HAWAIIAN BIOGEOGRAPHY Smithsonian Series Comparative Evolutionary Biology in V. A. Funk, Smithsonian Institution Peter F. The Cannell, Smithsonian Institution intent of this series is to publish innovative studies in the field of comparative evolutionary biology, especially by authors willing to introduce Within expand views now accepted. context, and with some preference toward the organismic new this ideas or to challenge or a diversity of viewpoints is level, sought. Advisory Board Richard Bateman Royal Botanic Garden, Edinburgh Daniel R. Brooks University of Toronto William DiMichele Smithsonian Institution Michael Donoghue Harvard University Douglas Erwin Smithsonian Institution David Grimaldi American David University of Texas at Austin Hillis Richard Mooi Robert Voss California American Museum of Natural History Academy Museum of Sciences of Natural History ALSO IN THE SERIES Parascript and the Language of Evolution Daniel R. Brooks and Deborah A. McLennan Parasites The Development and Evolution of Butterfly Wing H. Erederick Nijhout Patterns QH 198 H3H38' ]995X HAWAIIAN BIOGEOGRAPHY EVOLUTION ON A HOT SPOT ARCHIPELAGO EDITED BY WARREN L. WAGNER AND V. A. EUNK SMITHSONIAN INSTITUTION PRESS WASHINGTON AND LONDON © 1995 by the Smithsonian Institution All rights reserved Robynn K. Shannon Technical preparatory editing by Copyediting and typesetting by Peter Strupp/Princeton Editorial Associates Cover art by Alice Tangerini and Robynn K. Shannon Book design by Linda McKnight Proofreading by Eileen D’ Araujo Project management by Deborah L. Sanders Library of Congress Cataloging-in-Publication Data Hawaiian biogeography evolution on a hot spot archipelago / edited by Warren L. Wagner and V. A. Funk cm. (Smithsonian series in comparative evolutionary biology) p. Papers grew out of a symposium cosponsored by the American Society of Plant Taxonomists and the Association for Tropical Biology in 1 992. Includes bibliographical references (p. and index. : — ) ISBN 1-56098-462-7 (clothbound).—ISBN 1-56098-463-5 (paperbound) 1. —Hawaii. Biogeography Lambert. —Hawaii. Island ecology 2. Funk, V. A. (Vicki 11. A.), 1947- . III. 1. Wagner, Warren Series. QH198.H3H38 1995 574.9969— dc20 94-5353 British Library Cataloguing-in-Publication Throughout the text, places of deposit for plant herbarium abbreviations as given This material is Data available in Holmgren voucher specimens are indicated by et al. (1990). based upon work supported by the National Science Foundation under award no. DEB-9214264 to the American Society of Plant Taxonomists. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of the National Science Foundation. Manufactured in the United States of America 54321 01 00 99 98 97 96 95 © The paper used in this publication meets the minimum requirements of the American National Standard for Permanence of Paper for Printed Library Materials Z39.48-1984. For permission to reproduce illustrations appearing in this book, please correspond directly with the owners of the works (the authors of the chapter, unless otherwise indicated). The Smithsonian illustrations individually or Institution Press does not retain reproduction rights for these maintain a file of addresses for photo sources. Contents ix xiii XV Preface Acknowledgments Contributors 1 1 Introduction SHERWIN CARLQUIST 2 14 Geology and Biogeography of the Hawaiian Islands HAMPTON L. CARSON AND DAVID A. CLAGUE 3 30 Methods FUNK Cladistic V. A. 4 39 Biogeographic Patterns of Two Independent Hawaiian Cricket Radiations (Laupala and Prognathogryllus) KERRY L. SHAW 5 57 Chromosomes and Male Genitalia of Hawaiian Drosophila: Tools for Interpreting Phylogeny and Geography KENNETH Y. KANESHIRO, ROSEMARY AND HAMPTON L, CARSON G. GILLESPIE, VI CONTENTS 6 72 Molecular Approaches to Biogeographic Analysis of Hawaiian Drosophilidae ROB DESALLE 7 90 Evolution of Savona (Heteroptera, Miridae): Speciation on Geographic and Ecological Islands ADAM ASQUITH 8 121 Comparison of Speciation Mechanisms in Web-Building and Non-Web-Building Groups within a Lineage of Spiders ROSEMARY G. GILLESPIE AND HENRIETTA B. GROOM 9 147 Evolutionary Relationships of the Hawaiian Honey creepers (Aves, Drepanidinae) CHERYL L. TARR AND ROBERT C. FLEISCHER 10 160 Biogeography of Seven Ancient Hawaiian Plant Lineages V. A. FUNK AND WARREN L. WAGNER 11 195 Phylogeny, Adaptive Radiation, and Biogeography of Hawaiian Tetramolopium (Asteraceae, Astereae) TIMOTHY K. LOWREY 12 221 Phylogeny and Biogeography in Schiedea and Alsinidendron (Caryophyllaceae) WARREN L. WAGNER, STEPHEN ANN K. SAKAI G. WELLER, AND CONTENTS 13 259 Historical Biogeography and Ecology of the Hawaiian Silversword Alliance (Asteraceae): New Molecular Phylogenetic Perspectives BRUCE BALDWIN AND ROBERT G. H. ROBICHAUX 14 288 Molecular Evolution, Adaptive Radiation, and Geographic Cyanea (Campanulaceae, Lobelioideae) THOMAS J. GIVNISH, KENNETH J. SYTSMA, JAMES F. SMITH, AND WILLIAM J. HAHN Speciation in 15 338 Patterns of Speciation and Biogeography in Clermontia (Campanulaceae, Lobelioideae) THOMAS G. LAMMERS 16 363 Phylogenetic Analysis of Hawaiian and Other Pacific Species of Scaevola (Goodeniaceae) ROBERT PATTERSON 17 379 Biogeographic Patterns in the Hawaiian Islands V. A. FUNK AND WARREN 420 Postscript 423 Literature Cited 453 Index L. WAGNER vii >« I' Preface Isolated oceanic islands have long lured the evolutionary biologist. More than a century ago, Alfred Russell Wallace and Charles Dar’win stressed how much could be learned about evolution by studying plants and animals on volcanic high islands. archipelago is Of all the oceanic islands, the Hawaiian often considered an unparalleled example of insular evolu- making this so. The Hawaiian chain is the most massive oceanic archipelago and has extensive ecological variation— from dry and mesic coastal environments to a wide array of inland habitats ranging from arid to the wettest on earth and ranging in elevation. Several factors contribute to tion from sea level to 4,200 m. Moreover, the archipelago represents the longest, apparently regular, continuous formation of islands in a linear chronology in the world because of an incessant hot spot under the earth’s mantle. Former high islands in the Hawaiian-Emperor Chain, once in a geographic position similar to today’s eight high islands of the Hawaiian chain, were formed during the Tertiary Period, first at least 70 million years ago. Perhaps the most important aspect of the Hawaiian Islands for evolutionary studies is At more than 3,500 km most secluded archipelago their striking isolation. from the nearest continental land mass, this has been colonized exclusively by waif elements. Also, repeated colonization by the same confined to the species is less likely. first arrivals Stochastic colonizations were not to the archipelago. formed to the southeast of the existing As each new islands, new island was opportunities for colonization were constantly presented. However, successful establish- ment on a new island may be more dependent on ability to colonize ecologically younger sites or a wider range of habitats not available on older source islands. All these features combine to give the Hawaiian archipelago an extraordinary terrestrial biota that includes approximately 700 fungi, 800 lichens, land snails, 260 mosses, 180 pteridophytes, 1,000 angiosperms, 1,000 230 terrestrial arthropods (excluding insects), 5,000 insects. IX PREFACE X 112 birds, 5 fresh-water fishes, ical array ranges from about to 99% for insects. and 2 mammals. Endemism 50% for mosses and 89% in this biolog- for angiosperms Many of the terrestrial groups of Hawaiian organisms are represented here, including insects (Chapters 4 to 7), spiders (Chapter 8), birds (Chapter 9), and flowering plants (Chapters 10 to 16). There however, omissions including lichens, pteridophytes, bryophytes, and are, terrestrial snails, some of which have not radiated and others that have no appropriate data. Hawaiian geology continues to develop fortunately geology texts usually at a rapid pace, and un- do not emphasize the fundamental geologic features pertinent to the biologist, especially the biogeographer. We are fortunate to be able to include a chapter that provides date summary an up-to- of the features relevant to island biogeography. This volume, which grew out of a symposium cosponsored by the American Society of Plant Taxonomists and the Association of Tropical Biology in 1992, was catalyzed by the auspicious coincidence of three developments. First, during the three decades since the original articula- tion of the hot spot theory of mid-Pacific archipelago formation, tremen- dous advances have been made in our understanding of the geologic processes involved in the formation and history of this conveyor-like archipelago. Second, convenient phylogenetic methods are available and used by most researchers. Finally, the past now widely decade has witnessed a considerable increase in the number of researchers investigating a wide array of the evolutionary radiations of Hawaiian terrestrial organisms. The chapters of this collaborative work represent the to test the idea that independently derived groups of exhibit similar patterns of colonization directly to the unique geologic history of participant applied phylogenetic first attempt Hawaiian organisms and differentiation that relate this oceanic archipelago. Each methods to morphological or molecular data to generate phylogenetic hypotheses, secondarily deriving biogeographic hypotheses. Rather than mere summaries of the participant’s research, these studies mostly present tributor has used a consistent new data and analyses. Each con- methodology to allow evolutionary pat- terns of different groups to be directly compared in our search for common and discordant patterns. Patterns generated in these studies have been further manipulated to test ideas about evolution, such as innova- tion in breeding systems, behaviors, or ecology in an insular environment. This volume represents the Hawaiian organisms other than first detailed biogeographic study of isolated exceptions such as the Hawaiian PREFACE Drosophila and S. Carlquisfs innovative work culminating in his XI 1974 book Island Biology. This collaboration has brought together a majority of the contemporary biological researchers on the terrestrial Hawaiian biota who have appropriate and sufficient data. Indeed, this may represent the first attempt to analyze a significant proportion of the plants and animals of any natural area using a formal, rigorous approach, such that the results can be compared across different taxonomic groups. By collecting and synthesizing data for the Hawaiian biota, we not only add new understanding of the biogeography of the archipelago but may further kindle new ideas toward an understanding of evolution on islands. Acknowledgments grew out of the compatible idea for this volume The coeditors. One interests of the of us (W.L.W.) has a deep interest in Hawaiian biogeogra- phy, generated through years of collaboration with Derral Herbst many others (V.A.F.) on the classification of Hawaiian flowering plants. and The other has long been concerned with the use of phylogenetic patterns to study biogeography and speciation. Discussions between the coeditors in 1990 concerning application of phylogenetic systematics to the study of island species radiations resulted in a paper presented at the 1991 Society for the Study of Evolution meetings in Hilo (developed into Chapter 10). We are grateful to many colleagues, especially Sherwin Carlquist, Hamp Carson, Neal Evenhuis, Chris Haufler, Derral Herbst, Scott Miller, Lynne Parent!, Ann Sakai, ment during and Steve Weller for this project. astic participation in the We thank all insights, advice, and encourage- the contributors for their enthusi- 1992 American Institute for Biological Sciences (AIBS) symposium in Honolulu and their willingness to contribute, often unpublished We new data, to this volume. (W.L.W and DEB-9214264) from the National Science Foundation (NSF) to support the symposium and publication subsidy. We are grateful to the American Society of Plant Taxonomists, especially M. Denton, H. Eshbaugh, and R. Jensen, for the society’s support of the symposium held in 1992, for providing an award to V.A.E. and W.L.W. to help defray the costs of the symposium, and for appreciate the grant administering the NSE Each chapter reviewers, one V.A.E. co-PIs; grant for this project. in this volume was peer-reviewed by was also reviewed by both editors technical editor. Because this project brought together so biological expertise number of chapters. We two from among the volume contributors and the other an outside review. Each chapter able at least on the Hawaiian much of the we depended on a consider- the contributors as reviewers of other contributors’ thank them received one or Islands, and the all. In addition to these reviews, each chapter more reviews from a specialist outside of the project xiii XIV ACKNOWLEDGMENTS We contributors. M. Braun, P. Cannell, G. B. Dalrymple, D. Futuyma, D. Grimaldi, T. Henry, D. R. Herbst, M. Lane, F. R. Rabelei; We F. appreciate their interest and hard work. They are M. Hershkovitz, G. Hormiga, Lutzoni, Sheldon, J. Hunt, C. Labandeira, Manos, D. Olmstead, S. Olson, Soltis, B. Stein, and A. R. Templeton. P. P. thank the Bernice Bishop P. Museum J. use for Pakaluk, of plant from the Manual of the Flowering Plants of HawaPi in Chapters 10 and 13 and for the photograph by Joseph Rock in Chapillustrations ter 14. We the photograph by We Garden also thank the National Tropical Botanical S. Carlquist in Chapter 14. Lucy Julian and Jim Nix for especially thank encouragement, and patience during the preparation of The Hawaiian for use of Islands, unique among geographic are a fascinating place to study evolutionary biology. studies here not only spur new their support, this book. on earth, hope that the regions We insights for science but that the new understanding of the nature of the endemic organisms in the Hawaiian Islands will help from the promote sale of this their conservation. In that light, book will be contributed to the Botanical Garden research program, which ploration, study, is any royalties National Tropical contributing much to ex- and conservation of the plants of the Hawaiian and other Pacific islands and therefore also the fauna that depend on them. This comes at an especially tion of Hurricane ‘Iniki. critical time as they recover from the devasta- Contributors ADAM ASQUITH & Wildlife Service, Pacific Islands Office, 3 Waterfront Plaza, U.S. Fish 500 Ala Moana Boulevard, BRUCE G. Honolulu, HI 96813 Suite 580, BALDWIN Jepson Herbarium and Department of Integrative Biology, University of California, Berkeley, Berkeley, CA 94720 SHERWIN CARLQUIST 4539 Via Huerto, Santa Barbara, HAMPTON L. CA 93110 CARSON Department of Genetics and Molecular Biology, University of Hawai‘i, 1960 East- West DAVID Road, Honolulu, HI 96822 A. CLAGUE U. S. Geological Survey, Hav\^aiian Volcano Observatory, Hawai‘i National Park, HI 96718 HENRIETTA B. GROOM Department of Biology, The University of the South, Sewanee, TN 37375 ROB DeSALLE Department of Entomology, American Park West at 79th ROBERT C. Street, Museum New York, NY of Natural History, Central 10024-5192 FLEISCHER Department of Zoological Research, National Zoological Park, Smithsonian Institution, Washington, V. A. MRC 551, DC 20008 FUNK Department of Botany, MRC 166, National Museum of Natural History, Smithsonian Institution, Washington, ROSEMARY G. DC 20560 GILLESPIE Hawaiian Evolutionary Biology Program, University of Hawai‘i at Manoa, 3050 Made Way, Honolulu, HI 96822 XV CONTRIBUTORS XVI THOMAS GIVNISH J. Department of Botany, University of Wisconsin, Madison, WILLIAM HAHN J. Laboratory of Molecular Systematics, Museum Smithsonian Institution, Washington, DC 20560 KENNETH WI 53706 Support Center, MRC 534, KANESHIRO Y. Hawaiian Evolutionary Biology Program, University of Hawai‘i at Manoa, 3050 Maile Way, Honolulu, HI 96822 THOMAS LAMMERS G. Department of Botany, Center for Evolutionary and Environmental Biology, The Museum Field of Natural History, Roosevelt Road at Lake Shore Drive, Chicago, IL 60605-2496 TIMOTHY LOWREY K. Department of Biology, University of New Mexico, Albuquerque, NM 87131 ROBERT PATTERSON Department of Biology, San Francisco CA State University, San Francisco, 94132 ROBERT ROBICHAUX H. Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, ANN AZ 85721 SAKAI K. Department of Ecology and Evolutionary Biology, University of California, Irvine, CA 92717 KERRY L. SHAW Section of Ecology and Systematics, Corson Hall, Cornell University, Ithaca, NY 14853-2701 JAMES SMITH E. Department of Biology, 1910 University Drive, Boise State University, Boise, ID 83725 KENNETH J. SYTSMA Department of Botany, University of Wisconsin, Madison, CHERYL L. WI 53706 TARR Department of Biology and Institute of Molecular Evolutionary Genetics, Pennsylvania State University, University Park, PA 16802 CONTRIBUTORS WARREN L. WAGNER Department of Botany, MRC 166, National Museum of Natural History, Smithsonian Institution, Washington, STEPHEN G. DC 20560 WELLER Department of Ecology and Evolutionary Biology, University of California, Irvine, CA 92717 XVll Introduction SHERWIN CARLQUIST The chapters in this book represent a quantum advance in our know^ledge of Hawaiian organisms. Although advances have been made in this field in each decade since the European and American voyages that first brought specimens to interested scientists, advances in the past two decades have been quite phenomenal. Acceleration of our knowledge of Hawaiian organisms represents a number of coincident advances, mostly of a technological nature. Ease of travel to the Hawaiian Islands greater number of scientists to the islands. readily accessible for field nity to study its work that is responsible for bringing a The Hawaiian chain many workers have is now so taken the opportu- organisms, the most remarkable oceanic island biota in the when world. Ease of accessibility coincides with a time species are, to a large extent, still extant. Most native Hawaiian of the key genera and species necessary for development of a clear picture of the evolution of most groups in the cent: Hawaiian biota are still in existence. But one should not be compla- Although many species remain to be studied in many respects, some have gone extinct and a number are endangered, with doubtful prospects for persistence of many of these species into the next century. This book, then, should be treated as an accomplishment but also regarded as an urging for still more work on the Hawaiian chain as well as on other oceanic islands that face similar threats to native biota. Another advance represented by uniform analytic technique, this volume is the application of a cladistics. Cladistic results have been ex- 1 CARLQUIST 2 pressed both in taxonomic and in geographic terms: the latter as the sequence of colonization of islands and even areas within islands. The application of cladistics to all groups analyzed has permitted the compar- ison of patterns, so that the range of phyletic and geographic patterns in evolution of the groups can be analyzed. results, particularly the area ter by Funk and Wagner. results other their cladograms, comparison of offered in the terminal chap- than those Funk and Wagner analyze and to unsolved questions of special mind cladistic My purpose in this chapter is to cite noteworthy probable significance. In reader’s is A many interest. call attention to instances, attention is called to There should be no doubt that despite the dramatic results of this volume, fascinating studies remain to be formed quite simply and done in any many —and many of these can be per- easily. HISTORICAL PERSPECTIVE Our current knowledge of Hawaiian organisms has developed from and from newer approaches and technolo- traditional systematic studies gies. Before considering those advances, those earlier systematists who we must pay special tribute to not only prepared systematic monographs of value but also were careful observers of the biology of Hawaiian organisms. Notable (1948). we still It among these are Perkins (1913) and Zimmerman should be stressed that for almost any Hawaiian animal species, would like to know more about habitat preference, diet, and behavioral details. For almost any Hawaiian plant species, welcome more information on ogy, and dispersal biology. we would ecological preferences, reproductive biol- In plants, workers notable for contributing data on chromosome number and morphology include George W. Gillett and Carl Skottsberg and, more recently, Gerald D. Carr and D. W. Kyhos (1981, 1986). I attempted anatomical studies of the Hawaiian tarweeds (Carlquist, 1957b, 1959a,b) and other groups. In animals, work on Hawaiian species of Drosophila was a key to using techniques beyond morphology. Hawaiian species of Drosophila resisted the ordinary Drosophila laboratory cultural techniques because of their specialized food requirements. ever, when cultural How- media based on extracts of Hawaiian plants were developed, the study of polytene chromosomes was enabled (see Chapter 5). Using chromosome data, the phyletic and therefore geographic interrelationships of Drosophila species and species groups could be Introduction 3 names of the workers involved in the massive effort devoted to Hawaiian Drosophilidae can be found in Chapters 5 and 6). Polytene chromosomes are not, however, available for Hawaiian elucidated (the organisms other than Drosophilidae, so other tools for determination of must be sought; genetic interrelationships information DNA. One is ultimately, the most satisfying currently derived from molecular studies directly using notable example of this approach Fleischer (Chapter 9), which suggests revision is the work of Tarr and in the traditionally recog- nized subfamilies and genera of Hawaiian honeycreepers. LOCATION AND TIMING OF ORIGINS OF HAWAIIAN AND OTHER ISLAND FLORAS AND FAUNAS The recency of adaptive in such genera as radiation Drosophila (Chapters 5 and Tetramolopium (Chapter some and other types of speciation 11), 6), We can evident Geranium (Chapter and Clermontia (Chapter 10), 15), as well as in Hawaiian tarweeds species groups in other examples, such as the (Chapter 13). is certify the recency of this evolution because of the high degree of certainty of the area cladograms and the excellence of potassium-argon datings available for islands in the Hawaiian chain Chapter 2). No for speciation scientist —with (see reading the massive evidence this book presents many events of adaptive shifts, as evident in the tarweeds (Chapter 13) and Schiedea (Chapter 12) —could doubt that there has been, in genus after genus, autochthonous evolution, especially that involving adaptive radiation. Europeans studying Atlantic islands have views of evolution on islands that contrast starkly to the amazing patterns demonstrated in this book. For example. Berry (1992), in a section of his paper entitled '‘Modern Island Biology,” stated that “post-colonization adaptation probably plays little part in the origin of most endemism; although natural selection affects island biotas just as than much as —^perhaps —continental ones, the main differentiation of island forms is more usually the result of the chance characteristics of the original colonists of each species.” Likewise, Cronk (1992) generalized the relict nature of oceanic island endemics is for islands of the world, “If accepted, then these plants become a key to understanding biogeographical and taxonomic patterns. They indicate groups in which extinction has occurred, and the degree of taxonomic and geographical disjunction may reveal something of the extent of that extinction.” CARLQUIST 4 European workers apply the relict hypothesis of oceanic island endemics to manifestations that are demonstrated in this volume to result from adaptive radiation. For example, the cladograms for Tetramolopium (Chapter 11), the Hawaiian tarweeds (Chapter 13), and the Hawaiian lobelioids (Chapter 14) Hawaiian show evolution of increased woodiness on the However, for the Atlantic Islands. islands, Sunding (1979) interpreted similar patterns in exactly the reverse fashion: “That a large proportion of the Macaronesian vascular flora not only by its is of a great age is shown present-day distribution patterns with often large distribu- tion gaps to the nearest related taxa, but also by features like the prevail- ing woody life-form in genera elsewhere represented by herbs (Echium, Sonchus, Limonium, PlantagOy Sanguisorba, etc.). ...” Cronk (1992) regarded the St. Helena Island endemic monotypic probably older than 10 My,” genus Petrobium as being “Miocene . . . although the origins of the entire family Asteraceae are likely only a before Miocene —upper Oligocene—and Petrobium is little by no means prim- Cronk (1992) apparently was impressed by resemblances of Petrobium to the Polynesian genus Oparanthus and regards this two-ocean distribution as indicative of great age. In fact, both Petrobium and Oparanthus are recent derivatives of itive in the Bidens —so family (Jansen et 1991). al., recent, in fact, that Stuessy (1988) justifiably reduced both Oparanthus and Petrobium to Bidens. Those European workers cited who do believe in the relictual hypothesis above do not mention potassium-argon dates for islands. Moreover, molecular data, one of the great strengths of this book, have been applied to only a few situations on Atlantic islands. Very likely, the picture of more evolution on Atlantic volcanic islands will change to resemble closely the findings ogy and from the Hawaiian chain when evidence from DNA analysis One is available and when cladistic geol- methods are used. of the fascinating implications of this volume concerns the role of now-vanished islands of the Hawaiian chain in contributing to the flora and fauna we now see on the present high islands. familiar with the sequence beginning as early as ago (Ma) at Suiko the is Ma, are now 70 to 80 million years Seamount, proceeding through the bend Daikakuji Seamount at about 43 We in the chain at terminating at Kilauea and Lohhi, seamount that has not yet surfaced but is still actively growing. There a continuous chain, but has there been biological continuity? In other words, are there any plants or animals currently on the high islands whose ancestors were once on the earliest islands, such as Suiko? There are plants in the current Hawaiian flora one might be tempted to think Introduction 5 ancient in origin because of the families to which they belong, such as Eurya (Theaceae) and Cryptocarya (Lauraceae), but we have only one Hawaiian species each of these, and determining the date of their separation from the remainder of their respective genera may be impossible. However, both Eurya and several genera of Lauraceae are native to both Ogasawara (Bonin) and Volcano Islands, volcanic islands that are indubitably recent, have never had any contact with the Japanese mainland, and are far enough from Japan so that one must invoke long-distance dispersal as the mechanism for populating these islands (Kobayashi and Ono, 1987). If origins of the Hawaiian flora and fauna were on islands formed previous to Kaua‘i and Nihhau, we would expect to see, in an appreciable number of instances, two or more quite distinct lines present on Kaua‘i and post-Kauah islands, lines that on Kaua‘i already show long divergence. This would be evident in cladograms and in the nature of the the molecular data that might be used in cladograms. This kind of pattern is seen in several of the cladograms presented in this volume. Other cladograms suggest origin on Kauah or an even younger island the island of Maui in the case of —even on Tetramolopium. Another example is Clermontia, where the origin of at least the present extant species can be traced to the youngest island, Hawai‘i; however, the origin of the ancestors of Clermontia may have been on an older island (see Chapter 14). These patterns show that speciation and evolution on oceanic islands are quite recent, and we need not hypothesize radiation on continents lowed by dispersal of the products of that radiation to fol- islands, as Berry (1992) or Cronk (1992) did. There are no spectacular on the present Hawaiian chain, and calling any of them ancient is a misnomer if one compares Hawaiian plant or animal groups with those on continental islands. Even though the Emperor Seamounts could have theoretically served as steppingstones for plants and animals to migrate to the present-day high islands, we relicts lack evidence that they did so. Although volume mention possible some of the chapters in this origins of groups as far back in the chain as French Frigate Shoals, Necker, or Nihoa (when they were high islands), none of the Leeward Islands earlier than those have yet been placed into and indeed, the geologic evidence shows the chain was dormant before Kure long enough so that no colonization from pre-Kure consideration, would have eroded to atolls or seamounts before Kure or arose from the sea) can be hypothesized (see Chapter 2). A islands (which later islands few pre-Kaua‘i but post-Kure floristic or faunistic elements might have 6 CARLQUIST been displaced by and thereby extinguished by more recent colonists. Such genera as Alectryon, Hesperomannia, Hihiscadelphus, and Kokia suggest older immigrants that are no longer common and were already disappearing in prehuman times. NEW EVIDENCE, NEW IDEAS The geologic overview by Carson and Clague in Chapter 2 performs an enormous service by refining our ideas on age and size of islands in the Hawaiian chain. These are of great importance in providing a chronology for evolutionary events specified in the various cladograms. In this con- made by Carson and Clague that each of post-Wai‘anae volcanoes was at some earlier time coalescent with nection, the point volcano preceding islands, or is it the the The disappearance of huge portions of in the series. even the entire above-water portions of an island, by cata- strophic slumps will be a concept new to most readers. The most impor- tant contribution of Chapter 2, where biogeography summary Hawaiian chain was dormant of evidence that the is concerned, is the for long enough before the emergence of Kure so that no colonization from pre-Kure islands can be hypothesized. They had eroded to atolls or seamounts before Kure emerged, so that no high-island elements from older islands were available to Kure or the islands that emerged later than Kure. This should certify the role of long-distance dispersal in colonization of the Hawaiian chain, and the consequent importance of the Ha- waiian chain in discussions of long-distance dispersal has been magnified accordingly. Chapter 4 by Shaw deals with evolution of Hawaiian Noteworthy is the fact that demonstrated by strates that “the cricket species are one-island endemics, all DNA evidence Rapidity in species formation Hawaiian have diverged from crickets. but not always clear from morphology. is clearly suggested. Shaw also demon- tree crickets, as well as the swordtail crickets, their original founder lineages to such a degree that they were taxonomically misplaced by Perkins (1899) and Zimmerman (1948).” Chapter 5 on Hawaiian Drosophila by Kaneshiro, Gillespie, and Carson topic. is It notable for providing a thorough review of this fascinating updates the estimate of Hawaiian Drosophila species to a startling 1,000, of at present. This which a little more than half (511) have been described emphasizes the fact that the exploratory phase in Hawai- — Introduction ian biology is not yet complete in some groups where alpha taxonomy 7 is concerned and clearly not complete for any Hawaiian plant or animal group where understanding of species biology is concerned. Readers will be fascinated by the close relationship in chromosomal sequences be- tween the Hawaiian D. primaeva and D. colorata of Japan. When molecular approaches are used for study of Hawaiian Dro- sophilidae (see Chapter 6 by DeSalle), species hints group no single continental species or clearly appears as ancestral to the drosophilids. There are from molecular data that Hawaiian drosophilids might date back 20 to 40 Ma, although one would like more molecular clock indicators. At the other extreme, many readers will be surprised by the way molecular data can elucidate the “microbiogeographic” sequences in coloniza- on the island of Hawaii. Asquith (Chapter 7) shows that taxonomic identity of host tion of the newest areas plants is of great significance in analyzing phyletic patterns of Sarona (Heteroptera). Shifts in host plant preference can be placed on cladograms. Be- may mediate sympatric analysis may go well beyond cause Asquith claims these shifts in host plants events of speciation, the significance of his Sarona and may be applicable to other claimed instances of sympatric speciation. Asquith hypothesizes origin of Sarona on a pre-Kaua‘i island. and Croom (Chapter 8) offer the only contribution on spiders, whose evolution on islands has not been subject to much discussion. The Hawaiian species of Tetragnatha are diverse with respect to colors, shapes, sizes, ecological preferences, and behaviors, in contrast to Gillespie the relative uniformity of the genus elsewhere in the world. species appear to have originated ric, as in on Kauah; speciation has been most Hawaiian plant and animal groups, with the non-web-building group never occupying the the same The Hawaiian allopat- sister species in same volcano or even island. Tarr and Fleischer (Chapter 9) uncover some unexpected interrelationships in their analysis of Hawaiian honeycreepers. For example, the Kaua‘i Creeper {Oreomystis hairdii) thick-billed is apparently the Laysan Finch (Telespiza cantans). Melamprosops and Paroreomyza Two sister species of the of the honeycreepers —may not be honeycreepers at all but rather products of an independent colonization. However, the honey- creeper analysis makes one wonder how topology of the cladogram would change if DNA were available for very rare or extinct species. Acknowledging that their study is not definitive, Tarr and Fleischer offer a molecular clock figure of 3.5 Ma for the origin of the Hawaiian honeycreepers. This date is of considerable interest because the Hawaiian 8 CARLQUIST honeycreepers appear to be one of the older elements in the Hawaiian fauna. Funk and Wagner (Chapter lads with iar 10) analyze seven flowering plant phy- morphology-based cladograms. The results underline the famil- old island to young island progression seen in the majority of Hawaiian groups analyzed. However, the Hawaiian may well have originated on Maui and spread species of Geranium to both Kaua‘i Hawai‘i. Their analysis of Hesperomannia shows that the basal species, H. lydgatei, rich in is represents an extension of the Pacific genus Olearia, as hypothesized by Wagner and Herbst (1987), lined by the Funk and Wagner cladogram and this genus, earlier considered of uncertain Lowrey (Chapter Kauah autapomorphic characters. The likelihood Remya that the asteraceous genus and is under- resolves relationships of affinities. 11) analyzes the evolution of another genus of Asteraceae, Tetramolopium, with respect to geography, habitat, and ecological change. The results should be read by Europeans who have a relict view of oceanic island floras and faunas. Lowrey shows that Tetramolop- ium has traveled from New Guinea to Maui, with subsequent radiation to other main islands (except Kaua‘i, on which Tetramolopium is absent). Especially interesting Mitiaro in the Cook is the occurrence of T. sylvae both Islands. This of long-distance dispersal from sification on the is on Maui and on the result of a relatively recent event Maui or Moloka‘i to Mitiaro, and diver- latter island is already in progress. Lowrey’s results demonstrate clearly the occurrence of long-distance dispersal, the rapidity of adaptive radiation, the change from less to marked nature of ecological shift, more woody autochthonously and the in volcanic oceanic islands. Wagner, Weller, and Sakai (Chapter 12) study the complex formed by two endemic genera of Caryophyllaceae, Alsinidendron and Schiedea. They show that persistent field edly extinct species. The inclusion of these complex has resulted shifts work can uncover both new and supposin the latest cladograms of the in greater resolution. Schiedea has shown several between wet and dry habitats, so that the original habitat for the complex is uncertain. Likewise, there have been shifts from subdioecy to gynodioecy, as well as to hermaphroditism. Such changes should be kept in mind by those who tend to have a unidirectional view of these trends. Analysis of the Alsinidendron-Schiedea complex suggests a pre-Kaua‘i origin for the group, although not Kauah. markedly earlier than the emergence of Introduction The 9 Hawaiian tarweeds or silversword alliance (see Chapter 13 by Baldwin and Robichaux) is surely the most spectacular example of adaptive radiation on islands in the world. That fact alone makes any detail about this complex of great interest in contributing to this radiation of the amazing story. The ancestry of the Hawaiian tarweeds is in the Californian Floristic Province and very likely from subshrubby montane tarweeds of northern California such as Raillardiopsis or Madia species. The cladograms Argyroxiphium for the is complex an interesting paradox: Although offer the oldest of the genera to have originated single ancestral colonization, it from the does not occur on Kaua‘i or 0‘ahu. The authors entertain the possibility of a pre-Kaua^i origin for Argyroxiphium. pre-Kaua‘i, the origin If genetic distance between origin on Kaua‘i is was certainly not much earlier, because the Argyroxiphium and the other genera, for which indicated, is not very great. Montane Californian tarweeds likely to have been close to types ancestral to the Hawaiian complex exist in a climate at 2,200 to 2,800 1,000 to 1,500 m much like the climates at m in the Hawaiian Islands. This latter elevational range has been gone from Kaua‘i for perhaps the past from 0‘ahu also has been absent 1 to 2 million years and for that long as well. Transfer of Argyroxiphium from a now-vanished dry alpine zone on Kauah to Maui (where it has radiated into wet habitats from dry alpine we imagine an Argyroxiphium sites) seems on a pre-Kaua‘i island, we have to imagine it inhabiting Kaua‘i or both Kauah and 0‘ahu, yet not radiating into any of the habitats now present on these islands and entirely possible. If not surviving in those habitats. pre-Kaua‘i island, it is likely to origin Had Argyroxiphium have left some species originated on a on Kauah and/or 0‘ahu because of the capability of the genus for radiating rapidly into a A rapid diversification of a tarweed from montane range of habitats. California in a high-montane habitat on Kaua‘i into Argyroxiphium, Dubautia, and Wilkesia seems entirely conceivable on the basis of presently available data. In fact, the two islands before the Ni‘ihau-Kaua‘i complex would not have been suitable as sites of colonization for an alpine tarweed: The maximum height of Nihoa was 1,300 m and that of Necker 1,100 m (see Chapter 2), so those two islands would have lacked alpine regions in which Argyroxiphium could have originated if its ori- most of the evolution of the complex has taken place rapidly on Kauah, according to the available data. This gins were alpine. Certainly, scenario indicates this how much we know and other Hawaiian groups, even elusive. about the evolution and origin of if a few pieces of information are 10 CARLQUIST The Hawaiian weeds lobelioids are hardly less spectacular than the tar- in their adaptive radiation. contributed by Givnish et al. Information about the lobelioids (on Cyanea, Chapter 14) and by (on Clermontia, Chapter 15). The chapter on Cyanea is hammers also contains an account of the radiation that led to Brighamia, Delissea, Cyanea, Rollandia, and Clermontia. The fact that most of the be maintained despite their closeness able significance is is traditional genera can interesting. the fact that Brighamia However, of consider- and Delissea are and had a lowland (and therefore moderately dry affinities of still sister genera forest) origin. Brighamia have hitherto been uncertain because of its The highly distinctive features. Givnish et al. entertain the interesting speculation Cyanea that the prickly (thorny) sure from now-extinct browsing birds, the skeletons of which have large recently been found. Givnish et al. present carefully assembled circum- New arguments and draw on a parallel with the defenses of stantial now Zealand plants, defenses herbivory of the 1965). species originated in response to pres- What or other large birds, as Givnish do not address et al. Hawaiian species in the capabilities, moa generally accepted as response to the flora I is hinted earlier (Carlquist, why the prickly Cyanea have apparently increased in defensive whereas despite rapid plant evolution, no other Hawaiian angiosperms have increased either physical or chemical defenses, so we know. On far as Hawaiian angiosperms have plummeted to the lowest levels seen on any oceanic islands, suggesting that large herbivorous birds have had little or no effect. In New the contrary, defenses of Zealand, on the contrary, physical defenses are numerous and other families, spiny leaves in Aciphylla number of species of divaricating shrubs). (juvenile leaves of Araliaceae and Olearia, and a Chemical defenses large in the New Zealand flora are also much higher than in Hawaiian Islands. The moas undoubtedly had a much longer tenure on New Zealand than did the large herbivorous birds on Hawai‘i, but many of the New Zealand plant groups with defenses are the same or the close to the plant groups Olearia, Remya). ian flora as That such on Hawai‘i and are likely recent relatively recent (e.g., immigrants to the native Hawai- Argemone and Rubus have lost most of despite the presence of the large herbivorous birds earlier suggestion that is their prickliness mystifying. My land mollusks (but definitely not Achatinellidae or Succinidae, which graze on surface algae and fungi) might have posed a Cyanea was based on the tendency of mollusks to graze more on lower leaves of plants, neglecting upper leaves, which tend to be drier. Prickles in Cyanea and Rollandia tend to occur in wet habitats, which threat to 1 Introduction tend to be densely vegetated and perhaps thus 1 less accessible to large herbivorous birds than the more open habitats, which are lacking in prickly Cyanea species. The fascinating results of hammers in Clermontia suggest that this genus had its origin on the very new island of Hawaii. Lammers’s data seem to support this clearly, and it is entirely plausible to me. The rapidity of speciation and diversification of floral forms and sizes in Clermontia show how explosively plant evolution can occur on oceanic islands, countering the contentions of those who hold the relict hypothesis for oceanic island species. Migrations of a genus from a younger island to an older island require hypothesizing an open niche, and open niches are presumably fewer on older islands than on younger clermontioides, the sole species of the genus able to colonize Kaua‘i because epiphytic habitat is, its habit is islands. Clermontia on Kauah, may have been epiphytic or semiepiphytic; the by definition, a pioneering habitat. Patterson’s analysis of Scaevola (Chapter 16) clarifies the origins of the Hawaiian species of this group, confirming that at least two introductions account for the presently native species on the Hawaiian chain. One would like to DNA know how would change if widespread plumieri and the S. the topology of the cladogram data were available and many if species such as the Australian species (especially those with fleshy fruits) were included. Inclusion of these would clarify the migration of the genus into the Pacific widespread beach Scaevola, S. sericea. I and the origin of the note with interest Patterson’s observation, based on cladistic results, that dry country species of Scaevola have lost dispersibility to a lesser extent than have the wet which accords with my thesis that loss of dispersibility much more abundant in Hawaiian plants of wet forests than those forest species, is of dry forests (Carlquist, 1974). NATURE OF SPECIATION ON THE HAWAIIAN CHAIN One theme clear from these chapters is that evolutionary diversification on the Hawaiian Islands has been recent and has taken place autochthonously. Nevertheless, diversification has been profound, involving more than just a few morphological features. Robichaux et al. (1990) showed genuine physiological diversification among Hawaiian tarweeds: They are not doing similar things in different places; they are just as diverse in physiology as any selection of species from their various 12 CARLQUIST habitats. The leaf anatomy of the Hawaiian tarweeds also shows excep- tional diversification (Carlquist, 1957a, 1959a,b). Likewise, the ian honeycreepers differ not just in Hawai- shape but in such features as bill tongue morphology as related to their diverse food sources (Amadon, 1950). These examples are cited because there that adaptive radiation on islands is some tendency is to believe pervasive in diversification of less its products than adaptive radiation on continents. There are some spectacular examples of speciation adaptive shifts — on the Hawaiian chain that do not involve for example, the agate shells (Achatinellidae) would never be — ^but these examples of adaptive radiation but rather as cited as examples of speciation that does not involve radiation into different habitats. Despite the genuine morphological, ecological, physiological, and anatomical diversification of Hawaiian plants, there has apparently not been concomitant genetic change. For example, weeds can be crossed with each sterility with other, the all little if Hawaiian any tar- interspecific evident (Carr and Kyhos, 1981, 1986). This contrasts with the Californian tarweeds, in which strong common, and one can sterility barriers of strong cite instances among species are even sterility barriers within species (Clausen, 1951). This tendency also appears to be true of other groups in the Hawaiian flora, such as Bidens (Gillett and Lim, 1970). In part, these patterns may result from recency of fact that species are geographically isolated may speciation. The also be responsible: Closely related congeneric species in the Hawaiian flora are rarely sympatric. In general, lesser extent woody genera tend to develop interspecific barriers to a than do herbaceous species (e.g., Pinus, Quercus). This related not to woodiness per se but to the fact that cope with greater ecological diversity (e.g., woody plants is must plants with deep root systems encounter a greater range of soil structure and soil therefore successful occupancy of a diverse area moisture regimes), and may be related to reten- more heterogeneous gene pool. A higher proportion of Hawaiian plant species are more woody than are Californian plant species. One key group in which interspecific fertility is tion of interspecific fertility, widespread (not analyzed tree of all in this book) is Metrosideros (Myrtaceae), chief As the specific epithet (polymorpha) originally members of this complex suggests, the patterns of variation in Hawaiian given to resulting in a forests. Hawaiian Metrosideros involve the kind of diversity that active hybrid swarms exhibit, but some highly distinct populations have also been recognized as segregate species. Is this confusing pattern the result of 13 Introduction Hawaiian chain, or even recolonizations several colonizations of the within the chain, with subsequent hybridization among the various pop- ulations? Certainly, Metrosideros represents a remarkably successful sys- tem for occupancy by a tree species of various regimes to high bogs and deserves Metrosideros, we When we study. likely will better from new dry lava understand the genetics of understand the significance of genetic systems of other Hawaiian plants. PERSPECTIVES FOR CONSERVATION AND FUTURE RESEARCH One is At a key to their preservation. imply that if the conserva- Hawaiian organisms. Information about endangered organ- tion status of isms book regards of the indirect messages contained in this we know more about that can help us to manage its first glance, this statement seems to we a plant or animal, survival better. That will know likely true, is but facts I am also concerned with putting information about endangered species in the hands of the public—simple, appealing name facts they can associate with the of a plant or animal. Information of this sort support of conservation is essential to public because the public supports conservation efforts, of plants or animals about which it knows something. Species about which the public knows nothing or about which the public has no visual image are unlikely to be conserved. Therefore, the tremendous amount of information on Hawaiian groups in this book to conservation, and any future is of great potential value efforts to familiarize the public unique characteristics of Hawaiian biota are likely to with the have importance in gaining support for conservation efforts. Conservation efforts lar Hawaiian Faced with efforts is may species, regardless this, may not succeed in saving any particuof how intensive those efforts may be. or our best option other than continuation of conservation to study Hawaiian species intensively. Future generations may not fault us for failure to save a species that could not have been saved with reasonable, simple, practical measures. They will surely fault however, if simple measures were available but not used and may, ably, fault us for not gathering as tion that can only be gathered these species are still much from in existence. us, justifi- information (especially informa- living specimens) as Authors in this we can while book have made enormous contributions toward the goal of advancing our knowledge of Hawaiian organisms and are to be commended. store of Geology and Biogeography of the Hawaiian Islands 2 HAMPTON DAVID The diverse biota of the A. CARSON AND CLAGUE L. Hawaiian archipelago presents a geographic puzzles. Many oceanic islands are among forms have continental the most terrestrial immigrant they, or their ancestors, arrive? that has occurred in situ essentially lines affinities, isolated in the world. challenging questions deal with origins. endemic large and Where species number of but these The most did each of the many come from? When did Are they aboriginal products of evolution on the present islands or were they bequeathed, unchanged, from nearby ancient land masses that have since disappeared? Whatever the answers, the ancestral lineages that have led to the present-day endemic species need to be identified genetically. This process is now possible through the use of exquisite new techniques that use molecular markers of ancestry. Beyond these phylogenetic problems, there greater challenge. We may is another and even look at the population genetics of selected and try to identify the proximate causative factors that have promoted evolutionary character change in their populations. island forms Populations that have colonized the recent lava flows of the newer islands deserve special attention as possible sites of dynamic genetic change. To provide a background to these studies, we review geologic and geographic information that provides the time and spatial control for investigating further these challenging lines of inquiry, building on several other compilations and reviews that have dealt with Hawaiian biogeogra- 14 Geology and Biogeography phy (Zimmerman, 1948; Kay, 1972; and Dalrymple, 1987). Carlquist, 1980) 15 and geology (Clague HAWAIIAN ISLANDS AND PLATE TECTONICS Unique new data exist on the geologic history of the Pacific, the greatest of the earth’s oceans. Plate tectonics has revolutionized our understanding of the “ring of fire” that fringes the Pacific and of the many mid“Pacific islands. This volume focuses on the biological history of one such group, the Hawaiian archipelago. The geologic data permit us to formulate more realistic interpretations of both the biogeography and evolutionary patterns of the organisms present. The implications dis- cussed here deal particularly with terrestrial biota but are also relevant to marine organisms. Tuzo Wilson (1963) proposed an insightful new explain the origin of the Hawaiian Islands. This is now the Thirty years ago, hypothesis to main unifying theory wide (e.g., J. for the origin of many oceanic island groups world- Christie et ah, 1992). Simply stated for the Hawaiian Islands, the evidence indicates that the islands were formed successively over a fixed “hot spot” beneath the northwestward-moving Pacific tectonic plate. Morgan (1971) provided a physical model that consists of a ther- mal plume of material anomaly beneath the form plate. from the deep mantle that forms a melting The magma perforates the plate at intervals to discrete volcanoes as the plate slowly highest points The east arising may rise above sea level as moves over the hot spot. The emergent oceanic islands. eight current high islands of the archipelago end of a much longer and remarkably occupy the south- straight line of seamounts, the Hawaiian Ridge, extending 3,493 low km islands and northwest of Kilauea to Daikakuji Seamount (Figure 2.1). At this point, the orientation known Emperor Chain, continues for another 2,327 km, culminating at Meiji Seamount. The latter may have resulted from the initial volcanic activity over the Hawaiof the chain turns sharply northward and, as the ian hot spot about 75 to 80 million years ago (Ma). movement rate of of the Pacific plate has apparently undergone occasional slight changes, the direction of movement northwest about 43 Ma. This Pacific plate has shifted dramatically shift in direction is Hawaiian-Emperor Bend (Eigure at Although the from north to now marked by the 2.1). Since the time of the bend, the been moving 8 to 9 cm/year over a fixed hot spot located about 19° N, 155.5° W (Clague and Dalrymple, 1987). 120 ° 60 50 ° ° 40 ‘ 30 ° 20 ° 10 ° FIGURE 2.1. Bathymetry of the Hawaiian-Emperor volcanic chain. Con- 1-km and 2-km depths shown in area of chain only. Inset: Location of outlined by 2-km depth contour, in central North Pacific. From Clague tours at chain, and Dalrymple (1987). Each Fiawaiian volcanic island and submerged seamount in the Fiawaiian-Emperor Chain (Figure 2.1) appears to have been formed succession; all continue to drift Pacific plate. This process has is younger than Islands (i.e., its northwestward on the surface of the produced a series of islands, neighbor to the northwest. more than 400 each of which Of the eight high Fiawaiian m above sea level), Kaua‘i in the northwest is Ma. As diagrammed in Figure 2.2, this age when the island of Kauah was directly above the the oldest, formed about 5.1 corresponds to the time in Present 0.5 Kaua't 22 Ma ‘ 22 ' , Nihea Ife'yla Kaua'i Molote'i O'ahu Maui O'ahu Lanai Ka’ula Kaho'olawe 20 ' - 20 ' / J ^^Mahukona s f Loihi 156 ° 158 ° 160 ° Seamount 1.0 22 154 ° 2.5 22 Ma ' Nihoa ^ Kaua'i <pO 20 ' 156° Ma * . 158° 160 ° 154 ° O'ahu '^Ni'ihau Ka'ula 20 * 9 QNihoa teua'i ' 1 L/ Nilhau ^Ka'ula ^^Maui NuijX a 18‘ 154 ° 158 ° 160 ° 156 ° 154 ° 154 ° 7.5 Wla 22'' Gardner Pinnaeles La Perouse Pinnacles 20 o °. ^ Necker «— 'Nihoa, 18° 154 ° FIGURE 2.2. 156 ° 158 ° 160 ° Coordinate positions in the Pacific Ocean of the present Ha- waiian Islands and their reconstructed positions at five different times in the The rate of movement of the plate is assumed to be 9 cm/year. In each panel, the Hawaiian hot spot is indicated {light dashed circle) and lies at approximately 19° N, 155.5° W. Islands, past, as affected ridges, lier and by the movement of the reefs are identified Pacific plate. by outlines; their sizes and fused state at the ear- times are estimates based on bathymetry and GLORIA side-scan data. CARSON AND CLAGUE 18 Hawaiian hot spot. The southeast portion of the includes the three active volcanoes now m is Mauna Loa about 30 km of sea level (Figure 2.2, upper which Loa, Kilauea, and Hualalai, above the hot spot. Lo‘ihi Seamount sits the southeast flank of 950 Mauna island of Hawai‘i, a new active volcano offshore; it rises to on within Potassium-argon ages and left). paleomagnetic declination measurements confirm the recency of the land of Hawai‘i; no lava flows formed earlier than about 0.5 Ma is- have been found (McDougall and Swanson, 1972). HIGH AND LOW ISLANDS The height of the islands relative to sea level is inevitably reduced by two factors, erosion and subsidence. The greatest elevations above sea level are currently displayed by the islands at the southeast end of the archipel- ago; they decline in height to the northwest. For example, rises to Mauna Kea 4,205 m, whereas the highest points on Maui, 0‘ahu, and Kaua‘i and 1,598 m, respectively. Farther to the northwest, the elevations are low in comparison; examples include Ni‘ihau, 390 m; are 3,055, 1,231, Nihoa, 277 m; Necker, 84 m; Laysan, Kure Atoll, 6.1 m. In the Kure Atoll sea level. ian Ridge and the is latter three, the volcanic Emperor Chain rocks are well below Hawai- are currently submerged. Islands to the northwest of Kaua‘i have relatively sparse terrestrial biotas that nevertheless appear to include relict species m; and Atoll, 3.7 the northernmost island; the westernmost entire The Hawaiian 11m; Midway (Conant et al., 1984). It some recent seems hardly possible that pro- pagules arising from these low islands could be a source of a significant number islands. of suitable colonists for the higher-altitude areas of the newer The extraordinary diversity, specific and generic endemism, and discordant nature of these high-island biota require further explanation, as discussed below. Many volcanoes of the Hawaiian-Emperor Chain were at one time high islands. Direct evidence as to how high they were and what the nature of their terrestrial biota might have been, however, Estimates of their former height have been lines identified using al., made based on is lacking. their shore bathymetric and GLORIA side-scan data (Torresan 1991) and the slopes of the currently active volcanoes et (see island outlines in Figure 2.2). These volcanoes, at the time they were high islands, were located near the position of the present high tectonic reconstructions (e.g., islands. Plate Atwater, 1989) further indicate that the Geology and Biogeography islands were probably Pacific as are the just as modern 19 remote from the continents fringing the islands. DEVELOPMENT OF THE HAWAIIAN ISLANDS Hawaiian Islands, with the exception of Lohhi, Kilauea, and Mauna Loa, were once significantly higher than they are at present. Moore and Clague (1992) estimated the maximum heights that All the volcanoes in the 0‘ahu to Hawaih once attained by adding the current height above sea level and the depth of the deepest slope break below sea level (former shoreline). Their results are shown in Table 2.1, with the volcanoes from additional estimates for several older volcanoes in the chain; potassium- argon ages for each volcano are also given. As the volcanoes subsequently grew, many coalesced to form composite islands, and sank, they once again became separate islands as these islands then made more two separate islands that coalesced as Kauah grew, but later they separated to become separate islands once again as they subsided. The wide and deep Kauah Channel between Kauah and Wai‘anae volcano on 0‘ahu provided a volcanoes. For example, Kaua‘i and Nihhau formed of one or as formidable hurdle to species migration along the chain. However, the channel was not nearly as wide as it now appears because the Ka‘ena Ridge to the northwest of Wai^anae volcano was a subaerial ridge when it formed at perhaps 3.5 Ma. Thus, the channel between Kaua‘i and Ka‘ena Ridge was only about 48 km across, compared with the 116-km width of the present-day channel. In contrast, for the subsequent volcanoes to form, after Wai‘anae and each coalesced with the previously formed volcano. until Haleakala, The channel between Ko‘olau and Penguin Bank is only 690 m deep, yet the islands have subsided more than this (1,100 to 1,200 m), so these minimum elevation of comprise the Maui Nui complex, volcanoes were once joined above sea level with a about 400 m. Similarly, volcanoes that which consists of East and West Maui, eastern and western Moloka‘i, Lana‘i, and Kaho‘olawe, were one time joined by land bridges with minimum m. The elevations of 1,300 islands, at large Maui Nui complex became two first one consisting of Molokah and Lanah and a second consisting of Maui and Kaho‘olawe. This breakup happened sand years ago (ka). less than 300 to 400 thou- Kaho‘olawe then separated from Maui and Lana‘i separated from Molokah, both less than 100 to when Penguin Bank subsided to sea level, as 200 the coral cap ka. is It is of finally unclear unknown CARSON AND CLAGUE 20 TABLE 2.1. Present and Maximum Heights and K-Ar Ages of Hawaiian Volcanoes Height Volcano (mr Age Maximum^ Present (millions of years -950 -950 No K“Ar data Kilauea 1,247 1,247 0- 0.4 Mauna Loa Mauna Kea 4,169 4,169 0- 0.4 4,205 4,600 0.38 Hualalai 2,521 2,950 Lo‘ihi Kohala See Fig. 2.3 1,670 2,670 - 1,100 235 3,055 5,000 0.75 450 2,100 1.03 West Maui 1,764 3,400 1.32 Lana‘i 1,027 2,200 1.28 East Moloka‘i 1,515 3,300 1.76 421 1,600 1.9 -200 1,000 No K-Ar data 960 1,900 2.6 Wai‘anae 1,231 2,220 3.7 Kaua‘i 1,598 2,600 5.1 390 168 1,400 4.89 -200 800 277 1,300 7.2 84 1,100 10.3 Mahukona Haleakala Kaho‘olawe West Moloka‘i Penguin Bank ° 2 r N, 157 35 ' ( W ) Ko‘olau Ni‘ihau Ka‘ula 0.43 See Fig. 2.3 800 4.0 Unnamed “ 22 40 Nihoa Necker ' ( N , 161 “ W ) No K“Ar data ^Negative values represent meters below sea level (submarine volcanoes). ^From Moore and Clague (1992). ^Best K-Ar data on surface basalt (from Clague and Dalrymple, 1987). thickness. However, we assume that before Moloka‘i and Maui became it separated from West Moloka‘i separate islands. The channel between Haleakala and Kohala volcanoes is 1,890 m deep, and subsidence on the south side has been on the order of only 1,000 m. Therefore, Maui and Hawaih never had a land bridge between them, and species had to cross a narrow channel. The subsided shorelines on the two sides of the channel are only about 13 km apart, so the channel was this narrow when Kohala formed a high island about 370 ka (see also Figure 2.3), Geology and Biogeography 21 EROSION Erosion is an important process that reduces a new high island to sea Hawaih, the persistent northeast trade winds and the southerly “Kona” winds are heavily laden with level. At the latitude of present-day moisture that, as rain, brings about surface erosion. Catastrophic collapses of large sections of the present islands have been identified by systematic submarine mapping of the Hawaiian Ridge between Kaua‘i and Hawaih (Moore et al., 1989; Moore et aL, 1994). Slump and avalanche debris deposits are exposed over approximately 100,000 km^ of the ridge and adjacent sea floor or an area more than five times the surface area of the present islands. These slope failures begin before emergence of the island and continue after emergence and after dormancy. Seventeen such deposits have been recognized around the present-day high islands, involving areas adjacent to each. Examples for the island of Hawai'i The data indicate that these collapse events are diagrammed are a in Figure 2.3. normal part of the cycle of island growth and Such large landslides could affect decline. biogeography by removal of certain gene pools or even localized species. Further, they might be instrumental in introducing species from one island to another by rafting of debris after the slide has occurred. SUBSIDENCE Many data show that large volcanic islands sink below the ocean surface. not a linear process but occurs in two distinct Subsidence, however, is stages (Moore, 1987). The most rapid sinking comes about when a volcano is still an active in state of large growth. This early phase results from the flexure of the underlying plate caused by the added load of the volcano. The island of rates of at least 2.5 to 3 Within Hawaih, mm/year and has been 1 million years, the phase of subsidence is for example, is currently subsiding at for the past 0.5 million years. subsidence slows dramatically. A second slower mainly due to thermal contraction as the lithosphere ages with increasing distance from the hot spot. Observations indicate that subsidence of Maui, 0‘ahu, and Kauah, for example, is currently much slower than that of Hawai‘i. Although this process continues for tens of millions of years, the rate decreases so that nearly occurs in the 40 to first all the subsidence 10 to 20 million years. By the time the seamounts are 50 million years old, they are subsiding only very slowly. 22 CARSON AND CLAGUE FIGURE 2.3. half-million years Six stages in the growth of the island of Hawai‘i over the past on present-day base map. Existing bathymetric contours are shown by fine lines (depth in km). Island growth is shown at 100-ka intervals {bold numbers), with shoreline and volcano boundaries {heavy solid aerial volcanic centers {solid stars), dormant or waning subaerial volcanic vigorous sub- centers {open stars), feebly active subaerial volcanic centers {open circles), axis of Deep {cross-dashed line), and boundaries of landslides dashed lines), lines). The shorelines are mapped Hawaiian {stippled pattern delineated by as a break-in-slope offshore of the present coast and are inferred where buried by growth of subsequent volcanoes. Volcanoes (delimited by lines within island) are Ha, Haleakala; M, Mahukona; Ko, Kohala; MK, Mauna Kea; ML, Mauna Loa; H, Hualalai; Kl, Kilauea. From Moore and Clague (1992). LIFE HISTORY OF A PACIFIC OCEANIC ISLAND Separating the dual but basically parallel influences of erosion and subsi- dence of oceanic islands has been possible only with sophisticated new technology that involves deep-sea drilling and sonic imaging. Variations Geology and Biogeography in sea level further complicate the picture, compared with subsidence. From although this perspective the is 23 a small effect of the terrestrial biogeographer, however, the important point in this discussion is the gradual northward motion and reduction to sea level or below of former volcanic oceanic islands that once may have had an abundant and diverse terrestrial biota. may Such islands conveyor belt. The be viewed as having a motion like that of a island moving away from the hot is constructed by active volcanism and, before spot, For a period of several million may attain substantial mass and altitude. years, more or less, each new island can and serve as an active substrate for colonization agules may arrive diversification. Prop- by long-distance dispersal but are more from an adjacent older island. In the Flawaiian Islands, the island chain in motion toward the northwest likely to come one may visualize and most of the colonists moving southeast, in the opposite direction. As the island moves away from the source of magma, it becomes more stable biologically. At middle age, it may approach a biogeographic equilibrium of coloniza- manner proposed by MacArthur and Wilson (1967). Following this, as the island is further reduced by erosion and subsidence, this equilibrium will become perturbed and the overall biota gradually reduced by extinctions for which colonizations can no longer compensate. As old age sets in and the island approaches sea level, the diversity of terrestrial species would be expected to slowly become depauperate before submergence is complete. This is true for the tion and extinction in the Hawaiian archipelago because of its northerly location, which is marginal for coral growth. Other islands in more equatorial locations should become atolls that persist for long periods (e.g., the Marshall Islands). All such islands eventually end up below sea fate of an oceanic island calls for equilibrium theory, as suggested by From data on level. This view of the eventual a revision of the MacArthur- Wilson McKenna (1983) and Carson (1992a). the distribution of coral reefs, Darwin (1837) deduced that the subsurface platforms and seamounts in the Pacific were sunken islands. Modern study of the guyots of the western Hawaiian Ridge and Emperor Seamounts confirms these ideas, which have also been applied to seamounts found on the easternmost portion of the Galapagos Ridge (Christie et al., 1992). The implications of subsidence for the distribution of marine life, as Darwin recognized, are very great, because a sinking island in tropical waters acquires coral reefs. trial life. The main effect, however, is an irreversible loss of terres- CARSON AND CLAGUE 24 TERRESTRIAL BIOTA OF THE PRESENT HIGH HAWAIIAN ISLANDS The is modern high Hawaiian Islands world biogeography. The importance of diversity of the highland biota of the surely one of the wonders of the rapid rise and eventual subsidence of islands for the interpretation of island biotas can hardly be overestimated. their biota are If the Hawaiian Islands and viewed wholly from the perspective of graphic position, the biogeographer may their present geo- overlook the northwestern low and assume that the progenitors of any endemic or indigenous element must have arrived in Hawaih from fringing continents or islands biotic distant island sources. Furthermore, the surface of the present high waiian Islands is geologically youthful years), so that the recency of the (i.e., no older than 5 unique living species is Ha- to 6 million indeed striking, no matter where the founding propagules came from. A case can be tion by propagules made for some recent (i.e., less than 5 Ma) coloniza- from remote continental sources. With the advent of the use of molecular markers, this possibility can be tested; examples are presented in this volume. Nevertheless, the ancient high islands of the Hawaiian-Emperor Chain need to be considered as a source of propagules for at least some of the immigrant lines found on the present Hawaiian Islands. As has been indicated earlier, present low Hawaiian Islands is the endemic terrestrial biota of the depauperate compared with that of the high islands (Table 2.2), so that one cannot realistically search the present TABLE 2.2. Taxon or island Approximate Numbers of Endemic Species Biota of the High and Low Hawaiian Islands characteristic Southeast High Islands (Kaua‘i to Hawai‘i) in the Terrestrial Northwest (Nihoa to Kure Atoll) Insects 2,300 50 Land Land 1,000 8 70 4 120 850 4,340 0 snails birds Ferns and allies Flowering plants Total Area (km^) Endemic species/10 km^ 16,576 38 Low Islands 12 74 8.29 1 25 Geology and Biogeography low biota of the endemic high-island islands for clues to the origin of forms. Direct evidence, however, exists of ancient terrestrial biota. This has been obtained from cores that were drilled as part of the Deep Sea Drilling Project. In 3 of 46 samples taken from 20 cores from Koreneva (1980) reported this project, leg 55 of single spores of the fern families and Polypodiaceae, two pollen Pteridaceae, Schizaeaceae, Cyatheaceae, gymnosperm family Pinaceae, and one pollen grain from an angiosperm. This material was obtained from hole 43 3 B, drilled in Suiko Seamount in the Emperor Chain. This volcano was grains belonging to conifers of the determined to have a potassium-argon age of 64.7 million years (Dalrymple et ah, 1980). Although this evidence appears to support the theory that substantial islands with fernlike vegetation existed as long ago as the Paleocene, the possibility exists that these spores and pollen grains have arrived by long-distance dispersal. A rain of conifer pollens and fern spores over great distances, including oceans, 1962; Hirst et al., may is known (McDonald, well 1967). OLD LINEAGES AND NEW SPECIES High-altitude islands have existed in the present position of the Hawaiian Islands much of the time since the late Paleocene. islands harbor biota that contributed in high-altitude Hawaiian To what extent did such some way to that of present-day, Islands? In considering this question, a distinction must be made between immigrant lines of descent (lineages) that have evolved in situ (autochthonous species). present-day organisms may be traced back A it species line of descent of many millions perhaps to the Precam brian!), but at any one time, and of years (indeed consisted of a series of distinct species. These are basically genetically variable populations living under natural selection. Long-surviving individual species that remained unchanged over geologic time (“living exist, are likely to constitute When we examine ago, species endemism fossils”), if have they indeed only a small fraction of the total biota. the present-day terrestrial biota of the archipelis very high. Many organisms form clusters of phylogenetically closely related species. These clusters often have representatives In on all many or most of the present islands. cases, the individual species that make up these phylo- genetic clusters are found to be endemic to one island or even to individ- ual volcanoes. The conclusion have newly evolved in is usually drawn situ since the island or that such endemic species mountain was formed. An CARSON AND CLAGUE 26 alternative theory, vicariance, holds that these species might have pleted the speciation process on an older since disappeared. Accordingly, this is unchanged descendent from these older forms. This might, it is immigrant is older than the island a in on found. High species endemism on the present islands speaks against the vicariance view, as will now lines are relatively old, demism suggests rapid migrant lines. ENDEMISM A biota, has view contends that a species observed turn, lead to the conclusion that the species which its day might be considered to represent a taxon that at the present basically island that, with com- very large be discussed. At least some of the but the high degree of species en- and autochthonous speciation within these im- THE HAWAIIAN ISLANDS IN number of species have been accidentally or purposefully introduced into the Hawaiian Islands since the arrival of Polynesians about 2,000 years ago and Europeans a few hundred years ago. Most of these species are recognizable as recent introductions and would generally be excluded in a biogeographic study to concentrate attention on the species that arrived before identification of the fails endemic biota, disregarding an interesting problem to address change. Thus, humans. Although a necessary process for the it in the human introductions dynamics of evolutionary possible that significant genetic changes is may have occurred in their populations during the few hundred years since modern introduction. sets How occur? This is rapidly can evolution of new and character species an unexplored problem that greatly needs attention from the population geneticist. After eliminating the introduced species, one is left ments of the biota that are either indigenous or endemic with the ele- (see discussion by Carson, 1987a). The former designation characterizes species that colonized without the intervention of archipelago as well as in humans and some other live naturally in place or places. Some the of these indigenous species are strand or shore organisms that have wide distributions in the Pacific, being somewhat comparable with certain widespread marine organisms. Of greater significance, however, are the species or genera that are endemic (i.e., entirely restricted to the present islands and not naturally found elsewhere). In the Hawaiian biota, high levels of endemism occur in many groups of example, more than 90% of the species related species; are endemic. in insects, for Geology and Biogeography The vicariant explanation does not fit 27 well in the case of approxi- mately 100 intensively studied picture- winged Drosophila species (Car- son and Kaneshiro, 1976; Carson, 1983a, 1990a; see also Kaneshiro al., this volume. Chapter 5). These species are all et endemic to the existing high islands. With only a few exceptions, tracing by the use of chromo- somal markers indicates that a succession of new single-island endemic species have evolved as each new volcano and the southeast of an older one. plants of the island has been A similar pattern Hawaiian silversword of evolution is formed to shown by alliance (Asteraceae, Madiinae) which comparable crucial genetic data have been obtained (Carr 1989; Baldwin et al., 1991). In on et al., both cases, although the bulk of the colonizations has been from an older island to a newer one, colonizations in the reverse direction The reliability have also been recognized. of phylogenetic information provided by specific genetic markers, however, tends to be negatively correlated with the length of time since the cladistic event. In the case of the picture-winged Drosophila and the Hawaiian silversword alliance, all newer islands may be traced back genetically to species Kauah, the oldest high island. This does species still on the existing on not mean, however, that a continental ancestor of the observed line of descent must necessarily have originally colonized may have island. Kauah, an island formed 5.1 Ma. Ancestral forms colonized the present archipelago by Such an event is way of an older eroded indicated by molecular studies of the relationships some present-day forms. Although the picture-winged flies seem to form a very recently evolved group of species, a closely related group of Drosophila species that breed on fungus, also endemic to the islands, have DNA sequences that diverged from the picture-winged group about 10 Ma, twice the age of Kauah (Thomas and Hunt, 1991). These authors suggested that the divergence between the fungus feeders and the picturewinged group could have occurred on an island such as 10 million-yearold Necker, which is now reduced to the point that it does not support of any Drosophila species. SOURCES OF PROPAGULES FOR THE HIGHLAND HAWAIIAN BIOTA Although the vicariance idea has its advocates (see Melville, 1981), distributions interpretable as vicariant are infrequent in the terrestrial biota of the Hawaiian Islands. The present biota appears to trace to waifs 28 CARSON AND CLAGUE (Wagner et al., 1990). All such founding waifs, however, have not neces- come directly from the distant continents. With proper genetic markers, we may ultimately be able to recognize two kinds of waifs: those sarily originating on the continents or distant islands, and those that were derived from pre-existing, but now-foundered older islands. Propagules from older islands in the chain were probably not continuously reaching newer islands. After the formation of Koko volcano in Emperor Chain about 48 Ma, no islands higher than 1,000 m formed until Kure about 29 Ma. By the time Kure formed, all the previous high islands had subsided. Thus, there was at least one time in the southern when was no possibility to derive propagules from a previous island (which was already an atoll with depauperate biota) and the entire process of introducing waifs from far away would have rethe past there Hawaiian Bend, the volcanic started. Before the time of the a few exceptions, were high for only a short time, and it is the process of colonization from older to younger islands several times during the Paleocene The isolation of the must look to the islands, with possible that was cut off and Eocene. Hawaiian-Emperor Chain fringing continents and other is so great that one island groups as the ultimate source of the biota by long-distance dispersal (Carlquist, 1980). Data on the wide Hawaiian biota dispersibility of (e.g., reinforce this view, many of the immigrant lines of the Eosberg, 1963; Gressitt and Yoshimoto, 1963) and there is reason to believe that propagules from the continents have continually played a role in both ancient and times. Moreover, some of the island groups in the Pacific have served as “stepping stones” for immigrant lines modern (e.g., Eiji) might from more remote no evidence, however, for the existence of any “lost” or sunken mid-Pacific continents. Rotondo et al. (1981) cite the existence of two older seamounts near Necker and Kure atolls. They islands or continents. There is suggest that these could have contributed to the Hawaiian biota. ever, this How- seems unlikely because these seamounts were not islands at the same time as any nearby Hawaiian islands. The question becomes one of deciding from where and how long ago arrival in the Hawaiian-Emperor Chain took place. With molecular methods, precise identification of putative ancestral stocks for groups and species is separately lar now possible. Each immigrant group of examined from methods interest needs to be this point of view. If carefully applied, molecu- for estimating time since divergence should provide hard data indicating when such an event occurred. 29 Geology and Biogeography ISLAND OF HAWAri-»A PARADIGM OF “MICROBIOGEOGRAPHY” Hawai^i Island species and is both very large and very young. sets of characters that are biogeographer may endemic to lar also contains many this island alone. The thus use the species of this island as examples of dynamic and recent evolutionary process on a scale (Carson, It 1983b; see also DeSalle, markers such as mitochondrial this geographic relatively small volume, Chapter 6). Molecu- DNA (mtDNA) can provide historical data of extraordinary interest even for different populations of a single species, a field that has been called intraspecific phylogeography (Avise et ah, 1987). The newness of Hawaih reconstruction of its Island has permitted a detailed geologic development through a succession of volcanoes of varying very recent ages. Moore and Clague (1992) have traced the origin of the present five volcanoes of this island from their beginnings about 0.5 Ma, reached Mahukona volcano Kohala 245 ka, Mauna Kea 130 ka, using ancient shorelines (Figure 2.3). its largest size about 465 ka, and Hualalai 130 ka. The picture is completed by including the currently and very recent Mauna Loa and Kilauea. In view of these new data, the opportunities for tracing and timing the recent evolution and colonizactive ing history of selected indicator organisms with great accuracy are unparalleled. Even beyond the biogeographic genetic shifts most situation, it and formation of new character active at the time a new shield volcano is has been suggested that and species may be growing (Carson et ah, sets 1990). Single founder events of one or a few sexual propagules that are followed by immediate expansion of the new population to a large not appear to necessarily result in inbreeding depression or loss of icant quantitative genetic variability from the new population (see size do signif- review by Carson, 1990b). The as far as rise and fall of islands in an extended succession, back at least 65 Ma, appears to have mediated by stepwise founder of population sizes. Many elicited continued novel evolution effects separated by periods of expansion different lines of descent appear to have been affected by such events in a similar manner, indicating that many organ- isms have the capacity to retain extensive genetic variability despite occasional bottlenecks of population size. Cladistic V. A. The authors in this Methods FUNK volume have used the methods of phylogenetic and examine a rigorous way. Thorough systematics, also called cladistics, to develop phylogenies monophyletic groups (referred to as cladesY^ in explanations of cladistics can be extremely complex. This discussion not intended to be comprehensive; rather it is is an introduction to the concepts and terminology necessary for the reader inexperienced in phylogenetic theory to understand the analytic aspects of the chapters in this volume. Additional discussions can be found in Hennig (1966), Nelson and Platnick (1981), Wiley (1981), Swofford and Olsen (1990), Wiley et al. (1991), Forey et al. (1992), Maddison and Maddison (1992), Swofford (1993), and references cited therein. Cladistics seeks to of more than answer the following question: Given any group three taxa, which taxa are more another than to any other taxa? Relatedness is closely related to identified one by the sharing of one or more uniquely derived characters that other taxa outside the group do not possess. For example, within vertebrates the unique derived char- acter “feathers” identifies other. all birds as being The branching pattern of most closely related to each the tree that illustrates this relatedness formed by the distribution of the unique characters ^Most of the clades identified in this for emphasis, are given in italics name). 30 in the way is that book do not have formal taxonomic names and, and without capitalization (unless derived from a proper Methods Ciadistic FIGURE 3.1, whereas ters represent taxa, Group Y Group X Cladogram. Let- 31 let- ters in brackets are hypothetical ancestral taxa. sent Numbers repre- apomorphic characters of the transformation series; those with single bars are apomorphic, and those with double bars are independently derived. Group Group is paraphyletic, a grade. is monophyletic, a clade. requires the least ter loss. A tree cladogram but amount formed is X Y of convergent or parallel evolution solely and charac- by these unique characters can be called a also called a phylogenetic tree or tree (Figure 3.1). Cladograms are characterized by the fact that their information is con- tained in the branching sequence and not in the physical proximity of the terminal branches. For instance, Figure 3.2 shows the same branching sequence as Figure 3.1, and as far as information content is identical. In Figure 3.1, B is is concerned, next to D, but in Figure 3.2 B is next to it F. Neither of these physical locations gives the correct relationship because the branching sequence of both figures is shows that the actual relationship one of B being most closely related to the group of taxa discussion FIGURE on Venn diagrams below). 3.2. Cladogram with the same branching se- quence and the same information content as Figure 3.1. B A cladogram F in DGF (see which the branch G D FUNK 32 TABLE Character Matrix for Figures 3.1 to 3.3 3.1. Transformation series Taxon"^ 1 2 3 4 5 6 7 OG 0 0 0 0 0 0 0 B 1 0 0 0 0 1 0 D 1 1 0 0 0 0 1 F 1 1 1 1 0 1 0 G 1 1 1 1 1 0 0 [A] 1 0 0 0 0 0 0 [C] 1 1 0 0 0 0 0 [E] 1 1 1 1 0 0 0 D, F, G, and the OG (outgroup) are actual taxa, whereas A, C, and E are hypothetical taxa whose character data are inferred from the most-parsimonious and internode lengths internode is reflect the number tree. of characters on that branch or called a phylogram. Cladistics has as its basis three concepts: and parsimony. An apomorphy is apomorphy, monophyly, a uniquely derived evolutionary charac- apomorphous characters, but various other permutations of the term now include apomorphic character and apotipic. There are related terms; for instance, every apomorphy either is found in one taxon, an autapomorphy (Figure 3.1, apomorphic characters 5 and 7; Table 3.1), or is shared by more than one taxon, a synapomorphy (Figure 3.1, apomorphic characters 1 to 4). A synapoter. Hennig (1966) morphic character, called these in the true sense, is ancestor of a group of taxa marking a that group. Every from which it is one that has evolved once common apomorphous character is in the evolutionary history for paired with the character derived, the plesiomorphous character (or plesiomorphic character or plesiomorphy). In the bird example, “feathers” is the apo- morphic character, and because feathers are believed to be derived from scales, then “scales” becomes the plesiomorphic character. The apomorphic and plesiomorphic characters together form an evolutionary transformation series (often abbreviated TS) (Hennig, 1966; Wiley et al., apomorphic 1991). The transformation series can contain more than one character, provided they are evolutionarily Some authors homologous. and trans- in this book. refer to individual characters as character states formation series as characters, However, this alternative and both systems are used terminology necessitates placing apomorphic Cladistic and plesiomorphic character mation series. states into characters rather Methods 33 than transfor- Unfortunately, users of the term character state sometimes incorrectly shift to the term character in the discussion section. unambiguous, the transformation series concept is To be preferred. The terms apomorphy and plesiomorphy are dependent on their relative position on a cladogram. A character that is synapomorphous at a node when one is discussing group Y (Figure 3.1, apomorphic character 3) will be plesiomorphous that delimit taxon G. When if one is discussing the characters characters are found in more than one taxon, they are considered to be evolutionarily or phylogenetically homologous (Patterson, 1982, 1988). If what appears to be the same apomorphic character is found in two unrelated groups, it is considered to be nonhomologous and therefore not a single apomorphy (Figure 3.1, apomorphic character 6) and is referred to as a homoplasious character. If a character occurs as a synapomorphy on a cladogram and is subsequently lost in one or more taxa, then it is a character loss (also referred to as a reversal, but this term can be confused with genetic terminology). Homoplasious characters and character losses may obscure the phylogenetic pattern. These seemingly con- tradictory characters are referred to as character conflict. Such conflicts are resolved by parsimony analysis, and once they are recognized and understood, become apomorphic characters themselves. The parsimony It is criterion governs how cladograms are constructed. nearly identical to Flennig’s Auxiliary Principle: “Never assume convergence or parallel evolution, always assume homology in the ab- sence of contrary evidence” (Hennig, 1966, according to Wiley et al., 1991; Farris, 1983). This principle does not preclude the possibility of convergent or parallel evolution; it simply states that when there is no reason to think otherwise, two characters that appear to be the same are means that the character has the potential for grouping taxa if it is apomorphous. When characters support conflicting groups (Figure 3.1, apomorphic character 6), the explanation that is the simplest is chosen (i.e., the one that requires the smallest number of homoplasious characters and character loss). Therefore, the user of parsimony is not making any statement about the process of evolution. A monophyletic group is a group of taxa that share a common ancestor and includes all descendants of that ancestor, also referred to as a clade. On a cladogram, this translates into any group that includes all taxa that share at least one synapomorphy (Figure 3.1, group Y). Figure 3.3 is a Venn diagram for Figures 3.1 and 3.2; each ellipse represents treated as homologous. This FUNK 34 FIGURE 3.3. gram of Figure Venn dia- 3.1. a monophyletic group so that one can easily see three such groups, DFG, and the whole clade, BDFG. However, far more than a definition of a the concept of FG, monophyly group of taxa. Concomitant with it is is the notion that the only groups that are evolutionarily meaningful (natural) are monophyletic ones. Therefore, in this view, the only groups that can be recognized in formal classifications are monophyletic ones. cation for this position an ancestor and reflect a all common lies in its the nature of the groups. groups include evolutionary history and can be used to study specia- and other evolutionary concepts. Non-monophyletic groups are of two types is justifi- descendants (monophyletic), then the groups tion, biogeography, pollination biology, A If The the ancestor of taxa B, D, F, (Farris, 1974). In Figure 3.1, and G. Group X contains the common ancestor A, but only two of the descendants, B and D, and so monophyletic. Such a group, one that includes some but not descendants (Figure 3.1, group X), is called paraphyletic, it is all which not of the is also referred to as a grade. Polyphyletic groups have been defined several ways, but, in general, they consist of taxa taken from more than one monophyletic group. Under certain circumstances, rate paraphyletic to it is difficult to sepa- and polyphyletic groups, so often authors simply any group of taxa that does not satisfy the criterion of refer monophyly as non-monophyletic Both monophyly and parsimony depend on apomorphous characters; therefore, apomorphies are the central concept of process of assigning the status of apomorphy to a character determining polarity. Using an outgroup (or outgroups) mon way cladistics. is the is The called most com- of determining which characters are apomorphic (Watrous and Wheeler, 1981; Farris, 1982; Maddison et al., 1984). Characters found in some of the taxa of the group being studied (the considered to be plesiomorphous. Those characters found the outgroup as well as in ingroup) are only in some of the taxa of the ingroup but that are absent in the rest of Cladistic the ingroup Many and in the Methods 35 outgroup are considered to be apomorphous. exceptions and extenuating circumstances to be considered when using the outgroup criterion cannot be covered in this brief discussion. Additional information can be found in the general references listed in the first paragraph and in Watrous and Wheeler (1981), Farris (1982), and Maddison et al. (1984). An outgroup can be, but is not necessarily, the taxon most closely related to the ingroup, the sister group. In Figure 3.1, DFG is the sister group of B. The outgroup(s) should be a closely related taxon that does not contain large numbers of autapomorphous characters. Sometimes a specific outgroup cannot be identified, and a composite outgroup is constructed by evaluating each transformation series sepa- rately to determine which character (s) was apomorphic. Authors in this volume who use this approach have discussed how the composite outgroups were formed. Another method that is occasionally used to assign polarity is ontogeny (Patterson, 1982). The process of tree construction has changed greatly in the decades. Instead of the manual constructing of small character each transformation to decide if it is series, which necessitates struct now two trees for examining each character apomorphous and then looking can be nested, computer programs are past for groups of taxa that used. These programs con- networks based on the distribution of shared characters without assigning polarity or evolutionary direction, then root the tree based on the characters present in the outgroup(s), either by using the outgroup(s) as part of the analysis or is by attaching completed. The two most ford, 1993) and HENNIG86 increased the speed these it to the network after the analysis commonly used programs (Farris, 1988). are PAUP (Swof- These computer programs have and accuracy of cladogram production. Moreover programs have introduced many options that give the user a power- ful resource for investigating the phylogeny of the taxa in question. Another program available for analyzing characters and cladograms MacClade (Maddison and Maddison, 1992), which array of options. On is also has a broad occasion, different programs will give different answers to the same questions. It is the user’s responsibility to make sure she or he understands and endorses the assumptions that underlie the options in all the programs; otherwise, the results will be misleading (at best) or erroneous. For years, many phylogeneticists have tried to measure the robust- ness of data used to construct cladograms, to find a that would measure is indicate how way to assign a value robust” the cladogram was. The simplest the tree length, or total number of steps. The tree length is FUNK 36 number of characters actually on the tree, including all characters. The first index, and still the most popular, is the equal to the total conflicting consistency index (Cl) (Kluge and Farris, 1969). Currently, the index is and taking the minimum number the data agreed and dividing it by the actual calculated using only synapomorphies of steps necessary number of steps. if all The other commonly used index is the rescaled consis- tency index (RC) (Farris, 1989), v^hich multiplies the Cl by the retention index (RI; ratio of apparent synapomorphy to actual synapomorphy). The RC excludes characters that do not contribute to the “fit” of the tree by excluding autapomorphic characters as well as totally homoplasious The Cl and ones. RC the can be used for each individual transformation series (character) as v^ell as the proposed indices have been 1991) but are not used cladogram as a whole. Several other (e.g., in this F-ratios, d-measures) (Wiley et ah, volume. Each index has certain strengths and weaknesses, and no one index has been found that really gives us the information we seek, the answer to the question “Flow good is this cladogram?” Whereas the indices give information the individual transformation series), there on is the tree as a whole (or another approach to on esti- mating the value of a particular cladogram with respect to the data and that by placing confidence is limits on the individual branches. Some authors provide such values based on bootstrapping. This technique involves randomly sampling with replacement the character information from a data set to build the original data set, many “bootstrap” data sets of the same size as which are then analyzed to give one or more The percentage of occurrences trees. (usually out of 100) that a particular monophyletic group appears among the trees of the sample data sets can be considered an index of support for that monophyletic group. This technique does not result in true confidence limits in a One of the biggest problems the data set. Also, confidence level of plasious char- cters tional it is statistical sense. that the values can be related to the size of takes three synapomorphies at an internode for a 95% and these could all be homothat occur many times on the tree. There are addito be reached, problems with the assumptions required by bootstrapping that can cause either over- or underestimates of confidence (for further discussion, see Sanderson, 1989). As data sets grow, there is an ever greater chance of the analysis resulting in more than one equally parsimonious tree. A method of working with multiple trees is the implementation of consensus trees (Wiley et al., 1991; Swofford, 1993). Two types of consensus trees are Cladistic common in the literature, strict and majority only the groups that are found in reflect 37 consensus trees rule. Strict all Methods the equally parsimonious Majority rule consensus trees show the branching sequences that trees. are found in most of the Both consensus trees. tree methods have the potential of producing unresolved areas or branching patterns on the consensus tree that are not found in any of the equally parsimonious Although consensus trees. trees are useful in identifying the areas of agreement and conflict among the competing tree is identical to as a phylogeny one of the equally parsimonious result of conflicting in trees, it beyond the point of agreement found consensus cannot be used in all trees. more than two branches) instance, polytomies (nodes with found trees, unless a For that are the branching sequences in competing trees and are not any of the competing equally parsimonious used as part of the phylogeny. One should trees should not be consider selecting one of the equally parsimonious trees for use as a phylogenetic tree. Another option was used in this book is successive weighting based on the fit of the characters to for dealing with multiple trees that weighting, an a posteriori the trees (Farris, 1989; Swofford, 1993). There are several types of a priori weighting as well, but none were used by the authors in this volume. When many equally parsimonious trees are produced, especially with molecular data sets, the methods of bootstrapping and majority rule consensus trees are often combined to produce a must be used with such a relationship Once it tree, for there is tree. no way Extreme caution to gauge what holds with any of the equally most-parsimonious trees. a phylogenetic tree has been produced, one of the most interesting things to do with ability to ask questions it is to use it about evolution interested in producing phylogenies in the to study evolution. Indeed, the why many researchers are first place. One technique used is book to facilitate such evolutionary studies is optimization or mapping. The method is examined in detail in Funk and Brooks (1990), Brooks and McLennan (1991), and Maddison and Maddison (1992); a simplified explanation is offered here. Once a cladogram has been constructed, any feature or condition is selected to be examined in the light of the phylogeny of the group. Examples include habitat, habit, chromosome number, and home range. The condition of each terminal taxon is identified on the cladogram, and hypothetical conditions are assigned to in this the nodes that reflect the most-parsimonious arrangement of those conditions at each node. This allows conditions. In this volume, the one to determine the potential ancestral method is primarily used to examine FUNK 38 biogeography, but other features examined include speciation and habitat evolution as well as adaptive radiation and coevolution. features, only biogeography has its own special term: which the terminal taxa have been replaced by tions is called an area cladogram. Phylogenetic systematics changing field of study. hopes that the reader is A Of all these cladogram in their respective distribu- an interesting, growing, and constantly This brief discussion will be able to better is an introduction in the understand the chapters in this volume. ACKNOWLEDGMENTS I thank Paul Manos, Francois Lutzoni, Peter Cannell, and Warren L. Wagner for reading and commenting on drafts of the manuscript. Their willingness to offer criticism should not be taken to infer their acceptance of its contents. Biogeographic Patterns of Two Independent Hawaiian Cricket Radiations (Laupala and Prognathogryllus) KERRY The Hawaiian archipelago unique biological is L. well SHAW known to evolutionary biologists for diversity, particularly for its native insects (Zimmer- man, 1948; Howarth and Mull, 1992). Because of the successive emergence of new islands and habitats caused by the northwestern migration of the Pacific plate over a relatively stationary magmatic plume (Stearns, 1985), colonization opportunities are continually created for Hawaiian Compelling evidence biota. some groups such as the native Templeton, 1984) due to ila in the age and diversity Hawaiian geochronological pattern of island formaislands are hypothesized to and perpetuate the lineage of native Drosoph- northwest on the Pacific plate. In other Hawaiian which enduring and extensive diversity exists, the geologic development of these islands might have a similar relationships within those radiations. in concert means of diversification in Islands, in the face of island subsidence as the islands drift to the radiations in this mode Hawaiian Drosophila (Carson and from older to younger tion. Colonizations renew taxonomic exists for a special A effect on the pattern of phylogenetic frame of reference with knowledge of the geologic history of a region provides a for inferring the biogeographic history of a group, analysis is and this type of discussed here for a portion of the diverse native Hawaiian crickets (family Gryllidae). Many attributes of the native cricket fauna analysis attractive. diversity arising A relatively make phylogenetic thorough treatment of the extant cricket from colonizations of the Hawaiian Islands is possible 39 SHAW 40 because likely that all it is taxa arising in the Hawaiian cricket radiations communicate acoustically, distribution maps are easily compiled because the male calling songs are audible to humans and can often be heard from considerable are contained within the archipelago. In species that distances. Furthermore, a general phenomenon of single-island endemism pervades in the native cricket fauna, Perkins (1899, p. 3) noted that “the number is of species which quite remarkable, fail more to extend their range so, I believe, than other orders of insects.” In a recent and is beyond a single island the case with any of the more thorough taxonomic Hawaiian crickets, Otte (1994) found that all native species except Caconemobius sandwich ensis are single-island endemics. Hawaiian crickets are tremendously diverse, both in species richinvestigation of the ness the and first way in of life. Among closely related Hawaiian cricket species, distinguishable differences tend to be behavioral phenotypes involved in reproduction. Minimally, species hypotheses rely on the dif- ferences in the male calling song in groups that possess forewings (Otte, 1994). All native crickets exhibit classic “island flightlessness” (Carlquist, 1980; Williamson, 1981). However, (tegmina), many species retain the forewings which function as male sound-producing organs to which females respond. Only more distantly related species show differences in morphology of the male genitalia, size characters, and pigmentation patterns. Species that do not possess forewings and, therefore, do not sing are described largely on the basis of morphometric differences in the male genitalia. There have been three substantial and separate radiations into the Hawaiian Islands by the ground crickets (Nemobiinae), the tree crickets (Oecanthinae), and the swordtail crickets (Trigonidiinae) (Otte, 1989, 1994). Like so many other Hawaiian plants and animals (Carlquist, 1980; Simon et al., 1984; Carr et ah, 1989), adaptive shifts have occurred within each of these radiations. The nemobiines Thetella, and the conemobius is two genera, a widespread Pacific genus, genus Caconemobius (with 9 endemic species). Cainclude noteworthy for adaptive, potentially parallel shifts of from a shore form that inhabits the rocky coastal environment on all the main islands. Cave-inhabiting species, which exhibit eye reduction and a loss of pigmentation, have been discovered on Hawai‘i and Maui (Howarth and Mull, 1992; Otte, 1994). Only the lineages into lava tubes most diverse on the youngest islands (Table 4.1). This pattern may result from a loss of rocky and subterranean habitat as the older islands erode and subside. Caconemobius radiation is 41 Crickets TABLE Taxonomic 4.1. Diversity by Island in the Endemic Hawaiian Cricket Genera No. of species of: Prognathogryllus Thaumatogryllus Caconemobius Leptogryllus Trigonidium Prolaupala Laupala Island^ Nihoa 1 Kaua‘i 2 3 0‘ahu 1 7 Moloka‘i 5 Maui 2 2 Lana‘i 1 1 Hawai‘i 6 5 1 16 13 1 7 4 1 1 11 13 1 1 30 4 1 8 7 1 27 3 1 27 34 7 Note: All species noted here, except one widespread species of Caconemobius are single-island endemics. ^There are no known endemic species of crickets on the island of Kaho‘olawe. The tree crickets are represented in the Hawaiian Islands by a radiation that has resulted in three Thaumatogryllus (4 species), Perkins (1899) and later endemic genera: Leptogryllus (28 and Progmthogryllus (36 Zimmerman species) (Figure 4.1). (1948) considered these native genera to be closely alHed but contained within the Eneopterinae. tympana on the gland, species), The presence of foretibiae, vestigial forewings, the structure of the metanotal and the male genitaUa common to Thaumatogryllus and Leptogryllus suggest that these genera share a more recent does with Progmthogryllus (Otte, 1994). gryllus and Leptogryllus hide under bark or close to the ground. Some common Members in ancestor than either of both Thaumato- dead leaves and fern fronds species of Thaumatogryllus Hve in subterranean habitats or lava tubes, whereas members of the genus Progmthogryllus always found in bushes and treetops. Otte (1994) hypothesized that bers of Progmthogryllus form a monophyletic and Thaumatogryllus. Species greater diversity in on older islands, a pattern that is sister are mem- taxon to Leptogryllus Progmthogryllus is generally not consistent with the other genera in this radiation (Table 4.1). The Hawaiian ets, trigonidiines comprise the largest radiation of crick- with species classified into three genera: Trigonidium (135 species), Prolaupala (3 species), and Laupala (35 species) (Figure 4.2). Scudder (1868), Brunner (1895), Perkins (1899), and Zimmerman (1948) consid- SHAW 42 FIGURE 4,1. Prognathogryllus robustus (drawing by D. Otte). ered all the native Trigonidiinae to be allied to the genus Paratrigonidium, Hawaiian Trigonidium In Otte (1989), Species number although there is is species are referred to Anaxipha. roughly equal across the main islands (Table 4.1), a slightly higher number of on the middle-aged monophyletic. Laupala can be species The genus Laupala is clearly distinguished on the basis of the structure of the male genitalia and the female ovipositor and by increased venation in the lateral field of the tegmina (Perkins, 1899; Otte, 1994). Members of Prolaupala and Lauislands. pala are hypothesized to share an exclusive 1994; Shaw, 1993). They share common common ancestor (Otte, 1989, terrestrial habits slow-pulsing, diurnal singing behavior. According to p. 135), “It myth of is and exhibit Zimmerman the chirping of these crickets that gave rise to the the singing land snails.” Whether (1948, Hawaiian species of Trigonidium are a monophyletic or paraphyletic assemblage with respect to Prolaupala and Laupala is unclear. The predominance of suggests that inter-island migration with the observation of extensive in endemism in the native crickets rare. This phenomenon, combined single-island is diversity, offers appropriate conditions which Hawaiian geologic history may have had a on the process of diversification. Also, Otte (1989, significant influence 1994) proposed that speciation has occurred primarily within islands in the Hawaiian crickets. Crickets 43 / # FIGURE 4.2. Laupala paranigra (photo by D. Funk). This hypothesis is examined in the tree cricket compared with the the results are genus Laupala. The distributions main islands of the in genus Prognathogryllus, and patterns found in the swordtail cricket both these genera are limited to the eight Hawaiian archipelago (see Table 4.1). Otte (1989, 1994) proposed that the temporal sequence of colonization proceeded from geo- Maui Nui complex) more peripheral islands graphically central, middle-aged islands (0‘ahu or the where the taxonomic diversity is highest to the (Kaua‘i and Hawai‘i). In this chapter, this hypothesis PHYLOGENETIC ANALYSIS species. also addressed. IN P R O G N AT H O G RY L L U S Cladistic analysis of the genus Prognathogryllus variation in morphological is was undertaken using and pigmentation characters in 28 of the 36 These data were given in Otte (1994; reproduced in Appen- dixes 4.1 and 4.2), individuals. In where morphological measurements derive from two many cases, larger numbers of specimens do not collections (specimens are deposited at the Philadelphia exist in Academy of Natural Sciences). Most characters were measured on male specimens. Values for characters 11, 13, and 14 (see Appendixes 4.1 and 4.2) are missing for many taxa; excluding these characters had no effect on the SHAW 44 biogeographic patterns inferred from the most-parsimonious Some species were set of trees. out of the analysis because no mature male left specimens exist in collections. Species not included in the analysis are P. haupUy pus, P. P. wahiawa, P. makakapua, and Variation pararobustus, P. P. giganteus, P. polani, P. Olym- aphrastos. among Hawaiian cricket species is largely quantitative in nature. Differences scored distinguishing species of Prognathogryllus con- body size characters and pigmentation intensities. The character variation was coded in up to 10 states. Many characters were scored as sisted of multistate, with the assumption that these are polymorphic characters at the species level. Twenty-six characters were used in the present analysis, the number being limited by the extreme similarity among members the genus. This type of data, in addition to temporal song characters, of was used by Otte (1994) for species delimitation. Temporal song characters tend to be evolutionarily labile in crickets, particularly in the Hawaiian Islands (Otte, 1993; Shaw, 1993). The Song data were therefore not consid- and limitations of the using morphological data are discussed in more detail below. ered in the present analysis. benefits analysis Character distributions in Prognathogryllus and the nearest postulated outgroup (Xabea or Neoxabea) (Otte, 1994) provide little op- portunity to polarize the phylogenetic relationships. Therefore, a phylogenetic root for Prognathogryllus was not estimated by outgroup Hawaiian Islands forewings. Walker and Gurney comparison. Otte (1994) points out that only are there tree crickets with vestigial in the (1967) found the characteristics of the metanotal gland to be informative in systematic research. Thus, polarity of the cladogram was estimated by considering the high position of the metanotal gland and maximum forewing length to be ancestral (characters 8 and 10, respectively; see Appendixes 4.1 and 4.2 for description of the characters and the data matrix). Parsimony analysis was conducted with PAUP version mony algorithms are unfeasible with this many (OTUs) and the available computer power, tree space dom (Swof- 1990b) on a Quadra Macintosh computer. Although exact parsi- ford, units 3. Or operational taxonomic I sampled the possible by executing 1,000 replicate heuristic searches using the ran- addition option. All characters were designated as linearly ordered. The Lundberg rooting option in PAUP was used. The multistate characters were designated as polymorphic, as opposed to “uncertain.” Coding polymorphic characters as “polymorphic at the species to “multistate,” had some effect level,” as opposed on the topology but not on the biogeo- 45 Crickets alapa* awili* kahea* opua”^ alternatus* hana * kipahulu* o T3 waikemoi* I weli"^ C kohala • * kukui mauka 00 3 c T3 * puna* spadix * stridulans* makai”*^ oahuensis* elongatus epimeces erg o' O £ ^ kahili parakahili • alatus flavidus hea og i-t o, 5^ o c robustus § s pihea victoriae FIGURE 4.3. Strict maximum-parsimony cies consensus of 126 parsimony trees resulting from the analysis of the Prognathogryllus morphological data. Spe- groups are depicted to the right of the terminal taxa. opua group species; small closed circles refer to stridulans graphic implications discussed below. length trees above the An asterisk is shown 95% A strict group refers to species. consensus of 126 minimal The consistency index (Cl) is 0.531, for random data with an equivalent in Figure 4.3. confidence limits number of taxa and characters given by Klassen et al. (1991) in a recent study of Cl and random data. With a retention index (RI) of 0.725, it 46 SHAW alapa alternatus awili kahea opua hana kipahulu kohala kukui makai mauka puna spadix stridulans waikemoi well oahuensis elongatus epimeces hypomacron kahili parakahili alatus flavidus hea pihea robustus victoriae FIGURE 4.4. Strict consensus of the set of parsimonious and one-step-longer trees for Prognathogryllus. seems likely that only one tree island of minimal-length trees exists (as discussed by Maddison, 1991). The four species groups in Prognathogryllus designated by Otte (1994) are identified in Figure 4.3. The robustus group (found on Kauah) appears to be paraphyletic, by virtue of elongatus group, also found on Kaua‘i, its is basal position. How^ever, the a monophyletic group. The 47 Crickets Hawaii Maui 0‘ahu ] kipahulu Kauai ! waikemoi well equivocal BffiHfflffljB i mi mi i i kohala ffl IS mauka puna spadix stridulans makai oahuensis II elongatus epimeces I hypomacron kahili I parakahili II alatus flavidus hea robustus pihea FIGURE 4.5. Historical biogeographic reconstruction in the genus Prognathogryllus. opua group, from the islands of 0‘ahu, Maui, and Hawai‘i, could not be distinguished as a distinct historical group but rather was integrated within the stridulans group, also represented on 0‘ahu, Maui, and Hawai‘i. In the group of next-parsimonious trees (1,237 trees one step longer revealed by an heuristic search), the resolution degrades considerably (Figure 4.4) but primarily within the clade containing the The opua-stridulans group. elongatus group becomes paraphyletic with respect to the opua-stridulans clade, but all Trees two divisions members of steps longer the robustus group maintain their basal position. (more than 4,400 trees) main also retain these between the robustus, elongatus, and opua-stridulans groups. Geographic associations of lineages and patterns of colonization were inferred using the discrete character parsimony algorithm Clade 3.0 (Maddison and Maddison, 1992). Island affinities in Mac- of extant SHAW 48 species of Prognathogryllus terior nodes are presented and estimated geographic in Figure 4.5. All localities of in- Kaua‘i species are found in basal positions with respect to other Prognathogryllus species. treme basal position is occupied by the Kauah taxon robustus group. The basal position of level of trees two steps longer. Kauah taxa is F. ex- victoriae of the robust at least to the O'ahu taxa occupy the next in the phylogeny, a result that is The distal positions unambiguous when considering the most-parsimonious trees and trees one step longer (see Figure 4.4). Furthermore, although comprising an unresolved polytomy, the most-parsimonious trees shows Maui taxa tion (Figure 4,5). This degree of resolution in the is set of next more distal posi- degraded when considering the group of trees one step longer (Figure 4.4). PHYLOGENETIC ANALYSIS LAUPALA IN The swordtail cricket genus Laupala is morphologically the most cryptic group and epitomizes Zimmerman’s (1948) remark that ‘The Trigonidiinae is a systematically difficult assemblage.” Whereas the members of other groups of native crickets may vary to some extent in body proportions, in the presence or and differences absence of pigment patterns, in genitalic file teeth number, morphology, Laupala species show only minor differences of a morphometric nature. Pigment differences show continuous variation from dark to quantitative in nature less dark; genitalia differences are also and are highly correlated with body size. Systematic hypotheses were proposed by Otte (1989, 1994) based on differences in mitochondrial male genitalia and by Shaw (1993, DNA (mtDNA) sequence 16S rRNA, and tRNA'"^* regions. The icantly. The among several analyses explored, a results variation in press) based on from the 12S rRNA, results of these studies differ signif- from the molecular data are discussed here because maximum-parsimony analysis was performed, a higher degree of systematic resolution was proposed, and polarization of the phylogeny and subsequent biogeographic inference was by outgroup comparison. Otte (1989, 1994) hypothesized that the origin of the current distribution of Laupala is concordant with justified the geographic center of diversity of the genus (0‘ahu or Maui) his phylogenetic tree based Figure 4.6 shows a trees, on this strict and roots biogeographic proposition. consensus of eight equally parsimonious generated through a heuristic search routine using PAUP (as above, with 1,000 random addition replications). Thirty-six unique mtDNA Crickets 49 pruna-AF cerasina-AF cerasina-AF kona-HN hualalai-GW hualalai-GW cerasina-KW ] fugax-MLH IQ cerasina-KP ifl IB IB IGI IB ] IQ IB B ] ] ] ] ] pruna-ET paranigra-KW kohalaensis-KH cerasina-Kl paranigra-K2 koIea-M34 nigra-ET nigra-ET cerasina-ET prosea-HR prosea-HR eukolea-HR eukoIea-HR vespertina-MLH vespertina-MLH pacifica-MT pacifica-MT pacifica-MT tantalus-MT hapapa-KKP pacifica-MT |g hapapa-KKP kokeensis-AS kokeensis-KSP keahua-KA Prolaupala t FIGURE Trigonidium 4.6. Historical biogeographic reconstruction in the genus Laupala. Terminal taxon labels indicate the species followed by the population from which a haplotype was sampled. nucleotide sequences, sampled from 17 species of Laupala and 2 out- groups, were treated as the as outgroups, OTUs. Two Hawaiian were were included one from the circumglobal genus Trigonidium and the other from the endemic genus Prolaupala. ters species cladistically informative A total of 74 variable charac- and used to estimate the topologies represented by Figure 4.6 (see Shaw, 1993, in press, for further details). In Figure 4.6, a parsimonious reconstruction of geographic locality was superimposed using the discrete character parsimony algorithm in MacClade 3.0 (Maddison and Maddison, 1992). Only geographic data for the ingroup were used in the reconstruction. The basal position of the phylogeny is occupied by Laupala keahua, which occurs on Kaua‘i. The next clade leads to a group of species that inhabit 0‘ahu, with a KauaT SHAW 50 species, L. kokeensis, in the basal position. As more distal positions are considered, the most-parsimonious colonization pattern remains quite two simple, with the final clades containing groups of species with distributions confined solely to the island of within the island of Maui and primarily found Hawaih but with two back-migrations to Maui. DISCUSSION The most among striking general pattern endemism single-island the native cricket genera (the only exception being Thetella tarnis). Perkins phenomenon (1899) as well as Otte (1994) observed this species. On is many for morphology alone, one might doubt the singleof species of Laupala more than species from any the basis of island endemic status other endemic group. In fact, Perkins (1899) found these organisms so invariant morphologically that he designated only one species for is now recognized by Otte (1989, 1994) as a genus of 35 species. Single-island mtDNA what endemism in Laupala supported by the distribution of is variants across the archipelago. All sampled haplotypes were unique to single islands (Shaw, 1993, in The second common feature Laupala and Prognathogryllus is press). among that the generic radiations of Kauah, the geologically oldest island supporting a large extant native cricket fauna, harbors the basal lineages. This result was robust both groups. Furthermore, in for trees up to two most steps longer in both genera, the extant taxa descended from the next more-distal lineage are found on O'ahu, the next youngest and geographically nearest to Kauah. For the most clades contain species that inhabit the islands of generic radiations share the common part, the most distal Maui and Hawaih. Both feature that considerable inter-is- land exchange has occurred between the two youngest and geographically most proximate islands of Maui and Hawaih. These support the hypothesis of Otte (1989, 1994), who results do not proposed that the middle-aged islands (0‘ahu or possibly Maui in the case of Laupala)^ which harbor high taxonomic diversity, are the islands from which the would expect that species in the most basal positions would occur on 0‘ahu or Maui. A third common feature of the Laupala and Prognathogryllus radiations is that many species do find their closest relatives within the same island. Otte (1989) had hypothesized previously that most speciation in Hawaiian crickets occurs within islands, as opposed to the predominant extant radiations derived. Under Otte’s hypothesis, one 5 Crickets mode inter-island speciation of Hawaiian picture-winged the in Drosophila (Carson and Kaneshiro, 1976). Diversity in the two cricket genera studied here appears to have occurred via colonization to islands as they arose in geologic time but less frequently than in new Hawaiian Drosophila. Evidence for paleogeographic patterns in the relationships of native crickets is exciting for several reasons. First, the colonization patterns in and 4.6 Figures 4.5 raise the possibility that the origins of oecanthine and trigonidiine radiations may both the predate their current geo- graphic circumstances in the Hawaiian Islands. The oecanthine radiation which may provide further offers a compelling geographic distribution, insight. The group of Prognathogryllus, comprising the endemic sister genera Leptogryllus and Nihoa, a geologic relict Thaumatogryllus, has a representative on of the high-island part of the chain (Stearns, 1985; Carson and Clague, this volume. Chapter Nihoa species) Nihoa, it is conanti (the 2). If T. a biogeographic relict of a past oecanthine fauna on should reside in a basal phylogenetic position with respect to the Thaumatogryllus clade. suggest that the fauna Prognathogryllus is A Nihoa would derived and therefore past oecanthine fauna on on younger islands is also that a monophyletic sister group to Leptogryllus and Thaumatogryllus, providing evidence against a paraphyletic relationship. Furthermore, through phylogenetic investigation one might expect a similar older-to-younger island correspondence of basal genetic positions in the Leptogryllus Likewise, if and distal phylo- and Thaumatogryllus radiations. from a past fauna on Trigonidium species from the swordtail crickets are derived islands older than the current high islands, Kauah should occur in basal phylogenetic positions as discussed here for the Laupala radiation. Second, a geochronological influence is important because it creates the opportunity to investigate repeated patterns in history. Infrequent colonizations of and speciation, new islands, followed by adaptive intra-island radiations may offer circumstances in which similar evolutionary trends occur in parallel. A phylogenetic pattern with considerable inter- island resolution provides a context for focusing and ecological suites of characters. shifts implicates the The on shifts of reproductive relative evolutionary rate of these importance of different selective pressures in the diversification process. In Laupala, similar trends have already island, ent, one where may find anywhere from one species are distinguishable become clear. On any given to four sympatric species pres- by different songs (e.g., a slow, a 52 SHAW medium, and a fast singer). Convergent patterns in the temporal structure of the calling song and the communities that they comprise occur within Maui, and Hawaih) {Shaw, 1993). Ecological differences between species are subtle. Although Laupala species probably depend on certain elements of forest structure (e.g., suffithree of the high islands (O'ahu, cient understory or leaf plants. litter), they can survive independently of native Laupala species move into non-native forests such as guava or eucalyptus and thrive in the laboratory on a diet of standard Purina cricket chow. To what extent ecological boundaries exist sympatric species of Laupala is not these clear. In contrast to species of Laupala, nities among which occur in sympatric commu- apparently without host plant dependency, species in Prognatho- gryllus form identifiable ecological groups and occur in association with a variety of native plants, such as Metrosideros polymorpha Gaud. (Myrtaceae) and Freydnetia arborea Gaud. (Pandanaceae). The most closely related species of Prognathogryllus occur in disjunct ranges. patric associations are concomitant with association and greater phylogenetic members of more Sym- distinct differences in host diversity and, thus, greater island and elongatus groups on Kauah). Prognathogryllus robustus has been found exclusively in association with 'ohPa (Metrosideros polymorpha), where the purple or reddish individuals are cryptic against the red leaflets and small branches of 'ohPa trees. At night, males and females are often among the flowers in the terminal age (e.g., the robustus portions of the trees. Prognathogryllus robustus possesses cryptic coloration against the background of 'ohVa blossoms and foliage. Other members of the robustus group have similar stout differ to the largest degree in body proportions but pigmentation patterns. Members of the elongatus group, although only found in native forest, do not appear to be confined to any one particular native plant. They also cavort high in treetops at night but apparently pass the daylight hours closer to the ground in hollowed twigs or dried fern fronds. By contrast, members of the elongatus group are largely indistinguishable morphologically but Thus within Prognathogryllus, prominent patterns of diversity are apparent both in ecological and reproductive characters. The characters that make up the data set for Prognathogryllus are those used in the species-level taxonomy of the group. The benefits of this analysis are that it serves to establish explicitly hypotheses, which have have distinctive songs. previously only been suggested, and it allows the exploration of the phylogenetic information in species-level characters. A similar data set for Laupala proved uninformative. In Prognathogryllus, although some his- 53 Crickets torical information apparently exists in these characters, they are less than ideal for several reasons. As mentioned above, the variation among and thus a discrete coding system was chosen so that maximum-parsimony analysis could be performed. In cladistic species is quantitative, analyses, discrete characters with nonarbitrary categories are preferred (like the molecular characters used in the analysis of Laupala), but these kinds of data are currently unavailable for Frognathogryllus. The conclusions reached in the analysis of the Frognathogryllus data set are dependent on the quality of the phylogenetic inference, and there are several weaknesses in the present analysis. Partly due to the quantitative nature and arbitrary divisions imposed, polymorphisms had to be dealt with in the analysis of Frognathogryllus. Also, the characters that define different morphological groups in this genus, which often correspond to taxonomic species, are variable, and the distribution of variation in nature overlaps to and variable to some extent. To what extent they what extent natural gaps might exist in the distributions of these characters across the genus can only be ascertained through extensive population sampling. I are more chose to represent species as poly- morphic, as opposed to deciding on a single character value for each species based coding analysis size for (e.g., on one of the available methods for continuous character methods discussed by Archie, 1985). The strategy in this was chosen primarily because of the limited population sample each species. Recoding of polymorphisms at the species level had on the outcome of the biogeographic conclusions, although the topology was affected in minor ways. This alternative approach codes the species as polymorphic rather than allowing the maximum-parsimony algorithm to assign a monomorphic ancestral state as PAUP does under a polymorphic multistate designation (see Maddison and Maddison, 1992, no effect for a useful discussion of polymorphisms). The Hawaiian tree crickets, as well as the swordtail crickets, have diverged from their original founder lineages to such a degree that they were taxonomically misplaced by Perkins (1899) and Zimmerman (1948). The close relationship among species within the endemic genera further clouds estimating a root for the data. Thus, another polarity of the cladogram with morphological weakness of the Frognathogryllus analysis cladogram is inferred on the basis of is that the two characters (although a midpoint root did not change the biogeographic pattern inferred). in the A preferred outgroup comparison approach, such as that taken Laupala analysis, might be available. feasible if molecular data were 54 SHAW ACKNOWLEDGMENTS am making the Prognathogryllus data set available for me to analyze, for many informative discussions on Hawaiian crickets, and for comments on the manuscript. For useful discussions, I thank D. Baum and K. Crandall, and for insightful reviews of the manuscript, I thank R. DeSalle and C. Labandeira. I also thank V. Funk and W. L. Wagner for their efforts in organizing the symposium and this volume. The research discussed in this chapter was supported by I especially grateful to D. Otte for the National Science Foundation (BSR-9007117). APPENDIX 1. Character List for Prognathogryllus 4.1. Face, color of frons: 0 = pale; brown 3 = dark 1 = variegated pale and brown; 2 = brown; to black. 2. Head, color of rostrum: 0 = 3. Dorsum pale; 1 = brown; 2 = dark brown or black. of head, color of area medial to eyes: 0 = pale; 1 = brown; 2 = dark; 3 = all black. occipital stripes: 0 = absent; 1 = faint; 4. Head, 5. Pronotal length/greatest pronotal width: 0 = 0.80-0.84; 2 = present. 1 = 0.85-0.89; 2 = 0.90-0.94; 3 = 0.95-0.99; 4 = 1.00-1.04; 5 = 1.05-1.09; 6 = 1.101.14; 7 = 1.15-1.19; 8 = 1.20-1.24. 6. Pronotum, color of dorsal surface: 0 = mostly pale; 1 = slightly variegated; = highly variegated; 3 = mostly dark but with pale marks; 4 = 7. Pronotum, lateral lobe color: below; 2 = all 0 = pale or brown; black. = dark above, pale dark brown or black. 8. Metanotal gland, position of orifice: 0 = low; 9. Number of 1 all 2 file teeth: 0 = 50-99; 1 1 = moderately high; 2 = high. = 100-149; 2 = 150-199; 3 = 200-249; 4 = 250-299; 5 = 300-349; 6 = 350-399; 7 = >400. 10. Male forewing length/pronotal length: 1 = 2.0-2.4; 2 = 2.5-2.9; 3 = 3.0-3.4; 4 = 3.5-3.9; 5 = 4.0-4.4; 6 = 4.5-4.9; 7 = 5.0-5.4. 11. Female forewing length/pronotal length: 1 = 0.55-0.99; 2 = 1.00-1.49; 3 = 1.50-1.99; 4 = 2.00-2.49; 5 = 2.50-2.99; 6 = 3.00-3.49. 12. Male mirror length/mirror width: 0 = 1.00-1.09; 1 = 1.10-1.19; 2 = 1.20-1.29; 3 = 1.30-1.39; 4 = 1.40-1.49; 5 = 1.50-1.59; 6 = 1.60-1.69; 7 = 1.70-1.79; 8 = 1.80-1.89; 9 = 1.91-1.99. 13. Female cereal length/femur III length: 1 = 0.50-0.59; 2 = 0.60-0.69; 3 = 0.70-0.79; 4 = 0.80-0.89; 5 = 0.90-0.99; 6 = 1.00-1.09; 7 = 1.10-1.19. 14. Female ovipositor length/femur III length: 0 = 0.50-0.59; 1 = 0.60-0.69; 2 = 0.70-0.79; 3 = 0.80-0.89; 4 = 0.90-0.99; 5 = 1.00-1.09; 6 =1.101.19; 7 = 1.20-1.29; 8 = 1.30-1.39; 9 = 1.40-1.49. Crickets 15. Abdomen dorsum, last segments: 0 = pale; 1 55 = spotted; 2 = black. 16. Epiproct, central area: 0 = pale; 1 = black. 17. Subgenital plate color: brown or 18. Front = with small dark spots; 2 = dark 1 black. and middle femora 19. Tibiae: 0 = without 20. 0 = pale; Hind femur color: 0 = mostly pale; 1 = mostly dark. dark ring near knee; color: 0 = mostly pale 1 = with dark ring near knee. brown or tan; 1 = pale to dark; 2 = black. 21. Hind femur, patterning on outer 3 = 22. 23. all 0 = absent; face: 1 = faint; 2 = distinct; black. Femora III, Male tibiae knees: 0 = pale; 1 = dark. III length/femur III length: 0 = 0.30-0.39; 1 = 0.40-0.49; 2 = 0.50-0.59; 3 = 0.60-0.69; 4 ^ 0.70-0.79; 5 = 0.80-0.89; 6 = 0.90-0.99; 7 = 1.00-1.09; 8 ^ 1.10-1.19; 9 = 1.20-1.29. 24. 25. mm; 1 4 = 10-11 mm; 5 = 11-12 mm; 14-15 mm; 9 = 15-16 mm. Hind femur Number length: 0 = 6-7 of inner spines on tibiae mm; 2 = 8-9 mm; 3 = 9-10 mm; 12-13 mm; 7 = 13-14 mm; 8 = = 7-8 6 = III: 0 = 5-9; 1 = 10-14; 2 = 15-19; 3 = 20-24; 4 = 25-29; 5 = 30-34; 6 = 35-39; 7 = 40-44. 26. Number of middle spines on tibiae III: 0 = 0-9; 1 = 10-14; 2 = 15-19; 3 = 20-24; 4 = 25-29; 5 = 30-34. 27. Number of outer spines on tibiae 3 = 25-29; 4 = 30-34; 5 = 35-39. III: 0 = 10-14; 1 = 15-19; 2 = 20-24; ro ' o ro ro oooo o f\i r-. o rvj oo fv. ro ro ro n- 'o. Oo oo O ro o ro o ro «“ in 'Or ro ro ro o rs. rs- h- 'O'v. m Of st eo oooooo ro ro ro ro in in •O* ro \ N. o O 03 'o. ooooo O «“ ro ro ro '3- ro ro ro ro 03 in ro ro 00 'v. 'V. ro ro O >3- ro 'V, ro •“ ro ro ro ro in O'- '3- ro ooooooooo «- ro ro ro ro in oo >o r— ro >s ro ro ro ro ro ro ro ro r- ro o o 'J- ro ro ro ro ro ro ro >0 >» ro ro ro >» ro K> O ro ro ro ro in rsj 's*. ro ro >H. fv. fVJ in ru o ro O ro ro 'o ro ro ro ro ro (M oooooooOO o '3- OO ro S^ N. *“ ro 57 07o Ooooo «— fO o ro oooooo o ooo ro oooooo O ooooooo rsj oo •“ ns Prognathogryllus *” oo ro oooo fO •— o «— Oo ooo oooooO »— o f— *— r“ T— *” ooo ooooo ro o ooOo •— o *“ o *“ ooooo •— ro oo ro o o *“ oo *” o ro o ro *— o «*“ ro in oo fH, o- C“ ro n- o- ro ro ro c^- r^- ro o ro CO ro ro fH. fv. oooooooooooooooO ooo «- ro •— o «— •“ Xi ho ro ooo oo *“ *“ ooo oo ro ro *“ ro ooo oo ro (U nS NO oo o ns ro o T- o Oh (U (A <u Genus 4-> (0 i— CO C« JO 4-> nS iH (U 4-1 the O in ro ro •“ n- ro ro ro ro o n* o « •• n* n* h- f'- n- f'- ro m '*v o ro ro in 03 oO O" n- ro r“' ro ro n- n- 'O r- n- fs. oo Oo ro r^. oo ro 'O- •*3- r-* ro >4- ro r«- ro >0 >H. '*v. ro c^- 4- ro >3^ 'Sh, o N. o ro ro in 'v r^- O 's. 03 'v. o (U Oo ro ro ro r- ro ro fVJ c o ro X U States ro ro u ns sh nS for Bo »“ ro r- ro ro n- ro in r^' >4- ro in f\J o oooooo <^J ro o oo ro ooO ro ro o '4- ro f'- n- ro in ro ro ro ro in o- r- ro ro r- ro ro ro n* ro nO ro ro ro ro ro ro ro ro ro ro ro ro m ro 'J- 'v ro ro in »3 ro in o ooo o ro o »— o- N- o Oh Oh Character < o n- ro ro o ro oOoOooo r- ro o ro ro ro •- T- ooo to oooOo o o oo r- ro ro ro ro ooo ro oooo ro o ro St of o >0 m Matrix o oOo ro ro oo o ro o o in o ro ro »»» ro or o o ro ro r" «— ro o »“ ro o ro o ro o ooOooooo ro *“ *“ ro O ro ro ro <5 ro h- ro ** ro o o>. T— ro 03 o •“ o o ooo ro ro nS's. ro 03 ro ro C/3 (U ooooooo 4-> nS oooo ro ro 4-> ro C/3 Data Uh tU ro n* o ro ooo ro ro o ro ooo -M U u o *" ro ro ro ro o ro r^- ro ro ro ro ro o fO o ro ro ro o ro C'- ro o ro o ro ro ro ro o ro p>' o ro o ro ro ro ro ro o ro o ro ro o ro o ro ro ro o ro XD ri (N ro ro *" ro ro ro o -TO ro ro ooo o nS X Q Z w Oh Ch < c/3 3 u ns Wh nS XU (U c o u o cn kH OJ -)-> t/5 4-» 3 t/) Cfl ro 4^ ro 3 c 3 u c/5 (0 (0 (D $ Q. CO a •*-> (0 <0 ns • f— ro 3 (0 ro ro qJ a. Q. Cl. u •o ro c E > ro ro o c ro ro CT F O Q ro ro ro ro Q > </> 3 3 JC ro ai ro sz sz ro ro rol '4- • Q. a. CL a CL Q. CL CL Q c U) •f— c/5 C ro ro 3 o 3 ro JC ro 3 ro 3 x: 2 ro ro ro a E E o o1 ro in 3 xz ro JO ro ro 2 xz S ‘Z 2 qJ qI qJ ro i- ro X 3 "O v> EO "DQ ro L. c. 4-> <n ro ro o E o CU 4-* M U ro ro > 3 3 L. CL CL CL CL CL CL CL CL CL CL 0.1 a. clI clI CL 0.1 Chromosomes and Male Genitalia of Hawaiian Drosophila Tools for Interpreting Phylogeny and Geography KENNETH Y. KANESHIRO, ROSEMARY G. GILLESPIE, AND HAMPTON L. CARSON Zimmerman Nearly 35 years ago, cists and evolutionists to (1958) put out a challenge to geneti- investigate what he considered to be an “ex- He surmised that “It is possible that the Hawaiian drosophilid fauna may be the most remarkable in the world.” He was amazed that “there may be as many as 300 species traordinary” fauna of Drosophilidae. concentrated in an area smaller than the For three decades now (since the little summer state of Massachusetts. of 1962), a team of .” . . more than 75 evolutionary biologists from nearly every aspect of biology has studied this amazing group of the endemic mented by ogy, Hawaiian studies of behavior, biochemistry, insects, and as such, the taxonomic treatment of species in this family of flies has been supple- morphology, genetics, geographic distribution, ecol- developmental biology, cytology, population biology, and molecular biology. In Island Populations, Williamson (1981, p. 168) described the multidisciplinary approach to the study of the Hawaiian Drosophilidae being one of the most outstanding in evolutionary biology. Of all stated: the groups of organisms, plants or animals, that can be studied islands, the great He many Hawaiian Drosophilidae are supreme. This species; their ecology genetic analysis is is is as on why. There are a very varied; and, most important, a possible. In other groups of organisms, it is possible to study allozyme frequencies, metaphase chromosomes and so on. So far though, only in the Drosophilidae can we study evolution on an archipel- ago of a group with polytene chromosomes. These giant chromosomes. 57 KANESHIRO, GILLESPIE, AND CARSON 58 found in the salivary glands down sequences of bands and other parts of the their length whose show complex fly larvae, patterns differ in different banding produces detailed and sound evidence of species. This variation in phylogenetic history. In the Hawaiian Drosophilidae, this allows us to postulate at which points in evolutionary history the stock has immigrated to a different island In this chapter, and we what point at it has evolved on one island. discuss the significance of a comparative study of the banding patterns of the giant polytene preting phylogenetic relationships among chromosomes related species. for inter- When super- imposed on the biogeographic distribution of the species and a comparative study of the structures of the male genitalia, the chromo- somal phylogeny sets the foundation on which other sophisticated tools of evolutionary biology such as the DNA sequencing techniques can be applied to further our understanding of the evolutionary history of this remarkable group. THE GEOLOGY OF THE HAWAIIAN ISLANDS A key feature of the evolutionary biology of endemic fauna and flora of the Hawaiian Islands is the sequential formation of each of the high islands as the Pacific tectonic plate the large Pacific plate (Clague Clague, this volume. Chapter moved over a fixed hot spot beneath and Dalrymple, 1987; 2). see also Carson and Thus, approximately 5 million years ago (Ma), Kaua‘i, the oldest of the present high islands, was in the position where the island of HawaiT is currently situated. As the plate moved in a northwesterly direction at a rate of about 9 cm/year, the island of 0‘ahu emerged approximately 3.7 Ma, followed by the Maui Nui complex of islands (Molokah, Kaho‘olawe, Lanah, and Maui) 0.75 to 1.9 Ma. HawaiT, the youngest of the present high situated over the hot spot activity, began to form less islands, and which continues which to have is currently major volcanic than 0.5 Ma. Thus, evolutionary biologists are presented with a linear sequence of high islands and their constituent volcanoes, each formed in chronological sequence, with Kaua‘i currently the oldest gest. In most cases, of related species is it turns out that the and Hawaih the youn- most ancestral species of a group found on Kaua‘i and that the most derived species found on the island of Hawaih. Therefore, it is is possible to trace the evolutionary sequence of species formation with corroborative evidence from the geologic history of the Hawaiian Islands, which can serve as an Drosophila, Chromosomal and Morphological Analyses 59 important tool for increasing our understanding of the phylogenetic relationships among related taxa. TAXONOMIC STATUS OF THE HAWAIIAN DROSOPHILIDAE and Kaneshiro, 1981) have been named the family Drosophilidae from the Hav^aiian archipel- Currently, 511 species (Hardy and described in ago. Another 250 localities are sampled, to 300 undescribed new mates of 1,000 species 1993). species have been collected. species continue to be discovered, islands in historic times. endemic Drosophilidae were described in nine gen- era (Hardy, 1965), but several lines of evidence indicate that are part of only two esti- mostly widespread associates of species, humans, have been introduced into the Originally, the and fauna have been proposed (Kaneshiro, in this About 20 additional As new lineages. morton, 1966; Kaneshiro, 1976). all species Drosophila and Scaptomyza (Throck- On the basis of a comparative study of Throckmorton (1966) observed that although the Hawaiian species could be divided into two main groups, they showed distinct similarities, which suggested that the entire group may have arisen from a single introduction. He stated that “Hawai‘i must be considered to be the only place in the world where the otherwise sharp distinctions between Scaptomyza and Drosophila tend to disappear.” However, molecular studies of a larval protein (Beverley and Wilson, 1984) and recent DNA sequences of the alcohol dehydrogenase locus (Thomas and Hunt, 1991) suggest that the separation of scaptomyzoid and drosophiloid lineages took place at least 24 Ma. In view of the evidence that there were once high islands well to the northwest of Kaua‘i the internal anatomy, (see Carson and Clague, two lineages could either have occurred result of this volume. Chapter 2), divergence between the on older islands or may two independent introductions from continental Thus, despite extreme morphological diversity, which drosophilids appear to be very closely related phylogenetically. DNA (see ancestors. led earlier taxono- mists to divide the group into nine genera, the endemic ative studies of the be the Hawaiian Compar- review in DeSalle and Hunt, 1987) also appear to corroborate Throckmorton’s conclusions. Kaneshiro (1976), by pooling corroborating observations from studies of the internal (Spieth, anatomy (Throckmorton, 1966), mating behavior 1966, 1968), ecology (Heed, 1968, 1971), cytology (Stalker, 60 KANESHIRO, GILLESPIE, AND CARSON 1970, 1972; Yoon et 1972), and especially a comparative study of the al., external male genitalia (Kaneshiro, 1976), presented evidence for the existence of only two major lineages (genera) in the evolution of the Hawaiian Drosophilidae, Scaptomyza and Drosophila. It was demon- strated that the ''key” characters previously used to differentiate the more than two generic groupings were not characters and that most of these variations in external drosophilid species into "good” generic morphology were phylogenetically superficiaL Species in the genus Drosophila have been separated into species groups based primarily on external morphological characteristics of males (i.e., secondary sexual characters) (Hardy and Kaneshiro, 1981). These informal groupings of species have been designated primarily to facilitate discussions about species with synapomorphic characteristics, although these are not necessarily relevant for subgeneric classification. For example, about 100 species exhibit moderate to extreme modifications on the labellum of the mouthparts and have been grouped into the "modified mouthparts” species group. This large and heterogeneous group, however, Similarly, the is likely to be composed of many species complexes. "modified tarsus” species group can be further subdivided into the "split tarsus,” "spoon tarsus,” and "bristle tarsus” subgroups, THE PICTURE-WINGED SPECIES GROUP The "picture-winged” group of Hawaiian Drosophila, comprising 111 species, has attracted the most research attention. Most picture-winged species are large- bodied with striking maculations on the wings that vary from species to laboratory, and species. Many of these species can be reared in the detailed analyses of their morphology, behavior, genetics, cytology, proteins, and DNA can be conducted. These species provide extremely favorable cytological material and are particularly good subjects for comparisons of the banding sequences of the polytene chromo- somes. Carson and his collaborators conducted an extensive study (see reviews in Carson, 1987b, 1992b) of the inversion patterns of the giant polytene chromosomes of 106 picture- winged species, and they devel- oped a pattern of relationships based on the presence or absence of inversions relative to an arbitrary standard. Kaneshiro (1969) studied the male genitalic structures of this group and found that similarities in the shape of the phallic organs, especially that of the penis, are useful for separating the picture-winged species into species subgroups and com- Drosophila, most plexes. For the Chromosomal and Morphological Analyses on male part, the relationships based 61 genitalia complemented those of the chromosomal characters. We use examples from the picture-winged species group to illustrate the value of analyses of chromosomal inversion patterns together with a comparative study of conservative morphological characters, such as the male genitalia, for interpreting their phylogenetic relationships. CHROMOSOMAL TRACERS A main attribute of many OF PHYTOGENY dipteran groups, including that of Drosophilay chromosomes found in the cells of the salivary glands of the mature larvae. The banding patterns observed in the salivary gland chromosomes offer an abundance of details that can be used for comparative studies of gene order within and between species. Chromosomal is the giant polytene rearrangements, primarily a result of paracentric inversions, can be used to trace the evolutionary history of groups of closely related species that are similar in banding sequences. The chromosomal group (Carson, 1992b) assumed original unrooted phylogeny of the picture-winged species that paracentric inversions with two-break rearrangements are unique and species carrying this same arrangement in their chromosomes were presumed to have been derived from a common ancestor. More events, complex rearrangements with either overlapping inversions or inversions occurring within previously inverted sections with multiple breaks were assumed to reflect a step-wise evolution in the chromosomes. cases, a phylogenetic sequence of species formation In most can be traced by deciphering the sequence of step-wise rearrangements in the six polytene chromosomes, five long and one short. Each species was differentiated by a formula describing the relative to the arbitrarily number and position of inverted segments chosen D. grimshawi standard. PHYLOGENETIC ANALYSIS OF CHROMOSOMAL DATA Carson’s (1992b) data from 106 species of picture-winged Drosophila were coded for chromosomes X, and 5, according to whether an inversion sequence was absent (0), present and fixed (1), or polymorphic (0,1). Characters were analyzed using PAUP (Swofford, 1991), and character states 2, 3, 4, were polarized as primitive or derived by outgroup compari- son (Maddison et ah, the shortest trees. 1984). Heuristic searches were conducted to find The data were then reanalyzed by successive approxi- KANESHIRO, GILLESPIE, AND CARSON 62 mations, weighting characters according to their rescaled consistency index (RC) (Farris, 1969, 1989). Drosophila primaeva from the island of Kauah was chosen as an outgroup for the 106 species of picture-winged Drosophila. Although D. primaeva and sympatric close relative, D. attigua, are not “true” its picture-winged species in that they lack distinct maculations on the wings, their chromosomal banding sequences can be completely resolved in terms of the D. grimshawi standard. In D. primaeva, a sequence of bands on chromosome 5 has a gene order identical to the homologous sequence found on the same chromosome in D. colorata, a species from Japan (Stalker, 1972). The latter is now considered to be a member of the D. melanica group (Beppu, 1988), which is widespread on both the Asian and North American continents. Because all other picture-winged species of Hawaih have this sequence broken up by inversions, D. primaeva is clearly the Hawaiian most species closely related to continental species. This provides strong evidence that the “direction of evolution” has been from a continental species, through D. primaeva to the other picture- winged species of Kaua‘i and the newer Hawaiian considerations would seem to exclude the reverse order of evolution. For further discussion, see Carson To Islands. Geologic and Yoon (1982). establish the basic structure of the tree, represented most chromosomal types. we used 30 taxa that The taxa used were Drosophila D. bostrycha, D. clavisetae, D. crucigera, D. discreta, D. attigua, dis- tinguenda, D. engyochracea, D. fasciculisetae, D. flexipes, D. gradata, D. grimshawi, D. hawaiiensis, D. heteroneura, D. melanocephala, D. neopicta, D. nigribasis, D. oahuensis, D. obscuripes, D. ochracea, D. ornata, D. orphnopeza, D. pilimana, D. primaeva, D. psilotarsalis, D. punalua, D. setosifrons, D. setosimentum, D. spaniothrix, D. penna, and D. virgulata. A heuristic search generated which were then weighted according to the in the same 24 trees, RC 24 trunci- trees, length and reanalyzed, 147, resulting with a consistency index (Cl) of 0.998 and a retention index (RI) of 0.998. Strict consensus of these trees divided the taxa into three major clades and the D. primaeva-D. attigua sibling species pair. The first clade to branch off is the adiastola clade, which is defined by six nonhomoplasious characters (2d, 3k, 4o, Xu, Xx, and Xy). The second to branch off is the planitibia clade, nonhomoplasious characters clade, is (3d, Xj). final is defined by two group, the grimshawi characterized by three nonhomoplasious characters, the standard sequences corresponding to Xi, Xk, last The which group, there is Xo (Xi"^, Xk^, and Xo"^). Within this the grimshawi species group, characterized by the Chromosomal and Morphological Analyses Drosophila, 63 Standard fourth chromosomal sequence and the glabriapex species group. Within both these latter groups, several species complexes are poorly resolved by analysis of the chromosomal banding patterns (see belo\v). To determine taxonomic we used relationships w^ithin the grimshawi spe- 38 representatives of the group: Drosophila affinidisjuncta, D. atrimentum, D. balioptera, D. bostrychay D. ciliaticruSy D. claytonacy D. crucigeray D. disjunctay D. engyochraceay cies group, all D, flexipeSy D. formellay D. gradatay D. grimshawiy D. gymnobasiSy D. hawaiiensisy D. heediy D. hirtipalpus, D. lasiopoday D. limitatay D. musaphiliay D. obataiy D. ochraceay D. orphnopezay D. orthofasciay D. psilotarsalisy D. pullipeSy D. rectiD. D. mulliy ciliay murphyiy D. reynoldsiaCy D. sejunctay D. domaCy D. D. turbata, D. sproatiy silvarentiSy villitibiay D. sobrinay D. so- and D. Drosophila punalua was used as an outgroup. A villosipedis. heuristic search generated six trees, length 46, which were then weighted according to the RC cies group, and reanalyzed. The result was the same set of six trees, with a Cl of 1.000 and an RI of 1.000. There were nine terminal chromosomal groupings, represented in Figure 5.1 by D. ciliaticruSy D. engyochraceay D. flexipeSy D. gradatay D. grimshawiy D. hawaiiensiSy D. hirtipalpuSy D. murphyiy and D. ochracea. To determine the phylogenetic structure within the glabriapex spe- we used the remaining 33 species that grimshawi species group: Drosophila D. basisetaey D. divaricatay chaetaey D. D. conspicuay D. D. fasciculisetaey ineditay D. aglaiay digressay D. were not included D. alsophilay D. discretay in the assitay D. distinguenday D. glabriapeXy D. gymnophalluSy D. hexa- lineosetaey D. liophalluSy D. macrothriXy D. micro- myiay D. montgomeryiy D. ocellatay D. odontoph aliusy D. pauciciliay D. paucipunctay D. pilimanay D. vescisetay group. D. prostopalpisy D, D. punaluay D. spaniothriXy D. tarphytrichiay D. phalluSy D. prolaticiliay A and D. virgulata. psilo- uniseriatay Drosophila ornata was used as an out- heuristic search generated four trees, length 67. Weighting according to the RC generated the same four trees, with a Cl of 1.000 and an RI of 1.000. There were eight terminal chromosomal groupings, represented in Figure 5.1 by Drosophila assitay D. discretay D. distinguenday D. glabriapeXy D. gymnophallusy D. punaluay D. spaniothriXy and D. virgulata. Relationships using all ferenSy among species in the planitibia clade were determined 17 representatives of the clade: Drosophila cyrtolomay D. dif- D. hanaulaey D. hemipezay D. heteroneuray D. ingensy D. melano- KANESHIRO, GILLESPIE, AND CARSON 64 GRIMSHAWI CLADE GRIMSHAWI SPECIES GROUP Xa2j^u__ 3 hawaiiensis. heedi, silvarentis, recticilia, HAWAIIENSIS SUBGROUP musaphilia, gymnobasis, &. turbata eradata 2b Xp2. 2i hirtfpalpus, psilotarsalis . flexioes. formella. villitibia, lasiopoda (4b) 2 ^ orthofascia, sobrina murphvi ciliaticrus. reynoldsiae, 67% 5a ochracea. sejimcta, limitata, claytonae • . erimshawi, affinidisjuncta, disjuncta, bostrycha, pullipes obatai, GLABRIAPEX SPECIES Xe. Xf, 3z, 4e, 4f, mulli, villosipedis, atrimentum, orphnoDeza. crucieera. sodomae. soroati GROUP 4g puttalua. prostopalpis, prolaticilia, basisetae, uniseriata, ocellata, paucipuncta, paucicilia virculata. digressa, 5d (Xi. Xk, f"" |3. r"’ Ixh Xo) hexachaetae spaniothrix. rmcrothrix, odontophallus, psilophallus, tarphytrichia gvrmtophallus. Uophallus 4c discreta. lineosetae,fasciculisetae 75% elabrianex. pilimana. vesciseta, aglaia, conspicua, alsophila, micromyia Xk3, Xl3 assita. Xc3. Xd3. 2r montgomeryi dislinguenda. divaricata, inedita PLANITIBIA CLADE Xr Xt Xp, Xq, heteroneura. silvestris, planilibia, differens 3c2. 4w2. 4f3 86% oahuensis, hanaulae. neoperkinsi, Xs cyrtoloma, neopicta, hemipeza, nigribasis Xj, 3d substenoDtera, obscuripes setosifrons. picticomis (Xd2 Xe2 Xh 2 Xi2 Xp. Xk2 2f 4w, 4x, 4z Sh, 5i) , , . , ADIASTOLA CLADE , Xm 2 . 31, 4b2 ’ Xv, Xw, 3f. iXz. 2e 4o. 5f setosimentum. ochrobasis davisetae. neoclavisetae, neogrimshawi , adiastola. touchardiae, spectabilis, cilifera, Xu, Xx, Xv, 2d. 3k. 4o XV2 toxochaeta, peniculipedis hamifera. varipennis, paenehamifera, truncipenna orttata pnmaeva altigua FIGURE 5.1. Phylogeny of 106 species of picture-winged Hawaiian Dro- sophila. All resolved nodes have 100% support, except for the four marked. Solid lines lead to terminal groupings supported by specific chromosomal rear- rangements, using the terminology of Carson (1992). All marked characters dicate chromosomal inversion gains relative to the except for those in parentheses and in italics, in- “standard” D. grimshawi, such as {4b), which indicate an in- Dashed lines lead to groups of taxa within which relationships are unresolved and that do not form distinct groupings. Terminal taxa that are under- version loss. lined are those chosen to represent groups of species that are homosequential with the terminal taxon or that differ from it by autapomorphic character(s) only. Drosophila, Chromosomal and Morphological Analyses 65 cephala, D. neoperkinsi, D. neopictUy D, nigribasis, D. oahuensis, D. ob- D. scuripes, D. picticornis, D. planitibia, D. setosifrons, and silvestris, D. substenoptera. Drosophila ornata was used as an outgroup. A heuris- search generated 28 trees, length 59, which were then weighted tic RC was again 28 trees, with a Cl of 1.000 and an RI of 1.000. There were five terminal chromosomal according to the and reanalyzed. The result groupings, represented in Figure 5.1 by D. heteroneura, D. melanoceph- ahy D. oahuensis, D. setosifrons, and D. substenoptera. To establish the structure within the adiastola clade, we used 16 representatives of the clade: Drosophila adiastola, D. D. D. cilifera, D. hamifera, D. neoclavisetae, D. neogrimshawi, D. ochro- clavisetae, basis, D. ornata, D. paenehamifera, D. peniculipedis, D. setosimentum, spectabilis, D. touchardiae, D. toxochaeta, D. truncipenna, and A D. varipennis. Drosophila primaeva was used as an outgroup. ristic the all heu- search generated three trees, length 75. Weighting according to RC generated the same three trees, with a Cl of 0.991 and an RI of 0.963. There were five terminal chromosomal groupings, represented in Figure 5.1 by D. adiastola, D. clavisetae, D. hamifera, D. ornata, and D. setosimentum. Classical sibling species typically A feature of the show Hawaiian drosophilids is fixed inversion differences. the existence of 19 groups of chromosomally homosequential species that are nevertheless morphologically distinguishable (Carson fixed inversion differences characterizing species waiian Drosophila closest relatives inversions, is by a number of of picture-winged Ha- et al., 1967). In other cases, the quite variable. Some species may differ general, the more derived fixed inversions separating representatives (e.g., less species have fewer than separate species in the hawaiiensis subgroup), whereas species are characterized the island of Kaua‘i Hawaih by 18 is by many fixed differences their by two or more single inversion, others are separated and so on. In from (e.g., five inversions more ancestral D. ornata from separated from D. setosimentum from the island of fixed D. primaeva and D. inversions). attigua, The two most differ ancestral species, from each other by 13 fixed inversions. The fact that there are so patterns in the polytene necessarily many species that chromosomes show identical banding indicates that speciation accompanied by fixation of paracentric is not inversions. Indeed, inversions probably arise in clusters, under the influence of transposable elements (Carson, 1992b). KANESHIRO, GILLESPIE, AND CARSON 66 THE EXTERNAL MALE GENITALIA Sturtevant (1919) first mentioned the significance of the external male genitalia of Drosophilidae as a taxonomic tool for distinguishing between which other closely related species. Indeed, in cases of sibling species in external morphological characters are extremely similar, taxonomists have often had to rely on detailed comparisons of the structures of the male genitalia for distinguishing sibling species. Snodgrass (1957) stated that “the great diversity in structural detail of the genitalia gives these organs a value for identification of insect species almost equal to that of fingerprints for identification of human individuals.” Indeed, Hardy (1965) showed that the structures of the complex male genitalia of Scaptomyza species in the differentiating Hawaiian Islands are extremely important between closely related external features (i.e., nongenitalic) of species. for For the most part, other Scaptomyza species have been very conservative in the evolutionary history of these species and have not been useful for species identification. However, Kaneshiro (1969) observed that the male genitalia of the Hawaiian Drosophila species were not particularly useful for separating closely related species. Rather, he found that the strong similarities in the phallic organs, especially that of the penis, were extremely useful in showing phylogenetic relationships among many species of the large picture-winged group. Kaneshiro (1974) suggested that the dichotomy in the usefulness of the male genitalia for distinguishing species in Scaptomyza versus those in Drosophila may be correlated with the differences in mating behavior between the two groups. In Drosophila species, the plex courtship displays before attempting to mount males perform comthe female. Once the female provides an acceptance signal, the male mounts the female and, occasionally, copulation ensues. several minutes, male either and in most However, courtship can continue for cases, the female rejects the overtures of the by leaving the courtship arena or by physically aggressing against the male, inducing him to depart. Therefore, sexual selection in Drosophila appears to occur before the male mounting the female, explaining the elaborate secondary sexual structures observed in the external morphology of males of many of the in the species in this group. The species genus Scaptomyza^ however, display an assault-type courtship behavior with minimal premounting display. The males approach the female and immediately mount the female. In the mounted male appears to remain motionless while performing many position, the tactile stimuli involving the complex genitalic structures. Thus, in Scaptomyza, sexual Chromosomal and Morphological Analyses Drosophila, from the male selection appears to occur at the level of tactile stimuli genitalia presumably with corresponding receptors 67 in the female genitalia (ovipositor plates). SUBGROUPINGS BASED ON EXTERNAL MALE GENITALIA Based on comparisons of the external male formed subgroups that agree, chromosomal many cases, in general, genitalia, Kaneshiro (1969) with the groupings based on analyses. Nevertheless, the genitalic study provided, in a greater resolution in separating the species within the larger species groups into species subgroups (Table 5.1). For example, the grimshawi species group can be subdivided into three species subgroups, the grimshawi subgroup (7 species), the orphnopeza subgroup (17 species), and the hawaiiensis subgroup (14 The glabriapex glabriapex subgroup species species). group can be differentiated into the (five species), the punalua subgroup (eight species), the vesciseta subgroup (eight species), the conspicua subgroup (nine and the distinguenda subgroup species), Some sons of (three species). species can be placed in different groups based the aedeagus though even they are all on compari- chromosomally homosequential. For example, 13 homosequential species can be divided grimshawi subgroup (D. into the gera, (D. affinidisjuncta, D. bostrycha, D. cruci- D. disjuncta, and D. grimshawi) and the orphnopeza subgroup atrimentum, D. mulliy D. obatai, D. orphnopeza, D. pullipes, and D. There are also some discrep- D. sodomae, D. sproati, ancies between chromosomal and morphological among villosipedis. data. For example, four homosequential species, D. aglaia, D. glabriapex, D. pilim- ana, and D. vesciseta, D. vesciseta appears to be more closely related to and D. micromyia based on comparisons of the male genitalia even though they are two and four fixed inversions removed from D. vesD. assita ciseta, respectively. closely related to have as many D. Drosophila pilimana and D. glabriapex are more discreta, D. lineosetae, and D. fasciculisetae, which as four fixed inversion differences. Drosophila aglaia appears to be more closely related to D. conspicua, which is three fixed inversions removed. The point here agree well with the is that although the genitalic information appears to chromosomal phylogeny for higher-level group or clade designations, a comparative study of the external male genitalia can often resolve the groupings into species subgroups. In cases of chromo- 68 KANESHIRO, GILLESPIE, AND CARSON TABLE Hawaiian Drosophila Species Group and Subgroup Relationships Based on External Male Genitalia 5.1. Glade Species Species group subgroup Species primaeva primaeva primaeva Drosophila primaeva, D. attigua adiastola adiastola adiastola D. adiastola, D. cilifera, D. peniculipedis, D. ochrobasis, D. setosimentum, D. spectabilis, D. touchardiae, D. toxochaeta, D. ornata, D. clavisetae, D. neoclavisetae, D. neogrimshawi truncipenna D. truncipenna, D. hamifera, D. varipennis, D. paenehamifera planitibia D. planitibia planitibia D. differens, D. D. heteroneura, D. planitibia, silvestris, hemipeza grimshawi grimshawi cyrtoloma D. cyrtoloma, D. obscuripes, D. nigribasis, D. oahuensis, D. melanocephala, D. ingens, D. neoperkinsi, D. hanaulae, D. neopicta, D. substenoptera picticornis D. grimshawi D. grimshawi, D, crucigera, D. affinidisjuncta, D. disjuncta, D. bostrycha, D. balioptera, D. picticornis, D. setosifrons pullipes orphnopeza D. orphnopeza, D. mulli, D. villosipedis, D. atrimentum, D. sodomae, D. sproati, D. ochra- D. sejuncta, D. limitata, D. claytonae, D. ciliaticrus, D. reynoldisae, D. engyochracea, D. orthofascia, D. sobrina, D. murphyi, D. obatai cea, hawaiiensis D. hawaiiensis, D. heedi, D. silvarentis, D. musaphilia, D. gymnobasis, D, recticilia, D. turbata, D. gradata, D. hirtipalpus, D. psilotarsalis, D. flexipes, D. formella, D. villitibia, D. lasiopoda glabriapex glabriapex D. glabriapex, D. pilimana, D. discreta, D. fasciculisetae, D. lineosetae {Continued) Drosophila, Chromosomal and Morphological Analyses TABLE 69 {Continued) 5,1. Clade Species Species group subgroup grimshawi glabriapex (cont.) {cont.) Species punalua D. punalua, D. prostopalpis, D. prolaticilia, D. basisetae, D. uniseriata, D. ocellata, D. paucipuncta, D. paucicilia vesciseta D. conspicua D. conspicua, D. aglaia, D. spaniothrix, D. macrothrix, D. odontophallus, D. psilophallus, D. tarphytrichia, D. gymnophallus, D. liophallus distinguenda D. distinguenda, D. divaricata. D. inedita D. alsophila, D. assita, D. micromyia, D. montgomeryi, D. virgulata, D. digressa, D. hexachaetae vesciseta, somally homosequential species, the external male genitalia provide additional information with marked which the species can be separated into species subgroups. The examples described above significance of applying a combination of ting phylogenetic relationships among taxonomic clearly illustrate the criteria for interpre- related species. GEOGRAPHIC DISTRIBUTION OF THE PICTURE-WINGED DROSOPHILA We have generated an area cladogram based on the chromosomal data of the picture-winged Drosophila (Figure 5.2). This area cladogram directly comparable to the others in this volume because groups rather than individual taxa. Despite this, several it is not treats species general points can be made. The absence of chromosomal rearrangements at lower taxonomic levels suggests that these initial species 24 clades with two or more species: found on two or more islands, and only 4 are found on a divergence. There are 20 of these are single island, on 0‘ahu and two groups (two on Maui and one group (two species) on Hawai‘i. This one group (three and three species) pattern in is rearrangements are not involved in species) apparent agreement with the results of DeSalle (this volume. 70 KANESHIRO, GILLESPIE, AND CARSON FIGURE 5.2. General area cladogram of Hawaiian Drosophila based on chromosomal types. For any given clade, the ancestral species tend to be on Kaua‘i or 0‘ahu. The lack of chromosomal resolution at lower taxonomic levels suggests that chromosomal rearrangements divergence. Solid bar indicates presence Chapter 6), who found are not involved in initial species on an island. that each lineage of flies he examined had an inter-island distribution pattern rather than a radiation Of the 29 single island. terminal taxa, only 6 of the nonbasal ones have species on Kauah. This also agrees with DeSalle’s findings; distributed on a from 0‘ahu to the younger islands, all six of his clades were and none of them had on KauaT. Three of the basal groups, however, are on KauaT, so when the area cladogram is optimized, it gives a Kaua‘i ancestor for the species Drosophila, Chromosomal and Morphological Analyses 71 The lack of resolution in the area cladogram and the fact that the clades are found on several islands is consistent with the entire clade. nearly all pattern produced by many repeated introductions from older to younger islands. The results of this analysis do not conflict with the results of other studies in this volume. Also, they indicate that there are 10 clades of four or more species that have interesting distributional patterns that should be studied at the species level as soon as possible. These clades have the potential to establish whether the individual clades follow an older-to- younger island dispersal pattern and help to investigate whether there is repeated dispersal from older islands. Despite the lack of resolution of the chromosomal phylogeny, have reliable data on the geographic we distribution of the individual species of certain groups, especially the large picture-winged species, on the basis of extensive field work carried out during the many years of the Hawaiian Drosophila project, as well as the close scrutiny of the systematics and Most of these species, like many other extant terrestrial endemic fauna, show a very strong but by no means exclusive tendency to single-island endemism. Most species species identification of all specimens captured. thus appear to evolve on an island early in remain confined to that islands tend to form new and thereafter newer emerging species, a finding that has led to the serious may be somehow related to founder events Carson, 1990a, for discussion). These results are particularly vant, especially in view of the older, presently low, islands comparable this history island. Colonists arriving at consideration that speciation (see its on the moving sea level, from the information revealing that most of the northwest of Kaua‘i were once high islands in size to the present high islands (see volume. Chapter island new new 2). rele- Accordingly, it Carson and Clague, seems clear that as each Pacific tectonic plate rose new by volcanic action above populations became established from colonists stemming older, sinking islands. The important point is that these “founding” events have resulted in speciation on successively younger islands. Thus, active evolution, manifested by novel species and adaptations, has been most apparent at the newer, ecologically open lava flows that currently characterize the southeastern end of the archipelago. Molecular Approaches to Biogeographic Analysis of Hawaiian Drosophilidae ROB DESALLE Biogeographic patterns can be examined at several hierarchical levels using the Hawaiian Drosophilidae. The complexity of the patterns roughly coincides with the particular taxonomic levels of these overall phylogenetic relationships of flies living in flies The endemic to the archipelago to continental areas can best be examined at the generic and subgeneric levels in the family Drosophilidae. tionships of species examined flies. on the various The biogeographic rela- islands in the archipelago can be at the specific or infraspecific level. Possible relationships of examined using populations within a This study examines these three levels from a molecular perspec- areas within an island are best species. tive and attempts to detect biogeographic patterns mitochondrial The DNA first level at these levels using (mtDNA). concerns the origin of the Hawaiian Drosophilidae. Several authors have speculated on the origin of these flies. All have attempted to single out one or a few continental groups that might be the sister group of the Hawaiian lineage. Chromosomal (Stalker, 1972; Yoon, 1989; Carson, 1992b), behavioral (Spieth, 1982; Kaneshiro and Boake, 1987), morphological (Hardy, 1965; Throckmorton, 1966; Carson and Kaneshiro, 1976; Grimaldi, 1990), and recently, molecular techniques and Wilson, 1985; Thomas and Hunt, 1991; DeSalle, 1992) have been used to examine this question. (Beverley The second level concerns the detection of biogeographic patterns within the Hawaiian archipelago. These patterns will most likely reflect 72 FIGURE 6.1. Diagram showing the chromosomal relationships of the taxa examined The filled-in circles indicate the relevant chromosomal ancestor monophyly of the species in each group. in this study. mines the inter-island ecological that deter- founder events and have previously been examined from and chromosomal data. The chromosomal data (Carson, 1987b, 1992b) are probably the most enlightening for the examination of biogeographic patterns at this level. Several species groups exist within the Hawaiian picture-winged drosophilids that are ideal for this level of biogeographic analysis (Figure 6.1). The final level concerns the detection of biogeographic patterns within an island. HawaiT, the youngest of the current high islands in the 74 DESALLE archipelago, allows for the examination of patterns at this level. In particular, Drosophila the island of silvestris resides in most of the Hawaih. Previous analyses of D. morphological (Carson et al., rainforests that ring populations using silvestris 1982), chromosomal (Craddock and Car- and isozyme (Craddock and Johnson, 1979) techniques were able to detect a pattern of basal populations on the western side of the son, 1989), island and more derived populations on the eastern side of the island (Carson, 1992b). Kaneshiro and Kurihara (1981) used behavioral studies showing to establish mating asymmetries that they interpreted as bio- geographic patterns between areas from both sides of Hawai‘i. DeSalle and Templeton (1992) examined the relationship of D. silvestris populations on the eastern side of the island of Hawaih using molecular techniques and observed the same overall patterns for this side of the island as Kaneshiro and Kurihara (1981). MATERIALS AND METHODS The biogeographic relationship of Drosophila in the Hawaiian ago with those of continental areas has been examined using the 16S rDNA and ND-1 mtDNA archipel- in DeSalle (1992) sequences. Information for a more limited number of taxa for alcohol dehydrogenase sequences exists (Thomas and Hunt, 1991; DeSalle, 1992) and is also mentioned here. Parsimony trees were generated using PAUP version 3.0j (Swofford, 1990a). Hypotheses about the sister-group relationships of the Hawaiian taxa were examined, and an area cladogram was constructed. Because of number of taxa in some of the analyses, heuristic searches using a random addition option were performed. The taxa and outgroups for the study of the biogeographic relationships of six species groups among islands in the Hawaiian archipelago were chosen on the basis of chromosomal data. Only those groups of flies that were shown to be a monophyletic group on the basis of chromothe large somal inversions were used. Outgroups were always outside of these monophyletic groups on the basis of chromosomal data. Six groups of Drosophila and their outgroups were identified that these criteria fit (Figure 6.1, Table 6.1). Character state data were obtained in the mtDNA restriction form of fragment length polymorphisms (RFLPs). These data were collected using the methods outlined in DeSalle et al. DeSalle and Giddings (1986). For most data enzymes were used to sets, at least (1986b) and nine restriction collect character state information. These restric- Drosophilidae, Molecular Analysis TABLE 6.1. Drosophila Species Used in the Inter-island Biogeographic Analysis Species group or subgroup Species antopocerus Picture-winged hawaiiensis Picture-winged adiastola Picture- winged affinidisjuncta Picture-winged alpha planitibia Picture-winged beta planitibia Appendix Abbreviation"^ D. yooni yoon D. cognata cogn D. tanythrix tany D. adunca adun D. longiseta long D. arcuatus arcu D. hawaiiensis hawi D. gradata grad D. rect recticilia D. musaphila musa D. adiastola adia D. setosimentum seto D. clavisetae clav D. cili cilifera D. spectabilis spec D. peniculipedis peni D. affinidisjuncta affi D. bostrycha host D. disjuncta disj D. grimshawi grim D, cyrtoloma cyrt D. melanocephala mcph D. hanaulae hana D. neoperkinsi npki D. obscuripes obsc D. nigribasis nigb D. oahuensis oahu D. neopicta npct D. silvestris silv D. heteroneura hete D. planitibia plan D. diff differ ens D. hemipeza hemi D. neopicta npct ^Used in ^Used in the individual analyses. 6.1. Outgroup^ D. arcuatus D. musaphilia D. clavisetae D. grimshawi D, picticornis D. neopicta 75 DESALLE 76 FIGURE 6,2. Study areas for the inter-island analysis of six lineages of Drosophila. Area abbreviations are K, MK^ Kauah; O, 0‘ahu; Molokah; MEl^ East Maui ME2, Maui (Paliku); MW, West Maui (Hana‘ula); HH, Hawaii (Hilo side); HK, Hawaii (Kona side). (Waikamoi); tion enzymes varied from study to and scored Parsimony study. Restriction sites East were mapped as present or absent to generate the character state data. trees were generated from the data sets for each species Table 6.1 using PAUP (Swofford, 1990a). The areas subgroup listed in examined in this study are summarized in Figure 6.2. Area cladograms were then constructed from the taxon cladograms (Page, 1988, 1989). Because all the species in these analyses are single-island endemics, the analyses of these six groups are straightforward due to the lack of both widespread taxa and missing areas. The characters for the study of biogeographic relationships of sophila silvestris populations within the island of Hawaih are Dro- described in and Templeton (1992) and DeSalle et al. (1986a). Thirty- three characters are included from mapped four- base cutter endetail in DeSalle zymes (23 characters), several DNA sequence characters from ND-1, ND-2, and ND-5 mtDNA genes (2 characters), and characters from six-base cutter enzymes (8 characters). Taxon parsimony trees for the population level data were generated using PAUP. Area cladograms were constructed directly from the taxon cladogram using 1989) under assumptions 0, 1, and 2. COMPONENT These assumptions (Page, refer to the treatment of missing areas, widespread taxa, and redundant distributions in biogeographic analysis. Component considered a valid approach if analysis under assumption 0 the taxa under examination are neither widespread nor show redundant distributions endemic to is single areas). If there are (i.e., if the taxa are entirely widespread taxa or redundant distributions with respect to the areas under examination, then assumptions 1 and 2 are the more suitable approaches. Assumption 0 prohibitive of the three assumptions with respect to the is the number of most area cladograms allowed. Consequently, area relationships generated under assumption 0 will often show more resolution than under assumptions 1 77 Drosophilidae, Molecular Analysis and Assumption 2. 1 more is cladograms than assumption prohibitive w^ith respect to the 2. number of For a more detailed discussion of these assumptions, see Nelson and Platnick (1981) and Page (1988, 1990). RESULTS Origin of the Hawaiian Drosophilidae The Hawaiian Drosophilidae are monophyletic. There is no single continental form that can be designated as the sister to the Hawaiian lineages. The analysis for this level of biogeographic pattern is essentially the same as in DeSalle (1992). Taxa from the three main genera of Hawaiian Drosophilidae (Hawaiian Scaptomyza, Hawaiian Drosophila [or Idiomyia, Grimaldi, 1990] and Engiscaptomyza) were used to represent the Hawaiian mtDNA lineages. Figure sequences and 6.3A shows the phylogenetic analysis using 11 continental Drosophilidae candidates. Fig- Adh and mtDNA were common to both ure 6.3B shows a total evidence analysis in which both sequence data were combined for those taxa that studies. The pattern that emerges in these phylogenetic analyses addresses three important points. First, the three distinct Hawaiian Drosophila myza comprise 6.3. (or Idiomyia, Grimaldi, taxonomic lineages of 1990) and Engiscapto- a monophyletic group in both analyses depicted in Figure Second, Hirtodrosophila, the only non-Drosophila candidate for sister- group status to the Hawaiian Drosophila (Grimaldi, 1990), is shown to be basal in both analyses, as Grimaldi’s (1990) analysis also shows. Third, there is no single continental species or species clearly be designated as sister to the group that can monophyletic Hawaiian lineage. Inclusion of extra-Hawaiian Scaptomyza taxa in an analysis, which was not done here, could alter the tree topologies reported in this analysis. Biogeographic Patterns of Drosophila within the Hawaiian Islands The six species groups of Hawaiian Drosophila examined produce roughly similar inter-island phylogenetic patterns. Appendix 6.1 shows The phylogenetic patterns that arise from indicate a general trend of the most basal taxa the data used for these analyses. these data (Figure 6.4) occurring in the rainforests of the older or central islands in the archipel- ago (0‘ahu or Moloka‘i), with the more derived taxa residing younger islands (usually Hawaih). Tree mony analysis are shown in Table 6.2. statistics in the obtained from parsi- 78 DESALLE D. melanogaster^ Sophophora D. robusta D. melanica continental subgenus D. funebris Drosophila D. pinicola D. immigrans D. repleta D. mimica “ — “ D. sproati — E. crassifemur S. exigua Scaptomyza Hawaiian brosophilidae Drosophila Zaprionus Chymomyza Hirtodrosophila Scaptodrosophila subgenus Sophophora continental subgenus Drosophila Hawaiian Drosophilidae FIGURE Hawaiian 6.3. (A) species. mtDNA phylogeny of Numbers on 1 1 continental Drosophilidae and 4 the cladogram branches indicate the length of the branch. (B) Total evidence tree for nine Drosophilidae in mtDNA data and 238 bases of Adh sequence exist. The which 905 bases of limits of the Adh se- quences coincide with those reported in DeSalle (1992). Numbers on the clado- gram branches The indicate the length of the branch. patterns observed in these cladograms were used to construct area cladograms for each species group and subgroup (see Figure 6.3). Construction of area cladograms requires the consideration of several The methodology for construction assumptions (0, 1, and 2). Assumptions 1 aspects of the areas and taxa involved. of area cladograms uses three and 2 differ from assumption 0 and redundant in how they interpret widespread taxa distributions. In particular, assumptions 1 rate the existence of and 2 incorpo- widespread taxa and redundant distributions into the i i, side adiastola. Kaua (Hilo K, (F) Hawaih outgroup; hawaiiensts; HH, OG, (E) are (Hana'ula); planitibia; abbreviations Maui West alpha Area (D) MW, right. the (Paiiku); antopocerus; on Maui (C) cladograms East affinidisjuncta; ME2, area and (Waikamoi); (B) left the planitibia; on Maui East beta cladograms (A) MEl, side). mtDNA Moiokah; (Kona 6.4. MK, FIGURE Hawaii 0‘ahu; HK, DESALLE 80 TABLE 6.2. Tree Statistics for Individual Species group or subgroup Data Sets RI Cl Steps antopocerus 91 91 Picture-winged hawaiiensis 80 75 44 20 Picture-winged adiastola 60 51 71 Picture-winged affinidisjuncta 67 Picture-winged alpha planitibia 62 63 50 59 32 99 89 Picture-winged beta planitibia 61 Notes: Cl, consistency index; RI, retention index; steps, number of steps in the tree including uninformative characters. Cl and RI were computed ignoring uninformative characters. construction of area cladograms. Because all the taxa used to construct area cladograms for the separate Hawaiian species groups and subgroups are single-island endemics with no redundancy in distribution, the prob- lem of constructing a general area cladogram collapses to assumption What this means becomes COMPONENT (Page, 1989) the evident when 0. these data are analyzed using because generally all three assumptions give same area cladograms. A matrix with the relevant information for construction of a general area cladogram from the analyses of the six species groups and subgroups is shown in Appendix 6.2. The resulting area cladogram from this data matrix (Figure 6.5) indicates that the islands of 0‘ahu and Moloka‘i were most likely areas of original endemism for these taxa. The younger islands of Maui and Hawaii are observed as the islands on which more recent differentiation has occurred. FIGURE Ancestor 6.5. General area cladogram for the 0‘ahu six lineages in Figure 6.4 using the in Moloka‘i West Maui East Maui East Maui 2 1 Hawai‘i Kona Hawaii Hilo Appendix 6.1. matrix Drosophilidae, Molecular Analysis 81 Biogeographic Patterns in Drosophila silvestns on the Island of Hawai‘i An was area cladogram for the eastern side of the island of Hawai'i constructed using the data and analyses in DeSalle and Templeton (1992) and DeSalle et al. (1986a). In those studies, the congruence of the molec- ular cladogram with a behavioral hypothesis was the primary The characters and character states are shown in Appendix interest. 6.3. Area relationships for the eastern side of the island of Hawai‘i can be obtained DNA sequence data by constructing an area clado- from these RFLP and gram using COMPONENT (Page, 1989) under assumptions 0, 1, and 2 (due The results of this analysis (Figure 6.6) indicate that there is a pattern of more derived populations and individuals on the southernmost part of the island, with the more basal populations residing on the northernmost part of the island. to redundant distributions of some lineages). DISCUSSION General biogeographic patterns can be observed at several hierarchical levels in the Hawaiian Drosophilidae using different molecular tools. molecular approaches described in this report vary from level to The origin of the diverse ancestors Hawaiian was approached using evolving region of the mtDNA lineages DNA The level. from possible continental sequences of a relatively slowly (DeSalle, 1992) and a portion of the Adh gene (Thomas and Hunt, 1991; DeSalle, 1992). These gene regions appear to generate enough relationships. RFLP DNA sequence variability to resolve certain techniques were used to examine the other two hierarchical levels. Six-base cutter technology appears to be sensitive enough to detect patterns within sequences at this species groups level, in general, DNA generate nonsubstantial amounts of information in relation to effort (DeSalle et sequences of two rapidly evolving and subgroups. mtDNA al., 1987). In fact, DNA genes (ND-5 and ND-2) generated only two phylogenetically informative nucleotide positions among Drosophila silvestris populations (DeSalle and Templeton, 1992). Four-base cutters appear to be the most efficient means of generating molecular characters for within species questions in this study, as demonstrated by the analysis of D. ton, 1992). silvestris populations (DeSalle and Temple- 82 DESALLE Kilauea Pi'ihonua ‘Ola'a Maulua Kohala Kona A c Kilauea Pi'ihonua 'Ola'a Maulua Kohala Kona D B FIGURE 6.6. lineages. (A) study. Area cladogram for intra-island analysis of Drosophila Map The arrow of the east side of Hawaih showing tions 0 and 1. examined indicates the direction of divergence established havioral data of Kaneshiro and Kurihara (1981). (B) 10 individual the areas isolines examined in this study. (C) silvestris in this from the be- Taxon cladogram for the Area cladogram under assump- (D) Area cladogram under assumption 2. Drosophilidae, Molecular Analysis 83 Origin of Hawaiian Drosophilidae Three distinct questions Hawaiian lineages. can be asked to unravel the possible origin of the The first and most important point relevant to the validity of the other questions concerns the sister-group relationships of the Hawaiian taxa. Throckmorton (1966, 1975) first suggested that the Hawaiian taxa could be the product of a single or two introductions to the archipelago, primarily on the basis of internal morphological analysis. This conclusion stemmed from the existence of the Hawaiian drosophilid taxa (Hawaiian Drosophila or Idiomyia) and the Hawaiian scapto- myzoid (genus Scaptomyza) flies. If these two lineages are not sister taxa, then two or more continental groups may be sister to the Hawaiian Drosophilidae. Both DNA sequence data sets (Adh and mtDNA) support the notion that the Hawaiian Drosophila and Hawaiian Scaptomyza are taxa; thus, the next two questions can be approached under The second question concerns sister this hypothesis. the relationship of the mycophagous clade of Drosophilidae that Grimaldi (1990) indicated as the sister group to the Hawaiian Drosophila. Hirtodrosophila was chosen tative of this group and, in these analyses, is shown as a represen- to be basal in the drosophilid phylogeny, in agreement with Grimaldi’s (1990) placement. The Hawaiian taxa are not seen as a sister group to the Hirtodrosophila (see Figure 6.3). Also, closest sister Grimaldi (1990) suggested that Zaprionus is the group to the Hawaiian Scaptomyza and Engiscaptomyza. This hypothesis also is not supported by the molecular data, as the Hawaiian Scaptomyza and Engiscaptomyza are placed well within the genus Drosophila. The third question concerns the designation of a subgenus Dro- sophila or species group that could be the ancestor of the Hawaiian Drosophilidae. This question tive is best approached by placing several puta- subgenus Drosophila candidates along with the Hawaiian Droso- philidae in a phylogenetic analysis (see Figure 6.3). The results of this none of the continental subgenus Drosophila can be designated as the sole sister taxon of the Hawaiian lineages. analysis indicate that From these data, the Hawaiian Drosophila appears after Hirtodrosophila diverged from the ancestral drosophiline lineage and before the divergence of the major subgenus Drosophila species groups (see Figure 6.3). living taxon This observation, in effect, in the present analysis that means that there is no single can be assigned as the ancestor of Hawaiian Drosophilidae. Amber fossil subgenus Drosophila and genus Scaptomyza (Grimaldi, 1987) of 30 million years ago (Ma) exist the 84 that DESALLE would give a minimum age of the divergence of the Hawaiian lineage from the subgenus Drosophila based on sister-group dating 1992). This date of divergence (1985) estimate of 40 clock” and from Adh Ma is (Norell, roughly similar to Beverley and Wilson’s from a larval Thomas and Hunt’s hemolymph protein “molecular (1991) estimate of 20 Ma using an molecular clock. These divergence times are interesting because they imply that the colonization of the Hawaiian archipelago could have occurred well before the origin of the current oldest Hawaiian Island (Kauah) that has sufficient rainforests to harbor these tion is flies. This observa- not entirely surprising because the Hawaiian archipelago has been formed on a geologic “conveyor belt” (Carson and Kaneshiro, 1976; McKenna, 1983; Beverley and Wilson, 1985). Inter-island Biogeographic Patterns Most aspects of the mtDNA cladograms (see Figure 6.4) are in direct agreement with chromosomal data, but others show slight disagreement. For instance, the alpha sublineage of the Drosophila planitihia subgroup shows the two 0‘ahu basis of species as ancestral. Carson (1987b) argued on the chromosomal data that the area of origin for this lineage is Maui Nui complex (Maui, Molokah, Lana‘i, and Kaho‘olawe) 0‘ahu members of this lineage arose as back-migrants. If this is actually the and that correct, the patterns observed from mtDNA RFLP data might indicate that the back-migration events occurred very early in the differentiation of this clade. The general area cladogram in Figure 6.5 clearly 0‘ahu is placed in the most basal position in the tree and indicates that that Molokah appears next. These two areas can therefore be interpreted as being established before the remaining three. are observed as sister areas common The areas on Maui and Hawaih and are the last areas to be established. A misconception in the interpretation of these data would be to assume that vicariance events are responsible for the observed patterns. Endler (1982) argued that congruent area cladograms could reflect com- mon ecological processes. reflected in congruent area Common dispersal pathways may also be cladograms (Endler, 1983; Page, 1988). Intra-island Biogeographic Patterns The big island of Hawaih has also been formed as a consequence of the movement of the Pacific tectonic plate over a hot spot, resulting in a series Drosophilidae, Molecular Analysis 85 of volcanoes with decreasing age from the northern part of the island to the southern part of the island (Spieth, 1982). This in the archipelago and consequently has the is the youngest island greatest potential of the current high islands for the effects of this conveyor-belt island formation to be demonstrated in a cladistic analysis. Wet noes and consequently are colonized by these forests ring these volca- flies. Drosophila silvestris populations have been examined in these wet forests for the past 20 years, and a great deal of information regarding the chromosomes, isozymes, and behavior of this species has been collected. The chromosomal and isozyme data (Carson, 1983c) are inconclusive as to the phylogeny of these populations. However, behavioral studies (Kaneshiro and Kurihara, 1981) clearly show a pattern of ancestral populations residing in the wet and more derived populations forests of the northern older volcanoes residing in the wet younger volcanoes. The forests of the southern mtDNA data set (see Figure 6.6) also detects this pattern, although not in the same degree of detail as the behavioral data. the area relationships result of the more shown The lower under assumption 2 in Figure 6.6 inclusive nature of this level of resolution of is the assumption (Page, 1990). CONCLUSIONS Molecular techniques can generate characters for use in biogeographic Hawaiian Drosophilidae. Patterns can be detected at several biogeographic levels by using different taxonomic assemblages of these flies. Also, because different taxonomic levels are used to discover analysis of the the patterns, the molecular approaches levels. must change for the different DNA sequences of slowly evolving mtDNA genes (rDNA) are used to examine the patterns of the origin of these mtDNA data suggest that the Hawaiian lineages are that there is no The monophyletic and Hawaiian clear continental subgenus Drosophila taxon that can be designated as the sole ancestor of the Hawaiian lineage. mtDNA is amount of information all six and subgroups. This technique maximizes the for the effort used at this level. is a definite set of area relationships that are groups and subgroups of these relationships wet data from general area cladogram for six species groups (see Figure 6.5) indicates that there to RFLP used to examine the species-level phylogeny of several closely related species groups A fly lineages. may forests) to reflect the which these flies. The uniformity of these area narrow ecological ranges flies common (i.e., high-altitude are restricted. Also, once dispersal occurs. DESALLE 86 and behavioral attributes observed in these flies (Spieth^ 1982; Kaneshiro and Boake, 1987) contribute to the common pattern of area relationships from mtDNA cladograms. Mating asymmetries of these flies, discovered through experimental work (Giddings and Templeton, 1983; the it also possible that the strong mating asymmetries is Kaneshiro hypothesis, Kaneshiro, 1983; Kaneshiro and Giddings, 1987), imply strong behavioral isolation that might have a profound the phylogenetic patterns observed for ago an is mtDNA. The Hawaiian ideal system for demonstrating this possibility. effect on archipel- The sequence of formation of the Hawaiian archipelago would force colonization patterns to be these common flies. If in groups that have the same dispersal capabilities such as the mating asymmetries observed in the laboratory also affect these flies in nature, then an even stronger directional component would be enforced on the phylogenetic relationships of these flies. RFLP and DNA sequence data are used to examine the possible area relationships within an island. Area relationships within the island of Hawaih exist and, in general, follow a north-to-south direction. This result agrees with the temporal formation of the volcanoes and wet component analysis (see Figure 6.6) to diagnose the southernmost areas as distinct from each other is most likely due to the redundant distributions of the flies from these localities. The flow of genes among these populations is most likely forests on the island of Hawaii. The inability of the responsible for this redundant distribution, although stochastic branching processes (Avise, 1986; DeSalle et ah, 1986a) could also result in the types of distributions observed for Drosophila silvestris Hawaih. on the island of 03 d :5 ^ 05 IH d d o X 0 e cd c/5 w D 'O ^ 03 •M id CO O cd C/5 rd d X ^ 9 u CO Analysis e 1 X X d ' 3 d o a3 03 0) O g o * X T3 O ^ IH .a a e «sH IH d d cd . - 4.J > o ts «d 4-> mtDNA cd -2 o Mh CO cd for w u cd CO States 9 d O 4-i 2 ’> IH CO 3 O ui cd 3 O • o o 0 9 d cd CO 4) X H ^o -a N d qj X e X Q Z w cd d o ^ >2 « M M -Q cd TS d 03 ^ - U 9J r <» ON cd • £S > fH«HrH?HrHOOO ft. OOOOOMrHO OOOOfHrHOO 0000000«H « X X « d o "6 qj O u X u X d s IH .o " d ^ l-H cS ^ 'O 'S 2 a £ O^HOOOsHOO OOrHrHOOOO ^ CO U U Xo u X f 5 00110111 rH rH rH rH 0000000*H sHrH«HfHsHrH»HO d u ^ IH OOOOfH*H«H«H OOOOOOO^ JO«H*HOOO O o O O rH O o O O rH rH rH O rH oo O rH O rH O O rH rH OOO O O O rH rH rH rH rH rH oo oo rH rH rH o rH rH rH rH u 1«— OOOOOiHrHsH ?— CO W U W a o rH rH rH rH FHtHsH*HOOOO OOOOOrH^HO OOOOOOOrH 0000«HiH00 f-HrHrHiHs—iOOO 0000000«H rH rH rH rH rH rH rH rH rH rH rH rH £ OOOOOrHOO 0000000*H OOOOOOrHtH c/5 tu o d « 05 sHsH«HrHfHrHOO X O u s rH rH rH rH rH rH OrHOsHfHsHOO OO^OrHOOO OOOOrHO*HO 4) ^ 000000 110001 110001 100010 OrHrHrHsHfHOr—J OOOOO^’HrH rHfHsHrHOOOO u to f-HsHr-SiHOOOO d sHsHfHiHOOOO «HrHfHsHOOOO ca OOOOfHfHOO 0000«H«H*HtH M M M X IfHrHtHs-HOOO s— .2 sH»HOO«HOOO «H rH rH ?H «HI OO «H O >>a)(DCHH«H rH rH rH rH CO 00000«H«HrH .V c rH PC iHrHOrHOsHeHO fHsHsHfHOOrHO .2 ^ « O O rH o oo oo o O rH rH «H rH rH rH > sH«HrHsHsHO«HO • »' k_H cd 10111111 sHrHfHrHrHfHOO 03 4' -fi H -M H c ffl t-SOOO«H?HOrH 4-J e o »-t <— c« I.-N s 03 a;^s o uo 0«HfHfHOOOO 00000«HfH0 d — • O CO ^ w 03 w • XM W iH 9 * Oh Oh d 03 PS rH »H O O sH O ?H*HsH«HOOOO OOOOO^HiHfH CO td 4-i» ^ b. W IH u d 4J CO o ^ « u X ^ -M QJ \0 . 01110111 M OOOO^fHQfH cd flj 9 00110111 p=*i cd t/5 -d «d 1 fHrHrHsHrHrHrHO « IH ® «v pH pH pH —ifHsHfHeHeHrHO CO f— O O O pH o o 5 *-* cd O O O pH o o ooooooo^ sHfHrHfH«H«HOO > u O pH O o o OOOOOOOrH 0000000«H «h«h«h?hoo©o 6 & pH XS X s ^ o X X X X o cd es t_H IH < o X VO o 8 PQ dn*' d o 'M o 9^ fCd Characters u d a u ^ bD <M w u 3 3 m « U Oh cd and o qj « » Oh 3 td Mw u ,o d iS cd u « Character l-l U C C >. C O* 3 O O' C 3 C O 0 0 ta o o sh >, U 4J «J pH <a A- cd s a O « IH 9S Ck -M pH »** OOO O O OOO O OOOO O O O O O O O pH pH pH O O pH pH O O pH O pH O O O pH O O O O O pH O pH pH rH pH OOOO O pH O pH O O O O ooo oo o fH O pH pH pH pH > ^ o to <u e o O O o S K S o rH pH ?-H pH pH pH pH 3Oh 3 cd X o ^ J-l 3 RFLPs © w cd for oo O <H O pH oo pH pH O o o o o *-H e +.H o o o o OO O pH O pH pH pH OOO ooo ooo O pH O cd m "O d ^ O R s « d o ^ u CO 9 «o CO ® 0 ^ 8 O 9 r% u Inter-island a w < X d o ^ X 03 cd ?d QJ Ui o o o o a M 04 O 12 ° CO u 09 K s“H u ed « O 44 O a 3 Q w > « u d u 9 «-) -d W X CO o « ^ CO •-« ffi cd i-i a 0 U ^ « Q ® e © -o OOO O ooo ooo «H rH sH O sH O OOOOOO OOO OO OOOOO ooo ooo rH rH fH pH O oooooo oooooo ooo ooo pH O pH O O O rH iH i-t r-| sH jH «H sH sH sH <rH «H sH •M 0 rH rH rH rH ooo o ooo rH rH rH rH P-I rH rH rH rH o 4J 3^£0 ^ Qj •-HrH4J4J(t 3 •H *«H 0) W3 0) ^ ^ x: rH -rj Oj'O X c grad T3 d td td >• td rect hawi tnusa cd <d xt^xxx^xo xxxx^xoo U HH OG OH EM CL 3= o o r™ oo r” zu (/) CQ o o o V” o o o o r” r“ O o o o o o o— o o T” o o « ,- o o o o o o o o o o o o o o o «~ oo « oo o V— o o o o o o oo o oo o o o o o o o T— «“ oo « o o o o r“ o o o o o o o r— r“ o o o o o o o o £ X r— V— *— oo oo o o o o o oo oo o o o o o o o o o o o V— oo o o o— V” o o o o o o o o V" o r- T- V V u c/5 o oo oo o «“ o o o o v~ o ooo v~ o ^— o o % > a: o o V— o T" o « o V— o o o «— o « o o o o o ^— V o T— o o r— o t o o «— o o T— o o o T“ o o o _# CJ oo o o o o o oo o T- «- oo o o o o o o «— «— OO o o o o oo o oo oo o — r— T~ o o o o oo o o oo oo oooo o ooo »— «— o o »- »— o o oo o »- *o oo oo oo O O O 1»- o o o o o oo o oo o o o o o oooo oo o o o o o o o o o o «— o o «— o o oo »- »— o o o o ooo oo o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o a. Sc Si X Xh JD X ^r“ * o o o V— o oo oo o o o o o oo oo s ^ S o o o o o o o o o o o o ^ •-4 > au U « •-»l ^ a >» ^ "ts ^ R ' ^r“ «s ’T3 X HH Q (U :z: <u bC Q£ ’T3 > <u CL bP o ^ o o o o o «=” o o o o o o o o o o o o «“ o o o o o o -— o o oo o o o o o o o o o o o o o o o o o oo o o o o o o *" 00 »— «“ QC o o o o o o o o o o o o oo o o oo r™ oo «“ ^ • s ft- ’o o o o o o o o o o o o © o o o o © o o o o o o o o o o o o o o o o o o o © © o o o o o o © o o o o ^ — o o o o o o o «— o o o o o o »— «” -r* o © o © o »— o o o o o © o o *“ o o o o o o o o o o o o o o o o «— o o o o T~ -=- o o o o o o o o o T— o o o o o o o o© © V— o © o o o © V™ o w™ o ©o — o o oo © © o © o «™ o© © © «— © oo©oo o oo o o o o© o o o © o o o o © © o © -— o © © o o o o o o o o o o o o o o o o o o o o o e— o o «“ o o o © © o © © «— © © © © © © r* o o o © o o o o o o o o o o o oo o o o o o o o oo o o o o o o o o o o o o o T“ o o o o o o o o © © o «— o o © © © o o o © ©o © o © ©© © © © © o o o o o o o © o o o o o o o o o o oo o o o o o o o r“ o «“ o IP™- o o T“ o ©o © o »— o © © o o © © © »— © o © o © © © © r— o © «— IT™ IP* «-* w-’ «“ « «“ o o «— o o o o o o o o oo o o o oo oo T— « « V™ - <u bP *> !> <u iH w < o ^ O O o o «— o o '$ I Oh Oh o r^ ^ o o o o o c! C cs -C • " o o o o o «“ «“ V“ o o o oo o J r— ^ o ro o oo o o o oo «— o o o T— o «— «“ o o «” o o o o o oo r“ o o o o o o o o o o o *“ o o oo o <— r— o oo o o V— o «» o o o ooo o ooo oo o oo o »— o »— o « o o o o o o «— ,p». r- «-=“ T“ T“ o oo oo oo o oooo W“ Q o oo V— oo o Q ^ o^ o o o o ^ o o o o o o o 0 0^0 o r— o o o »“ «- oc o o o o o o o o o o o o o o o o o o «“ o o o o o «“ o o o o o «” oo ©© o o o o oo o © © o *— r™ oooo oo©©o o 10 10 10 o o o o o T™ o o o o •bb4 oooo o ooo ^ «— o o oo «*“ o o o o r— o o o o o o o oo oo o o o o o o o o o o T— o o o o o o o o o o o o o o o o o © o © o o © V— «“ o o © o o o o o o o «“ o o o © © © © o © ©© ©o «— «— »— T”* o o o o o CD CD CD ^D o o o «“ o oo o o »— o o o o «“ o o o o o oo oo o «“ oo o o o o o o o o o o o o o o o T” o «— o oo o «— o o oo o oo oo o o o o o o o o o o oo o «“ o o o ooo o oo oo o ^ > o oo o© oo o o o o © o o o «— o «— T«— «— r«" o £ adia seto clav i 1 i a u ^ 1 <u i f *-> spec peni • disj • . e? grin host af c o V ^ Sh B u £a c o > o E sz *-> (tl L. CD (B Lu S S Vo LU 3 ^ O Ph :e x: LU UJ i • o Si 3 t-> w u> o CD c o c o c Q r- *-> jyL CL jQ X LU OH HO V OG Drosophilidae, Molecular Analysis APPENDIX 6.2. Characters of the Six Selected Species Groups Listed in Table 6.1 Used General Area Cladogram An, antopocerus; affinidisjuncta; in the Construction of the adi, adiastola; bet, beta planitibia; Alp, Haw, 89 hawaiiensis; ?, alpha planitibia; Aff, missing data. Area An Adi Bet Alp Aff Haw Ancestor 000 001 Oil 000 00 ?? 01 01 ?? ??? }}} ??? 0000 0001 0011 0111 00 Oil 000 001 001 11 ?? Ill 111 oil 11 11 }}} ?p? ?p? nil nil ?? ?? ?p? ??? Ill ???? ?? ?? Ill Ill 111 ???? ?? 11 0‘ahu Molokai West Maui East Maui 1 East Maui 2 Hawaii Kona side Hawaii Hilo side APPENDIX 6.3. ??? Characters and States for the Intra-island Study Abbr, abbreviations used in Figure 6.6; four-base cutter data from DeSalle and Templeton (1992); six-base cutter data from DeSalle et al. (1986b); Seq, two characters from the DNA sequencing study (DeSalle and Templeton, 1992). Characters Area Abbr 4-base cutter 6-base cutter Seq ‘01a‘a 1 Kilauea H5 11000101100101010110101 11000101101101101000101 11100101100101101000101 00101100100001000000001 10000011100001000010001 10001100100001000000001 10000101101001101000101 10011110010001010111010 00011000010010000001010 1000101 00000 1 00000 1 1 000 01000000 01100110 01101110 01111010 11000100 01001111 01100110 00001100 01100110 10010101 11 ‘01a‘a 2 H4-1 H4-2 Pi‘ihonua 1 Pi'ihonua 2 Pi‘ihonua 3 Pi‘ihonua 4 Maulua Kohala Kona H3-1 H3-2 H3-3 H3-4 H2 HI K 11 11 11 11 11 11 11 00 00 / Evolution of Sarona (Heteroptera, Miridae) Speciation on Geographic and Ecological Islands ADAM ASQUITH Very often, however, we find species, extremely closely allied species, occurring habitually in the same locality and not geographically -R. C. In is most models of allopatric required for, isolated. PERKINS, L. 1913 speciation, a geographic barrier to gene flow or greatly facilitates, speciation (Mayr, 1963; Carson and Templeton, 1984; Carson, 1987a; Barton, 1988). Nowhere the pattern is of gene flow barriers so conspicuous, discrete, and repeated than in island The conventional model of allopatric speciation in an archipelago with an ontogeny, such as the Hawaiian Islands, is that when a new island is formed, it is colonized by founders from a species on the nearest island (Zimmerman, 1948; Carson, 1987b; see also Carson and archipelagoes. Clague, this volume. Chapter founders, and the process is 2). Speciation ensues repeated when among the next island these isolated is formed. This process produces a pattern of single-island endemics in which the sister species to This Islands any taxon occurs on the most proximate, older is the most simplistic model of speciation in the Hawaiian and probably explains the evolution of many groups of such as the orthopteran genus Banza (Tettigoniidae) publ.) island. (J. insects, Strazanac, un- and the heteropteran genus Kamehameha (Miridae) (A. Asquith, unpubl.). This pattern can be complicated by back-dispersal to older islands (Carson, 1987b; see also Lowrey, this volume. Chapter 11) or allopatric speciation within islands, such as 90 among volcanoes, isolation in Sarona (Zimmerman, kipuka to restriction 1948), ecological 91 communities (Howarth, 1991), or social or sexual selection (Kaneshiro, 1983; Carson, 1986; Otte, 1989). Whether they represent inter- or intra-island speciation, some degree of geographic processes mentioned above involve Theoretically, however, complete isolation is all the isolation. not necessary for speciation (Endler, 1977; Wright, 1982). Ecological factors such as microhabitat specialization (Kaneshiro (Wood, 1980; Wood and et al., 1973) or phenological partitioning Guttman, 1982) can prove to be barriers sufficient to precipitate divergence. versification of many Hawaiian fying insular evolution, rarely is insect extrinsic Although the ecological groups is di- often touted as exempli- speciation argued as being directly linked to these ecological radiations. Sympatric speciation by host plant or some Hawaiian Drosophila species (Kaneshiro et al., 1973; Carson and Ohta, 1981), but putative sister species typically have similar if not identical ecological traits (Carson and habitat shifts has been suggested for Kaneshiro, 1976). The potential for sympatric, ecological barriers among phytophagous is perhaps greatest insects that are host plant-specific (Bush, 1974, 1975; Bush and Diehl, 1982; Bush and Howard, 1986). phytophagous Hawaiian insects Many groups of have been noted for their radiations onto different host plants (Usinger, 1942; Zimmerman, 1948; Gressitt, 1978), but species relationships in most groups are poorly understood at best. In this chapter, I examine the evolution of the endemic phytophagous plant bug genus Sarona, With only one exception, Sarona species are island endemics, and each species feeds, breeds, single- and develops on a single species of host plant. Using cladistic analysis for the identification of sister taxa, I attempt to elucidate the relative roles of geographic versus ecological barriers in speciation in this genus. For the purposes of zoological is nomenclature none of the names for the permanent in Sarona mentioned in this chapter scientific record. ORIGINS The orthotyline (Table 7.1), is plant bug genus Sarona, with 40 known endemic to the Hawaiian Islands. species The North American genera Slaterocoris and Scalponotatus together have been identified as The outgroup to these three another North or Central American the sister group to Sarona (Asquith, 1994b). genera is unknown but is likely 92 ASQUITH TABLE 7.1. Species S. adonias Island Distributions and Host Plants of Sarona Species Host Island"* MoMLH S. akoko K S. alani H S. annae K S. antennata S. aula Mo L Family Plant*’ Metrosideros polymorpha Gaud. Chamaesyce Melicope Zanthoxylum Euphorbiaceae Rutaceae Rutaceae Pipturus Ilex Myrtaceae Urticaceae anomala Hook. &; Aquifoliaceae Arnott S. azophila L Nestegis sandwicensis (A. Gray) Degener Oleaceae et al. S. beardsleyi S. dakine M M S. flavidorsum H Korthalsella Viscaceae S. gagnei O Korthalsella complanata Viscaceae Nestegis sandwicensis Oleaceae Melicope Rutaceae } (Tiegh.) Engl. S. haleakala EM S. hamakua H S. hie O S. hiiaka K Dubautia menziesii (A. Gray) D. Keck Myrsine Melicope ? Melicope clusiifolia (A. Gray) T. Hartley Asteraceae Myrsinaceae Rutaceae Rutaceae & B. Stone S. iki S. koala S. kanaka kane kau kohana kuaana kukona S. S. S. S. S. H O EM EM H O O K Unknown Broussaisia arguta Gaud. Hydrangeaceae Cheirodendron Myrsine Dubautia Araliaceae ? Myrsinaceae Asteraceae Unknown Metrosideros Myrtaceae Rutaceae ? Melicope barbigera A. Gray S. laka K Claoxylon sandwicense Mull. Arg. Euphorbiaceae S. lanaiensis L Pipturus Urticaceae S. lissochorium O Broussaisia makua mamaki K Unknown S. S. S. S. maui mokihana H EM K Hydrangeaceae ? Pipturus Urticaceae Pipturus Urticaceae Melicope anisata Rutaceae (H. Mann) T. Hartley & B. Stone S. myoporicola H Myoporum sandwicense Myoporaceae A. Gray S. ( oahuensis Continued) O Coprosma } Rubiaceae Sarona TA B L E 7 1 . ( . Continued) Species Host Plant^ Island"^ O oloa S. 93 Family Neraudia Urticaceae melastomifolia Gaud. O Unknown S. palolo S. pittospori S. pookoi H Mo S. pusilla M Pipturus S. saltator K Melicope S. usingeri O Claoxylon S. xanthostelma O Unknown Pittosporum Pittosporaceae Unknown Urticaceae Euphorbiaceae sandwicense Kaua'i; O, O'ahu; Mo, Moloka'i; Rutaceae clusiifolia M, Maui; EM, ? East Maui; L, Lanai; H, Hawaii. (in some ^Confirmation of host plants was based on the collection of more than six adults cases, (i.e., many more) or the presence of immatures. ?, a questionable or unconfirmed host fewer than six adults have been collected from the plant). taxon, as there are no Asian or Indo-Pacific genera with any affinities to this group (Asquith, 1994b). Making the assumption that continent-to-is- land colonization is more likely than the reverse (Ward and Brookfield, 1992; Asquith, 1994a), then the ancestor of Sarona colonized the Hawaiian Islands from western North America. This places Sarona group of Hawaiian insects believed to be derived in a minority from North America, including the plagithmysine Cerambycidae (Gressitt, 1978), the oecanthine and trigonidiine crickets (Otte, 1989), and the metrargine Lygaeidae (Asquith, 1994a). The identification of the sister taxon of Sarona gives us a base from which to make a comparative analysis of insular versus continental evolution. For example, because of the allopatric barriers inherent in archipelagoes, it is sometimes argued that they facilitate and increase the rate of speciation (Mayr, 1942; Williamson, 1981), with the radiation of the Hawaiian Drosophilidae held as an example. The sister group to Sarona (Slaterocoris and Scalp onotatus) contains approximately 50 species; if sister taxa are of equal geologic age (Hennig, 1966), then the archipelago endemic Sarona, with 40 species, has not undergone a greater speciation rate than on its continental sister taxon. However, extinction rates islands are probably greater than in continental areas (Mac Arthur and Wilson, 1967), and additional species of Sarona were probably present ASQUITH 94 FIGURE 7.1. Sarona saltator. Dorsal habitus. on older, ally once-emergent islands, so that cumulatively, Sarona may actu- have had more species than the extant taxa alone indicate. In contrast to groups such as the Drosophilidae (Hardy, 1965), Cosmopterigidae (Zimmerman, 1978), and the plagithmysine Ceram bycidae (Gressitt, 1978), \vhich have undergone spectacular morphological radiations in the Hawaiian Islands, the evolution of Sarona has been morphologically conservative (Figures 7.1 and 7.2, illustrating the extremes of body form), with most interspecific variation occurring in male genitalic structures. Slaterocoris Also, the genitalic differences and Scalponotatus involve the same among structures species of and are of the same general magnitude as those seen in Sarona. Sarona has apparently undergone an extensive ecological radiation, however. Its continental sister taxon is known to breed on 13 genera of plants in four families, predominantly however, are in the Asteraceae (Kelton, 1968, 1969). Species of Sarona, known from 17 genera of plants in 14 families (Table 7.1). would be more convincing compare the radiation of Sarona with the sister group consisting of Sarona, Scalponotatus, and Slaterocoris, this relationship is not known. Most other genera of North Although it to Sarona FIGURE 7.2. 95 Sarona oloa. Dor- sal habitus. American orthotylines, however, typically have taxonomically restricted host plant associations of one to four families, certainly much narrower than those displayed by Sarona, Thus, the only distinct difference nental radiations in this group number of new host is between the archipelago and conti- the ecological colonization of a large plant families by Sarona, This is in contrast to Gagne’s (1983) contention that the diversity of host plant families that Sarona uses a consequence of its is having evolved from a polyphagous ancestor. PHYLOGENETIC ANALYSIS The genera Scalponotatus and Slaterocoris were used as a composite outgroup to polarize characters analyzed with the phylogenetics program HENNIG86 (Farris, 1988). in a two-state ters, Most of the character information format (Appendixes 7.1 and 7.2). Most was coded multistate charac- with the exception of characters 4 and 20, were linearly ordered. Other multistate characters were multiple-branched; because HENNIG86 does not support complex branching characters, a form of additive binary coding was used states The ordering of character was derived from hypothesized transformation series based on (e.g., characters 4 and 5). 96 ASQUITH FIGURE monious Numbers indicates Three variations (A-C) in 7.3. trees of tree Sarona species, based on analysis of a 31 -character data in small type are characters followed homoplasy or reversal. most conservative variation is Numbers trees to An asterisk node designations. The mokihana and and bb**' 5. alani arise from options (Fitzugh, 1989) with lengths of 62 and consistency indexes (Cl) of 0.64. Successive weighting procedures of the initial number of trees states. set. on next page) morphoclines. Analyses using the trees, all by character in circles are (A), because S. the basal polytomy. (Continued produced 910 topology from 56 equally parsi- to 56. This procedure was applied 910 trees reduced the to reduce the number of be examined and to choose those trees with the most consistent characters (Carpenter; 1988). among 56 trees involved the relationships of Sarona alani, S. mokihana, and S. annae to nodes 1 and 4 (Figure 7.3). Four of the 56 trees