William E. Duellman Editor Museum of Natural History The University of Kansas Monof^aph No. 7 HARVARD UNIVERSITY Library of the Museum of Comparative Zoology THE SOUTH AMERICAN HERPETOFAUNA: ITS ORIGIN, EVOLUTION, AND DISPERSAL THE SOUTH AMERICAN HERPETOFAUNA: ITS ORIGIN, EVOLUTION, AND DISPERSAL WILLIAM E. DUELLMAN EDITOR Museum of Natural History and Department of Systematics and Ecology The University of Kansas Lawrence, Kansas 66045, USA MONOGRAPH OF THE MUSEUM OF NATURAL HISTORY, THE UNIVERSITY OF KANSAS NUMRER 7 1979 MONOGRAPH OF THE MUSEUM OF NATURAL HISTORY, THE UNIVERSITY OF KANSAS Number 7, pages 1^85, 172 figures in text Issued December 28, 1979 © 1979 by The Museum of Natural History, The University of Kansas, Lawrence, Kansas. All rights reserved. No part of this book may be reproduced in any form or by any means without permission in writing from the publisher. ISRN Number: 0-89338-008-3 MUS. COMP. ZOOL LIBRARY Cover design by Linda Trueb JUN 51985 HARVARD UNIVERSITY PRINTED BY UNIVERSITY OF KANSAS PRINTING SERVICE LAWRENCE, KANSAS, USA Dedicated to the memories of three herpetologists who contributed so much to our knowledge of the South American herpetofauna: Roberto Donoso-Barros (1922-1975) Bertha Lutz (1894-1976) James A. Peters (1922-1972) PREFACE This volume is the result of a symposium of the same title held on 11-13 August 1977 in conjunction with the joint annual meetings of the Herpetologists' League and the Society for the Study of Amphibians and Reptiles at Lawrence, Kansas. I originally conceived the idea for such a symposium in August 1975 while returning from a 15-month sojourn in South America. My interactions with many South American biologists during that trip had convinced me that the time was appropri- ate for a thorough discussion of ideas and presentation of our existing knowledge of the South American heq^etofauna. The initial response from colleagues was heartening, so during the following year the symposium was organized. Unfortunately, owing to various circumstances not all subjects were covered; obvious omissions in this volume are chapters on the South American-North American her- petofaunal relationships and the herpetofau- nas of the Brasilian Highlands, the Atacama Desert, and the caatinga and campos cerrados of Brasil. This volume is organized in much the same way as was the symposium, except that my introductory chapter provides an overview of the South American herpetofauna. Chapter 2 deals with the fossil record of amphibians and reptiles in South America, and Chapters 3 and 4 are concerned with the relationships of the South American herpetofauna with those of Africa and Australia. The Quaternary bio- geography of the continent is the subject of Chapters 5-7. Treatments of regional herpeto- faunas are found in Chapters 8-15, and the Participants in the Symposium on the South American Herpetofauna held in Lawrence, Kansas, 11-13 August 1977. Front row (left to right): Alberto Veloso M., Beryl B. Simpson, Jaime E. Pefaur, Ana Maria Baez, Jose M. Cei. Second row: Lars Brundin, Thomas E. Lovejoy, Donn E. Rosen, Jiirgen Haffer, Thomas H. Fritts, Ramon Formas, Raymond F. Laurent. Back row: William E. Duellman, James R. Dixon, Marinus S. Hoogmoed, John D. Lynch, W. Ronald Heyer, Michael J. Tyler, Jose M. Gallardo. final chapter is devoted to the conservation of the herpetofauna. I am grateful to the contributors to this volume for their scholarly efforts and for their patience and understanding while it was being produced. For their participation in the symposium, I thank the contributors and Lars Brundin, Thomas H. Fritts, W. Ronald Heyer, Jaime E. Pefaur, Donn E. Rosen, and Alberto Veloso M. Their enthusiastic partici- pation contributed a high level of scholarly interaction, as well as much good cheer. During the editing of this volume I called upon many colleagues to review manuscripts. The quality of the papers included herein benefited from reviews by Avelino Barrio, Lars Brundin, Richard Estes, Thomas H. Fritts, Steven Gorzula, W. Ronald Heyer, Philip S. Humphrey, Jean Lescure, Alan E. Leviton, John D. Lynch, Larry D. Martin, Braulio Orejas-Miranda, Jaime E. Pefaur, Alan H. Savitzky, Beryl B. Simpson, Linda Trueb, T. van der Hammen, Alberto Veloso M. and Richard G. Zweifel. The drawings for many of the papers were executed by Debra K. Bennett, Staff Illustrator of the Museum of Natural History at The University of Kansas. Jaime E. Pefaur translated many of the sum- maries and edited the Spanish of others. Lin- da Trueb 's competent editorial review of the manuscripts is evident in their consistency and style. Rose Etta Kurtz retyped many pages of manuscript, and Rebecca A. Pyles painstakingly worked on the index. To all of these persons I owe a debt of gratitude for their endeavors in behalf of this volume. Throughout the early phases of develop- ment and organization of the symposium, as well as during the production of this volume, Philip S. Humphrey, Director of the Museum of Natural History, has provided advice, en- couragement and support. Ronald K. Cal- gaard. Vice Chancellor for Academic Affairs, and George R. Waggoner, Associate Vice Chancellor for International Programs, The University of Kansas, gave enthusiastic sup- port for the symposium. Richard F. Treece of the Bureau of Conferences and Institutes coordinated the logistics of the meetings. Without their interest and aid the symposium and this volume would not have been pos- sible. Financial support for bringing together the participants in the symposium was gen- erously provided by the National Science Foundation (DEB 76-16767), the World Wildlife Fund (WWF-US-71) and the Office of Academic Affairs, The University of Kansas. Support for the preparation of the index was provided by a grant from the General Re- search Fund of The University of Kansas. William E. Duelhnan Lawrence, Kansas September 6, 1979 CONTENTS 1. The South American Heqjetof auna : A Panoramic View. William E. Dnellman 2. The South American Herpetofauna: An Evaluation of the Fossil Record. Ana Maria Bdez and Znlma B. de Gasparini 29 3. Herpetofaunal Relationships Between Africa and South America. Baymond F. Laurent 55 4. Herpetofaunal Relations of South America with Australia. Michael J. Tyler 73 5. Quaternary Biogeography of Tropical Lowland South America. Jiirgen Haffer 107 6. Late Cenozoic Environmental Changes in Temperate Argentina. Ana Maria Bdez and Gustavo Juan Scillato Yane 141 7. Quaternary Biogeography of the High Montane Regions of South America. Beryl B. Simpson ... 157 8. The Amphibians of the Lowland Tropical Forests. John D. Lynch 189 9. Origin and Distribution of Reptiles in Lowland Tropical Rainforests of South America. James R. Dixon 217 10. The Herpetofauna of the Guianan Region. Marinus S. Hoogmoed 241 11. Origin and Distribution of the Herpetofauna of the Dry Lowland Regions of Northern South America. Carlos Rivero-Blanco and James R. Dixon 281 12. Composition, Distribution y Origen de la Herpetofauna Chaquefia. Jose M. Gallardo _ 299 13. The Patagonian Herpetofauna. Jose M. Cei 309 14. La HeqDetofauna de los Bosques Temperados de Sudamerica. /. Ramon Formas 341 15. The HeqDetofauna of the Andes: Patterns of Distribution, Origin, Differentiation and Present Communities. William E. Duellman 371 16. Refugia, Refuges and Minimum Critical Size: Problems in the Conservation of the Neotropical Herpetofauna. Thomas E. Lovejoy 461 Subject Index 465 Taxonomic Index 470 1. The South American Herpetofauna: A Panoramic View William E. Duellman Museum of Natural History and Department of Systematics and Ecology The University of Kansas Lawrence, Kansas 66045 USA A vast array of dinosaurs still inhabited the earth, ratite birds watched curiously as furry mammals experimented with new ways of reproduction, and varieties of anurans and squamates set out on diverse evolutionary courses, while turtles continued their conser- vative approach to a changing world. They witnessed the breakup of the earth as Gond- wanaland was split by the magma, giving birth to a new ocean, and somewhat later the fracture of the land again to create a large island continent — South America — destined to drift in a northwestward arc for nearly fifty million years before establishing a narrow connection with a neighbor of long ago and far away — North America. During that long period of isolation of South America, some of the archaic groups of plants and animals became extinct; some dwindled in numbers leaving only a few scat- tered relicts, and others prospered and gave rise to new and diverse kinds in the face of changing environments, for the island was not static. Tectonic events and climatic changes shaped the landscapes in an ever-changing scene. Great areas of the land were innun- dated by epeiric seas, the southern end of the continent cooled and desiccated as world zonation of climates was established; the rise of a gigantic mountain chain interrupted the winds and modified the climates and gave birth to thousands of small streams that coa- lesced in their descents to the lowlands and formed huge rivers, and finally, cooling brought glaciation and fluctuations in climate and sea level that brought about changes in the drainages and the biota. Late in this scenario man entered South America and began preying upon the animals, clearing land, cultivating plants, and building temples. So it was when Columbus "discovered" South America on his third voyage in 1498, when Ferdinand Magellan arrived in 1520, and at the beginning of the conquest of the New World by Francisco Pizarro in 1532. For nearly two centuries the decimation of the native human populace (by the sword, the Bible, and disease) was second in importance only to the frenetic search for El Dorado — the real and fabled materialistic riches of South America. With the exception of pitifully few men, no attention was given to the natural riches of the continent. But in the 18th Cen- tury, European naturalists began exploring South America, as related so eloquently by von Hagen (1948:xiii). For it was the explorer-naturalists who opened South America. It was these knowl- edge-thirsting men who, because they were deemed harmless, were permitted entry be- hind the Green Curtain when others were not. It was the naturalists who methodically and systematically pushed aside the frontiers of South America and dug it from its oblivion. With an enthusiasm that bridged every bar- rier, they climbed the Andes, they swept down the dark mysterious rivers, they trekked across the deserts and struggled through the Laocoon entanglements of its fire-fly spangled jungles. They dispelled legends, they uncovered facts, they rediscovered rubber, studied quinine and the coca leaf. They measured the earth's sur- face, they crawled into the jungle and col- lected plants, they studied the animals, they measured the tides. ... It was the naturalists who opened South America. Early collections reaching Europe formed substantial natural history cabinets, and some of these were illustrated and described; the most ambitious undertaking was that by Al- bertus Seba, who in the 1730's published his classic "Thesaurus" in four volumes. The illus- MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 (rations in Seba's work and specimens that reached Upsala formed the basis for names of South American species by Linnaeus in 1758. In the 19th Century, some of the world's most famous naturalists worked in South America — Charles Marie de La Condamine, Alexander von Humboldt, Alfred R. Wallace, Henry W. Rates, Richard Spruce, and of course Charles Darwin. These men made extensive natural history collections, but these included few, if any, amphibians and reptiles. Plants, insects, and birds were the chief goals of most of the collectors. Five European naturalists made important contributions to the early knowledge of the South American herpetofauna in the early 1800's through their collections and their writings — Maximillian A. P. zu Wied-Neu- wied, Alcide D. D'Orbigny, Johann R. von Spix, Marcos X. Jimenez de la Espada, and Johann J. Tschudi. Some of the collections made by those men, plus many small collec- tions that reached European museums pro- vided the basis for countless papers on South American amphibians and reptiles by Albert Giinther, Wilhelm Peters, Oskar Boettger, Franz Steindachner, and the most prolific of European herpetologists — George Boulenger. During the latter part of the 19th Century, South American specimens reached the United States; most of these were reported on by Edward D. Cope. Ry the beginning of the present century several centers of biological research had been established in South America. Early pioneers in herpetological research included Julio Koslowsky in Argentina, R. A. Philippi in Chile, and Alipio de Miranda-Ribeiro and Adolfo Lutz in Rrasil. By the mid-20th Cen- tury investigations on the South American herpetofauna flourished. But as research on amphibians and reptiles broadens to include studies on the ecology, life history, and be- havior, the need still remains for descriptive morphology and systematics. Ever increasing human disturbance of natural environments, especially the rainforests, eliminates forever many components of the biota before they be- come known to science. COMPOSITION OF THE HERPETOFAUNA The complex history and diverse topog- raphy and climate of South America have produced an extraordinarily rich and diverse herpetofauna. Currently more than 2,200 species are recognized in more than 300 gen- era in 37 families (Tables 1:1-1:2). These numbers are bound to increase with future discoveries. The rate of discovery of new species in South America is astonishing. As examples, of the 313 species of hylid frogs now known from South America, 100 have been named in the last two decades (1960- present); Peters and Donoso-Barros (1970) listed 71 species of Anolis, and 14 additional species have been named. Likewise, many new species of frogs, especially Centrolenella, Colostethus, and Eleatherodactylus, and of salamanders (Bolitoglossa) are being discov- ered and named yearly. A far higher percentage of the living am- phibians of the world than of the reptiles inhabits South America. In this respect am- phibians are more like birds, whereas reptiles are more like mammals (Table 1:3). Review of the Families In this brief review, each family is dis- cussed with respect to its origin (Table 1:4), temporal and geographic distribution in South America (Table 1:5, Fig. 1:1), and differen- tiation and dispersal in South America. Ma- rine reptiles are not included. Plethodontidae. — Known from Pliocene and Pleistocene deposits in North America, the family is highly differentiated there (23 genera, about 200 species). Two genera that are most speciose in Central America ( Bolito- glossa and Oedipina) also occur in South America. There, Oedipina (2 species) occurs only in the Choco, whereas Bolitoglossa (24 species) inhabits the Choco, Amazonia, and the northern Andes. Plethodontid salaman- ders entered South America from Central America after the closure of the Panamanian Portal (Wake, 1966). 1979 DUELLMAN: SOUTH AMERICAN HERPETOFAUNA Table 1:1. — Taxonomic Composition of the South American Herpetofauna. ( ° = endemic to South America) Table 1:3. — Comparison of Numbers of Species of Tetrapod Vertebrates in South America with World Fauna. Genera Species 37 4 2 1 2 15 1 13 2 4 6 Family Total Endemic Amphibia Plethodontidae 2 Pipidae 1 Leptodactylidae 41 Bufonidae 7 Brachycephalidae" .. 2 Rhinodermatidae° .. 1 Dendrobatidae — 3 Pseudidae" 2 Hylidae 22 Centrolenidae 2 Ranidae 1 Microhylidae 16 Rhinatrematidae° .... 2 Typhlonectidae" .... 4 Caeciliidae 9 Reptilia Pelomedusidae 1 Chelidae 7 Kinostemidae 1 Chelydridae 1 Emydidae 2 Testudinidae 1 Gekkonidae 16 Iguanidae 27 Teiidae 38 Scincidae 1 Anguidae 2 Amphisbaenidae 6 Anomalepidae 4 Leptotyphlopidae _ 1 Typhlopidae 1 Boidae 5 Aniliidae 1 Tropidophiidae 2 Colubridae 77 Micruridae 2 Viperidae 3 Crocodylidae 4 Total Endemic 7 20 28 1 5 1 2 1 39 1 24 21 5 4 411 396 95 88 2 2 2 2 75 69 4 4 313 285 57 52 1 31 28 10 10 18 18 47 44 7 15 3 1 5 5 64 240 151 8 8 45 17 34 3 10 1 4 409 32 46 7 7 15 2 1 2 4 56 220 140 7 7 44 16 32 3 6 1 3 357 27 39 5 Class Total South American Percentage South American Amphibians Reptiles 3,307 5,954 8,656' .. 4,060' 1,095 1,115 2,780 b 810 d 33 19 Birds 32 Mammals 20" 1 Brodkorb (1972). '" Meyer de Schauensee ( 1964). c Anderson and Jones ( 1967). d Hershkovitz (1972). " The figure 810 includes Central America and the West Indies, so the total number and percentage for South America will be lower. Leiopelmatidae. — With living representa- tives only in North America (Ascaphus) and New Zealand ( Leiopelma ) , this family is rep- resented in South America only by the fossils of Vieraella and Notobatrachus from the Ju- rassic of Patagonia (Estes and Reig, 1973). Pipidae. — At present restricted to sub- Saharan Africa and tropical America, pipids have an extensive fossil record in Gondwana- land — Early Cretaceous of Israel, Upper Cre- taceous-Miocene of southern Africa, and Late Cretaceous-Eocene of South America (Estes, 1975). Early Cenozoic fossils from South America are representatives of the living Afri- can genus Xenopus. One South American genus, Pipa, contains five species in the tropi- cal lowlands (three genera recognized by some authors). Pipa parva enters eastern Panama. Leptodactylidae. — This is the most diverse and speciose anuran family in South America, where it is known back as far as the late Paleo- Table 1:2. — Summary of the Taxonomic Composition of the South American Herpetofauna. Ordinal group Families Genera Species Total Endemic Total Endemic Total Endemic 1 2 24 21 11 3 98 75 996 930 3 2 15 12 75 72 15 5 115 87 1,095 1,023 6 13 7 36 31 5 84 56 471 430 1 6 5 45 44 9 96 44 556 484 1 4 2 7 5 22 203 114 1,115 994 37 5 318 201 2,210 2,017 Salamanders Anurans Caecilians ... total amphibians Turtles Lizards Amphisbaenians Snakes Crocodilians total reptiles total MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 cene (Baez and Gasparini, this volume). Al- though the family is unquestionably of Gond- wanan origin, the relationships of the lepto- dactylids are not clear. Lynch (1971) recog- nized the South American and Australian frogs plus the South African Heleophryne as one family, the Leptodactylidae, but he (1973) separated the Old World genera into the Myobatrachidae. This arrangement was fol- lowed generally by Savage (1973), Duellman (1975), Heyer (1975), and Heyer and Liem ( 1976), but not by Tyler (this volume). Hey- er ( 1975 ) suggested that leptodactylids might have evolved from leiopelmatids; this idea was elaborated upon by Lynch (1978). Within South America, the primitive Tel- matobiinae are primarily distributed in tem- perate regions — the tribe Telmatobiini in Pat- agonia, austral forests, and the high Andes. More advanced telmatobiines are in temper- ate and tropical regions — Odontophrynini in the Chaco, southeastern Brasil, and nonfor- ested regions in eastern Brasil, Grypiscini on the Brasilian Shield, Eleutherodactylini most diverse in northwestern South America but also occurring on the Brasilian and Guianan shields and in Amazonia, and also speciose in Middle America and the West Indies. The diversity of eleutherodactyline genera and the differentiation of Eleiitherodactyhis in Middle America are indicative of immigration of eleu- therodactylines into Central America prior to the establishment of the isthmian link in the late Pliocene (Savage, 1973; Lynch, 1976). The Ceratophryinae are widespread in Cha- coan, Amazonian, and Guianan lowlands. The Elosiinae are restricted to the Brasilian Shield. The Leptodactylinae are widespread in tropical and subtropical lowlands, with a primitive genus (Pleurodema) also inhabiting Patagonia, austral forests, and the Andes (Duellman and Veloso, 1977). Physalaemus, Pleurodema, and Leptodactylus have entered Central America, and the latter also is in the West Indies. Bufonidae. — The earliest fossil bufonids are from the Paleocene of Brasil (Estes and Reig, 1973), followed by the Oligocene Neo- procoela, which is a member of the Eurasian Bufo calamita group, according to Tihen ( 1962) and Baez and Gasparini (this volume) but referred to the telmatobiine leptodactylids by Lynch ( 1971 ) . By the Miocene, Bufo was present in South America, North America, Europe, and Africa (Tihen, 1972). The ab- sence of bufonids from the Australo-Papuan Region (except for the introduced Bufo ma- rinus), combined with the fossil history of the group, strongly suggests a western Gondwana- land origin of the family ( Blair, 1972; Savage, Table 1:4. — Postulated Geographic Origins of Families of Amphibians and Reptiles Inhabiting South America. (NA = North America; SA = South America; f= Extinct in South America) Pangaea Leiopelmatidae 1 Boidae Laurasia Plethodontidae (NA) Kinosternidae (NA) Chelydridae (NA) Emydidae Trionychidaet Anguidae Aniliidae Viperidae Gondwanaland Uncertain Pipidae Leptodactylidae Bufonidae Brachycephalidae (SA) Rhinodermandae ( SA ) Pseudidae (SA) Hylidae Centrolenidae (SA) Ranidae Microhylidae Rhinatrematidae (SA) Typhlonectidae (SA) Caeciliidae Pelomedusidae Chelidae Meiolaniidaef Iguanidae (SA) Teiidae (SA) Anomalepidae ( SA ) Tropidophiidae (SA) Micruridae (SA) Testudinidae Gekkonidae Scincidae Amphisbaenidae Typhlopidae Leptotyphlopidae Colubridae Crocodylidae 1979 DUELLMAN: SOUTH AMERICAN HERPETOFAUNA Table 1:5. — Distribution of Herpetofaunal Family Groups in Major Eco-physiographic Regions in South America. Family group J3 o g •c o ■c O c < 13 c/j 3 O o 2 c 03 s O 3 c u Amphibia Plethodonndae Pipidae Ceratophryinae Telmatobiinae Elosiinae Leptodactylinae . Bufonidae B rachycephalidae Rhinodermatidae Dendrobatidae Pseudidae Phyllomedusinae Hemiphractinae Amphignathodontinae Hyhnae Centrolenidae Ranidae Microhylidae Rhinatrematidae Typhlonectidae Caeciliidae Reptilia Pelomedusidae - Chelidae Kinosternidae Chelydridae Emydidae Testudinidae Gekkoninae Sphaerodactylinae I guanines Basiliscines Anolines Tropidurines Teiidae Scincidae Anguidae Amphisbaenidae Anomalepidae .._ Leptotyphlopidae Typhlopidae Boidae Aniliidae Tropidophiidae Xenodontinae Colubrinae Micruridae Crotalinae Crocodylidae + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + — + + + + — + + — — + — — + — — + + — + + + + — + + — — + — + + — + + + + — + + + — — + — — + — — + — + + + + + — — + + — + — + + + + + — + + + + + + + + — — + MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 1979 DUELLMAN: SOUTH AMERICAN HERPETOFAUNA 1973; Laurent, this volume). Bufo and six other genera occur in South America, and Bufo and seven other genera occur in Africa (principally tropical western Africa), but Bufo and five other genera inhabit southeast- ern Asia and adjacent islands. One genus (Crepidophryne) is endemic to Central America, and Bufo is widespread in the Hol- arctic Region. Differing views have been ex- pressed on the dispersal of Bufo ( Blair, 1972; Savage, 1973; Duellman, this volume; Lau- rent, this volume). Bufo occurs throughout South America, but only members of the Bufo spinulosus group are present in Patagonia, the austral forests, and the high Andes (Cei, 1968, 1972). Rhamphophryne and Atelopus are primarily northern Andean; Dendrophryniscus is in Amazonia and the Brasilian Shield, Melano- phryniscus in the Chaco and adjacent areas, and Oreophrynella in the Guiana Highlands (Trueb, 1971; McDiarmid, 1971). Brachycephalidae. — Unknown in the fossil record, the two small frogs comprising this family are restricted to humid coastal low- lands of southeastern Brasil (Izecksohn, 1971). Although superficially resembling a specialized bufonid, brachycephalids lack Bidder's Organs (McDiarmid, 1971), an uniquely derived character in the Bufonidae (Lynch, 1973). The phylogenetic position of this endemic South American family is not clear, but presumably it arose from a lepto- dactylid-primitive bufonid stock. Rhinodermatidae. — Known from two spe- cies restricted to austral forests (Formas, et al., 1975), Rhinoderma is considered to be most closely related to the bufonids by Lynch ( 1971, 1973 ) and must be considered as of temperate South American origin. Dendrobatidae. — Lacking a fossil record but composed of three Recent genera, the dendrobatids are especially speciose in the northern Andes, Choco, and western Ama- zonia, but also occur in eastern Amazonia and on the Guianan and Brasilian shields. Lynch ( 1971 ) demonstrated that the dendrobatids are derived from the elosiine leptodactylids and thus arc of South American origin. Spe- cies of all three genera occur in lower Central America, presumably having arrived there after the closure of the Panamanian Portal in the late Pliocene. Pseudidae. — An autochthonous South American family containing two genera and four species (Gallardo, 1961) and widely dis- tributed in tropical and subtropical cis-An- dean lowlands, these aquatic frogs have been considered as relatives of the leptodactylids (Savage and Carvalho, 1953) or hylids (Lynch, 1973). HyUdae. — Although Estes and Reig ( 1973 ) mentioned the existence of Paleocene hylid material from Brasil, these specimens have not yet been described. The mid-Mio- cene Australobat melius from Australia has been referred to the Hylidae by Tyler ( 1974 ) ; a presumed hylid is known from the Oligo- cene of North America (Holman, 1968), and Hyla is known from the Miocene of Europe (Noble, 1928). By far the greatest diversity of hylids is in South America ( 22 genera, 313 species), as compared with Middle America ( 15 genera, 129 species; Duellman, 1970). Six Hyla, plus two endemic genera (Osteopdus and Calyptahyla) inhabit the West Indies ( Trueb and Tyler, 1974 ) . The Holarctic hylid fauna is depauperate, but in the Australo- Papuan Region 118 species are known in the genera Litoria and Nyctimystes (Duellman, 1977; Tyler and Da vies, 1978), and nine more if Cyclorana is included in the family (Tyler, et al., 1978; Tyler, this volume). Like the leptodactylids, the Australian hylids are of questionable relationship with the South American hylids. Savage (1973) resurrected the family name Pelodryadidae for the Aus- tralo-Papuan "hylids" and considered them to be derived independently from the Neotropi- cal hylids. Tyler (this volume) emphasized the lack of evidence for such an arrangement. In Australia, Cyclorana seems to be intermedi- ate between the Australian "leptodactylids" and "hylids" and may prove to establish a phylogenetic link between the two families on that continent. No such intermediates are Fig. 1:1. Major eco-physiographic regions of South America. Temperate regions: Austral forests (AF), Patagonia (PAT). Tropical evergreen forests: Amazonia (AM), Choco (CH), Atlantic coast (AC). Tropical and subtropical nonforests: Caribbean coastal desert (CD), Llanos (LL), Savannas (black), Caatinga (CA), Cerrados (CE), Gran Chaco (GC), Pampas (PA), Monte (MO), Espinal (ES), Matorral (MA), Atacama Desert (AD). Mountains (stippled): Andes (A), Guiana Highlands (G), Brasilian Highlands (B). Regiones ecofisiograficas mayores de Sudamerica. 8 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 known in South America, and even the mon- ophyly of the Neotropical hylids has been questioned. Maxson (1976) provided im- munological evidence that the phyllomedu- sines were not closely related to the other hylids. Within South America, the phyllomedusine hylids are widespread in Amazonia, Atlantic forests, and the Guianan and Brasilian shields. PhyUoinedusa is primarily South American, with only two species (one endemic) in Cen- tral America; the Central American Agalych- nis is represented by three species in the Choco and one endemic species in western Amazonia. The hemiphractines are restricted to northwestern South America with one spe- cies entering Central America (Trueb, 1974). The amphignathodontines are most speciose in northwestern South America (two species enter Central America ) , but with three genera on the Brasilian Shield and one in the Guiana Highlands. Among the hylines, all of the South American genera are endemic to the continent, except one species of Phrynohyas and several species groups of Hyla that enter Central America. Two species of the Middle American Smilisca enter South America. Assuming that at least the Neotropical hylids arose in South America, some stocks must have entered Central America by waif dispersal prior to the closure of the Panaman- ian Portal in the late Pliocene. These stocks were the ancestors of the several genera and species groups of Hyla endemic to Middle America. After the closure of the portal sev- eral groups dispersed northward into Central America (PhyUomedusa, Hemiphractus, Gas- trotheca, Phrynohyas, Hyla albomarginata, H. boans, H. bogotensis, H. leucophyllata, and H. rubra groups) and representatives of two Middle American genera (Agalychnis and Smilisca) dispersed into South America. Centrolenidae. — No fossils are known. Two genera and 46 species inhabit cloud for- ests in the Andes, whereas a few species occur on the Guianan and Brasilian shields and in Amazonia, the Choco, and Central America (Duellman, 1977). Obviously of South Amer- ican origin with Late Cenozoic dispersal into Central America, the relationships of the cen- trolenids usually are thought to be with the hylids, but no convincing evidence is avail- able. Ranidae. — Although no fossils are known before those in the Oligocene of North Amer- ica (Holman, 1968), the center of origin and dispersal of ranids quite clearly is in Africa (Savage, 1973), and presumably occurred af- ter the rift of South America and Africa in the Cretaceous. The single South American ranid, Rana palmipes, is widespread in Cen- tral America and must have entered South America after the establishment of the isth- mian link. The species is widespread in the tropical lowlands of South America. Microhylidae. — This large and diverse family presents one of the most controversial issues in anuran phylogeny and classification. Although the family is clearly of Gondwanan origin, the present interpretations of phylog- eny and biogeography are in conflict at times. Savage ( 1973 ) based his biogeography of the microhylids on Starrett's ( 1973 ) interpreta- tion of anuran phylogenv as demonstrated by larvae. Zweifel (1972), Lynch (1973), Sokol ( 1975 ) , and Tyler ( this volume ) provided compelling arguments based on diverse mor- phological, developmental, and biogeographic evidence against the Starrett and Savage model. All South American microhylids belong to the subfamily Microhylinae, which is shared with North America, tropical southeastern Asia, and the Malayan Archipelago. It is most logical biogeographically and phylogenetically that the Neotropical microhylids evolved in isolation in South America and that a stock that subsequently gave rise to the North American Ga.strophryne and Hypopachus managed to enter Central America from the south during mid-Cenozoic times; microhy- lines are known from the Miocene of Florida (Holman, 1967). The 16 genera of South American microhylids occur throughout the tropical and subtropical lowlands with the greatest diversity in the southern part of their range, particularly on the Brasilian Shield. Three genera (3 species) entered Central America after the connection of the conti- nents. The monotypic Geobatrachus in the Sierra Nevada de Santa Marta in northern Colombia tentatively was referred to the Microhylidae by Lynch (1971), but Duellman (1975) showed that this small frog has a combination of characters that precludes its assignment to 1979 DUELLMAN: SOUTH AMERICAN HERPETOFAUNA 9 any family as presently defined. Geobatra- chus has not been included in the numerical account of the microhylids. Rhinatrematidae. — Lacking fossils and en- demic to South America, these primitive cae- cilians are a sister group of the ichthyophiids (Nussbaum, 1977), a family restricted to In- dia, tropical southeastern Asia, Malayan Arch- ipelago, and the Philippines. Roth South American genera occur on the Guianan Shield and one (Epicrionops) has four species on the forested slopes of the northern Andes. Typhlonectidae. — The specialized aquatic caecilians are autochthonous to South Amer- ica, where they are distributed discontinuous- ly in the Caribbean and Amazonian lowlands and in the Parana Rasin. Caeciliidae. — The presence of a single fos- sil from the Paleocene of Rrasil (Estes and Wake, 1972 ) , possibly referable to this family, signifies a long history of caecilians in South America, where nine genera and 46 species now occur in the humid lowland tropics and forested slopes of the northern Andes. Four genera inhabit Middle America; the endemism there in Dermophis and Gymnopis indicates that a caecilian stock entered Central America prior to the establishment of the isthmian link, whereas the other two (Caecilia and Oscaecilia) are both widespread in South America, and evidently dispersed into Central America after the closure of the Panamanian Portal. The presence of caeciliids in tropical Africa (6 genera), India (3 genera), and the Seychelles Islands (3 genera), as well as in South America, is indicative of a widespread Gondwanan distribution prior to the Late Cretaceous. Pelomednsidae. — The classical, present Gondwanaland distribution pattern of pelo- medusid turtles is complicated by their occur- rence in Cretaceous deposits in Europe and North America and in the Eocene of Asia. However, the family has an extensive fossil record beginning in the Cretaceous in both South America and Africa (Wood, 1970). The single South American genus, Podocnemis, is widely distributed in the cis-Andean tropical lowlands. Chelidae. — Considered to be a derivative of the Pelomedusidae (Gaffney, 1975, 1977), the chelids are known are fossils only from Australia (early Tertiary to Pleistocene) and South America (Eocene to Pleistocene). The seven genera endemic to South America are widely distributed in the cis-Andean tropics, mostly in Amazonia. Kinosternidae. — Known as far back as the Oligocene in North America, the great ma- jority of kinosteraids (4 genera, 19 species) occur in North America and northern Central America. Only three species of Kinosternon occur in lower Central America; one of these also is widespread in cis-Andean South Amer- ica, and two vicariant species occur in the Choco. The kinosternids obviously are a post- portal entrant into South America from the north. Chelydridae. — The snapping turtles have an extensive fossil record throughout the Cenozoic in North America, where two genera are extant. One species of Chelydra inhabits lower Central America, and one occurs in the Choco in South America. Chelydra obviously is a recent immigrant into South America. Emydidae. — This family has an extensive fossil record in the Holarctic Region and to- day is distributed mainly in North America and the Oriental Region. The genus Chry- semys (=Pseudemys) is speciose in North America and in the West Indies, occurs in Central America (same species in northern South America), and is represented by anoth- er species in the Parana Rasin. Two species of the Central American Rhinoclemys occur in the Choco, and Rhinoclemys is known from the Pleistocene of Ecuador. Possibly an early Chrysemys stock waifed to South America, but Rhinoclemys and Chrysemys scripta cer- tainly entered the continent from the north subsequent to the establishment of the isth- mian link. Meiolaniidac. — This extinct family known from South America (Late Cretaceous to early Eocene) and Australia (Miocene to Pleistocene) seems to antedate the testudi- nids. Their fossil record suggests a Gondwa- nan history similar to that of the chelids. Testudinidae. — This family is cosmopoli- tan, except in the Australo-Papuan Region, and possibly had its initial radiation in Lau- rasia (Cracraft, 1974). The genus Geoche- lone occurs today in South America, Africa, India, southeastern Asia, and on the Gala- pagos Islands; it is known from the late Oligo- cene through the Pleistocene in South Amer- 10 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 ica, where it presently occurs throughout cis- Andean lowlands southward to northern Pata- gonia. Auffenberg ( 1971 ) suggested that the South American stock probably entered the continent from the north in the Oligocene; this implies waif dispersal. Trionychidae. — Although presently wide- spread in sub-Saharan Africa and in the Ori- ental Region, trionychids have an extensive fossil history in the Holarctic Region, where Trionyx occurs today in North America. Wood and Patterson ( 1973 ) reported a tri- onychid from the late Pliocene of Venezuela; probably this was a waif, because no other fossil or Recent trionychids are known south of northeastern Mexico. Gekkonidae. — Although the gekkonids were considered to be of uncertain geographic origin by Cracraf t ( 1974 ) , reevaluation of Kluge's (1967) phylogenetic scheme of the family suggests that the gekkonids are early Gondwanan. The primitive eublepharines oc- cur in Africa, southern Asia, and North Amer- ica. The diplodactylines are dominant in Australia and have dispersed onto islands in the southwest Pacific. Sphaerodactylines are restricted to tropical America. The gekko- nines are pantropical, being most diverse in the Indian, Oriental, and Ethiopian regions but also with many representatives in Aus- tralia and on Pacific islands. Of the 11 gen- era of gekkonines in South America, only eight are endemic to the American tropics, if the Old World species presently assigned to Phyllodactyhis are not considered to be con- generic, as suggested by Dixon and Anderson (1973). Except for the speciose Phyllodac- tyhis in dry habitats in western and northern South America (also Middle America, West Indies, and Galapagos Islands), most of the endemic genera are represented by only one or two species and all live in eastern South America, save for the monotypic Thccadacty- lus in western Amazonia, the Choco, Guianan Shield, Central America, and Lesser Antilles. The only genus with more than two species is Homonota (8 species) in temperate cis-An- dean areas; the related Garthia with two spe- cies occurs in the southern Atacaman Region. The other gekkonines in South America ( Gymnodactylus — 3 species, Hemidactylus — 5, and Lygodactylus — 2) are most speciose in Africa. 1 In fact, all but one of the species of Hemidactylus in America are widespread in Africa and elsewhere. The presence of gekkonids in the Paleo- cene of Brasil (Estes, 1970) suggests that gekkonines may have been present in South America prior to the separation of Africa and South America in the Cretaceous ( Bons and Pasteur, 1977). Trans-Atlantic waifing could explain the presence in South America of Af- rican genera, such as Gymnodactylus and Hemidactylus, but some of the latter most likely were transported by man. Of the five genera of sphaerodactyline geckos in South America, two (Coleodactylus and Pseudogonatodes) are endemic to for- ested cis-Andean regions in northern South America. Gonatodes is widespread in tropical lowlands, and one species has dispersed north- ward into Central America and the West In- dies. Lepidohlcpharis occurs in northwestern South America and has two species in Pana- ma. Sphaerodactylus is most speciose in the West Indies but with a few species in Central America, two of which extend into South America. Iguanidae. — This large and diverse family is first known in the fossil record from the Upper Cretaceous of Brasil (Estes and Price, 1973) and is diverse in the late Paleocene of Brasil (Estes, 1970). The earliest North American fossils are from the Eocene. Evi- dently an early iguanid stock reached North America prior to the separation of the conti- nents or waifed between the two. Informally, the iguanids are divided into five major groups (Etheridge, 1964, 1967). The sceloporines dif- ferentiated in North America and were paral- leled by an extensive radiation of tropidurines in temperate South America. The iguanines and basiliscines are primarily Middle Amer- ican with genera of the former endemic to the Galapagos Islands (Amblyrhynchus and Conolophus) and the West Indies (Cyclura); one species of Iguana and three of Basiliscus 'Smith et al. (1977) recognized the South American Lygodactylus as the sole representatives of the genus Vanzoia, but Kluge (pers. coram.) informed me that the South American species are perfectly good exam- ples of Lygodactylus. 1979 DUELLMAN: SOUTH AMERICAN HERPETOFAUNA 11 have entered northern South America. The anolines are widely distributed in the tropics of South America, Middle America, and the West Indies. Obviously waif dispersal be- tween South America and the emerging An- tilles and Central America during the Tertiary accounted for some of the patterns of distribu- tion, many of which have been masked by more recent dispersal after the closure of the Panamanian Portal. It certainly seems safe to assume that the origin of tropidurines and some anolines was in South America. The presence of two genera of iguanids on Madagascar has been interpreted as evidence for the former occurrence of the family in Africa with subsequent extinction there (per- haps owing to competition with agamids and chamaeleontids). The similarity of caudal structure of the Madagascaran iguanids to the tropidurines (Etheridge, 1967) and the pres- ence of iguanids in the Cretaceous of South America do not contradict that hypothesis. The iguanine Brachylophus, endemic to islands in the southwest Pacific, apparently is an example of long-distance rafting via the Trans-Pacific Current (Cogger, 1974). Teiidae. — Although presumed teiids are known from the Late Cretaceous in North America, those do not seem to be ancestral to living North American teiids, whereas late Paleocene teiids of South American resemble extant Neotropical genera (Estes, 1970). All living teiid genera occur in South America, where they are widespread throughout the continent, except for Patagonia and the aus- tral forests. Five genera have representatives in the West Indies, and 10 genera extend into Central America. With the exception of Cne- midophorns, which is widespread and spe- ciose in North America, all other teiids prob- ably arrived in Central America after the formation of the isthmian link. Scincidae. — Although skinks are known from the Paleocene of Rrasil and the Late Cretaceous of North America (Estes, 1976), only four genera presently occur in the Amer- icas; this is only a small fraction of the family containing at least SO genera and more than 1,000 species (Greer, 1970). Occurring throughout South America, except for the Andes and cool temperate regions, are eight species of Mabuija, a genus containing about 75 additional species in Africa, Madagascar, southern Asia, and the Pacific islands, plus an additional two species in Central America and the Antilles. Tihen (1964) suggested an Eur- asian origin of the North American skinks (Eumeces and Scincella), but an African ori- gin of the South American Mabmja seems to be reasonable. Anguidae. — Presently distributed primar- ily in North America, western Eurasia, and southeastern Asia, the anguids have an ex- tensive fossil history dating from the Late Cretaceous in North America (Meszoely, 1970). Two genera occur in South America. Diploglossus is speciose in Central America and the West Indies, and one of the South American species is shared with Central America. The endemic South American Ophi- odes (4 species) in the south-central part of the continent apparently evolved from an anguid stock that entered South America from the north in the Cenozoic. Amphisbaenidae. — Numerous fossil am- phisbaenians are known from the Paleocene to Miocene in North America, Eocene to Plio- cene of Europe, and Oligocene of Mongolia. With the exception of Bipes and the Pale- arctic Blanus, the amphisbaenids (sensu Ber- man, 1973) are Neotropical and African — 10 genera in Africa (one ranging into Europe) and six in South America (plus one endemic to Cuba). The generic and specific differen- tiation in Africa (10 genera, 52 species) and South America (6 genera, 45 species), and the possible presence of Amphisbaena in both Africa and South America (Gans, 1967), plus 10 species endemic to the West Indies and one South American species extending into Central America, are suggestive of an African- South American amphisbaenid interchange. However, the place of origin of the amphis- baenids still remains problematic. Anomalepidae. — Unknown as fossils, 17 of the 20 species and all four genera of anoma- lepids occur in South America. One species each of Anomalepis, Helminthophis, and Liotyphlops occurs in Central America, and another species of Liotyphlops ranges from Costa Rica into northern South America. On the basis of present distributions, the anoma- lepids seem to be a South American group that only recently invaded Central America. 12 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 In South America the family is widespread in trans-Andean and eis-Andean tropical low- lands. Leptotyphlopidae. — The single genus in this family containing about 64 species is widespread in tropical and subtropical South America, Middle America, southwestern United States, Africa, and southwestern Asia. No fossils are known. On the basis of present distribution it is reasonable to suggest that the leptotyphlopids had a western Gondwa- naland origin and subsequently spread north- ward into Central and North America and independently into Asia. Typhlopidae. — These fossorial snakes are known from two genera ( Typhlina, 33 species in Australia, New Guinea, Solomon and Fiji islands) and Typhlops (about 114 species throughout tropical and subtropical parts of the world, except Australia). Only three spe- cies occur in South America; another five are in Central America, and 16 occur in the West Indies. Thus, with respect to the total differ- entiation of the family, the Neotropics are poor in typhlopids. The only fossils (Eocene- Miocene of Europe) are of no help in inter- preting the paleobiogeography of the group. In the absence of any evidence for the occur- rence of typhlopids in North America, a west- ern Gondwanaland origin for the Neotropical stocks might be suggested. Boidae. — Represented by an extensive, world-wide fossil record from the Upper Cre- taceous through the Eocene (only Pleistocene in Australia ) , the boids seem to have been the dominant snakes throughout the world in the Early Cenozoic. Evidently they had dispersed widely before the breakup of Pangaea. The Early Cenozoic South American boid Madtso- ia also is known from the Late Cretaceous of Madagascar, and a related boid, Wonambi, is known from the Pleistocene of Australia (Ty- ler, this volume). In some respects the dis- tribution of these fossils parallels that of liv- ing boines — eight genera in the Neotropics, two in Madagascar, one in New Guinea and islands in the southwest Pacific. Presently the family is widespread in tropical South Amer- ica and especially diverse in Amazonia. Aniliidae. — These problematic fossorial snakes have a long history in northern conti- nents, dating from the Middle Cretaceous in North America and the Eocene of Europe. A Late Cretaceous snake, Dinilysia from Pata- gonia, is considered to be related to aniliids ( Rage, 1977 ) , and hue aniliids were reported from the Eocene of Brasil by Baez and Gas- parini (this volume), who support Cracraft's ( 1974 ) contention that aniliids are of Laura- sian origin. Nonetheless, entry into South America possibly was by way of Africa. The one living South American aniliid is wide- spread in cis-Andean tropical lowlands. Tropidophiidae. — Structurally, the tropi- dophiids are intermediate between the boids and colubroid snakes. Tropidophis has three widely dispersed species in South America (northern Andes and southeastern Brasil) and 12 species in the West Indies. Trachyboa and Ungaliophis occur in the Choco and Central America. The tropidophiids are considered to be of South American origin with subsequent northward dispersal. Colubridae. — The poor fossil record and taxonomic chaos of the colubrids (sensu lato) permit only the most general comments to be made about this immense and important fam- ily. My use of subfamilial designations fol- lows that of Dowling (1975) but eliminates some apparent misapplications not especially germane to the Neotropical colubrids. A brief summary of the colubrid snakes follows. 1. Xenodontinae: 93 genera, about 570 species. Sixty genera occur in South America; 22 of these are shared with Central America. Seven genera are endemic to the West Indies; 13 are restricted to North America (north of the Isthmus of Tehuantepec in southern Mex- ico), and 13 are Middle American. Included in this group of rear-fanged genera are fossor- ial, terrestrial, aquatic, and arboreal snakes. Most arboreal xenodontines are nocturnal, and few (Alsophis, Dromicus, Leimadophis, Ly- gophis) are diurnal racer-like snakes. Most xenodontines feed on frogs, lizards, or other snakes. 2. Lycodontinae: 79 genera, about 285 species. These rear-fanged snakes are distrib- uted primarily in the African and Oriental regions and peripherally in northern Australia and the Palearctic Region. 3. Colubrinae: 74 genera, about 440 spe- cies. Widespread in the Holarctic Region, some genera occur in the Ethiopian and Ori- 1979 DUELLMAN: SOUTH AMERICAN HERPETOFAUNA 13 ental regions, and two genera reach Australia. Thirty-two genera occur in the New World; of these, 12 are in South America, but only one of those (Drymoluber) does not occur in Central America. Most of the colubrines in South America (Chironius, Dendrophidion, Dn/marchon, Drymobius, Drymoluber, Lep- tophis, Masticophis, Mastigodryas, Spilotes) are diurnal racer-like snakes that are terres- trial or arboreal. 4. Natricinae: 34 genera, about 170 spe- cies (excluding from the Natricinae the Old World snakes more appropriately referred to Acrochordidae and Homalopsinae). The na- tricines are widely distributed in the Holand- ric and Oriental regions, with a few represen- tatives in Africa and one in northern Aus- tralia. Nine genera of natricines occur in North America, with Thamnophis extending to Costa Rica. Even if these subfamilial groups are mon- ophyletic, the historical biogeography of the colubrids still remains shrouded. It is evident from the distributions of the subfamilies that centers of dispersal (and perhaps of origin) can be ascertained, but ancestors cannot. Ap- parently the xenodontines evolved in South America and the lycodontines in the African- Indian-Asian Arc and had corresponding par- allel radiations in the New World and Old World, respectively. Colubrines and natri- cines probably are Holarctic in origin. If these truly are the centers of origin and dis- persal, it is possible to make some reasonable generalizations about the South American colubrid fauna. 1. Although colubrids are known from the Eocene (Rage, 1975a,b), they be- came dominant in the upper Miocene (Holman, 1976), when they first ap- pear in South America (Baez and Gasparini, this volume). 2. Xenodontines evolved in South Amer- ica, where they are the dominant snakes today and the only colubrids in the southern part of the continent. 3. Some xenodontine snakes were present in Middle America in the Cenozoic; these differentiated, and some of them dispersed as far as northeastern North America. 4. Colubrine stocks in North America in- vaded Central America, and some of them may have waifed to South Amer- ica before the closure of the Pana- manian Portal in the late Pliocene. 5. With the closure of the Panamanian Portal in the late Pliocene there was an interchange of South American xenodontines northwards and Middle American colubrines and xenodontines southward. 6. Natricines never have extended farther south than Central America. 7. The West Indian colubrid fauna is com- posed of xenodontines derived either from Central or South America (some widespread colubrine species are recent immigrants into the Lesser Antilles ) . Micruridae. — Formerly associated with the Elapidae, micrurids have been shown to be an independently derived group from Ela- pomorphus-Apostolepis xenodontines in South America (Savitzky, 1978). An upper Miocene fossil from Nebraska (Holman, 1977), to- gether with the presence of Micruroides and numerous species of Micrurus in Central America, is suggestive of dispersal of micru- rids into Central America in the Cenozoic with additional interchange after the closure of the Panamanian Portal. Micrurus has 27 endemic species in South America and ranges throughout the lowlands and moderate eleva- tions south to northern Patagonia; the mono- typic Leptomicrurus is restricted to Amazonia. Viperidae. — Although the vipers are pri- marily an Old World group, which apparently originated in the Palearctic Region (Marx and Rabb, 1965), one lineage — the crotaline vipers — may have evolved in the Oriental Re- gion and dispersed via Beringia to North America (Burger, 1971). The earliest North American fossil crotalines are of Miocene age (Holman, 1977). The presence of the pre- sumably primitive crotaline Lachesis muta in northwestern South America and lower Cen- tral America suggests that the crotaline vipers entered South America after the establishment of the isthmian link. On the other hand, the presence of many species of Bothrops throughout South America as far south as Patagonia, as well as many different species in Middle America, is suggestive of an earlier dispersal into South America. Bothrops is 14 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 known as a fossil in South America only from the Pleistocene of Bolivia. The presence of Bothrops in the West Indies attests to their abilities of over-water dispersal. Crotalus cer- tainly is a post-Pliocene immigrant into South America, where it has spread throughout non- forested tropical areas. Crocodylidae. — Present distributions and fossil records indicate that the alligatorines and Crocodyhts entered South America from the north ( Sill, 1968; Baez and Gasparini, this volume), although Paleocene crocodylines in South America are suggestive of possible Afri- can derivation. Likewise, the presence of gavial-like crocodilians in South American de- posits (Baez and Gasparini, this volume) im- plicates at least the early crocodilians in a Gondwanan distribution. Possibly the croco- dilians presently living in South America were derived from North American stocks, whereas the ancient Gondwanan crocodilians are ex- tinct in South America. Obviously, entry of crocodilians into South America from the north prior to the formation of the isthmian link was facilitated by their abilities at tra- versing open water. Only Caiman is wide- spread throughout the tropical lowlands of South America; Crocodijlus is restricted to the northern part of the continent (one spe- cies endemic to the llanos). The other two genera (Melanosuchus and Paleosnchus) are in western Amazonia. EXTRA-CONTINENTAL RELATIONSHIPS Elsewhere in this volume detailed com- parisons of the origins of the African and Australian herpetofaunas with respect to that of South America have been made by Lau- rent and Tyler, respectively. In this section I compare the compositions and taxonomic di- versities of those three faunas. Furthermore, I provide a discussion of the herpetofaunal relationships between South America and North America and between South America and the West Indies. Herpetofaunas of Gondwanan Continents South America contains 37 living families of amphibians and reptiles, Africa 26, and Australia 17. Among the amphibians only three families are shared by the three conti- nents — Leptodactylidae (only one genus with three species in Africa), Ranidae (only one species each in South America and Australia), and Microhylidae (only two genera with seven species in Australia). Two additional families are shared by South America and Africa (Pi- pidae and Bufonidae) and one (Hylidae) by South America and Australia. Among the nonmarine reptiles, six families are common to the three continents — Crocodylidae, Gek- konidae, Scincidae (only one genus with eight species in South America), Typhlopidae, Boi- dae, Colubridae (only four genera with six species in Australia). Five additional families are shared by Africa and South America (Pe- lomedusidae, Testudinidae, Amphisbaenidae, Leptotyphlopidae, and Viperidae). Three other families are shared by Africa and Aus- tralia (Agamidae, Varanidae, Elapidae), whereas only one other (Chelidae) is shared by South America and Australia. Faunal re- semblance factors (Duellman, 1966) at the family level are highest between Africa and Australia (0.56), followed by Africa and South America (0.54) and Australia and South America (0.40). South America has five endemic families of amphibians plus two others that have dispersed only to lower Cen- tral America, but no endemic families of reptiles. Africa has one endemic family of lizards (Cordylidae), one of caecilians (Sco- lecomorphidae), and one of frogs that has dispersed to Madagascar and the Seychelles Islands ( Hyperoliidae ) . Australia has no en- demic families, but the Pygopodidae is shared only with New Guinea. Examination of the amount of taxonomic diversity in anurans and lizards on each conti- nent reveals that the South American anuran fauna is much more diverse than that on the other continents but that Africa has the most species of lizards.- Analyses of diversity in- L ' Data for Australia were gathered primarily from Cogger ( 1975); for African lizards chiefly from Mer- tens (1963, 1966), Wermuth (1965, 1967, 1968), and Greer (1970, 1974); for African frogs and South American frogs and lizards from my personal compilations. Owing to the absence of modern com- prehensive works on African snakes, a complete com- pazine analysis of snakes was not attempted. South America has 9 families, 96 genera, and 556 species of snakes, compared with 4 families, 36 genera, and 104 species of nonmarine snakes in Australia. The species/area values for snakes are 31.2 for South America and 13.5 for Australia. 1979 DUELLMAN: SOUTH AMERICAN HERPETOFAUNA 15 eluded the differentiation of genera and spe- cies on each continent and the numbers of species per unit area (Table 1:6). The taxonomic diversity among South American frogs is extremely high in two fami- lies — Leptodactylidae and Hylidae; these con- tain 64 percent of the genera and 73 percent of the species of South American frogs. These same two families are the dominant compo- nents of the Australian amphibian fauna, ac- counting for 88 percent of the genera and 95 percent of the species. The dominant African families are Bufonidae and Ranidae, together accounting for 44 percent of the genera and 73 percent of the species. In South America the dominant families of lizards are the Iguanidae and Teiidae (com- bined, 77% of genera, 83% of species). In Australia the Scincidae alone accounts for 37 percent of the genera and 54 percent of the species, whereas the Gekkonidae and Agami- dae are secondary (combined 48% of the gen- era, 32% of species). Gekkonids and scincids are the most diverse African families (60% of genera, 56% of species) followed by cordylids and lacertids ( 35% of genera, 28% of species ) . The presence or absence of families on the three continents relates primarily to historical factors, whereas the diversity within families may be dependent upon the amount of time that the family has occupied the continent or perhaps also the size of the area. Further- more, ecological factors may be extremely important in the evolutionary diversity of a family provided that the family has been established on the continent for a sufficient period of time. There is no easy or objective way to measure habitat diversity within and between the three continents. One method is to compare the sizes of the continents with respect to taxonomic diver- sity. Africa is by far the largest of the con- tinents (30,264,000 km 2 ) followed by South America (17,793,000 km 2 ) and Australia (7,687,000 km 2 ). Analyses of numbers of species per unit area show that South Amer- ica has an excessive number of frogs and Australia an excessive number of lizards; all other values are negative (Table 1:6). Lizards usually are most diverse and nu- merous in xeric areas, and the vast majority of Australia is xeric. Likewise, most of Africa is arid, and this is reflected in the large number of lizards (506 species) on that continent. Table 1:6. — Taxonomic Diversity of Anurans and Lizards on Gondwanan Continents. South America Africa Australia Anurans Families 11 98 996 8.9 7 63 355 9.0 4 Genera ... 26 Species 155 Genera/Family 6.5 Species/Genus 10.2 5.6 6.0 Species/Family 90.5 50.7 38.8 Species/million km" . 56.0 11.7 20.1 Deviation from expected .. . +26.7 -17.6 -9.2 Lizards Families .. 5 84 471 16.8 7 78 506 11.1 5 Genera 54 Species 364 Genera/Family 10.8 Species/Genus 5.6 6.5 6.7 Species/Family 94.2 72.3 72.8 Species/million knr 26.5 16.7 47.3 Deviation from expected _ -3.7 -13.5 + 17.1 However, proportional to size, Africa has fewer species of lizards than either Australia or South America. The comparatively few lizards in Africa might be related to the vast areas of extreme deserts (Sahara and Kala- hari), which although inhabited by some spe- cialized lizards are not species rich. Another factor might be the presence of herds of large mammals on the plains (Janzen, 1976). Fur- thermore, Pianka (1971) and Pianka and Huey (1971) suggested that lower species di- versity of lizards in the Kalahari Desert, as compared with central Australian deserts, may be a result of the comparatively richer avian fauna in the Kalahari. Frogs are most diverse in humid tropical forests. By comparison with the South Amer- ican forests, the lowland tropical forests in Africa and Australia are much smaller; the Congo Basin has about 2,000,000 km 2 of rainforest, whereas the Amazon Basin has 4,500,000 km 2 (Richards, 1973). Further- more, the South American lowland tropical rainforests are in three distinctly separate units — Amazonian, trans-Andean, and Atlan- tic coastal, each harboring 203, 111, and 168 endemic species of amphibians, respectively (Lynch, this volume). Quaternary climatic- vegetational changes in west African rainfor- ests (Moreau, 1963, 1969) apparently resulted in the elimination of proportionately more of the lowland tropical forests there than in South America (Haffer, 1974). Montane rain- forests (cloud forests) are much more exten- 16 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 sive in South America than in Australia, and especially, in Africa. In South America these forests support rich, localized anuran faunas. The comparative herpetofaunal diversity in the three Gondwanan continents first is the result of historical components that were ei- ther present on the continents when they were formed or emigrated there after the continents became discrete units. However, the taxonomic diversity is primarily a factor of habitat diversity. In comparison with the other continents, South America offers a much more diverse landscape, climate, and vegeta- tion. The cool temperate rainforests, Patagon- ian steppes, and high Andean punas and paramos stand in marked contrast to the pampas, llanos, caatinga, and Atacama Desert, which in turn harbor distinctly different bi- otas than do the lowland and montane rain- forests. Herpetofaunal Comparisons of North and South America It has been nearly half a century since Dunn's (1931) then classic essay on the herpetofauna of the Americas; Dunn's ap- proach was based entirely on Matthew's (1915) hypothesis of northern origin and southward dispersal of mammalian orders. Savage (1966) provided a well-documented account of the distribution patterns of am- phibians and reptiles in Central America and emphasized the degree of differentiation and endemism in that fauna by recognizing a Mesoamerican herpetofauna distinct from the Nearctic and Neotropical faunas. Although the patterns delineated by Savage are realistic, the interpretation of the origins and times of dispersal can now be modified by new paleon- tological and geomorphological information. Of primary importance to a biogeographic- al analysis of the Central-South American re- gion is the history of the connection of the two major continental masses. According to Dietz and Holden (1970), after the initial breakup of Pangaea in the Early Jurassic ( =T80 m.y.b.p. ) there was no direct land connection between North America and South America until the Late Tertiary, although the positions of the two continents converged beginning in the mid-Cretaceous. An island arc, the proto- Antilles, existed between nuclear Central America and South America in the Cretaceous and Early Tertiary (Holden and Dietz, 1972; Malfait and Dinkelman, 1972); this arc moved eastward, relative to the westward drift of the American continents, through the Tertiary and formed the present Lesser Antilles. The region of lower Central America (Costa Rica and Panama), or the isthmian link, formed as a volcanic archipelago in the Oligocene; addi- tional land emerged, and the archipelago co- alesced with nuclear Central America 10-12 m.y.b.p. and finally with South America about 5.7 m.y.b.p. During the late Mesozoic ( 180-90 m.y.b.p.) South America had direct land con- nections with Africa (Grant, 1971; Reyment and Tait, 1972; Larson and Ladd, 1973) and with Australia via Antarctica until the Eocene or Oligocene («50 m.y.b.p.) (McGowran, 1973; Veevers and McElhinny, 1976). Throughout the Cenozoic until the late Plio- cene (5.7 m.y.b.p.) there was no land con- nection with North and Central America. Thus, for about 45 million years South Amer- ica was isolated from other continents. How- ever, the island arc ( proto- Antilles ) in the Late Cretaceous and Early Tertiary and the Central American Archipelago in the Middle to Late Tertiary provided opportunities for limited faunal exchange between the conti- nents. Fossil evidence (albeit scanty or non- existent for some groups) and present pat- terns of distribution and speciation provide a basis for analysis of the herpetofaunal inter- change between Central America and South America (see preceding review of families and Savage, 1966). Examination of the family groups of amphibians and reptiles that have entered into the exchange (Table 1:7) shows two modes of entry. The first is by means of dispersal across one of two archipelagos — the early proto-Antillean island arc or the later Central American Archipelago. The second is by direct dispersal after the establishment of the isthmian link. My analysis shows no amphibians entering South America from the north via the archi- pelago but five amphibian family groups dis- persing northward via the island route. Pre- 1979 DUELLMAN: SOUTH AMERICAN HERPETOFAUNA 17 Table 1:7. — Postulated Herpetofaunal Exchange Be- tween South America and Central America. Across Via Family Panamanian Isthmian group Portal Link N^SS ->~N N -> S S -> N Plethodonhdae — — + Pipidae — — ~ Eleutherodactylini — + +? + Leptodactylinae — — — + Bufonidae — — + + Dendrobatidae — — — + Phyllomedusinae — + + + Hemiphractinae — — — + Amphignathodontinae — — — + Hylinae — + + + Centrolenidae — — — + Ranidae — — + — Microhylidae — + — + Caeciliidae — + + + Kinostemidae — — + — Chelydridae — — + — Testudinidae + — — — Gekkoninae — + — + Sphaerodactylinae — + + + Iguanidae (primitive) — + — — Iguanines + — + — Basihscines — — + — Anolines — + + + Teiidae — + + + Scincidae — — — + ■ Anguidae + — + ~ Amphisbaenidae — — — + Anomalepidae — — — + Leptotyphlopidae — + — — Typhlopidae — + — — Tropidophiidae — — — + Xenodontinae — + + + Colubrinae +? + +? Micruridae — +? +? + Crotalinae +? - + +? Crocodylidae + — + + sumably all of these (Eleutherodactylini, Phyllomedusinae, Hylinae, Microhylidae, Caeciliidae ) dispersed northward via the Cen- tral American Archipelago, which emerged in the Oligocene. However, the Hylinae may have dispersed earlier via the proto-Antilles, for hyline frogs are known from the Oligocene in North America, have had an extensive radi- ation in North America and nuclear Central America, and have dispersed into Eurasia (presumably via Reringia). Also, it is pos- sible that a proto-pipid frog entered North America via the proto-Antillean arc; this frog could be the ancestor of the Rhinophrynidae now restricted to Mexico and nuclear Central America but known from the Paleocene-Oli- gocene of North America. Reptilian dispersal via the islands appar- ently was much more extensive than that of the amphibians. Probable dispersers via the proto-Antillean island arc are primitive igua- nid lizards ( south to north ) and anguid lizards (north to south). Dispersal via the Middle to Late Tertiary Central American Archipelago included testudinids, iguanines, crocodylids, and perhaps some colubrines and crotalines ( all north to south ) and gekkonines, anolines, teiids, leptotyphlopids, typhlopids, xenodon- tines, and perhaps sphaerodactylines and mi- crurids (all south to north). The iguanine dispersal is postulated for the migration of an Ambhjrhynchus-Conolophus stock to the Galapagos Islands from the South American mainland, but possibly this stock waifed di- rectly from Central America ( Avery and Tan- ner, 1971). Colubrine southward dispersal probably was late in the history of the archi- pelago, if indeed these snakes did enter South America prior to the isthmian link. The evo- lution of the alpha and beta groups of Anolis north and south of the isthmus bespeaks the separation of the two groups on the two land masses (Etheridge, 1959; Savage, 1966). Pos- sibly crocodilians and gekkonines also dis- persed via the proto-Antillean island arc. Overland dispersal after the establishment of the isthmian link involved more northward than southward dispersal by amphibians, but it did permit entry into South America for the first time of plethodontid salamanders (2 gen- era) and ranid frogs (1 species), all wide- spread taxa in Central America. Other groups moving southward were some phyllomedusine and perhaps some eleutherodactyline frogs that were part of the Mesoamerican fauna evolved from South American stocks that earlier had invaded Central America. Also, a member of the Bufo valliceps group (B. coniferus) invaded South America. The north- ern infusion of South American taxa includes some groups that have speciated (S) and/or dispersed widely (D) in Central America — Eleutherodactylus (SD), Leptodactylus (SD), Physalaemus (D), Bufo marinus (D), Hyla ebraccata (D) and microcephalia (SD) groups, and Centrolcnella (SD). Most of the other South American amphibians have dis- 18 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 persed only into lower Central America and have not speciated there — Pipa (eastern Pan- ama); Pleurodema, Bufo typhonius, Elachis- tocleis, Relictivomer, Caecilia, and Oscaecilia (central Panama); Hemiphractus and Gastro- theca (western Panama); Glossostoma (Costa Rica); and Bufo haematiticus (Nicaragua). The South American Phyllomedusa buckleiji group has a species endemic to lower Central America (Duellman, 1970). All three genera of South American dendrobatids occur in low- er Central America; each has undergone some speciation (Savage, 1968). The Atelopus vari- us group has invaded Central America (to Costa Rica) and has undergone a bewilder- ing diversification (Savage, 1972). Late Tertiary and Quaternary overland dispersal amongst reptiles also was extensive. Southward dispersal brought chelydrid and kinosternid turtles, basiliscine lizards, and possibly colubrine and crotaline snakes into South America for the first time, whereas si- multaneously the first Mabuija, Amphisbaena, and anomalepid and tropidophiid snakes reached Central America. Mesoamerican groups originally derived from South Amer- ican stocks (beta anoles, Cnemidophorus, and many xendodontine snakes) dispersed into South America. Many South American taxa (Thecadactylus, alpha anoles, Ameiva, micro- teiids, and xenodontine and micrurid snakes) moved northward. The herpetofauna of eastern Panama con- tains many genera that are chiefly South American — Pipa, Rhamphophryne, Hemi- phractus, Gastrotheca, Elachistocleis, Caeci- lia, Geochelone, Lepidoblepharis, Enyalioides, Echinosaura, Amphisbaena, CoraUus, Trachy- boa, Atractus, Diaphorolepis, and Pseudoboa. The herpetofauna of the Chocoan lowlands of northwestern South America contains many species that are familiar to the herpetologist working in lower Central America, whereas the fauna in the Amazon Basin is greatly dif- ferent (at least at the species level); com- pare data given by Savage (1966) with those presented by Lynch ( this volume ) and Dixon (this volume). Herpetofauna of the West Indies Although the West Indies are peripheral to a discussion of the South American herpeto- fauna, it is germane to this essay to ascertain the herpetofaunal relationships of the two regions inasmuch as many genera and some species are common to the two. The history of the Caribbean Plate and the tectonic move- ments in the Antillean-Caribbean region have not been resolved, but Rosen (1975) sum- marized (and extended) the existing geologi- cal data and proposed a plausible vicariance model of Caribbean biogeography. Excluding introduced taxa, the herpeto- fauna of the West Indies ( not including Trini- dad, Tobago, Bonaire, Aruba, and Curacao) consists of 505 species in 57 genera ( Schwartz and Thomas, 1975 ) ; 476 of the species and 18 of the genera are endemic to the West Indies. Schwartz ( 1978 ) gave a brief description of the herpetogeography of the West Indies. Twenty-two genera of reptiles are primar- ily mainland taxa having one or two species extending into the West Indies. Fifteen of these are South American taxa that extend into the Lesser Antilles — Phyllodactylus (also Greater Antilles), Thecadactylus, Iguana, Ba- chia, Cnemidophorus, Gymnophthabnus, Kcn- tropyx, Mabuya (also Greater Antilles), Boa, CoraUus, Chironius, Clelia, Mastigodryas, Pseudoboa, and Bothrops. Five are Central American taxa that extend into the Greater Antilles — Gonatodes, Tretanorhinus, and Ctenosaura, Boa, and Coniophanes only reach- ing Isla San Andres and/or Isla Providencia. Three are North American taxa that reach Cuba — Natrix fasciata, Kinosternon bauri, and Crocodylus acutus; the latter also has in- vaded the Lesser Antilles from South Amer- ica, and there is an endemic species of Croco- dylus on Cuba. The geckos Tarentola and Hemidactylus may have arrived by waifing from any one of many sources, although the other species of Tarentola are circum-Medi- terranean. Two of the Leptodactylus and one each of Eleuthcrodactylus and Hyla are main- land species. The five west Indian Chrysemys probably stem from an invasion from North America. Among the endemic or taxonomically rich genera in the West Indies, the hylid genera (Calyptaliyla, Osteopilas, and Hyla) were studied by Trueb and Tyler (1974), who in- ferred five invasions of the Greater Antilles 1979 DUELLMAN: SOUTH AMERICAN HERPETOFAUNA 19 by separate hylid stocks, probably from South America by rafting; however, it is possible that the stocks for some of these were on the proto-Antilles and drifted part of the way to their present positions. The same might be true for the monotypic Cuban eleutherodac- tyline Sminthillus and some of the West In- dian stocks of Eleutherodactylus. Lynch (1971) suggested that most of the West In- dian Eleutherodactylus, plus Sminthillus and the Mexican Syrrhophus and Tomodactylus possibly represented one eleutherodactyline lineage and that the Eleutherodactylus inop- tatus group of Hispaniola and the mainland Eleutherodactylus formed another lineage. If these suppositions are correct, minimally two eleutherodactyline invasions of the West In- dies are required. The nine Greater Antillean Bufo seem to be related (Schwartz, 1972), but their affinities with mainland taxa have yet to be determined. Amongst the lizards, the dominant genus is Anolis, represented by two groups of spe- cies (alpha and beta, fide Etheridge, 1959), plus two endemic genera (Chamaeleolis and Chamaelinorops in Cuba and Hispaniola, re- spectively). The alpha anoles inhabit the Greater and Lesser Antilles and are wide- spread in South America, whereas the beta anoles occur in Central and South America and the Greater Antilles. Williams ( 1969 and in Trueb and Tyler, 1974 ) required minimally two invasions of the Greater Antilles by Anolis and two for the endemic genera. Cyclura is related to Ctenosaura of Middle America (Avery and Tanner, 1971); the ancestral stock of Cyclura presumably arrived in the Greater Antilles from Central America. This also probably is true for the ancestral xantusiid stock that gave rise to Cricosaura endemic to Cuba (Savage, 1963) and that of Diploglossus represented by some Central American and many West Indian species. Possibly an earlier or separate invasion was responsible for the endemic Hispaniolan anguid Wetnwrena. Two of the speciose Antillean genera seem to be of South American origin — Ameiva and Leiocephalus. Etheridge ( 1966 ) showed Leiocephalus to be a tropidurine related to Liolacmus (restricted to temperate South America). Sphaerodactylus, with 56 species in the West Indies, may have evolved there from an early sphaerodactyline invasion; if so, the few mainland species (Central America and Choco) are the result of dispersal of stocks back to the mainland. All of the colubrid snakes ( save the North American Natrix fasciata) in the West Indies are xenodontines. Maglio (1970) demon- strated relationships of the seven endemic genera and Alsophis with diverse mainland xenodontines and concluded that four sepa- rate xenodontine invasions of the West Indies from either Central or South America were necessary in the evolution of the West Indian xenodontines. Presumably the tropidophiid stock that gave rise to the 12 species of Tropi- dophis in the Greater Antilles and the Ba- hamas came from South America, perhaps via Central America. Too little is known about the relationships of the Antillean Aristeltiger, amphisbaenids, Leptotyphlops, and Typhlops to speculate on their origins, except that it is unlikely that they invaded from North America. CONTINENTAL PATTERNS OF DISTRIBUTION It is becoming increasingly evident that the patterns of climate and vegetation have changed drastically in South America since the beginning of the Cretaceous. Axelrod ( 1972 ) argued convincingly that the interior of the large African-American continent was arid prior to the birth of the South Atlantic Ocean, which brought maritime and mesic climates to western Africa and eastern South America for the first time. Some elements of the arid-adapted west Gondwanan flora sur- vived in South America (and Africa) (Sol- brig, 1976; Sarmiento, 1976), while much of the continent was mesic. Subsequent to the Eocene, temperate South America gradually became cooler and drier ( Axelrod and Bailey, 1969; Wolfe, 1971; Baez and Scillato Yane, this volume). The uplift of the Andes begin- ning in the Miocene drastically modified wind patterns and resulted in great changes in cli- mate and vegetation (Simpson, this volume), and the formerly widespread Tertiary-Chaco Paleoflora (Solbrig, 1976) became fragmented on the Pacific slopes as the climatic effects of 20 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 the developing Humboldt Current resulted in desiccation of the land in the Late Tertiary (Jeannel, 1967). Pleistocene climatic fluctua- tion effected the entire continent with cool and warm periods in the south (Baez and Scillato Yane, this volume), humid and dry periods in the lowland tropics (Haffer, this volume ) and extensive glaciation in the Andes (Simpson, this volume). Although the fossil record of the herpeto- fauna in South America is still fragmentary, sufficient material exists, especially when placed with the better data from mammals and the paleofloras, to give a faint impression of past distributions, especially in the southern part of the continent (Baez and Gasparini, this volume). It is evident that there has been a northward retreat of the tropical biota, espe- cially those types requiring mesic environ- ments. Conceivably much of the present arid- adapted temperate heqietofauna has evolved rather recently in response to increasingly xeric conditions, as postulated for xeric floras by Axelrod ( 1967 ) . Thus, the archaic frogs in the austral forests are relicts, like the forests themselves ( Vuilleumier, 1968; Lynch, 1971). The fossil record is especially secretive about the presently large and diverse herpeto- fauna of the tropical forests. Presumably most of this fauna evolved at the generic level by the mid-Tertiary, or at least by the Pliocene. Endemic Andean groups apparently evolved with the uplift of the Andes and probably are not older than the Pliocene. Speciation in many lowland tropical groups (Haffer, 1974, this volume) and Andean groups (Simpson, 1975, this volume; Duellman, this volume) seems to have occurred in the Pleistocene. Thus, we are faced with contrasting pic- tures — presumed recent speciation and appar- ently rapid evolutionary rates in the lowland tropics and in the Andes, as well as in some temperate groups adapted to xeric conditions, versus the survival of many old taxa in habitat refugia in the austral forests and also in the ancient Brasilian and Guianan highlands (Hoogmoed, this volume). Various contributors to this volume have analyzed distributions within certain regions (e.g., Patagonia) or biotopes (e.g., lowland tropical rainforests); here I attempt to provide a broad synthesis of patterns in the entire continent. Data for many groups and/or re- gions are inadequate for a detailed analysis; instead I present a general picture of the di- verse distribution patterns and give examples of each. Temperate Herpetofaunas Austral Forests. — The cool, moist forests of southern Chile and adjacent Argentina repre- sent an unique biotope in South America, characterized by a highly endemic herpeto- fauna composed mostly of primitive lepto- dactylid frogs ( Alsodes, Batrachyla, Caudi- verbera, Eupsophus, Hylorina, Tehnatobufo) mostly restricted to forests south of 37°S Lat (Fig. 1:2A). The distributions of all of the species are mapped by Formas ( this volume ) . The herpetofauna of the austral forests mostly is relictual and presumably consists (at least in amphibians) of remnants of groups that were widespread in temperate South America in the Early Tertiary. With few minor excep- tions (Bufo spinidosus, Tachymenis peruvi- ana), none of the species extends beyond the present limits of the region, but some species of Liolaemus and Pleurodema have congeners in adjacent regions. The Atacama Desert to the north and the Patagonian steppe to the east are effective barriers to the dispersal of groups inhabiting the austral forests. Patagonian Steppe. — The cool, dry steppes of southern Argentina interdigitate in the north with the monte ( Cei, this volume ) . The Patagonian herpetofauna contains some an- cient relicts ( telmatobiine leptodactylid frogs and some tropidurine iguanid lizards) but also many species of Liolaemus that have dif- ferentiated in the Pleistocene (Cei, this vol- ume). Some Patagonian groups have relatives in the adjacent monte and the pampas, or in the austral forests, but the major latitudinal expansion has been northward in the Andes, best exemplified by Liolaemus (Fig. 1:2B). Tropical and Subtropical Herpetofaunas Herein distinction is made between two primary biotopes, as follow: 1) Forests — tropical evergreen forests, including rainforest and cloud forest, and 2) Nonforests — the de- ciduous, scrub or thorn forests, and savannas, grasslands and deserts. Although each of these categories, especially the latter, contains diverse vegetation formations, they seem to have reality with respect to major patterns of distribution of the herpetofauna. 1979 DUELLMAN: SOUTH AMERICAN HERPETOFAUNA 21 Fig. 1:2. Distribution patterns of the South American herpetofauna: A. Batrachyla in austral forests (black); Centrolcnclla principally in cloud forests but also entering lowland tropical rainforest (stippled). B. Liolac- mus, an austral group entering adjacent subtropical areas and extending northward in the Andes (stippled); Pleurodema brachijops, widespread in nonforests in northern South America (black). C. Hyla parviceps group with vicariant species in lowland rainforests; three species are in the Amazon Basin and one each in the other areas (stippled); Phyllopezus with disjunct populations in nonforested areas of the caatinga, cerrados, and pampas (black) (after Vanzolini, 1974). D. Osteocephalus with species on Andean and Guianan slopes and others in lowland rainforests. E. Tropidurus with species inhabiting diverse nonforested environments through- out tropical South America and the Galapagos Islands. F. Pleurodema, a temperate South American genus with vicariant species in the Andes and in nonforested environments to Panama. Patrones de distribution de la herpetofauna sudamericana. A. Batrachyla en los bosques australes (negro); Centrolenella principahncnte en bosques neblinos pero tambien entra las tierras bajas de la selva lluviosa tropical (punteado). B. Liolaemus, un grupo austral entra los areas subtropicales adyacentes y se extende hacia el norte en los Andes (punteado); Pleurodema brachyops diseminado en ambientes no forestales en el norte de Sud- america (negro). C. Hyla parviceps con especies vicarias en las tierras bajas del bosque lluvioso, tres especies en la Amazonia y una especie en cada una de los otros areas (punteado); Phyllopezus poblaciones disjuntas en areas sin bosque de caatinga, cerrado y pampas (segun Vanzolini, 1974) (negro). D. Osteocephalus con espe- cies en las laderas andinas y guianense y otras especies en las tierras bajas del bosque pluvial. E. Tropidurus con especies habitando diversos ambientes no forestales atraves de Sudamerica tropical y las Islas Galapagos. F. Pleurodema, un genero de la region templada de Sudamerica con especies vicarias en los Andes y en ambi- entes no forestales hasta Panama. 22 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 Forests. — The vast Amazonian rainforests and the smaller areas of rainforest in the Choco and along the southeastern coast of Brasil contain the richest herpetofaunas in South America (Lynch, this volume; Dixon, this volume). Especially diverse in these for- ests are dendrobatid and hylid frogs, anoline and teiid lizards, and xenodontine colubrid snakes. The herpetofaunas of the montane rainforests or cloud forests in the Andes, Gui- ana Highlands, and the Brasilian Highlands are primarily altitudinal extensions of the lowland groups (Hoogmoed, this volume; Duellman, this volume). However, in the montane forests, certain groups are either en- demic or far more diverse than in the low- lands — frogs of the families Centrolenidae (Fig. 1:2A), Dendrobatidae (Colostethus), and Leptodactylidae (Eleutherodactylus) and salamanders of the genus Bolitoglossa (Andes only). Distribution patterns are highly variable (see Duellman, 1978, Fig. 197, for examples of Amazonian distributions). A few wide- spread species, such as Boa constrictor, in- habit all of the lowland forests, but these species usually also inhabit intervening non- forest areas. More commonly, vicariant spe- cies occur in the different areas of rainforest (Fig. 1:2C). Widespread and speciose gen- era, such as Eleutherodactylus, Hyla, and An- olis, are found throughout the lowland and montane forests, but usually there are distinct combinations of species at different elevations, as shown for Eleutherodactylus by Lynch and Duellman (1979). These patterns are more readily discernible in smaller genera, such as Osteocephalus (3 Amazonian species, 1 Guianan, 2 Andean, and 1 coastal Brasilian; Fig. 1:2D) or Enyalioicles (2 Amazonian spe- cies, 2 Chocoan, and 3 Andean). The differentiation of populations in Qua- ternary forest refugia (Haffer, 1969, 1974) has been postulated for lizards (Vanzolini and Williams, 1970), frogs (Duellman, 1972; Heyer, 1973; Duellman and Crump, 1974) and snakes (Dixon, this volume). Nonforests. — The tropical and subtropical nonforested biotopes are more extensive, di- verse and fragmented than the forests. In northern South America are the coastal des- erts, savannas, and the extensive llanos; south and east of the Amazon Basin are the dry areas of northeastern Brasil (Caatinga), the interior savannas (cerrados), and the Gran Chaco. Subtropical, cis-Andean, nonforests include the pampas, espinal, and monte in Argentina; west of the Andes are the matorral and the Atacama Desert (Fig. 1:1). Distributions of many species of plants and animals indicate that various combina- tions of these nonforest environments were continuous with one another in the not-too- distant past. Pleistocene climatic fluctuations resulted in drier periods ( interglacials ) that allowed for expansion of the nonforests ( Haf- fer, 1974; Vanzolini, 1976). Gallardo (1969, 1971, this volume) emphasized the faunal re- lationships among the chaco, pampas and monte, and Vanzolini (1968, 1974, 1976) demonstrated distribution patterns in the cer- rados and caatinga. The herpetofaunal rela- tionships among the coastal deserts, llanos and savannas of northeastern South America are analyzed by Rivero-Blanco and Dixon (this volume) and Hoogmoed (this volume). The trans-Andean area consists of a nar- row coastal strip and Andean slopes to 2000- 3000 m, 1-37°S Lat. The dry upper Maranon Valley and associated valleys in the Huanca- bamba Depression are separated from the trans-Andean arid zone by passes at less than 3000 m. The coastal deserts and matorral have a small, but largely endemic, herpeto- fauna including three endemic genera of liz- ards — Garthia (2 species), Callopistes (2) and Dicrodon (3). The dominant groups are two genera of lizards, Tropidurus ( Dixon and Wright, 1975), and Phylhdactylus (Dixon and Huey, 1970), both of which have repre- sentatives in the dry valleys east of the Andes and on the Galapagos Islands. For much of the length of the coastal desert in northern Chile and southern Peni, the entire herpeto- fauna is composed of solely two species of Tropidurus and one of Phyllodactylus. The Humboldt Current sweeps the coast of Chile and Peru and swings westward past the Galapagos Islands, 600 km off the coast of Ecuador. This current has been important in rafting stocks of Atacaman reptiles — Tropi- durus, Pkyllodactylus, Alsophis (=Dromi- cus) — to the Galapagos. 1979 DUELLMAN: SOUTH AMERICAN HERPETOFAUNA 23 Several patterns of distribution are evi- dent in the nonforested regions. Some species, such as Pleitrodcma brachyops and Phimophis guianensis, are widespread in the northern regions (Fig. 1:2B), whereas others are re- stricted to the coastal deserts or llanos. Some ceratophryine frogs, some snakes, and several lizards have series of populations, subspecies or species distributed from the caatinga south- westward in the cerrados to the chaco or even into the monte or pampas (Fig. 1:2C). Such open habitats obviously were continuous, or nearly so, in the past so as to permit the dis- persal of such nonforest taxa as Tropidurus, Cnemidophorus, and Bufo granulosus (Webb, 1978 ) . Some of the taxa in the Atacama Des- ert are not represented east of the Andes, except in the upper Mar anon Valley, but Phyllodactylus also is diverse in northern South America, Middle America, and the West Indies. Tropidurus is widely distributed in tropical nonforested environments on both sides of the Andes (Fig. 1:2E). At least one temperate group (Pleurodema) has dispersed northward in nonforests to northern South America (Duellman and Veloso, 1977) (Fig. 1:2F). Montane Herpetofaunas The three major highland regions of South America — the ancient Guianan and Brasilian highlands and the young Andes — have little in common herpetologically. With few excep- tions, there are no isolated sister groups at high elevations that do not have relatives at low elevations. Hylid frogs of the genera Cryptobatra chits (northern Andes) and Ste- fania (Guiana Highlands) and teiid lizards, Euspondylus in the same regions, are primary examples. Hylid frogs of the genera Gastro- theca and Flectonotus occur in the Andes and in the Brasilian Highlands, but some of these species occur at low to moderate elevations, even though at present none lives in the inter- vening lowlands. The herpetofaunas of the highland regions seem to have been derived independently from the adjacent lowlands. In the case of the Andes, the fauna is composed of a southern assemblage derived from Pata- gonia and a northern assemblage derived from the tropical lowlands (Duellman, this vol- ume). FUTURE OF THE HERPETOFAUNA I have attempted to interpret the past and to describe the present; now I provide a prog- nosis for the study of the South American herpetofauna. South America has the richest herpetofauna of any continent, but the fauna is still poorly known taxonomically. Our knowledge of systematic and ecological rela- tionships is even less. Human devastation of vast areas of forest that a few years ago were unexplored is eliminating forever important, and in many cases unknown, aspects of the biota. Although biologists have had some in- fluence on the control of this exploitation, there is little hope that we can preserve all that we may wish to save. Thus, we are faced with two courses of action — salvage collecting and preservation of diverse natural preserves. Random collecting of the biota, even in reasonably well known areas, commonly re- sults in new information on distributions, tax- onomy, or life histories. However, collecting efforts need to be intensified and planned to sample biota before they are destroyed. In the case of amphibians and reptiles, efforts must be made to obtain not only series of well- preserved specimens but also tissues for karyo- logical and biochemical studies, colored pho- tographs, tape recordings of calls (of frogs), life history data, and extensive notes on habi- tats and behavior. We cannot necessarily ex- pect that a visit to the same region five or ten years hence, or even next year, will permit the collection of these data. The collection of these kinds of materials and data must be encouraged and supported at every level inter- nationally. The impending biological crisis has no national boundaries; responsible and effective collectors should be encouraged to make adequate collections throughout the continent. Systematic collections under re- sponsible direction of trained biologists will be one of the most important biological re- sources of the future; materials in these col- lections made available internationally to qualified investigators will be the basis for not only systematic studies but much evolutionary synthesis. The establishment of large natural reserves in areas of high species richness and ende- mism can preserve large segments of the her- 24 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 petofauna. However, such reserves should not be established strictly for conservation pur- poses. It is in these reserves that biologists can undertake long-range studies of communi- ties, population dynamics, behavior, and life history. The resulting kinds of information complement those obtained from salvage col- lecting and contribute substantially to our total understanding of the biota. Actions of these kinds are necessary now; a few years hence will be too late for some of the regions and their herpetofaunas. Without such actions the papers assembled in this vol- ume will not be the preliminary assessments as intended, but instead they might be the last word on the South American herpeto- fauna. ACKNOWLEDGMENTS In the preparation of this paper I have drawn freely on the manuscripts submitted by other contributors to this volume. I am in- debted to Richard Etheridge and John D. Lynch for some of the data and to Juan Man- uel Renjifo for translating the summary. An earlier draft of the manuscript benefited from critical review by John D. Lynch, Gregory K. Pregill, Linda Trueb, and Margaret Davies, whose austral invectives in the pits decidedly influenced the effectiveness of my writing. RESUMEN La herpetofauna sudamericana se corn- pone de 1,095 especies de anfibios distribuidos en 115 generos y 15 familias, 1,115 especies de reptiles en 203 generos y 22 familias (exclu- yendo los taxa marinos). De los 318 generos y 2,210 especies, 201 generos y 2,017 especies son endemicas de este continente. Entre las familias de reptiles, no encontramos ninguna endemica en sudamerica; en cambio existen cinco familias endemicas de anfibios. Durante 4.5-50 millones de aiios la fauna sudamericana evoluciono aislada del resto de los continentes formaban Gondwanalandia, solo hasta la relativamente reciente conexion (5.7 millones de aiios) con Norte America por la via del Lstmo de Panama. Antes del estab- lecimiento de la conexion terrestre, un archi- pielago proveo las veces de filtro entre Norte y Sur America, el cual fue cruzado en ambas direcciones por algunos grupos. El mayor in- tercambio entre las dos faunas se llevo a cabo una vez fue establecida la conexion entre los dos continentes. Seis familias de anfibios y 15 de reptiles son compartidas por Norte y Sur America. Ademas cuatro familias de anfibios y cuatro familias de reptiles sudamericanas tambien se encuentran en Centro America. Gran parte de la herpetofauna Antillana se compone de grupos neotropicales, algunos de los cuales invadieron las islas, especialmente las islas menores de las Antillas provenientes de Sudamerica; otros grupos invadieron las islas desde Centro America. De las 37 familias sudamericanas, tres de anfibios y seis de reptiles son compartidas con Africa y Australia. Un total de cinco familias de anfibios son compartidas con Africa y cu- atro familias con Australia. Entre las familias de reptiles, 11 son comunes con Africa y siete con Australia. De este modo, las relaciones a nivel de familias entre las herpetofaunas son mayores entre Sudamerica y Africa que entre Sudamerica y Australia. La mayor semejanza existe entre Australia y Africa. Comparado con los otros continentes que formaban Gond- wanalandia, Sudamerica tiene un numero des- proporcionadamente alto de anuros y Aus- tralia de saurios. La presencia o ausencia de las familias en los tres continentes se debe principalmente a factores historicos, mientras que la diversidad dentro de las familias depende del tiempo que estas hayan estado en el continente, el tamano del area de este y su diversidad ecologica. En Sudamerica la herpetofauna evoluci- ono respondiendo a los cambios de las condi- ciones climaticas durante el Cenozoico; apar- entemente muchas de las especies que existen actualmente evolucionaron en el Quaternario. Las herpetofaunas de las regiones templadas incluyen aquellas encontradas en los bosques australes y las estepas patagonicas. Los bosques estan restringidos a zonas aisladas, mientras que las estepas se han dispersado hasta el monte subtropical adyacente y hacia el norte en los Andes. Las herpetofaunas tropicales y subtropi- cales incluyen aquellas asociadas con bosques tropicales siempre verdes y bosques montano- 1979 DUELLMAN: SOUTH AMERICAN HERPETOFAUNA 25 sos y agregaciones asociadas con ambientes sin bosqucs grandes — pampas, monte, espinal, chaco, matorral, cerrado, caatinga, sabanas, llanos y desiertos. Cada una de estas regiones tiene su fauna caracteristica. Las herpeto- faunas de las tierras altas en los Andes, en Guiana y en Rrasil tienen poco en coniun; aparentemente cada una se derivo independi- entemente de las faunas encontradas en las tierras bajas adyacentes. Con el fin de adquirir material para estu- dios en el futuro y conservar agregaciones de la herpetofauna sudamericana antes de que la destrucion causada por el hombre elimine aun mas la biota, es necesario hacer extensas y bien documentadas colecciones de la herpeto- fauna a lo largo del continente asi como establecer grandes reservas naturales en areas ricas en especies y endemismos donde estudios biologicos de la fauna puedan ser llevados a cabo en el futuro. LITERATURE CITED Anderson, S., Jones, J. K., Jr. 1967. Recent mam- mals of the world: A synopsis of families. Ronald Press, New York, 453 p. Auffenberg, W. 1971. A new fossil tortoise, with remarks on the origin of South American testu- dines. Copeia 1971 ( 1 ) : 106-117. Avery, D. F., Tanner, W. W. 1971. Evolution of the iguanine lizards ( Sauria, Iguanidae ) as deter- mined by osteological and myological characters. Brigham Young Univ. Sci. Bull. Biol. Ser. 12(3): 1-79. Axelrod, D. I. 1967. Drought, diastrophism, and quantum evolution. Evolution 21:201-209. Axelrod, D. I. 1972. Edaphic aridity as a factor in angiosperm evolution. Amer. Nat. 106:311-320. Axelrod, D. I., Bailey, H. P. 1969. Paleotempera- ture analysis of Tertiary floras. Paleogeogr. Paleo- climatol. Paleoecol. 6:163-195. Berman, D. S. 1973. Spathorhynchus fossorium, a middle Eocene amphisbaenian ( Reptilia ) from Wyoming. Copeia 1973 (4):704-721. Blair, W. F. 1972. Summary, pp. 329-343 in Blair, W. F. (ed.). Evolution in the genus Bufo. Univ. Texas Press, Austin, 459 p. Bons, J., Pasteur, G. 1977. Solution histologique a un probleme de taxonomie herpetologique interes- sant les rapports paleobiologiques de I'Amerique du Sud el de l'Afrique. Acad. Sci. Paris, Ser. D 284:2547-2550. Brodkorb, P. 1971. Origin and evolution of birds, pp. 20-55 in Farner, D. S., King, J. R. (eds.). Avian biology, I. Academic Press, New York, 586 p. Burcer, W. L. 1971. Genera of pitvipers (Serpen- tes: Crotalidae). PhD Dissert. Univ. Kansas, 186 p. Cei, J. M. 1968. Remarks on the geographical dis- tribution and phyletic trends of South American toads. Pearce-Sellards Ser. Texas Mem. Mus. (13): 1-20. Cei, J. M. 1972. Bufo of South America, pp. 82-92 in Blair, W. F. (ed. ). Evolution in the genus Bufo. Univ. Texas Press, Austin, 459 p. Cogger, H. G. 1974. Voyage of the banded iguana. Australian Nat. Hist. 18:144-149. Cogger, H. G. 1975. Reptiles and amphibians of Australia. Reed, Sydney, 584 p. Cracraft, J. 1973. Continental drift, paleoclimatol- ogy, and evolution and biogeography of birds. J. Zool. 179:455-545. Cracraft, J. 1974. Continental drift and vertebrate distribution. Ann. Rev. Ecol. Syst. 5:215-261. Dietz, R. S., Holden, J. C. 1970. Reconstruction of Pangaea: Breakup and dispersion of continents, Permian to present. J. Geophys. Res. 75:4939- 4956. Dixon, J. R., Anderson, S. C. 1973. A new species and genus of gecko (Sauria: Gekkonidae) from Iran and Iraq. Bull. South. California Acad. Sci. 72:155-160. Dlxon, J. R. ; Huey, R. B. 1970. Systematics of the lizards of the gekkonid genus Phyllodactylus of mainland South America. Nat. Hist. Mus. Los Angeles Cty. Contrib. Sci. (192):l-78. Dlxon, J. R., Wright, J. W. 1975. A review of the lizards of die iguanid genus Troj)idurus in Peru. Ibid. (271): 1-39. Dowling, H. G. 1975. A provisional classification of snakes. 1974 Yearbook Herpetology. Herpetol. Infor. Search Syst., New York, pp. 167-170. Duellman, W. E. 1966. The Central American herpetofauna: An ecological perspective. Copeia 1966(4):700-719. Duellman, W. E. 1970. The hylid frogs of Middle America. Univ. Kansas Mus. Nat. Hist. Monogr. (l):l-753. Duellman, W. E. 1972. South American frogs of the Hyla rostrata group (Amphibia, Anura, Hy- lidae). Zool. Meded. Rijks Mus. Nat. Hist. Leiden 47:177-192. Duellman, W. E. 1975. On the classification of frogs. Univ. Kansas Mus. Nat. Hist. Occas. Pap. (42): 1-14. Duellman, W. E. 1977. Liste der rezenten Amphi- bien und Reptilien: Hylidae, Centrolenidae, Pseu- didae. Das Tierreich (95) : 1-225. Duellman, W. E., Crump, M. L. 1974. Speciation in frogs of the Hyla parviceps group in the upper Amazon Basin. Univ. Kansas Mus. Nat. Hist. Occas. Pap. (23): 1-10. Duellman, W. E., Veloso M., A. 1977. The phy- logeny of Pleurodema (Anura: Leptodactylidae ) : A biogeographic model. Ibid. (64): 1-46. Dunn, E. R. 1931. The herpetological fauna of the Americas. Copeia 1931 (3): 106-119. 26 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 Estes, R. 1970. Origin of the Recent North Amer- ican lower vertebrate fauna: An inquiry into the fossil record. Forma Functio 3:139-163. Estes, R. 1975. Fossil Xenopus from the Paleocene of South America and the zoogeography of pipid frogs. Herpetologica 31:263-278. Estes, R. 1976. Middle Paleocene lower vertebrates from the Tongue Formation, southeastern Mon- tana. J. Paleontol. 50:500-520. Estes, R., Price, L. I. 1973. Iguanid lizard from the Upper Cretaceous of Rrasil. Science 180:744- 751. Estes, R. Reig, O. A. 1973. The early fossil record of frogs: A review of the evidence, pp. 11-63 in Vial, J. L. (ed.). Evolutionary biology of the anurans: Contemporary research on major prob- lems. Univ. Missouri Press, Columbia, 470 p. Estes, R., Wake, M. 1972. The first fossil record of caecilian amphibians. Nature 239:228-231. Etheridge, R. 1959. The relationships of the anoles (Reptilia: Sauna: Iguanidae): An interpretation based on skeletal morphology. PhD Dissert. Univ. Michigan, 236 p. Etheridge, R. 1964. The skeletal morphology and systematic relationships of sceloporine lizards. Copeia 1964(4):610-631. Etheridge, R. 1966. The systematic relationships of West Indian and South American lizards referred to the iguanid genus Leiocephalus. Ibid. 1966 (1):79-91. Etheridge, R. 1967. Lizard caudal vertebrae. Ibid. 1967(4) :699-721. Formas, R., Pugin, E., Jorquera, R. 1975. La iden- bdad del batracio chileno Heminectes rufus Phi- lippi, 1902. Physis 34:147-157. Gaffnev, E. S. 1975. A phylogeny and classifica- tion of the higher categories of turtles. Bull. Amer. Mus. Nat. Hist. 155:387-436. Gaffney, E. S. 1977. The side-necked turtle family Chelidae: A theory of relationships using shared derived characters. Amer. Mus. Novit. (2620):1- 28. Gallardo, J. M. 1961. On the species of Pseudidae. Bull. Mus. Comp. Zool. Harvard Univ. 125:108- 134. Gallardo, J. M. 1969. Las especies de saurios (Reptilia) de la Provincia de Santa Fe, Argen- tina, y consideraciones sobre su ecologia y zoo- geografia. Neotropica 15:73-81. Gallardo, J. M. 1971. Composicion faunistica de los saurios de la Provincia de La Pampa, Repvib- lica Argentina. Ibid. 17:44—48. Gans, C. 1967. A check list of Recent amphisbae- nians ( Amphisbaenia, Reptilia). Bull. Amer. Mus. Nat. Hist. 135:61-105. Grant, N. K. 1971. South Atlantic, Benue Trough, and Gulf of Guinea Cretaceous triple junction. Geol. Soc. Amer. Bull. 82:229.5-2298. Greer, A. E. 1970. A subfamilial classification of scincid lizards. Bull. Mus. Comp. Zool. Harvard Univ. 139:151-183. Greer, A. E. 1974. The generic relationships of the scincid lizard genus Leiolopisma and its relatives. Australian J. Zool. Suppl. Ser. (31): 1-67. Haffer, J. 1969. Speciation in Amazonian forest birds. Science 165:131-137. Haffer, J. 1974. Avian speciation in tropical South America. Publ. Nuttall Ornith. Club (14): 1-390. Hagen, V. W. vox. 1948. The green world of the naturalists. Greenberg, New York, 392 p. Hershkovitz, P. 1972. The Recent mammals of the Neotropical Region: A zoogeographic and eco- logical review, pp. 311-^131 in Keast, A., Erx, F. C., Glass, B. (eds. ). Evolution, mammals, and southern continents. State Univ. New York Press, Albany, 543 p. Heyer, W. R. 1973. Systematics of the marmoratus group of the frog genus Leptodactylus. Nat. Hist. Mus. Los Angeles Cty. Contr. Sci. (251): 1-50. Heyer, W. R. 1975. A preliminary analysis of the intergeneric relationships of the frog family Lep- todactylidae. Smithsonian Contrib. Zool. (199): 1-55. Heyer, W. R., Liem, D. S. 1976. Analysis of the intergeneric relationships of the Australian frog family Myobatrachidae. Ibid. (233): 1-29. Holdex, J. C, Dietz, R. S. 1972. Galapagos core, NazCoPac triple junction and Carnegie/Cocos ridges. Nature 235:266-269. Holmax, J. A. 1967. Additional Miocene anurans from Florida. Q. J. Florida Acad. Sci. 30:121- 140. Holmax, J. A. 1968. Lower Oligocene amphibians from Saskatchewan. Ibid. 31:273-289. Holman, J. A. 1976. Snakes from the Rosebud For- mation (Middle Miocene) of South Dakota. Her- petologica 32:41-48. Holmax, J. A. 1977. Upper Miocene snakes (Rep- tilia, Serpentes) from southeastern Nebraska. J. Herpetol. 11:323-335. Izecksohx, E. 1971. Novo genero e nova especie de Brachycephalidae do Estado do Rio de Janeiro, Brasil (Amphibia, Anura). Bol. Mus. Nac. Rio de Janeiro, Zool. (280): 1-12. Jaxzex, D. H. 1976. The depression of reptile bio- mass by large herbivores. Amer. Nat. 110:371- 400. Jeannel, R. 1967. Biogeographie de l'Amerique australe, pp. 401-460 in Delamere Deboutte- ville, C, Rapoport, E. (eds.). Biologie de l'Amerique australe, 3. C. N. R. S. Groupe Fran- cais Argiles C. R. Reun. Etud., Paris, 834 p. Keast, A. 1972. Continental drift and the evolution of the biota on southern continents, pp. 23-87 in Keast, A., Erk, F. C, Glass, B. (eds.). Evolu- tion, mammals, and southern continents. State Univ. New York Press, Albany, 543 p. Kluge, A. G. 1967. Higher taxonomic categories of gekkonid lizards and their evolution. Bull. Amer. Mus. Nat. Hist. (135): 1-60. Larson, R. L., Ladd, J. W. 1973. Evidence for the opening of the South Atlantic in the Early Creta- ceous. Nature 246:209-212. 1979 DUELLMAN: SOUTH AMERICAN HERPETOFAUNA 27 Lynch, J. D. 1971. Evolutionary relationships, oste- ology, and zoogeography of leptodactyloid frogs. Univ. Kansas Mus. Nat. Hist. Misc. Publ. (53): 1-238. Lynch, J. D. 1973. The transition from archaic to advanced frogs, pp. 133-182 in Vial, J. L. (ed.). Evolutionary biology of the anurans: Contem- porary research on major problems. Univ. Mis- souri Press, Columbia, 470 p. Lynch, J. D. 1976. The species groups of the South American frogs of the genus Eleutlicrodactijlus ( Leptodactylidae). Univ. Kansas Mus. Nat. Hist. Occas. Pap. (61): 1-24. Lynch, J. D. 1978. A re-assessment of the telmato- biine frogs of Patagonia. Ibid. (72): 1-57. Lynch, J. D., Duellman, W. E. 1979. The Elcu- therodactylus of the Amazonian slopes of the Ecuadorian Andes (Anura: Leptodactylidae). Univ. Kansas Mus. Nat. Hist. Misc. Publ. (in press ) . Malfait, B. T„ Dinkelman, M. G. 1972. Circum- Caribbean tectonic and igneous activity and the evolution of the Caribbean Plate. Geol. Soc. Amer. Bull. 82:251-272. Maclio, V. J. 1970. West Indian xenodontine colu- brid snakes: Their probable origin, phylogeny, and zoogeography. Bull. Mus. Comp. Zool. Har- vard Univ. (141): 1-53. Marx, H., Rabb, G. B. 1965. Relationships and zoo- geography of the viperine snakes ( Family Viperi- dae). Fieldiana Zool. 44:161-206. Matthew, W. D. 1915. Climate and evolution. Ann. New York Acad. Sci. 24:171-318. Maxson, L. R. 1976. The phylogenetic status of phyllomedusine frogs ( Hylidae ) as evidenced from immunological studies of their serum albu- mins. Experimentia 32:1149-1150. McDiarmid, R. W. 1971. Comparative morphology and evolution of frogs of the Neotropical genera Atelopus, Dendrophryniscus, Melanophryniscus, and Oreophrynclla. Nat. Hist. Mus. Los Angeles Cty. Sci. Bull. (12): 1-66. McGowran, B. 1973. Rifting and drift of Australia and the migration of animals. Science 180:759- 761. Mertens, R. 1963. Liste der rezenten Amphibien und Reptilien: Helodermatidae, Varanidae, Lan- thanotidae. Das Tierreich (79): 1-26. Mertens, R. 1966. Liste der rezenten Amphibien und Reptilien: Chamaeleonidae. Ibid. (83): 1-37. Meszoely, C. 1970. North American fossil anguid lizards. Bull. Mus. Comp. Zool. Harvard Univ. 139:87-149. Meyer de Schauensee, R. 1964. The birds of Co- lombia. Livingston Publ. Co., Narberth, Pa., 427 p. Moreau, R. E. 1963. Vicissitudes of the African biomes in the late Pleistocene. Proc. Zool. Soc. London 141:395-421. Moreau, R. E. 1969. Climatic changes and distribu- tion of forest vertebrates in West Africa. J. Zool. 158:39-61. Noble, G. K. 1928. Two new fossil Amphibia of zoogeographic importance from the Miocene of Europe. Amer. Mus. Novit. (303): 1-13. Nussbaum, R. A. 1977. Rhinatrematidae: A new family of caecilians (Amphibia: Gymnophiona). Occas. Pap. Mus. Zool. Univ. Michigan (682): 1-30. Patterson, B., Pascual, R. 1968. The fossil mam- mal fauna of South America. Q. Rev. Biol. 43: 409-451. Peters, J. A., Donoso-Barros, R. 1970. Catalogue of the Neotropical Squamata: Part II. Lizards and amphisbaenians. U.S. Natl. Mus. Bull. (297): 1-293. Pianka, E. R. 1971. Lizard species diversity in the Kalahari Desert. Ecology 52:1024-1029. Pianka, E. R., Huey, R. B. 1971. Bird species di- versity in the Kalahari and the Australian deserts. Koedoe 14:123-130. Rage, J.-C. 1975a (1974). Les serpents des Phos- phorites du Quercy. Paleovert. 6:273-303. Rage, J.-C. 1975b. Un caenophidien primirif ( Rep- tilia: Serpentes) dans l'Eocene inferieur. C. R. Somm. Seances Soc. Geol. France 1975:46-47. Rage, J.-C. 1977. La position phyletique de Dini- lysia patagonica, serpent du Cretace superieur. C. R. Acad. Sci. Paris, Ser. D, 284:1765-1768. Raven, P. H., Axelrod, D. I. 1974. Angiosperm bio- geography and past continental movements. Ann. Missouri Bot. Card. 61:539-673. Reyment, R. A., Tait, E. A. 1972. Biostratigraphic dating of the early history of the South Atlantic Ocean. Philos. Trans. Roy. Soc. London 264:55- 95. Richards, P. W. 1973. The tropical rain forest. Sci. Amer. 229:58-67. Rosen, D. E. 1975. A vicariance model of Caribbean biogeography. Syst. Zool. 24:431—464. Sarmiento, G. 1976. Evolution of arid vegetation in tropical America, pp. 65-99 in Goodall, D. W. (ed.). Evolution of desert biota. Univ. Texas Press, Austin, 250 p. Savage, J. M. 1963. Studies on the lizard family Xantusiidae IV. The genera. Nat. Hist. Mus. Los Angeles Cty. Contrib. Sci. (71): 1-38. Savage, J. M. 1966. The origins and history of the Central American herpetofauna. Copeia 1966(4): 719-766. Savage, J. M. 1968. The dendrobatid frogs of Cen- tral America. Ibid. 1968(4):745-776. Savage, J. M. 1972. The harlequin frogs, genus Atelopus of Costa Rica and western Panama. Herpetologica 28:77-94. Savage, J. M. 1973. The geographic distribution of frogs: Patterns and predictions, pp. 351-445 in Vial, J. L. ( ed. ) . Evolutionary biology of the anurans: Contemporary research on major prob- lems. Univ. Missouri Press, Columbia, 470 p. Savage, J. M. 1974. The isthmian link and the evo- lution of Neotropical mammals. Nat. Hist. Mus. Los Angeles Cty. Contrib. Sci. (260):1-51. 28 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 Savage, J. M, Carvalho, A. L. de. 1953. The fam- ily position of Neotropical frogs currently referred to the genus Pseudis. Zoologica 38:193-200. Savitzky, A. H. 1978. The origin of the New World proteroglyphous snakes and its bearing on the study of venom delivery in snakes. PhD Dissert. Univ. Kansas, 387 p. Schwartz, A. 1972. The native toads (Anura: Bu- fonidae) of Hispaniola. J. Herpetol. 6:217-231. Schwartz, A. 1978. Some aspects of the herpeto- geography of the West Indies, pp. 31-51 in Gill, F. B. (ed.). Zoogeography in the Caribbean. Acad. Nat. Sci. Philadelphia Spec. Publ. (13): 1-128. Schwartz, A., Thomas, R. 1975. A check-list of West Indian amphibians and reptiles. Carnegie Mus. Nat. Hist. Spec. Publ. (1):1-216. Sill, W. D. 1968. The zoogeography of the croco- dilians. Copeia 1968( 1 ):76-88. Simpson, B. B. 1975. Pleistocene changes in the flora of the high tropical Andes. Paleobiology 1:273-294. Smith, H. M., Martin, R. L., Swain, T. A. 1977. A new genus and two new species of South Amer- ican geckos (Reptilia: Lacertilia). Pap. Avulsos Zool. (Sao Paulo) 30:195-213. Sokol, O. M. 1975. The phylogeny of anuran lar- vae: A new look. Copeia 1975( 1 ) :l-23. Solbric, O. T. 1976. The origin and floristic affini- ties of the South American desert and semi-desert regions, pp. 7—49 in Goodall, D. W. (ed. ). Evo- lution of desert biota. Univ. Texas Press, Austin, 250 p. Starrett, P. H. 1973. Evolutionary patterns in lar- val morphology, pp. 251-271 in Vial, J. L. (ed.). Evolutionary biology of the anurans: Contem- porary research on major problems. Univ. Mis- souri Press, Columbia, 470 p. Tihen, J. A. 1962. A review of New World fossil bufonids. Amer. Midi. Nat. 68:1-50. Tihen, J. A. 1964. Tertiary changes in the herpeto- faunas of temperate North America. Sencken- bergiana. Biol. 45:265-279. Tihen, J. A. 1972. The fossil record, pp. 8-13 in Blair, W. F. (ed.). Evolution in the genus Bufo. Univ. Texas Press, Austin, 459 p. Trueb, L. 1971. Phylogenetic relationships of cer- tain Neotropical toads with the description of a new genus (Anura: Bufonidae). Nat. Hist. Mus. Los Angeles Cty. Contrib. Sci. (216): 1^40. Trueb, L. 1974. Systematic relationships of Neo- tropical horned frogs, genus Hcmiphraetus (Anu- ra: Hylidae). Univ. Kansas Mus. Nat. Hist. Occas. Pap. (29): 1-60. Trueb, L., Tyler, M. J. 1974. Systematics and evo- lution of the Greater Antillean hylid frogs. Ibid. (24): 1-60. Tyler, M. J. 1974. First fossil frogs from Australia. Nature 248:711-712. Tyler, M. J., Davies, M. 1978. Species groups with- in the Australopapuan hylid frog genus Litoria Tschudi. Australian I. Zool Suppl. 63:1—17. Tyler, M. J., Davies, M., King, M. 1978. The Aus- tralian frog Chiroleptes dahlii Boulenger: Its sys- tematic position, morphology, chromosomes and distribution. Trans. Roy. Soc. South Australia 102:17-23. Vanzolini, P. E. 1968. Geography of the South American Gekkonidae (Sauria). Arq. Zool. (Sao Paulo) 17:85-112. Vanzolini, P. E. 1974. Ecological and geographical distribution of lizards in Pemambuco, northeastern Brasil (Sauria). Pap. Avulsos Zool. (Sao Paulo) 28:61-90. Vanzolini, P. E. 1976. On the lizards of a cerrado- caatinga contact: Evolutionary and zoogeographi- cal implications (Sauria). Ibid. 29:111-119. Vanzolini, P. E., Williams, E. E. 1970. South American anoles: The geographic differentiation and evolution of the Anolis chrysolepis species group (Sauria, Iguanidae). Arq. Zool. (Sao Paulo) 19:1-298. Veevers, ]. J., McElhinny, M. W. 1976. The sep- aration of Australia from other continents. Earth Sci. Rev. 12:139-159. Vuilleumier, F. 1968. Origin of frogs of Patagonian forests. Nature 219:87-89. Wake, D. B. 1966. Comparative osteology and evo- lution of the lungless salamanders, family Pletho- dontidae. Mem. South. California Acad. Sci. (4): 1-111. Webb, S. D. 1978. A history of savanna vertebrates in the New World. Part II: South America and the great interchange. Ann. Rev. Ecol. Syst. 9: 393-436. Wermuth, H. 1965. Liste der rezenten Amphibien und Reptilien: Gekkonidae, Pygopodidae, Xantu- siidae. Das Tierreich (80) : 1-246. Wermuth, H. 1967. Liste der rezenten Amphibien und Reptilien: Agamidae. Ibid. (86):1-127. Wermuth, H. 1968. Liste der rezenten Amphibien und Reptilien: Cordylidae. Ibid. (87): 1-30. Williams, E. E. 1969. The ecology of colonization as seen in the zoogeography of anoline lizards on small islands. Q. Rev. Biol. 44:345-389. Wolfe, J. A. 1971. Tertiary climatic fluctuations and methods of analysis of Tertiary floras. Paleo- geogr. Paleoclimatol. Paleoecol. 9:27-57. Wood, R. C. 1971. The fossil Pelomedusidae (Testu- dines, Pleurodira) of Africa. PhD Dissert. Har- vard Univ., 345 p. Wood, R. C, Patterson, B. 1973. A fossil triony- chid turtle from South America. Breviora (415): 1-10. Zweifel, R. G. 1972. Results of the Archbold Ex- peditions No. 97. A revision of the subfamily Asterophryinae, family Microhylidae. Bull. Amer. Mus. Nat. Hist. 148:411-546. 2. The South American Herpetofauna: An Evaluation of the Fossil Record Ana Maria Baez Departamento de Geologia Facultad de Ciencias Exactas y Naturales Univcrsidad de Buenos Aires Buenos Aires, Argentina Zulma B. de Gasparini Facultad de Ciencias Naturales y Museo Universidad Nacional de La Plata La Plata, Argentina The presence of fossil remains of amphib- ians and reptiles related to living taxa in South America has been documented since the last century. Nevertheless, examination of this literature reveals that in many cases the names are only mentioned and the ma- terial has not been studied; in many other cases it is evident that a revision is badly needed. Partial reviews concerning several South American countries have been carried out (Argentina: Pascual, 1970; Pascual and Odre- man Rivas, 1971; Gasparini and Baez, 1975; Brasil: Paula Couto, 1970; Colombia: Hoff- stetter, 1970a; Ecuador: Hoffstetter, 1970b; Peru: Hoffstetter, 1970c; Uruguay: Mones, 1972; 1975). Also, Baez and Gasparini (1977) critically examined the Cenozoic record of amphibians and reptiles in that continent and analyzed distributional shifts, relating them to the Cenozoic environmental changes that oc- curred as a result of different geological events. In this paper, the available paleontological data are summarized in an attempt to evalu- ate the information that the fossil record can provide about the historical development of these groups in South America. The taxo- nomic assignment and the geographic and stratigraphic references are sometimes doubt- ful; questionable referrals are not considered. Some recent finds are currently under study, and identification of genera and species is not yet available; thus, only an analysis at the family level is possible at present. The main areas that have yielded fossil amphibians and reptiles referable to modern groups are shown in figures 2:4-7. A checklist of the material recorded there and the corresponding litera- ture references appear in Appendix 2:1. Inspection of the fossil record reveals that many families comprising the present South American heqoetofauna have a considerable antiquity in that continent. Their presence, and even that of some recent genera, extends back to the late Mesozoic and early Tertiary (Figs. 2:1-3). Thus, it is essential to con- sider the past changes in the geographical position of South America and its connections with other continents, especially from the Middle Mesozoic onward, for those changes must have affected distribution patterns. According to recent paleomagnetic data (Vilas and Valencio, 1978a), South America was part of Gondwanaland, which may have existed up to the Late Jurassic. It was still joined to Africa during the Early Cretaceous ( Reyment and Tait, 1972; Douglas, Moullade and Nairn, 1973; Vilas and Valencio, 1978), while faunal relations with Antarctica-Aus- tralia seem to have been possible up to the Early Tertiary (Raven and Axelrod, 1975). There is no clear evidence that a direct land connection between North and South America existed from the Jurassic to the end of the Tertiary (Malfait and Dinkelman, 1972; McKenna, 1973). Nevertheless, a discontin- uous pathway along a chain of volcanic is- lands could have been established at different times (Haffer, 1970; Malfait and Dinkelman, 1972 ) . There is no agreement concerning the time of final reconnection between both Americas. Some evidence indicates that the Isthmus of Panama was established during 29 30 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 FAMILIAL GROUPS RECENT PLEISTOCENE TERTIARY CRETACEOUS PLIOCENE MIOCENE OLIGOCENE EOCENE PALEOCENE PIP1DAE LEPTODACTYLIDAE — — ? BUFONIDAE HYLIDAE RHINODERMATIDAE PSEUDIDAE CENTROLENIDAE DENDROBATIDAE BRACHYCEPHALIDAE MICROHYLIDAE RANIDAE CAECILIDAE — Fig. 2:1. Chronological range (Cretaceous-Recent) of Anura and Gymnophiona in South America. Distribution cronologica (Cretdcico-Reciente) de Anura y Gymnophiona en America del Sur. the late Pliocene-early Pleistocene ( Patter- son and Pascual, 1968); others favor its earlier existence (Emiliani et al., 1972; Savage, 1974). Thus, it is implied that biogeograph- ical relationships with Africa and Australia may have been close until the end of the Mesozoic, whereas those with North America became more important by the Late Tertiary. Available paleomagnetic data suggest that the latitudinal position of South America did not alter significantly since the latest Paleo- zoic, although displacements of about 5°, at most, may have existed in some areas because of the different orientation of the continent (Vilas and Valencio, 1978, 1979). Thus, these changes are disregarded here because they are insufficient to account for the different distributions of many groups in the past as evidenced by the fossil record. The dispersal history of amphibians and reptiles also has been very closely related to past climates. In South America climatic- ecologic conditions were more equable dur- ing at least the Cretaceous and early Tertiary. Various kinds of evidence indicate that a warm and humid climate prevailed then at high latitudes (Pascual and Odreman Rivas, 1971; Archangelsky and Romero, 1974; Petri- ella and Archangelsky, 1975). Throughout the Cenozoic, geologic events of different magnitude provoked physiographic, and con- sequently, climatic and floristic changes. Al- though these changes restricted the dispersal of some groups, they multiplied the available environments and thereby promoted the ap- pearance of new adaptive types. 1979 BAEZ & GASPARINI: FOSSIL RECORD 31 FAMILIAL GROUPS RECENT PLEISTOCENE TERTIARY CRETACEOUS PLIOCENE MIOCENE OLIGOCENE EOCENE PALEOCENE MEIOLANIIDAE ? PELOMEDUSIDAE - CHELIDAE - TESTUDINIDAE EMYDIDAE P TRIONYCHIDAE — CHELYDRIDAE KINOSTERNIDAE SEBECIDAE CROCODYLIDAE 9 ALLIGATORIDAE GAVIALIDAE NETTOSUCHIDAE Fig. 2:2. Chronological range (Cretaceous-Recent) of Testudines and Crocodilia in South America. Fam- ily groups without Cenozoic records have not been included. Distribution cronologica (Cretdcico-Reciente) de Testudines ij Crocodilia en America del Sur. Las familias sin registros cenozoicos no han sido incluidas. THE FOSSIL RECORD OF MODERN GROUPS OF AMPHIBIANS AND REPTILES IN SOUTH AMERICA The oldest known fossils referable to fam- ily groups that comprise the present South American herpetofauna appear in that conti- nent in Cretaceous deposits. The presence of pipid frogs, iguanid lizards and pelomedusid turtles in the Late Cretaceous is documented by the fossil record. Pelomedusid turtles were quite diversified and widely distributed at that time (Fig. 2:4). It is noteworthy that the extant pelomedusid genus Podocnemis was present then. Leptodactylid frogs perhaps also are present in Cretaceous horizons. The earliest Cenozoic records of amphib- ians and reptiles come almost exclusively from southern South America — Patagonia and southeastern Brasil (Fig. 2:5). The early and late Paleocene Patagonian localities have yielded remains of turtles ( only pelomedusids can be ascertained definitely), eusuchian crocodilians (including crocodylids) and boid snakes. Most of this material is very fragmen- tary, but these records are quite interesting from a paleoclimatic point of view, for they indicate humid subtropical conditions at lati- tudes of about 45°S. On the other hand, the rich and diversified assemblage of late Paleo- cene age of Itaborai, Brasil, provides more comprehensive information about the family groups that were inhabiting South America at that time. However, most of this material has vet to be studied. 32 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 FAMILIAL GROUPS RECENT PLEISTOCENE T E RT 1 A R Y CRETACEOUS PLIOCENE MIOCENE OLIGOCENE EOCENE PALEOCENE IGUANIOAE TEIIDAE GEKKONIDAE ANGUIDAE SCINCIDAE AMPHISBAENIDAE ANILIIDAE BOIIDAE ? ? COLUBRIDAE VIPERIDAE Fig. 2:3. Chronological range (Cretaceous-Recent) of Squamata, exclusive of scolecophidian snakes, in South America. Distribution cronologica (Cretdcica-Reciente) de Squamata, cxcluyendo serpientes scolecofidias, en America del Sur. In the known Paleocene sites, fourteen families have been recorded, all of them (ex- cept the extinct sebecid crocodilians ) are rep- resented in the present-day fauna of South America. They include the Caeciliidae, Pipi- dae, Leptodactylidae, Bufonidae, Hylidae, Pelomedusidae, Crocodylidae, Alligatoridae, Iguanidae, Teiidae, Gekkonidae, Aniliidae, and Boidae. Among these families we can recognize elements that are of diverse histori- cal backgrounds, and that joined the fauna of present-day South America at different times and developed in situ. Presently-known Eocene faunas also were found mainly in the southern part of the con- tinent (Fig. 2:5). Sebecid and alligatorid crocodilians are still recorded at latitudes of about 46°S, thereby indicating that at least warm temperate conditions persisted there. Noteworthy is the appearance of chelid tur- tles, whose extinct and living representatives have been found only in Australia and South America. The Eocene remains have been re- ferred to the extant genus Hydromedusa, which now extends as far south as 36°S. The living leptodactylid genus Caudiverbera is known from Eocene deposits of Patagonia, where it was associated with crocodilian re- mains. The earliest known Cenozoic records of amphibians and reptiles in northernmost South America come mainly from Oligocene sites, especially several localities in the upper and middle Magdalena River Valley, Colom- bia (Fig. 2:6). These assemblages are char- acterized by the abundance of mesosuchian 1979 BAEZ & GASPARINI: FOSSIL RECORD 33 (Sebecidae) and eusuchian (Alligatoridae, Crocodylidae, Gavialidae) crocodilians, which is explained by the presence there of an extensive lowland with local swamps and lakes (Van Houten and Travis, 1968). All of the families recorded from those horizons also are represented in older deposits of more southern latitudes, except gavialid crocodil- ians, whose affinities are still uncertain. Oligocene localities in central Patagonia (Fig. 2:6) have yielded remains of the extant lepodactylid genera Eupsophus and Caudi- verbera, whose presence there was made pos- sible by the prevalence of more mesic condi- tions at that time. The oldest known testu- dinid turtle in South America, a species of the genus Geochelone related to the living G. chilemis (Auffenberg, 1971), is recorded in the upper Oligocene of Patagonia. Very rich, late Miocene faunas are known from the upper Magdalena River Valley, Co- lombia, where the general environmental con- ditions of Oligocene times did not change significantly, although important tectonic events took place in the Miocene (Irving, 1971). Most of the families present there have previous records in the continent, the majority being represented by extant genera or species: Bufo marinus, Podocnemis expan- sa, Tupinambis cf. T. teguixin, Dracaena, Eu- nectes, Caiman. Also a turtle, Geochelone hesterna, closely related to the living G. car- bonaria and G. denticulata (Auffenberg, 1971) was recorded. The appearance of two families, Colubridae and Nettosuchidae, in the fossil record is noteworthy. The latter are endemic eusuchian crocodilians seemingly re- stricted to tropical regions from Miocene to Plio-Pleistocene times. The southernmost records of Cenozoic am- phibians or reptiles on the continent come from early to middle Miocene deposits of southern Patagonia (Fig. 2:6). Iguanids, teiids and other lizards, turtles and snakes have been found there, but all of this ma- terial has yet to be restudied. The post-Mio- cene history of the Patagonian herpetofauna is not documented in the fossil record. The increasing aridity that developed there as a consequence of the Andean tectonic move- ments and the continuous uplift of that area throughout the late Cenozoic made the pres- ervation of remains unlikely. Numerous crocodilian fossils, all referable to the suborder Eusuchia (Alligatoridae, Gav- ialidae, Nettosuchidae and probably Croco- dylidae) are recorded in the upper Pliocene of northern Venezuela (Fig. 2:7). The gigan- tic size exhibited by many members of these taxa and the appearance of the extant alli- gatorid, Melanosuchus, is noteworthy. The possible presence of trionychid turtles there should be noted. This group is not repre- sented in the present fauna of South America, and may have reached the continent from the north. In Pliocene times the Rio Parana consti- tuted, as it does today, an important pathway for the southward migration of elements of the subtropical biota. This explains the rec- ord of alligatorid and gavialid crocodilians in Pliocene deposits near the city of Parana, Argentina (Fig. 2:7). Among them, Caiman latirostris and possibly Rhamphostomopsis were already represented in the late Miocene faunas of Colombia. Numerous remains of amphibians and rep- tiles have been reported from horizons of late Pliocene age in Monte Hermoso, Argentina (Fig. 2:7). The presence there of the living teiid genus Callopistes is especially significant considering its present range, which is on the arid Pacific lowlands of Peru and Chile. The relatively few Pleistocene sites that have yielded amphibians or reptiles are im- portant because they document the past pres- ence of families that currently inhabit South America, and that have not been recorded in older deposits. Among these families are emy- dids, viperids, and amphisbaenids. The Pleis- tocene records also suggest that significant environmental changes occurred at that time, because the distribution of some taxa is incon- sistent with the present distribution of their habitats. For example, the Pleistocene faun- ules from Talara, northwestern Peru (Lemon and Churcher, 1961; Hoffstetter, 1970c) and from the Santa Elena Peninsula, southwestern Ecuador (see Appendix 2:1), indicate that more mesic environments prevailed in those areas in the near past. 34 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 JO» 1 2 3 4 5 6 CRETACEOUS Alemanio Laquna Umayo Mossoro Vilo-Vilo Sao Jose do Rio Preto Lago Colhue Huapi Fig. 2:4. Major areas in South America that have yielded Cretaceous amphibians and/or rep- tiles of modern groups. Principales areas en America del Stir que han brindado anfibios y/o reptiles cretdcicos de grupos modernos. 1979 BAEZ & GASPARINI: FOSSIL RECORD 35 J0° A PALEOCENE • EOCENE i Golfo de San Jorge 2 Itaboroi' 3 Cerro Pan de Azucar 4 Laguna del Hunco 5 Mina Aguilor 6 Canadon Hondo -Cana- do'n Vaca 7 Lago Colhue- Huapi 8 Negritos 9 Divisadero Largo / 20" Fic. 2:5. Major areas in South America that have yielded amphibians and/or reptiles of Paleo- cene and Eocene age. Principales areas en America del Sur que han brindado anfibios y/o reptiles de cdad paleocena y eocena. 36 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 ao° A 0LIG0CENE • MIOCENE i Compo Waldo 2 Scorritt Pocket 3 Tremembe' 4 Colhue' Huopi' 5 Chaparral 6 Gaiman 7 Santa Cruz 8 Coyaima 9 Carmen de Apicola 10 La Venta n Logo Buenos Aires 12 Barranca de los Loros 13 Ingeniero Jacobacci / Fie. 2:6. Major areas in South America that have yielded amphibians and/or reptiles of Oligo- eene and Miocene age. Principales areas en America del Sur que han brindado anfibios tj/o reptiles de edad oligocena y miocena. 19(79 BAEZ & GASPARINI: FOSSIL RECORD 37 ^0° A PLIOCENE • PLEISTOCENE 1 Urumoco 2 V. de Santa Maria 3 Parana' 4 Monte Hermoso 5 Queque'n Salado 6 Chapadmalal 7 Rio Jurud 8 Rio Aguaytia 9 Arroyo Perico Flaco to Santa Elena 11 Tarijo 12 Nuapua i 40° / 20° / Fie. 2:7. Major areas in South America that have yielded amphibians and/or reptiles of Plio- cene and Pleistocene age. Principales areas en America del Stir que han brindado anfibios ij/o reptiles de edad pliocena y pleistocena. 38 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 AMPHIBIANS The fossil record of modern amphibian groups in South America is largely restricted to anurans. The only known caecilian fossil is from the late Paleocene of Brasil; this rec- ord implies that caecilians are old components of the South American herpetofauna. No salamander fossils are known from South America. The considerable diversity of anuran groups already present by the early Tertiary in South America is noteworthy. Further- more, their phylogenetic relationships and the fact that some of them were represented by taxa similar to extant South American genera or species indicate that they have had a long history there. However, many families that comprise the present anuran fauna are un- known as fossils. The aquatic frogs of the family Pipidae now live in South America and Africa; they constitute an ancient, independent lineage. The earliest pipids are from the Early Creta- ceous of Israel (Nevo, 1968), but their origin probably extends back into the Jurassic. Un- questionable pipids occur in the Cretaceous and Tertiary of both Africa and South Amer- ica; this strongly suggests that pipids were components of the Gondwanan fauna. Fur- thermore, remains referable to the African genus Xenopus have been recorded from both continents (Ahl, 1926; Vergnaud-Grazzini, 1966; Broin et al., 1974; Estes, 1975a,b; Baez, 1976). The fossil record indicates that differ- ent phyletic lines of pipids have existed in South America. The oldest known example there is Saltenia ibanezi from the Late Cre- taceous of northwestern Argentina. Members of the extant genus Xenopus have been re- corded in the late Paleocene of Brasil and Ar- gentina. No fossil taxa directly ancestral to the living Neotropical Pipinae are known. The Leptodactylidae are another ancient component of the South American fauna. The oldest known remains unquestionably refer- able to the family come from late Paleocene deposits in Brasil; undescribed taxa close to living genera and perhaps even an extant genus are present there (Estes and Reig, 1973). This group probably is represented in the Late Cretaceous of Peru (Sige, 1968), but that record has not been confirmed. Other early presumed records outside of South America have been discarded or questioned (Hecht, 1963; Estes, 1964, 1969; Reig, 1968; Lynch, 1971; Savage, 1973). Available evi- dence indicates that the Leptodactylidae, in its restricted sense (Lynch, 1973) may have differentiated in South America from a stock associated with temperate forests (Lynch, 1971; Savage, 1973; Heyer, 1975). The pro- posed relationships to leiopelmatids (Heyer, 1975) are significant in that the latter inhab- ited Patagonia by Jurassic times (Estes and Reig, 1973). Leptodactylids must have dis- persed northward early in their history. Wide- spread occurrence of subhumid environments in middle latitudes in the Late Jurassic and Early Cretaceous might have been influential in the development of xeric-adapted types of leptodactylids (Baez and Gasparini, 1977). Telmatobiine leptodactylids, which represent an ancient radiation, are recorded in the Ter- tiary of Patagonia. The extant genus Caudi- verbera appears in early Eocene, early and late Oligocene, and late Miocene deposits. By the early Oligocene two other genera were present — the living Eupsophus and the ex- tinct Neoprocoela; the assignment of Neopro- coela to this family has been discussed by Tihen (1962, 1972) and Lynch (1971). These records furnish evidence of their former wider distribution east of the Andes, and therefore reflect the mesic climate that prevailed in those now arid regions (Gasparini and Baez, 1975). Ceratophryine leptodactylids, which some authors give familial status, are first recorded in the late Miocene of northern Patagonia, but they must have originated much earlier. The basic adaptation of this group to an arid environment ( Heyer, 1975 ) could account for the paucity of paleontologi- cal evidence for its early evolution, because few fossil-bearing horizons representing those environments are known, particularly in the Late Mesozoic and Early Tertiary. Remains referable to the extant genus Ceratophrys (specific allocation undetermined) are known from the late Pliocene and middle Pleistocene of Argentina (Baez and Gasparini, 1977). The living C. ornata probably is represented in the late Pliocene of Argentina (Reig, 1958) 1979 BAEZ & GASPARINI: FOSSIL RECORD 39 and late Pleistocene of Bolivia, and C. aurita is known from the latest Pleistocene in Brasil (Lynch, 1971). Records of other leptodac- tylids are limited to the living genus Lepto- dachjlus from Pleistocene deposits. The earliest known record of the cosmo- politan bufonid toads (absent from Australia and Madagascar) is from the late Paleocene of Brasil. Although this material has not yet been described, the presence of members of living species groups of Bufo has been recog- nized (Estes and Reig, 1973). This supports the proposal that bufonids are ancient com- ponents of the South American fauna and that they could have had a southern origin (Lynch, 1971; Reig, 1972; Savage, 1973). No unquestionable bufonids are known in North America prior to the early Miocene (Tihen, 1972). Moreover, assignment of specimens of Early Tertiary age from Europe to the Bufo- nidae is highly doubtful; remains referable to the genus Bufo are unknown there before the middle Miocene (Tihen, 1972). Different evidence suggests that South America is the most likely area of origin of Bufo (Blair, 1972). Although sparse, the fossil record in- dicates that considerable diversification with- in that genus took place in South America at least since the Early Tertiary. Members of the marinus group of Bufo have been reported as far back as the late Miocene, although an earlier differentiation of that group seems likely. Tihen (1972) and Estes and Reig (1973) considered Neoprocoela to be a mem- ber of the Bufo calamita group, which is now confined to the Old World (also see Gal- lardo, 1962). The Hylidae is poorly represented in the fossil record. The earliest remains referred to this family are from the late Paleocene of Brasil, but the material is still undescribed (Estes and Reig, 1973). This testifies to the early presence of hylid frogs in South Amer- ica, and is consistent with the high degree of differentiation that they attained there during the time of isolation, as well as their possible origin on that continent (Savage, 1973). The now widespread genus Hula has been re- corded from the early Oligocene of Canada (Holman, 196S) and the early Miocene of Florida in the United States (Tihen, 1964). The paleontological data add little informa- tion to the biogeographical history of hylids. Frogs are ancient members of South Amer- ican fauna. Pipids, leptodactylids, bufonids and hylids are old components and are the only frog families represented in the South American fossil record. Other groups such as rhinodermatids and dendrobatids, differenti- ated on that continent probably during the Tertiary. On the other hand, ranids are late immigrants from the north. REPTILES Pelomedusid turtles presently inhabit South America, Africa and Madagascar, but they were more widely distributed in Late Mesozoic and Early Tertiary times (Romer, 1966; Gaffney and Zangerl, 1968; Wood, 1970, 1976b; Jimenez Fuentes, 1971, 1975; Gaffney, 1975). The earliest-known pelomedusids are from Early Cretaceous deposits in Africa (Broin et al., 1974), and remains of Late Cretaceous age have been reported from South America, North America, Africa, and Europe. The greater proximity of continents at that time (Smith, Briden and Drewry, 1973), coupled with the presumed marine habits of some of these turtles (Wood, 1972, 1976b) could have favored such wide distri- bution. The majority of the South American fossil pelomedusids that have been described are referred to the genus Podocnemis, which is still present in South America and Mada- gascar. The presence of that genus in South America extends back to the Late Cretaceous, at which time the genus was represented by a species, P. elegans, noted as being "strikingly modern in aspect" (Wood, 1971:27s). 1 Nu- merous fossil remains have been assigned to Podocnemis, but there is no general agree- ment concerning the validity of species (Wood and Gamero, 1971; Baez and Gaspari- ni, 1977). By the late Miocene the extant P. expanse was already in existence. Remains referable to the genus Podocnemis are cited outside of South America, from Africa and western Europe. Pelomedusids referable to other genera have been described from the 1 Dr. F. de Broin (pers. coram.) considers that Po- docnemis elegans may belong to the extinct pelome- dusid genus Roxochehjs, related to Podocnemis. 40 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. Cenozoic of South America. A species of the genus Taphrosphys occurs in the Eocene of coastal Peru. That genus, which includes ma- rine forms, is also recorded from the Creta- ceous of North America and Paleocene of Africa (Wood, 1975). The gigantic Pliocene Stupendemys geographicus exhibits many un- usual characteristics, and its affinities are still not clear (Wood, 1976b). The freshwater turtles of the family Cheli- dae occur today in South America, Australia and New Guinea. The geographical location of fossils assigned to this group indicates that they could have had an essentially similar former range, because the only known un- questionable records come from the early Eo- cene of southern Argentina, middle Tertiary of Australia, and Oligocene or Miocene of Tasmania (Warren, 1969). That evidence suggests an area of dispersal that included also Antarctica and that the past relations between South American and Antarctica-Aus- tralia (McGowran, 1973; Dalziel et al., 1973; Sclater and Fisher, 1974) could account for the present range. A species of the extant South American genus Hydro medusa is the oldest member ( Eocene ) of the family re- corded so far (Wood and Moody, 1976). Two extinct species of the now monotvpie genus Chelus have been described — C. co- lombianus from the late Miocene of Colombia and C. leuisi from the Pliocene of Venezuela (Wood, 1976a). Neither of these seem to have been directly ancestral to the living species, thereby indicating that different line- ages evolved within the genus (Wood, 1976a). Testudinid turtles are represented in South America by the genus Geochelone, with a world-wide distribution in the Early Tertiary. Fossils having similarities with an extant Asi- atic species have been recorded from Eocene deposits in North America and Africa, and also from the early Oligocene of North Amer- ica, Asia and Western Europe (Auffenberg, 1974 ) . The earliest representative of the genus Geochelone in South America is G. gringorum from the late Oligocene of Patagonia (Simp- son, 1942). Auffenberg (1971) suggested that the ancestors of the South American species, comprising the distinctive subgenus Cltclo- noides, probably entered that continent from the north during the Oligocene or even earlier. It is noteworthy that even though chelonians are known since the Late Creta- ceous, no testudinids have been reported be- fore the late Oligocene, whereas their pres- ence is documented frequently since the late Miocene. Assuming an entrance from the north, this partially could result from the fact that most known early Tertiary reptile-bear- ing sites are in the southern part of the conti- nent. However, an arrival earlier than late Eocene times seems improbable. Geochelone gringorum is closely related, and may be an- cestral, to the living G. chilensis, which be- came adapted to drier conditions (Auffenberg, 1971). Increasingly xeric environments devel- oped in western mid-latitudes east of the Andes since the Pleistocene (Baez and Scil- lato Yane, this volume). The related Pliocene G. gallardoi has been recorded in areas where a marked dry season presumably existed. Geochelone hesterna from the late Miocene of Colombia is thought to be ancestral to the extant species G. carhonaria and G. denticu- lata that now live in the northern part of the continent (Auffenberg, 1971). Although emydid turtles now occur in South America, their presence there during the Tertiary is still uncertain. The existence of remains referable to the Emydidae among the material collected from late Miocene de- posits in Colombia was mentioned by Medem (1966, 196S), but the record has not been substantiated. The extant genus Geocmyda is known from the late Pleistocene of Ecuador. No fossil records of the Chelydridae and Kinosternidae are known in South America. Trionychids, old members of the North Amer- ican faunas, are not present today in South America. Their presence in northern Vene- zuela in the late Pliocene could have resulted from waif dispersal, but their colonization was not successful (Wood and Patterson, 1973). In the Late Cretaceous and Early Tertiary, a peculiar group of land turtles of uncertain affinities, the Meiolaniidae, was also part of the herpetofauna of southern South America. They also have been recorded in Australia, where they survived until Pleistocene times. In summary, pelomedusids are ancient components of the ehelonian faunas of South 1979 BAEZ & GASPARINI: FOSSIL RECORD 41 America, being continuously represented since the Late Mesozoic. Podocnemis, the only liv- ing genus in South America, is known since the Late Cretaceous; it has undergone con- siderable speciation. Chelids also are old members of the fauna, even though no occur- rence prior to the early Eocene is known. No land turtles are known from the Late Meso- zoic and Early Tertiary besides meiolaniids. Their ecological role was assumed later by the testiudinid Geochelone, which entered the continent probably in Eocene-Oligocene times. The fossil record of the lizard family Iguanidae in South America is fragmentary, a situation that contrasts with its significance in the present Neotropical herpetofauna. Al- though the former presence of this group in Asia and Europe has been discarded (Hoff- stetter, 1962; Estes, 1970), fossils referable to this family have been recorded in North America, but none antedates the early Eocene (Estes and Price, 1973). The earliest record of an iguanid is from the Upper Cretaceous of Brasil, and the family is represented by at least five species in the late Paleocene fauna from Itaborai, Brasil, thereby indicating an early radiation in South America (Estes, 1970). Presently-known data are consistent with the suggested Gondwanan origin of iguanids (Cracraft, 1973). The group could have entered North America from the south by waif dispersal, becoming quite diversified by Miocene times ( Robinson and Van Deven- der, 1973). By the Miocene, iguanids attained a wide distribution in South America, but little is known of the taxa present at that time (Baez and Gasparini, 1977). Extant repre- sentatives of the family are recorded from late Pleistocene deposits: Iguana in coastal Ecua- dor ( Hoffstetter, 1970b) and Leiosaurus bellii in central-western Argentina (Van Devender, 1977). The evolution of teiids took place primar- ily in South America, where they are most diverse and widely distributed today. In the Late Cretaceous teiids were present in North America (Estes, 1964, 1969), although these early records do not seem to be of direct sig- nificance in the establishment of the Recent teiids there (Estes, 1970). On the other hand, fossil remains referable to the Teiidae and resembling primitive living South Amer- ican taxa have been recorded from the late Paleocene of Brasil (Estes, 1970). Their ab- sence in the Cretaceous is not surprising be- cause continental Middle and Late Mesozoic faunas are still poorly known in South Amer- ica. Practically all later teiid fossils (Mio- cene-Pleistocene) have been referred to the extant South American genera Tupinambis, Dracaena, Callopistes, and Dicrodon, al- though differences in distributions are evident when compared with their present ranges (Baez and Gasparini, 1977). Very few fossil records of the other lizard families now inhabiting South America are known from that continent. However, gek- konids occur in the upper Paleocene of Brasil. Although this is the earliest known record of the family, the origin of gekkonids could ex- tend back to the Late Mesozoic, upon consid- eration of their relationships with some Juras- sic European and Asiatic taxa (Hoffstetter, 1964; Kluge, 1967). Living South American representatives might have been derived from different sources; the strong African affinities of sphaerodactylines (Kluge, 1967) could be interpreted as resulting from the past connec- tions of those continents (Estes, 1970). The presence of scincids in South America was extended back to the Paleocene ( Estes, 1976 ) . These lizards are seemingly already repre- sented in the Late Cretaceous North Amer- ican faunas (Estes, 1976). Even though the fossil record of lizards in South America is extremely poor, it is evident that iguanids and teiids, both well repre- sented there today, were characteristic com- ponents of the Cenozoic herpetofaunas of that continent, where they underwent consid- erable diversification. Those groups, and also gekkonids, seem to comprise an ancient faunal element. The presence of anguids in South America has not yet been documented by the fossil record. They are present in North America since the Late Cretaceous, and seem to have had an essentially northern dispersal (Hoffstetter, 1962; Estes, 1964; Meszoely, 1970). Snake remains have been recorded in South America from the Late Cretaceous through the Pleistocene. With a few excep- tions, no good descriptions are available, and 42 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 much of the material has been assigned only to families. At least in the Early Tertiary the ophidian fauna was comprised mostly of boids. They were represented by the large snakes of the genus Madtsoia, of the extinct subfamily Madtsoiinae, which were recorded from the late Paleocene and early Eocene of Patagonia. Species of that genus also are known from the Late Cretaceous of Mada- gascar ( Hoffstetter, 1961) and of Niger (Broin et al., 1974), a distribution that sug- gests their derivation from a Gondwanan stock. Remains assigned to the Boidae have been cited from the late Paleocene of Brasil and late Pleistocene of Bolivia. The living South American genus Eunectes occurs in late Miocene deposits of Colombia. Boids probably were also represented in the early- middle Miocene of southern Patagonia and the Pliocene of Parana, Argentina, but those records have not been substantiated (Gas- parini and Baez, 1975; Baez and Gasparini, 1977). Aniliids, now restricted to the Oriental and Neotropical regions, were represented in the Early Tertiary faunas of South America. A snake of controversial phylogenetic position, Dinilysia patagonica, was recorded from the Late Cretaceous of central-western Argentina (Smith Woodward, 1901; Huene, 1929). Its affinities to modern aniliids were pointed out by Estes, Frazzeta and Williams (1970), but it was retained in the monotypic family Dini- lysiidae. However, Dinilysia was considered to be closer to boids, although not ancestral to them, by Rage ( 1977). True aniliids are pres- ent in the late Paleocene of Brasil, but the material has not yet been described. The late Miocene Colombophis portai from Colombia more closely resembles the extant Asiatic Cy- lindrophis than Anilhis, now living in South America (Hoffstetter, 1967b; Hoffstetter and Rage, 1977). The earliest known record of aniliids is from the Middle-Late Cretaceous of Canada (Fox, 1975); their presence in North America is documented from that time through the Early Tertiary (Estes, 1976). The group also occurs in Eocene deposits of Eu- rope (Hoffstetter, 1962; Rage, 1974). Rein- terpretation of relationships of Dinilysia and the known fossil record of aniliids seems to be more consistent with the postulated Laurasian origin of that group (Cracraft, 1973). The evidence, however, is still meager. The earliest known colubrid snake in South America is from late Miocene deposits in Colombia; other fossil colubrids in South America are from the late Pleistocene. True colubrids appear in the late Eocene of Europe (Rage, 1974). In North America their pres- ence has been documented since the early Miocene (Holman, 1976b); they become dominant elements of snake faunas by the late Miocene (Holman, 1976a). Thus, an entry in South America from the north during the Miocene was suggested (Hoffstetter, 1967b; Hoffstetter and Rage, 1977). However, it is noteworthy that the relationship of some Mio- cene colubrids from the United States to living Central or South American forms could indi- cate that they were derived from a more southern source (Tihen, 1964). Fossil viperids, which were referred to the Crotalinae, are known in South America only from late Pleistocene deposits in Bolivia. They could be late immigrants from the north (Reig, 1962; Baez and Gasparini, 1977), al- though their arrival in South America should have preceded that record. In the early Mio- cene, typical viperids appear in Europe (Hoffstetter, 1962). Viperid remains, tenta- tively assigned to the Crotalinae, were re- ported from the middle Miocene of North America ( Holman, 1976c ) . All evidence indi- cates, as in the case of colubrids, that the early history of this group is still largely unknown. The fossil record of snakes in South Amer- ica is not only meager but remains practically unstudied. At least during the Early Tertiary henophidians predominated, although lin- eages different from those evolving in the northern continents seem to have been present there. Available paleontological data do not give much information concerning the origin of cenophidians in the South American con- tinent. Scolecophidians are not yet recorded. The Crocodylidae in South America now are restricted to the northern part of the con- tinent. However, among the early Paleocene taxa assigned to that group, Necrosuchus ion- en-sis Simpson, 1937, inhabited southern Ar- gentina. The affinities of this crocodilian have not been clearly established. Crocodylids ap- pear in the fossil record in Cretaceous de- 1979 BAEZ & GASPARINI: FOSSIL RECORD 43 posits of comparable age in North America, Africa and Asia (Sill, 1968). Thus, it is doubtful if the ancestors of Necrosuchus came from North America, at this time is is impos- sible to determine their area of origin. From the Oligocene onward, crocodylids are re- corded north of the Amazonian Basin, with a latitudinal distribution similar to the present one. No fossil Crocodylus is known from South America, except for a doubtful record from the Pliocene of Maranhao, coastal Brasil (Maury, 1923). Available evidence suggests that the living representatives of that genus are late immigrants from the north. The rela- tionships of the peculiar crocodylid, Charac- tosuchus fieldsi Langston, 1965, from the late Miocene of Colombia are still unknown. Alligatorid crocodilians are distributed more widely than crocodylids in South Amer- ica. The earliest known representatives in that continent occur in the late Paleocene of Brasil, and they probably also are present in deposits of that age from southern Argentina. In both cases, the material has not been studied yet. The validity of the Paleocene "Notocaiman stromeri" from Patagonia, pro- posed as an ancestor of Caiman, has been dis- carded (Gasparini and Baez, 1975). The presence of alligatorids in North America ex- tends back to the Cretaceous, and a center of origin there was suggested by Sill (1968). Nevertheless, available data are inconclusive. The Eocene Eocaiman cavernensis is the old- est member of the family described from South America. Langston ( 1965 ) pointed out its modern aspect and suggested that it could have been ancestral to Caiman, Melanosuch- us, and perhaps also to the peculiar Balanero- dus. Kalin ( 1955 ) did not rule out the pos- sibility that it could be referred to the extant genus Caiman. This indicates that taxa close- ly related to living South American represen- tatives of this family were already present on that continent by the Early Tertiary. In the late Miocene of Colombia an extant genus and probably also a species ( Caiman latiro- stris) are known (Gasparini and Baez, 1975). The recent Caiman yacare is known from the Pliocene of Argentina and Melanosuchus from the Pliocene of Venezuela. Today, gavialids are restricted to the Ganges Basin, India, but they may have in- habited South America in the Tertiary. It is still controversial if the South American Ion- girostrine crocodilians are true gavialids or constitute a different lineage ( Langston, 1965; Gasparini, 1968; Sill, 1968, 1970; Hoffstetter, 1970a; Hecht and Malone, 1972; Baez and Gasparini, 1977). The oldest gavialids are known from Eocene deposits in Egypt (Hecht and Malone, 1972). The Asiatic and South American forms could have originated from an African stock; their dispersal would have been facilitated by their capacity to swim in marine waters. In South America the first gavialids appear in deposits of late Oligocene- early Miocene age from the northern part of the continent; they became extinct by the Plio-Pleistocene. The referral of South Amer- ican forms to the genus Gavialis is doubtful, because they are different from the Asiatic members of that genus. The extinct Nettosuchidae is an endemic group from mesic tropical environments of the northern and central part of the continent from the late Miocene to the earliest Pleisto- cene. Crocodilians of the more primitive and extinct suborder Mesosuchia are known in South America during a large part of the Ter- tiary. These forms belong to the family Se- becidae ( Sebecosuchia, Gasparini, 1972) and are known from the Paleocene through the Miocene. They were peculiar, mostly terres- trial crocodiles, and inhabitants of the tropical forests (Langston, 1965; Molnar, 1969, 1977; Neill, 1971; Gasparini, 1972). Sebecids seem to have had a long history on the continent, where they probably originated from bauro- suchids. The great diversity of Cenozoic crocodil- ian faunas in South America is noteworthy. Species referable to five families (Sebecidae, Crocodylidae, Alligatoridae, Gavialidae, Net- tosuchidae) have lived there, but only two of those groups (Crocodylidae, Alligatoridae) are present now. Available evidence suggests that the differentiation of sebecids and netto- suchids occurred in South America. The most conspicuous groups were the Sebecidae and Alligatoridae, the latter being represented throughout the Cenozoic and the most im- portant crocodilian family there today. On the other hand, the fossil record of crocodylids is 44 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 comparatively poor; it seems likely that they never were important components of the South American herpetof auna. The living rep- resentatives of this family are not related to the earliest forms recorded from early Terti- ary deposits in the southern part of the continent. DISCUSSION This review of the fossil record of amphib- ians and reptiles of South America discloses how poor that record actually is and how many questions concerning the history of the groups still remain unanswered. The available paleontological data are insufficient to pro- vide a comprehensive picture of the evolu- tionary, faunal, and distributional changes that occurred in South America. It is perti- nent to emphasize here that practically no special techniques for fossil collecting have been applied; for the most part, the discov- ery of amphibians and reptiles has been hap- hazard. Comparison of local faunas is diffi- cult, for in many cases little weight can be given to the differences in composition of known assemblages; thus, evolutionary and zoogeographic interpretations are still tenta- tive. The records of certain groups, such as tur- tles and crocodilians, greatly outnumber those of other groups which could be partially the result of the fact that most known sites that have yielded amphibian or reptile remains represent lowland and aquatic habitats. This could also explain the absence of other groups associated with different ecological condi- tions. In general, little is known about the heqoetofaunas that inhabited the northern and central part of the continent, not only in the cratonic areas but also in the intercratonic basins. These regions have a considerable significance from the zoogeographic point of view in that they have been considered areas of origin of diverse groups. The oldest known remains referable to modern groups of amphibians and reptiles occurring in South America are from the Late Cretaceous deposits, except for the record of leiopelmatid frogs (Vieraella and Notobatra- chus) in the Jurassic of Patagonia. The next important sample is the late Paleocene rec- ords from Itaborai, Brasil, the earliest assem- blage which offers a broad picture of the groups already comprising the fauna of the continent. According to available data from that sample, little similarity to the North American assemblages of comparable age are evident (Estes, 1976; Baez and Gasparini, 1977). Basically, the family groups recorded from the known Paleocene and Eocene sites evolved in isolation in South America and comprise the present fauna. Some of them such as leptodactylids, bufomds, pelomedu- sids, chelids, teiids, alligatorids, were already represented by forms related to their living representatives on that continent. It is evi- dent that the early history in South America of many of those groups extends back well into the Mesozoic. Unfortunately, Mesozoic records are very scarce, and the earliest known Tertiary samples (except that from Itaborai, Brasil) are badly preserved and not diverse (Gasparini and Baez, 1975). Of the additional family groups recorded from Oli- gocene and Miocene deposits, testudinids and perhaps emydids could have arrived from the north by overwater transport at different times. The known late Cenozoic records pro- vide little information concerning the faunal interchange between the Americas when the isthmian link was established. Different lines of evidence indicate that in the Early Tertiary a humid and warm tem- perate climate prevailed at high latitudes in South America as well as in North America (Dawson, et al., 1976). In the southernmost part of South America, the known assem- blages of that age are archaic, being com- prised of extinct groups that do not now exist in the area. Tectonic events throughout the Cenozoic altered the physiographic condi- tions, and the subsequent climatic and Holistic changes affected the composition of the local faunas. The modification of the Patagonian herpetofauna from the early Eocene onward clearly illustrates this point. The subtropical elements, such as crocodilians, disappeared from that region. Furthermore, the gradual desiccation of climate related to the Andean uplift presumably resulted in the confinement of taxa adapted to humid and aquatic environ- ments to the more mesic western areas. The 1979 BAEZ & GASPARINI: FOSSIL RECORD 45 change of the biota in the region of the pres- ent upper Magdalena Valley, Colombia, in relation to the uplift of the Eastern Cordillera and increasing aridity ( Howe, 1974 ) is an- other example. Many components of the rich late Miocene La Venta fauna, living on the broad flood plains that prevailed there are absent from that now semiarid area (Fields, 1959). Many fundamental aspects in the history of the South American herpetofauna have yet to be elucidated. The amount of paleontolog- ical information that has accumulated in re- cent years enables us to expect that future studies will clarify these problems. ACKNOWLEDGMENTS For their valuable comments and provision of data, we are indebted to Drs. France de Broin, Richard Estes, J. Alan Holman, Bryan Patterson and Roger Wood. Special thanks are extended to Dr. William E. Duellman, who kindly revised and corrected the manu- script. RESUMEN En este trabajo se sintetiza la information disponible sobre el registro fosil de los grupos que integran la actual herpetofauna de Amer- ica del Sur, en un intento de valorar el aporte que el mismo brinda al conocimiento del desarrollo historico de dichos grupos en ese continente. En tal sentido, se han tornado en cuenta no solo los registros sudamericanos, sino tambien aquellos otros directamente rela- cionados y provenientes de otras partes del mundo. Para integrar los resultados en un contexto coherente se considero muy espe- cialmente la disposition y relation de Amer- ica del Sur con respecto a otras masas con- tinentales a partir del Mesozoico medio. Tam- bien se tomaron en cuenta los principales eventos geologicos acaecidos desde fines del Cretacico a la actualidad en dicho continente, valorando su incidencia en los cambios fisio- graficos los que, evidentemente, actuaron sobre la composition y distribution de la herpetofauna. La mention de anfibios y reptiles ceno- zoicos en America del Sur es relativamente frecuente, sin embargo, muchas de las asigna- ciones taxonomicas como las referencias geo- graficas y cronoestratigraficas son dudosas. Ello, conjuntamente con el hecho de que la mayoria de los hallazgos son aislados, sin formar parte de asociaciones representativas, hace que las conclusiones resulten aiin tenta- tivas. En general los datos disponibles per- miten un analisis a nivel familiar. El examen del registro senala la notable antigiiedad, en America del Sur, de muchas de la familias de anfibios y reptiles que viven en ese continente. Todas las familias regis- tradas en el Terciario temprano son inte- grantes de la herpetofauna actual sudameri- cana, excepto las tortugas meiolanidas y los cocodrilos sebecidos, ambos e.xtinguidos. Es evidente que esas familias tienen distinto abolengo, habiendose integrado o diferenci- ado in situ alocronicamente, desde fines del Cretacico al menos. De acuerdo a las eviden- cias paleontologicas la antigiiedad en Ameri- ca del Sur de los pipidos, iguanidos, pelo- medusidos, meiolanidos, muy probablemente los quelidos, y posiblemente los leptodacti- lidos se remonta al Mesozoico tardio. De estas familias, los pipidos, meiolanidos y quelidos son de origen gondwanico. La diferenciacion de los sebecidos y leptodactilidos, en sentido estricto, habria tenido lugar en America del Sur. Los datos paleontologicos son aun in- suficientes para dilucidar el origen de otras familias ya presentes en el Terciario tem- prano. En el Terciario medio y tardio se constata la presencia de familias no registra- das en sedimentos mas viejos. Algunos de esos grupos pudieron haber arribado a Ameri- ca del Sur por medios de dispersion accidental en diferentes momentos durante su aislami- ento. Tal seria el caso de los testudinidos, trioniquidos, gavialidos, crocodilidos directa- mente relacionados a las formas actuales y tal vez los emididos. Los cocodrilos nettosuqui- dos, actualmente e.xtinguidos, se diferenciaron in situ. De la confrontation con la herpeto- fauna actual se desprende que numerosas familias no estan presentes en el registro. Al- gunas de ellas tales como los ranidos, angui- dos, chelidridos y tal vez kinosternidos po- 46 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 drian ser invasores tardios, llegados a traves del Istmo de Panama. Otras por el grado de endemismo y relaciones filogeneticas serian mas antiguas integrantes de la fauna de este continente, no obstante no haberselas regis- trado fosiles hasta el momento. Tal es el caso, por ejemplo, de dendrobatidos y rinoderma- tidos. La localization geografica de los depositos portadores de anfibios y reptiles ha sido tam- bien tomada en consideration. Resulta evi- dente que, con poeas excepciones, es escasa o nula la information disponible sobre las fau- nas de anfibios y reptiles que habitaron la parte norte del continente, tanto en las areas cratonicas como en las cuencas intercraton- icas. Estas regiones revisten especial interes por cuanto se ha sefialado su importancia como areas de origen de diversos grupos. Re- cien a partir del Oligoceno se conocen regis- tros de los grupos considerados en el extremo mas septentrional, fundamentalmente en Co- lombia y Venezuela. En cambio, en la parte sur del continente, en la Patagonia extran- dina, varias han sido las localidades donde depositos del Terciario temprano brindaron restos de anfibios y reptiles. Segun diversas evidancias, en ese tiempo las condiciones am- bientales fueron mas benignas que las actuales en esa region, por lo que la herpetofauna fosil que alii se registra es muy distinta de la que habita ese area en nuestros dias. El registro fosil es aim inadecuado para brindar un panorama integral de los cambios ocurridos en la composition de las diversas faunas regionales en consonancia con las modificaciones ambientales; del mismo modo limita el conocimiento de la evolucion intra- familiar a traves del tiempo. La busqueda sistematica y la aplicacion de tecnicas adecu- adas permitiran, sin duda, un mayor aporte de la paleontologia al conocimiento de la historia de la herpetofauna sudamericana. LITERATURE CITED Ahl, E. 1926. Anura; Aglossa, pp. 141-142 in Kaiser, E., Beetz, W. (eds. ). Die Diamanten- wiiste Siidwest Afrikas 2. D. Reimer, Berlin, 535 p. Ambrosetti, J. 1890. Observaeiones sobre los rep- tiles fosiles oligocenos de los terrenos terciarios antiguos del Parana. Bol. Acad. Nac. Cien. Cor- doba 10:489-499. Ameghixo, F. 1893. Sobre la presencia de verte- brados de aspecto mesozoico en la formacion Santacrueeiia de Patagonia austral. Rev. Jardin Zool. Buenos Aires 1:75-84. Ameghino, F. 1899. Sinopsis geologico-paleontolog- ica. Suplemento. La Plata, Folio 13 p. Antunes, M. T. 1975. Iberosuchus, crocodile sebe- cosuchien nouveau, de l'Eocene iberique au Nord de la Chaine Centrales et l'origine du Canyon de Nazare. Comun. Serv. Geol. Portugal 59:285-330. Archangelsky, S., Romero, E. 1974. Los registros mas antiguos del polen de Nothofagus (Fagaceae) de Patagonia (Argentina y Chile). Bol. Soc. Bot. Mexico 33:13-30. Arid, F., Vizotto, L. 1966. Um quelonio fossil de Sao Jose do Rio Preto. Rev. Cien Cultura 18: 422^28. Auffenberg, W. 1971. A new fossil tortoise, with remarks on the origin of South American testu- dines. Copeia 1971( 1 ) : 106-H7. Auffenberg, W. 1974. Checklist of fossil land tor- toises (Testudinidae). Bull. Florida State Mus. Biol Ser. 18:121-251. Baez, A. M. 1975. Los anuros de la Formacion Las Curtiembres (Cretacico superior), provincia de Salta, Republica Argentina. Evolucion de la familia Pipidae (Amphibia, Anura) en relation a la historia paleogeografiea. Tesis Doctoral Univ. Buenos Aires, 144 p. Baez, A. M. 1976. El significado paleogeografico y paleoecologico de los pipidos (Amphibia, Anura) fosiles de America del Sur. VI Congr. Argent. Geol., Bahia Blanca, 1975, 1:333-340. Baez, A. M. 1977. Sobre Teracopluys Ameghino, nomina nuda (Anura, Leptodactylidae ) de la Formacion Colhue-Huapi (Oligoceno superior), Provincia del Chubut, Republica Argentina. Rev. Asoc. Geol. Argentina, 32:145-151. Baez, A. M., Gasparini, Z. B. de. 1977. Origenes y evolucion de los anfibios y reptiles del Cenozoico de America del Sur. Acta Geol. Lilloana 14:140- 232. Blair, W. F. 1972. Summary, pp. 329-343 in Blair, W. F. (ed. ). Evolution in the genus Bufo. Univ. Texas Press, Austin, 459 p. Bravard, A. 1858. Monografia de los terrenos ma- rinos terciarios del Parana. Diario Oficial de Gobiemo "El Nacional Argentino." (Not seen) Broin, F. de. 1971. Une espece nouvelle de tortue pleurodire (?RoxocIielys vilavilensis n.sp.) dans le Cretace superieur de Bolivie. Bull. Soc. Geol. France (7)13:445-452. Broin, F. de, Buffetaut, E., Koeniguer, J. C, Rage, J. C, Russell, D., Taquet, P., Vergnaud- Grazzini, C, Wenz, S. 1974. La faune de verte- bres continentaux du gisement din Beceten ( Se- nonien du Niger). C. R. Acad. Sci. Paris, Ser. D 279:469-472. 1979 BAEZ & GASPARINI: FOSSIL RECORD 47 Burmeister, H. 1883. Monografia de los terrenos marinos terciarios del Parana. An. Mus. Puhl. Buenos Aires 3:45-94. Burmeister, H. 1885. Examen eritico de los niamf- feros y reptiles fosiles denominados por don Augusto Bravard. Ibid. 3:95-174. Casamiquela, R. 1958. Un anuro gigante del Mio- ceno de Patagonia. Rev. Asoc. Geol. Argentina 13:171-183. Casamiquela, R. 1960. Datos preliminares sobre un pipoideo fosil de la Patagonia. Actas y Trabajos I Congr. Sudamer. Zool., La Plata 1959, 4:17-21. Casamiquela, R. 1963. Sobre un par de anuros del Mioceno de Rio Negro (Patagonia) Wawclia ger- lioldi n. gen. et. sp. ( Ceratophrydidae ) y Gigan- tohatrachus parodii ( Leptodactylidae ) . Ameghini- ana 3:141-162. Casamiquela, R. 1967. Sobre un nuevo Bufo fosil de la Provincia de Buenos Aires ( Argentina ) . Ibid. 5:161-168. Cattoi, N., Freibebc, M. 1958. Una nueva especie de "Podocncmis" del Cretaceo argentine Physis 21 (60): 58-66. Cei, J. 1972. Bufo of South America, pp. 82-90 in Blair, W. F. (ed.). Evolution in the genus Bufo. Univ. Texas Press, Austin, 459 p. Cham, J. 1976. Relaciones de un nuevo Teiidae (Lacertilia) fosil del Plioceno superior, Callopistes bicuspidatus n. sp. Inst. Miguel Lillo, Univ. Nac. Tucuman, Publ. Especial, pp. 133—153. Cracraft, J. 1973. Vertebrate evolution and bio- geography in the Old World tropics, pp. 373-393 in Tarlinc, D., Runcorn, S. (eds.). Implica- tions of continental drift to the earth sciences, Vol. 1. Academic Press, New York, 622 p. Dalziel, I., Kligfield, R., Lowrie, N., Opdyke. N. D. 1973. Palaeomagnetic data from the southernmost Andes and the Antarctandes. Ibid.: 87-101. Dawson, M., West, R., Lancston, W., Hutchison, J. 1976. Paleogene terrestrial vertebrates: north- ernmost occurrence, Ellesmere Island, Canada. Science 192:781-782. Douglas, R., Moullade, M., Nairn, A. 1973. Causes and consequences of drift in the South Atlantic, pp. 513-534 in Tarling, D., Runcorn, S. (eds.). Implications of continental drift to the earth sciences, Vol. 1. Academic Press, New York, 622 p. Emiliant, C, Gaertner, S., Lidz, B. 1972. Neo- gene sedimentation on the Blake Plateau and the emergence of the Central American Isthmus. Palaeogeogr., Palaeoclimatol., Palaeoecol., 11:1- 10. Estes, R. 1961. Miocene lizards from Colombia, South America. Breviora ( 143) : 1—11. Estes, R. 1964. Fossil vertebrates from the late Cretaceous Lance Formation eastern Wyoming. Univ. Calif. Publ. Geol. Sci. 49:1-180. Estes, R. 1969. Relationships of two Cretaceous lizards (Sauria, Teiidae). Breviora (317): 1-8. Estes, R. 1970. Origin of the Recent North Amer- ican lower vertebrate fauna: An inquiry into the fossil record. Forma Functio 3:139-163. Estes, R. 1975a. Xenopus from the Palaeocene of Brazil and its zoogeographic importance. Nature 254:48-50. Estes, R. 1975b. Fossil Xenopus from the Palaeo- cene of South America and the zoogeography of pipid frogs. Herpetologica 31:263-278. Estes, R. 1976. Middle Paleocene lower vertebrates from the Tongue Formation, southeastern Mon- tana. J. Paleontol. 50:500-520. Estes, R., Frazzetta, T., Williams, E. 1970. Stud- ies on the fossil snake Dinilysia patagonica Wood- ward. Part I. Cranial morphology-. Bull. Mus. Comp. Zool. Harvard Univ. 140:25-74. Estes, R., Price, Ll. 1973. Iguanid lizard from the upper Cretaceous of Brazil. Science 180:748-751. Estes, R., Reig, O. 1973. The early fossil record of frogs: A review of the evidence, pp. 11-63 in Vial, J. L. (ed. ). Evolutionary biology of the anurans: Contemporary research on mayor prob- lems. Univ. Missouri Press, Columbia, 470 p. Estes, R., Wake, M. 1972. The first fossil record of caecilian amphibians. Nature 239:228-231. Estes, R., Wassersug, R. 1963. A Miocene toad from Colombia, South America. Breviora (193): 1-13. Fields, R. 1959. Geology of the La Venta badlands, Colombia, South America. Univ. Calif. Publ. Geol. Sci. 23:403-444. Fox, R. 1975. Fossil snakes from the Upper Milk- River Formation (Upper Cretaceous), Alberta. Canad. J. Earth Sci. 12:1557-1563. Gaffney, E. 1975. A revision of the side-necked turtle Taphrosplujs sulcatus (Leidy) from the Cretaceous of New Jersey. Amer. Mus. Novit (2571): 1-24. Gaffney, E., Zangerl, R. 1968. A revision of the chelonian genus Bothremys ( Pleurodira: Pelome- dusidae). Fieldiana Geol. 16:193-239. Gallardo, J. M. 1962. A proposito de Bufo varie- gatus (Giinther), sapo del bosque humedo An- tartandico, y las otras especies de Bufo neotropi- cales. Physis 23 (64):93-102. Gasparini, Z. B. de. 1968. Nuevos restos de Rham- plwstomopsis neogaeus (Burm.) Rusconi 1933 ( Reptilia, Crocodilia) del "Mesopotamiense" (Pli- oceno medio-superior) de Argentina. Ameghini- ana 5:299-311. Gasparini, Z. B. de. 1972. Los Sebecosuchia (Croco- dilia) del territorio argentino. Consideraciones sobre su "status" taxonomico. Ibid. 9:23-34. Gasparini, Z. B. de. 1973. Revision de los Crocodilia (Reptilia) fosiles del territorio argentino. Su evolucion, sus relaciones filogeneticas, su clasifica- cion y sus implicancias estratigrahcas. Tesis Doc- toral, Univ. Nac. La Plata, 169 p. Gasparini, Z. B. de, Baez, A. M. 1975. Aportes al conocimiento de la herpetofauna terciaria de la Argentina. Actas I Congr. Argent Paleontol. Bioestr., Tucuman 1974, 2:377-415. 48 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 Haffer, J. 1970. Geologic-climatic history and zoo- geographic significance of the Uraba Region in northwestern Colombia. Caldasia 10:603-636. Hecht, M. 1963. A reevaluation of the early history of the frogs. Part II. Syst. Zool. 12:20-35. Hecht, M., Malone, B. 1972. On the early history of the gaviaiid crocodilians. Herpetologica 28: 281-284. Heyer, W. R. 1975. A preliminary analysis of the intergeneric relationships of the frog family Lep- todactylidae. Smithsonian Contr. Zool. (199): 1-55. Hoffstetter, R. 1959. Un dentaire de Madtsoia ( Serpent geant du Paleocene de Patagonie ) . Bull. Mus. Natl. Hist. Nat. Zool., Paris, (2) 31:379-386. Hoffstetter, R. 1961. Nouveau.x restes dun ser- pent Boide ( Madtsoia madagascaricnsis nov. sp. ) dans le Cretace inferieur de Madagascar. Ibid. (2) 33:152-160. Hoffstetter, R. 1962. Revue des recents acquisi- tions concernant l'histoire et la systematique des Squamates, pp. 243-279 in Problemes actuels de Paleontologie (Evolution des Vertebres). Colloq. Int. C.N.R.S. Paris 104, 474 p. Hoffstetter, R. 1963. La faune Pleistocene de Tarija (Bolivie). Note preliminaire. Bull. Mus. Natl. Hist. Nat. Zool., Paris (2) 35:195-203. Hoffstetter, R. 1964. Les Sauria du Jurassique superieur et specialment les Gekkota de Baviere et de Mandchourie. Senckenbergiana Biol. 45: 281-324. Hoffstetter, R. 1967a. Remarques sur les dates d'implantation des differents groupes de serpents terrestres en Amerique du Sud. Convp. Rend. Somm. Soc. Geol. France 1967:93-94. Hoffstetter, R. 1967b. Observations additionelles sur les serpents du Miocene de Colombie et recti- fication concernant la date d'arrivee des colubrides en Amerique du Sud. Ibid. 1967:209-210. 32:419-426. Hoffstetter, R. 1968. Nuapua, un gisement de vertebres pleistocenes dans le Chaco Bolivien. Bull. Mus. Natl. Hist. Nat. Zool., Paris (2) 40: 823-836. Hoffstetter, R. 1970a. Vertebrados cenozoicos de Colombia. Actas IV Congr. Latinoamer. Zool., Caracas 1968, 2:931-954. Hoffstetter, R. 1970b. Vertebrados cenozoicos de Ecuador. Ibid. 2:955-970. Hoffstetter, R. 1970c. Vertebrados cenozoicos y mamiferos cretacicos del Peru. Ibid. 2:971-984. Hoffstetter, R., Rage, J. C. 1977. Le gisement de vertebres miocenes de La Venta (Colombie) et sa faune da Serpents. Ann. Paleontol. (Vertebres) 63:161-190. Hol.man, J. 1968. Lower Oligocene amphibians from Saskatchewan. Q. J. Florida. Acad. Sci. 31:273-289. Holman, J. 1976a. Snakes from the Rosebud For- mation ( Middle-Miocene ) of South Dakota. Her- petologica 32:41—48. Holman, J. 1976b. A boid and a colubrid snake from the Harrison Formation ( Lower Miocene, Arikareean) of Sioux County, Nebraska. Ibid. 32:387-389. Holman, J. 1976c. Snakes of the Split Rock Forma- tion ( Middle Miocene ) , Central Wyoming. Ibid. 32:419-426. Howe, M. 1974. Nonmarine Neiva Formation (Plio- cene?) Upper Magdalena Valley, Colombia: re- gional tectonism. Geol. Soc. Amer. Bull. 85:1031- 1042. Huene, F. von. 1929. Los sarisquios y ornitisquios del Cretaceo argentino. An. Mus. La Plata 3:1- 196. Irving, E. M. 1971. La evolucion estructural de los Andes mas septentrionales de Colombia. Inst. Nac. Invest. Geol. Min. Colombia, Bol. Geol. 19: 1-90. Jimenez Fuentes, F. 1971. Los reptiles fosiles del Valle del Duero: Podocnemis carbajosai nov. sp. del Eoceno de Salamanca. Estud. Geol. (Madrid) 27:85-93. Jimenez Fuentes, E. 1975. Duerochelys arribasi, nov. gen., nov. sp., Pelomedusidae du Ludien du Bassin du Duero (Espagne). Bull. Geol. Soc. France (7) 17:410-415. Kalin, J. 1955. Crocodilia, pp. 695-784 in Piveteau, J. (ed.). Traite de Paleontologie, Masson et Cie., Paris 5:1-1113. Kluge, A. 1967. Higher taxonomic categories of gekkomd lizards and their evolution. Bull. Amer. Mus. Nat. Hist. 135:1-59. Langston, W. 1965. Fossil crocodilians from Co- lombia and the Cenozoic history of the Crocodilia in South America. Univ. Calif. Publ. Geol. Sci. 52:1-157. Langston, W. 1966. Mourasuchus Price, Ncttosuch- us Langston, and the family Nettosuchidae ( Rept. Crocodilia). Copeia 1966(4) :882-885. Lemon, R., Chubcher, C. 1961. Pleistocene geology and paleontology of the Talara region, Northwest Peru. Amer. J. Sci., 259:410-129. Lynch, J. 1971. Evolutionary relationships, osteol- ogy and zoogeography of leptodactyloid frogs. Univ. Kansas Mus. Nat. Hist. Misc. Publ. (53): 1-238. Lynch, J. 1973. The transition from archaic to ad- vanced frogs, pp. 133-182 in Vial, J. L. (ed. ). Evolutionary biology of the anurans: Contempo- rary research on major problems. Univ. Missouri Press, Columbia, 470 p. Malfait, B., Dinkelman, M. 1972. Circum-Carib- bean tectonic and igneous activity and the evolu- tion of the Caribbean Plate. Geol. Soc. Amer. Bull. 83:251-272. Maury, C. 1924. Fosseis terciarios do Brasil com descripgao de novas formas cretaceas. Serv. Geol. Min. Brasil, Monogr. 4:1-711. McGowran, G. 1973. Rifting and drift of Australia and the migration of mammals. Science, 180: 759-761. 1979 BAEZ & GASPARINI: FOSSIL RECORD 49 McKenna, M. 1973. Sweepstakes, filters, corridors, Noah's arks, and beached Viking funeral ships in palaeogeography, pp. 295-308 in Tarlinc, D., Runcorn, S. (ed.). Implication of continental drift to the earth sciences. Vol. 1. Academic Press, New York, 622 p. Medem, F. 1966. Contribuciones al conocimiento sobre la ecologia y distribution geografica de Phrynaps (Batrachemys) dahli; (Testudinata, Pleurodira, Chelyidae). Caldasia 9:467-489. Medem, F. 1968. Desarrollo de la herpetologia en Colombia. Rev. Acad. Colombiana Cien. Exactas, Fis. Nat. 13:149-200. Medina, C. 1976. Crocodilians from the Late Terti- ary of Northwestern Venezuela: Melanosuchus fisheri sp. nov. Breviora (438):1-14. Meszoely, C. 1970. North America fossil anguid lizards. Bull. Mus. Comp. Zool. Harvard Univ. 139:87-149. Molnar, R. 1969. Jaw musculature and jaw mechan- ics of the Eocene crocodilian Sebecus icaeorliinus. M.A. Thesis. Univ. Texas, Austin, 89 p. Molnar, R. 1977. Crocodile with laterally com- pressed snout: First find in Australia. Science, 197:62-64. Mones, A. 1972. Lista de los vertebrados fosiles del Uruguay, I. Chondrichthyes, Osteichthyes, Rep- ulia, Aves. Comun. Paleontol. Mus. Hist. Nat. Montevideo 1:23-36. Mones, A. 1975. Notas paleontologieas uruguayas, III. Vertebrados fosiles nuevos o poco conocidos ( Chondrichthyes, Osteichthyes, Amphibia, Mam- malia). Ameghiniana 12:343—349. Mook, C. 1921. Bracliygnalliosuchus braziliensis, a new fossil crocodilian from Brazil. Bull. Amer. Mus. Nat. Hist. 44:43-50. Mook, C. 1941. A new fossil crocodilian from Co- lombia. Proc. U.S. Natl. Mus. 91:55-58. Neill, W. 1971. The last of the ruling reptiles: Alligators, Crocodiles and their kin. Columbia University Press, New York, 486 p. Nevo, E. 1968. Pipid frogs from the Early Creta- ceous of Israel and pipid evolution. Bull. Mus. Comp. Zool. Harvard Univ. 136:255-318. Parodi Bustos, R., Figueroa Caprini, M., Kraglie- vich, J., Del Corro, G. 1960. Noticia preliminar acerca del yacimiento de anuros extinguidos de Puente Morales (Depto. de Guachipas, Prov. de Salta). Rev. Fac. Cienc. Nat. Salta Univ. Nac. Tucuman 1:5-25. Pascual, R. 1970. Evolution de comunidades, cam- bios faunisticos e integraciones biocenoticas de los vertebrados cenozoicos de Argentina. Actas IV Congr. Latinoamer. Zool., Caracas 1968, 2: 991-1089. Pascual, R., Odreman Rivas, O. 1971. Evolucion de las comunidades de los vertebrados del Ter- ciario argentino. Los aspectos paleozoogeograficos y paleoclimatieos relacionados. Ameghiniana 8: 372-412. Patterson, B., Pascual, R. 1968. The fossil mam- mal fauna of South America. Q. Rev. Biol., 43: 409-451. Paula Couto, C. de. 1970. Evolucao de comuni- dades, modificacoes faunisticas e integracoes bio- cenoticas dos vertebrados cenozoicos do Brasil. Actas IV Congr. Latinoamer. Zool., Caracas 1968, 2:907-930. Petkiella, B., Archangelsky, S. 1975. Vegetation y ambiente en el Paleoceno de Chubut. Actas I Congr. Argent. Paleontol. Bioestr., Tucuman, 1974, 2:257-270. Price, Ll. 1953. Os quelonios de Formacao Bauru, Cretaceo terrestre do Brasil meridional. Dept. Nac. Prod. Min., Div. Geol. Min., Rio de Janeiro 147:1-34. Price, Ll. 1954. Um quelonio pleurodiro no Cal- cario da Serie Apodi, Cretaceo do Estado do Rio Grande do Norte. Dept. Nac. Prod. Min., Div. Geol. Min., Notas Prelim. Estud, Rio de Janeiro 85:1-12. Price, Ll. 1964. Sobre o cranio de um grande croco- dilideo extinto do Alto rio Jurua, Estado de Acre. An. Acad. Brasil. Cien. 36:59-66. Rage, J.-C. 1974. Les serpents des Phosphorites du Quercy, Palaeovert, 6:274-303. Rage, J.-C. 1975. Un serpent du Paleocene du Niger. Etude preliminaire sur l'origine des cae- nophidiens ( Reptilia, Serpentes ) . C. R. Acad. Sci. Paris, Ser. D 281:515-518. Rage, J.-C. 1977. La position phyletique de Dinily- sia patagonica, serpent du Cretace superieur. Ibid. 284:1765-1768. Raven, P., Axelrod, D. 1975. History of the flora and fauna of Latin America. Amer. Sci. 63:420- 429. Reic, O. 1958. Notas para una actualization del conocimiento de la fauna de la Formation Chapad- malal. Acta Geol. Lilloana, Tucuman 2:255-283. Reic, O. 1959. Primeros datos descriptivos sobre los anuros del eocretacico de la provincia de Salta, Republica Argentina. Ameghiniana 1:1—8. Reig, O. 1962. Las integraciones cenogeneticas en el desarrollo de la fauna de vertebrados de Ameri- ca del Sur. Ibid. 2:131-140. Reig, O. 1968. Peuplement en vertebres tetrapodes de TAmerique du Sud, pp. 215-260 in Delaware Deboutteville, C, Rapoport, E. (eds.). Biolo- gie de l'Amerique australe. C. N. R. S. Groupe Francais Argiles, C. R. Reun. Edud. 4:1-472. Reig, O. 1972. Macrogenioglottus and the South American bufonid toads, pp. 14—36, in Blair, W. F. (ed.). Evolution in the genus Bufo. Univ. Texas Press, Austin, 457 p. Reyment, R. Tatt, E. 1972. Biostratigraphical dat- ing of the early history of the South Atlantic Ocean. Phil. Trans. Roy. Soc. London, B. Biol. Sci. 264:55-95. Roblnson, M., Van Devendeb, T. 1973. Miocene lizards from Wyoming and Nebraska. Copeia 1973:698-704. Romer, A. 1966. Vertebrate Paleontology. 3rd. ed. Univ. Chicago Press, Chicago, 468 p. Rovereto, C. 1914. Los estratos araucanos y sus fosiles. An. Mus. Nac. Hist. Nat. Buenos Aires 25:1-247. 50 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 Royo Y Gomez, J. 1945-46. Los vertebrados del Terciario continental colombiano. Rev. Acad. Co- lombiana Cien. Exactas Fis. Nat. 6:496-512. Rusconi, C. 1933. Observaciones criticas sobre rep- tiles terciarios de Parana ( Alligatoridae) . Rev. Univ. Nac. Cordoba 20:57-106. Rusconi, C. 1935. Observaciones sobre los gaviales fosiles argentinos. An. Soc. Cient. Argentina 119: 203-214. Rusconi, C. 1946. Ave y reptil oligoceno de Men- doza. Bol. Paleontol., Buenos Aires 21:1-3. Savage, J. 1973. The geographic distribution of frogs: Patterns and predictions, pp. 351—445 in Vial, J. L. (ed.). Evolutionary biology of the anurans: Contemporary research on major prob- lems. Univ. Missouri Press, Columbia, 470 p. Savage, J. 1974. The isthmian link and the evolution of Neotropical mammals. Nat. Hist. Mus. Los Angeles Cty. Contrib. Sci. (260): 1-51. Schaeffer, B. 1949. Anurans from the Early Ter- tiary of Patagonia. Bull. Amer. Mus. Nat. Hist. 93:41-68. Sclater, J., Fisher, R. 1974. Evolution of the East Central Indian Ocean, with emphasis on the tec- tonic setting of the Ninetyeast Ridge. Geol. Soc. Amer. Bull. 85:683-702. Segura, P. 1944. Estudio de la primera especie de tortuga fosil de Costa Rica con algunas generali- dades sobre el orden Testudines. Esc. Farm. Guatemala, 6:9-29. (Not seen) Sige, B. 1968. Dents du micromammiferes et frag- ments de coquilles d'oefs de dinosauriens dans la faune de vertebres du Cretace superieur de La- guna Umayo (Andes peruviennes). C. R. Aead. Sci., Paris, Ser. D 267:1495-1498. Sill, W. 1968. The zoogeography of the crocodiles. Copeia 1968(1 ):76-88. Sill, W. 1970. Nota preliminar sobre un nuevo gavial del Plioceno de Venezuela y una discusion de los gaviales sudamerieanos. Ameghiniana 7: 151-159. Simpson, G. 1933a. A new fossil snake from the Notostylops Beds of Patagonia. Bull. Amer. Mus. Nat. Hist. 67:1-22. Simpson, G. 1933b. A new crocodilian from the Notostylops Beds of Patagonia. Amer. Mus. Novit. (623): 1-9. Simpson, G. 1935. Early and middle Tertiary geol- ogy of the Gaiman region, Chubut, Argentina. Ibid. 775:1-29. Simpson, G. 1937a. New reptiles from the Eocene of South America. Ibid. 927:1-3. Simpson, G. 1937b. An ancient eusuchian crocodile from Patagonia. Ibid. 965:1-20. Simpson, G. 1938. Crossochelys, Eocene horned turtle from Patagonia. Bull. Amer. Mus. Nat. Hist. 74:221-254. Simpson, G. 1942. A Miocene tortoise from Pata- gonia. Amer. Mus. Novit. (1209): 1-6. Simpson, G., Minopiuo, L., Patterson, B. 1962. The mammalian fauna of the Oivisadero Largo Formation, Mendoza, Argentina. Bull. Mus. Comp. Zool. Harvard Univ. 127:239-293. Smith, A., Briden, J., Drewry, G. 1973. Phanero- zoic world maps, pp. 1-42 in Hughes, N. (ed.). Organisms and continents through time. Spec. Pap. Palaeontol. ( Paleo. Assoc. London) 12:1- 334. Smith Woodward, A. 1901. On some extinct rep- tiles from Patagonia of the genera Miolania, Dini- hjsia and Genyodcctes. Proc. Zool. Soc. London, 1901:169-184. Staesche, K. von. 1937. Podocnemis brasiliensis n. sp. aus der Oberen Kreide Brasiliens. Neuens Jahrb. Mineral. Geol. Palaeontol. Abh. Abt. A, 77:291-309. Reprod. by Minist. Agricult., Div. Geol. Min., Brasil, Bol. 114 (1944). Stirton, R. 1953. Vertebrate paleontology and con- tinental stratigraphy in Colombia. Geol. Soc. Amer. Bull. 64:603-622. Suarez, J. 1969. Urn quelonio da formacao Bauru. Dept. Geograf., Fac. Filos. Cien. Letras, Pres. Prudente 2:35-54. Tihen, J. 1962. New world fossil bufonids. Amer. Midi. Nat. 68:1-50. Tihen, J. 1964. Tertiary changes in the herpeto- faunas of temperate North America. Senckenber- giana Biol. 45:265-279. Tihen, ]. 1972. The fossil record, pp. 8-13 in Blair, W. F. (ed. ). Evolution in the genus Bufo. Univ. Texas Press, Austin, 457 p. Van Denvender, T. 1977. Observations on the Ar- gentine iguanid lizard Leiosaurus bellii Dumeril and Bibron ( Reptilia, Lacertilia, Iguanidae). J. Herpetol. 11:238-241. Van Houten, F., Travis, R. 1968. Cenozoic de- posits of the Upper Magdalena Valley, Colombia. Bull. Amer. Soc. Petrol. Geol. 52:675-702. Vergnaud-Grazzint, C. 1966. Les amphibiens du Miocene de Beni-Mellal. Notes Serv. Geol. Ma- roc 27 (198): 4.3-69. Vercnauu-Grazzini, C. 1968. Amphibiens pleisto- cenes de Bolivie. Bull. Soc. Geol. France, (7) 10:688-695. Vilas, J., Valencio, D. 1978. Paleomagnetism of the South American and African rocks and the age of the South Atlantic. Rev. Brasil. Geocien- cias 18:3-10. Vilas, J., Valencio, D. 1979. Paleomagnetism of South American rocks and the Gondwana Conti- nent. Seminar on Past configuration of Gondwana and geological correlation through time. IV In- termit. Gondwana Symp., Calcutta, 1977 (in press ) . Warren, J. 1969. Chelid turtles from the mid-Ter- tiary of Tasmania. J. Paleontol. 43:179-182. Wieland, G. 1923. A new Parana Pleurodira. Amer. J. Sci. 5:1-14. Williams, E. 1950. Testudo cubensis and the evo- lution of Western Hemisphere tortoises. Bull. Amer. Mus. Nat. Hist. 95:1-36. Williams, E. 1956. Podocnemis bassleri, a new species of pelomedusid turtle from the late Terti- ary of Peru. Amer. Mus. Novit. (1782): 1-10. 1979 BAEZ & GASPARINI: FOSSIL RECORD 51 Wood, R. 1970. A review of the fossil Pelomedusi- dae ( Testudines, Pleurodira ) of Asia. Breviora (357): 1-23. Wood, R. 1971. The fossil Pelomedusidae (Testu- dines, Pleurodira) of Africa. PhD Dissert. Har- vard Univ., 345 p. Wood, R. 1972. A fossil pelomedusid turtle from Puerto Rico. Breviora (392): 1-13. Wood, R. 1975. Redescription of "Bantuchelys" con- golensis, a fossil pelomedusid turtle from the Paleocene of Africa. Rev. Zool. Bot. Africaines 89:127-144. Wood, R. 1976a. Two new species of Chelus (Tes- tudines, Pleurodira) from the Late Tertiary of northern South America. Breviora (435): 1-26. Wood, R. 1976h. Stupendemys geographicus, the world's largest turtle. Ibid. (436):1-31. Wood, R., Gamero, M. de. 1971. Podocnemis vene- zueletisis, a new fossil pelomedusid ( Testudines, Pleurodira ) from the Pliocene of Venezuela and a review of the history of Podocnemis in South America. Ibid. (376)1-23. Wood, R., Moody, R. 1976. Unique arrangement of carapace bones in the South American chelid turtle Hydromcdusa maximiliani (Mikan). J. Zool. 59:69-78. Wood, R., Pattersox, B. 1973. A fossil trionychid turtle from South America. Breviora (415):1-10. Zancerl, R. 1947. Redescription of Taphrosphys olssoni, a fossil turtle from Peru. Fieldiana Geol. 10:29-40. APPENDIX Appendix 2:1. — The fossil amphibians and reptiles recorded from the areas shown in figures 4-7 and their cor- responding bibliographic references are listed below; the numbers correspond to those on the maps (Figs. 4-7). The areas have been designated by conspicuous geographic names. Taxonomic entities not recognized in re- cent studies are excluded. CRETACEOUS 1. Alemania, Provincia de Salta, Argentina (Late Cretaceous). anura: Pipidae: Saltcnia ibanezi Reig, 1959 (Reig, 1959; Parodi Bustos et al., 1960; Baez, 1975). 2. Laguna Umayo, Departamento de Puno, Peru (Late Cretaceous). anura: Leptodactylidae ? (Sige, 1968). crocodilia (Sige, 1968). 3. Mossoro, Estado do Rio Grande do Norte, Brasil ( Late Cretaceous ) . testudines: Pelomedusidae: Apodichelys lucianoi Price, 1954. 4. Vila- Vila, Departamento de Cochabamba, Bolivia ( Late Cretaceous ) . testudines: Pelomedusidae: ? Roxochelys vilavilcmis Broin, 1971. 5. Sao Jose do Rio Preto, Estado de Sao Paulo Brasil (Late Cretaceous). testudines: Pelomedusidae: Podocnemis brasilieniis Staesche, 1937 2 (Price, 1953; Arid and Vizotto, 1966; Broin, 1971); Roxochelys wanderleyi Price, 1953; Podocnemis elegans Suarez, 1969. 2 Pieropolis, Estado de Minas Gerais, Brasil (Late Cretaceous). squamata: Sauria: Iguanidae: Prist iguana brasilicnsis Estes and Price, 1973. 6. Northwest of Lago Colhue-Huapi, Provincia del Chubut, Argentina (Late Cretaceous ?). testudines: Meiolaniidae: Niolamia patagonica Ameghino, 1899 (Smith Woodward, 1901; Simpson 1938). PALEOCE NE-EOCE NE 1. Golfo de San Jorge, Provincia del Chubut, Argentina (early Paleocene). testudines: (Gasparini and Baez, 1975; Baez and Gasparini, 1977). crocodilia: Crocodylidae: Necrosuchus ionensis Simpson, 1937 (Gasparini and Baez, 1975). 2. Itaborai, Estado de Rio de Janeiro, Brasil (late Paleocene). anura: Pipidae: Xenopus romeri Estes, 1975 (Estes, 1975a, b); Leptodactylidae (Estes, 1970); Bufoni- dae (Estes, 1970); Hylidae (Estes, 1970). gymnophiona: Caeciliidae: Apodops pricei Estes and Wake, 1972. testudines: Pelomedusidae: Podocnemis sp. (Paula Couto, 1970). squamata: Sauria: Iguanidae (Estes, 1970); Teiidae (Estes, 1970; Paula Couto, 1970); Gekkonidae (Estes, 1970). Serpentes: Boidae (Estes, 1970; Paula Couto, 1970); Aniliidae (Estes, 1970; Hoffstetter and Rage, 1977). crocodilia: Sebecidae: Sebecus sp. (Paula Couto, 1970); Alligatoridae (Paula Couto, 1970). J F. de Broin ( pers. comm. ) considers the assignment of P. brasilicnsis and P. elegans to the genus Podocnemis to be questionable. 52 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 3. Cerro Pan de Azucar, Provincia del Chubut, Argentina (late Paleocene). testudines ( Simpson, 1935). squamata: Serpentes: Boidae: Madtsoia cf. M. bai Simpson, 1933 (Simpson, 1935; Hoffstetter, 1959). crocodilia (Simpson, 1935). 4. Laguna del Hunco, Provincia del Chubut, Argentina (late Paleocene). anura: Pipidae: Xenopus pascuali ( Casamiquela, 1960) (Estes, 1975b; Gasparini and Baez, 1975; Baez, 1976). testudines: Pleurodira (Gasparini and Baez, 1975). 5. Mina Aguilar, Provincia de Jujuy, Argentina (late Paleocene-early Eocene). testudines (Gasparini and Baez, 1975). crocodilia (Gasparini and Baez, 1975). Quebrada de Humahuaca, Provincia de Jujuy, Argentina (late Paleocene-early Eocene). testudines: Pelomedusidae: Podocnemis argentinensis Cattoi and Freiberg, 1958 (Gasparini and Baez, 1975; Baez and Gasparini, 1977). 6. Canadon Hondo, near Paso Niemann, Provincia del Chubut, Argentina (early Eocene). anura: Leptodactylidae: Caudivcrbera casamayorensis (Schaeffer, 1949) (Lynch, 1971). testudines: Meiolaniidae : Crossochelys corniger Simpson, 1937 (Simpson, 1937a, 1938); Chelidae: Hij- dromcdusa sp. (Wood and Moody, 1976). crocodilia: Sebecidae: Sebecus icaeorhinus Simpson, 1937. 7. Lago Colhue-Huapi, Provincia del Chubut, Argentina ( early Eocene ) . crocodilia: Alligatoridae: Eocaiman caverensis Simpson, 1933b. 8. Negritos, Departamento de Piura, Peru (middle Eocene). testudines: Pelomedusidae: Taphrosphtjs olssoni (Schmidt) (Zangerl, 1947; Gaffney, 1975). 9. Divisadero Largo, Provincia de Mendoza, Argentina (late Eocene). testudines (Simpson et ah, 1962). squamata: Serpentes (Simpson et al., 1962). crocodilia: Sebecidae ?: Ilchunaia parca Rusconi, 1946 (Gasparini, 1972). OLIGOCE NE-MIOCE NE 1. Campo Waldo, Departamento de Santander, Colombia (Oligocene). 3 crocodilia: Sebecidae: Sebecus sp. (Langston, 1965); Crocodylidae ( Langston, 1965). testudines (Stirton, 1953). 2. Scarrit Pocket, Provincia del Chubut, Argentina (early Oligocene). anura: Leptodactylidae: Caudiverbera caudiverbcra (Linnaeus) (Schaeffer, 1949; Lynch, 1971); Eu- psophus sp. (Schaeffer, 1949); Neoprocoela edentatus (Schaeffer, 1949). 3. Tremembe, Estado de Sao Paulo, Brasil (early Oligocene). testudines: Chelidae (Wood and Patterson, 1973). 4. South of Lago Colhue-Huapi, Provincia del Chubut, Argentina (late Oligocene). anura: Leptodactylidae: Caudivcrbera sp. (Schaeffer, 1949; Baez, 1977). 5. Chaparral, Departamento de Tolima, Colombia (late Oligocene-early Miocene). crocodilia: Alligatoridae: Balancrodus logimus Langston, 1965; Gavialidae (Langston, 1965). 6. Gaiman, Provincia del Chubut, Argentina (late Oligocene). testudines: Testudinidae: Geochelonc gringorum (Simpson, 1942) (Williams, 1950; Auffenberg, 1971; de la Fuente, pers. comm.). 7. Southern Provincia de Santa Cruz, Argentina (early -middle Miocene). squamata: Sauria: Iguanidae (Ameghino, 1899; Gasparini and Baez, 1975; Baez and Gasparini, 1977); Teiidae: Diasemosaurus occidentalis Ameghino, 1893 (Gasparini and Baez, 1975); Serpentes (Ameghino, 1899). 8. Coyaima, Departamento de Tolima, Colombia (late Miocene). testudines: Chelidae: Chelus colombianus Wood, 1976a. squamata: Sauria: Teiidae: cf. Tupinambis (Estes, 1961). crocodilia: Sebecidae: Sebecus sp. (Langston, 1965); Alligatoridae (Langston, 1965); Crocodylidae (Langston, 1965); Gavialidae: ? Gavialis colombianus Langston, 1965. 9. Carmen de Apicala, Departamento de Tolima, Colombia (late Miocene). testudines: Pelomedusidae (Royo y Gomez, 1945-1946; Stirton (1953); Chelidae: Chelus colombianus Wood, 1976a. crocodilia: Alligatoridae: Eocaiman sp. (Langston, 1965); Caiman neivensis Mook, 1941 (Langston, 1965). 3 The Tertiary amphibian and reptile bearing deposits of Colombia are assigned chronologically according to Van Houten and Travis (1968) and Irving (1971); in an earlier paper the authors (1977) followed Stirton (1953). 1979 BAEZ & GASPARINI: FOSSIL RECORD 53 10. Quebrada La Venta, Villavieja, Departamento de Huila, Colombia (late Miocene). anura: Bufonidae: Bufo marinus Linnaeus (Estes and Wassersug, 1963). testudines: Pelomedusidae: Podocnemis expansa ( Schvveigger, 1912) (Medem, 1966, 1968); Chelidae: Chelus colombianus Wood, 1976a; Testudinidae: Geochelone (Chelonoides) hesterna Auffenberg, 1971; Em- ydidae (Medem, 1968). squamata: Sauria: Iguanidae (Estes, 1961); Teiidae: Tupinamhis cf. T. tequixin (Estes, 1961); Dra- caena colombiana Estes, 1961; Serpentes: Aniliidae: Colombophis portai Hoffstetter and Rage, 1977; Boi- dae: Eunectes stirtoni Hoffstetter and Rage, 1977; Colubridae ( Hoffstetter, 1967b; Hoffstetter and Rage, 1977). crocodilia: Sebecidae: Sebecus huilensis Langston, 1965; Scbecus sp. (Langston, 1965); Alligatoridae: Eocaiman sp. (Langston, 1965); Caiman neivensis Mook, 1941 (Langston, 1965); Caiman cf. C. latirostris ( Daudin, 1802) (Langston, 1965; Baez and Gasparini, 1977); Crocodylidae: Charactosuchus fieldsi Lang- ston, 1965; Nettosuchidae: Mourasuchus atopus Langston, 1965 (Langston, 1966); Gavialidae: cf. Rham- phostomopsis (Langston, 1965). 11. North of Lago Buenos Aires, Provincia de Santa Cruz, Argentina (late Miocene). anura: Leptodactylidae: Caudiverbera caudiverbera Linnaeus ( Casamiquela, 1958; Lynch, 1971). 12. Barranca de los Loros, Provincia de Rio Negro, Argentina (late Miocene). anura: Leptodactylidae: Caudiverbera caudiverbera Linnaeus (Casamiquela, 1963; Lynch, 1971). 13. Ingeniero Jacobacci, Provincia de Rio Negro, Argentina (late Miocene). anura: Leptodactylidae: Wawelia gerholdi Casamiquela, 1963. PLIOCENE-PLEISTOCENE 1. Urumaco, Estado de Falcon, Venezuela (middle Pliocene). testudines: Pelomedusidae: Stupetidemys geograpliicus Wood, 1976b; Chelidae: Chelus lewisi Wood, 1976a; Trionychidae (Wood and Patterson, 1973); Testudinidae (Wood and Patterson, 1973). crocodilia: Alligatoridae: Melanosuchus fisheri Medina, 1976; Crocodylidae ?: Gryposuchus sp. (Pat- terson, pers. comm.); Gavialidae: Ikanogavialis gameroi Sill, 1970; Nettosuchia: Mourasuchus amazon- ensis Price, 1964 (Patterson, pers. comm.). 2. Valle de Santa Maria, Provincia de Catamarca, Argentina ( middle Pliocene ) . testudines: Testudinidae: Geochelone gallardoi ( Rovereto, 1914) (Auffenberg, 1974). 3. Parana, Provincia de Entre Rios, Argentina (middle-late ? Pliocene). testudines: Testudinidae : Geochelone sp. (Gasparini and Baez, 1975); Chelidae (Wieland, 1923). squamata: Sauna: Teiidae (Ambrosetti, 1890; Ga;parini and Baez, 1975; Baez and Gasparini, 1977); Ser- pentes: Boidae (Bravard, 1858; Burmeister, 1883, 1885). crocodilia: Alligatoridae: Caiman latirostris (Daudin, 1802) (Gasparini and Baez, 1975); C. australis (Burmeister, 1885); cf. C. jacare (Daudin, 1802) ( Gasparini and Baez, 1975); C. sp. (Gasparini and Baez, 1975; Baez and Gaspirini, 1977); Gavialidae: Rhamphostomopsis neogaeus (Burmeister, 1885) (Rusconi, 1933, 1935; Gasparini, 1968). 4. Monte Hermoso, Provincia de Buenos Aires, Argentina (late Pliocene). anura: Leptodactylidae: Ceratophrys prisca Ameghino, 1899 (Rovereto, 1914); Bufonidae (Gasparini and Baez, 1975). testudines: Testudinidae: Geochelone gallardoi (Rovereto, 1914) (Auffenberg, 1974). squamata: Sauria: Teiidae: Tupinamhis sp. (Rovereto, 1914); Callopistes bicuspidatus Chani, 1976. 5. Rio Quequen Salado, Provincia de Buenos Aires, Argentina (late Pliocene). anura: Bufonidae: Bufo pisanoi Casamiquela, 1967. 6. Chapadmalal, Provincia de Buenos Aires, Argentina (late Pliocene). anura: Bufonidae: Bufo pisanoi Casamiquela, 1967; Leptodactylidae: Ceratophrys sp. ( Reig, 1958). 7. Rio Jurua, Estado do Acre, Brasil (Plio-Pleistocene). testudines: Pelomedusidae: Podocnemis sp. (Paula Couto, 1970); Chelidae: Chelus sp. (Paula Couto, 1970); Testudinidae: Geochelone sp. (Paula Couto, 1970). squamata: Serpentes (Paula Couto, 1970). crocodilia: Alligatoridae: Brachygnatosuchus brasiliensis Mook, 1921; Purussaurus sp. (Paula Couto, 1970); Crocodylidae ?: Gryposuchus (Paula Couto, 1970); Gavialidae: ? Gavialis (Paula Couto, 1970; Baez and Gasparini, 1977); Nettosuchidae: Mourasuchus sp. (Paula Couto, 1970); Mourasuchus amazon- ensis Price, 1964. 8. Rio Aguaytia, West of Rio Ucayali, Peru (Pliocene ?; Pleistocene ?). testudines: Pelomedusidae: Podocnemis bassleri Williams, 1956. 9. Arroyo Perico Flaco, branch of Rio Negro, Departamento de Soriano, Uruguay (Pleistocene). anura: Leptodactylidae: Leptodactylus sp. (Mones, 1975). 54 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 Geoemyda ( = Callopsis) sp. Dicrodon (Hoffstetter, 1970b). Bufonidae: Bufo cf. B. marinus 10. Peninsula de Santa Elena, Provincia de Guayas, Ecuador (late Pleistocene). testudines: Testudinidae: Geoclielone sp. (Hoffstetter, 1970b); Emydidae: (Hoffstetter, 1970b). squamata: Sauria: Iguanidae: -Iguana ( Hoffstetter, 1970b); Teiidae: crocodilia: Alligatoridae (Hoffstetter, 1970b). 11. Tarija, Departamento de Tarija, Bolivia (late Pleistocene). anura: Leptodactylidae; Ceratophrys sp. ( Vergnaud-Grazzini, 1968); horribilis Weigmann (Vergnaud-Grazzini, 1968). squamata: Sauria: Teiidae: Tupinambis tequixin (Hoffstetter, 1963). 12. Nuapua, near Carandaiti, Bolivia — Horizon 1 (middle Pleistocene). testudines: Testudinidae: Geoclielone sp. (Hoffstetter, 1968). Nuapua, near Carandaiti, Bolivia — Horizon 2 (late Pleistocene). anura: Leptodactylidae: Leptodactylus cf. L. ocellatus (Linnaeus) (Vergnaud-Grazzini, 1968); Cera- tophnjs cf. C. oiuata (Bell) (Vergnaud-Grazzini, 1968); Bufonidae: Bufo cf. B. paracnemis (Vergnaud- Grazzini, 1968). squamata: Sauria: Teiidae: Tupinambis tequixin (Linnaeus) (Hoffstetter, 1968); Serpentes: Boidae (Hoffstetter, 1968); Colubridae (Hoffstetter, 1968); Viperidae (Crotalinae) (Hoffstetter, 1968). amphisbaenia: Aiuphisbaenidae : Leposternon ? (Hoffstetter, 1968). 3. Herpetofaunal Relationships Between Africa and South America Raymond F. Laurent Investigator Titular Fundacion Miguel Lillo Miguel Lillo 205 4000 Tucumdn, Argentina At the height of the Matthewsian theory of continental biogeography ( Matthew, 1915; Darlington, 1957), the very title of this paper would have been almost preposterous, at least in the influential herpetological centers of North America dominated by Noble (1931), Dunn (1923, 1931), and Schmidt (1946). The dissident voices of Jeannel (1942), Du Toit (1937) and others were stifled as inconse- quential. Africa and South America were sup- posed to have had no relationships whatever for a very long time. Indeed, the differences between the Ethiopian and Neotropical her- petofaunas are striking and seem to support the idea of independent histories. Most domi- nant groups are represented in Africa and South America by pairs of adaptive equiva- lents or vicarious groups in Simpson's (1965) sense (Table 3:1). Faunal similarities between Africa and South America have been explained by ex- tinctions in the Holarctic Region. In some cases the fossil record seems to bear out this explanation (e.g., turtles of the family Pelo- medusidae). When such paleontological evi- dence was lacking, it often was implied; Dunn ( 1931 ) emphasized that it would not be sane reasoning to postulate a trans-Atlantic land bridge just because no fossils were known from the northern continents. When Dunn made his statement the fossil record was Table 3:1. — Family Groups of Amphibians and Rep- tiles that have Trans-Atlantic Counterparts. South America Africa Leptodactylidae Hylidae Iguanidae Teiidae Boinae Crotalinae Ranidae (sensu Liem, 1970) Hyperoliidae (sensu Liem, 1970) Agamidae + Chamaeleontidae Lacertidae Pythonini Viperinae Investigator Principal del CONICET. much poorer than it is now. Presently, some credence can be given to negative evidence, namely, the lack of fossils, in North America and Europe of groups that elsewhere have extensive fossil records (e.g., turtles and crocodilians). New evidence definitely points to the union of Africa and South America into a sin- gle continent from at least the Late Carbonif- erous until the Cretaceous. Then a graben formed a narrow gulf north and south of a residual bridge between northeastern Brasil and Nigeria. During the last half of the Tu- ranian, the continents were split with the birth of the South Atlantic Ocean, which initially was quite narrow. This scenario was de- scribed by Reyment ( 1975 ) and is supported by convincing geological evidence — paleo- magnetism ( Creer, 1973 ) , fit of the continents ( Dietz and Holden, 1970 ) , sea-floor spreading (Heirtzler, 1968; Francheteau, 1973), and plate tectonics with the fitting of cratons and rocks ( Hurley, 1968; Dietz and Holden, 1970; Douglas et al., 1973). The stratigraphic con- cordances are especially striking (Reyment and Tait, 1972), as well as the structure of the coastal basins with their immense amounts of fresh water sediments (Martin, 1968), which suggest a graben phase like that of the African Rift Valleys. The salt deposits in An- gola, Gabon and Brasil are reminiscent of later phases like that of Lake Turkana or the Red Sea (Reyment, 1975). Paleontological data are still more con- vincing (Baez and Gasparini, this volume). The amphi-Atlantic distribution of the meso- saurians of the early Permian (Romer, 1966) is strong evidence for a Gondwanan land mass ( Colbert, 1973 ) . The early Triassic of Argen- tina has revealed a Cynognathus fauna almost identical to that of the upper Beaufort beds of South Africa (Bonaparte, 1967). Two 55 56 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 groups of mammals provide evidence that the Atlantic was not the wide ocean of today in the Eocene and Oligocene. Primates and cav- iomorph rodents, conspicuously lacking in the early South American mammalian fauna, nev- ertheless entered the continent long before the Late Cenozoic invasion via Panama. Classically, it was assumed that they entered South America by island-hopping from North America (Wood, 1950; Simpson, 1965; Patter- son and Pascual, 1968). However, another hypothesis assuming an African origin and a trans-Atlantic migration has been maintained by Lavocat (1974) and Hoffstetter (1972, 1977). Krommelbein (1971) showed that the fossil, fresh water ostracods from the coastal basins of Reconcavo, Sergipe, Brasil, and of Gabon are almost identical. Two different am- monite faunas (one in the Potiguar Basin in northern Brasil and another in the Sergipe- Alagoas Basin) existed until the lower Turan- ian. Then the African-Brasilian isthmus disappeared, and the two faunas mingled (Beurlen, 1961; Reyment, 1958, 1970). These data provide evidence for the birth of the Atlantic Ocean and the separation of Africa and South America in the middle Turanian, about 90-95 m.y.b.p. This establishes a firm basis on which to proceed to determine which groups existed in both continents before their separation and which emigrated from one to the other before or after their separation. The Hennigian sys- tematists and biogeographers ( Brundin, 1966; Croizat, 1964; Nelson, 1973; Croizat et al., 1974) have insisted on the necessity of vi- cariance events in determining generalized tracts in reconstructing biogeographic his- tories. Their conclusions are in general agree- ment with plate tectonics, mainly because their work has been based on these premises rather than because they emphasize vicari- ance at the expense of geocenters and dis- persal. (Appendix 3:1). Their approach is applicable to the study of relationships be- tween the African and South American faunas. There was an old African-Brasilian (Inabre- sian, fide Jeannel, 1942) fauna, and there also are vicariant groups. SUMMARY OF DISTRIBUTION PATTERNS The first step in a biogeographic analysis must be a summary of distribution patterns or "generalized tracts" (Croizat, 1964). These patterns are listed below. 1. Gondwanan or West Gondwanan groups. — Geotry petes- Apodops, Pipi- dae, Pelomedusidae, Amphisbaenidae, Typhlopidae, Lcptotyphlopidae. 2. South American groups that invaded Africa before the separation of the con- tinents. — Bufonidae, Iguanidae ( ex- tinct in Africa but surviving in Mada- gascar). 3. Presumed Indian groups that invaded South America via Africa before the separation of the continents. — Microhy- lidae. 4. African groups that invaded South America after the separation of the continents. — Gavialidae (?), Gekkoni- nae, Scincidae, Amphisbaenidae (?), Colubrinae ( ? ) . 5. Holarctic groups tliat invaded Africa and South America from the north. — Testudinidae, Crocodylidae, Colubri- nae (?). 6. African groups that recently invaded South America by a northern route. — Ranidae. 7. Neotropical groups absent from Afri- ca. — Rhinatrematidae. Dermophiinae, 1 Caeciliidae, Typhlonectidae, Bolito- glossini, Leptodactylidae,-' Hylidae, 3 Centrolenidae, Pseudidae, Dendrobati- dae, Sphacrodactylinae, Iguanidae, Teiidae, Anguidae, Boini, Xenodonti- nae, Micrurinae, Crotalinae. 1 The reasons for splitting the Caeciliidae and recog- nizing the Herpelinae are given by Laurent (in press ) . - The African Heleophryninae are considered to be members of the Myobatrachidae by Lynch (1973). 3 The Hylidae, as well as other groups, like Disco- glossidae, Pelobatidae, Salamandridae, and Angui- dae, are present in northern Africa. That part of Africa belongs to the Palaearctica Region and is not considered to be relevant here, although the past existence of some of its fauna in the Ethiopian Re- gion cannot be ruled out entirely. 1979 LAURENT: AFRICA AND SOUTH AMERICA 57 8. African and Old World groups absent from South America. — Scolecomorphi- dae, Herpelinae, Heleophryninae, Hy- peroliidae, Agamidae, Chamaeleonti- dae, Lacertidae, Cordylidae, Varanidae, Pythonini, Lycodontinae, Dasypeltinae, Elapidae, Viperinae. 9. Groups that apparently once lived in Africa but are now extinct there. — Igua- nidae. 10. Groups that presumably icere pan- Gonduanan in the Jurassic. — Leiopel- matidae. Two patterns are more prevalent than others — groups present in South America but not in Africa, and present in Africa but not in South America. A third pattern involves groups that are present in both continents. RELATIONSHIPS OF PATTERNS TO CONTINENTAL HISTORIES Amphibians Caecilians. — The distribution of caecilians indicates that they are a typical Gondwanan group. Estes and Wake (1972) described the only known fossil caecilian, Apodops, from the Paleocene of southeastern Brasil. The fossil tends to confirm a Gondwanan distribution, especially because Apodops resembles the Af- rican Geotrypetes and may be closely related to it. Possibly Geotrypetes is more archaic than either of the primitive families Rhinatremati- dae and Ichthyophiidae. It is said to have a free, dentate ectopterygoid, a bone generally lost and always edentulous in other caecilians. Furthermore, it has two pairs of openings in the skull (temporal and interpterygoidal ) . According to the Lissamphibian hypothesis, such vacuities should be primitive and their disappearance a secondary adaptation to fos- sorial habits. 4 ' This seems more likely than the theory- of Carroll and Currie (1975), who related the Gymnophiona to microsaurians, because the loss of so many cranial bones is more easily visualized as the result of fene- stration than an effect of further solidification of an already continuously roofed skull. Geotrypetes has as many chromosomes as the ichthyophiids, which also include the In- dian Uraeotyphlus (Nussbaum, pers. comm. ). Also, caecilians, like other lissamphibians, demonstrate a negative correlation between chromosome number and the number of de- rived character states (Laurent, in press). Thus, Apodops might belong to an African- Brasilian group of primitive caecilians, from which the African and Neotropical caecilians descended. By this reasoning, the other cae- cilians are actually different in South America and Africa. The South American Rhinatrema- tidae is more primitive than the Ichthyophii- dae (Nussbaum, 1977). The aquatic South American Typhlonectidae and the fossorial African Scolecomorphidae are specialized groups apparently derived from old stocks. The remaining genera have been placed in the Caeciliidae. The specialized Caecilia and Oscaecilia should be separated from the bulk of the family (Laurent, in press). The other genera can be divided into an Old World subfamily Herpelinae with splenial teeth (ex- cept in the specialized Boulengerula) and a New World subfamily Dermophiinae without splenial teeth (except in Gymnophis, there- fore deemed primitive ) ( Laurent, in press ) . These subfamilies are most likely sister groups. Salamanders. — All South American sala- manders are members of the Plethodontidae and obvious immigrants from Central Amer- ica (Wake, 1966). However, I must stress a recent discovery, as interesting as it is unex- pected, of a salamander in Senonian (Upper Cretaceous) beds of Niger (de Broin et al., 1975). The genus and even the family have not been determined. The age is not fixed precisely, because the Senonian is a long period following the Turanian (85 m.y.b.p.) and lasting until the end of the Mesozoic ( Maastrichian, about 65 m.y.b.p.). Salaman- ders have been considered as exclusively Hol- arctic. Therefore, their presence in equatorial Africa, perhaps quite soon after the formation of the Atlantic Ocean, indicates the possibility of an older salamander fauna in Africa and perhaps in South America. Archaeobatrachians. — The only living archaeobatrachians in South America are the 58 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 pipids, which apparently are an early special- ized derivative from Jurassic frogs. The pipids have a typical west Gondwanan distribution and are now common to Africa and South America. They are known from both conti- nents by a number of Mesozoic and Cenozoic fossils, beginning with the Early Cretaceous of Israel (Nevo, 1968). It has been argued that the pipids might have lived in the north- ern continents and subsequently migrated in- dependently into South America and Africa. However, the family is conspicuously absent in the fossil record in the north; it is replaced there by the ecologically similar palaeobatra- chids, now extinct, but apparently common from the Jurassic to the Pliocene. 5 Estes (1975a) reported a fossil of Xenopus from the Paleocene of Brasil and referred the Eocene Shelania from Patagonia to the same genus. Therefore, there is little reason to doubt the Inabresian origin of the family. Estes (1975b) synthesized the phylogeny of the pipids. Of the two lower Cretaceous gen- era from Israel, Cordicephalus seems to be ancestral to Xenopus, whereas Thoraciliacus seems to be an early specialized derivative of the primitive stem of the family. No other fossil is known from the Cretaceous before the separation of Africa and South America. In the Late Cretaceous (Senonian, 80-85 m.y.b.p. ) there is Saltenia in northwestern Argentina and Xenopus and other pipids simi- lar to Pipa or Hijmenochirus in western Af- rica. Eoxenopoides from southwestern Africa (about the limit of the Mesozoic and Creta- ceous, 65 m.y.b.p.) is a specialized derivative of Xenopus. All more recent fossils are re- ferred to Xenopus — two in South America [Paleocene (±60 m.y.b.p.) of southeastern Brasil (Estes, 1975a) and Eocene (±50 m.y.b.p.) of Patagonia (Casamiquela, 1961)] and two in Africa [Miocene ( ±20 m.y.b.p) of southwestern Africa (Ahl, 1926) and of Mo- = According to some workers (e.g., Estes and Reig, 1973), the Palaeobatrachidae is related to the Pipi- dae, but Vergnaud-Grazzini and Hoffstetter (1972) believed that the similarities are the result of con- vergence. However, Estes (1975b) argued con- vincingly that they are in the same superfaniily. "The term "Inabresia" was coined by Jeannel (1942) for the African-Brasilian continent. rocco (Vergnaud-Grazzini, 1966)]. Xenopus presently is speciose in Africa, where the spe- cialized Hijmenochirus and Pseudhymenochi- rus occur in forests. Xenopus is extinct in South America, but perhaps through a Sal- teniaAike stock it gave rise to the modern South American pipids, now restricted to for- ested regions of northern and eastern South America (Estes, 1975b). Other archaeobatrachians have lived in South America ( Baez and Gasparini, this vol- ume). Vieraella (Lower Jurassic) and Noto- batrachus (Upper Jurassic) were placed in the Leiopelmatidae by Estes and Reig ( 1973 ) , who suggested that the leiopelmatids radiated during the Jurassic in Gondwana- land and that Ascaphus is a remnant of a northward migration at the end of the Meso- zoic. Leiopelma survived in New Zealand, where it is the only frog. Discovery of fossil leiopelmatids in other parts of Gondwana- land, especially Africa, is expected. Neobatrachians. — Estes and Reig (1973) and Laurent ( in press ) believed that the neo- batrachians are a Gondwanan group. Possibly the neobatrachians were derived from the southern Discoglossoidea ( Leiopelmatidae ) through a grade exemplified by the Australian family Myobatrachidae, which perhaps for- merly had a pan-Gondwanan range. Among the neobatrachians, only the Bu- fonidae, Microhylidae and Ranidae are pres- ent on both Africa and South America. Of these, the ranids are represented in South America only by Rana pahnipes, a Central American species of recent entry. According to Noble (1931), the bufonids and microhy- lids originated in the Holarctic and subse- quently invaded the southern continents. Blair ( 1972b ) suggested that Bufo originated in South America. Estes and Reig (1973) re- ported Bufo from the Paleocene of Brasil, whereas no other bufomd fossils are known before the Miocene or Oligocene (Hecht, 1963). The fossil record and the rich Neo- tropical radiation of bufonids (9 genera, more than 120 species) support Blair's conclusions. From a South American center of origin, three avenues of bufonid dispersal are con- ceivable — 1) through Antarctica and Austral- ia, 2) through Africa, and 3) through North America. The first obviously is out of the 1979 LAURENT: AFRICA AND SOUTH AMERICA 59 question, because the Australian region is de- void of bufonids. The last was suggested bv Rlair (1972b). Laurent (1972, 1975) did not reject dispersal through North America but proposed that bufonids also dispersed through Africa. Laurent was influenced by Estes' ( 1970, pers. coram. ) insistence that no bufo- nid entered Nortli America before the Mio- cene. There is further evidence in favor of an African dispersal route. 1) The South Amer- ican bufonid radiation is the largest (Trueb, 1971; McDiarmid, 1971; Cei, 1972) and the African is next with seven genera and about 50 species (Tihen, 1960; Tandy and Keith, 1972), followed by Eurasia with only six gen- era and about 40 species and finally North America with one genus and some 20 species. 2) Few bufonids have retained an omoster- num, a plesiomorphic character for the fam- ily. These include the Bufo haematiticus group in northern South America, the African genus Nectophrynoides, and possibly Wemer- ia in west Africa. 7 3) Relationships between the Neotropical and African bufonids is sup- ported by the high degree of genetic compat- ibility between the South American Bufo arenarum and the African B. regularis (Blair, 1972a), by the striking similarity of peculiar species like the African B. superciliaris and the Neotropical B. blombergi (Blair, 1972b), and by serological affinities (Cei, 1977). Such evidence induced Laurent (in press) to em- phasize the African route rather than the North American one. A dispersal following the described route need not exclude a Mio- cene invasion of North America from South America, but the African dispersal was much earlier (±90 m.y.b.p. versus ±25 m.y.b.p.) and therefore much more important to the evolutionary biogeography of the family. The systematic position of the Microhyli- dae is the most controversial matter in the taxonomy of frogs. Boulenger (1882) con- sidered them (as the Engystomatidae) to be related to the Ranidae, because of their firmi- 'Andersson (1903) mentioned the presence of a vestigial omosternum in his description of Steno- glossa, a synonym of Wemeria, but Amiet (1976) said that the omosternum is absent in the genus. Nonetheless, the species of Wemeria resemble toads of the Bufo haematiticus group. sternal pectoral girdle. Noble (1931) sup- ported Boulenger. Orton (1957) emphasized the apparent primitiveness of the microhylid tadpoles, which are similar in many respects to those of pipids. Orton believed that it was unlikely that such an adaptive complex of features as the larval mouth in most anurans would be lost; therefore, she thought that the microhylids were related to the pipids and represented an early radiation among frogs. This contention was resisted by several herpe- tologists beginning with Griffiths (1963). The traditionalists include Griffiths and Car- valho (1965), Tihen (1965), and Kluge and Farris (1969). The "Ortonists" include Hecht (1963), Inger (1967), and Starrett ( 1973 ) , who based her opinion on a detailed study of tadpoles. Savage (1973) enthusi- astically based a new zoogeographic scheme on the apparent strengths of Starrett's conclu- sions, and even included Australia in the original Gondwanan realm of the family ( see Tyler, this volume, for contrary zoogeographic arguments ) . Lynch ( 1973 ) showed that such an evolutionary scheme of the microhylids re- quired the independent acquisition of no less than 13 characters present in the ranids. In a detailed study of tadpole structure, Sokol ( 1975 ) demonstrated that the microhylids had lost the larval papillae, denticles, and horny beaks.- Therefore, the microhylids are related to the ranoids. Microhylids have a semirelictual distribu- tion, intermediate between the scattered pat- terns of some old families, such as the Disco- glossidae and Pelobatidae, and the compact patterns of the more modern, still radiating groups, like the Ranidae and Bufonidae. Therefore, the microhylids must be older than the ranids. This idea is supported by the pres- ence of 28 chromosomes in the microhylid Kaloula; the number is not the result of sec- ondary fusions and therefore is a plesiomor- phic feature similar to the karyotype of Dk- coglossus. The presence of a variety of micro- hylids in South America also supports the ' Blommers-Schlbsser (1975) confirmed Sokol's con- clusions by discovering that in the Scaphiophryninae, the most primitive subfamily of microhylids, the tadpoles have papillae and a slightly sinistral spiracle (median in other subfamilies). 60 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 antiquity of the family; the ranids barely enter South America. Assuming that the microhylids evolved, like the ranoids, in the eastern part of Gond- wanaland, their radiation was around the Indian Ocean ( tropical Asia, East Indies, and Madagascar). The most primitive subfamilies ( Scaphiophryninae and Dyscophinae) live in Madagascar and tropical Asia. The eastern groups in New Guinea ( Sphenophryninae and Asterophryinae) are not relevant here, but the other subfamilies are — Rrevicipinae, Hop- lophryninae 9 and Phrynomerinae in Africa, Cophylinae in Madagascar, and Microhylinae in Asia and America. We can dismiss as improbable an Holarc- tic origin of the family, for the primitive sub- families live in the tropics. Thus, only Mada- gascar and the Gondwanan part of Asia are likely centers of origin of the microhylids. Both are possible, for during its northward drift India apparently was connected at times with Madagascar by garlands of islands (Mc- Kenzie and Sclater, 1973), some of which continue to exist as the Seychelles, Amirante, Mascarene, Maldive and Laccadive Islands (Appendix 3:2). Later, contacts were with Malaysia and Indochina through the Nicobar and Andaman islands, allowing the Indian fauna to invade eastern Asia, Indonesia, and even the East Indies. The African invasion likely passed through the Mozambique Chan- nel and its islands (e.g., Comores). The Af- rican groups of microhylids are now highly differentiated. The American microhylines are a problem. According to the Matthewsian theory, the microhylines originated in tropical Asia, in- vaded the eastern Palaearctic Region, and passed into North America by the Bering isth- mus and into South America by island-hop- ping well before the Pliocene. There are seri- ous objections to this hypothesis. In Asia, as well as in America, the primitive genera hav- ing a complete pectoral girdle are in the trop- ics; these are Kalophrynus, Chaperina, Me- lanobatrachus and Gastrophrynoides in Asia and Otophryiie and Dermatonotus in South Mclanohatrachus is included in the Microhylinae (Savage, 1973; Laurent, in press). America. The genera living in temperate Asia (Microhijla) and North America {Gas- trophryne) have reduced pectoral girdles. Carvalho (1954), Nelson (1966), Nelson and Cuellar ( 1968 ) questioned the affinities be- tween the Asiatic and American microhylines. There is a possible trans-Gondwanan path- way from India to Madagascar to Africa to South America that is marked by a series of genera having complete pectoral girdles — Kalophrynus (tropical Asia), Mclanohatra- chus (India), Scaphiophryninae and Dysco- phus (Madagascar), Brevicipinae and Parho- plophryne (Africa) and Otophryne and Dermatonotus (South America). If such a dispersal took place, it might have occurred before or slightly after the birth of the At- lantic Ocean. Chromosome numbers support this hypoth- esis ( Morescalchi, 1973; Bogart and Nelson, 1976; Bogart et al., 1976). In Asia, Kaloula has 28 chromosomes, which equals the rela- tively primitive number of Discoglossus; other Asian genera ( Uperodon, Ramanella, Micro- hijla ) are known to have 26 chromosomes. In Africa there are 26 in Phrynomerus and 24 in Breoiceps. In America there are 26 chromo- somes in the primitive genera Otophryne and Glossostoma, 24 in Chiasmocleis and 22 in seven genera, including the widespread Gas- trophryne and Elachistocleis. Considering now only those families that are present on one side of the Atlantic, we see some striking parallelisms. South American leptodactylids, rhinodermatids and dendro- batids are paralleled by the terrestrial ranids and hyperoliids in Africa; the hylids and cen- trolenids in South America are paralleled by the arboreal ranids (Chiromantis) and hyper- oliids in Africa. Moreover, some peculiar adaptations in one continent have counter- parts in the other — the aquatic South Amer- ican pseudids (coexisting with pipids) versus African pipids; the rheophilous South Ameri- can telmatobiines versus African heleophry- nines; atelopine bufonids and brachyecpha- lids in South America versus Didynamipus in Africa. In some cases the resemblances are striking. For example, compare Physalacmus biligonigerus in South America with Tomop- terna delalandii (and congeners) in Africa; compare the South American Leptodactylus 1979 LAURENT: AFRICA AND SOUTH AMERICA 61 fuscus with African species of Ptycltadena, and the Hyla leucophyllata group in South America with species of Afrixalus. On the other hand, some adaptations are unique to one continent. Africa has nothing like the marsupial tree frogs (Amphignathodontinae); South America has no frog emulating the sex- ual dichromatism of the tribe Hyperoliini (for other examples, see Laurent, 1973). Reptiles Chelonians. — The Pelomedusidae is a clas- sical case of an amphi-Atlantic distribution, which has been explained by Matthewsians as an Holarctic origin and southward migra- tions. Others have explained the distribution by the fragmentation of a primitively Gond- wanan range. The alternatives are not clear, for there are northern fossils. Some very old turtles from Germany generally classified in other suborders are really pleurodires (de Broin, pers. comm.). These are the Triassic Proterochersis ( Proganochelydia ) and the Ju- rassic Platychelys ( Amphichelydia). 10 Quite an array of other northern genera are known from Upper Cretaceous to Oligocene beds be- longing to the shores of the young Atlantic Ocean. According to de Broin (pers. comm.), the marine coastal Bothremydidae is a sister family of the Pelomedusidae. The oldest pelo- medusid fossil is Platycheloides from the Low- er Cretaceous of Africa. Thus, the family was in existence before the birth of the Atlantic Ocean. Later, the exclusively African pelo- medusines (not known before Oligocene) were separated from the Podocneminae, which flourished in South America, Africa, Europe (Neochelys, Eocene-Oligocene), and even India (Schwoeboemys, Pliocene), and survived in Africa until the Pleistocene, in South America and in Madagascar (Erymno- chelys). The Cryptodira, rather common in Laura- sia in Jurassic and Cretaceous times, are pres- ent in Africa and South America; they seem to be rather recent immigrants into South America — Oligocene for Geochelone (Simp- son, 1942), Miocene for Emydidae (Medem, 1968), a short Pliocene apparition for the Tri- onychidae in Venezuela (Wood and Patter- son, 1973). Only the Testudinidae and Tri- onychidae became established in Africa, where they are known since the Miocene (Romer, 1966). Crocodilians. — The living crocodilians do not show significant similarities between South America and Africa. The Crocodylidae is present in both continents but relatively unimportant in South America; the Alligator- idae, absent from Africa, radiated impressive- ly in South America. However, before the formation of the Atlantic Ocean in the Creta- ceous, the crocodilian fauna, composed ex- clusively of mesosuchians, was much the same in South America and Africa. Thus, the Afri- can Libycosuchidae were small, blunt-snouted crocodiles, very similar to the South American Notosuchidae (Sill, 1968; Buffetaut, 1976)." The gigantic, long-snouted pholidosaurid ge- nus Sarcosuchus was common to Brasil and west Africa (Buffetaut, pers. comm.). After the severance of the last remnants of a bridge, the faunas gradually became different. Al- though the mesosuchians were dominant over the eusuchians until well into the Cenozoic in the scattered Gondwanan continents, they became subordinate to them in Laurasia ( Buf- fetaut, pers. comm.). The Dyrosauridae (Up- per Cretaceous and Early Cenozoic mesosuch- ians), extremely long-snouted, gavial-like creatures, although essentially African, also lived at the end of the Cretaceous on the west side of the still trench-like Atlantic, but this can be ascribed to their littoral habits (Buf- fetaut, 1976). Other gavial-like crocodilians, now extinct, were part of the Neotropical eu- suchian radiation in the Tertiary. Sill (1968) considered them as Gavialidae, but possibly they are another case of parallel evolution (Baez and Gasparini, this volume). If they are true gavialids, their dispersal through a 'Gaffney (1975) specifically removed Proterochersis from the Proganochelydia, because they have a fused pelvis like the pleurodires; he explicitly in- cluded Platychelys in the Pleurodira. 11 Steel (1973) suggested that the groups might have evolved in parallel, but Buffetaut (1976) and Sill ( 1968) recognized two families while admitting that they probably have a common ancestor. Nopsca (1928) recognized them as subfamilies, but Mook (1936), von Huene (1956) and Romer (1956, 1966) did not even make that distinction. 62 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 still narrow ocean in the Eocene seems to be the only explanation of their distribution, for the family is first known in the Eocene ( Hecht and Malone, 1972) of Africa. Lizards. — Only the Gekkonidae and Igua- nidae are involved in the Afro-American sep- aration. Other families of lizards seem to be parallel radiations after the separation — Aga- midae and Chamaeleontidae similar to the Iguanidae; Lacertidae, Cordylidae and Scin- cidae emulating the Teiidae and some terres- trial iguanids; the Varanidae copied by large teiids, like Tiipinaml)is. Also there is the eco- logical similarity between the numerous "mi- croteiids" and the lygosomine skinks (Lau- rent, 1973). Although some archaic lizards were pres- ent in the Triassic, these apparently have little in common with Jurassic ones, which belong to the modern infraorders (Robinson, 1967; Hoffstetter, 1955, 1967). Some of the modern families were present in the Cretaceous. Thus, when Africa and South America drifted apart, several recent families were already in existence. The iguanid, Pristiguana (Estes and Price, 1973 ) , from the Cretaceous of Bra- sil has some characters of the Teiidae. Chrom- osome morphology is similar in iguanids and teiids (Gorman, 1970). The Gekkonidae may be the oldest family of modern lizards. It existed in Brasil in the Paleocene (Estes, 1970). The Jurassic Ardeo- sauridae presumably is ancestral to the gek- konids and so similar to them (Hoffstetter, 1964) that in a cladistic system they could be included in the gekkonids. Therefore, the presence of gekkonids in western Gondwana- land before the formation of the Atlantic graben is realistic. The relict and disjunct distribution of the Eublepharinae is best explained by a northern origin. On the other hand, the Sphaerodac- tylinae is likely to be a strictly Neotropical derivative of a Gondwanan stock. Most of the South American Gekkoninae seem to have come from Africa by waif dispersal after the formation of the Atlantic Ocean. This is fairly certain and recent for the species common to both continents, such as Ilemidactylus hrooki and //. mabouia (Kluge, 1969) and hardly less obvious for Turentohx (Kluge, 1967; Vanzolini, 1968). This dispersal was easier when the Atlantic Ocean was narrower. Other gekkonine stocks (e.g., Briba and Bogertia) may have entered South America in the Late Cretaceous or Early Cenozoic. Possibly some (e.g., Homonota) immigrated into South America before the separation of the conti- nents. Bons and Pasteur (1977) suggested an early immigration for the two Neotropical species assigned to the African genus Lygo- dactylus. Estes and Price ( 1973 ) believed that the Iguanidae originated in South America when it was still united to Africa and invaded Af- rica, where they became extinct, and Mada- gascar, where they survived. Alternatively, they could have originated in Africa, where the related agamids and chamaeleontids sup- planted them. If the Iguanidae and Teiidae have a common ancestor, the presence of teiids in North America in the Cretaceous, contrasting to the absence of iguanids there (Estes, 1970), is puzzling and cannot be ex- plained with our present data. The past existence of iguanids in Africa is hardly questionable, for they are still living in Madagascar. Is their extinction in Africa the result of competition with agamids? Not likely, because the African agamids are not diverse and therefore unlikely to out-compete the diversified iguanids. Also, it is unlikely that the chameleons out-competed the igua- nids, except for possibly some arboreal types, for the chameleons are a highly specialized group of lizards. Furthermore, the agamids are probably relatively recent immigrants into Africa from Eurasia. Rafting of iguanids and teiids from South America to Africa is not possible now, because of the direction of the ocean currents, but 50-80 m.y.b.p. such an event was more likely. In the reverse direc- tion, the feasibility of a successful crossing has been proved by Ilemidactylus (Kluge, 1969), but no cases are documented for agamids, chamaeleontids, lacertids, or cordylids. If, as indicated by Estes and Price (1973), iguanids and teiids are respectively the roots of the Iguania and Scincomorpha radiations, an eastern invasion of "eoteiids" is suggested. Estes ( pers. comm. ) sees the teiids as an essentially Neotropical radiation with a lacer- toid derivation through northern Gondwana- 1979 LAURENT: AFRICA AND SOUTH AMERICA 63 land. The Scincidae have relationships be- tween the southern Atlantic continents, but the few South American species of Mahmja probably came from Africa long after the sep- aration of the continents. Amphisbaenians.- — The worm-lizards have a typical western Gondwanaland distribution. They are an old group, and their existence in the Inabresian continents is likely. Fossils are known from North America (Eocene to Pleis- tocene) and belong to several extinct genera, as well as to the extant Rhincura and Lcpo- sternon, now surviving only in South America (Romer, 1966). The extinct Omoiotyphlops is from the Eocene-Pliocene of Europe ( Rom- er, 1966; Hoffstetter, 1962). Therefore the range of the amphisbaenids underwent a con- traction similar to that of many tropical groups that lived in Europe and North Amer- ica. 12 Snakes. — Among the primitive scoleco- phidians, the Leptotyphlopidae has about the same range as the amphisbaenians and pre- sumably the same history. The Typhlopidae has a pantropical range, which can be deemed pan-Gondwanan until there is contrary evi- dence. The Boidae also is considered to be a Gondwanan group; nonetheless, the present distribution is the result of complex migra- tions. Primitive fossil genera have been found in southern continents — Laparrcntophis in the Lower Cretaceous of northern Africa, Dini- Jysia in Patagonia (Upper Cretaceous), and Madtsoia in South America, Africa, and Mad- agascar (Upper Cretaceous to Paleocene). Presently, the surviving boids in South America and Africa are not closely related. The Neotropical tribe Boini may be de- scended in situ from archaic South American boids, but the African Erycinae and Pythonini probably came from elsewhere. Hoffstetter and Rage (1972) believed that the Erycinae, which may have originated in North America from a South American boine stock, was pres- ent in North America in the Paleocene or earlier and in Europe in the Eocene. Another lineage (Rage, 1977), using the Bering path- way, entered Africa in the lower Miocene. Rage ( pcrs. comm. ) contemplates three pos- sible origins for the Pythonini — Africa or Aus- tralasia, both of which he considers to be doubtful, and Asia, a choice also favored by Underwood (1976, pers. comm.). Thus, boids supposedly migrated from North America to Asia via the Bering land bridge. A fourth possibility is the Indian raft (see Appendix 3:2), which may explain why boids flourish in Australasia and how they later came to Africa from Asia, presumably in Miocene times. 13 The least understood of all groups is the vast array of higher snakes, the caenophidians or Colubroidea. Until recently, the fossil rec- ord of caenophidians was exclusively Holarc- tic and only back to the Miocene. New data show that caenophidians existed in Europe in the lower Eocene, and the Colubridae (sensu lato) is known from the middle Oligocene. 14 Rage (1975) described Nigerophis from the Paleocene of Africa; this genus seems to be intermediate between the caenophidians and the Palaeophidae. Such a systematic position suggests an aquatic origin of modern snakes and makes their paleogeographic history even more difficult to interpret. Rabb and Marx ( 1973 ) suggested that the group perhaps had a tropicopolitan distribution before the Gond- wanan fragmentation, but Rage (1976) dis- agreed. The scarcity and primitiveness of the few colubroids from the early Cenozoic sup- port Rage, rather than Rabb and Marx. The problem is compounded by the taxo- nomic uncertainty that prevails within the Colubridae (sensu lato). Most attempts to clarify the systematica have resulted in 1) recognition of small groups that can be sep- arated from the bulk of the genera, or 2 ) par- tition of larger groups that are highly contro- versial (Dunn, 1928; Bogert, 1940; Bourgeois, 1968; Underwood, 1967; Dowling, 1975; Smith et al., 1977). Dowling's (1975) classification ' The Oligocene fossils, Changlosaurus and Cnjthio- saurus, from Mongolia are not amphisbaenians (Gans, pers. comm.). "Rage (pers. comm.) still prefers the Bering route, but he does not reject the Indian hypothesis. "A record from the upper Eocene (Rage, 1974) is doubtful (Rage, pers. comm.). 64 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 is beautifully simple, resulting in geograph- ically discrete groups, namely New World Xenodontinae ( following Dunn's scheme ) and Old World Lycodontinae. This concept is in agreement with the probable belated birth and radiation of the family as suggested by the fossil record. Thus, parallel radiations in- to terrestrial, aquatic and fossorial groups took place in South America and Africa. On the other hand, Underwood's ( 1967 ) revolutionary classification, which was severe- ly criticized by others (namely, Dowling, 1975; Hoffstetter, 196S; Smith et al., 1977), implied trans-Atlantic relationships. The simi- larities between the fossorial African Cala- melaps and the South American Atractus, the aquatic Limnophis (or Hydraethiops) with Helicops, or the terrestrial Lycophidion and Oxyrhopus simply may be the result of con- vergence. Of course, this is the most likely hypothesis, but rafting should not be dis- missed off-handedly. Three adaptive types of arboreal colubrids correspond to three groups distinguished by Bourgeois (1968), as follows: 1) large-headed snakes with vertical pupils and slender necks (Boiginae); 2) streamlined but rather robust green snakes with round pupils (Philotham- ninae); 3) very slender snakes with round or horizontal pupils, commonly with pointed heads and venom ( Dispholidinae ) . The simi- larities between the Neotropical Leptodeira and the African Boiga-Dipsadoboa-Crotapho- peltis group are probably only convergence, but on the basis of Bourgeois' ( 1968 ) criteria, the Neotropical Oxybelis can be grouped with the African Thelotornis, in spite of its round pupil. 15 Likewise, genera such as Leptophis and Chironius in South America would be- long to the African Philothamninae. Because of the relatively recent develop- ment of the colubrids, it is unlikely that any groups had an Inabresian distribution. Trans- Atlantic rafting may have occurred, but prob- ably only in one direction (Africa to South America). Hoffstetter (1972) convincingly argued that hystricomorph rodents and mon- keys entered South America across the At- ' For osteological reasons, Bourgeois (1968) put Rhamnophis and Thraso))s in the Dispholidinae, notwithstanding their round pupils. lantic Ocean. The reverse migration has never been advocated, except in the beginning of the Atlantic era, when the sea was very nar- row. Now the ocean currents are favorable for westward rafting in the tropics. If this situation prevailed for a long time, it may explain why the Neotropical fauna is so obvi- ously richer than the Ethiopian fauna. South America might have received a sizable fau- nistic contribution from Africa without giving anything in exchange, at least for the last 50 million years. Savitzky (1978) provided evidence that the micrurines are a derivative of Neotropical rear-fanged colubrids, such as Elapomorplms, rather than relatives of the Old World cobras. This removes a zoogeographic problem. Thus, Africa contributed no venomous snakes to South America, for the Crotalinae are absent in Africa. The Viperidae appears in the fossil record in the lower Miocene in northern con- tinents. Its general range suggests an open radiation without insular or peninsular traps. A Laurasian origin is likely, as clearly deduct- ible from the study of Azemiops by Liem et al. (1971). CONCLUSIONS Present geological knowledge indicates that Africa and South America were united and formed a single continent from at least Carboniferous times, during most of the Mesozoic until the Turonian in the Cretace- ous. The final split of the continents and the birth of the Atlantic Ocean occurred 90-95 m.y.b.p. Fossil evidence shows the existence of a common fauna before the Turonian — mes- osaurians of the Permian, the Cynognathns fauna of the Triassic, and mesosuchian croco- diles of the Jurassic and Cretaceous. Both pre-Atlantic and post-Atlantic distributions are hypothesized for groups of amphibians and reptiles (Figs. 3:1). Leiopelmatid frogs existed in the Jurassic in South America, and pipid frogs and pelo- medusid turtles existed in Africa in the Early Cretaceous. The continents were united then, so it is reasonable to assume that leiopelmatids also lived in Africa, and that pipids and pelo- medusids were present in South America be- 1979 LAURENT: AFRICA AND SOUTH AMERICA 65 fore the birth of the Atlantic Ocean. In Late Cretaceous beds of South America, there are doubtful leptodactylids and iguanids. In the Paleocene beds of Brasil there are caecilians, leptodactylids, hylids and bufonids. In Afri- ca there are only an unidentified salamander in the Senonian and a primitive colubroid in the Paleocene. In some groups (bufonids and iguanids) the characteristics of the fossils and the pres- ent range of the group and its inferred phylog- eny are suggestive of a pre-Atlantic western Gondwanan range. In other cases, similar conclusions can be assumed on distributional data alone without the benefit of pertinent paleontological data. Amphi-Atlantic ranges of the gekkonids, amphisbaenians, leptoptyph- lopids and typhlopids are examples. The myo- batrachid frogs have a doubtful fossil in Cre- taceous beds of India. Their presence in western Gondwanaland in Early Cretaceous times is assumed. My hypothesis is that, ac- cording to vicariance principles, a myobatra- chid stock could have become leptodactylids in western Gondwanaland (South America), ranoids in mid-western Gondwanaland (Af- rica), microhyloids in mid-eastern Gondwa- naland (Madagascar, India), and pelodrya- dids in eastern Gondwanaland (Australia). The microhylids are supposed to have radi- ated early enough to spread to Africa and South America just before or after the separa- tion of the continents. The other families generally are quite dif- ferent in Africa and South America, suggest- ing a post-Atlantic radiation. The Neotropical leptodactylids, alligatorids, iguanids, podo- cnemine turtles (which must have crossed Af- rica to reach Madagascar), teiids, anguids, Fie. 3:1. A. Pre-Atlantic Gondwanan distribu- tions — groups present in South America and Africa before the separation. B. Last pre-Atlantic faunistic exchanges — groups that migrated from one continent to the other just before the separation or perhaps soon afterwards. C. Post-Atlantic distributions and later one-way dispersals. A. Distribuciones gondwanense preatldnticas — grupos prescntes en Sudamerica y Africa largo tiempo antes de la division. B. Ultimo intercambio faunistico preatldntico — grupos que migraron de un continente al otro, justo antes dc la division o, tal vez, pronto despues de clla. C. Distribuciones postatldnticas y despues dispersion "waif" unidireccional. MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 boids, xenodontine colubrids, micrurines and crotalines are paralleled, respectively, by Ethi- opian ranoids, crocodylids, agamids, chamae- leontids, pelomedusines, lacertids, cordylids, scincids, pythonines, lycodontine colubrids, elapids and viperines. Some measure of competitive exclusion cannot be ruled out completely, although defi- nite misgivings have been expressed about the universality of Gause's principle (Fryer and lies, 1972). The expansion of three important Neotropical families outside of South America suggests the presence of rival groups is indeed an obstacle. The bufonids have no obvious competitors and are nearly cosmopolitan. The hylids invaded the entire Holarctic Region, where no other tree frogs exist, but failed to spread into the Old World tropics, where other groups of tree frogs occur. The lepto- dactylids, which met ranids when they were barely out of their Neotropical stronghold, barely encroached upon southern North America. Such an effect must be considerably magnified in colonization by rafting, because the indigenous populations have an over- whelming advantage in numbers. Colonizers can succeed only if they enter an empty eco- logical niche or if they are superior to the indigenous species. ACKNOWLEDGMENTS I am grateful to William E. Duellman for his critical comments and to the Consejo Na- cional de Investigaciones Cientificas y Tec- nicas of Argentina and the Fundacion Miguel Lillo, which authorized this work. I am also indebted to various colleagues for valuable help and information — F. de Rroin, E. Ruf- fetaut, R. Estes, C. Gans, J. C. Rage, A. Sa- vitzky, and G. Underwood. RESUMEN El contraste evidente entre las faunas herpetologicas sudamericana y africana par- ece a primera vista apoyar las teorias zoogeo- graficas de Matthew (1915) y Darlington (1957). Los grupos dominantes son complet- amente distintos aunque en general ecologica- mente similares y a menudo relacionados: Dermophiinae al oeste del Oceano y Herpe- linae al este, y asi en seguida, Leptodactylidae y Ranoidea terrestres, Hylidae y Ranoidea arboricolas, Iguanidae y Agamidae (con Chamaeleontidae), Teiidae y Lacertidae, Podocneminae y Pelomedusinae, Alligatoridae y Crocodylidae, Boinae y Pythoninae, Xeno- dontinae y Lycodontinae. Sin embargo, la tectonica de las placas y otros progresos recientes de la geologia com- probaron sin dejar lugar a duda que Africa y Sud America estaban unidas en un solo con- tinente hasta el Turoniano, es decir hasta hace mas o menos 90 millones de anos. Rastros de la comunidad faunistica de esta epoca remota persistieron en antiguos grupos que no dominan la escena, como los Gimno- fionos, Pipidae, Pelomedusidae, Gekkonidae, Amphisbaenidae, Leptotyphlopidae y Typhlo- pidae. Pero, aim en estos ejemplos, la diver- gencia debida a su evolucion por separado durante cerca de 100 millones de anos es generalmente obvia. Hay tambien familias que aparentemente despues de haber nacido en una region occi- dental o oriental del Continente de Gondwana invadieron el resto poco antes de su frag- mentacion o tal vez poco despues, ya que travesias de mares estrechos como son oceanos recien nacidos no presentan difficultades ma- yores. Asi, aparentemente los Bufonidae, Iguanidae, quizas los Teiidae nacidos en Sud- america invadieron Africa, los primeros para seguir en conquista del mundo, los lagartos para evolucionar en otros grupos y/o estar desplazados por ellos ultimamente ( Agamidae y Chamaeleontidae, Lacertidae, Cordylidae y Scincidae). Los Microhylidae, que el autor considera como un antiguo grupo de Neobatracios de origen Indico-Malgache hicieron, al parecer, el viaje inverso, ya que la selva amazonica alberga generos bastante primitivos, como Otophryne. Aun mas tarde hay pruebas de que la travesia del Atlantico no fue imposible, ya que la invasion de America por Gekkonidae de los generos Hemidactylus (Kluge, 1969) y Tarentola se hizo a fines del Cenozoico. Por consiguiente se puede suponer que tales mi- graciones ocurrieron durante todo el Ceno- 1979 LAURENT: AFRICA AND SOUTH AMERICA 67 zoico, con frecuencia decreciente, por supues- to, a medida que los continentes se alejaban. La direction de los corrientes favorece clara- mente las travesias de Este a Oeste, de ma- nera que America del Sur ceso temprano de enriquecer la fauna africana, mientras que al contrario Africa mando probablemente emi- sarios bastante numerosos a Sudamerica, no solamente salamanquesas, sino tambien escin- cidos del genero Mabuija y, tal vez varios gru- pos de culebras arboricolas. Es interesante notar que los anfibios no tuvieron exito en este tipo de aventura, porque, en general no aguantan bien el agua salada y, tal vez la competencia de los Anfibios sudamericanos no permitia lo que se llamo el acceso ecologico en su continente. LITERATURE CITED Ahl, E. 1926. Anura: Aglossa, Xenopodidae, pp. 141-142 in Kaiser, B. (ed. ). Die Diamanten- wiiste Siidwest Afrikas 2. D. Reimer, Berlin. Amiet, J. L. 1976. Observations anatomiques et biologiques sur le genre Werneria Poche, 1903 ( Amphibiens, Anoures, Fam. Bufonidae) . Rev. Zool. Africaine 90:33-45. Andersson, L. C. 1903. Neue Batrachier aus Kam- arun von den Herren Dr. Y. Sjostedt und Dr. S. Junger gesammelt. Verh. Zool. Bot. Ges. Wien 53:141-145. Beurlen, K. 1961. Die palaogeographische Entwick- Iung des siidatlantischen Ozeans. Nova Acta Leo- poldina 24:1-36. Blair, W. F. 1972a. Evidence from hybridization, pp. 196-232 in Blair, W. F. (ed.). Evolution in the genus Bufo. Univ. Texas Press, Austin, 459 p. Blair, W. F. 1972b. Summary, pp. 329-343 in Blair, W. F. (ed.). Evolution in the genus Bufo. Ibid. Blommers-Schlosser, R. M. A. 1975. Observations on the larval development of some Malagasy frogs, with notes on their ecology and biology (Anura: Dyscophinae, Scaphiophryninae and Co- phylinae). Beaufortia 24:7-25. Bogart, J. P., Nelson-, C. E. 1976. Evolutionary implications from karyotypic analysis of frogs of the families Microhylidae and Rhinophrynidae. Herpetologica 32:199-208. Bogart, J. P., Pyburn, W. F., Nelson, C. E. 1976. The karyotype of Otophrync robusta (Anura: Microhylidae). Ibid. 32:208-210. Bocert, C. M. 1940. Herpetological results of the Vernay Angola Expedition, with notes on African reptiles in other collections. Part I. Snakes, in- cluding an arrangement of African Colubridae. Bull. Amer. Mus. Nat. Hist. 77:1-107. Bonaparte, J. F. 1967. New vertebrate evidence for a southern transatlantic connection during the lower and the middle Triassic. Palaeontol. 10: 554-63. Boss, J., Pasteur, G. 1977. Solution histologique a un probleme de taxinomie herpetologique interes- sant les rapports paleobiologiques de l'Amerique de Sud et de l'Afrique. C. R. Acad. Sci. Paris, Ser. D 284:2547-50. Boulenger, G. A. 1882. Catalogue of the Batrachia Salientia S. Ecaudata in the collection of the British Museum, London, 503 p. Bourgeois, M. 1968. Contribution a la morphologie comparee du crane des ophidiens de l'Afrique Centrale. Publ. Univ. Off. Congo Lubumbashi 18:1-293. Bhoin, F. de, Buffetaut, E., Koeniguer, J. C, Race, J. C, Russell, D., Taquet, P., Vercnaud-Graz- zini, C, Wenz, S. 1975. La faune de vertebres continentaux du gisement d'ln Beceten (Senonien du Niger). C. R. Acad. Sci. Paris, Ser. D 279: 469-472. Brundin, L. 1966. Transantarctic relationships and their significance. K. Sven. Vetenskapsakad. Handl. (4)ll(l):l-472. Buffetaut, E. 1976. Sur la repartition geographi- que hors d'Afrique des Dyrosauridae, Crocodiliens Mesosuchiens du Cretace terminal et du Paleo- gene. C. R. Acad. Sci. Paris, Ser. D 283:487-90. Carroll, R. L., Currie, P. J. 1975. Microsaurs as possible apodan ancestors. J. Zool. 57:229-247. Carvalho, A. L. de. 1954. A preliminary synopsis of the genera of American microhylid frogs. Occas. Pap. Mus. Zool. Univ. Mich. (555): 1-19. Casamiquela, R. 1961. Un pipoideo fosil de Pata- gonia. Rev. Mus. La Plata, Secc. Paleontol. 4:71- 123. Cei, J. M. 1972. Bufo of South America, pp. 82-90 in Blair, W. F. (ed.). Evolution in the genus Bufo. Univ. Texas Press, Austin, 459 p. Cei, J. M. 1977. Serological relationships of the Patagonian toad Bufo variegatus (Giinther). Bull. Serol. Mus. 52:2. Colbert, E. H. 1973. Continental drift and the distribution of fossil reptiles, pp. 395—412 in Tarling, D. H., Runcorn, S. K. (eds.). Implica- tions of continental drift in the earth sciences, Vol. 1. Academic Press, New York, 622 p. Creer, K. M. 1973. On the arrangement of the landmasses and the configuration of the geomag- netic field during the phanerozoic, pp. 47-76 in Tarling, D. H., Runcorn, S. K. (eds.). Implica- tions of continental drift in the earth sciences. Ibid. Croizat, L. 1964. Space, time, form: The biological synthesis. Published by author, Caracas, 881 p. Croizat, L., Nelson, G., Rosen, D. E. 1974. Cen- ters of origin and related concepts. Syst. Zool. 23:265-287. Darlincton, P. J., Jr. 1957. Zoogeography: the geographical distribution of animals. John Wiley and Sons, New York, 675 p. Dietz, R. S., Holden, J. C. 1970. The break-up of Pangaea. Sci. Amer. 223:30-41. 68 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 Douglas, R. G., Moullade, M., Nairn, A. E. M. 1973. Causes and consequences of drift in South Atlantic, pp. 517-537 in Tarlinc, D. H., Run- corn, S. K. (eds. ). Implications of continental drift in the earth sciences. Vol. 1. Academic Press, New York, 622 p. Dowling, H. G. 1975. A provisional classification of snakes. 1974 Yearbook Herpetology Herpetol. Infor. Search Syst., New York, pp. 167-170. Duellman, W. E. 1978. The biology of an equa- torial herpetofauna in Amazonian Ecuador. Univ. Kansas Mus. Nat. Hist., Misc. Publ. (65): 1-352. Dunn, E. R. 1923. The geographical distribution of the Amphibians. Amer. Nat. 57:129-136. Dunn, E. R. 1928. A tentative key and arrangement of the American genera of Colubridae. Bull. An- tivenin Inst. Amer. 2:18-24. Dunn, E. R. 1931. The herpetological faunas of the Americas. Copeia 1931(3) : 106-119. Du Toit, A. 1937. Our wandering continents. Oli- ver and Boyd, Edinburgh, 366 p. Estes, R. 1970. Origin of the North American lower vertebrate fauna: An inquiry into the fossil rec- ord. Forma Functio 3:139-163. Estes, R. 1975a. African frog Xenopus (Pipidae) from the Palaeocene of Brasil and its zoogeo- graphic importance. Nature 254:48-50. Estes, R. 1975b. Fossil Xenopus from the Palaeo- cene of South America and the zoogeography of pipid frogs. Herpetologica 31:263-278. Estes, R., Price, L. I. 1973. Iguanid lizard from the upper Cretaceous of Brasil. Science 180: 748-751. Estes, R., Reig, O. 1973. The early fossil record of frogs: A review of the evidence, pp. 11-63 in Vial, J. L. (ed.). Evolutionary biology of the anurans: Contemporary research on major prob- lems. Univ. Missouri Press, Columbia, 470 p. Estes, R., Wake, M. 1972. The first fossil record of caecilian amphibians. Nature 239:228-231. Francheteau, J. 1973. Plate tectonic model of the opening of the Atlantic ocean south of the Azores, pp. 197-202 in Tarlinc, D. H., Runcorn, S. K. (eds.). Implications of continental drift in the earth sciences. Vol. 1. Academic Press, New York, 622 p. Fryer, O, Iles, T. D. 1972. The cichlid fishes of the great lakes of Africa. T. F. H. Publ., Neptune City, N.J., 641 p. GafFx\ey, E. S. 1975. A phytogeny and classifica- tion of the higher categories of turtles. Bull. Amer. Mus. Nat. Hist. 155:391-436. Gorman, G. C. 1970. Chromosomes and the sys- tematics of the family Teiidae (Sauria, Reptilia). Copeia 1970(2):230-245. Griffiths, I. 1963. The phylogeny of the Salientia. Biol. Rev. 38:241-292. Griffiths, I., Carvalho, A. L. de. 1965. On the validity of employing larval characters as major phyletic indices in Amphibia Salientia. Rev. Bra- sileira Biol. 25:113-121. Hecht, M. 1963. A reevaluation of the early history of the frogs. Part II. Syst. Zool. 12:20-35. Hecht, M., Malone, B. 1972. On the early history of the gavialid crocodilians. Herpetologica 28: 281-284. Heirtzler, J. R. 1968. Sea-floor spreading. Sci. Amer. 219:60-70. Hoffstetter, R. 1955. Squamates de type moderne, pp. 606-662 in Piveteau, J. (ed.). Traite de Paleontologie, Paris, 1113 p. Hoffstetter, R. 1962. Revue des recentes acquisi- tions concernant l'histoire et la systematique des squamates. Colloq. Int. C. N. R. S. Paris 104, 474 p. Hoffstetter, R. 1964. Les Sauria du Jurassique superieur et speeialment les Gekkota de Baviere et de Mandchourie. Senckenbergiana Biol. 45: 281-324. Hoffstetter, R. 1967. Coup d'oeil sur les Sauriens ( = Lacertiliens) des Couches de Purbeck (Juras- sique superieur dAngleterre). Colloq. Int. C. N. R. S. Paris 163:349-71. Hoffstetter, R. 1968. [Review of] A contribution to the classification of snakes by G. Underwood. Copeia 1968(1 ):201-213. Hoffstetter, R. 1972. Relationships, origin and history of the ceboid monkeys and caviomorph rodents: A modem reinterpretation. Evol. Biol. 6:323-348. Hoffstetter, R. 1977. Primates: Filogenia e his- toria biogeografica. Stud. Geol. Univ. Salamanca 13:211-253. Hoffstetter, R., Race, J.-C. 1972. Les Erycinae fossiles de France ( Serpentes, Boidae ) . Compre- hension et histoire de la sous-famille. Ann. Paleon- tol. Vertebr. 58:81-124. Huene, F. von. 1956. Palaontologie und Phylogenie der niederen Tetrapoden. Fischer, Jena, 761 p. Hurley, P. M. 1968. The confirmation of continen- tal drift. Sci. Amer. 218:52-64. Inger, R. F. 1967. The development of a phylogeny of frogs. Evolution 21:369-383. Jeannel, R. 1942. La genese des faunes terrestres. Presses Univ. France, Paris, 514 p. Kluge, A. G. 1967. Higher taxonomic categories of gekkonid lizards and their evolution. Bull. Amer. Mus. Nat. Hist. 135:1-59. Kluge, A. G. 1969. The evolution and geographical origin of the New World Hemidactylus nxabouia- brooki complex (Gekkonidae, Sauria). Misc. Publ. Mus. Zool. Univ. Michigan (138): 1-78. Kluge, A. G., Farris, S. J. 1969. Quantitative phy- letics and the evolution of the anurans. Syst. Zool. 18:1-32. Khommelbein, K. 1971. Non-marine Cretaceous os- tracodes and their importance for the hypothesis of Gondwanaland. Proc. 2nd. IUGS Gondwana Symp. South Africa, pp. 617-619. Laurent, R. F. 1972. La distribution des amphib- ious et les translations continentales. 17th. Inter. Cong. Zool., Monaco, 1:1-16. Laurent, R. F. 1973. A parallel survey of equatorial amphibians and reptiles in Africa and South America, pp. 259-266 in Mecgers, B. J., Ayensu, E. S., Duckworth, W. D. (eds.). Tropical forest ecosystems in Africa and South America. Smith- sonian Inst. Press, Washington, D.C., 350 p. 1979 LAURENT: AFRICA AND SOUTH AMERICA 69 Laurent, R. F. 1975. La distribution des amphib- iens et les translations continentales. Mem. Mus. Natl. Hist. Nat. Paris Ser. A Zool. 88:176-191. Laurent, R. F. In press. Systematique et distribu- tion geographique (des Amphibiens ) . in Grasse, P. P. (ed.). Traite de Zoologie 14, fasc. 2 Am- phibiens. Laughton, A. S., Sclater, J. G., McKenzie, D. P. 1973. The structure and evolution of the Indian ocean, pp. 203-212 in Tarling, D. H., Runcorn, S. K. (eds.). Implications of continental drift in the earth sciences. Vol. 1. Academic Press, New York, 622 p. Lavocat, R. 1974. What is an hystricomorph ro- dent? Symp. Zool. Soc. London 34:7-20. Liem, K. F., Marx, H., Rabb, G. B. 1971. The viperid snake Azemiops: Its comparative cephalic anatomy and phylogenetic position in relation to Viperinae and Crotalinae. Fieldiana Zool. 59: 65-126. Liem, S. S. 1970. The morphology, systematics and evolution of the Old World tree frogs (Rhaco- phoridae and Hyperoliidae ) . Fieldiana, Zool. 57: 1-145. Lynch, J. D. 1973. The transition from archaic to advanced frogs, pp. 133-182 in Vial, J. L. (ed.). Evolutionary biology of the anurans: Contempo- rary research on major problems. Univ. Missouri Press, Columbia, 470 p. Martin, H. 1968. A critical review of the evidence for a former direct connection of South America with Africa, pp. 25-53 in Fittkau, E. J., Illtes, J., Klinge, H., Schwabe, G. H., Sioli, H. (eds.). Biogeography and ecology in South America. W. Junk, The Hague, 946 p. Matthew, W. D. 1915. Climate and evolution. Ann. New York Acad. Sci. 24:171-318. McDiarmid, R. W. 1971. Comparative morphology and evolution of frogs of the Neotropical genera Atelopus, Dendrophryniscus, Melanophnjniscus and Oreophrynella. Nat. Hist. Mus. Los Angeles Cty. Sci. Bull. (12;: 1-66. McKenna, M. C. 1973. Sweepstakes, filters, corri- dors, Noah's arks and beached Viking funeral ships in palaeogeography, pp. 295-308 in Tarl- inc, D. H., Runcorn, S. K. (eds.). Implications of continental drift in the earth sciences. Vol. 1. Academic Press, New York, 622 p. McKenzie, D. P., Sclater, J. G. 1973. The evolu- tion of the Indian Ocean. Sci. Amer. 228:62-72. Medem, F. 1968. Desarrollo de la herpetologia en Colombia. Rev. Acad. Colombiana Cien. Exactas Fis. Nat. 13:149-200. Mook, C. C. 1934. The evolution and classification of the Crocodilia. J. Geol. 42:295-304. Morescalchi, A. 1973. Amphibia, pp. 233-348 in Chiarelli, A. B., Cat-anna, E. (eds.). Cyto- taxonomy and vertebrate evolution. Academic Press, New York, 783 p. Nelson, C. E. 1966. The evolution of frogs of the family xMicrohylidae. PhD Dissert. Univ. Texas Diss. Abst. 27, 3719 B. Nelson, C. E., Cuellar, H. S. 1968. Anatomical comparison of tadpoles of the genera Hypopachus and Gastrophryne ( Microhylidae). Copeia 1968 (2):423-424. Nelson, G. 1973. Comments on Leon Croizat's bio- geography. Syst. Zool. 22:312-320. Nevo, E. 1968. Pipid frogs from the Early Cretace- ous of Israel and pipid evolution. Bull. Mus. Comp. Zool. Harvard Univ. 136:255-318. Noble, G. K. 1931. The biology of the Amphibia. McGraw-Hill Book Co., New York, 577 p. Nopsca, F. 1928. The genera of Reptiles. Paleobiol. 1:163-188. Nussbaum, R. A. 1977. Rhinatremidae: A new fam- ily of Caecilians (Amphibia: Gymnophiona ) . Occas. Pap. Mus. Zool. Univ. Mich. (682): 1-30. Orton, G. L. 1957. The bearing of larval evolution on some problems in frog classification. Syst. Zool. 6:79-86. Patterson, B., Pascual, R. 1968. The fossil mam- mal fauna of South America. Q. Rev. Biol. 43: 409-451. Rabb, G. B., Marx, H. 1973. Major ecological and geographic patterns in the evolution of colubroid snakes. Evolution 27:69-83. Rage, J.-C. 1974. Les Serpents des phosphorites du Quercy. Palaeovert. 6:274-303. Rage, J.-C. 1975. Un serpent du Paleocene du Niger. Etude preliminaire sur l'origine des cae- nophidiens (Reptilia, Serpentes). C. R. Acad. Sci. Paris, Ser. D 281:515-18. Rage, J.-C. 1976. Paleontologie, phylogenie et paleo- biogeographie des serpents. PhD Dissert. Univ. Paris. Rage, J.-C. 1977. L'origine des Colubroi'des et des Acrochordoides (Reptilia, Serpentes). C. R. Acad. Sci. Paris, Ser D 286:595-597. Reyment, R. A. 1958. Factors in the distribution of fossil cephalopods. Stockholm Contrib. Geol. 1: 91-184. Reyment, R. A. 1970. Vertically inbedded cepha- lopods shells. Some factors in the distribution of fossil cephalopods. Palaeogeogr. Palaeoclimatol. Palaeoecol. 7:103-111. Reyment, R. A. 1975. Paleontologie evolutive et nouvelle tectonique. Mem. Mus. Natl. Hist. Nat. Paris Ser. A Zool. 88:8-18. Reyment, R. A., Tait, E. A. 1972. Biostratigraph- ical dating of the early history of the South Atlantic Ocean. Phil. Trans. Roy. Soc. London, B. Biol. Sci. 264:55-95. Robinson, P. 1967. The evolution of the Lacertilia. Colloq. Int. C. N. R. S., Paris 163:395-407. Romer, A. S. 1956. Osteology of the reptiles. Univ. Chicago Press, Chicago, 772 p. Romer, A. S. 1966. Vertebrate Paleontology. 3rd ed. Univ. Chicago Press, Chicago, 468 p. Savage, J. M. 1973. The geographic distribution of frogs: Patterns and predictions, pp. 351-445 in Vial, J. L. (ed.). Evolutionary biology of the anurans: Contemporary research on major prob- lems. Univ. Missouri Press, 470 p. 70 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 Savitzky, A. H. 1978. The origin of the New World proteroglyphous snakes and its bearing on the study of venom delivery in snakes. PhD Dissert. Univ. Kansas, 387 p. Schmidt, K. P. 1946. On the zoogeography of the Holarctic Region. Copeia 1946 (3): 144-152. Sill W. D. 1968. The zoogeography of the Croco- dilia. Copeia 1968(1 ):76-88. Simpson, G. G. 1942. A Miocene tortoise from Pata- gonia. Amer. Mus. Novit. (1209): 1-6. Simpson, G. G. 1965. The geography of evolution. Chilton, Philadelphia, 249 p. Smith, H. M., Smith, R. B., Sawin, H. L. 1977. A summary of snake classification ( Reptilia, Ser- pentes). J. Herpetol. 11:115-121. Sokal, R. R., Crovello, T. J. 1970. The biological species concept: A critical evaluation. Amer. Nat. 104:127-153. Sokol, O. M. 1975. The phylogeny of anuran lar- vae: A new look. Copeia 1975 ( 1 ) :l-24. Starbett, P. H. 1973. Evolutionary patterns in lar- val morphology, pp. 251-271 in Vial, J. L. (ed. ). Evolutionary biology of the anurans: Contem- porary research on major problems. Univ. Mis- souri Press, Columbia, 470 p. Steel, R. 1973. Crocodylia. Handbuch der Paliio- herpetologie. 16:1-116. Tandy, M., Keith, R. 1972. Bufo of Africa, pp. 119-170 in Blair, W. F. (ed.). Evolution in the genus Bufo. Univ. Texas Press, Austin, 459 p. Tihen, J. A. 1960. Two new genera of African bufo- nids with remarks on the phylogeny of related genera. Copeia 1960 (3):225-233. Tihen, J. A. 1965. Evolutionary trends in frogs. Amer. Zool. 5:309-318. Trueb, L. 1971. Phylogenetic relationships of cer- tain Neotropical toads with the description of a new genus ( Anura:Bufonidae). Nat. Hist. Mus. Los Angeles Cty. Contr. Sci. (216):l-40. Underwood, G. 1967. A contribution to the classi- fication of snakes. Brit. Mus. (Nat. Hist.), Lon- don, 179 p. Underwood, G. 1976. A systematic analysis of boid snakes, pp. 151-175 in Bellairs, A., Cox, C. B. (eds.). Morphology and Biology of Reptiles. Linnean Soc. London Symp. Series 3, 290 p. Vanzolini, P. E. 1968. Geography of the South American Gekkonidae (Sauria). Arq. Zool. (Sao Paulo) 17:85-111. Vergnaud-Grazzini, C. 1966. Les amphibiens du Miocene de Beni-Mellal. Notes Serv. Geol. Maroc 27:43-69. Vergnaud-Grazzini, C. Hoffstetter, R. 1972. Presence de Palaeobatrachidae (Anura) dans des gisements tertiaires francais. Caracterisation, dis- tribution et affinites de la famille. Palaeovert. 5:157-177. Wake, D. 1966. Comparative osteology and evolu- tion of the lungless salamanders, family Pletho- dontidae. Mem. South. Calif. Acad. Sci. (4):1- 111. Wood, A. E. 1950. Porcupines, paleogeography and parallelism. Evolution 4:87-98. Wood, R., Patterson, B. 1973. A fossil trionychid turtle from South America. Breviora (415): 1-10. APPENDICES Appendlx 3:1. — Vicariance, dispersal and centers of origin. According to the Hennigian and cladistic theories, the concepts of dispersal and centers of origin, cher- ished by so many zoologists from Darwin to Darling- ton, Simpson and Mayr, are useless (Croizat et al., 1974), whereas the concept of vicariance is the master key to biogeographic problems. Perhaps it is true that the importance of vicariance has been over- looked in the zoogeographic conjectures of the past decades, but surely not to the extent implied by these authors. The model of allopatric speciation, one of the main tenets of the synthetic evolutionary theory, is the very basis of the vicariance concept. But this does not necessitate the rejection of other causes of biotic distributions. Dispersal occurs, for there is more than one species at almost every locality (more than 80 species of frogs at Santa Cecilia, Ecuador, as noted by Duellman, 1978). Croizat and his co- workers admit that distributions are not static but fail to allow that dispersal is as important as vicariance. Furthermore, they belittle the concept of the geno- center under the pretext that a species may have a huge array of disjunct populations over an entire con- tinent. Such enormous genetic pools are not espe- cially productive evolutionarily. They impose a great deal of inertia to the spreading of genetic changes. On the contrary, innovative processes, such as genetic drift, genetic revolution, quantum and tachytelic evo- lution, occur in small, transitory and local popula- tions. Nevertheless, it is futile to seek genocenters when adequate data are lacking. It might be said for genocenters, as Sokal and Crovello ( 1970 ) did for the biological species concept, that the concept is not operational. However, this does not preclude the existence of centers of origin, even if we are unable to discover their location, exactly as the non-opera- tionality of the biological species concept does not eliminate the fact that the cessation of gene exchange between two populations is such a momentous event in evolution that it is inconceivable to ignore it, even if we are unable to pinpoint its occurrence. Appendix 3:2. — The Indian raft. It is now generally believed that India drifted away from Antarctica, Madagascar and Africa some- time at the end of the Mesozoic and travelled north- wards through the India Ocean to collide with Laura- sia in the Miocene. Tire dating of the separation is still doubtful — about 100 m.y.b.p. from Antarctica and maybe the Paleocene (60 m.y.b.p) from Mada- gascar. Little attention has been given to the impact of the Indian fauna on the evolutionary zoogeography in the Tertiary. This is an unfortunate omission, not justified by lack of evidence. The Eocene lndo- batrachus seems to belong to the Myobatrachidae, and the primitive snake family Uropeltidae survives in southern India and Sri Lanka. The rationale for my hypothesis is as follows: 1979 LAURENT: AFRICA AND SOUTH AMERICA 71 1. India broke from Madagascar and Africa dur- ing the late Cretaceous or early Tertiary. 2. As a large island, it rafted away from Mada- gascar northwards along the Mascarene Ridge, leaving behind the Seychelles, Amirante and finally the Laccadive and Maldive islands ( Laughton et al., 1973; McKenzie and Sclater, 1973). 3. The Indian fauna evolved in isolation for about 50 million years; evolution was enhanced by the changing climates. 4. Some faunistic exchanges remained with Mada- gascar and Africa through the intervening islands, like the Seychelles, and possibly others that have since disappeared. 5. These small islands provided opportunities for genetic drift and quantum and tachytelic evo- lution favoring major adaptive shifts (e.g., Microhylidae, Savage, 1973). 6. Perhaps other exchanges took place in front of and/or on the eastern side when the Indian Noah's Ark (McKenna, 1973) drew near Lau- rasia, gliding along the Ninety-east Ridge and finally along the Nicobar and the Andaman islands. 7. The fate of much of the Indian fauna must have been extinction. 8. Some elements escaped early to Madagascar and Africa and proved successful in their ex- pansion (e.g., Microhylidae). 9. Other elements escaped later in northern, northeastern and eastern directions (e.g., other Microhylidae and perhaps Agamidae, Varani- dae, Pythonini, Elapidae). Four groups of reptiles (Agamidae, Varanidae, Pythonini, Elapidae) have patterns of distribution that can be explained by an Indian differentiation. Each has a strong Indo-Malaysian component, another strong Australasian component, and a weak African component, as if there had been a late invasion of Africa from Asia. Subsequent to writing this account, I have been informed by R. Hoffstetter that recent data show that the Indian Subcontinent collided with Laurasia not later than the Eocene. 4. Herpetofaunal Relationships of South America With Australia Michael J. Tyler Department of Zoology University of Adelaide Adelaide, South Australia 5001 Australia In reviewing the extent of South American herpetofaunal relationships with Australia, while simultaneously considering South Amer- ican-African relationships (Laurent, this vol- ume), it is helpful to recognize that vast dif- ferences have existed in the opportunity for faunal exchange. South America and Africa may be regarded as lovers who experienced and exploited a large zone of contact and had considerable opportunity for interchange and exchange across it. In contrast, the South American-Australian relationship suffered from being in the form of an arranged engage- ment of longer duration. The couple never so much as touched one another at any time. The only contact was via a related intermedi- ary named Aunt Arctica, whose presence be- tween them effectively prevented a compar- able degree of intimacy, and who is now outwardly cool and distinctly secretive about revealing what took place between them. The benefit of employing such an analogy lies in emphasizing the fact that Australia and South America have always been physically separated. This separation always has been extensive, because the intervening Antarctica is a vast continent with a surface area of 1,165,500,000 km 2 , comparable in size to South America north of the Tropic of Capri- corn, and considerably greater than Australia (7,700,000 km 2 ). A North to South traverse of Antarctica involves a distance of approxi- mately 4,000 km. When Antarctica was an integral com- ponent of Gondwanaland, the herpetofaunal elements shared at any one time by South America and Australia also would have oc- curred on Antarctica. Certainly a topography, climate, and vegetation equable to the main- tenance of reptiles and amphibians had to exist on Antarctica, and at least some of the modern families could just as well have orig- inated there as on the adjacent landmasses. Thus any realistic concept of intercontinental exchange avoids reference to "journeys" along "routes," and only visualizes the expansion and retraction of populations. Cartoons in an otherwise serious paper by Rich ( 1975 ) on the origins of the Australian nonpasserine avi- fauna, illustrate the errors to which some in- vestigational philosophies may have suc- cumbed. The study of intercontinental herpeto- faunal relationships faces problems of varia- tion of systematic interpretation of taxa, and these materially influence the degree of faunal similarity. For example, if the numerically dominant Australian terrestrial and arboreal frogs are regarded as members of the Lepto- dactylidae and Hylidae, respectively, all anu- ran families found in Australia are shared with South America. Superficially at least, the anuran relationship appears likely to prove a close one. However, if the names Myobatrachidae and Pelodryadidae are em- ployed for these same groups, it is difficult to avoid a bias towards a quite different inter- pretation. In fact it would appear that, for the purposes of intercontinental comparisons, there is a mystique surrounding a family name that does not extend to other nomen- clature. Over the past few years there have been substantial contributions to the study of plate tectonics, continental drift, palaeoclimate and the past flora and fauna of Australia. Many of these papers are highly relevant to the inter- pretation of evolutionary opportunities and the nature of the diversification of the herpe- tofauna. Here I have attempted to bring to- gether the most recent literature as a general background before examining the evidence to 73 74 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 l Al » 0t Fig. 4:1. Australia, New Guinea and adjacent landmasses. Australia, Nueva Guinea tj iierras adijacentes. establish the origins of the modern, non- marine Australian herpetofauna, and its af- finities to the herpetofauna of South America. A map of Australia and associated landmasses is shown in figure 4:1. PALAEOENVIRONMENTAL CONSIDERATIONS Onset of drifting. — Previously there has been considerable variation in estimates of 1979 TYLER: SOUTH AMERICA AND AUSTRALIA 75 the onset of the northwards drift of Australia away from east Antarctica, ranging from a low of 43 m.y.b.p. (Jardine and McKenzie, 1972) to a high of 180-100 m.y.b.p. (Fooden, 1972; Savage, 1973). However, it is now placed at 55-52 m.y.b.p., with most authors favoring 53 m.y.b.p. ( McGowran, 1973; Sclat- er et al., 1974; Coleman and Packham, 1976; Veevers and McElhinny, 1976). McGowran's studies of the Antarctic-Australian suture led him to suggest that, despite the onset of drift, there was no substantial barrier to the pas- sage of land animals prior to the early Eocene (49 m.y.b.p.). Climatic, floral and faunal changes. — Be- cause Nothofagus forests now occur in some temperate areas, such as southeast Australia including Tasmania, and in New Zealand, it has been possible to deduce that Nothofagus is associated classically with temperate cli- matic conditions (Axelrod, 1975). Thus, with evidence of Nothofagus occurring in the Eo- cene at several localities in southern Australia, the inference might be drawn that, at the time of the separation of Australia from East Antarctica, the southern Australian fauna was probably cool-temperate. Certainly this as- sumption would be valid for N. fusca and N. menziesi, which now exist in Australia, New Zealand, Chile and Argentina. However, the important species is N. brassi, which now exists in New Guinea and New Caledonia and clearly is a subtropical species. Formerly, its distribution was far more extensive, being known in Australia and New Zealand from the Early Cretaceous to the mid-Pliocene, in West Antarctica from the early Palaeocene to the mid-Eocene, and from Chile and Argen- tina from the Early Cretaceous to the late Oligocene (Schlinger, 1974). Further evidence of the southern Austral- ian climate being subtropical has been estab- lished by Lange (1976) from his study of microfossil epiphyllous germlings, and by Christophel and Blackburn (1978) from their assessment of the Eocene South Australian Maslin Bay flora. The geomorphological evi- dence of widespread subtropical conditions are summarized by Bowler (1976). Central Australia is another portion of the continent whose palaeoclimatic conditions have been misinterpreted. Axelrod (1960) envisaged deterioration throughout the Ceno- zoic leading to arid to semiarid conditions by the Miocene. There is now evidence that cen- tral Australia bore large, permanent lakes in the Miocene. Lungfishes, teleosts, turtles, crocodiles, lizards, and frogs shared the site with a vast diversity of marsupials and birds. The surrounding vegetation was dense, rang- ing from rainforests to extensive areas of grassland. Gallery forests extended along the watercourses, and the presence of Nothofagus and Podocarpus are interpreted to be evi- dence of high rainfall. There was southern communication with the sea at some stage (Callen and Tedford, 1976). The records of crocodiles and turtles at former freshwater sites in central Australia are particularly nu- merous (see also Newsome and Rochow, 1964). However, these represent the most conspicuous and most readily recognized rep- tile fossils; the search for smaller material has only just begun. At some stage, the area between central Australia and the north coast also was moist. This is demonstrated by the cabbage palms, Livistona mariae, now restricted to a colony of 3,000 at Finke River in central Australia. Their nearest relatives lie 1,000 km away in the northwest of the continent ( Latz, 1975 ) . Pa- laeontological and geomorphological evidence demonstrate that central Australia provided numerous niches for mesic animals until the end of the Pleistocene ( Wopfner and Twidale, 1967; Mabbut, 1967; Twidale, 1972). Cer- tainly deserts have featured in Australia for a long period and have had an essential role in lizard speciation (Pianka, 1972). It has been suggested that in the Quaternary much of the now moist extreme southwest of the continent was arid (Glassford and Killigrew, 1976). However the extent of Australia affected by aridity appears to have been exaggerated. The minimal morphological differentiation of the central Australian hylid frog fauna is wholly consistent with aridity being a late Pleistocene feature. Thus the species Litoria caerulea and L. rubella are relicts of a much richer fauna, surviving because of tolerance of adults or larvae to high temperature or 76 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 possibly for some other reason. But for the contrary evidence of Glassford and Killigrew ( 1976 ) , it is possible that Australia has not been any more arid than it is today and, to judge from the nature and abundance of vegetation cover now stabilizing sand dunes in some areas, a trend towards climatic ame- lioration has already begun. However, the Pleistocene record is of minimal relevance to this review. THE NATURE OF THE AUSTRALIAN-ORIENTAL COLLISION Modern New Guinea is composed of three distinct and roughly longitudinally arranged portions. The southern portion and the inter- vening Arafura Sea originally represented the leading edge of the Australian continental plate. The central cordillera is predominantly a much younger feature; uplift commenced in the Miocene. Finally there is a row of iso- lated mountain ranges on the north coast, each of which is composed of older volcanic rocks. The history of the evolution of New Guin- ea, and of the area to the east and west is extremely complex. Similarly only the broadest of principles of the nature of the collision of the plates has yet been established. The fol- lowing contributions provide a brief spectrum of opinions and are a source of many other references : Thompson ( 1967 ) , Falvey and Taylor (1974), Coleman (1975), Denham (1975), Mackenzie (1975), Taylor (1975), Tilbury (1975) and Coleman and Packham (1976). Only recently attempts have been made to reconstruct the nature of the plate collision. Mackenzie (1975) suggested that mountain ranges now on the north coast of New Guinea represent an arc of islands that persisted through to the Miocene, and became accreted during the collision of the plate margins (Fig. 4:2). Coleman and Packham (1976:204) favored this concept: "For the moment, we accept the likelihood that north coastal New Guinea is a piece of crust, prob- ably an arc segment, in collision with Aus- tralia-New Guinea." New Britain, to the east of New Guinea, therefore represents an island of the same arc, but which did not come di- rectly into contact with New Guinea. ^---^ TORRICELLI ~\TORRICELLI ^^^ 1 -^FINISTERRE D \=INISTERRE -uu/j \~\<3c: . NEW BRITAIN '■•■. \ ^J^riEDGE OF ^ »-* ': \F\_) MIOCENE LANDMASS /f^~~\ yi MIOCENE PLEISTOCENE * ;/ \ v3^ • EDGE OF ^""-—^V^ yr :<£ PALAEOZOIC Via r\ : ' j ( CRUST n ; ^Mi^^H Fig. 4:2. Reconstruction of the collision of eastern figure represents the early Miocene prior to collision, right figure shows two of the islands ( now known as accreted into New Guinea, thereby becoming part of 1975.) Reconstruction dc la colision del cste de Nueva Gui izquierda representa el Mioccno inferior previo a la de islas. La figura de la derecha muestra dos de las islas sterre) adheridas a Nueva Guinea, llegando a formar McKenzie, 1975.) New Guinea with islands in the mid-Miocene. The left with the landmass approaching a chain of islands. The the Torricelli Mountains and the the northern New Guinea coastline Finnisterre block ) (After McKenzie, nea eon islas en el Mioceno medio. La figura de la colision, con la mass de tierra alcanzando una cadena (conocidos como Montanas Torricelli y el Bloque Finni- parte de la costa del norte de Nueva Guinea. (Dc 1979 TYLER: SOUTH AMERICA AND AUSTRALIA 77 HERPETOFAUNAL ORIGINS In very broad terms, the ancestral stocks of the South American and Australian faunas were derived from two distinct centers. For South America there was an initial Gondwa- nan source, reinforced following the drifting of Africa, followed by a later North American infusion. For Australia there was similarly an initial Gondwanan source followed by an Oriental one. A basic problem is to determine for each continent which taxa are of Gondwanan an- cestry. The recent literature includes several assessments (Keast, 1971, 1973; Cracraft, 1973, 1974, 1975; Savage, 1973). For Aus- tralia the distinction between the Gondwanan element and the more recent Oriental one should be distinguishable on the basis of morphological divergence and by the nature of geographic distribution. This is because the Australian-Oriental collision occurred in the mid-Miocene, so that animals in Australia of Oriental origin should have distinct affinities and comparable geographic distributions with animals in the Oriental Region. Conversely, such relationships should be lacking among the Gondwanan component and there should be minimal geographic distribution outside the Australian continent. Thus I propose to establish the constituents of the Gondwanan element of the Australian fauna primarily by a process of identifying, and so eliminating, the Oriental element. On an historic and biogeographic basis the Oriental element of the Australian herpe- tofauna will fit into one of two categories, as follows: 1) Animals that occurred within the northern chain prior to the mid-Miocene col- lision with Australia. All of these would have entered that area from the west. They could range from the Philippine Islands to Fiji. To be recognizable as pre-collision components, they should be more abundant on islands east of New Guinea than in New Guinea itself. 2) Animals that have dispersed from west to east following the accretion of the chain within northern New Guinea. Such animals are likely to show a progressive west to east reduction in diversity and to be poorly repre- sented on the islands east of New Guinea. It is worth contemplating that some of the deficiencies of Wallace's Line and of other attempts to delineate the Oriental and Aus- tralian faunas exist because in reality there are three components. Hence to the recog- nized Australian and post-collision Oriental colonizers, biogeographers have failed to rec- ognize the existence of the additional pre- collision Oriental unit. ORIENTAL ELEMENTS Ranidae The overall distribution of the Ranidae in the Australian Region is wholly consistent with the concept of entry from the adjacent Oriental Region to the west. What is less satisfactorily explained is the existence of two endemic species of Platymantis in Fiji far to the east, whereas none occurs in Australia. In terms of diversity and abundance of ranid species, New Guinea is equally anomalous, to the extent that this component of its fauna is depauperate when compared with those of smaller islands to the west and to the east. Thus there are 20 ranids in the Philippine Islands, 10 on New Guinea, but 24 on the Solomon Islands. Those anomalies are high- lighted by the study of the genus Platy mantis, including species previously referred to Cor- nufer (Fig. 4:3). Viewing such a distribution pattern has led to the assumption that the distribution of Platymantis in New Guinea is relictual (Zweifel, 1969). In support of a concept that Platymantis was formerly far more widely distributed in New Guinea than it is today, there is evidence of close phylogenetic relationships existing between species that are geographically iso- lated from one another. An example is P. batantae of Batanta adjacent to the Vogelkop Peninsula of Irian Jaya (West New Guinea), which Zweifel ( 1969 ) considered most closely related to P. giUiardi and P. mimicus of New Britain about 2000 km distant. Platymantis punctata of northern New Guinea and P. myersi of Bougainville, Solomon Islands, with which it has affinities, are separated by a simi- lar distance. Inger (1954:355) evidently drew comparable conclusions when he suggested that the closest relations of P. meyeri of the Philippine Islands ". . . are not with other 78 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 B O R N E O p 11 H L 1 1 S 1 P A V 1 N N s E B PALAU ISLANDS BATANTA/WAIGEU N.W. NEW GUINEA CYCLOPS MTS. r 1 2 m |_ NEW GUINEA 1 ^; EW BRITAIN AUSTRALIA \ A SOLOMON V \ ISLANDS FIJI {2} Fig. 4:3. Modern distribution and numbers of species of frogs of the genus Platymantis (Ranidae). Distribution actual y numeros de especies de batracios del genero Platymantis (Ranidae). Philippine Corniifer [Platymantis] but instead seem to be with non-Philippine species." Brown and Alcala ( 1970 ) proceeded a step further and declared that the distribution of Platymantis within the Philippines is relictual, thereby accounting for the predominance of the genus in the north of that group of islands. An inteq:>retation of Platymantis as relicts is most readily made if the relevant landmasses are visualized as being static and the animal populations conveniently mobile. An alterna- tive interpretation is one in which it is pos- sible to contemplate mobile landmasses and relatively static insular populations. Clearly, there are several major centers of ranoid evolution in different parts of the world; the Philippine Islands with seven gen- era and 29 species is one of them. The adja- cent and larger land mass of Sabah (Borneo) to the southwest has fewer. Platymantis prob- ably evolved within the Philippines in the Late Tertiary and subsequently dispersed southeastwards into New Britain, the Solo- mon Islands and Fiji by rafting. Presumably this would require marine equatorial currents following patterns essentially similar to those today. The primary radiation is from the Philippines to New Britain and the Solomon 1979 TYLER: SOUTH AMERICA AND AUSTRALIA 79 Islands (Fig. 4:4). The direction of second- ary radiations is a reflection of demonstrable pliylogenetic affinities of the extant species. New Britain evidently was a major source of colonizing species, leading to the occurrence of P. batantae on Batanta and P. punctata on Waigeo Island off the Vogelkop Peninsula, P. cheesmanac on the Cyclops Mountains, and possibly P. papuensis on the Finisterre Mountains (subsequently extending through- out northern New Guinea). When the Australian continental mass col- lided with the Oriental island chain, it there- fore acquired four species (or stocks) of Platymantis (Fig. 4:4). Three of them (ba- tantae, cheesmanae and punctata) have re- mained almost entirely within the original confines of the islands on the north coast and have not spread appreciably in New Guinea. The fourth (papuensis) has extended as far as the south coast in the extreme west of New Guinea. This species ranges to New Britain and the Solomon Islands, but it is known to inhabit the intertidal zone and is well suited to dispersal by land and by sea (Tyler, 1976a). Nevertheless, with the time scale available, its dispersal in New Guinea remains modest (Fig. 4:5). Perhaps this im- plies the existence of an Australopapuan com- petitor and hence an ecological, rather than a physical, barrier to dispersal. The source of the stock that gave rise to the two endemic species on Fiji is uncertain. The intervening and florally rich New Hebri- des lacks Platymantis or any other endemic species of frogs. The Australian hylid Litoria aurea has been introduced there recently, pos- sibly from New Caledonia, where it was in- troduced at the turn of the century (Tyler, 1976a, 1979 ) . The striking success of the New Hebrides introduction tends to eliminate any possibility of extinction as an explanation for the absence of frogs there. Rana represents a more recent ranid ar- rival. The number of species on the various landmasses north of Australia exhibits a pro- gressive reduction from west to east in accord with an Oriental origin (Fig. 4:6). Microhylidae Microhylids occur in South and North America, Africa, Madagascar, Asia including Indonesia and the Philippines, New Guinea and northern Australia. The Australopapuan unit is the most prolific, with 13 genera and 102 species ( Zweifel, 1972; Menzies and Tyler, 1977), compared with 16 genera but only 32 species in South America (Walker, 1973; Walker and Duellman, 1974; Nelson, 1975). Parker (1934) was the last contributor to treat this family in its entirety. Subsequent contributions have tended to examine single geographic components, and the overall phy- logenetic relationships of the diverse genera remain obscure. Cracraft ( 1973 ) and Bogart and Nelson ( 1976 ) outlined the principal issues, of which contention has centered on interpretation of the origin of the family, and particularly its relationship to the Ranidae. In reality, the wide variety of opinions that have been offered on the origin of this family reflects the extreme morphological complexity of the constituent members and the absence of a satisfactory, modern synthesis. This is demonstrated particularly well by the varying subfamilial classifications that have been pro- posed. Within the context of South American- Australian faunal studies, the contributions of Savage (1973) must be considered in detail. Savage's conclusions differed quite strikingly from those of Parker (1934). Whereas the latter recognized two subfamilies occurring within and confined to the Australopapuan area (Asterophryinae and Sphenophryninae), Savage recognized only one, to which the name Asterophryinae was applied. Savage (1973:355) considered that the only distinc- tion between the Asterophryinae and the Sphenophryninae was that ". . . the former usually have an amphicoelous vertebra just anterior to the sacrum and the other presacral vertebrae procoelous (diplasiocoelous), while the latter have all presacral vertebrae procoe- lous." Savage reinforced his argument of the inherently trivial nature of such a distinction, by pointing out that although Genyophryne has uniformly procoelous vertebrae, it had been referred to the amphicoelous Astero- phryinae by Parker (1934). Savage wrote without the benefit of access to a contemporary study by Zweifel (1971), who reexamined the diagnostic characteristics of both subfamilies in general and of Genyo- 80 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 p H 1 1 S 1 L 1 A P N P D 1 S N E SECONDARY ^ □ RADIATION SECONDARY RADIATIONS NEW GUINEA AUSTRALIA Fig. 4:4. Dispersal routes and the distribution of Plat center of the figure represents the original Oriental is tanta + Waigeo Island; hatched = Torricelli Moun collided with these islands, the latter occupied the in leading to the situation shown in figure 3. Rutas de dispersion y la distribution de Platymantis centro de la figura represente la sistema original de Batanta + Isla Waigeo; achurado = Montanas Torri Guinea choco con est as islas, la ultima ocupada las irregu conduciendo a la situation mostrada en la figura 3. ymantis by the early Miocene. The row of squares in the land arc system. (From left to right: stippled = Ba- tains; open = Finnisterre Range.) When New Guinea dentations shown on the north coast of New Guinea, en el Mioceno inferior. La fila de cuadrados en el islas Orientales. (De izquierda a derecha: punteado = celli; bianco = Cerros Finnisterre.) Cuando Nueva laridades mostrados en la cosia norte de Nueva Guinea, 1979 TYLER: SOUTH AMERICA AND AUSTRALIA 81 Fig. 4:5. Distribution of Platymantis papuensis on New Guinea and adjacent islands. Populations on islands east of the mainland may represent an undescribed species (R. G. Zweifel, pers. comm. ). Distribution dc Platymantis papuensis en Nucva Guinea y islas adyacentes. Las poblaciones de las islas al este del contincnte pudieran ser una especie no descrita (R. G. Zweifel, pers. com.). phryne in particular. Zweifel ( 1971 ) con- cluded that, despite certain equivocal fea- tures, Genyophryne was properly considered a member of the Sphenophryninae, and he proceeded to redefine the Asterophryinae and Sphenophryninae on the basis of distinctions of maxillae, dentaries, vertebral column and tongue. My studies of superficial mandibular musculature (Tyler, 1974a) provide addition- al data supporting such a recognition of two subfamilial units (Table 4:1). A further action by Savage ( 1973) of con- siderable impact was that of including within the Asterophryinae (sensu lato) Calluella, an Asian genus uniquely associated by Parker (1934) with the Malagasy Dyscophinae. Table 4:1. — Diagnostic Characters of Australopapuan Microhylid Frogs. (Data from Zweifel, 1971, and Tyler, 1974a) Character Asterophryinae Sphenophryninae Maxillae Dentaries Vertebral column Tongue Interhyoideus muscle Often overlapping premaxillae, and usually in contact In contact anteriorly (except in Hylophorbus) Diplasiocoelous Subcircular, entirely adherent, often with a median furrow and posterior pouch Anteriorly underlies intermandibularis ( except in Hylophorbus ) Not overlapping premaxillae Never in contact medially Not in contact Procoelous Oval, half-free behind, lacking median furrow and posterior pouch Does not underly intermandibularis 82 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 18 p H I L I P : 13 N E S (JJ PALAU BATANTA/WAIGEU NW. NEW GUINEA NEW GUINEA 6 JEW BRITAIN \ 1 AUSTRALIA *T\ SOLOMON \A ISLANDS FIJI [O] Fig. 4:6. Distribution and numbers of species of fi ana in Australia and the adjacent Oriental Region and Pacific area. Distribution y numcros de especies de Rana en Australia y la adyaccnte Region Oriental y areas del Pacifico. Unquestionably, the former union provided a biogeographic disjunction that was difficult to interpret. Nevertheless, to associate CallueUa with the Asterophryinae introduces new anomalies of even greater magnitude. This is because the Asterophryinae and Spheno- phryninae are composed exclusively of frogs exhibiting direct development. The inclusion of CallueUa among the Australopapuan spe- cies introduces species with a free-living larval stage. The magnitude of this introduc- tion can only be appreciated when the con- siderable diversity of scansorial, terrestrial, semi-aquatic and fossorial frogs is seen to be united by sharing uniformly similar ontog- enies. To accommodate CallueUa in the As- terophryinae ( sensu lato ) also has a profound biogeographic impact, extending the range of the subfamily from the Australian Region to as far as western China. Evidence contra- dicting this step can be obtained from bio- geographic and from morphological sources, 1979 TYLER: SOUTH AMERICA AND AUSTRALIA 83 but it involves resurrecting and modifying concepts refuted by Savage. Few authors have contemplated the pos- sibility of microhylids being widely distrib- uted in Gondwanaland in the Cretaceous. Admittedly, the concept of the entry of the family into Australia from Indonesia ante- dated acceptance of continental drift and sea- floor spreading. However such an entry has been supported by many authors, of which Laurent (1975) is the most recent. Savage (1973) proposed the interesting hypothesis that the Microhylidae was present in the trop- ical portions of each of the southern land masses prior to the fragmentation of Gond- wanaland. He further put forward an in- genious account of a turbulent climatic his- tory for Australia, so as to account for the family's present abundance in New Guinea (13 genera, 95 species) and almost total ab- sence from Australia (2 genera, 7 species). Savage's hypothesis demands considerable mobility for the Australopapuan populations. In particular there is the need for extinction of the Australian component, followed by di- versification in New Guinea, and the subse- quent recolonization of Australia by immi- grants from New Guinea. The distribution of the Microhylids in Australia, New Guinea and the adjacent por- tion of the Oriental Region is shown in figure 4:7. The frogs are predominantly montane. Interpretation of the origin of the Australo- papuan microhylids must accommodate three important facts. 1) The highly adapted montane microhylids of New Guinea are unlikely to be any older than the orogeny of the mountains that they inhabit. 2) The geographically most widely distributed gen- era (Cophixalus, Oreophryne, and Spheno- phryne) all have representatives at low alti- tudes. 3) Microhylids do not occur in any part of southern Australia [therefore exclud- ing geographic areas affected by the aridity that Savage (1973) believed to have caused their demise]. Whereas Savage visualized the Australo- papuan microhylids as a Gondwanan element and thus a group whose origins involved a direct ancestry to South American frogs, I subscribe to the more orthodox opinion of an Oriental ancestry for the Papuan stock. This view can be supported on historical, biogeo- graphic, and morphological grounds. Pro- vided with the evidence of the nature of the collision of the Australian continental plate with the pre-existing chain of Oriental islands, an Oriental origin seems highly likely for the Australopapuan stock. Thus, Cophixalus, Oreophyrne, and Sphenophryne were prob- ably established within the chain at the time of the collision. The absence of these genera in the Solomon Islands and islands farther south, and the presence in New Britain of only a single species each of Oreophryne and Sphenophryne (Tyler, 1967) indicate that microhylids passed eastwards after the colo- nization of the same areas by ranids. It fol- lows that the ancestry of the Papuan micro- hylid fauna must be far less complicated than an examination of the diverse modern genera would indicate. The important criterion for their success appears to have been the in- herent ability to colonize the New Guinean montane environments that evolved during the rapid elevation immediately after the col- lision. Cophixalus, Oreophryne, Sphenophryne, and in fact all Australopapuan microhylids exhibit direct development. In this regard they differ from all Oriental microhylids. Di- rect development has enormous selective ad- vantage in situations where there is a shortage of suitable aquatic breeding sites. The first step in its evolution is probably acquisition of macrolecithal eggs without altering the met- abolic demands of the embryo. The potential for delayed emergence from the vitelline membranes would result. In terms of the anatomical structure of the tadpole, it follows that any deferment of the onset of larval life is most likely to permit economy in the elab- oration of the vast digestive system. This re- sults from the existence of increased food reserves, and of decreased demands upon the use of the larval digestive apparatus. There are several Oriental microhylids that exhibit trends towards delayed emer- gence. Inger (1966) provided ecological notes of species in which enlarged and un- pigmented ova have been found. In the genus Kalophrynus the eggs of some species are pig- mented, whereas in others such as K. pleuro- stigma they are unpigmented, and the larvae 84 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 C5 Fig. 4:7. The Australian and adjacent Oriental and Pacific areas showing the distribution of the Microhyli- dae (broken line) and the range of the most widely distributed Australopapuan genus, Oreophryne (con- tinuous line ) . Distribution de Microhylidae (linea entrccortada) y el rango del mas extendido de todos los generos austral- opapua, Oreophryne (linea continua) en Australia y las areas Oriental y Pacifico adtjacentes. have poorly developed intestines: ". . . only two loops visible ventrally and appears to be full of yolk." (Inger, 1966:135). The origin of the Asterophryinae from sphenophrynine ancestors has been consid- ered previously. Zweifel (1972:431) observed that of the asterophryine genera Hylopharbus ". . . differs from Cophixalus of the Spheno- phryninae only in having the tongue less free, and in having a diplasiocoelous rather than procoelous vertebral column. Therefore it may be that the Asterophryinae sprang from stock much like the present-day Cophixalus." Tyler ( 1974a ) similarly concluded that the super- ficial mandibular musculature of Ihjlophorhus is comparable to the uniform condition of Cophixalus and other sphenophrynines, and that the various conditions in the Astero- phryinae can be derived from the generalized sphenophrynine muscle pattern. The superficial mandibular musculature of South and North American microhylids ex- hibits a progressive trend of elongation of a single pair of slender, supplementary elements of the intermandibularis muscle (Emerson, 1976). In many respects these structures re- semble those found in Papuan sphenophry- nines, but the interhyoideus muscle is more closely involved in the vocal sac, and the structure of the vocal sac is distinctive. It forms an involuted pouch dorsal to the inter- mandibularis in at least some of the South American taxa, but there is no such trend in sphenophrynines, and in the asterophryines the interhyoideus forms a single sheet lying ventral to the intermandibularis (Fig. 4:8). Variation in microhylid muscle architecture on each of the major continents is shown in figure 4:9; the nature of the diversity is in- dicative of complex separate radiations. 1979 TYLER: SOUTH AMERICA AND AUSTRALIA 85 suppl mand. interhy. Fig. 4:8. Superficial mandibular musculature of the Papuan mierohylid frog Phrynomantis stictogaster; ventral view with skin removed. Interhy. = inter- hyoideus; inland. = intermandihularis; mand. = man- dible; subm. = submentals; suppl. = supplementary element of intermandibularis. Musculatura mandibular superficial del batracio microhylido Phrynomantis stictogaster; visto desde la cara ventral con la piel rcmovida. Recently, our knowledge of mierohylid karyotypes has been extended to many addi- tional species, particularly by the contri- butions of Rogart and Nelson ( 1976 ) and Rlommers-Schlosser (1976). It is generally accepted that a diploid karyotype of 26 is primitive for several families, including the microhylids. This number is retained in the three Papuan and 13 Madagascan species studied, but is of variable occurrence else- where (Table 4:2). Varanidae The varanid lizards constitute a group of small to large animals that are more diversi- fied in Australia than elsewhere. The modern distribution of the family (and sole genus, Varanns) forms a broad arc from Africa to Australia; King and King (1975) supported an Asian origin. Formerly, varanids were dis- tributed far more extensively, extending far- ther north and occupying Mongolia through to Europe and to North America (McDowell and fiogert, 1954; Hoffstetter, 1961). The 19 Australian species of Varanus are placed in two groups — 1) small species lacking lateral compression of the tail and often referred to the subgenus Odatria, and 2) the large, typi- cal monitors. Hecht ( 1975 ) reviewed the fairly exten- sive history of the Varanidae, noting that the Late Cretaceous, Paleocene, and Eocene rec- ords from North America, Europe, and Mon- golia, combined with the absence of the fam- ily in South America are indicative of a Laurasian origin. The fossil record in Aus- tralia suggests that varanids may have entered Australia on two occasions. The giant Mega- lama prisca of southeastern Australia is known only from Pleistocene deposits. Al- though McDowell and Rogert (1954) synony- mized Megalania with Varanus, Hecht ( 1975) redescribed and redefined the fossil genus and provided adequate evidence to merit its recognition. Thus, it is possible that Mega- lania and Varanus represent separate inva- sions of Australia. Megalania prisca is the largest lizard known (total length up to 8 m and an estimated weight up to 600 kg). Fos- sil Varanus have been reported from the Mio- cene by Stirton, Tedford, and Miller (1961), from the Pliocene by Archer and Wade (1976), and from the Pleistocene by Smith (1976) and Molnar (1978). Thus Megalania and Varanus were contemporaneous. The success of the carnivorous and car- rion-consuming varanids in Australia was at- tributed by Storr (1964) to the absence of eutherian carnivores. Hecht (1975) suggested that Megalania represented the carnivore of the Australian megafauna preying upon some of the large herbivorous marsupials that were so abundant in the Pleistocene. Within South America the teiid genus Tupinambis is the only lizard approaching the niche filled by Varanus in Australia; Tu- pinambis is omnivorous. Scincidae Throughout the world there are over 800 species of skinks unevenly distributed among four subfamilies. The widespread distribu- tion of the family and varying interpretations of its systematics and phylogeny are biogeo- graphically undesirable attributes. The major systematic treatments are those of Mittleman (1952) and Greer (1970); I have adopted Greer's scheme. 86 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 FORMS OF MANDIBULAR MUSCULATURE IN MICROHYLIDS VENTRAL VIEW North America South America Madagascar Asia New Guinea Australia Africa Madagascar New Guinea New Guinea Fig. 4:9. Schematic representation of the orientation In each figure the muscle slip is shown on a single man plest form of this muscle is a slip at the apex of the the mandible as shown in A and B, or partially migrates and D). Representation esquenuitica cle la orientation de los crohylidos. En eada figura la portion muscular sc mucs tral. ha forma mas simple de este musculo cs como una 8. Esto migra posteriormente a lo largo de la mandi y se divide en dos porciones de varias formas (C y D). Greer considered the Scincinae to be the most primitive group within a distribution that is predominantly Laurasian but also oc- cupies the entire African continent and Mada- New Guinea of supplementary muscle elements in microhylid frogs. dible and is viewed from the ventral surface. The sim- mandibles as in figure 8. This migrates posteriorly along and then divides into two slips of various forms ( C elementos del musculo suplementario en batrachios mi- tra en una sola mandibula, y sc lo vc desde la cara ven- porcion en la punta de las mandibulas como en la figura bula como se muesira en A y B, o migra parcialmcnte gascar as well. Eumeces has a remarkably disjunct distribution, with isolates ranging from North Africa to India, China to Vietnam, and Middle to North America. Disjunctions 1979 TYLER: SOUTH AMERICA AND AUSTRALIA 87 Table 4:2. — Microhylid Karyotypes Continent Genus Species 2n Authority North America Gastrophryne Hypopachm 2 2 22 22 Morescalchi 1968b, Bogart & Nelson 1976 Leon 1970, Bogart & Nelson 1976 South America Arcovomer Dermatoiwtus Elachistocleis Glossostoma Hamptophnjne Otophryne Stercocyclops 1 1 1 1 1 1 1 22 22 22 26 22 26 22 Bogart & Nelson 1976 Rabello 1970, Becak et al. 1970 Bogart & Nelson 1976 Bogart & Nelson 1976 Bogart & Nelson 1976 Bogart, Pyburn & Nelson 1976 Bogart & Nelson 1976 Asia Kaloula Microhyla Ramanclla Uperodon 2 1 1 1 24,28 26 26 26 Morescalchi 1968a, 1968b, Bogart & Nelson Bai 1956 Bai 1956 Natarjan 1953 1976, Sato 1936 Africa Breviceps Phrynomerus 1 1 24 26 Bogart & Nelson 1976, Morescalchi 1968a, . Morescalchi 1968a, 1968b, Bogart & Nelson 1968b 1976 Madagascar Anodontohyla Dyscophus Mantipus Platyhyla Platypelis Paracophyla Plethodontoh yla 2 4 1 1 2 1 o 26 26 26 26 26 26 26 Blommers 1971, Blommers-Schlosser 1976 Blommers 1971, Blommers-Schlosser 1976 Blommers 1971, Blommers-Schlosser 1976 Blommers 1971, Blommers-Schlosser 1976 Blommers 1971, Blommers-Schlosser 1976 Blommers 1971, Blommers-Schlosser 1976 Blommers 1971, Blommers-Schlosser 1976 New Guinea Cophixalus Phrynomantis" 2 1 26 26 Cole & Zweifel 1971 Cole & Zweifel 1971 ° Identified as "Asterophrys sp." by Cole and Zweifel, this individual was subsequently referred to Phrynoman- tis stictogaster by Zweifel (1972, p. 504). in at least some other components of the Scincinae are explained most readily in terms of independent origins from the Eumeces stock. The Acontinae and Feylininae are con- fined to southern, central, and western Africa and are regarded as independent derivatives of the Scincinae, whereas the Lygosominae are found throughout Australia, in all but the extreme south of South America, Middle America, southern North America, in most of Africa, and the entire Oriental Region. Greer (pers. comm.) considers that the present distribution and radiation of the Scin- cidae can be explained without reference to continental drift. Greer (1970:178) consid- ered the lygosomines ". . . are clearly derived from scincines and are morphologically the most advanced skinks." Thus, within Australia the problem of interpretation of the vast de- gree of diversification is the time available for that radiation and diversification. However, skinks evidently are adept at dispersal over sea water and the timing of their arrival in Australia remains unknown. No pre-Pleisto- cene fossils are known from the Australian Region which is scarcely surprising, consider- ing the state of the fossil record. Agamidae The absence of agamids from South Amer- ica and their present, virtually continuous range from Africa to Australia are highly in- dicative of the Oriental origin of the Aus- tralian component of the family. The rela- tively poor representation of the family in New Guinea reflects the ecological differences between New Guinea and Australia. The highly successful radiation of agamids in Aus- tralia is not unique to agamids, and probably reflects the wide variety of niches open to the early colonizers. Archer and Wade (1976) reported a small undetermined species similar to Amphibolurus from the Pliocene. MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 Carettochelyidae and Trionychidae The carettochelyid turtles have an unusual distribution pattern, but, as yet, there is no evidence of their former presence in South America. Known from the Eocene of North America and the late Miocene to Present of New Guinea (Glaessner, 1942), the sole liv- ing representative Carettochelys insculpta was first reported from rivers of northern Aus- tralia by Cogger (1970), and since has been shown by Schodde, Mason, and Wolfe (1972) to be distributed quite widely in the Northern Territory. This species evidently has a high tolerance to salt water. Carettochelyids rep- resent a relict family, but it is not necessarily one of any great antiquity within the region. They probably represent the first trionychoid invasion that was followed by Pelochelys bi- broni, which did not extend beyond New Guinea. Both species probably entered the Australian Region in the Miocene. A triony- chid has been found in the middle Pliocene of Venezuela (Wood and Patterson, 1973). Elapidae Within Australia and New Guinea the elapid fauna is exceptionally diverse. Authors vary in the number of species and genera that they recognize, but Cogger (1975) recog- nized 26 genera and 61 species in Australia. Such numbers and diversity would seem to require a great evolutionary time span. This seems to conflict with the existence of endem- ism within the Solomon Islands (Salomone- laps and Loveridgelaps) and even as far as Fiji (Ogmodon) (McDowell, 1970). More- over, elapids extend in a slightly disjunct arc through to Africa; an Oriental origin seems likely for many of them. Whether this applies to all components of the Australian elapid fauna will have to await completion of the splendid work commenced by McDowell (1967, 1970). The elapid fossil record currently is con- fined to a species of Pseudonaja differing from P. nuchalis from the Pleistocene deposits at Naracoorte, South Australia, by Smith ( 1975 ) and Pliocene vertebrae from northern Queens- land tentatively referred to the Elapidae by Archer and Wade (1976). Colubridae The colubrids occur mainly in the northern and eastern portions of the Australian con- tinent. Some of the genera, such as Stegono- tus, range through Australia and New Guinea into the Oriental Region. The phylogenetic affinities of Stegonotus appear to be with the Oriental Dinodon (McDowell, 1972). An Oriental route of entry for the colubrids as a whole seems to be unquestionable. Acrochordidae and Uropeltidae Two acrochordids occur in New Guinea; in Australia they are restricted to the extreme north of the continent. These aquatic species are either both referred to Acrochordus, or one to that genus and the other to Chersydrus. Each species is distributed extensively in In- donesia and farther west, and they represent a recent Oriental invasion. The uropeltid genus Cylindrophis can be included within the fauna of the Australian Region because it reaches the Am Islands between Australia and New Guinea (McDowell, 1975). In other respects, it is an exclusively Oriental genus, and it has certainly entered the Aus- tralian Region very recently. Typhlopidae Typhlopids occur on almost all continents. Typhlops has a range almost equivalent to the entire family, extending throughout Asia to New Guinea. The Australian species now are referred to the genus Typhlina, which resem- bles Typhlops in external features, but differs substantially in the nature of the male geni- talia (Guibe, 1948; Robb, 1966). Cogger (1975) listed 22 species of Typhlina in Aus- tralia, and McDowell (1974) listed 11 from New Guinea and the Solomon Islands. Ty- phlina extends as far south as Fiji. The nature of the distribution pattern indicates an Orien- tal origin for the Australian component, but the date of its entry is uncertain. Its distri- bution in the southwest Pacific is extensive, and it may be well disposed to sea dispersal. Boidae The Boidae is described by Cracraft 1979 TYLER: SOUTH AMERICA AND AUSTRALIA 89 (1973:384) as ". . . an excellent example of Gondvvanan dispersal." However, for Aus- tralia the boids may have had a rather check- ered history. The distribution of Python in an almost continuous arc from Africa to Aus- tralia provides further evidence of origin out- side Australia. There has been a fairly pro- nounced successful radiation within Australia, attributed by Storr (1964) to the absence of the Felidae. McDowell (1975) recognized three groups among the Australian Pythoni- nae, distinguished by the presence or absence of labial scale pits and prehensile adaptations to the structure of the tail. The first of these contains Liasis, and the second Python, More- lia and Chondropython. McDowell consid- ered the latter two genera to be only weakly defined and maintained that a good case could be made for referring them all to Python. Such is the state of herpetological exploration in Australia that a giant new species of Python was discovered in the Northern Territory in 1975 (Gow, 1977). However, the third group containing Aspi- clites was defined more satisfactorily by pos- sessing several features not exhibited by other genera. The Boinae is represented in New Guinea by two species of Candoia. Smith (1976) upset the concept of all of the Australian boids being completely attrib- utable to radiation from a northern entry of a Python ancestor. From the Naracoorte Caves in the southeast of South Australia, she described the fossil genus and species Wo- nambi naracoortensis. She considered Wo- nambi to be related most closely to Madtsoia bai ( Palaeocene-Eocene of Patagonia) and M. madagascariensis (Cretaceous of Mada- gascar). The Naracoorte Caves provide an ex- ceptionally rich fossil fauna and although of only late Pleistocene age, many forms of verte- brates recovered from them are now extinct. Wonambi naracoortensis was described from a series of vertebrae; Smith estimated that they were derived from a snake with a body length of approximately 5 m. A frag- ment of left maxilla associated with the verte- brae bore teeth approximately 7 mm in length. It differs from extant Australian boids in lack- ing accessory processes beneath the prezy- gapophyses, in possessing weak subcentral ridges and paracotylar foramina, and in ex- hibiting a slight posterior slope to the neural spine. If Wonambi is correctly associated with the Boidae, Australia may have been colo- nized by the family twice — an initial entry antedating the drift of Australia from Antarc- tica, and a second entry, presumably in the Miocene. If so, when Python first appeared in the north of the continent, Wonambi or its ancestors inhabited at least the southeastern part. The only other fossil record of Australian boids is the report by Archer and Wade ( 1976 ) of three vertebrae of a very large species in the lower Pliocene Allingham For- mation in north Queensland. These authors do not associate it with a modern species, but noted (op. cit: 385) that it is ". . . mor- phologically very similar to modern species of Morelia." GONDWANAN ELEMENTS As demonstrated here, the Gondvvanan elements are predominantly anuran. The fos- sil record is only just being assembled, and it is best dealt with here rather than within the individual families. Australian Fossil Frog Record Until recently there were no known fossils of frogs in the Australian Region. Tyler ( 1974a ) reported the discovery of an isolated left ilium amongst a rich vertebrate assem- blage taken at Lake Palankarinna, north of Lake Eyre, in the northern part of South Aus- tralia. Subsequently, this ilium became the holotype of the new genus and species Aus- tralohatrachus ilius, tentatively referred to the Hylidae (Tyler, 1976). More recently, 19 more ilia have been found in sediments from Lake Palankarinna (Tyler, unpublished). These include a number of additional speci- mens of A. ilius, several specimens of an un- described species of the leptodactylid Limno- dynastes and also a Litoria species closely re- lated to, and possibly representing L. caeru- lea. The age of the Lake Palankarinna fauna is uncertain but is considered to be most likely mid-Miocene. Thus, the age coincides with the collision of the Australopapuan and Ori- ental plates. 90 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 A rich Pleistocene fauna, including 166 frog ilia of extant species, has been taken at two cave sites at Naracoorte in the lower southeast of South Australia (Tyler, 1977). The fossils include Litoria ewingi, Limno- dijnastes cf. dumerili, L. tasmaniensis, Rani- della signifera, and Geocrinia cf. laevis, all of which occur in that area today. Unfortunate- ly, as yet there are no known frog fossils of an age predating the Australopapuan-Orien- tal collision. However, it is noteworthy that Limnodijnastes and Litoria were established in the Miocene at Lake Palankarinna. Hylidae Until about 1970 the confamilial status of the hylids of Australia and New Guinea ( "Australopapuan" ) with those of South America had not been disputed, and in fact Hyla was applied quite uniformly to Aus- tralopapuan and Neotropical species (creat- ing numerous problems of homonymy but ap- parently not of biogeography). The disjunct nature of the distribution of Hyla (sensu lato) was the principal area of interest long before the homogeneity of the genus (and later of the family) was really questioned. Within the framework of a concept of static continents, Parker (1929) suggested that the Hylidae might be of North American origin, radiating from there in several different direc- tions, of which South America and Australia were the ultimate destinations of two. The major gap between the depauperate Oriental Hyla and the relatively numerous Australian species remained a serious obstacle to ade- quate explanations of dispersal, although Dar- lington (1957) suggested that the arboreal rhacophorids have displaced and now replace the hylids in that intermediate area. The first suggestion of a direct association between South American and Australian hylid species was made by Reaufort ( 1951 ) , who visualized an Antarctic land bridge as a route for entry to Australia from South America. During the past decade fresh interpretations of the phylogeny and systematics of the hy- lids of Australia and South America have far exceeded the comparative faunal studies de- sirable to support some of the conclusions reached. Tyler (1971) studied a suite of characters associated with the superficial mandibular musculature, and the vocal sac that it con- tains, in representatives of numerous families and including almost all known hylid genera. He noted that anatomical divergence occurred principally in association with taxonomic units recognized as genera. He demonstrated that the Australopapuan Hyla constituted a single morph distinguishable from species from other parts of the world. Consequently he proposed the resurrection of Litoria Tschu- di to accommodate the Australopapuan spe- cies and further considered Litoria and Nyc- timystcs to be a unique, monophyletic group. Cracraft (1973) reported the above find- ings without variation, but in 1974 suggested only that the distinction of Asiatic and Aus- tralopapuan species had been demonstrated, and he introduced the topic of Australo- papuan-South American hylid affinities as though it were a new proposal. Savage ( 1973) acknowledged some of my data in press as evidence refuting the confamilial status of the two groups of frogs, and accordingly resur- rected the family name Pelodryadidae Giin- ther to accommodate them. Ragnara ( 1974 ) published rather conflict- ing data. From his discovery of the occur- rence of rhodomelanochrome in the skin of Neotropical phyllomedusines and some, but not all, Australian hylids, he made two radi- cal proposals. These were 1) that the Aus- tralian species of Litoria exhibiting rhodo- melanochrome are more closely related phylogenetically to South American phyllo- medusines than to sympatric congeners, and suggested therefore, 2) that Litoria was a highly heterogeneous assemblage. More re- cently, Ragnara (1976) reiterated the com- mon origin of the phyllomedusine-L/Yon'a component more forcefully. Laurent ( 1975 ) envisaged a totally differ- ent origin for the Australopapuan species, sug- gesting independent origin in situ from a for- merly world-wide leptodactylid ancestor, which he visualized as the ancestral stock of several families on different continents. Con- templating a wholly autochthonous origin he employed the name Nyctimystinae for the Australopapuan fauna. In an attempt to stabilize the classification 1979 TYLER: SOUTH AMERICA AND AUSTRALIA 91 of these and other frogs, Duellman ( 1975 ) offered a number of solutions to controversial issues. Among his actions, he continued to include Australopapuan species in the Hy- lidae. In order to crystalize the current contro- versy and to aid the transcontinental study of the Hylidae, it is necessary to restate or clarify the following issues: 1) the phylogenetic re- lationships of the Indonesian hylids ( those on the periphery of the Australian population and geographically closest to the Oriental species); 2) whether the Australopapuan species genuinely constitute a monophyletic group; 3) the phylogenetic relationships of the Australian species; and 4) the phylo- genetic relationships of the Australian and South American hylid faunal units. The Indonesian hylid fauna. — The north- western geographic limit of Litoria occurs in the Indonesian islands of Timor and the Less- er Sunda Islands of Sumba, Savu and Alor. This latter assemblage represents the eastern end of an archipelago forming an intimate link to the Malaysian Peninsula far to the northwest. Therefore, the Litoria fauna of the Timor-Lesser Sunda group is of impor- tance to any contemplation of entry of hylids into Australia from the northwest. The only species (L. everetti) occurring in the relevant area is a member of the L. peroni group represented elsewhere in the Australopapuan area by five described species — peroni, everetti, amboinensis, rothii and darlingtoni. The total geographic range of this species group is exceptionally extensive (Tyler and Da vies, 1978a). It appears that the group evolved in Australia or New Guinea and is now radiating in several directions and extending its range. Thus, it is confirmed that the phylogenetic affinities of L. everetti are with other Australopapuan species and not with the southernmost Oriental hylid (Hyla chinensis). Monophyletic or polyphyletic origins. — In- sofar as the Australopapuan fauna is con- cerned, the issue is whether the frogs referred to the Hylidae are a monophyletic group. Ecologically, at least, they are incredibly di- verse, filling a spectrum of niches occupied on other continents by different families. A su- perficial examination of the diversity of struc- ture may render the casual observer critical of Australopapuan hylid systematics. Neverthe- less the magnitude of diversity need be no indication of polyphyly. Tyler ( 1971 ) pro- posed the concept of monophyletic origin on the basis of his studies of superficial mandibu- lar musculature and vocal sac structure. Sub- sequent karyotypic data assembled by Steph- enson and Stephenson (1970), Woodruff ( 1972 ) , Morescalchi and Ingram ( 1974 ) , and Menzies and Tippett (1976) have in no way caused this concept to change. All of the hylids karyotyped to date have 2n = 26, ex- cept L. infrafrenata (2n = 24), and in that instance a model for derivation from 2n = 26 has been proposed (Menzies and Tippett, 1976). Phylogenetic relationships of Australian species. — Tyler and Davies ( 1978a ) examined the morphology, osteology, myology, distri- bution, and biology of 92 of the 94 species of Litoria currently recognized. They found that these species can be associated in no less than 37 species groups. Insofar as all geographic areas occupied by Hyla are concerned, this total of groups is not exceptionally high. In reality the number of species per species group is remarkably similar in several geo- graphic areas (Table 4:3). Although more concerned with the initial step of establishing phonetic groupings, Tyler and Davies demon- strated that Australian hylids occupy an in- credible gamut of niches. Tyler (1970, 1972a) suggested that there exists a close phylogenetic relationship be- tween Australian hylids and leptodactylids. The core of this suggestion related to the lep- todactylid genus Cyclorana. That genus as then constituted comprised a group of squat- bodied and also some elongate species. Lynch ( 1971 ) regarded them an integral component of the Australian leptodactylid fauna. Tyler (1971, 1972a) studied superficial mandibular musculature in all Australian leptodactylids and his conclusions differed from those of Lynch only in his appraisal of Cyclorana, in which he noted distinct hylid affinities. Tyler (1970) suggested that the similarities of the Australian hylids and leptodactylids implied the existence of a single common ancestor. 92 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 Table 4:3. — Relationship Between Number of Species and Number of Species Groups of Selected Hylid Genera on Different Continents. Geographic area Genus No. of species No. of groups Species/Group Australia Litoria 46 22 2.09 : 1 New Guinea Litoria 51 24 2.12 : 1 Australia & New Guinea" Litoria 94 37 2.54 : 1 Middle America" * Hyla 73 28 2.61 : 1 North America"" Hyla 12 4 3.00 : 1 " Includes a component common to Australia and New Guinea. *• Data from Duellman (1970). Robinson and Tyler ( 1973 ) examined the relative predominance of the catecholamines epinephrine and norepinephrine (dopamine was not detected) in the adrenal glands of various Australian hylids and leptodactylids, including Cijclorana. The found that epi- nephrine was the predominant transmitter in all hylids examined and that norepinephrine was predominant in all leptodactylids, except Cijclorana. In the absence of any biochemical convergence associated with ecological con- vergence, they regarded catecholamine selec- tion as an exceptionally conservative feature. The major issues to be explored are, as follow: 1) Do the Australopapuan hylids ex- hibit a close phylogenetic relationship with any of the sympatric leptodactylids, or 2) is there a closer relationship with South Amer- ican hylids or leptodactylids? Hence it is conceivable that Australopapuan hylids and South American hylids enjoy a reasonably close relationship, or that the Australopapuan hylids are independently derived from a South American leptodactylid stock, and that resemblance only reflects convergence. Before exploring the nature of the resem- blance of hylids from each of the continents, it is worthwhile discussing here the novel proposition introduced by Bagnara and Ferris (1975) on the basis of the significance of the presence of rhodomelanochrome in some but not all Litoria (see p. 90). Maxson (1976) rejected this conclusion on the basis of im- munological data, and Tyler and Davies (1978b) reexamined the phylogenetic rela- tionships of the same species studied by Bag- nara and Ferris, together with additional spe- cies. Employing features of adult myology, osteology, larval structure, and reproductive biology, as well as pigment, their findings conflict with those of Bagnara and Ferris. They found that Australian congeners are genuinely more closely related with one another than some Litoria are with phyllo- medusines. Their findings also tend to re- inforce further the recognition of the phyllo- medusinae as a subfamilial unit. Some of the biochemical features employed by Bagnara as indicators of close phylogenetic relationships tend to be particularly subject to convergence, and Guttman ( 1973 ) outlined the problems that arise when the relationships of higher taxa are based on such characters. Phylogenetic relationships of Australian and South American hylids. — Numerically at the level of species, genera, and families, South America is clearly the most major cen- ter in the world of evolution of modern taxa of anurans. There is no reason to suppose that the Hylidae evolved in Australia and migrated to South America, but the obvious reverse option raises a number of most inter- esting questions. When some South American hylid frogs are placed directly beside Aus- tralian frogs and compared one with another, the confamilial association is stretched to the limit. South America simply does not possess frogs that resemble the Litoria aurea, L. cae- rulea and L. freycineti groups. Similarly, there are numerous dominant components of the South American fauna that do not have coun- terparts in Australia. Irrespective of the magnitude of diver- gence between South American and Australian species, the fact remains that all possess inter- calary structures, and so are referred to the family Hylidae. On a continental basis the species are indeed different. Savage (1973), using Tyler's ( 1971 ) data, considered the dis- tinction adequate to merit family status, and so termed the Australopapuan unit "Pelodrya- didae." I too recognize the morphological distinction and the monophyletic origin of the Australopapuan unit, but I find only adequate 1979 TYLER: SOUTH AMERICA AND AUSTRALIA 93 grounds for an intrinsic division within the current concept of the Hylidae. Therefore, I favor recognition of a subfamily to accommo- date Litoria, Nyctimystes and (as indicated in later discussion) Cyclorana. Because all family group names are of equal status for the purposes of nomenclatural priority, Pelo- dryadinae (derived from Pelodryadidae Giinther, 1858) takes priority over Nyctimy- stinae Laurent, 1975. It is worth noting that when Maxson and Wilson (1975) implemented Savage's concept of the Pelodryadidae, because of results of estimated mean albumin-immunological dis- tances between continental populations, they overestimated the continental divergence time. Their immunological distance of 100 units equates with 60 million years, so that the existence of 100 immunological units be- tween any two taxa involves acceptance of 60 million years isolation between the popula- tions. Their calculation of an immunological distance of approximately 129 units between the relevant Australian and South American populations can be interpreted in two ways, but may well be excessive. The physical separation of Australia from Antarctica is now established at 52-55 m.y.b.p. This total com- pares with 77 m.y.b.p. calculated by immu- nological techniques. If the latter is the period of isolation of the stocks, ecological or physical barriers on the Antarctic land mass are called for to explain the separation of populations prior to rifting. Wallace, Maxson, and Wilson (1971) found greater immunological distances exist- ing between South American and the adjacent North American species, than between North American and the geographically distant sin- gle Australian species examined. Maxson (1978) demonstrated a high degree of compatability between North American and European Hyla. Maxson and Wilson (1971) noted that where discrepancies exist between organismal resemblance and albumin resem- blance, it is to be attributed to differential rates of organismal evolution. As an example they cited Acris, which exhibits albumin and haemoglobin affinities to North American Hyla, and yet is strikingly different from such species in anatomy, gross structure, biology, and ecology. Duellman (1970:647) accepted such evidence with considerably less toler- ance: "Despite the divergent nature of Acris with respect to other hylids, and the super- ficial similarity of Acris to ranids, the inescap- able fact remains that Acris has procoelous vertebrae, an arciferal pectoral girdle, inter- calary cartilages and claw-shaped terminal phalanges — a combination of characters that seemingly inextricably ally the genus with the hylids." It is equally reasonable to suggest that organismal evolution is unlikely to be con- strained along any linear path of morpho- logical divergence as assessed by human ob- servers. Hence, systematists have a quandry that is of their own making, and while the Hylidae remains defined as it is now, the Pelodryadinae remains an integral component of it. Cyclorana is a problematic genus. By virtue of the fossorial habit of most of its species and the absence of intercalary struc- tures it formerly has been accommodated in the Leptodactylidae (Parker, 1940; Lynch, 1971). More recent evidence has demon- strated similarities between Cyclorana and pelodryadine hylids in myology (Tyler, 1972a), adrenal catecholamines (Robinson and Tyler, 1973), in larval structure and biol- ogy (Watson and Martin, 1973), and in cranial osteology (Fig. 4:10). A closer exam- ination of the moqjhology of Cyclorana spe- cies resulted in the discovery of intercalary structures in C. inermis, C. alboguttatus, and C. dahlii, and led to these species being re- ferred to the hylid genus Litoria by Straughan (1969), Tyler (1974b) and Tyler, Davies and King (1978) respectively. 1 In consequence of these actions and of the resurrection of one species and the descrip- tion of five new species (Tyler and Martin, 1975, 1977), Cyclorana now is composed ex- clusively of robust fossorial frogs lacking in- tercalary structures but retaining a closer af- finity to hylid than to leptodactylid frogs. Awareness of this presumably led Heyer and Liem (1976) to omit Cyclorana from their 1 The customary term "intercalary cartilages" is not used because these structures are bony in 46 of 71 Australopapuan hylid species studied (Tyler and Davies, 1978a). Ossification bears no correlation with finger length, habits or geographic distribution. However, all large or moderately large arboreal species retain a cartilaginous state. 94 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 Cyclorana brevipes Cyclorana australis 5mm 5mm Litoria raniformis Litoria alboguttata Fig. 4:10. Skulls of certain species of Cyclorana and Litoria. Crdneos de ciertas especies de Cyclorana ij Litoria. phylogenetic analysis of the Australopapuan leptodactylidae ( Myobatrachidae ) . Because all hylids exhibit axillary amplex- us and Australian leptodactylids (except Mix- ophyes) inguinal, it follows that the embrace of Cyclorana should be of relevance in de- termining its phylogenetic relationships. I have observed amplexus in four species of Cyclorana. Initially the grasp is high in a circumcervical position, as though the male intends to strangle his mate. However the grasp slides posteriorly to an inguinal position. Within the Hylidae, Cyclorana appears to be related most closely to the Litoria aurea species group, which now includes two spe- cies previously referred to Cyclorana. This group may prove to be the sister group of Cyclorana and merit elevation to distinct ge- neric identity. Leptodactylidae (Myobatrachidae) It is valid to describe the current state of nomenclature and phylogcny of Australian leptodactylid frogs as distinctly unstable. Even to refer to them here as Leptodactyli- dae rather than Myobatrachidae is in total opposition to the sincere efforts of many workers. I do so now because I seek a re- evaluation of the steps that led to the nomen- clatural change, and because I suspect that the examination of South American-Australian 1979 TYLER: SOUTH AMERICA AND AUSTRALIA 95 relationships is best served by nomenclatural conservatism. Parker's (1940) review of Australasian species brought a considerable degree of stability to the nomenclature of many Aus- tralian genera. Parker recognized two sub- families — Cycloraninae (which he erected) and the Myobatrachinae. Lynch (1971) pro- vided a splendid historical account of the classification of the Leptodactylidae, and in so doing, placed Parker's contribution in an historical perspective. Following the publica- tion of Parker's monograph, A. R. Main and his colleagues undertook the first detailed biological, herpetofaunal studies. The result of their work (and of that of their students) was the description of numerous new species. Nevertheless, the genera recognized and sus- tained by Parker were in no way challenged. Certainly Neobatrachus was resurrected from the synonymy of Heleioporus by Main, Lee, and Littlejohn ( 1958 ) , but in over 25 years only one new genus was erected — Tauclac- tyJus by Straughan and Lee ( 1966 ) . At that time it was tempting to assume that Austral- ia's leptodactylid fauna was already reason- ably well established. However, as late as 1960, only 59 of the currently recognized total of 79 species had been discovered. ( Cyclo- rana has been excluded from these totals.) Close examination of some of the numerically large genera then recognized (e.g., Crinia) indicated an unsuspected heterogeneity. In fact, because most Crinia were small species, the genus had in reality become a repository for a great variety of frogs only sharing small stature. It followed that closer examination led to the erection of some new genera and the resurrection of others (Tvler, 1972b; Blake, 1973). Collection of Rheobatrachus silus by Liem (1973) represented one of the most extra- ordinary herpetological discoveries of this century. In its gross morphology as an aquatic frog with profuse dermal, mucous glands, fully webbed toes, long pointed fingers and incredible aquatic maneuverability, the re- semblance to pipids such as Xenopus, and particularly to the South American lepto- dactylid Telmatobius, is extremely striking. Rheobatrachus silus was found to be even more noteworthy when Corben, Ingram, and Tyler ( 1974 ) reported that the female broods the larvae within her stomach. Subsequently, a robust-bodied, fossorial frog was found living in coastal sandhills in a remote and arid part of Western Australia. Named Arenophryne rotunda by Tyler ( 1976c ) , this genus has affinities with both Myobatrachus and Pseudophryne. More re- cently, another new and as yet undescribed genus (Tyler et al., 1979) was discovered in the northern portion of Western Australia and the Northern Territory. This form produces a foam nest, has tadpoles with suctorial mouths and elongate tails, and the adult exhibits enor- mous tympana. The subdivision of existing genera ini- tiated by Tyler (1972b) and Blake (1973) took a further, and more radical, step with actions of Heyer and Liem ( 1976 ) who de- scribed three more new genera to accommo- date known species — Paracrinia for Crinia haswelli; Australocrinia to accommodate two southeastern species referred to Ranidella by Blake (1973); and KankanopJiryne for Pseu- dophryne occidentalis. They also resurrected Platyplectron without defining it or naming the constituent species. Unfortunately the data on which this study- by Heyer and Liem is based are, as yet, un- published, being available only in a paper by Liem cited as "in press." Thus, it is simply not possible to comprehend or assess several of the decisions reached by these authors and I am unable to recognize the genera in the following discussion. In attempting to provide a brief resume of the case for the familial and subfamilial status of the Australian frogs, I must at the outset put forth the evidence that has been provided for family distinction. Lynch ( 1973 ) and Savage (1973) well may be considered the prime initiators of the concept of the Myobatrachidae as a family unit distinct from the Leptodactylidae. In their respective views of the classification of the Anura they, and other authors, differed in many respects, and Duellman ( 1975 ) attempted to synthesize the various expressed opinions and produce a classification that constituted a compromise. Duellman recognized the distinctness of the Myobatrachidae, but his brief diagnoses did not include a single nongeographic feature by 96 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 which the majority of species of one family could be distinguished from those of the other. I draw attention to this fact not in an attempt to score a point, but solely to high- light the fact that the case for considering the Myobatrachidae a separate family still needs to be substantiated. Duellman (1975) was able to accommo- date Rheobatrachas within the Myobatrachi- dae, arguing that neither the suite of primi- tive morphological character states, nor the bizarre reproductive mode should exclude it. Lynch ( 1971 ) employed a graphic technique enabling him to compare the relative primi- tiveness of a large number of nonarchaic frogs. If Rheobatrachus is added to Lynch's frog groups and the 13 nonreproductive charac- ters are scored, Rheobatrachus has a total score of 0. It is the bizarre reproductive state that produces a positive sum. I wholly sup- port Heyer and Liem's ( 1976 ) action of plac- ing Rheobatrachus in a separate subfamily, the Rheobatrachinae. Therefore, within the Australian Region there is an enormous diver- sity of animals in terms of the nature of char- acter states, and a crucial question is whether the disjunction between the Australian and South American subfamilies is best reflected by regarding them as members of different families. Intrasubfamilial variation is more extensive among the Australian subfamilies than in any other comparable units on other continents, and it is this variation that renders their definition so difficult (Lvnch, 1973:170- 171). The South African Heleophryne forms the Heleophryninae, which Lynch placed in the Myobatrachidae. It would be extremely in- teresting to test this assignment by means of the comparison of serum albumins employing the techniques for frogs of Wallace, Maxson, and Wilson (1971). Because the African con- tinent separated from Gondwanaland be- tween the mid-Jurassic and mid-Cretaceous (100-155 m.y.b.p.), but South America from Gondwanaland in the Cenozoic (25-45 m.y.b.p.), the absence of myobatrachids from South America is curious indeed. Morescalchi ( 1973 ) and Morescalchi and Ingram ( 1974 ) noted that Cyclorana alboguttatus (now Li- toria albo guttata) is karyologically less dif- ferentiated than many Australian leptodac- tylid genera and approaches some primitive leptodactylids from other geographical areas ( Heleophryne, the Ceratophryinae, many Tel- matobiinae, all with 2n = 26). On gross morphological grounds, the re- semblance between some Australian and South American genera is striking. If the South American Ratrachyla should be found tomorrow in the cool, temperate forests of the southern section of the Australian Great Di- viding Range, it would be compared with Kyarranus and Philoria, found to be highly similar, and attract little comment. Lynch ( 1973 ) pointed out that there were several systematic options in any cladistic study of the leptodactyloid frogs, and he maintained that separating the Limnodynas- tinae ( "Cycloraninae" of Lynch without Cyclorana), Heleophryninae, and Myobatra- chinae from the Leptodactylidae, reduced the gradation of characters within the latter fam- ily. The knowledge of each of these units remains incomplete, and other more radical options are still open. For example, on bio- geographic grounds, frogs associated with he- leophrynines or myobatrachines should occur among the cool temperate Austral fauna of South America. Alternatively, other data might reinforce the existing myobatrachine- telmatobiine links, or the degree of distinc- tion between the Myobatrachinae and the Limnodynastinae, thereby meriting independ- ent family status for each of the latter. For the present, there seems to be a good case for including the Australian species in the Lep- todactylidae while the other avenues are being explored. Gekkonidae In recent years the status of the higher taxa of gekkonid lizards has attracted consid- erable attention. Underwood (1954) recog- nized three families — Eublepharidae, Sphae- rodactylidae, and Gekkonidae, with two subfamilies (Gekkoninae and Diplodactyli- nae ) . Kluge ( 1967 ) recognized only one family (the Gekkonidae) containing each of the other four units as subfamilies. Some of Kluge's concepts of the origins, dispersal, and evolutionary relationship of these subfamilies have been variously criticized by Maderson 1979 TYLER: SOUTH AMERICA AND AUSTRALIA 97 (1972), Moffat (1973), and Russell (1976). Kluge's ( 1967 ) interpretation was made with the assumption that continental drift was neither tenable as an hypothesis, nor germane to the resolution of the study. This led to conclusions that must be reexamined within the context of continental drift as an accept- able hypothesis. For example, there is his assumption that the Australian Diplodactyli- nae evolved from a primitive southeast Asian gekkonid stock during the Late Mesozoic. However, for that entire era, Australia lay far to the south and was still united to western Antarctica; the rift and long northwards drift towards southeast Asia only commenced in the early Cenozoic. Therefore, as Cracraft ( 1975) pointed out, the prospect of a success- ful overwater dispersal of geckos from Asia to Australia in the Mesozoic is remote indeed. In addition to hosting the pantropical gek- koninae, the endemic subfamilies of South America and Australia probably exhibit simi- lar historical patterns of gekkonid evolution. Certainly a significant feature of the disjunct- ly distributed eublepharines is their absence from South America, Madagascar, and Aus- tralia. Such a distribution is explained most readily by adopting Kluge's concept of an African or Asian site of origin. The general consensus of opinion is that the sphaerodac- tylines of South America and the diplodacty- lines of Australia evolved within their present geographic ranges. The diplodactylines ex- tended to New Caledonia, the Loyalty Islands, and New Zealand evolving in New Zealand to form an ovoviviparous group. Arrival of the diplodactylines in New Zealand has been sug- gested to be a Miocene event, but if the subfamily existed in Australia in the Cenozoic it could have entered New Zealand via the Lord Howe Rise. Many of the gekkonines are remarkably well suited to transoceanic dispersal. Phyllo- dachjlus has an incredible geographic range, occurring in the Americas, Africa, Madagas- car, Australia, New Caledonia, New Zealand, and the Galapagos Islands. However, Dixon and Anderson (1973) indicated that Phyllo- dactylus may be heterogeneous; species in the Eastern Hemisphere should be separated ge- nerically from PhyUodactyhis. A number of gekkonines are dispersed by man. Recent karyotypic studies in Australia have demon- strated within well-established gekkonine and diplodactyline species the existence of numer- ous biological species, none of which has yet been accorded formal taxonomic status ( King, 1973, 1975, 1977; King and Rofe, 1976). For example, King (1977) recognized five dis- cretely distributed and chromosomally dis- tinct populations within what is now termed Diplodactylus vittatus. When all of these new taxa are named, it is clear that the Australian gekkonid fauna will be vast numerically. Chelidae Presently, the pleurodire chelonians are restricted to South America, Africa, Australia and New Guinea. Of the constituent families now extant, the Chelidae occurs in the Neo- tropical and Australian Regions, whereas the Pelomedusidae occurs in Africa, Madagascar and South America. In terms of diversity and radiation, the South American genera and spe- cies have been more labile, and study of the extant Australian members suggests an ex- ceptionally conservative history. In a phylo- genetic study, Gaffney (1977) envisaged the Australian genus Pseudemydura as possibly the sister group of all other Australian and South American genera. At present, there are four recognizable genera of extant Australian chelids — Chelo- dina, Elseya, Emydura and Pseudemydura. Fossil records are becoming quite numerous, but the majority of them are based upon totally disarticulated shell fragments lacking sufficient detail to provide adequate diagnos- tic characters for generic determinations. The first report of fossil chelids from Aus- tralia is that of Lydekker (1889) who re- ported fragments of CheJodina and Emydura from various localities. De Vis (1894) pro- posed Trionyx australiemis for a substantial quantity of fragments (probably chelid) taken at Darling Downs, and then (1897) described one new genus and four new spe- cies from the same or adjacent localities. These specimens are in urgent need of re- view. Records of Pleistocene chelids from Queensland reported by Longman ( 1929 ) are fragmentary and lack diagnostic characters (Warren, 1969). Evidence of the conservative nature of the 98 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 chelid fauna was provided by Warren ( 1969 ) who reported Emydura sp. aff. macquari from siltstones in Tasmania. Reported to be of Oligocene-Miocene age they are now con- sidered somewhat older and of Early Tertiary age (J. W. Warren, pers. coram.). Chelids are no longer extant in Tasmania, and E. mac- quari is now confined to the Murray-Darling drainage system of the southeastern Australian mainland. With the discovery of freshwater turtle remains from Tasmania and also from various Miocene to Pleistocene deposits on the mainland (Callen and Tedford, 1976; Archer and Wade, 1976), it is likely that previously chelids have occupied almost all of the Aus- tralian continent and New Guinea as well. Contraction of their ranges is probably a Pleistocene phenomenon. The intercontinental relationships of che- lids is, superficially at least, extremely close (e.g., Elseya and Platemys). However, the Lower Cretaceous Chelycarapookus arcuatus Warren (1975) (Chelycarapookidae) (previ- ously identified erroneously as Emydura mac- quari by Chapman, 1919) needs to be exam- ined. Unfortunately, with the posterior por- tion of the plastron of that fossil missing, whether the pelvic girdle was attached or not remains unknown, so that even the infra-order position of the family is uncertain. Warren ( 1975 ) noted that in Chelycarapookus neurals probably were present between all costals, and suggested tentatively that with loss of the neurals, Australian chelids could have been derived from a chelycarapookid ancestor. However Rhodin and Mittermeier (1977), ap- parently without sighting the description of Chelycarapookus, reported that neurals occur regularly in one species of Australian chelid and irregularly in several other species. Gaff- ney's ( 1975 ) study of the phytogeny and classification of turtles almost exclusively re- lied upon cranial characters, and, in common with most Australian chelid remains, the head and neck of Chelycarapookus remain un- known. Crocodiles The crocodiles include two quite distinct components differing in their ancestry and arc probably of separate Gondwanan and Orien- tal origins. The extant species of Crocodylus are clearly of Oriental origin or derived from an Oriental stock, and are confined to the north of the Region — C. porosus of south- eastern Asia, New Guinea and Australia, C. novaeguineae of New Guinea and C. johnsoni of Australia. The fossil fauna is substantial both in quantity and diversity, and a number of highly significant finds has been reported re- cently. Molnar ( 1977 ) described from Chil- lagoe in North Queensland an incomplete skull, with a high and laterally compressed snout and probably xiphodont dentition. The subsequent discoveiy of Pleistocene Palor- chestes cf. P. azael at the same site was in- terpreted as evidence of a Pleistocene age for the xiphodont (Molnar, 1978), and indi- cated that xiphodonts had survived in Aus- tralia long after their extinction elsewhere in the world. Hecht and Archer ( 1977 ) reported two forms of xiphodonts from the Pleistocene of South Australia and southeast Queensland, re- spectively. The former is reported to compare favorably with the type of the sebecosuchian Sebecus icaeorhinus from the Eocene of Pata- gonia. The authors suggested that the as- sumed sebechosuchian Planocrania datangen- sis from the Early Tertiary of China is in reality probably an eusuchian, so that origin of the Sebecosuchia is clearly from Gond- wanaland. AUTOCHTHONOUS ELEMENT Pygopodidae The legless, fossorial lizard family Pygo- podidae is the only reptilian family restricted to the Australian Region. Kluge (1974) rec- ognized eight genera and 30 species of pygo- podids. Kluge's (1976) analysis of phylo- genetic relationships within the family led to the recognition of only six genera. Following the work of Underwood (1957), there has been general acceptance that the pygopodid- gekkonid relationship is extremely close, and it follows that the most likely origin for the pygopodids is within Australia directly from a gekkonid stock. Many pygopodid species are restricted 1979 TYLER: SOUTH AMERICA AND AUSTRALIA 99 either to the southwest or to the southeast of the continent, a distribution pattern common to numerous vertebrates and attributed to speciation in the Pleistocene, associated with the glacial-interglacial climatic oscillations. Thus, the implication of the existence of such distribution patterns is that speciation of these same populations is not of any great antiquity. Somewhat in conflict is the total absence of pygopodids from the southern island of Tas- mania and the presence of only a single spe- cies on Kangaroo Island. If these indicate that pygopodids occupied the adjacent main- land only after isolation of these islands from the mainland (8,000-10,000 y.b.p.), the southern speciation pattern evidently is more complex. Western Pacific Island Faunas and Oceanic Dispersal Although South America now is separated from Australia by the vast expanses of the Pacific Ocean, the nature of the major oceanic surface currents provides a mechanism for westward dispersal of animals by rafting from South America in the direction of Australia. The west coast of South America is swept by the north flowing Humboldt Current which meets the transpacific southern equatorial cur- rent at mid-latitudes. The latter current travels westward and eventually disperses around all of the islands of the Pacific south of the Equator. Thirty years ago Thor Heyerdahl's raft Kon-Tiki demonstrated the transport potential of the southern equatorial current by travel- ling from Peru to the Tuamotu Archipelago in French Polynesia south of the Marquesas Islands. Had his vessel been driven two or three degrees northwards, he woidd have passed between the Marquesas and the Tua- motu Archipelago, and continued much far- ther west on a longer journey terminating in Western Samoa, Fiji, or Tonga. This longer journey is probably the route of the ancestral stocks of the iguanids of Fiji and Tonga — Brcichylophus fasciatus and B. brevicephalus, respectively. South American derivates are not evident elsewhere, and the remaining ele- ments of the herpetofauna of the islands in the west and southwest Pacific area are repre- sentative of two other sources — 1) overwater dispersal principally from the north and north- east, and 2) land communication with Aus- tralia via the exposed Lord Howe Rise. Fiji and New Zealand are the most south- erly of the large landmasses in the Pacific and merit specific mention. Fiji. — In addition to the iguanid Brachy- lophus (discussed by Cogger, 1974), Gorham ( 1965) stated that an additional 14 lizards oc- cur on Fiji. These are gekkonids and scincids, but only one possibly is endemic. The pres- ence of two boids, one elapid (endemic), and one typhlopid so far southeast tends to sup- port the hypothesis that these are all of Ori- ental source and not Gondwanan elements. As noted previously, the two endemic ranids on Fiji (Platy mantis vitiensis and P. vitianus) were derived from the same source route. In the case of the ranids, the absence of frogs from the intermediate potential stepping stones of New Hebrides and New Caledonia remains an apparently inexplicable anomaly. Additional references dealing with the Fijian herpetofauna are Barbour ( 1923 ) , Brown and Myers (1949), and Gorham (1968). Neic Zealand. — The most well known com- ponent of the New Zealand herpetofauna is the rhynchocephalian reptile, the Tuatara Sphenodon punctatus. This relic was prob- ably quite widely distributed elsewhere in the Mesozoic, as evidenced by the extensive fos- sil record of rhynchocephalians at that time. The remaining terrestrial reptiles are 35 spe- cies of lizards (almost all are endemic) of the families Gekkonidae and Scincidae. The geckos include three endemic genera that are unique among gekkonids in being ovovivipar- ous, whereas the endemic skinks are members of genera widely distributed outside New Zealand. In their recent survey of New Zealand vertebrates, Bull and Whittaker (1975) ac- cepted Kluge's (1967) interpretations of the origin of the gekkonids, resulting in their statement (op. cit: 239): "The geckos form part of the Malayo-Pacific element of the New Zealand fauna and probably entered New Zealand in the Miocene when the climate was warmer and the land more extensive than now." They further visualized fairly extensive oceanic dispersal by rafting from New Cale- 100 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 donia or directly from Australia. A totally dif- ferent interpretation would result by consider- ing the connection between New Zealand and the Lord Howe Rise and the exceptionally close proximity of the Lord Howe Rise to Australia in Early Cenozoic (Griffiths and Varne, 1972). If geckos really arrived in New Zealand no earlier than the Miocene, they have been evolving rapidly ever since along a unique path. Therefore, acceptance of an Early Ceno- zoic entry avoids any concept of an explosive radiation, and provides an adequate time span for speciation in situ along novel lines. Kluge's (1967) hypothesis of gekkonid evolution and dispersal did not accommodate continental drift. Accordingly, it is not surprising that his interpretations of the source of the faunal ancestors is at variance with one incorporat- ing this phenomenon. The endemic frog fauna of New Zealand is even more bizarre than the reptile fauna and is represented by three species of Leio- pelma. Whether Leiopelma and Ascaphus of North America should be placed in a single family ( Ascaphidae), or whether Leiopelma should constitute the Leiopelmatidae remains a matter of debate. However, there is no argument to the concept that these genera represent relics of a fauna that was widely distributed. Estes and Reig ( 1973 ) referred Vieraella and Notobatrachus of the Early and Late Jurassic of Patagonia to the Ascaphidae. Until recently, Leiopelma hamiltoni was known only from a small heap of stones oc- cupying one quarter of a hectare on the South Island. Now it is known from an additional fifteen hectares on the small Maud Island off the coast (Rull and Whitaker, 1975). Leio- pelma archeyi and L. hochstetteri are dis- tributed somewhat more widely on the North Island, but Bull and Whitaker (op. cit: 235) stated that in the recent past (". . . probably within the last 1,000 years. . . .") Leiopelma was far more widespread, being known from five sites where it no longer occurs today. They described subfossil material as being about twice the size of the extant species, but otherwise similar in skeletal features. Thus, the animals involved would have been as much as 100 mm in length. Either the mor- phological change to existing species was ef- fected in a millenium or the subfossils repre- sent extinct species; in either case it seems to be unprofitable to speculate about the char- acteristics of their immediate postdrift an- cestors. Within the context of discussion of south- west Pacific biological origins, the concepts of Nur and Ben-Avraham (1977) on a lost Pa- cific continent (formerly lying close to the east coast of Australia) must be considered. Nevertheless the present study of the anurans has not required such a landmass to explain their origins. CONCLUSIONS When the amphibian and reptile families now found in Australia are examined, one by one, to determine whether their affinities lie with South American or with Oriental stocks, it rapidly becomes apparent that Oriental sources predominate, and that Australian- South American links are few indeed. At the commencement of drifting in the Eocene, the Australian herpetofauna included the follow- ing families shared with South America. Gekkonidae (Diplodactylinae). — The na- ture of the extensive radiation within Aus- tralia may well support the concept that the Gekkonidae was the only lizard family pres- ent in Australia. Boidae (P Madtsoinae). — The presence of this family hinges upon the Pleistocene Wo- nambi whose phylogenetic affinities are with snakes outside the Oriental Region. Extant genera certainly arrived in the Miocene. Chelidae. — A Gondwanan component probably of considerable antiquity. Within Australia, fossils extending to the Early Ter- tiary represent modern species. The origin of the family is uncertain, but the Lower Cre- taceous Chehjcarapookus arcuatus, for which Warren (1975) erected the Chelycarapooki- dae, exhibits postcranial features that render it a potential chelid ancestor. Crocodijlidac. — The first fossils of a sebe- cosuchian crocodile fauna have just been dis- covered. Hylidae. — Previous concepts of South American and Australian tree frogs represent- ing a single family do not as yet appear to have been refuted. 1979 TYLER: SOUTH AMERICA AND AUSTRALIA 101 Leptodactylidae. — Irrespective of the final assessment of the familial disposition of the Australian genera and species, the South American affinities of the Australian members are indisputable. These six families represent the elements of the Australian herpetofauna destined to persist through to the Holocene. It follows that by modern standards the total herpe- tofauna was incredibly depauperate, and lacked many significant elements. Pygopodids evolved in Australia probably during the period of isolation. The serious deficiencies were remedied from the Miocene onwards as Australia and the eastern outliers of the Oriental Region approached one another on their collision course. The commencement of colonization by families such as the Scincidae probably antedated the mid-Miocene. Others such as the Typhlopidae, Carettochelyidae, Microhy- lidae, Varanidae, and Crocodylidae probably were acquired at the time of the collision, while the Ranidae entered then and again in a subsequent wave. In broad terms, the Ori- ental influence upon Australia probably was more significant than the North American in- fluence upon South America, because South America had a more diverse and better estab- lished herpetofauna before the time of con- tact. ACKNOWLEDGMENTS For the invitation to participate in the symposium "The South American Herpeto- fauna," and for the funding making this pos- sible, I am greatly indebted to The University of Kansas, and particularly to the convenor, William E. Duellman. Many of the ideas and assessment pro- posed in this paper arose from or were stimu- lated by discussions or correspondence with many of my colleagues. For this great help I gratefully acknowledge the contributions of Bob Lange, John M. Legler, Brian McGow- ran, Allen E. Greer, Rowley Twidale and George R. Zug. Margaret Davies prepared figures 4:1 and 4:10, whereas figures 4:3-4:7 and 4:9 are the work of Debra Bennett. The manuscript was typed by Mrs. J. Russell-Price. I am also indebted to Richard G. Zweifel for permis- sion to reproduce the map of New Guinea upon which figure 4:5 was prepared. RESUMEN Las oportunidades para un intercambio herpetofaunistico gondwanico entre Sud America y Australia fueron muchos menores que aquellos entre Sud America y Africa, debido a la expansion de la Antartica. Al comienzo de la separation de Australia de Sud America (53 m.a.a.p.) el continental su- reno Australiano era subtropical. Condiciones humedas con una diversidad de anfibios y reptiles occupaban el centra de Australia hasta el Pleistoceno superior cuando un incremento de la aridez elimino muchos sitios acuaticos. De todos los posibles modos de establecer cuales segmentos de la herpetofauna australi- ana eran compartidas con Sud America como resultado de un intercambio gondwanico, se ha seleccionado determinar que familias que estan hoy presentes en Australia y Nueva Guinea, fueron adquiridas cuando Australia coludio con la Region Oriental en el Mioceno medio. Un analisis de tal evento demuestra que un grupo de islas adheridas a lo que es la actual costa norte de Nueva Guinea. Por ende, si estas islas estaban poblados con an- fibios y reptiles algunos, al menos, debieron haberse integrado a la heipetofauna australo- papua a traves de esta via. Se propone un modelo que involucra la dispersion de algunos generos a traves de las Islas Orientales desde las Filipinas a Fiji en el Mioceno (el genera de batracios ranidos Platy mantis). La colision de tierras arego este rana a la fauna de Nueva Guinea y los actuales extremos de diversidad al Oeste y al Este de Nueva Guinea reflejan la edad de especiacion. En contraste otros generos arri- baron desde el Oeste despues de la colision, resultando en una progresiva reduction del numero de especies desde el Oeste al Este. Entonces, los patrones de distribution geo- grafica de los elementos Orientales en Aus- tralia varian de acuerdo con la fecha de en- trada relativa a la colision de Australia. Sin embargo, el topico mas importante es las rela- 102 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 cion de los anfibios y reptiles de Australia que aquellos de la Region Oriental adyacente. Bajo esta criterio, las Ranidae, Microhylidae, Varanidae, ? Scincidae, Agamidae, Caretto- chelyidae, Trionychidae, Elapidae, Colubri- dae, Acrochordidae, Uropeltidae, Typhlopi- dae, y Boidae entraron a la region geografica australiana despues del Mioceno medio. Los Boidae son linicos, por tener un genuino com- ponente gondwanico (si Wonambi del Pleis- toceno se relaciona con Madtsoia), y otro Oriental. Correspondientemente, la hcrpetofauna gondwanica de Australia no poseia la gran mayoria de los elementos de esa fauna. Un buen ejemplo esta dado por la presencia de los Hylidae y Leptodactylidae ( y por la man- tencion del uso de estos nombres familiares para los animales modernos de Australia). Aparte de estos, existian solo los Cheliidae y los geckos diplodactylinos, de los cuals los Pygopodidae probablemente evolucionaron como los linicos representantes australianos. Una deriva pasiva permitio a los largartos iguanidos viajar largas distancias desde la costa oeste de Sud America a Fiji y a Tonga. A pasar de que la direccion de las corrientes transpacificas (Este a Oeste) favorece Is dis- persion desde Sud America a Australia mas que la direccion opuesta, los inmigrantes sud- americanos han alcanzado las islas del Pa- cifico Sur, pero aim no han filtrado has al- canzar las costas del continente australiano. LITERATURE CITED Archer, M., Wade, M. 1976. Results of the Ray E. Lemley Expeditions, Part 1. The Allingham For- mation and a new Pliocene vertebrate fauna from northern Queensland. Mem. Queensland Mus. 17(3): 379-397. Axelrod, D. I. 1960. The evolution of flowering plants, pp. 227-305 in Tax, S. (ed.). Evolution after Darwin, 1, Univ. Chicago Press, Chicago, 629 p. Axelrod, D. I. 1975. Plate tectonics and problems of angiospeim history. Mem. Mus. Natl. Hist. Nat., Paris Sec. C 88:72-86. Bagnara, J. F., Ferris, W. 1975. The presence of phyllomedusine melanosomes and pigments in Australian hylids. Copeia 1975(3) :592-595. Bagnara, J. T. 1976. Color change, pp. 1-52 in Lofts, B. (ed.). Physiology of the Amphibia. Vol. Ill, Academic Press, New York, 597 p. Bai, A. R. K. 1956. Analysis of chromosomes in two genera of Microhylidae (Amphibia: Anura). Proc. Natl. Inst. Sci. India, Part B 22:1-5. Barbour, T. 1923. The frogs of the Fiji Islands. Proc. Acad. Nat. Sci. Philadelphia 75:111-115. Beaufort, L. F. de. 1951. Zoogeography of the land and inland waters. Sidgwick and Jackson, London, 208 p. Becak, M. L., Demaro, L., Becak, W. 1970. Poly- ploidy and mechanisms of karyotypic diversifica- tion in Amphibia. Cytogenetics 9:225-238. Blake, A. J. D. 1973. Taxonomy and relationships of myobatrachine frogs ( Leptodactylidae ) : A numerical approach. Australian ]. Zool. 21:119- 149. Blommers, R. 1971. Karyotype de Anodontohyla montana Angel (Anura, Microhylidae) de massif de TAndringitra (Madagascar). Terre Mai. 10: 261-265. Blommers-Schlosser, R. M. A. 1976. Chromosomal analysis of twelve species of Microhylidae (Anu- ra) from Madagascar. Genetica 46:199-210. Bocart, J. P., Nelson, C. E. 1976. Evolutionary implications from karyotypic analysis of the fami- lies Microhylidae and Rhinophrynidae. Herpeto- logica 32:199-208. Bogart, J. P., Pyburn, W., Nelson, C. E. 1976. The karyotype of Otophnjne robusta (Anura: Micro- hylidae). Ibid. 32:208-210. Bowler, J. M. 1976. Aridity in Australia: Age, ori- gins and expression in aeolian landforms and sedi- ments: Earth-Sci. Rev. 12:279-310. Brown, W. C, Alcala, A. C. 1970. The zoogeog- raphy of the herpetofauna of the Philippine Islands, a fringing archipelago. Proc. California Acad. Sci. 38(6) : 10.5-130. Brown, W. C, Myers, G. S. 1949. A new frog of the genus Cornufer from the Solomon Islands, with notes on the endemic nature of the Fijian frog fauna. Amer. Mus. Novit. (1418): 1-10. Bull, P. C, Whitaker, A. H. 1975. The amphib- ians, reptiles, birds and mammals, pp. 231-276 in Kuschel, G. (ed.). Biogeography and ecology in New Zealand. W. Junk, The Hague, 689 p. Callen, R. A., Tedford, R. H. 1976. New late Cainozoic rock units and depositional environ- ments, Lake Frome area, South Australia. Trans. Roy. Soc. South Australia 100:125-168. Chapman, F. 1919. On a fossil tortoise in ironstone from Carapook, near Casterton. Proc. Roy. Soc. Victoria 32 (newser.) (1):11-13. Christophel, D. C, Blackburn, D. T. 1978. Terti- ary megafossil flora of Maslin Bay, South Australia: A preliminary report. Alcheringa 2:311-319. Cogger, H. G. 1970. First record of the pitted- shelled turtle Carcitocliehjs insculpta from Aus- tralia. Search 1:41. Cocger, H. G. 1974. Voyage of the banded iguana. Australian Nat. Hist. 18:144-149. Cogger, H. G. 1975. Reptiles and amphibians of Australia. Reed, Sydney, 584 p. Cole, C. J., Zweifel, R. G. 1971. Chromosomes of a New Guinean microhylid frog, Cophixalus ri- IHirius Zweifel. Herpetol. Rev. 3:15-16. 1979 TYLER: SOUTH AMERICA AND AUSTRALIA 103 Coleman, P. J. 1975. The Solomons as a non-arc. Bull. Australian Soc. Explor. Geophys., 6(2-3): 60-61. Coleman, P. J., Packham, G. H. 1976. The Melane- sian borderlands and India-Pacific plates' bound- ary. Earth-Sci. Rev. 12:197-233. Corben, C. J., Ingram, G. J., Tyler, M. J. 1974. Gastric brooding: Unique form of parental care in an Australian frog. Science 186:946-947. Cracraft, J. 1973. Vertebrate evolution and bio- geography in the Old World tropics: Implications of continental drift and palaeoclimatology, pp. 373-393 in Tarling, D. H., Runcorn, S. K. (eds. ). Implications of continental drift to the earth sciences. Vol. 1. Academic Press, London, 622 p. Cracraft, J. 1974. Continental drift and vertebrate distribution. Ann. Rev. Ecol. Syst. 5:215-261. Cracraft, J. 1975. Mesozoic dispersal of terrestrial faunas around the southern end of the world. Mem. Mus. Natl. Hist. Nat. Paris, Ser A 88:29- 54. Darlington, P. J. 1957. Zoogeography: The geo- graphical distribution of animals. Wiley, New York, 675 p. Denham, D. 1975. Distribution of underthrust litho- spheric slabs and focal mechanisms — Papua New Guinea and Solomon Islands Region. Bull. Aus- tralian Soc. Explor. Geophys. 6(2-3) : 78-79. De Vis, C. W. 1894. The lesser chelonians of the Nototherian drifts. Proc. Roy. Soc. Queensland 10:123-127. De Vis, C. W. 1897. The extinct fresh-water turtles of Queensland. Ann. Queensland Mus. 3:3-7. Dixon, J. R., Anderson, S. C. 1973. A new species and genus of gecko (Sauria: Gekkonidae) from Iran and Iraq. Bull. South. California Acad. Sci. 72:155-160. Duellman, W. E. 1970. The hylid frogs of Middle America. Univ. Kansas Mus. Nat. Hist. Monogr. ( 1 ) : 1-753. Duellman, W. E. 1975. On the classification of frogs. Univ. Kansas Mus. Nat. Hist. Occas. Pap. (42): 1-14. Emerson, S. B. 1976. A preliminary report on the superficial throat musculature of the Microhylidae and its possible role in tongue action. Copeia 1976 (3):546-551. Estes, R., Reig, O. A. 1973. The early fossil record of frogs: A review of the evidence, pp. 11-63 in Vial, J. L. (ed.). Evolutionary biology of the anurans: Contemporary research on major prob- lems. Univ. Missouri Press, Columbia, 470 p. Falvey, D. A., Taylor, L. W. H. 1974. Queens- land Plateau and Coral Sea Basin: Structure and time — stratigraphic patterns. Bull. Australian Soc. Explor. Geophys. 5(4) : 123-126. Fooden, J. 1972. Breakup of Pangaea and isolation of relict mammals in Australia, South America, and Madagascar. Science 175:894-898. Gaffney, E. S. 1975. A phylogeny and classification of the higher categories of turtles. Bull. Amer. Mus. Nat. Hist. 155:387-436. Gaffney, E. S. 1977. The side-necked turtle family Chelidae: A theory of relationships using shared derived characters. Amer. Mus. Novit. ( 2620 ) : 1-28. Claessner, M. F. 1942. The occurrence of the New Guinea turtle (Carettochchjs) in the Mio- cene of Papua. Rec. Australian Mus. 21:106-109. Glassford, D. K., Killigrew, L. P. 1976. Evidence for Quaternary westward extension of the Aus- tralian desert into south-western Australia. Search 7:394-396. Gorham, S. W. 1965. Fiji frogs (with synopses of the genera Cornufcr and Plat y mantis). Zool. Beitr. 11:381-435. Gorham, S. W. 1968. Fiji frogs. Life history data from field work. Ibid. 11:427-446. Gow, G. F. 1977. A new species of Pytlion from Arnhem Land. Aust. Zool. 19:133-139. Greer, A. E. 1970. A subfamilial classification of scincid lizards. Bull. Mus. Comp. Zool. Harvard Univ. 139:151-184. Griffiths, J. R., Barne, R. 1972. Evolution of the Tasman Sea, Macquarie Ridge and Alpine Fault. Nature (Phys. Sci.) 235:83-86. Guibe, J. 1948. Contribution a l'etude de l'appareil genital des tvphlopides (Ophidiens). Bull. Soc. Zool. France 73:224-228. Guttman, S. I. 1973. Biochemical techniques and problems in anuran evolution, pp. 183-203 in Vial, J. L. (ed. ). Evolutionary biology of the anurans: Contemporary research on major prob- lems. Univ. Missouri Press, Columbia, 470 p. Hecht, M. K. 1975. The morphology and relation- ships of the largest known terrestrial lizard, Mega- lania prisca Owen, from the Pleistocene of Aus- tralia. Proc. Roy. Soc. Victoria (new ser.) 87(2): 239-250. Hecht, M. K, Archer, M. 1977. Presence of xipho- dont crocodilians in the Tertiary and Pleistocene of Australia. Alcheringa 10:383-385. Heyer, W. R., Liem, D. S. 1976. Analysis of the intergeneric relationships of the Australian frog family Mvobatrachidae. Smithsonian Contri. Zool. 233:1-29. Hoffstetter, R. 1961. Nouveaux restes d'un ser- pent boide (Madtsoia madagascahcnsis nov. sp.) dans la Cretace superieur de Madagascar. Bull. Mus. Natl. Hist. Nat. Zool, Paris (2) 33:152-160. Inger, R. F. 1954. Systematics and zoogeography of Philippine amphibia. Fieldiana Zool. 33(4): 181- 531. Incer, R. F. 1966. The systematics and zoogeog- raphy of the amphibia of Borneo. Ibid. 52:1-402. Jardine, N., McKenzie, D. 1972. Continental drift and the dispersal and evolution of organisms. Nature 234(5332) : 20-24. Keast, A. 1971. Continental drift and the evolu- tion of the biota on southern continents. Q. Rev. Biol. 46:335-378. Keast, A. 1973. Contemporary biotas and the sep- aration sequence of the southern continents, pp. 309-343 in Tabling, D. H., Runcorn, S. K. ( eds. ) . Implications of continental drift to the earth sciences. Vol. 1. Academic Press, London, 622 p. 104 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 King, M. 1973. Chromosomes of two Australian lizards of the families Scincidae and Gekkonidae. Cytologia 38:20.5-210. Kinc, M. 1975. Chromosomal studies on Australian lizards. Ph.D. Dissert. Univ. Adelaide, 150 p. King, M. 1977. Chromosomal and morphometric variation in the Gekko Diplodactylus vittatus Gray. Australian J. Zool. 25:43-57. King, M., King, D. 1975. Chromosomal variation in the lizard genus Varanus (Reptilia). Australian J. Biol. Sci. 28:89-108. King, M., Rofe, R. 1976. Karyotypic variation in the Australian Cecko Phyllodactylus marmoratus (Gray) (Gekkonidae: Reptilia). Chromosoma 54:75-87. Kluge, A. G. 1967. Higher taxonomic categories of gekkonid lizards and their evolution. Bull. Amer. Mus. Nat. Hist. 135:1-60. Kluge, A. G. 1974. A taxonomic revision of the lizard family Pvgopodidae. Misc. Publ. Mus. Zool. Univ. Michigan (147): 1-221. Kluge, A. G. 1976. Phylogenetic relationships in the lizard family Pygopodidae: An evaluation of theory, methods and data. Ibid. (152): 1-72. Lance, R. T. 1976. Fossil epiphyllous "Germlings," their living equivalents and their palaeohabitat indicator value. Neues Jahrb. Geol. Palaeontol. Abh. 151:142-165. Latz, P. K. 1975. Notes on the relict palm Livistona mariae F. Muell. in central Australia. Trans. Roy. Soc. South Australia 99:189-195. Laurent, R. F. 1975. La distribution des amphib- iens et les translations continentales. Mem. Mus. Natl. Hist. Nat., Paris 88:176-191. Leon, P. E. 1970. Report of the chromosome num- bers of some Costa Rican anurans. Rev. Biol. Trop. 17:119-124. Liem, D. S. 1973. A new genus of frog of the family Leptodactylidae from SE. Queensland, Australia. Mem. Queensland Mus. 16:459-470. Longman, N. A. 1929. Palaeontological notes. Specimens from a well at Brigalow. Ibid. 9:247- 252. Lydekker, R. L. 1889. Catalogue of fossil amphib- ians and reptiles in the British Museum ( Natural History) Part 3. British Museum, London, 254 p. Lynch, J. D. 1971. Evolutionary relationships, oste- ology and zoogeography of leptodactyloid frogs. Univ. Kansas Mus. Nat. Hist. Misc. Publ. (53): 1-238. Lynch, J. D. 1973. The transition from archaic to advanced frogs, pp. 133-182 in Vial, J. L. (ed.). Evolutionary biology of the anurans: Contem- porary research on major problems. Univ. Mis- souri Press, Columbia, 470 p. Mabbutt, J. A. 1967. Denudation chronology in Central Australia. Structure, climate and land- form inheritance in the Alice Springs area, pp. 144-181 in Jennings, J. N., Mabbutt, J. A. (eds.). Landform studies from Australia and New Guinea. Australian Natl. Univ. Press, Can- berra, 434 p. MacKenzie, D. E. 1975. Volcanic and plate tectonic evolution of central Papua and New Cuinea. Bull. Australian Soc. Explor. Geophys. 6(2-3) :65-67. Maderson, P. F. A. 1972. The structure and evolu- tion of holocrine glands in sphaerodaetyline and eublepharine gekkonid lizards. Copeia 1973 (3): 559-571. Main, A. R., Lee, A. K, Littlejohn, M. J. 1958. Evolution in three genera of Australian frogs. Evolution 12:224-233. Maxson, L. R. 1976. The phylogenetic status of phyllomedusine frogs ( Hylidae ) as evidenced from immunological studies of their serum albu- mins. Experientia 32:1149-1150. Maxson, L. R., Wilson, A. C. 1975. Albumin evo- lution and organismal evolution in tree frogs (Hylidae). Syst. Zool. 24:1-15. McDowell, S. B. 1967. Aspidomorphus, a genus of New Guinea snakes of the family Elapidae, with notes on related genera. J. Zool. 151:497-543. McDowell, S. B. 1970. On the status and relation- ships of the Solomon Island elapid snakes. Ibid. 161:145-190. McDowell, S. B. 1972. The species of Stegoiwtus (Serpentes, Colubridae) in Papua New Guinea. Zool. Meded. Rijks Mus. Nat. Hist. Leiden 47: 6-26. McDowell, S. B. 1974. A catalogue of the snakes of New Guinea and the Solomons, with special reference to those in the Bernice P. Bishop Mu- seum. Part 1, Scolecophidia. J. Herpetol. 8:1-57. McDowell, S. B. 1975. A catalogue of the snakes of New Guinea and the Solomons, with special reference to those in the Bernice P. Bishop Mu- seum. Part II, Anilioidea and Pvthoninae. Ibid. 9:1-79. McDowell, S. B., Bocert, C. M. 1954. The sys- tematic position of Lanthanotus and the affinities of the anguimorphan lizards. Bull. Amer. Mus. Nat. Hist. 105:1-142. McGowran, B. 1973. Rifting and drift of Australia and the migration of mammals. Science 180: 759-761. Menzies, J. I., Tippett, J. 1976. Chromosome num- bers of Papuan hylid frogs and the karyotype of Litoria infrafrenata (Amphibia, Anura, Hvlidae). J. Herpetol. 10:167-173. Menzies, J. I., Tyler, M. J. 1977. The systematica and adaptations of some Papuan microhylid frogs which live underground. J. Zool. 183:431-464. Mittleman, M. B. 1952. A generic synopsis of the lizards of the subfamily lygosominae. Smithson- ian Misc. Collect. 117:1-35. Moffatt, L. 1973. The concept of primitiveness and its bearing on the phylogenetic classification of the Gekkota. Proc. Linn. Soc. New South Wales 97:275-301. Molnar, R. E. 1977. Crocodile with laterally com- pressed snout: First find in Australia. Science 197:62-64. Molnar, R. E. 1978. Age of the Chillagoe crocodile. Search 9(4) : 156-158. 1979 TYLER: SOUTH AMERICA AND AUSTRALIA 105 Morescalchi, A. 1968a. Initial cytotaxonomic data on certain families of amphibious Anura ( Diplas- iocoela, after Noble, T. C. ). Experientia 24:280- 283. Morescalchi, A. 1968b. Hypothesis on the phytog- eny of the Salientia, based on karyological data. Ibid. 24:964-966. Morescalchi, A. 1973. Amphibia, pp. 233-248 in Chiarelli, A. B., Capanna, E. (eds.). Cytotax- onomy and vertebrate evolution. Academic Press, London, 783 p. Morescalchi, A., Ingram, G. J. 1974. New chro- mosome numbers in Australian Leptodactylidae (Amphibia, Salientia). Experientia 30:1134— 1135. Natarjan, R. 1953. A note on the chromosomes of Cacopus systoma. Proc. 40th Indian Sci. Congress, Lucknow pp. 180-181. Nelson', C. E. 1975. Another new miniature 4-toed South American microhylid frog (Genus: Syn- cope). J. Herpetol. 9:81-84. Newsome, A. E., Rochow, K. A. 1964. Vertebrate fossils from Tertiary sediments in central Aus- tralia. Australian J. Sci. 26:252. Nur, A., Ben-Avraham, Z. 1977. Lost Pacifica con- tinent. Nature 270:41-45. Parker, H. W. 1929. Two fossil frogs from the Lower Miocene of Europe. Ann. Mag. Nat. Hist. (10) 4 (21):270-281. Parker, H. W. 1934. A monograph of the frogs of the family Microhylidae. British Museum (Na- tural History), London, 208 p. Parker, H. W. 1940. The Australasian frogs of the family Leptodactylidae. Novit. Zool. 42:1-106. Pianka, E. R. 1972. Zoogeography and speciation of Australian desert lizards: An ecological perspec- tive. Copeia 1972 (1):127-144. Rabello, M. N. 1970. Chromosomal studies in Bra- zilian anurans. Caryologia 23:45-59. Rhodin, A. G. J., Mittermeier, R. A. 1977. Neural bones in chelid turtles from Australia and New Guinea. Copeia 1977 (2):370-372. Rich, P. V. 1975. Antarctic dispersal routes, wan- dering continents, and the origin of Australia's non-passeriform avifauna. Mem. Natl. Mus. Vic- toria (36):63-125. Robb, J. 1966. The structure and possible function of the cloacal pouches of male Australian typhlo- pids. Australian J. Zool. 14:27-30. Robinson, R. L., Tyler, M. J. 1972. The catecho- lamine content of the adrenal glands of frogs as an index of phylogenetic relationships. Comp. Gen. Pharmacol. 3:167-170. Russell, A. P. 1976. Some comments concerning interrelationships amongst gekkonine geckos, pp. 217-244 in Bellaxrs, A. d'A., Cox, B. (eds.). Morphology and biology of reptiles. Linn. Soc. Symp. (Ser. 3), 290 p. Savage, J. M. 1973. The geographic distribution of frogs: Patterns and predictions, pp. 351-445 in Vial, J. L. (ed.). Evolutionary biology of the anurans: Contemporary research on major prob- lems. Univ. Missouri Press, Columbia, 470 p. Schllnger, E. I. 1974. Continental drift, Notho- fagus, and some ecologically associated insects. Ann. Rev. Entomol. 19:323-343. Schodde, R., Mason, I., Wolfe, T. O. 1972. Fur- ther records of the pitted-shelled turtle (Caretto- chclys insculpta) from Australia. Trans. Roy. Soc. South Australia 96:115-117. Sclater, J. G., Von Der Borch, C, Veevers, J. J., Hekinian, R., Thompson, R. W., Pimm, A. C., McGowran, B., Gartner, S., Johnson, D. A. 1974. Regional synthesis of the deep sea drilling results from leg 22 in the eastern Indian Ocean, pp. 815-831 in Von Der Borch, C, Sclater, J. G., Veevers, J. J. Initial reports of the Deep Sea Drilling Project 22, 890 p. Smith, M. J. 1975. The vertebrae of four Australian elapid snakes. Trans. Roy. Soc. South Australia 99:71-84. Smith, M. J. 1976. Small fossil vertebrates from Victoria Cave, Naracoorte, South Australia. IV Reptiles. Ibid. 100:39-51. Stephenson, E. M., Stephenson, N. G. 1970. Kar- yotypes of two Australian hylid frogs. Chromo- soma 30:38-50. Stirton, R. A., Tedford, R. H., Miller, A. H. 1961. Cenozoic stratigraphy and vertebrate paleontology of the Tirari Desert, South Australia. Geogr. Rev. 62:40-70. Storr, G. M. 1964. Some aspects of the geography of Australian reptiles. Senckenbergiana Biol. 45: 577-589. Strauchan, I. R. 1969. Hyla inermis (Peters), a species hitherto erroneously referred to the lepto- dactylid genus Cyclorana (Anura, Hylidae/Lep- todactylidae). Zool. Meded. Rijks Mus. Nat. Hist. Leiden 43:207-212. Straughan, I. R., Lee, A. K. 1966. A new genus and species of leptodactylid frog from Queens- land. Proc. Roy. Soc. Queensland 77:6.3-66. Taylor, L. W. H. 1975. Deposition and tectonic- patterns in the western Coral Sea. Bull. Australian Soc. Explor. Geophys. 6:33-35. Tilbury, L. A. 1975. Lineations in the Bismarck Sea. Ibid. 6:72. Twidale, C. R. 1972. Evolution of sand dunes in the Simpson Desert, Central Australia. Trans. Instit. British Geogr. (56):77-109. Tyler, M. J. 1967. Microhylid frogs of New Britain. Trans. Roy. Soc. South Australia 91:187-190. Tyler, M. J. 1970. Patterns of distribution and the origins of the Papuan hylid frog fauna. Search 1:246-247. Tyler, M. J. 1971. The phylogenetic significance of vocal sac structure in hylid frogs. Univ. Kansas Mus. Nat. Hist. Occas. Pap. ( 19) :319-360. Tyler, M. ]. 1972a. Superficial mandibular muscu- lature, vocal sacs and the phylogeny of Australo- Papuan leptodactylid frogs. Rec. South Australian Mus. 16:1-20. Tyler, M. J. 1972b. A new genus for the Australian leptodactylid frog Crinia darlingtoni. Zool. Me- ded. Rijks. Mus. Nat. Hist. Leiden 47:193-201. 106 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 Tyler, M. J. 1974a. Superficial mandibular mus- culature and vocal sac structure in the Anura. M.Sc. Thesis, Univ. Adelaide, 134 p. Tyler, M. J. 1974b. First frog fossils from Aus- tralia. Nature 248 (5450) :711-712. Tyler, M. J. 1974c. The systematic position and geographic distribution of the Australian frog Chirolcptes alboguttatus Gunther. Proc. Roy. Soc. Queensland 85:27-32. Tyler, M. J. 1976a. Frogs. William Collins, Syd- ney, 236 p. Tyler, M. J. 1976b. Comparative osteology of the pelvic girdle of Australian frogs and description of a new fossil genus. Trans. Roy. Soc. South Australia 101:3-14. Tyler, M. J. 1976c. A new genus and two new spe- cies of leptodactylid frogs from Western Aus- tralia. Rec. Western Australian Mus. 4:45-52. Tyler, M. J. 1977. Pleistocene frogs from caves at Naracoorte, South Australia. Trans. Roy. Soc. South Australia 101:85-89. Tyler, M. J. 1979. The introduction and current distribution in the New Hebrides of the Aus- tralian hylid frog Litoria aurea. Copeia. (In press ) Tyler, M. J., Davies, M. 1978a. Species groups within the Australopapuan hylid frog genus Li- toria Tschudi. Australian J. Zool. Suppl. 63:1—47. Tyler, M. J., Davies, M. 1978b. Phylogenetic rela- tionships of Australian hyline and Neotropical phyllomedusine frogs of the family Hylidae. Her- petologica 34:219-224. Tyler, M. J., Davies, M., King, M. 1978. The Aus- tralian frog Chiroleptes dahlii Boulenger: Its sys- tematic position, morphology, chromosomes and distribution. Trans. Roy. Soc. South Australia 102:17-23. Tyler, M. J., Martin, A. A. 1975. Australian lepto- dactylid frogs of the Cyclorana australis complex. Ibid. 99:93-99. Tyler, M. J., Martin, A. A. 1977. Taxonomic studies of some Australian leptodactylid frogs of the genus Cyclorana Steindacfmer. Rec. South Australian Mus. 17:261-276. Tyler, M. J., Rorerts, J. D. 1973. Noteworthy range extensions for some South Australian frogs. South Australian Nat. 99:143-147. Tyler, M. J., Martin, A. A., Davies, M. 1979. Biol- ogy and systematics of a new limnodynastine genus (Anura: Leptodactylidae) from northwestern Australia. Australian J. Zool. 27 ( 1 ) : 135-150. Underwood, G. L. 1954. On the classification and evolution of geckos. Proc. Zool. Soc. London 124: 469-492. Underwood, G. L. 1957. On lizards of the family Pygopodidae. A contribution to the morphology and phvlogeny of the Squamata. J. Morphol. 100:207-268. Veevers, J. J., McElhinny, M. W. 1976. The sep- aration of Australia from other continents. Earth- Sci. Rev. 12:139-159. Walker, C. F. 1973. A new genus and species of microhylid frog from Ecuador. Univ. Kansas Mus. Nat. Hist. Occas. Pap. (20): 1-7. Walker, C. F., Duellman, W. E. 1974. Descrip- tion of a new species of microhylid frog, Cliiasmo- cZeis, from Ecuador. Ibid. (26): 1-6. Wallace, D. C, Maxson, L. R., Wilson, A. C. 1971. Albumin evolution in frogs: A test of the evolutionary clock hypothesis. Proc. Natl. Acad. Sci. USA 68:3127-3129. Warren, J. W. 1969. Chelid turtles from the mid- tertiary of Tasmania. J. Palentol. 43:179-182. Warren, J. W. 1975. A fossil chelonian of probably Lower Cretaceous age from Victoria, Australia. Mem. Natl. Mus. Victoria 29:23-28. Watson, G. F., Martin, A. A. 1973. Life history, larval morphology and relationships of Australian leptodactylid frogs. Trans. Roy. Soc. South Aus- tralia 97:33-45. Wood, R., Patterson, B. 1973. A fossil trionychid turtle from South America. Breviora 415:1-10. Woodruff, D. S. 1972. Australian anuran chromo- some numbers. Herpetol. Rev. 4:208. Wopfner, H., Twidale, C. R. 1967. Geomorpho- logical history of the Lake Eyre Basin, pp. 119- 143 in Jennings, J. N., Mabrutt, J. A. (eds.). Landform studies from Australia and New Guinea. Australian National Univ. Press, Canberra, 434 p. Zweifel, R. G. 1969. Frogs of the genus Platyman- tis ( Ranidae ) in New Guinea, with the descrip- tion of a new species. Amer. Mus. Novit. (2374): 1-19. Zweifel, R. G. 1971. Results of the Archbold Ex- peditions. No. 96. Relationships and distribution of Genyophryne tliomsoni, a microhylid frog of New Guinea. Ibid. (2469): 1-13. Zweifel, R. G. 1972. Results of the Archbold Ex- peditions. No. 97. A revision of the frogs of the subfamily Asterophryinae Familv Microhvlidae. Bull. Amer. Mus. Nat. Hist. 148:411-546. 5. Quaternary Biogeography of Tropical Lowland South America Jiirgen Haffer Tommesweg # 60 4300 Essen-1 West Germany Research into the Quaternary biogeog- raphy of the Neotropical Region has been intensified during recent years as biologists became increasingly aware of the fact that Pleistocene climatic-vegetational fluctuations caused vast changes in the distribution of forest and nonforest biotas. Comparatively restricted populations of previously widely distributed plants and animals were isolated in remnant habitats during adverse climatic periods and differentiated at a varying rate depending upon the size of the restricted population (i.e., the size of the "refuge" area), the degree of isolation, and the varying "plasticity" of systematic groups following the model of geographic speciation (Mayr, 1942, 1963). The interpretation of Quaternary for- est and savanna fragmentation provides biol- ogists with a mechanism to explain extensive recent speciation in the South American low- lands, the occurrence of widely disjunct popu- lations of related taxa and other biogeo- graphical phenomena that could not be ac- counted for in the absence of natural barriers to interbreeding and dispersal (Meggers, 1977). The studies recently completed in the fields of Neotropical ornithology, herpetology, entomology, and botany yield comparable re- sults and, hopefully, will stimulate further investigations needed to test the model of Quaternary differentiation proposed. Prob- ably only few Neotropical plants and animals have survived from the Late Tertiary until the present time without evolutionary change — a notion popular among biologists only one or two decades ago. It is held that speciation events leading to extensive differentiation of faunas and floras during the Tertiary con- tinued into the Quaternary, possibly at a somewhat accelerated pace because of rapid environmental changes. In this review I characterize, in a brief introductory section, the climate and vegeta- tion of tropical lowland South America and their Quaternary history as far as it is known today. The main portion of this paper is a detailed discussion of recent research into the Quaternary biogeography of various groups of South American animals and plants. The re- sults of these studies concern biologists in- terested in the historical aspects of tropical biotas and should prove useful for compara- tive purposes to students of the Neotropical herpetofauna in particular. CLIMATE AND VEGETATION The climate of the lowlands of tropical South America varies from wet, humid, or moist, especially near mountain ranges and in the equatorial Amazon region, to dry or even arid in northeastern Brasil and along the Car- ibbean coast of northern Venezuela and Co- lombia. The northern tradewind belt moves southward into northern South America dur- ing the northern winter causing a pronounced dry season. Alternating wet and dry seasons occur over most of tropical South America, even to some extent in portions of the lower Amazon Valley. The dry seasons are least pronounced and the annual rainfall corre- spondingly high in the upper Amazon Valley, near the Atlantic coast of northeastern and southeastern tropical South America as well as in the vicinity of the Andes, especially in the Pacific Choco region of western Colombia, and along the Caribbean slope of the Middle American mountains. Tall evergreen forests grow in areas of high rainfall (Fig. 5:1) and grade through semi-evergreen and deciduous forests into 107 108 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 thorn forest and scrub as the annual rainfall decreases and the influence of a prolonged dry season increases. Variations are caused by local conditions of soil and topography. Characteristic plant associations of the non- forest regions of interior Brasil are the cer- rado, a typical woodland savanna, and the caatinga, an open thorn woodland rich in cacti. Extensive grass savannas are found in the north and south of the Amazon forest in areas where flooding lasts several months each year and alternates with a severe dry season (eastern Colombian and central Venezuelan llanos; the savannas of eastern Bolivia and the varzea campos of the lower Amazon Val- ley). Extensive discussions of the climate and vegetation of South America have been pub- lished by Schwerdtfeger (1976), Hueck (1966), and Hueck and Seibert (1972); see also the recent literature review by Haffer (1974). The vast Amazonian forest covers some 6,000,000 km 2 of central South American low- lands from the Andes to the Atlantic coast including the upper Orinoco region of south- ern Venezuela and the Guianas to the north- east of the Solimoes-Amazon basin itself. Riv- ers bordered by characteristic vegetation zones of varying width and isolated savanna en- claves interrupt the immense and superficially uniform forests. Interspersed savannas are concentrated in a transverse zone of reduced annual precipitation extending from southern Venezuela across the lower Amazon River into northeastern Brasil; others occur between the upper Madeira and Purus rivers (Fig. 5:1). Extensive forests in Amazonia grow on infer- tile leached soils of the terra firme in areas where Tertiary strata and "basement" rocks of the Guianan and Brasilian shields form the subsoil. Only rather small areas of Amazonia are underlain by fertile soil, especially along major river valleys and in the Andean fore- land. Erosive material from the Andes is transported eastward into the Amazonian low- lands and forms soils that are considerably richer in nutrients than the soils of the adja- cent terra firme between large river courses (Fittkau, 1969, 1974). This simplified scheme is currently being modified through detailed interpretation of radar images, field controls, and extensive mapping in Amazonia (Ham- mond, 1977). CLIMATIC-VEGETATIONAL HISTORY OF THE NEOTROPICAL LOWLAND REGION DURING THE QUATERNARY Humid tropical vegetation, perhaps some- what drier in midlatitudes, covered most of the exposed land area of South America dur- ing early Tertiary time (Wolfe, 1971; Solbrig, 1976). Forests slowly retreated northward in the Patagonian region during the second half of the Tertiary when the Andes were gradu- ally uplifted and the climate became cooler and drier, possibly in response to periodic polar glaciations which began during the Miocene. The glacial phases gained momen- tum with time until, during the Quaternary, vast polar glaciers repeatedly advanced to- ward lower latitudes and extensive montane glaciers covered the higher slopes of tropical mountains. Extensive continental shelf areas were emergent during the glacial periods of lowered world sea level and were submerged during interglacial phases. Interglacial seas even encroached over low lying coastal plains, such as northern Colombia, and covered a huge portion of the Amazon Valley. Although temperatures in the tropical lowlands remained "tropical" during glacial periods («3°C lower than today), alternating humid and arid climatic phases of the Quater- nary caused vast changes in the distribution of forest and nonforest vegetation. Forests broke into isolated remnants during cool dry periods (glacial phases) and expanded and coalesced during warm, humid phases (inter- glacial periods). Conversely, nonforest vege- tation expanded during glacials and retreated during interglacial phases. Geoscience data are insufficient so far to map the changing distribution of forest and nonforest vegetation during the various climatic periods and, in particular, to locate accurately areas of rem- nant forests during arid phases which served as "refugia" for animal populations. From the location of current rainfall maxima and the topographic relief (which was already in existence during most of the Pleistocene) one would tentatively conclude that several areas along the northern slopes and foreland of the mountains in the interior Guianas, along the eastern slopes and foreland of the Andes, as well as along the northern margin of the Brasilian tableland remained humid and for- 1979 HAFFER: QUATERNARY OF TROPICAL LOWLANDS 109 Fig. 5:1. Distribution of humid tropical lowland forest and location of rainfall centers in Middle and South America. Explanations: Shaded = humid forest, often semideciduous around savanna regions. Hatched vertically = areas receiving over 2500 mm of rain per year. Solid = Andean Cordilleras and Middle American mountains of more than 2000 m elevation. Heavy dashed lines delimit the dry transverse zone of lower Ama- zonia characterized by numerous isolated savanna enclaves. Letters designate areas of paleoecological research. (See text for details.) Distribution de selva humeda y position de centros de Uuvia en las tierras bajas tropicales de Centro y Sud America. Explicaciones: Matizado = selva humeda; frecuentemente selva semidecidua alrededor de regiones de savanas. Rayado vertical = areas recibiendo mas que 2500 mm de Uuvia anuales. Negro = Cordilleras andinas y montafias de Centro America con alturas mayores que 2000 m. Lineas rayadas anchas delimitan la zona seca transversal de Amazonia baja caracterizada por numerosas savanas aisladas. Las letras indican las areas de investigaciones paleoecologicas. (Ver el texto para detalles.) 110 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 Copoera + Katiro (Rondonia. Brasil) I0 2O3O4O5O6O7O809O KDO% Loguxi de Agua Suoa (Llanos Onerrfoles .Gotomtoo ) Loke Moreru iRupsKn. Guyana) 21601 23401 ■ i^ Eem of tie wt tropical foresf .'. , [ Maunna I 1 Savarmo elem (pnnc Gtamineae) 50 100%^° 38301 4IK)| 0-H00 j Oftier forest elements Gramineae (savannas) □ B G/amneae , Cyperoceoe orel crtTier 5 / / • x y v 4. ^ 400 KM Fie. 5:5. Reconstruction of Quaternary refugia in tat (dashed outline). Schematic representation. Explan parapatric species and hybridizing subspecies. Contact suture zones. 2. Mapping of distribution centers for ( contours indicate numbers of sympatric species and of location of suture zones and distribution centers. 4. ing into consideration all available data on relief, dim under consideration. Reconstruction de refugios cualernarios en medio de savana; delimitada por un margen raijado). Explication especies parapdtricas ij subespecies con hibridacion. Zo ureas formando asi zonas de sutura. 2. Mupeo de cen subespecies localizadas (Hncas de contorno indican el grupos de organismos bajo consideration). 3. Compara de distribution. 4. Mapeo de refugios rclacionados a datos disponibles sobre el relieve, clima, geomorfologia, an extensive and fairly uniform forest or nonforest habi- ations: 1. Mapping of secondary contact zones of allied zones often cluster in certain regions forming faunal med by fairly localized species and subspecies clusters subspecies in the groups considered). 3. Comparison Mapping of refugia related to the dispersal centers tak- ate, geomorphology and palynology related to the areas tin extenso ambiente y mas o menos uniforme (sclva o es: 1. Mapeo de zonas de contacto secundaria entre nav de contacto frccucntcmente se encuentran en ciertas tros de distribution formadas por grupos de especies y niimero de especies y subespecies simpdtricas en los cion de la position de zonas de sutura y de los centros los centros de dispersion teniendo en cucnta todos los y palinologia de la region. 120 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 contact zones and distribution centers. — Often clusters of secondary contact zones fall be- tween core areas, thereby supporting the in- terpretation that the distribution centers func- tioned as centers of dispersal in the past. This interpretation is particularly applicable in the case of hybridizing subspecies and parapatric species that characterize neighboring centers and meet in the intervening area where other contact zones of more wide ranging forms are clustered as well. Recause the contact zones and centers often involve different species, full complementarity between them cannot be expected. Thus, there are many contact zones between forms whose present ranges comprise more than one distribution center. This is schematically indicated in the diagram (Fig. 5:5, no. 3), where several forms occupying the entire eastern portion of the habitat with a northern and a southern center meet western forms along contact zones that are not re- stricted to the area between the centers. If biogeographical analyses are based ex- clusively on the population structure of sev- eral widespread and geographically variable species, a patchwork of subspecies ranges sep- arated by more or less extensive hybrid zones exists. Superimposition of the various species maps may show that the ranges of pure sub- specific populations with uniform character expression and the location of separating hy- brid belts more or less coincide in the differ- ent species (Fig. 5:6). Contour lines illustrate varying hybrid levels of populations around the centers. Analyses also may be based on the population structure of a single species, and distribution maps of several individual characters may be prepared (e.g., Vanzolini and Williams, 1970, for the lizard Anolis chrij- solepis). The character maps may then be superimposed and contoured in a similar manner as those of subspecies. Coinciding areas of uniform characters with low variabil- ity often cluster in core areas that are sep- arated by zones of high character variability coinciding with hybrid belts between sub- species ranges. Step 4: Mapping of the refngia related to the dispersal centers. — Ideally, this can be accomplished using geoscience data exclu- sively. Having established on the basis of zoogeographical data that a given distribution 300 KM Fig. 5:6. Schematic representation of centers of subspecies endemism (stippled) in an extensive and fairly uniform habitat (dashed outline). Explana- tions: Superimposed ranges of "pure" subspecies populations and separating hybrid belts in different species often cluster in certain areas. Number of dif- ferentiated forms of various species superimposed is indicated (n = 11 or 12). Average hybrid level of combined species populations is mapped by contour lines ( = "pure" ) . A barrier to gene-flow ( broad river or mountain range) is schematically shown as a black bar; the populations on either side of the barrier are mostly pure, as limited gene-flow takes place between subspecies of only a few species whose ranges have been superimposed. Diagrama esquemdtico de centros de endemismo subespecifico (puntcado) dentro de tin habitat rela- tivamente extenso y uniforme (delimitado por una linea entrecortada). Explicaciones: Rangos supcr- impuestos de poblaciones de subespecies "puras" y los cinturones de hibridizacion en varias especies muchas veces se juntan en determinadas rcgiones. El numero de jormas difcrenciadas de las especies superpuestas estd indicado (n = 11 6 12). El nivel promedio de hibridizacion en las poblaciones de especies combinadas estd mapcado por tineas de con- lomo (0 = "puro"). Una barrera para el flujo genetico (un gran rio 6 una montana) estd indicada esquematicamente por una faja ncgra. ha matjoria de las poblaciones en ambos lados de la barrera son "puras"; un flujo genetico restringuido ocurrc sola- mente entre unas pocas poblaciones separadas. center probably was a center of dispersal in the past, we proceed to postulate the approxi- mate location, size and shape of the corre- sponding refuge area of forest or nonforest vegetation within the central portion of the center, taking into consideration all available geoscience data from the region, however de- ficient these data may be. Medium elevation areas with high rainfall near the base of mountains or plateau regions are prime candi- dates for forest refugia whose postulated ex- tent during the maximum of the arid climatic 1979 HAFFER: QUATERNARY OF TROPICAL LOWLANDS 121 period remains highly speculative. The model assumes that the species and subspecies char- acterizing a given center were confined to the postulated refuge prior to dispersal and prior to establishing secondary contact with other forms spreading from distant centers. Obvi- ously, the former existence and changing size of the refugia ultimately can be traced only through detailed palynological, pedological, and geomorphological studies rather than through zoogeographical analyses. After all, the refugia are geological-paleoclimatological, not biological, phenomena. However, as long as only scattered geoscience data are avail- able, zoogeographical analyses as outlined above will help our understanding of biotic differentiation in lowland tropical South America during the Quaternary. In general, the locations of distribution centers of the Amazonian forest biota correlate well with the locations of forest refugia tentatively de- rived from rainfall, relief and geoscience data alone (p. 10S), thus strengthening an histor- ical interpretation of the biogeographical core areas (Simpson and Haffer, 1978; Brown and Ab'Saber, 1979). Endler (1977) suggested that parapatric speciation occurs frequently in nature and that the hybrid zones in Amazonia might ac- tually be zones of primary intergradation caused by strong environmental gradients. However, the geoscience data reviewed above (but not discussed by Endler) favor the inter- pretation of allopatric rather than parapatric speciation for the many hybridizing and non- hybridizing populations of plants and animals in contact. The Amazonian forest refugia are geological-paleoclimatical, not biological phenomena. Biogeographical Studies on the Forest Fauna and Flora Some of the major regional conclusions re- garding the Quaternary differentiation and dispersal of Neotropical biotas are summar- ized, as follows: 1) Extensive speciation took place in many groups of Neotropical animals and plants that were repeatedly isolated in refugia and later expanded their ranges dur- ing favorable expansive phases. 2) During humid climatic periods, extensive forests probably permitted a direct connection of the Amazonian fauna and flora with those of the Atlantic forest in eastern Brasil across the central Brasilian Plateau and along the coastal region of northeastern Brasil, where scattered forests are still preserved on isolated moun- tains (Muller, 1973; Haffer, 1974; Brown, 1976, 1977b). 3) The cis- Andean and trans- Andean forest biotas probably were connected repeatedly both north of the Andes in the Caribbean lowlands of northern Colombia and through the north Peruvian Andes from the upper Maranon Valley via the low Por- culla Pass so as to reach the forested Pacific lowlands of Colombia from the south (Chap- man, 1926; Haffer, 1967, 1975; Muller, 1973). 4) The relations between the Middle and South American nonforest bird faunas are less pronounced than those of the forest faunas of these two areas (Mayr, 1964). 5) There are numerous conspecific popula- tions inhabiting the nonforest regions of northern South America and central Brasil, respectively, being separated by the entire width of the Amazon forest. Because their dispersal under present climatic-vegetational conditions is highly unlikely, a rather recent direct communication of nonforest faunas across Amazonia during one or more arid cli- matic periods seems probable. This theory also explains the close relationship of the fauna and flora of the Amazonian savanna enclaves with those of the nonforest regions to the north and south of Amazonia ( Hueck, 1966; Haffer, 1969; Muller, 1973). The authors of the studies reviewed below are aware of the tentative nature of the results obtained and of the suggestions made. More- over, the nascent theory of ecological refugia during the Quaternary, like any theory, cannot be proven but only disproven by the results of additional studies on the paleo- climatology, vegetational history, and zooge- ography of South America. I wish to empha- size with Meggers (1977) that the efforts by biologists and archeologists at present are not more than a search for correlations and pat- terns useful as guides for investigation. The strikingly similar results as to the basic pat- terns of differentiation in various unrelated groups of Neotropical animals tend to sup- port the refuge theory and justify continued 122 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 research into this field of enquiry. Biogeog- raphers working on the Neotropical fauna proposed for the lowlands of Middle and South America a total of 40 areas that are assumed to have served as refugia for the forest fauna at various times during the Qua- ternary. These areas are listed and briefly described in Appendix 5:1. Birds. — Numerous secondary contact zones of avian species and subspecies pairs are clustered in north-central Amazonia — in southern Venezuela and in the Rio Negro region of northern Brasil (Fig. 5:7), where Guianan forms from the east established con- tact with western forms that had spread from upper Amazonia (Haffer, 1969, 1974). These contact zones represent range limits either of hybridizing subspecies or of non-hybridizing competing species; the ranges were mapped in conjunction with studies of patterns of geographic variation in several groups of Amazon forest birds. South of the Rio Ama- zonas, contact zones are scattered over a more extensive area, where we may distin- guish an upper Amazonian and a south-cen- tral Amazonian suture zone. A distributional analysis of the Amazonian forest avifauna (Haffer, 1978) indicated the existence of six core areas of distribution (Fig. 5:7) or local- ized species clusters, each of them composed of 10 to 50 species. The six clusters together are characterized by a total of around 150 species or about 25 percent of the Amazon forest bird fauna. Most of the remaining forest birds (75%) are more widely distrib- uted, their ranges comprising two or more distribution centers. There are conspicuous clusters of lower and upper Amazon forest birds, as well as smaller groups of northern and southern Amazon forest birds, in addition to those species that inhabit even larger areas of Amazonia and beyond (see Haffer, 1978, for more details). A number of other species have an irregular, spotty distribution or are known from only a single locality. The mapped contact zones in Amazonia (Fig. 5:7) are mostly located between distributional core areas which probably functioned as centers of dispersal. The complementarity of contact zones and distribution centers is less well de- veloped in central Amazonia, where a num- ber of birds spreading from the species-rich Fig. 5:7. Location of secondary contact zones (above) and of distribution centers in the Amazon forest avifauna (below). Explanations: Secondary contact zones (dashed) are concentrated in north- central Amazonia (a), upper Amazonia (b), and south-central Amazonia (c). A continuous line de- limits the Amazon forest (above). Fairly localized species clusters (each composed of 10 to 50 species) form the distribution centers A to F (below). Adapted from Haffer, 1974, Fig. 9.13 and Haffer, 1978. Additional clusters of endemic species char- acterize the Atlantic forests of southeastern Brasil and the trans-Andean forests of northwestern South America. Position de zonas de contacto secundario (arriba) y de centras de distribution en la avifauna de la sclva amazonica (abajo). Explicaciones: Zonas de contacto secundario (ttneas rayadas) se encucntran en la Ama- zonia norcentral (a), en la alta Amazonia (b) y en la Amazonia surccntral (c). Una linea continua de- limita la sclva amazonica (arriba). Grupos de especies hastante localizadas forman los centros de distribu- tion (A-F; abajo) cada uno de cllos formado pot It) a 50 especies. Adaptado de Haffer, 1974, Fig. 9.13 y Haffer, 1978. Otros grupos de especies en- demicas caracterizan la selva atldntica de Brasil meri- dional y la sclva transandina de Sudamerica norocci- dental. upper and lower Amazonian centers are in contact within the area of the "weak" Imeri and Rondonia centers (characterized by com- paratively few species). 1979 HAFFER: QUATERNARY OF TROPICAL LOWLANDS 123 Based on geosciencc data, Quaternary for- est refugia in Amazonia probably were lo- cated near the windward base of hilly ranges and mountains, such as the northern margin of the Brasilian Plateau, the eastern base of the Andes, and the northern base of the moun- tains in the interior Guianas. These areas correlate well with the location of the zoo- geographically defined distribution centers which, therefore, are assumed to indicate the approximate location of Quaternary forest refugia. Additional refugia for the avifauna (Fig. 5:8) based on similar geological and zoogeographical criteria have been proposed for the Atlantic forest of southeastern Brasil (Miiller, 1973; Jackson, 1978) and for the trans-Andean forest region in western Colom- bia and in Middle America (Haffer, 1967, 1969, 1974, 1975). A recent study of a com- plex Neotropical avian genus on the basis of the refuge concept is Fitzpatrick's ( 1976 ) analysis of the Todirostrum flycatcher group; also see the recent review by Dorst ( 1976 ) . Lizards and frogs. — Analysing the popula- tion structure and character variation in the Amazonian Anolis chrysolepis group, Vanzo- lini and Williams (1970) and Vanzolini (1970, 1973) arrived at conclusions regarding the history of faunal differentiation during the Quaternary that are strikingly similar to those reached by Haffer (1969, 1974). Sev- eral extensive areas of uniform character ex- pression (core areas) in Anolis chrysolepis are separated by regions where complex char- acter variation suggests hybridization or in- trogression along zones of secondary contact. Vanzolini and Williams (1970) interpreted this situation as being the result of secondary intergradation of populations that had differ- entiated in geographic isolation, and they as- sumed changes in the distribution of forest in Amazonia during the course of climatic fluctu- ations. These authors reconstructed several forest refugia around the periphery of Ama- zonia (Fig. 5:8), most of which coincide closely with those proposed by Haffer ( 1969, 1974) for birds. Additional forest refugia probably determined the differentiation of Neotropical forest reptiles and will be iden- tified when other species are studied in detail. Miiller (1972, 1973) using mostly herpetolog- ical and ornithological data described several broad centers of dispersal but did not sug- fest specific refugia. A comparison shows that the boundaries of Miiller's centers rarely co- incide with those of avifaunal core areas but fall in the peripheral gradient of decreasing species numbers or coincide with river bar- riers. Miiller (1973) did not state the criteria he used to delimit the various broad centers, which on his maps are frequently separated only by narrow corridors. In recent years, several herpetologists ac- cepted the notion of a Pleistocene origin of numerous species and subspecies of Neo- tropical frogs, lizards and snakes in forest refugia: Duellman (1972, 1978), Heyer (1973), Duellman and Crump (1974), and Silverstone (1975, 1976) for certain Amazon- ian and trans-Andean frogs, Gallardo ( 1965, 1972 ) and Lynch ( this volume ) for the South American amphibian fauna generally, Dixon (this volume) for Amazonian reptiles, Ech- ternacht (1973) for Middle American lizards (Ameiva), Hoogmoed (1973, this volume) for the herpetofauna of the Guianas, Jackson (1978) for two genera of eastern Brasilian iguanid lizards, and C. W. Myers (1973, 1974) for two genera of snakes. Dixon (this volume) and Lynch (this volume) mapped clusters of endemic reptile and amphibian species in upper and lower Amazonia. These centers of distribution more or less coincide with similar centers established for the forest avifauna. At this stage, a more quantitative treatment of available distributional data would be desirable, such as the derivation of herpetological distribution centers by means of contoured diversity maps of localized spe- cies clusters and a comparison of the location of secondary contact zones with that of dis- persal centers and postulated refugia. Insects. — The distribution and population structure of two groups of Neotropical insects are relatively well known and permit initial biogeographic analyses — 1) Certain Dro- sophila flies that have been collected and analysed extensively in the course of genetic investigations, and 2) butterflies of the genus Heliconius. Spassky et al. (1971) summarized the distribution of several semispecies and closely related species of Drosophila. These authors and Winge (1973) concluded that the various forms may have originated in for- 124 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 7 Fig. 5:8. Location of presumed Quaternary forest refugia (hatched) in tropical lowland South America. Explanations: 1. Reconstruction based on the distribution patterns of Neotropical birds ( Haffer, 1967, 1969, 1974). 2. Reconstruction based on the population structure of Amazonian lizards (Anolis chnjsolcpis group, Vanzolini and Williams, 1970, Vanzolini, 1970; wide hatches indicate core areas). 3. Reconstruction based on the distribution patterns of subspecies and species of Heliconius butterflies ( Brown et al., 1974, Brown 1977a,b). 4. Reconstruction based on the distribution patterns of four families of Amazonian trees (Prance, 1973). Position de presuntos refugios cuaternarios de selva humeda (rayado) en las Metros hajas tropicales de Sudamerica. Explicaciones: 1. Reconstruction basada en los patrones de distribution de aves neotropicales (Haffer, 1967, 1969, 1974). 2. Reconstruction basada en la estructura de poblaciones de lagaitos amazonicos (grupo de Anolis chrysolepis, Vanzolini y Williams, 1970, Vanzolini, 1970; Hneas verticales indican areas nu- clearcs). 3. Reconstruction basada en la distribution de subespicies y espeties