bone measurements and more intact specimens. These data can, in turn, be integrated with collection data to estimate population size (Anderson et al., 1985), and to establish browse profiles for individuals and populations (Coe et al., 1987). Browse profiles compare the height of vegetation at a particular site to the size of herbivorous animals present (Gierli n´ ski and Pie´nkowski, 1999). When plotted through geologic time, it is possible to detect changes in the height of feeding levels. Such a procedure provides a method of determining changes in the coevolution of plant and animal size through time. Rees et al. (2004), however, suggested that dinosaur preservation may represent a taphonomic bias for environments that were drier and thus not necessarily representative of the vegetation at the time. Skeletal changes, including the presence of grinding dentition, are apparent in certain dinosaurs during the interval between the Late Jurassic and Late Cretaceous. Although some have suggested that the extinction of the dinosaurs was coupled to changes in the vegetation, others suggest this hypothesis may not be valid (Coe et al., 1987).

1018

Paleobotany: the biology and evolution of fossil plants

Wing and Tiffney (1987) proposed that data from a variety of sources can be used to develop a synthesis of tetrapod– plant interactions over a continuum of geologic time. Using herbivore size, locomotion, diet, diaspore size, and flowering plant stature, they categorized terrestrial biotas since the Jurassic into four stages. In the first stage, herbivory by large herbivores, such as dinosaurs, produced a continually disturbed habitat. In this type of environment, the early flowering plants would have been successful because they were small in stature and produced a large number of small propagules that allowed them to rapidly colonize new habitats (r-selection). This is consistent with the suggestion of Bakker (1978), who viewed the Cretaceous low-browsing dinosaurs (ornithischians) as closely correlated with the rapid radiation of the angiosperms. With the increased geographic distribution of angiosperms in the Late Cretaceous, Wing and Tiffney (1987) hypothesized that selection began to favor smaller dinosaurs that could browse on the smaller plants. They suggested that the extinction of large herbivores at the Cretaceous–Paleogene boundary modified the selective pressures on angiosperms, resulting in greater competition among plant species, coupled with an increase in seed dispersal by small animals. The increased size of some flowering plants in the Cenozoic suggests that some groups were becoming K-selected, that is, they produced fewer, larger propagules in more stable environments. It is at this point in geologic time that seed dispersal through interactions with animals became important to the radiation of angiosperms. Finally, climatic influences in the Neogene favored plants with shorter life histories, and thus herbaceous forms prospered. Several authors have plotted the diversity of floras through time, in particular those that may have served as food for herbivorous dinosaurs (Krassilov, 1981b, 1987). Generally, the Middle Triassic floras that supported dinosaurs consisted of cycadophytes and ferns. These gave way to conifer forests in the Jurassic, and eventually to angiosperms in the Cretaceous. Precisely what types of plants were eaten by dinosaurs is not known in detail, but the large amount of plant material needed to sustain life necessitated that these animals most certainly were indiscriminate foragers. DENTITION The analysis of teeth in determining the diet of various fossil vertebrates is also a viable source of information about plant–animal interactions in the geologic record. Collinson and Hooker (1987, 1991) correlated mammalian faunal changes with those of associated plants from Cenozoic rocks of southern England. By charting tooth structure with the occurrences of megaspores, seeds, and fruits, these authors were able to provide information about the diet of the animals

extending from the late Paleocene into the Oligocene. For example, there is a high proportion of soft fruit-eating frugivores in the early–middle Eocene that gives way in the Oligocene to forms adapted to eating stony fruits. Such an approach, in which a large data matrix is used, provides the only opportunity to understand some of the complex ecological interactions that were operative over an extended period of geologic time. As noted by Tiffney (2004), dentition and dispersal may not always be linked. Imprints of certain types of phytoliths have been identified on the fossil teeth of Gigantopithecus, a large Pleistocene fossil ape from China (Ciochon et al., 1990). Based on the correlation of particular types of phytoliths with their parent plants, these researchers hypothesize that this animal ate large amounts of fibrous grasses, such as bamboo, as well as certain types of tropical fruits. COPROLITES AND STOMACH CONTENTS There are only a few reports of actual stomach contents of plants from dinosaurs. Kräusel (1922b) described conifer leaves and shoots as well as fruits and seeds from the stomach of Annatosaurus, and Stokes (1964) reported woody stems up to 1 cm in diameter. A few coprolites containing plant remains have been described that may have been produced by dinosaurs (Hill, 1976; Viera and Torres, 1979), but they contribute little information about food preferences of the animals because the remains are too fragmentary to identify the plants present (Thulborn, 1991). The presence of leaf-borne fungi in coprolites, however, has been used to suggest that the animals were tree browsers (Kar et al., 2004). Chin and Kirkland (1998) reported on an assemblage of coprolites from the Upper Jurassic Morrison Formation in which they found seeds, pieces of wood, fragments of leaf cuticle, and cycadophyte leaves and petioles. In thin section, it was possible to see mesophyll and vascular bundles within the leaves. One of the more interesting coprolite finds is the presence of decayed wood in coprolites of herbivorous dinosaurs (Chin, 2007). Since wood holds little nutritional value and cannot be digested by most herbivores, Chin suggested that perhaps the animals ate wood that had been partially rotted by fungi only when other food was scarce (Chapter 3). Phytoliths recovered from dinosaur and other vertebrate coprolites have been used to examine the coevolutionary interactions between vertebrate herbivores and grasses (Prasad et al., 2005). Dispersal

Plants, especially angiosperms, use vertebrates for fruit and seed dispersal in several ways. There are numerous adaptations on the surface of seeds and fruits that allow the seeds to be attached to animals and thereby be widely

CHAPTER 23 Interactions between plants and animals

disseminated. Frugivorous animals also play an important role in the distribution of plant seeds, as some seeds are designed to pass through the gut of the animal unharmed, while the fruit wall is digested. Janzen and Martin (1982) reported that the extinction of a large number of herbivorous vertebrates in the Pleistocene had an effect on the distribution of several plant species in Central American lowland forests. They further note that several of the existing plant species continue to produce spines, although the plants are no longer browsed by herbivores. In this study, several unusual modern fruit and seed traits could be explained, not as examples of coevolution among living species, but rather as evolutionary holdovers from previous interactions (Janzen and Martin, 1982). Although frugivory did not exist prior to the evolution of the angiosperms, early seed plants may also have had adaptations to disperse their seeds or alternatively to prevent herbivory. Seeds with a fleshy outer integumentary layer occur early in the late Paleozoic seed ferns and extend into the Mesozoic. Some medullosan seeds are very large, up to 8 cm long, and the parenchymatous sarcotesta may have been a food source for Paleozoic arthropods or tetrapods. Although cycad seeds possess an outer fleshy layer that is rich in sugar, and brightly colored, the leaves of the plant and megagametophyte of the seeds contain toxins that discourage herbivory (Brenner et al., 2003). It has been suggested that the large populations of dinosaurs and cycads indicate coevolutionary adaptations in which the outer integument of the cycad seeds served initially as a visual attractant but that the toxins within the seed served to discourage certain types of herbivores from gnawing or masticating the seeds. Early angiosperms in the Cretaceous also had fleshy fruit walls; however, whether any of these plant adaptations served as dispersal mechanisms remains a difficult question to answer (Tiffney, 2004).

1019

Figure 23.37 Arc-like oviposition sites of damselflies on dicot

leaf (Cretaceous). Bar  5 mm. (Courtesy C. C. Labandeira and K. R. Johnson.)

Plants as habitat Cuticles of various arthropods associated with plant remains have been identified as early as the Devonian, and no doubt they existed earlier. A more passive level of plant–animal interaction certainly involved situations in which plants served as refuges in predator–prey relationships. By their existence, plants most certainly assisted in ameliorating the microclimates where various types of animals lived, ranging from the litter to the undersides of leaves or along stems. Plants also served as sites for ovipositing insects to lay eggs (FIGS. 23.37, 23.38) (Krassilov et al., 2007). For example, distinctive lenticular scars caused by ovipositor insertion of

Figure 23.38 Insect eggs on surface of Nilssoniopteris sp. leaf

(Triassic). Bar  50 μm. (Courtesy C. Pott.)

1020

Paleobotany: the biology and evolution of fossil plants

insect eggs (ovipositional scars) have been described from Mesozoic Equisetites stems and Taeniopteris-type leaves from Europe and North America; the scars are generally interpreted as having been produced by Protodonata or Odonata (Kelber and Geyer, 1989; Grauvogel-Stamm and Kelber, 1996; Ash, 2005). To date, the oldest evidence of oviposition scars are from latest Pennsylvanian (late Gzhelian) Calamites stems from the Graissessac Basin of France (Béthoux et al., 2004). These authors noted that sphenophyte axes are the most commonly found hosts in Late Triassic occurrences as well. Schmeissneria leaves from the Jurassic of Germany display similar surface structures that have been interpreted as odonatan egg remains (Van Konijnenburg-Van Cittert and Schmeissner, 1999). Actual insect eggs of uncertain affinity that are preserved on leaf cuticles (FIG. 23.38) of the bennettitalean foliage Nilssoniopteris haidingeri from the Upper Triassic of Austria have been documented by Pott et al. (2008). These authors showed that the eggs were attached to the lower leaf surface and arranged in circles. Insect galls or eggs have been reported on the underside of the Jurassic dipterid fern leaf Dictyophyllum (Webb, 1982), and from conifers and ginkgophytes from the Jurassic– Cretaceous of eastern Siberia (Vasilenko, 2005, 2007). At some stage in the evolution of plant–animal interactions, some plants evolved strategies to mimic egg cases. For example, in modern Passiflora numerous nectaries are produced on the leaves and petioles that look like larval egg cases. Such structures are believed to discourage ovipositing insects from laying on sites already populated by eggs. In the case of many plants with extrafloral nectaries (EFN), ants and aphids become associated with the plants, using the EFNs as a source of sugar. Some of these insects, especially the ants, are well known for these interactions, and they serve to keep other, herbivorous insects off the host plant. These secondary levels of interaction may be impossible to resolve in fossil associations. There is compelling evidence that caddis flies constructed egg cases from Eocene leaf fragments (Berry, 1927). In Cretaceous deposits, these egg cases have been found with Karkenia seeds attached to them. Some of these are believed to have functioned as floats (Krassilov, 1987). Labandeira and Phillips (1996b) described the activities of larva of the Holometabola forming galls in tissues of tree-fern fronds in the Pennsylvanian. Galls are swellings or spheroidal structures (FIG. 23.39) that develop on plants in response to insects or mites. Both the plant and the insect are involved in the structure of the gall, which includes internal nutritive tissue and outer tissues that protect the larvae inside (Labandeira and Allen, 2007). Galls represent another foliar feature that can be detected in fossil plants,

Figure 23.39 Columnar galls (Paleocene). Bar  1 mm. (Courtesy A. Iglesias and C. C. Labanderia.)

especially leaves, and are known from late Pennsylvanian and Mesozoic seed ferns and conifers (FIG. 23.40). The structures have been assigned to ichnogenera such as Acrobulbillites and Pteriditorichnos. Several examples of galls have been reported from Upper Triassic leaves from the famous Chinle Formation in the southwestern United States (Ash, 1996). In addition to feeding traces on several foliage types (e.g., ferns, Bennettitales), galls have also been identified on the enigmatic gymnosperm Dechellyia (see Chapter 19). These are morphologically similar to galls produced by eriophyid mites today. The fact that only one gall is present on each leaf suggests additional levels of interactions among the gall producers. A number of galls have been reported on Cenozoic leaves ( Berry, 1916b), and Brooks (1955) described and illustrated numerous crown galls on one Eocene leaf type. Morphologically, they are similar to those produced by gall mites today. Spindle-shaped galls have been reported on a single compressed Quercus leaf from the Miocene of Nevada (Waggoner and Poteet, 1996). Extant galls that are most similar to the fossil Antronoides schorni are produced by certain cynipids (gall wasps) (Erwin and Schick, 2007). Since modern gall production appears to be correlated with certain environmental factors including plant water stress, the presence of galls on leaves may be useful in documenting not only plant–animal interactions but also the relationship between gall formers and their predators (Fernandes and Price, 1992).

CHAPTER 23 Interactions between plants and animals

Figure 23.40 Gall on leaf of Taxodium wallisii (Cretaceous). Bar  0.5 mm. (Courtesy R. Serbet.)

Eocene leaves that contain acarodomatia or mite houses have been described from Australia (O’Dowd et al., 1991). These structures, located in vein axils on the abaxial leaf surface, represent sites where mite eggs were deposited. These mite–plant associations can be related to modern taxa that inhabit rainforests in northeastern Australia. Experimental evidence indicates that plants with domatia have increased levels of predatory mites, but that the leaf domatia function as a defense against herbivores (Walter, 1996).

Other plant–animal interactions Large accumulations of grass androecia have been found in arthropod burrows in the early Miocene of Nebraska (Thomasson, 1982). Based on the behavior of modern arthropods, the fossil burrows are believed to have been formed by harvester ants and carabid beetles. The presence

1021

of husks in the burrows suggests that the lemma and palea, bracts at the base of the flower, were removed below ground in chambers that afforded protection against predation (Thomasson, 1982). A somewhat similar example of plant– animal interaction comes from the Miocene of Germany and occurs in the form of a nut cache assembled by a rodent in coastal dunes of the Lower Rhine Embayment. Based on the arrangement of the nuts, the rodent was probably a hamster; this occurrence represents one of the earliest records of larder keeping in the fossil record (Gee et al., 2003). Another interesting example of plant–animal interactions is an assemblage of more than 100 Mississippian bivalves (Caneyella, ?Ptychopteria) attached to what is interpreted as a brown alga (McRoberts and Stanley, 1989). It is suggested that the alga may have been either planktonic or rooted in a substrate, since no flotation structures were present on the thallus. Although modern algae do not typically serve as hosts for bivalves, perhaps in this instance the bottom conditions were unfavorable for growth and the algae were the only suitable substrate available. Another marine example of a similar interaction involves the Eocene seagrass Thalassodendron (Ivany et al., 1990). Associated with the fossil leaf blades are numerous epibionts and invertebrates. Vertebrates also occur at this locality, thus mirroring the interrelationships that exist in modern seagrass communities. Moissette et al. (2007) reported on a late Pliocene seagrass community from the Isle of Rhodes that included rhizomes and leaves of Posidonia. The community of invertebrates and coralline algae that lived on the seagrasses was also fossilized. This study found a higher diversity of organisms on the leaves (121 species) than on the rhizomes (57 species), and some organisms are comparable to extant species that live exclusively on Posidonia. Mimicry

Plants and animals sometimes interact in relatively subtle ways. One of these is the process of mimicry, in which an organism mimics, in coloration, behavior or habit, another organism or a part of the physical environment. The result is that a predator of prey may be unable to differentiate this organism from the physical environment (Wiens, 1978). One easily detectable form of mimicry in the fossil record involves the general camouflaging that allows the organism to blend into a particular environment. For example, the wings of some insects mimic leaves of plants with which they are associated. An extraordinary leaf mimic specimen was recovered from the famous Messel (Eocene) site in Germany (Wedmann et al., 2007). Eophyllium messelensis is nearly identical in foliaceous morphology to modern male

1022

Paleobotany: the biology and evolution of fossil plants

23.41 Pinnules of Neuropteris (left) and the Carboniferous cockroach Phylomylacris villeti. (From Scott and Taylor, 1983.)

Figure

leaf insects of the order Phasmatodea. The fossil has a leafshaped abdomen and perhaps a behavior that allowed it to remain still on the plant. In this way, it could elude predators that hunted by visual acuity. Using various sources of information, including coprolites, feeding traces, and dentition, Ferguson (2003) has reconstructed a preliminary food web for the Eocene Messel lake site. Several hypotheses have been advanced to explain the origin of wings in arthropods (Dudley, 2000), an evolutionary innovation that may have appeared as early as the Early Devonian (Engel and Grimaldi, 2004). These ideas, however, suggest that initially wings may have been fixed so that they could not be folded back over the abdomen. One of the earliest forms of leaf-wing mimicry is seen in the shapes and venation patterns of a number of Carboniferous foliage types, such as Neuropteris, Linopteris, and Odontopteris, and the analogous morphologies seen in the wings of early insects (FIG. 23.41). It is not difficult to envision some Carboniferous insect resting on the frond of a seed fern, its fixed wings mimicking the two pinnules below its body. Similar associations must have also existed during the Permian, since Glossopteris leaves morphologically mimic some insect wing venation patterns (Cúneo, 1986). Such camouflage would have provided the individual the opportunity to go undetected as its mouthparts penetrated the phloem of the pinna axis. This type of camouflage would also allow predatory insects to wait for prey undetected. Another example of mimicry has been suggested by Fisher (1979), who hypothesized that the prosomal spines of the Carboniferous horseshoe crab Euproops danae are morphologically similar to the leaves of an arborescent lycopsid

Figure 23.42 Suggested reconstruction of Euproops danae

(horseshoe crab) clinging to a vegetative axis of an arborescent lycopod. (From Scott and Taylor, 1983.)

such as Lepidodendron. This may have provided the animal some form of protection if it crawled in among the lycopsid leaves (FIG. 23.42). Pollination

One of the most important types of interaction between plants and animals is the transfer of pollen from the pollen sacs to the stigma in the case of flowering plants, or to the micropylar area in gymnosperms. Modern angiosperms exhibit a variety of complex relationships with their pollinators and some species, for example, are pollinated by a single species of insect. Plants encourage animal foraging behaviors that serve to disperse pollen either passively or actively, for example, with attractants or even mimicry. Although it has generally been postulated that the earliest pollination syndrome within the seed plants (seed ferns, Chapter 14) was anemophily (Taylor and Millay, 1979), there is some indirect evidence to suggest that entomophily was in existence as early as the Carboniferous (Taylor, 1978).

CHAPTER 23 Interactions between plants and animals

Terrestrial arthropods can be traced from the Devonian, with a major explosive radiation taking place during the Pennsylvanian (Rolfe, 1980). Two lines of evidence suggest that at least one of these organisms was perhaps involved in pollination in the medullosan seed ferns. Arthropleura was a millipede-like animal that was the largest terrestrial arthropod during the Carboniferous. It has been suggested that Arthropleura was herbivorous or detritivorous (Rolfe and Ingham, 1967). The animal inhabited the Pennsylvanian Euramerican province and had numerous body segments with complex limbs consisting of many segments (Rolfe and Ingham, 1967). The potential involvement of Arthropleura as a pollinator is based on two lines of evidence. One is the discovery of numerous seed fern pollen grains of the Monoletes type on a leg segment of an animal from the Middle Pennsylvanian (Westphalian D) Francis Creek Shale of northeastern Illinois. Strengthening the possible pollinator role of Arthropleura is the fact that pollen of Monoletes is extremely large (200–550 μm) and obviously not well adapted to wind dispersal. These independent lines of evidence suggest that Arthropleura may have acted as a pollinator in the Euramerican swamps. What is not known is whether such a pollination system evolved passively through casual contact of a foraging herbivore or whether the earliest biotic pollinators were attracted to the plant by some chemical or morphological cue. The Mesozoic bennettitaleans (Chapter 17) are another group of seed plants that may have had insect pollination. The general construction of the cones, which have fibrous interseminal scales with expanded tips between the ovules, is a feature sometimes found in pollination syndromes involving insects with chewing mouthparts. Approximately 22% of the bisporangiate (hermaphroditic) cones of Cycadeoidea examined in one study show extensive tissue disruption that is believed to have been caused by chewing prior to preservation (Crepet, 1974). In many instances, tunnels 1 mm in diameter containing masticated tissue extend through parts of the fleshy receptacle. The areas containing the pollen sacs and young ovules are especially distorted. Crepet hypothesized that the cycadeoids were actively visited by chewing beetles. During such random foraging there was probably some outcrossing achieved as pollen from one plant was carried to the ovules of a different plant. Pollination syndromes involving insects are believed to have played a significant role in the evolution and radiation of the angiosperms (Crepet, 1997a,b; Crepet and Friis, 1987; Grimaldi, 1999), although some data challenge this long-held assumption (Gorelick, 2001; Percy et al., 2004). There are well-preserved examples of angiosperm pollen types as early

1023

as the Aptian (Doyle et al., 1982), with easily recognizable flowers appearing in the Albian (see Chapter 22). Several types of evidence have been examined relative to investigating early angiosperm pollination syndromes in the fossil record (Crepet et al., 1991), including the presence of coprolites containing specific pollen that can be matched to the pollen in a particular fossil flower (Lupia et al., 2002). Another method is to examine the stratigraphic distribution of potential pollinators to infer the advent of particular pollination syndromes. Certain types of flies, beetles, moths, and butterflies are well represented by the Early Cretaceous (Crepet and Friis, 1987; Grimaldi, 1999). The suggestion that many of the earliest angiosperms were pollinated by insects is strengthened by the tectate-columellate organization of the earliest pollen and the morphology of some of the early flower types, such as some of the simple flower types with conspicuous perianths (Crane et al., 1989). During the mid-Cretaceous, angiosperms evolved adaptations to permit the clumping of pollen and, at the same time, there was an increase in specialized pollinators. Hu et al. (2008) examined modern basal angiosperms and found that 86% of the families have zoophilous species, whereas only 17% have wind pollination. Knowledge of pollinators can also potentially provide important information about more subtle aspects of early flowering plants, such as flower color. Since the first angiosperms were presumably pollinated by insects like beetles and flies, that are incapable of resolving color, it has been assumed that the flowers of these plants were white or otherwise dull (Harborne, 1990). Brightly colored floral parts may have evolved later and thereby increased the opportunity for additional relationships with pollinators (Crepet, 1996). Although virtually all modern cycads are insect pollinated (Norstog and Nichols, 1997; Kono and Tobe, 2007), there is to date no persuasive evidence from the fossil record as to when and how insect pollination evolved in this group of plants. The discovery of coprolites inside the pollen sacs of the Middle Triassic cycad pollen cone Delemaya spinulosa (Chapter 17), however, suggests that at least one cycad– insect interaction existed around 235 Ma. Although there is no evidence of pollination, pollinivory may represent a precursory stage in the establishment of a more complex cycad– pollinator relationship (Klavins et al., 2005). Another method of determining pollinator interactions among early flowering plants is by analyzing the structural and morphological features of flowers through geologic time. In Chapter 22 we discussed a number of Mesozoic seed plants that demonstrate various stages of seed enclosure (also see Chapter 15), and offered the two most widely cited theories for the evolution of the carpel: protection against predation and/or the placement of sterile tissue around ovules as

1024

Paleobotany: the biology and evolution of fossil plants

an initial stage in the evolution of an incompatibility system. The evolution of floral structures was of profound importance to certain groups of flowering plants because it provided a method of attracting faithful pollinators. This method was interpreted as being energetically more efficient, whereas at the same time it promoted outcrossing and therefore increased genetic variability. Documenting certain floral patterns in time can thus be valuable in inferring pollinator interactions; these include characteristics of the perianth, position of the floral parts, floral symmetry, degree of fusion of parts, and number of parts (Friis and Crepet, 1987; Crepet, 1996). Studies of extant pollinators, however, indicate that flowers attract many visitors, and that flowers may diverge without excluding one type of pollinator over another (Fenster et al., 2004). It is argued that functional groups of similar pollinators (e.g., long-tongued flies), as opposed to pollinator species, provide the selective pressure on the evolution of floral traits. It is hypothesized that entomophily evolved as a direct consequence of insect foraging (FIG. 23.43), either for pollen or nectar. Many extant angiosperm species produce flowers that offer a surplus of pollen as a reward to insect visitors, instead of nectar or similar substances (Vogel, 1978). The relative amounts of pollen produced per flower and the occurrence of nectaries are aspects of entomophily that can be documented with fossil evidence. Heteranthy is another type of plant– pollinator interaction in which both functional stamens and structures that mimic anthers but do not contain pollen are produced. In this system, pollinators are attracted and pollen transfer effected, but less pollen is provided as a reward. Information about the diversity of fossil flowers is still too incomplete to recognize the existence of heteranthy in fossils. Structures such as nectaries and certain types of glands (Krassilov and Golovneva, 2001) have been reported as early as the late Paleozoic and, in several instances, the presence of identical structures on different plant organs has been used to indicate biological affinities. The capitate glands used by Oliver and Scott (1904) in the establishment of the Lyginopteridales (Chapter 14) represent structures that may have been attractants for potential pollinators, but may also have served as defense structures against herbivores (discussed above). Mamay (1976a) commented on the glandular bodies, that is, the remains of secretory cavities, between the seeds in the fertile leaves of the Paleozoic cycad Phasmatocycas as potentially functioning in an entomophilous pollination syndrome. It is important to understand that ideas about pollination syndromes in extant plants may also change. At one time it was assumed that modern cycads were wind pollinated, but now all cycads are thought to be dependant on mutualisms with specialized insect pollinators, especially beetles (Terry et al., 2005), and in one instance, thrips (Norstog and Nichols,

Figure 23.43 Biotic interaction of Homaloneura lehmani feeding on a cordaitean reproductive organ. (From Shear and Kukalová-Peck, 1990.)

1997; Terry et al., 2007). Other studies with Cycas revoluta suggest that both wind and insect pollination may be involved. These studies indicate that caution needs to be used in analyzing certain biological interactions, including pollination, not only because of the patchy fossil record for these interactions but also because of changing biotic and abiotic influences during the millions of years of plant–animal associations.

Conclusions As paleobotanists have learned more about various groups of fossil plants, increasing amounts of data from other disciplines have been incorporated into our reconstructions of ancient plants within their environments, including information on the relationships among organisms that existed at a particular point in time and space. This paleoecological approach has provided a wealth of information about the structure of various plant communities and how they have changed through time. Some of the most important, yet least understood, components of these paleoecological studies are the interactions that existed between plants and animals (FIG. 23.44), and their cumulative effect on the ecosystem and the composition of the community in which they lived. The patterns and mechanisms used

CHAPTER 23 Interactions between plants and animals

Figure 23.44 herbivory.

Decorative lapel pin showing artistic effect of

by various species to adapt to their surroundings necessitate an understanding of all biotic and abiotic interactions that existed in an ecosystem. Some of these, such as herbivory (FIG. 23.44), often had a dramatic influence on a population and can be documented in the fossil record. Others, such as pollination and dispersal syndromes, can be more difficult to evaluate in fossils. Detailed knowledge of fossil plants can also be used to ask questions about plant–animal interactions within certain groups and even organs. For example, there are no reports of sporangial predation in lycopsids, sphenophytes, or conifers in the Carboniferous, despite the rich fossil record and the fact that these would seem to be plentiful sources of food for herbivores. Does this reflect synchronized and episodic production, spatial distribution of the cones, mechanical barriers to predation, or some other factors? Although there are numerous studies that focus on predation of angiosperm fruits and seeds, there is little information available about extant conifer pollen and seed cones in plant–animal interactions. The information that is available focuses principally on the entomofauna in the Pinaceae (Turgeon et al., 1994). An interesting study might be

1025

a quantitative analysis of the many permineralized pinaceous seed cones described to date for cone and seed predation. We have discussed a small number of examples of plant– animal interactions that might be determined from the fossil record. It is especially important, however, not to assume that fossil evidence of an interaction found in a single group of plants or from a single specimen can necessarily be used to infer broad generalizations about first occurrence in the fossil record or phylogenetic patterns that focus on diversity (Labandeira et al., 2007b). As noted above, interactions with certain plants have provided evidence of a fossil record for some insects earlier than the currently known body fossil record (Hasiotis and Dubiel, 1995; Hasiotis, 2003). In contrast, several molecular phylogenetic studies of congruence between insects and their plant hosts have found that host-specific insect taxa radiated after their host plants (Percy et al., 2004; LopezVaamonde et al., 2006). Countless questions can be asked and hypotheses advanced regarding these complex associations. In the future, a combination of paleobotanical, paleontological, ichnological, and phylogenetic data will no doubt be applied to resolve some of the interesting questions posed by such research. Evolutionary ecology and genetics, reproductive biology, and an ever-expanding database of fossils will provide new dimensions to increase our understanding of organism interactions in the fossil record. For many questions about plant–animal interactions, the answers are not yet available from studies on modern ecosystems; nevertheless, to increase our hypotheses that rely on the fossil record can at least provide data on these occurrences in past communities. The interactions between plants and animals represent one of many components that have influenced the distribution and abundance of various plant species throughout geologic time. This is a relatively new approach in paleobotany that has come about because of a more focused appreciation of the interplay between the environment and genome during the evolution of the world’s floras, both past and present. Perhaps Charles Darwin, with his extraordinary grasp of biological interactions, best described the scope of paleobotany today when he wrote in 1872: It is interesting to contemplate a tangled bank, clothed with many plants of many kinds, with birds singing on the bushes, with various insects flitting about, and with worms crawling through the damp earth, and to reflect that these elaborately constructed forms, so different from each other, and dependent upon each other in so complex a manner, have all been produced by laws acting around us…from so simple a beginning endless forms most beautiful and most wonderful have been, and are being evolved.

This page intentionally left blank

appendix

1

CLASSIFICATION OF ORGANISMS c. Class Xanthophyceae (yellow-green algae) Order: Vaucheriales d. Class Phaeophyceae (brown algae) Orders: Ectocarpales, Laminariales, Fucales 4. Phylum Prymnesiophyta (haptophytes) Order: Coccolithophorales 5. Phylum Rhodophyta (red algae) a. Class Rhodophyceae Orders: Bangiales, Corallinales, Nemaliales, Ceramiales, Gigartinales *Solenoporaceans (artificial group) 6. *Acritarcha

Those groups marked with an asterisk (*) are extinct. I. DOMAIN BACTERIA (Woese et al., 1990; Eubacteria in Woese, 1987) A. Kingdom Eubacteria 1. Cyanobacteria Orders: Chroococcales, Oscillatoriales, Pleurocapsales B. Other bacteria (purple bacteria, Gram-positive bacteria, green nonsulfur bacteria, sulfate-reducing bacteria, and others) II. DOMAIN ARCHAEA, methanogens Archaea III. DOMAIN Eukaryota A. Kingdom Fungi 1. Phylum Chytridiomycota (Chytridiomycetes) 2. Phylum Zygomycota a. Subclass Zygomycetes b. Subclass Trichomycetes 3. Phylum Glomeromycota 4. Phylum Ascomycota a. Subphylum Pezizomycotina 5. Phylum Basidiomycota a. Subphylum Agaricomycotina 6. Fungal-like organisms: Peronosporomycetes (Oomycota) B. Algae (artificial group) 1. Phylum Euglenophyta (euglenoids) 2. Phylum Dinophyta (dinoflagellates) 3. Phylum Heterokontophyta (heterokonts) a. Class Bacillariophyceae b. Class Dictyochophyceae Order: Dictyochales

C. Kingdom Plantae (subkingdom Viridaeplantae) 1. Phylum Chlorophyta (green algae) a. Class Prasinophyceae Order: Pyramimonadales b. Class Charophyceae Orders: Charales, *Moellerinales, *Sycidiales, *Trochiliscaceae, *Chovanellaceae, *Pinnoputamenaceae, Zygnematales, Coleochaetales c. Class Chlorophyceae Orders: Chlorococcales, Tetrasporales, Chaetophorales, Volvocales d. Class Ulvophyceae Orders: Cladophorales, Dasycladales, *Cyclocrinales, Caulerpales, Ulvales LAND PLANTS 1. *Nematophytes (artificial group) incertae sedis 2. Phylum Anthocerotophyta (hornworts) 3. Phylum Bryophyta

1027

1028

Appendix 1 classification of organisms

4. 5. 6. 7.

8.

9.

Subphylum Marchantiophytina (liverworts) a. Classes: Treubiopsida, Marchantiopsida (thalloid liverworts), Jungermanniopsida (leafy liverworts) Subphylum Bryophytina (mosses) a. Classes: Sphagnopsida, Takakiopsida, Bryopsida *Phylum Rhyniophyta *Phylum Zosterophyllophyta *Phylum Trimerophytophyta Phylum Lycophyta (lycophytes) a. *Order Drepanophycales b. *Order Protolepidodendrales c. *Order Lepidodendrales Families: Lepidodendrales, Diaphorodendraceae, Sigillariaceae d. Order Lycopodiales Family: Lycopodiaceae e. Order Selaginellales Family: Selaginellaceae f. *Order Pleuromeiales g. Order Isoetales Families: *Chaloneriaceae, Isoetaceae Phylum Sphenophyta a. *Order Pseudoborniales b. *Order Sphenophyllales c. Order Equisetales Families: *Calamitaceae, *Tchernoviaceae, *Gondwanostachyaceae, Equisetaceae Phylum Pteridophyta a. *Class Cladoxylopsida 1) Order Pseudosporochnales 2) Order Iridopteridales b. *Early Fernlike Plants (not a formal group) 1) Order Rhacophytales 2) Order Stauropteridales Family Stauropteridaceae 3) Order Zygopteridales Family Zygopteridaceae c. Eusporangiate Ferns (not a formal group) 1) Order Marattiales Families: *Psaroniaceae, Marattiaceae 2) Order Ophioglossales d. Leptosporangiate Ferns (not a formal group) 1) Order Osmundales Families: *Guaireaceae, Osmundaceae 2) Order Filicales (ordinal name not currently accepted)

10.

11.

12.

13.

14.

Families: *Botryopteridaceae, *Anachoropteridaceae, *Kaplanopteridaceae, *Psalixochlaenaceae, *Sermayaceae, *Tedeleaceae, *Skaaripteridaceae, *Tempskyaceae, Schizaeaceae, Hymenophyllaceae, Gleicheniaceae, Dicksoniaceae, Cyatheaceae, Matoniaceae, Loxsomataceae, Dipteridaceae 3) Order Polypodiales Basal polypod families: Dennstaedtiaceae, Pteridaceae. Eupolypod families: Onocleaceae, Blechnaceae, Polypodiaceae 4) Order Salviniales (heterosporous ferns) Families: Marsileaceae, Salviniaceae *Phylum Progymnospermophyta (progymnosperms) a. Order Archaeopteridales b. Order Aneurophytales c. Order Protopityales d. Noeggerathians *Phylum Pteridospermophyta a. Order Calamopityales b. Order Buteoxylonales c. Order Lyginopteridales d. Order Medullosales e. Order Callistophytales f. Order Glossopteridales g. Order Peltaspermales h. Order Caytoniales i. Order Corystospermales j. Order Petriellales Phylum Cycadophyta a. Order Cycadales b. *Order Bennettitales Families: Cycadeoidaceae, Williamsoniaceae Phylum Ginkgophyta a. Order Ginkgoales Families: Ginkgoaceae, *Karkeniaceae, *Umaltolepidiaceae, *Yimaiaceae, *Schmeissneriaceae Other Gymnosperms a. *Order Gigantopteridales b. *Order Vojnovskyales c. *Order Czekanowskiales

Appendix 1 classification of organisms

d. *Order Iraniales e. *Order Pentoxylales f. *Order Hermanophytales g. Order Gnetales h. *Family Dirhopalostachyaceae 15. Coniferophytes a. *Order Cordaitales b. *Order Voltziales Families: Utrechtiaceae, Thucydiaceae, Emporiaceae, Majonicaceae, Ullmanniaceae, Bartheliaceae, Ferugliocladaceae, Buriadiaceae c. Order Coniferales Families: Palissyaceae, *Cheirolepidiaceae, Podocarpaceae, Araucariaceae, Cupressaceae, Sciadopityaceae, *Pararaucariaceae, Pinaceae, Cephalotaxaceae, Taxaceae 16. Angiosperms a. Basal angiosperms Families: Amborellaceae, Hydatellaceae, *Archaefructaceae, Chloranthaceae 1) Order Nymphaeales Family: Nymphaeaceae 2) Order Austrobaileyales Families: Austrobaileyaceae, Iliciaceae, Schisandraceae 3) Order Ceratophyllales Family: Ceratophyllaceae b. Magnoliids 1) Order Canellales Family: Winteraceae 2) Order Laurales Families: Calycanthaceae, Lauraceae 3) Order Magnoliales Families: Annonaceae, Magnoliaceae, Myristicaceae 4) Order Piperales Families: Lactoridaceae, Saururaceae c. Monocotyledons 1) Order Alismatales Families: Alismataceae, Araceae, Hydrocharitaceae, Zosteraceae 2) Order Asparagales Families: Agapanthaceae, Hemerocallidaceae, Orchidaceae 3) Order Dioscoreales Family: Dioscoreaceae

d.

e.

f.

g.

h.

1029

4) Order Liliales Family: Petermanniaceae 5) Order Pandanales Families: Pandanaceae, Triuridaceae Commelinids 1) Order Arecales Family: Arecaceae (Palmae) 2) Order Commelinales Family: Commelinaceae 3) Order Poales Families: Cyperaceae, Poaceae (Gramineae) 4) Order Zingiberales Families: Musaceae, Zingiberaceae Eudicots Families: Buxaceae, Trochodendraceae 1) Order Proteales Families: Nelumbonaceae, Proteaceae, Platanaceae 2) Order Ranunculales Families: Berberidaceae, Ranunculaceae Core eudicots 1) Order Gunnerales Family: Gunneraceae 2) Order Caryophyllales Family: Phytolaccaceae 3) Order Saxifragales Families: Cerciphyllaceae, Haloragaceae, Hamamelidaceae, Iteaceae, Saxifragaceae Rosids Family: Vitaceae 1) Order Myrtales Families: Lythraceae, Trapaceae, Myrtaceae, Onagraceae Eurosids I (Fabids) 1) Order Fabales Family: Fabaceae (Leguminosae) 2) Order Fagales Families: Betulaceae, Casuarinaceae, Fagaceae, Juglandaceae, Myricaceae, Nothofagaceae 3) Order Malpighiales Families: Clusiaceae, Euphorbiaceae, Salicaceae, Malpighiaceae 4) Order Oxalidales Families: Cunoniaceae, Elaeocarpaceae

1030

Appendix 1 classification of organisms

5) Order Rosales Families: Moraceae, Rhamnaceae, Rosaceae, Ulmaceae i. Eurosids II (Malvids) 1) Order Brassicales Families: Capparaceae 2) Order Malvales Families: Tiliaceae, Sapindales, Anacardiaceae, Meliaceae, Rutaceae, Sapindaceae j. Asterids 1) Order Cornales Families: Cornaceae, Curtisiaceae, Hydrangeaceae, Ericales, Ebenaceae, Ericaceae, Theaceae k. Euasterids I (Lamids) Family: Icacinaceae 1) Order Garryales Family: Eucommiaceae

2) Order Gentianales Families: Gentianaceae, Rubiaceae 3) Order Lamiales Families: Avicenniaceae, Byblidaceae, Lentibulariaceae, Oleaceae 4) Order Solanales Family: Solanaceae 1. Euasterids II (Campanulids) Families: Bruniaceae, Quintiniaceae 1) Order Apiales Family: Araliaceae 2) Order Aquifoliales Family: Aquifoliaceae 3) Order Asterales Families: Asteraceae (Compositae), Menyanthaceae 4) Order Dipsacales Family: Caprifoliaceae

GLOSSARY abaxial The lower surface; on the side away from the axis of the plant. abracteate cone A sphenophyte cone exclusively composed of sporangiophore whorls (lacking bracts). abscission zone The region at the base of a plant part (e.g., leaf, flower, fruit) that is involved in the loss (shedding) of that plant part. absolute dating Quantitative dating of rocks utilizing decay of radioactive elements (see radiometric dating). acanthomorphic acritarch An acritarch with a clear distinction between the central body and radially oriented processes attached to a central body accessory pigments Light-absorbing compounds, found in photosynthetic organisms, that work in conjunction with chlorophyll a. achene A simple, dry, one-seeded, and indehiscent angiosperm fruit in which the seed coat is not fused to the fruit wall (pericarp). acritarch A generalized term for a unicellular, morphologically variable microfossil whose affinities cannot be determined; many acritarchs represent cysts of planktonic algae. acrodromous A leaf venation pattern in which two or more primary veins run from the base of the leaf to the apex in convergent arches. acrolamella In certain megaspores, a leaflike extension of the laesura. acroscopic Pointing toward the tip or apex (e.g., of a pinna or a frond). actinocytic stomata Stomatal arrangement in which the guard cells are surrounded by radially arranged subsidiary cells. actinodromous A leaf venation pattern in which three or more primary veins extend radially from a central point. actinomorphic flower A flower with radial symmetry; with three or more planes of symmetry (compare regular flower) (contrast zygomorphic flower). actinostele A protostele that appears star shaped or fluted in cross section. acuminate Tapering to a slender point. adaptation A structure or feature that performs a particular function and which results in increased survival or reproduction. adaxial Upper surface; the side toward the axis; facing the stem. adnate Fusion of unlike parts, for example, stamens adnate to petals (contrast connate). adventitious A structure arising from an unusual or abnormal position, for example, roots growing from leaves. aerenchyma Tissue with many air spaces, as in the rootlets of Psaronius. aerial Living above the surface of the ground or water (as in aerial roots).

aerobes Organisms that require free oxygen (O2). aerobic Requiring free oxygen (O2). akinete Thick-walled, nonmotile spore in cyanobacteria alation An irregular flap or winglike extension of tissue. alete A spore having no haptotypic mark or laesura. alginite An organic component of coal that consists of algal matter. alternate Arrangement of leaves or other appendages in which only one appendage is produced at each node. alternation of generations A type of reproductive cycle in which a haploid or gametophyte (n) phase alternates with a diploid or sporophyte (2n) phase. alveolate A type of sporoderm infrastructure with many small spaces; honeycombed. amb The outline or shape of a spore (or pollen grain) as viewed in the proximo-distal plane, that is, from one of the poles. amber Solidified fossil plant resin, usually yellow-brown in color. amentiferous With the flowers in catkins or aments. amino acid racemization A method of dating based on the natural racemization of amino acids (the compounds that make up proteins) over time. amphicribal A vascular bundle in which the phloem surrounds the xylem; sometimes listed as amphiphloic. Compare amphivasal. amphiphloic Having phloem on both sides of the xylem (in a stele or a single vascular bundle). amphistomatic Bearing stomata on both sides of a leaf. amphivasal A vascular bundle in which the xylem surrounds the phloem. Compare amphicribal. anaerobes Organisms that can live in environments lacking free oxygen. anaerobic Occurring in the absence of free oxygen (O2). analogous Structures with a similar function that do not have a common evolutionary history (i.e., structures resulting from convergent evolution) (contrast homologous). anamorph The asexual, conidial, or the so-called imperfect state of a fungus in which the spores are produced by mitosis. anastomose To fuse together. anatropous Attachment of a seed or ovule so that it is reflexed and the micropyle is directed toward the point of attachment (contrast orthotropous). ancestral A preexisting condition or character state. androecium Collectively all of the stamens in a flower. anemophilous Wind pollinated.

1031

1032

glossary

anisocytic stomata Three subsidiary cells of unequal size surrounding guard cells. anisophyllous Having at least two types of leaves. annual A plant living one year or less. annual ring The accumulation of secondary xylem (or phloem) over a single growing season; in temperate trees evident because of the demarcation between the last cells of the summer (latewood) and the first cells of the subsequent spring (earlywood). annular In rings; a type of secondary wall pattern in tracheids. annulate sporangium A sporangium with an annulus. annulus (pl. annuli) A row of specialized cells in some fern sporangia that is involved in the opening of the sporangium. anomocytic stomata A stomatal pattern in which there are no obvious subsidiary cells; guard cells are apparently surrounded by normal epidermal cells. anoxygenic photosynthesis Photosynthesis that does not produce oxygen anther The upper, pollen-bearing part of a stamen; the pollen sacs. antheridiophore A stalk that bears antheridia, for example, in the liverworts. antheridium (pl. antheridia) A unicellular (in most fungi and algae) or multicellular (in land plants) reproductive structure that produces sperm (male gametes); male gametangium. anthracite A hard black, metamorphic coal with a high percentage of fixed carbon (92–98%) and a low percentage of volatile matter. anticlinal Perpendicular to the surface (contrast periclinal). antithetic (interpolation) theory Theory for the origin of the alternation of generations in which the gametophyte is primitive and the sporophyte evolved through a delay in meiosis of the diploid zygote (contrast homologous theory). aperture A site through which the contents of a spore or pollen grain exit (e.g., laesura, pore, colpus). aphlebia (pl. aphlebiae) Anomalous pinnae on the rachis of some ferns. apical cell A single meristematic cell initial in the apical meristem of an axis that divides to form new tissues for the organ; occurs in vascular cryptogams and mosses (see also segment cells and sextant cells). apical prominence see gula. apiculate Ending in a sharp, flexible point. apocarpous Having free (i.e., unfused) carpels. apomixis Reproduction without meiosis, that is, without fertilization; a type of asexual reproduction. Also called apogamy apomorphy In cladistics, a derived feature that provides information on evolutionary relationships (contrast plesiomorphy). apophysis A natural swelling or enlargement, for example, at the base of the capsule in certain mosses or on the cone scale of certain conifers. apothecium (pl. apothecia) In fungi, an open ascocarp. apotracheal parenchyma In certain types of angiosperm wood, parenchyma that is arranged independently of the vessels (contrast paratracheal parenchyma). apoxogenesis A type of determinate growth that results in a continual decrease in the diameter of the primary xylem in distal parts of the plant, for example, in the Lepidodendrales (contrast epidogenesis). araucarioid pitting A type of tracheid pitting in which the aperture is narrow and included in the border, pits arranged in alternate rows of three or more, and appear polygonal in outline.

arborescent Treelike. arbuscular mycorrhizae Endomycorrhizal fungi forming arbuscules within host plant cells. arbuscule Hyphae that are shrubby and highly branched that occur inside the host root and rhizome cells of endomycorrhizal fungi associations. archegoniophore A stalk that bears archegonia; found in some liverworts and early land plants. archegonium (pl. archegonia) Multicellular structure in land plants that produces an egg; consists of a swollen venter that surrounds the egg and a narrow neck. areole The smallest area of leaf tissue surrounded by veins. aril A fleshy, often colored, outer envelope in certain types of seeds. arillate Having an aril. aroid Referring to members of the family Araceae. artifact In biology, a structure or material not normally present, not naturally occurring, for example, produced by preparation techniques. ascidial Pitcher or flask shaped. ascidiate A carpel that is not leaflike, but rather develops from a ring of tissue and assumes a flask shape ascocarp A fungal fruiting body containing asci. ascospore A spore produced within an ascus. ascostroma (pl. ascostromata) A fungal fruiting body of the Ascomycotina that consists of a stroma that bears asci. ascus (pl. asci) The saclike reproductive structure of the Ascomycotina. aseptate See nonseptate. assemblage zone In biostratigraphy, a biozone that is characterized by a particular group of organisms rather than a single organism. atactostele A type of eustele with scattered bundles; stem stele in monocotyledonous angiosperms. authigenic cementation A type of preservation that involves soft sediment cementation by iron and carbonate compounds; results in replicas of surface features including molds and casts (e.g., ironstone concretions from Mazon Creek area of Illinois). autotroph A form of metabolism in which the organism synthesizes its own food from inorganic compounds; photosynthetic and chemosynthetic organisms are. axil The upper angle between a stem and leaf or other appendage. axile placentation The attachment of ovules within a fruit to a central axis. axillary position Occurring within the axil, for example, the sporangial position of some early lycophytes. azonate Lacking an equatorial extension or zone that encircles the equator of a spore or pollen grain (contrast zonate). baculae Small rods. baculate A type of ornamentation in pollen and spores consisting of small rods. bark The complement of tissues external to the vascular cambium. basionym In nomenclature, the name under which the taxon was first described. basipetal Toward the base. basiscopic Pointing toward the base (e.g., of a pinna or a frond). benthic Living on or attached to the bottom in aquatic habitats (contrast pelagic, planktonic).

glossary

berry A type of simple, fleshy fruit that includes more than one seed embedded in a fleshy fruit wall; many have one or more carpels. bicollateral Having phloem on both the outside and inside of a vascular bundle (contrast collateral). bifacial Having two faces, for example, a cambium that produces cells on both sides (contrast unifacial cambium). bifid Divided into two parts. bifurcate To divide into two halves; dichotomize. bilateral symmetry Capable of being divided into two equal halves by only one longitudinal plane (contrast radial symmetry; compare zygomorphic flower). biodictyon A type of biofilm characterized by a network of filamentous organisms that become embedded in their substrate(s). biofilm A community of microorganisms embedded in extracellular biopolymers that form a thin film. bioherm A mound- or reeflike mass of rock, built up by organisms (e.g., coralline algae) and composed primarily of their calcareous remains. biomarker A chemical signature of an organism in the sedimentary record; a product derived from biochemical precursors that provides evidence of the past presence of an organism or group of organisms. biostratigraphy Relative dating and/or correlation of rocks by the type of organisms (plants, animals, and/or microfossils) that they contain. biotroph An organism that can live and multiply only on another organism; a condition in which only living cells are used as a nutrient source. biozonation See biostratigraphy. bipartite Split into two almost to the base; as in fronds in which the petiole (or stipe) forks and forms a two-parted blade with two rachides. bisaccate A pollen grain with two sacci. biseriate In two ranks or rows. biseriate frond Pinnae borne in two ranks or rows and flattened in the same plane. bisporangiate Bearing two types of sporangia, that is, microsporangia and megasporangia, in the same fructification. bitegmic An ovule having two integuments or seed coats. bordered pit A structure in the wall of tracheids and vessels consisting of a central thin area surrounded by a thickened border. bract A modified leaflike structure subtending a lateral branch, flower, cone, sporangiophores, etc. bracteate cone A sphenophyte cone in which whorls of sterile bracts and whorls of sporangiophores occur (contrast abracteate cone). bracteole A small bract; a scale. brochidodromous Venation pattern in which the pinnate secondary veins do not terminate at the margins of the leaf, but are joined in a series of arches. bulb A short, thick, underground stem surrounded by scalelike leaves; the leaves may be fleshy or membranous. bulbil A deciduous bud usually formed on the aerial part of a plant in the axils of leaves bundle cap An outer cluster of sclerenchymatous fibers associated with a vascular bundle. caducous Deciduous, falling from the plant at some time; not persistent.

1033

calcification A type of fossilization in which calcium carbonate (CaCO3) either infills intercellular spaces and other voids (permineralization) or replaces the original organic material (petrifaction) of an organism. calyptra A structure or tissue that covers something as in a root cap. calyx The outermost series of modified leaves in the perianth; sepals. cambium A lateral meristem, responsible for growth in the diameter of vascular plants. camptodromous Venation pattern in which the secondary veins are pinnate and do not terminate at the margin of the leaf; brochidodromous venation is a type of camptodromous pattern. capitulum See head. cappa Thickened portion of the proximal surface of the corpus of a saccate pollen grain. capsule A moss sporangium; a type of dry, many-seeded, dehiscent angiosperm fruit that forms from a compound ovary (more than one carpel); the sporangium of a ferm without its stalk. carinal canal A canal in the xylem of certain sphenophytes (e.g., Equisetum) that results from extension and rupture of protoxylem elements; sometimes termed protoxylem canal. carpel A modified leaf in the center of a flower that encloses one or more ovules; a carpel consists of an ovary, style, and stigma; all the carpels together form the gynoecium or pistil. carpellate A flower that only has carpels and no stamens; female flowers only; pistillate (contrast staminate). Casparian strip A secondary thickening found on the anticlinal (radial and transverse) walls of endodermal cells that consists of suberin and lignin, and serves to force water and mineral solutes to pass through the plasma membrane of these cells. cast A preservation type that forms within a mold; casts are usually three dimensional. cataphyll A reduced, scalelike leaf (common in members of the Cycadales). catkin A lax, spikelike inflorescence consisting of apetalous (naked), unisexual (i.e., either carpellate or staminate) flowers. centrarch A type of primary xylem maturation in which the protoxylem is central and surrounded by metaxylem. chalaza The proximal or basal region of an ovule. checking Small separations in the tracheid wall that occur in a spiral pattern as they follow the helical microfibril orientation of the secondary wall. chemosystematics The study of plant classification based on chemical data. chert A hard, brittle sedimentary rock consisting of microcrystalline quartz. chitinozoan A vase-shaped microfossil of an extinct marine group (Cambrian–Devonian) with uncertain affinities. chlamydospore A thick-walled, asexual fungal resting spore. chloranthoid teeth Leaf margin type in which a medial vein thickens below a small triangular gland that makes up the tooth; typical of members of the Chloranthaceae. chlorenchyma Parenchymatous tissue that contains chloroplasts. chromatin The darkly staining material that makes up eukaryotic chromosomes consisting of protein and DNA. cingulum (pl. cingula) A thick flange on certain spores, usually in an equatorial position. circinate Coiled.

1034

glossary

circinate vernation The coiled arrangement of immature leaves or leaflets, for example in ferms and cycads. clade A monophyletic group or lineage. cladistic A type of methodology for inferring the evolutionary history of a group of organisms by grouping taxa based on shared, derived characters (apomorphies). clamp connection A specialized connection between two adjacent hyphal cells in a member of the Basidiomycotina. clavate Club shaped. cleistothecium (pl. cleistothecia) An ascocarp in which the ascospores are completely enclosed. clepsydroid Dumbbell or hourglass shaped. coal ball A type of permineralization known from Carboniferous and Permian coal measures in which the plants are preserved by calcium carbonate and other minerals (e.g., pyrite). coccoid Spherical, used to describe the shape of certain bacterial cells. coccolith A calcified (CaCO3) scale of a coccolithophore. coccolithophore A small, marine planktonic alga, with affinities to the Prymnesiophyta; has a complex wall that includes small, calcareous plates (coccoliths). coenobium (pl. coenobia) A colony of unicellular organisms, for example, green algae, that are surrounded by a common membrane. coenocytic A thallus in which the nuclei are contained in a common cytoplasm without being separated by septa or cross walls. collateral Having phloem only on the outside of a vascular bundle (contrast bicollateral). collenchyma A cell with unevenly thickened primary walls; important in support in the primary growth phase. colpate A pollen grain with one or more elongate furrows (colpi). colpus (pl. colpi) A longitudinal furrow in some pollen grains; when used strictly, colpi cross the equator and are confined to angiosperms; sometimes used in synonymy with sulcus. columella (pl. columellae) A central, sterile axis in bryophyte sporangia; columns that support the tectum in a tectate pollen grain. columellate Containing a columella. composite teeth Multiple serrations per secondary vein along the margin of a leaf, for example, in Nothofagus. compound Made up of a number of parts, for example, a compound ovary consists of two or more ovaries. compression A preservation type in which the organic matter consists of a thin film of carbon. conceptacle An open cavity that bears reproductive structures (e.g., in some algae). concretion A nodule that forms in an accretionary manner around a nucleus or center; concretions generally differ chemically from the rocks in which they are found; many are known to contain plant and animal fossils, such as the ironstone nodules from Mazon Creek, Illinois. conduplicate Folded together lengthwise, as in conduplicate carpels. cone dome In cycads with terminal cones, an adventitious meristem initiates growth and as a result forms a sequence of domelike profiles of vascular tissue. conidium (pl. conidia) In fungi, a nonmotile, asexual spore usually formed at the tip or side of a sporogenous cell. connate Fusion of like parts, such as petals fused to each other (contrast adnate). conus (pl. coni) A type of sculpture in pollen and spores consisting of small, cone-shaped projections.

convergent evolution Evolution that results in similar structures in organisms that do not share a common evolutionary history. coprolite Fossil fecal material. coriaceous Leathery; with a thick cuticular layer. cork A type of secondary tissue produced by a cork cambium; cork cells are nonliving at maturity and have waxy or fatty deposits in their walls. cork cambium A type of lateral meristem that gives rise to cork (phellem) externally and sometimes phelloderm internally. corm An upright, thickened, underground stem that functions in food storage. cormose Shaped like a corm, for example, the cormose lycopsids. corolla All the petals in a flower; often colored; the floral whorl inside the calyx. coronula In charophyte oogonia, one or two tiers of small cells resting on the apical ends of the enveloping cells to form a ring around the summit. corpus The central body of a saccate pollen grain or pseudosaccate spore. corymb An inflorescence type that is indeterminate, consisting of a central stalk that bears pedicelled flowers along its sides; the marginal flowers bloom first and the top of the corymb is flattened or slightly rounded. costa (pl. costae) A central rib or ridge. costapalmate leaves Fan-shaped leaves with a petiole that extends through the leaf as a midrib; found in certain palms. craspedodromous A leaf venation pattern in which the pinnate secondary veins terminate at the leaf margin. crassula (pl. crassulae) A thickened portion of a tracheid wall above and below a pit pair; found in some gymnosperms; also called bars of Sanio. crenate A leaf margin type characterized by smoothly rounded teeth. crenulations Minute notches. cross-field pits In secondary xylem, the pits between axial tracheids and ray parenchyma; they are often different from typical tracheid pits and may be diagnostic for the particular taxon. crotonoid Pollen sculpture type consisting of triangular–rectangular shaped units that are fused together to form a reticulum (e.g., Stellatopollis). crozier An immature, coiled fern frond (also crosier), a fiddlehead cruciate Four-armed, having the form of a cross. cryptogam A seedless plant. cryptopore The pore covered by the tectum on the distal face of a Classopollis pollen grain. cryptospore A spore type for which the source plant is not known; mostly used for spores older than the earliest land plant megafossils. cucullate Shaped like a hood. culm The jointed, usually hollow stem found in certain grasses (Poaceae) and sedges (Cyperaceae). cupressoid pitting A type of cross-field pitting in which the elliptical aperture is included within the border and is narrower than the border. cupule An accessory set of structures that surrounds one or more ovules or seeds; the cupule lobes may be free or united. curvatura (pl. curvaturae) A line present in some trilete spores that connects the ends of the laesurae; curvaturae perfectae join to connect the three arms of the trilete together; curvaturae imperfectae are just short, forked extensions at the end of each trilete arm.

glossary

cuticle The amorphous, waxy layer that covers and impregnates the walls of the epidermal cells on the aerial parts of plants; made of cutin and waxes. cutin The waxy, impermeable substance that makes up the cuticle. cygneous Curved downward like a swan’s neck. cyme A type of determinate, flat-topped inflorescence with few flowers; the central (terminal) flowers open slightly before the outer ones. cymule A small cyme. cypsela An achene that is attached to the surrounding floral tube; that is, it is formed from an inferior ovary; found in the Asteraceae. decortication The stripping away of bark, outer cortex, or periderm. decurrent Extending downward below the point of insertion, as certain leaves extend along the stem like a wing. decussate Appendages attached in pairs alternately and at right angles to one another; the result is four ranks or rows of appendages. dehiscent fruit A fruit that naturally opens at maturity to release seeds. dentate Having teeth or with a toothed margin. desmocytic stomata A single cell enclosing the guard cells that has one anticlinal wall extending to divide the cell once. determinate A type of growth and development that has a definite limit; typical of flowers, most inflorescences and leaves, and the vegetative organs of the Lepidodendrales (contrast indeterminate). dextrorse Turned to the right; clockwise. diagenesis The physical and chemical changes that occur to sedimentary deposits after deposition. diarch Having two protoxylem poles. diacytic A stomatal type in which there are two large subsidiary cells that completely surround the guard cells and are aligned perpendicular to them. diatomaceous earth A sedimentary deposit made up of siliceous diatom frustules. diatomite See diatomaceous earth. dichasium (pl. dichasia) A type of determinate inflorescence in which the main axis terminates in a flower and gives off two lateral axes that each terminate in a flower and give off two lateral axes etc. dichotomize To divide into two equal parts; bifurcate. dichotomous Divided into two equal parts. dictyostele A dissected amphiphloic siphonostele. dictyoxylon A type of cortical organization characterized by a netlike system of hypodermal fiber strands (contrast sparganum); sometimes Dictyoxylon. dictyoxylic siphonostele A stele in which large overlapping leaf gaps dissect the vascular system into strands, each with phloem surrounding the xylem. dicyclic Having two whorls (as a dicyclic perianth); having two rows of cells (as a dicyclic stomatal arrangement). diffuse porous A type of secondary xylem organization in which the vessels are scattered throughout the growth of one season or there is little difference in the size of the vessels throughout the season (contrast ring porous) digitate With parts that diverge from a single point; palmate. dinocyst A contraction of dinoflagellate cyst, a thick-walled resting stage. dinosterane Steroidal hydrocarbon produced by dinoflagellates. dioecious A condition in which ovule- and pollen-bearing structures are borne on separate plants; all dioecious plants have unisexual reproductive structures (contrast monoecious).

1035

dithecal Having two compartments. diploid Having two sets of chromosomes in each body cell (2n) (contrast polyploid; haploid). disk flower One of the central, tubular flowers in a head (Asteraceae); usually regular in their symmetry (contrast ray flower). distal Farther away from the axis, center, or point of attachment. distal surface The surface of a spore or pollen grain away from the center of a tetrad (contrast proximal surface). distichous Producing appendages in two opposite, vertical, or longitudinal rows. DNA Deoxyribonucleic acid. di-upsilon shaped Bar shaped with two oppositely recurved arms at each end; for example, the cross-sectional view of petiole traces in Ankyropteris. doliporus septum (ⴝ septum with dolipores) A type of septum found in basidiomycetes that is characterized by a central pore covered on both sides by a septal pore cap (parenthesome). drupe A one-seeded, indehiscent, fleshy fruit with a stony endocarp, a fleshy mesocarp, and a membranous exocarp, for example, a plum or cherry. dubiofossil A fossil whose biogenic origin (i.e., resulting from a onceliving organism) is in question. earlywood Wood produced early in the growing season; usually characterized by larger diameter, more thin-walled tracheary elements; also called spring wood (contrast latewood). ectomycorrhiza A type of mycorrhiza in which the fungus surrounds the root tip with a sheath and penetrates into intercellular spaces; roots with ectomycorrhizal infections are short, swollen, and branched; common in modern pines (contrast endomycorrhiza). ectopic Arising in an unusual place. ectotrophic Deriving nourishment from the surface; a mycorrhiza in which the mycelium forms an external covering on the root but does not penetrate the host cells, i.e., ectomycorrhizae. ectophloic Having phloem only on the outer side of the stele. edaphic conditions The physical and chemical nature of the soil and/ or water in which a plant is growing. elaters Elongated structures that may be attached to the spore wall (as in some sphenophytes) or present in the sporangium (in some liverworts) that are believed to aid in spore dispersal. eligulate Lacking a ligule (contrast ligulate). embolism The occurrence of an air bubble within the water-conducting system of a plant and the subsequent cavitation (collapse) of the system containing the bubble. enation A nonvascularized, epidermal outgrowth found in some early land plants. encyclocytic stomata Single row of five or more cells enclosing the guard cells. endarch A type of xylem maturation in which the protoxylem is situated internal to the metaxylem and development proceeds centrifugally (contrast exarch and mesarch). endexine The inner, typically homogeneous layer of the exine in pollen and spores that normally stains less deeply. endobiotic A parasite or saprotrophic organism that develops in the interior of the host. endocarp The inner layer of the pericarp (mature ovary wall) (compare exocarp and mesocarp). endodermis A single layer of cells immediately outside of the stele; Casparian strips or thickenings may occur on their anticlinal walls.

1036

glossary

endohydric Conducting water on the inside of the plant body as in certain bryophytes. endolithic Living within rocks, usually in the pore spaces within the rock, as in endolithic algae, bacteria and fungi. endomycorrhiza The most common type of mycorrhizal infection in which the fungus grows intracellularly within the cortex of the root; the root is not conspicuously different in morphology from uninfected roots (contrast ectomycorrhiza). endoreticulations Ornamentation on the inside of the saccus walls in certain pollen grains; true endoreticulations are not continuous between the inner saccus wall and the corpus (contrast protosaccate). endosperm The triploid (3N) storage tissue found only in the seeds of angiosperms. endosporangium (pl. endosporangia) A spore-containing unit in a fossil cyanobacterium. endospore A spore formed inside a mother cell. endosporic development The development of the gametophyte generation within the confines of the spore wall (contrast exosporic development). endotesta The delicate, innermost layer of the integument or seed coat in some seeds (see also sarcotesta and sclerotesta). entomophilous Pollination by insects. epibiotic An organism that develops its reproductive organs on the surface of the host. epicotyl The portion of the stem above the cotyledons and below the first true leaves in a seedling. epidogenesis A type of growth that results in a continual increase in the diameter of primary xylem at the base of a plant organ (contrast apoxogenesis). epigeal germenation A type of germination in which the cotyledons appear above the ground; through the elongation of the hypocotyl (contrast hypogeal germination) epigynous Having the floral parts (calyx, corolla, and androecium) attached to the top of the ovary; the ovary is said to be inferior (contrast hypogynous). epimatium A fleshy outgrowth covering the ovule of members of the podocarpacean conifers. epiphyllous Upon a leaf; an organism that lives on leaves; some epiphylls may also be parasitic on the leaf itself. epiphyte Growing upon another plant and not attached to the soil. epithelium A layer of parenchyma cells that lines the inside of an intercellular canal or cavity. equisetiform hair A multicellular, uniseriate epidermal hair characterized by an extension of the end walls that gives a toothlike appearance; each hair has the appearance of a miniature Equisetum stem. ergastic deposits Cellular waste products, either crystalline or noncrystalline. eucamptodromous Venation pattern in which the pinnate secondary veins turn up at the ends and gradually diminish; they connect to the secondaries above by cross veins but do not form prominent arches. eukaryote An organism that possesses a membrane-bound nucleus and other membrane-bound organelles (contrast prokaryote). eurytopic A plant that is able to adapt to a wide range of environmental conditions and so is geographically widespread. eusaccate A pollen grain that possesses a true saccus with endoreticulations (contrast protosaccate). eusporangium (pl. eusporangia) A sporangium that arises from a group of superficial cells (contrast leptosporangium).

eustele A stele type that consists of distinct strands called sympodia embedded in parenchymatous tissue; the stele type of most seed plants. exannulate Lacking an annulus. exarch A type of xylem maturation in which the protoxylem is toward the outside of the metaxylem and development is centripetal (contrast endarch and mesarch). exindusiate Without an indusium; a naked sorus. exine The outermost wall layer of the sporoderm of pollen and spores, outside the intine; made of sporopollenin. exocarp The outer layer of the pericarp or fruit wall (compare endocarp and mesocarp). exon A portion of DNA that is transcribed into RNA; characteristic of eukaryotes (contrast intron). exosporic development Development of the gametophyte outside the parent spore, that is, a free-living gametophyte; typical of homosporous ferns (contrast endosporic development). exothecium Thick-walled cells of a moss capsule; outer coat of an anther in angiosperms. exserted Projecting beyond the usual distance, as stamens that extend out beyond the corolla tube at maturity. exstipulate Without a stipule. extrorse Directed outward (contrast introrse). facies The appearance and all of the characteristics of a rock unit that typically reflect its condition of origin. falcate Sickle shaped. false stem An axis that superficially resembles a true stem but structurally does not, for example, the false stem of Psaronius consists of a small true stem with a large, adventitious root mantle. fascicular Arising within a vascular bundle, as a fascicular cambium; in a bundle. feeder An outgrowth of the hypocotyl that serves as a temporary organ of absorption in certain gnetophytes. female gametophyte See megagametophyte. filament (1) The usually slender stalk of a stamen that bears the anther sacs at its tip; (2) a string of cyanobacterial cells surrounded by an investment or sheath. filiform Threadlike or filamentous shape. fimbrils Delicate strands of wall material that extend between the scalariform bars in certain fossil lycopsid tracheids (also termed Williamson striations). flabelliform Fan shaped. floret A small, often imperfect flower that is found in the inflorescence of the Asteraceae (a head) or of the Poaceae (a spike). foliar member A generalized term used to refer to frond parts of certain fossil ferns; most demonstrate bilateral symmetry and unusual branching patterns. foliar physiognomy The analysis of leaf morphology. foliose Leaflike in appearance, as in the foliose lichens. follicle A dry, unilocular angiosperm fruit type characterized by longitudinal dehiscence via a suture on one side; a follicle is formed from a single carpel and contains more than one seed. foot The base of a sporophyte that physiologically and physically connects it to the gametophyte. Foraminifera A group of heterotrophic, marine protists with calcareous shells (tests); some are planktonic and some benthic; also called forams. foveolate

Ornamented with small pits; pitted.

frass Fossilized fecal remains, typically of a small, burrowing arthropod; see also coprolites.

glossary

free sporing Non-seed plants that produce spores (e.g., bryophytes, lycophytes, ferns). frond The leaf of a fern or seed fern; a large divided leaf, as a cycad frond. frugivorous Fruit-eating. frustule The shell (cell wall) of a diatom, composed of silicon dioxide (SiO2). fusain A component of coal characterized by its silky luster, fibrous structure, and black color; fusinized (or fusainized) refers to fossil plants preserved in fusain, that is, preserved as charcoal or charcoalified. fusiform initials A cambial initial that gives rise to axial (vertical) elements of the secondary xylem and phloem; it appears fusiform (elongated with tapered ends) in tangential section. fusinite A type of inertinite with some cellular structure, a reflectance above vitrinite, and a particle size 50 μm. gametangiophore A stalk that bears a gametangium. gametangium (pl. gametangia) A structure that bears gametes or sex cells. gamete A haploid (n) reproductive cell or sex cell; sperm and egg are two types of gametes. gametophore Gamete-bearing structure in bryophytes. gametophyte The haploid (n) phase of a plant life cycle on which gametes are produced (contrast sporophyte). ganodermatoid Fungal spore type in which two wall layers are separated by delicate inner wall pillars as in spores of most members of the polypore family Ganodermataceae. GC–MS Gas chromatography–mass spectrometry. gemma (pl. gemmae) A multicellular, asexual reproductive structure or propagule found in bryophytes, which consists of a small piece of the plant body (thallus). geochronology The science of dating rocks. glochidium (pl. glochidia) A hairlike structure with a barbed or hooked tip that forms on the microspore massulae of certain heterosporous ferns and that serves to attach the microspores to the megaspore. glume A bract that occurs at the base of spikelets in grasses (Poaceae). gradate A type of sorus development in which the sporangia within the sorus mature from the center toward the periphery. grana Stacks of membrane-bound disks within chloroplasts that contain the chlorophyll and accessory pigments; in palynology, a type of surface ornamentation that consists of small, rounded warts 1.0 μm in diameter (sing. granum). granulose Granular, finely roughened. graptolite An extinct, chitinous, colonial animal that is useful in biostratigraphy; a member of the class Graptolithina. guard cell One of two chlorophyllous cells that flank a stoma; changes in turgor pressure in the guard cells open and close the stomata. gula A neck-like projection or extension on the proximal face of a trilete spore; an apical prominence. gynoecium All the carpels in a flower. gyrogonite A fossilized oogonium of the Charophyceae (Chlorophyta). halophyte A plant that grows in a saline environment. haplobiontic life cycle A life cycle in which there is only one freeliving organism; if zygotic meiosis is present, the free-living organisms are haploid and the zygote is the only diploid cell. haplocheilic A type of stomatal ontogeny pattern in which the guard cells arise from a common cell and the subsidiary cells are derived from other epidermal cells; typical of members of the Cycadales (contrast syndetocheilic)

1037

haploid Having a single set of chromosomes in a cell (n) (contrast polyploid; diploid). haplostele A cylindrical protostele with a smooth margin (contrast actinostele). haptotypic mark A characteristic mark or marks on spores or pollen grains that result from contact within a tetrad, for example, the laesurae and contact areas, or a trilete mark; the presence of a haptotypic mark is sometimes used in fossils as evidence for the occurrence of meiosis. haustorium (pl. haustoria) The structure that penetrates a host and absorbs nutrients in a parasitic relationship. head An inflorescence type found in the Asteraceae that consists of a dense, terminal collection of flowers surrounded by an involucre; both ray and disk flowers can occur in a head; also called a capitulum. helophyte A perennial plant with renewal buds, commonly on rhizomes, buried in soil or mud below the water level. hemiparacytic stomata Astomatal type in which one guard cell is flanked by a parallel subsidiary cell whereas the other guard cell has normal epidermal cells surrounding it. herbaceous A plant having little or no secondary development; nonwoody. heteroblastic development Programmed changes in shoot axis and leaf morphology during plant ontogeny. heterocellular Having cells of more than one type (contrast homocellular). heterocyst A specialized, often thick-walled cell found in the filaments of cyanobacteria; represents the site of N2 fixation. heterophyllous Having leaves of different size and/or form. heterosporous Having two types of spores: microspores and megaspores. heterotrichous A thallus consisting of two different types of growth: prostrate and upright, filamentous growth. heterotroph An organism that obtains its food by ingesting organic materials; heterotrophs are incapable of producing their own food; also saprotrophism; also mixotrophy, osmotrophy, myzotrophy, and phagotrophy, which are types of heterotrophy. Contrast with autotrophy. hilum The central point in a starch grain; the scar left on a seed indicating the attachment point. holdfast The basal portion of an algal thallus that serves to attach the plant to its substrate. holocarpic An organism whose entire thallus is converted into one or more reproductive organs. holophytic Synthesizing organic compounds from inorganic substrates, as in a green plant. holozoic Feeding entirely in the manner of an animal by ingesting complex organic matter. homocellular Having only one type of cell (contrast heterocellular). homoiohydric A plant able to regulate the rate of water loss so that it can remain hydrated when the water supply is restricted, such as vascular plants (contrast poikilohydric) homologous Having a shared evolutionary history (contrast analogous). homologous (transformation) theory Theory for the origin of the alternation of generations in which the sporophyte evolved by a modification of the gametophyte and not as a new phase in the life cycle (contrast antithetic theory). homosporous Having one type of spore. hydathode A specialized structure that exudes liquid water from a leaf. hydrasperman reproduction A type of reproductive system in early seed plants that includes a pollen chamber often with the tip of the

1038

glossary

chamber elongate, and a prominent central column arising from the membranous pollen chamber floor. hydroid A water-conducting cell in bryophytes. hydrophilous Pollinated by water. hydrophyte A land plant adapted to wet conditions (contrast mesophyte and xerophyte). hydropote A cell or group of cells found on the lower epidermis of members of the Nymphaeaceae which are thought to function in the uptake of ions from the water. hypanthium Cuplike or floral tubelike structure atop the ovary that bears sepals, petals, and stamens. hyperplasia Abnormal increase in the number of cells that increase the size of an organ. hypertrophy An abnormal increase in the size of cells. hypha (pl. hyphae) A single tubular filament that represents the structural entity of most fungi. hyphopodium (pl. hyphopodia) A short, lateral hyphal branch represented by one or two cells, and characteristic of certain obligately parasitic fungi. hypocotyl The portion of the axis in an embryo or seeding below the cotyledons and above the root or radicle (embryonic root). hypodermis The outer cortex immediately beneath the epidermis; often serves a support function and may include collenchyma or sclerenchyma. hypogynous Having the floral parts (sepals, petals, and stamens) attached at the base of (below) the ovary; the ovary is said to be superior (contrast epigynous). hypogeal germination A type of seedling germination in which the cotyledons remain below ground level (contrast epigeal germination). hypostomatic Bearing stomata only on the abaxial (lower) surface. ichnofossil A trace fossil. ichnogenus, ichnospecies Names given to morphotaxa of trace fossils. idiocuticular Formed by, or composed of, only the cuticle, for example, warts or ridges on the cuticle surface. imbricate Overlapping. imparipinnate A frond with a single apical pinnule; contrast paripinnate. imperfect A flower in which one of the sexual parts (either stamens or carpels) is missing; imperfect flowers are unisexual (contrast perfect). impression A preservation type that represents a negative imprint of an organism; no organic material remains. in situ In place, as pollen grains and spores found within the sporangium that produced them (contrast sporae dispersae). inaperturate Of spores and pollen, having no aperture or other opening. indehiscent Not naturally opening at maturity. indeterminate Unlimited growth and development, as of the vegetative growth of seed plants (contrast determinate). index fossil A geographically widespread fossil that is diagnostic of a particular restricted time period and therefore useful in biostratigraphy. indumentum A hairy covering in plants and animals indusium (pl. indusia) The shieldlike covering over the sorus in a fern. inertinite A type of organic component of coal (a maceral) with a moderately high carbon content that is relatively inert during the carbonization process.

inertodetrinite A type of inertinite with a reflectance greater than associated vitrinite; occurs as angular fragments in coal. inferior ovary A flower in which the floral parts (sepals, petals, and stamens) are attached to the top of the ovary and fused to it (see epigynous). inflorescence A cluster of flowers or a single flower. infructescence An inflorescence after fruit development; a cluster of fruits. integument Seed coat or testa; tissue that covers the megasporangium (nucellus) in a seed plant. intercalary growth Growth occurring within the plant body, that is, the production of new tissues between old tissues; for example, growth occurring within the internode of a vascular plant or within a thallus. intercostal areas/fields The spaces between the veins on a leaf. interfascicular ray The region between the vascular bundles in a stem (compare pith ray). internode The portion of an axis between two nodes. interpolation theory See antithetic theory. intine Generally the innermost layer of the sporoderm in pollen and spores, inside the exine; made up of cellulose and pectates and not present in many fossils. intron A section of mRNA (messenger RNA) that is removed prior to protein transcription (contrast exon). introrse Directed inward; toward the center (contrast extrorse). involucre A cluster of bracts or modified leaves that subtend the inflorescence in the Asteraceae. iridoid compounds A group of monoterpenes that often occur as glycosides; they are water soluble. irregular flower A flower with bilateral symmetry (compare zygomorphic flower; contrast regular flower). isodiametric Having the same diameter in each direction; typical of parenchymatous cells. isogametes Gametes that are of the same size and appearance. isomorphic A type of alternation of generations in which the haploid and diploid generations are morphologically similar. isophyllous Producing one type of leaf. isopolar pollen Pollen with radial symmetry and no conspicuous difference between the two poles. isospore A generalized term for a spore produced by a homosporous plant. isotomous Dividing equally and repeatedly. kelp A large marine alga of the Laminariales (Phaeophyceae). kerogen The insoluble, amorphous organic matter present in sediments. Kranz anatomy A type of leaf anatomy characterized by leaf veins that are surrounded by bundle sheath cells, a low ratio of mesophyll to bundle sheath area, and vascular bundles that are separated from each other by few mesophyll cells; characteristic of plants with C4 photosynthesis. K-selected A term used to describe the reproductive strategy of organisms that live in less disturbed, more stable communities and that produce fewer, larger propagules (contrast r-selected). lacuna (pl. lacunae) A cavity or space within a plant; lacunate  containing lacunae. lacustrine deposits Lake bed deposits. laesura (pl. laesurae) A haptotypic mark. laevigate With a smooth surface, for example, of pollen and spores.

glossary

lagenostome A small extension of the distal end of the megasporangium (nucellus) of certain fossil seed plants. Lagerstätte (pl. Lagerstätten) An exceptionally preserved or rich fossil deposit. lamina (pl. laminae) Flattened blade portion of a leaf or a thallus. lanceolate Longer than wide and tapering to a point. latewood Wood produced later in the growing season; usually characterized by smaller diameter, more thick-walled tracheary elements; also called summer wood (contrast earlywood). lateral meristem See cambium. laterocytic stomata Three or more subsidiary cells, each with their long axis parallel to the long axis of the guard cells. laticifer A cell or cavity containing latex. lax Drooping or relaxed. leaf gap A region of parenchyma within a stele above the point of departure of a leaf trace; the parenchyma appears as an interruption in the vascular tissue of the stele. lemma The lower or outer bract at the base of the flower in the grasses (Poaceae) (see also palea). leptoid A food-conducting cell in bryophytes that has a similar structure to the sieve cell in vascular plants (contrast hydroid). leptoma A slightly thinner region on the distal surface of a pollen grain through which germination takes place. leptosporangium (pl. leptosporangia) A sporangium derived from a single superficial cell; typical of the filicalean ferns (contrast eusporangium). liana A vine or woody climbing plant. lianescent Growing like a liana or vine. lignin A biopolymer of phenylpropanoid units that surrounds the cellulose microfibrils in the secondary wall of many plant cells; it occurs in all plants that possess vascular tissue and provides structural strength, resistance to decay, and a barrier to water permeability. lignophyte A clade composed of the seed plants and the progymnosperms characterized by a bifacial vascular cambium. lignite A brownish-black coaly deposit in which the plant material is not highly consolidated; intermediate in coalification between peat and subbituminous coal. ligulate Having a ligule (contrast eligulate). ligule A small flap of tissue on the adaxial surface of leaves of certain members of the Lycophyta. liguliform Strap shaped, flat and narrow. limonite Hydrated iron oxides; limonitized fossil plants are preserved in limonite. lithofacies A lateral, mappable subdivision of a designated stratigraphic unit distinguished on the basis of lithology. locule A cavity within the ovary of a flower; a cavity in a sporangium. lumina (pl.) In pollen ornamentation, low areas between striae or muri. lunate Shaped like a crescent. lysigenous Formed from the breakdown or lysis of cells, as in lysigenous lacunae. maceral One of the organic constituents that comprise the coal mass. macrosclereid A sclereid cell that is longer than wide; often rod or barrel shaped. male gametophyte See microgametophyte. manoxylic Wood type that contains abundant parenchyma, for example, in the cycads (contrast pycnoxylic).

1039

margo See pit membrane. massa An extension on the proximal surface of a megaspore (e.g., in Achlamydocarpon) that is constructed of irregularly shaped sporopollenin globules. massula (pl. massulae) A large mass of mucilaginous material that encloses the microspores in some water ferns (e.g., Azolla) that develops from the tapetum of the microsporangium; the term also refers to a mass of microspores in orchids. medulla Pith (in vascular plants); a mass of loose hyphae in a fungal or lichen thallus; in algae the central portion of the thallus. medullary ray See interfascicular ray. medullated Having a pith. megagametophyte In heterosporous plants and seed plants, the haploid (n) plant body produced by a megaspore; the female gametophyte. megaphyll A leaf with more than one vein and a trace associated with a leaf gap in the stele; it may or may not be large in size; characteristic of the ferns and seed plants. megasporangium (pl. megasporangia) A sporangium in which megaspores are produced (compare nucellus). megaspore A large, haploid (n) spore of a heterosporous plant that produces a megagametophyte, or female gametophyte; the single, functional spore in seed plants. In sporae dispersae, megaspores are arbitrarily defined as spores 200 μm in diameter. megasporophyll A leaflike organ that bears one or more megasporangia or ovules. meiosis Reduction division; two successive nuclear divisions that reduce the ploidy level of a cell from diploid (2n) to haploid (n); in plants, spores are the products of meiosis. melasmatic tissue Large cells that occur in calamites around the vascular bundles; typically contain dark contents. meristele A segment of vascular tissue, surrounded by an endodermis, that makes up part of a dictyostele, for example, in the extant fern Pteridium. merom, merome Lateral branches in the receptaculitids. mesarch A type of xylem maturation in which the protoxylem is embedded in the metaxylem and development proceeds centripetally and centrifugally. mesocarp The middle layer of the pericarp or fruit wall; between the endocarp and the exocarp. mesofossil Intermediate size (0.25 mm to several millimeters) plant fossils (e.g., megaspores, small seeds, flowers) that require microscopic study. mesophyll The parenchymatous, chlorophyllous tissue in a leaf located between the upper and lower epidermis; may consist of palisade parenchyma and spongy parenchyma. mesophyte A land plant that is adapted to humid conditions (contrast xerophyte and hydrophyte). mesophytic An informal period of geologic time that extended from the dominance of conifers and other gymnosperms (cycadophytes, ginkgophytes) (about mid-Permian) until the dominance of the angiosperms (the cenophytic; about mid-Cretaceous). mesothermic forest A forest that occurs in moderate heat and moisture conditions; a temperate forest. metaphyte A multicellular plant with some degree of organ differentiation. metaxylem A component of the primary xylem that differentiates and matures later than the protoxylem; metaxylem matures after the plant has essentially ceased elongation (contrast protoxylem).

1040

glossary

metazoa (pl.) Multicellular animals. (sing metazoan) microbial mats An accumulation of films of communities of microorganisms that become more than a centimeter thick. microgametophyte The haploid (n) plant body produced by the microspore in heterosporous and seed plants; the male gametophyte. microphyll A small leaf vascularized by a single vascular bundle that is not associated with a gap in the stele; typical leaf type of the lycophytes. micropyle A small opening in the integument at the apex of a seed through which either the pollen (in gymnosperms) or the pollen tube (in angiosperms) enters. microsporangium (pl. microsporangia) A sporangium in which microspores are produced. microspore In heterosporous plants, a small spore that produces the microgametophyte. microsporophyll A leaflike organ that bears one or more microsporangia (contrast megasporophyll). miospore A neutral term for fossil spores smaller than 200 μm in diameter whose biological function (as micro- or megaspore) is not necessarily known. mixotrophy Obtaining nutrients by both photosynthesis (autotrophy) and ingesting organics or prey (heterotrophy). mold A three-dimensional preservation type that represents a negative imprint of the plant (compare cast). molecular clock A technique used to relate the divergence time of two species based on DNA sequences or proteins measured against a reliable fossil record. monarch Having a single protoxylem group. monocarpic Producing reproductive propagules only once in the lifetime of the plant. monocolpate A pollen grain with a single, meridionally positioned colpus; often used interchangeably with monosulcate, which is preferred. monocyclic stomata A stomatal arrangement with a single row of subsidiary cells surrounding the guard cells. monoecious Bearing both pollen- and ovule-producing organs on the same plant; the organs can be either unisexual or bisexual (both types in a single fructification) (contrast dioecious). monolete A spore with a single, straight laesura on the proximal surface (contrast trilete). monophyletic A group of organisms that consist of a common ancestor including all descendants of that ancestor. monopodial Having a single, main trunk or axis. monoporate A pollen grain with a single pore; typical of the grasses (Poaceae). monosporangiate Bearing only a single type of spore in one reproductive organ (i.e., either megaspores or microspores). monosulcate Pollen grain with a single sulcus; often used as being equivalent to monocolpate. morphotaxa (e.g., morphospecies and morphogenus) taxonomic units defined based on macromorphological features. They are used for, and assigned to, isolated parts of fossil plants (e.g., leaves, cones, stems), and indicate structural similarity but not necessarily biological relationship. motile Having the ability to swim, for example, some algal spores or gametes. mucronate Ending in a sharp point. multicellate A fungal palynomorph having six or more connected cells or chambers separated by septa. murus (pl. muri) A pollen or cyst (algae) ornamentation type characterized by a series of ridges or walls that make up a reticulate sculpture pattern.

mutualism A symbiotic relationship that is not detrimental to either organism involved. mycelium (pl. mycelia) The body of a fungus, made up of hyphae which may be aggregated into different morphologies. mycoparasite A fungus parasitic on another fungus. mycorrhiza A symbiotic association between fungi and the roots of vascular plants (see ectomycorrhiza and endomycorrhiza). myzotrophy A method of feeding in some heterotrophic organisms in which a feeding tube is inserted into the prey and the cellular contents are sucked out; found in some dinoflagellates and euglenoids. nannoplankton Nannofossils; very small (usually