aa. Ss eee = =: 2.2. —s if Tae ae eee SES > i... ase te i] nt WY Nia sian Taviesy sat : v Mi F bret tye he i NAW Sara Loi tei i i i wa Se see . 0 { ; h iia Wa i i Mee Bat Dida detn) Vota v hei pate Poles Ha Oey eeuleshy ty SPU 2 he vadae ¢ $ +) ty asa wis nad us hi pric sey teusbbead: (ited tena) APE bit i ae i ites ' ret! tf statis il hishes Ait ars 3 Seesees SE Mere epee =: = > Ss Sa ae aes : S = eet st etree ies : Steers SeseSesrcssist : Se me : : ae = : . = s =i : ys - t a = = a Stte a ee. cf at : Vea i bi Ht "yd = SS= nee 2 ae Speeerany T=y Sere et ce eee ee soe — Sas Eererees ae = SESS : Ss = ~~ fest ceatte: =s5 Sse Se eters eee we ae = tS S a Se eee HA tree! ui sin % tea aS Me tell ) i aan i} rt hd } Pub etayt AT et i i ae { iriLaiae ty in ie ut ie He ot i if iY “ 52225 ee eee ree 2 aS. —a SS Se in ff oie att has } tity 1s See eee Pe Se a : Fen ee eed = : : St = = ee reed : 5 —— ee Hai 1 ; ei ei esidh te Hi ub rant Bi (i) 4 H dit i 4 tins essai Hee ™ We f Miaahne if CESS a Og wn ae gattener 7 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES No. XXI The Galapagos Finches (Geospizinae) A Study in Variation By DAVID LACK SAN FRANCISCO PUBLISHED BY THE ACADEMY - 1945 TRUSTEES, OFFICERS, COUNCIL, AND MUSEUM STAFF OF THE CALIFORNIA ACADEMY OF SCIENCES BOARD OF TRUSTEES NORMAN B. LIVERMORE, President........cccccccccccecscccocccses Term expires 1950 BRUCE, ‘CORNWALL, ‘Vice-President: . suc ste dia sin ovaioie'e daisinintnvialeish sie o.n%0 Term expires 1946 BOUISE: As BOYD 55 0553 Hoicleiee nics Ooveisiein sjale's/avenusintelitiy wine! ones oie ...Term expires 1949 Fo WV) WALEED, TS wei G us Be celled wiele Aims Siar pia ai at RAID: Rtetentatiohe al a ote . Term expires 1948 CHARLES PAGE ey. bicianis ois. clerehtcs (h'e ein tie sinaje dom inaare any 'atwioleial ote SALE Term expires 1947 BM MACH ARLAND SG), 6:00 ete iesareo aie. oie iciele we ae totale aleteaiefaty einiotiie me\e/a sips Term expires 1946 — President of the Academy and ex officio member PRANCIS Po FARQUBARS: sicccins we csc phenome oe a ean cauite genre cites sim Term expires 1946 Treasurer of the Academy and ex officio member H. S. ANDERSON, Secretary to the Board OFFICERS AND COUNCIL F. M. MacFARLAND, President of the Academy E. P. MEINECKE, First Vice-President GEORGE S. MYERS, Second Vice-President CHARLES L. CAMP, Corresponding Secretary RICHARD M. EAKIN, Recording Secretary FRANCIS P. FARQUHAR, Treasurer CHARLES H. DANFORTH, Librarian ROBERT C. MILLER, Director of the Museum and of the Steinhart Aquarium SCIENTIFIC STAFF ROBERT C. MILLER, Director and Executive Curator Department of Botany.......... Auice Eastwoop, Curator Joun THomas HoweELL, Assistant Curator Lewis S. Rost, Research Associate Department of Entomology...... E. C. VAN Dyke, Honorary Curator Epwarp S. Ross, Assistant Curator FRANK E. BLAIsDELL, Research Associate IsaBEL McCrackEN, Research Associate in Hymenoptera Department of Ichthyology...... Wicsert M. CuHapman, Curator Department of Herpetology...... JosepH R. SLevin, Curator EIBEAEG Te Peon Saletan hae bce Cartes H. Danrortu, Librarian VERONICA J. SEXTON, Assistant Librarian Department of Ornithology é and Mammalogy ............ Rosert T. Orr, Assistant Curator JOSEPH MAILLiaRD, Curator Emeritus Mitton S. Ray, Research Associate A. S. Loukasnxin, Research Associate Department of Paleontology..... G. Dattas Hanna, Curator Leo Georce HERTLEIN, Assistant Curator Frank M. AnpErson, Honorary Curator M. VonseENn, Honorary Curator in Mineralogy WitraM B. Pirts, Honorary Curator of Gem Minerals Attyn G. Situ, Research Associate in Conchology A. M. Strone, Research Associate in Conchology Steinhart Aquarium ............ Rosert C. Miter, Director Joun L. Kask, Curator of Aquatic Biology Witeert M. Cuapman, Curator of Fishes ALviIn SEALE, Superintendent Emeritus 4CULPEPPER 1. SWENMAN I. (‘\ ABINGDON I. BINDLOE LOD JAMES I. NARBOROUGH I. BARTHOLOMEW I. JERVIS LH SEYMOUR IS. Dun, ANG, INDEFATIGABLE 1. ALBEMARLE I. BARRINGTON LQ} ee NEL CHARLES ee C4 Hoop I. Map of the Galapagos Islands The Galapagos Finches (Geospizinae) A Study in Variation BY DAVID LACK SAN FRANCISCO CALIFORNIA ACADEMY OF SCIENCES 1945 OCCASIONAL PAPERS NO. XxI OF THE CALIFORNIA ACADEMY OF SCIENCES Issued May 30, 1945 COMMITTEE ON PUBLICATION Dr. F. M. MacFaruanp, Chairman and Editor Dr. Cuarves L. Camp Dr. E. P. MEINECKE &, j { At ty \ y JUNG 1945 } Nv : / ion we AL MUSEUE CONTENTS INTRODUCTION : Procedure and Metco wledements The Galapagos Islands Section I. Taxonomy . . The Family ee Genera Species Geospiza saneneecae Gould’ Geospiza fortis Gould : Intermediates between Geospiza sone oat Games one Intermediates between Geospiza fortis and Geospiza scandens Geospiza fuliginosa Gould Intermediates between Geospiza fortis aad Copepiae hdcaone Geospiza difficilis superspecies . AN al Bs aor re Geospiza nebulosa Gould Geospiza scandens superspecies . Platyspiza crassirostris (Gould) Camarhynchus psittacula superspecies . Camarhynchus parvulus (Gould) Intermediates between Camarhynchus psitacula aqme ses ad Camarhynchus parvulus . Cactospiza pallida (Sclater and Seba) : Cactospiza heliobates (Snodgrass and Heller) . Certhidea olivacea superspecies . Pinaroloxias inornata (Gould) . Swarth’s New Species Section II. BREEDING BEHAVIOR . Pair-Formation : Territory-Aggressive Seis ; Functions of Song . . Postures—Secondary Sacral Chsnieiee Nest-Building Sexual Flights Copulation : Courtship Feeding . Order of Breeding Cycle ' Incubation and Feeding of the vie Nature of Adult Songs . d Song Differences and eoainticn Juvenile Song Call Notes Attacks on Alien pees vi CALIFORNIA ACADEMY OF SCIENCES Species Recognition . Field Experiments ; Breeding Behavior and iyahuitaan in tthe Canenicinee : Section III. Econocy . Food’ x!) Summary of F Wee ad F eeding Habis Drinks". : : Effects of Ban Nests . Habitats Discussion of Habitats Influence of Habitat on Geographical Diebueont Breeding Season . : i saline Molt Predators : Population Density SECTION IV. COLORATION Black Male Plumage in Cadac Platyspiza, aa Camarkynchus 59 Description s Seasonal and Ecological Viewtaton } Inter-Island Variation in Male Plumage Adult Molts . . ae The Change from Sinealed 4 Black Plumage g Hormonal and Genetic Control . i Evolutionary Significance Black Male Plumage in Parole: Rufous Under-Tail Coverts . Chestnut Throat Patch : Female Plumage (Omitting Carlier) Plumage Variations in Adult Certhidea ; Discussion of Plumage Variations in the Gecepicmn : Coloration of the Bill SECTION V. VARIATIONS IN BILL AND Wine Methods of Measurement Size Variations Due to Age . Sex Differences Specimens of Guesioceble identity, Differences between Island Populations of the Sane Se aie : Correlations between Bill and Wing Measurements Differences in Mean of Bill and Wing between Species Differences in Variability between Species . Inter-Island Wandering . OUI RP RE ae: SEcTION VI. GENERAL EVOLUTIONARY PROBLEMS . Previous Views . History of the Talands [Oc. Papers PACE No. 21] LACK: VARIATION IN GALAPAGOS FINCHES Comparison of the Different Galapagos Islands Evolutionary Factors Affecting the Geospizinae Evolution of Island Forms . Function of the Bill Differences aneee the eee: The Origin of Closely Related ona ; Hybridization 3... : Partial Isolation the ke F sae The Closely Related Species Origin of the Larger Units . ; An Evolutionary Tree of the Esioiies ; Comparison with Other Insular Species SUMMARY LITERATURE CITED MAIN TABLES . PiatEes ]—4, * ig xe 1 ape kas tS uy ah, & iN ' ‘ The Galapagos Finches (Geospizinae) A Study in Variation * INTRODUCTION PROCEDURE AND ACKNOWLEDGMENTS The present study of variation in the remarkable endemic finches of the Galapagos Archipelago is based, first, on an expedition to the islands from mid-December 1938 to early April 1939, and second, on a study of the skins of Geospizinae in American and British museums. The writer wishes gratefully to acknowledge grants made by the Royal Society, the Zoological Society of Lon- don, and the Elmgrant Trustees, Dartington Hall, which made this study possible. The expedition had three main fields of inquiry: (1) breeding behavior, (2) ecology, (3) hybridization. The first two, breeding behavior and ecology, concerning which extremely little was known before, were covered adequately, but far less suc- cess attended breeding experiments. Attempts to cross-breed the birds in aviaries on the Galapagos failed. Further, when the time came to capture birds for transportation to England, they proved much harder to catch than earlier in the year, so that only thirty- one individuals of four geospizid species were obtained. These traveled badly, so that on arrival in Panama it seemed extremely improbable that they would survive the journey to England. Ac- cordingly we cabled to Dr. Julian S. Huxley, and, with his ap- proval, the birds were transferred to the California Academy of Sciences in San Francisco. The field expedition consisted of: W. Hugh Thompson, who assisted the writer with the field study of the finches; L. S. V. Venables, who worked on birds other than the Geospizinae; Richard * The present paper was written by David Lack in the fall of 1939, which time marked the beginning of World War II. Working under pressure, the author left the rough manuscript in this country and returned to England to enter military service before the close of the year. During the year following he attempted to revise the copy but under the circumstances this proved a difficult task. Members of the staff of the California Academy of Sciences have made a number of changes and corrections and prepared the tables, histograms, and other illustrative features incorporated into the work. As would be expected, however, many inconsistencies and conflicts in style have arisen. The reader is asked, therefore, to bear this in mind when critically reviewing the results contained herein.—EbIToR Be 2 CALIFORNIA ACADEMY OF SCIENCES [Oc. PaPErs Leacock, photographer; and Mr. and Mrs. T. W. J. Taylor, bota- nists. The main party arrived at Guayaquil, Ecuador, in late No- vember 1938, and reached the Galapagos on December 14, leaving again on April 3, 1939. The writer worked on Chatham Island from arrival until January 29, January 30 was spent on Hood, January 31 to April 2 on Indefatigable, and April 3 on Tower Island. W. H. Thompson worked with the writer on Chatham until January 10, then on Indefatigable from January 11 until the middle of February, when he became ill. Up to the time of his illness, W. H. Thompson shared equally in the field observations on breed- ing behavior and ecology, and may therefore be considered equally responsible with the writer for the data upon which Sections II and III are based. The majority of the life histories had been worked out before he became incapacitated. Unfortunately, since my re- turn to England after the war started, I have been unable to get in touch with him for writing up these sections. However, I have had a copy of most of his original field notes from which to work. I should like to record here my gratitude to him for his invaluable assistance. I also thank L. S. V. Venables and the other members of the party for their assistance with some of the observations, and help in catching the birds. T. W. J. Taylor is responsible for the botanical identifications in Section III, and some of the habitat photographs. Thanks are particularly due to Dr. Julian S. Huxley, who took a continuous and extremely stimulating interest in the expedition and helped in innumerable ways. We must also thank the Ecuadorian government for permission to stay on the islands, and the British Consul and Vice Consul in Guayaquil for their assistance in arrang- ing our passage to the Galapagos, permits, customs, and other for- malities. The British Ornithologists’ Union, the British Ornitholo- gists’ Club, and certain private individuals further assisted the expedition by grants to L. S. V. Venables. We also express our appreciation to various inhabitants of Chatham and Indefatigable for their assistance, in particular Senor Cobos, in whose house we stayed on Chatham, and the Angermeyers and the Kiiblers on Inde- fatigable. On arrival in San Francisco, the thirty surviving captive birds were placed in the care of Mr. Eric C. Kinsey. Subsequently they were moved to aviaries in the California Academy of Sciences. When in the United States, the writer took the opportunity of No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 3 studying various collections of Geospizinae, for which thanks are due especially to the California Academy of Sciences,* with its large series which made the statistical investigation of variation possible, also to the American Museum of Natural History’ includ- ing the Rothschild Collection, the United States National Museum,’ Stanford University Museum of Natural History,* the Museum of Comparative Zoology,’ and later, the British Museum (Natural History ),° for their kind hospitality and for providing the writer with every facility for work. The Carnegie Museum,’ the Univer- sity of Michigan,*° the Museum of Vertebrate Zoology,’ the Field Museum of Natural History,"’ and the Philadelphia Academy of Natural Sciences’* kindly allowed me to examine their smaller series. Many American workers helped me to clarify my ideas in dis- cussion of various aspects, particularly the staff of the California Academy of Sciences, including Dr. Robert C. Miller, the late Mr. James Moffitt, Dr. Robert T. Orr, Mr. John Thomas Howell, and others; Dr. Alden H. Miller of the University of California; and Dr. Ernst Mayr of the American Museum of Natural History. After my return to England on September 30, in addition to the continued interest of Dr. Julian S. Huxley, I have particularly to thank Dr. R. A. Fisher, of the Galton Laboratories, for his help with the statistical data. The manuscript was sent to the California Academy of Sciences in May 1940, and I am deeply indebted to the editors for prepar- ing the final copy and for seeing the paper through the press, a task which, owing to the war, I could not undertake myself. THE GALAPAGOS ISLANDS The Galapagos Islands are a group of volcanic islands lying on the Equator, some six hundred miles to the west of Ecuador, to whom they belong. They have been described so often that little need be said here. For those seeking a general description, Murphy (1936, pp. 296-303) gives an accurate summary with references. The islands are in the coolest equatorial region in the world, being in the path of the Humboldt current coming up from the Antarctic. Southeast trade winds blow for three-quarters of the year, while from about December to March, but varying greatly in different 1¢.A.S.; 2A.M.N.H.; 2U.S.N.M.; 4Stanford Coll.; 5M.C.Z.; °B.M.; 7C.M.; 8U.M.; ' 9M.V.Z.; 19F.M.; 11P.A.N.S. 4 CALIFORNIA ACADEMY OF SCIENCES [Oc. Papers years, there is a rainy season with alternating periods of thunder storms and calms. The coastal belt of all the islands is thornbush in which tree cacti (Opuntia and Cereus) are conspicuous; the largest islands have humid forests at higher altitudes, and the tops of Chatham, Indefatigable, and Albemarle are covered with grass, ferns, etc. The map shows the positions and English names of the islands. Although all now have Spanish names, the English names are so much a part of the literature that it seems advisable to re- tain them. The distances between certain of the islands, which are of interest concerning the production of island forms, are given in Text Table 1. Table 1 Distances between various Galapagos Islands (expressed in nautical miles and, for larger distances, to nearest 5 miles) Wenman Tower Bindloe James Indefatigable Albemarle Chatham Hood Charles Crossman Culpepper 18.5 Tower 125 28 50 45* 70 Abingdon 1 45 3 40 38 105 James 50 32 10 10 Jervis 5 3 4 Daphne 3 4 42 Indefatigable 45* 50* 10 15 40 30 20 Duncan 12.5 | 10 Albemarle 15 42 10 15 30 4 Narborough 2 Barrington 25 30 Chatham 70 40 25 50 Charles 30 30 50 37 30 Gardner, near 4 Charles *Estimated from South Seymour, The islands first became famous through the visit of Charles Darwin in H.M.S. “Beagle” in 1835. Darwin was the first to collect the endemic Galapagos finches, and, as one can gather from The Origin of Species ..., they were one of the most important influ- ences in assisting him with his conclusions on evolution. The spe- cial peculiarity of these finches is that, unlike most insular birds, where one different form is found on each island, “each separate island of the Galapagos Archipelago is tenanted, and the fact is a marvellous one, by many distinct species...” (Darwin, 1888, No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 5 p- 355). This state of affairs has given rise to much speculation which is here discussed primarily in the last section. Further, the taxonomy of the group is abnormally complicated. In this respect, the present study owes much to the extremely sound and careful work of the late Mr. H. S. Swarth (1931), who cleared up most of the pre-existing confusion. The Galapagos finches are not, of course, the only interesting birds on the islands, but they are the only ones considered here; for a general account of other bird problems, see Swarth (1934). The present study is divided into the following six sections: I. Taxonomy II. Breeding behavior III. Ecology IV. Coloration V. Variations in bill and wing VI. General evolutionary problems SECTION I. TAXONOMY This paper is concerned primarily with variation in the Geo- spizinae, not with their nomenclature. As a basis for the latter, I have used the latest review of Swarth (1931, pp. 137-270), slightly modified by Hellmayr (1938, pp. 130-146). At first I intended to make no changes; but consistency in nomenclature is a valuable guide to the degree of variation, hence I have made certain modi- fications, mostly relating to the merging of island forms admitted by Swarth. In each instance, my larger series of measurements shows that there is a significantly different mean size for the popu- lations concerned; but in cases where less than 75 per cent of the specimens are safely distinguishable (and often less than 50 per cent are distinguishable) the forms have been merged. Otherwise, for consistency, one would have to name a number of new forms which differ to about the same extent. The main object of classifi- cation is convenience of reference, and all such incompletely dif- ferentiated forms seem best designated by the name of the species together with that of the island or islands on which they occur. Many species of Geospizinae show slight differences in mean size of bill and wing on almost every island, but with wide overlapping in extremes between birds from different islands. Thus it would be absurd to designate all such cases by formal scientific names. Although the present classification seems more consistent than any previous one, complete consistency is unattainable. The treat- ment of Geospiza scandens (sensu stricto) is particularly difficult. In the Camarhynchus psittacula group, I have provisionally re- tained Swarth’s nomenclature, although it is not altogether con- sistent with that adopted elsewhere, since further collecting is needed to establish the position of some forms. Evidence for nomenclatural changes is only summarized in this section, details, including measurements, being given later, while for full descriptions of the genera and species the reader is re- ferred to Swarth (supra cit.). THE FAMILY ‘‘GEOSPIZIDAE”’ Swarth (1929) placed the Galapagos finches, together with Pi- naroloxias of Cocos Island, into a separate family, but the charac- ters he gave hardly justify this. The only diagnostic feature, the [6] No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 7 long fluffy feathers of the lower back, is found in many other tropi- cal birds. On anatomical grounds, Sushkin (1925) and Lowe (1936) agree that the Galapagos finches are true Fringillidae, closely related to the West Indian Tiaris or Euetheia. Hence they can be placed in one of the existing subfamilies of the Fringillidae. Hellmayr (1938, p. 130), however, prefers to use the subfamily designation, Geospizinae, for the group, a term which for conven- ience is employed in this paper. GENERA The writer is in full agreement with Swarth (1931) as regards the application of generic names. The following field observations confirm these conclusions in several ways: 1. Certhidea has breeding habits almost identical with the other Geospizinae, being clearly of this group and not, as was once thought, a member of the family Mniotiltidae. 2. Cactospiza pallida was formerly united in the same genus with Geospiza scandens. Song and feeding habits fully confirm the view that scandens is related to other species of Geospiza, while Cactospiza seems most closely related to Camarhynchus. 3. Platyspiza has a different song and different feeding habits from Camarhynchus, with which it has often been united. SPECIES With reference to the arrangement and concept of species, Swarth (1931) cleared up most of the previous confusion, and, with few exceptions, his treatment is followed here. The distribu- tions are summarized in Text Table 2, in which the birds are classi- fied under “‘superspecies,”’ the convenient term proposed by Mayr (19316, p. 2) for related forms which replace each other geo- graphically, even though sometimes distinct enough to be named as separate species. The term closely corresponds with Rensch’s ““Artenkreis.” The Geospizinae have been so thoroughly collected that the table probably contains few omissions. There are large series of each form available from the various islands except in a few cases indicated in the table by an asterisk, or by a small letter in brackets. These exceptions probably represent stragglers from one island to another which have not established themselves. Some of these may possibly prove to be resident. 8 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS Table 2 Distribution of the Geospizinae F a z z 8 a8 : 4 aie ee ete ee eine be eae fe By O8) CORI Ve deh eee ee Species oO = & 4 OQ kh Leal Oo ie] oO o Geospiza magnirostris x x x x x x x x x Geospizs fortis x* xt es x x x x x x x x Geospiza fuliginosa x x x x x x x x x x Geospiza difficilis difficilis x x nm i debilirostris x x x " " septentri onalis x x Geospiza scandens x xe x x x x x x Geospiza conirostris conirostris x bs My propinqua x -) ui) darwini x Platyspiza crassirostris ac x x x x x x Camarhynchus psittacula psittacula x x x < me a iy habeli x x " " affinis x ack = r = pauper x Camarhynchus parvulus parvulus x* x x x x x x 2 " galvini x Cactospiza pallida pallida x x x x* " a striatipecta x Cactospize heliobates x Certhidea olivacea becki x x " " mentalis x ty n fusca x x Ls olivacea x x x U) ul bifasciata x f W luteola x a iW cinerascens x tN " ridgwayi x Pinaroloxias ijnornata x *Species has occurred on island, but perhaps only as a straggler. Geospiza magnirostris Gould Plate 1, figures 1 and 2 Not separated into island forms. The birds from Culpepper Island placed by Swarth (1931, pp. 149-150) in this species are here transferred to the conirostris group. The series in the Roth- schild Collection shows that the bill, although heavy, is much more No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 9 like conirostris than magnirostris in shape and the female plumage also differs from that of magnirostris. By referring them to magni- rostris, Swarth (op. cit., p. 206) also had to list G. conirostris pro- pinqua for Culpepper Island. Actually, all seem assignable to one variable form for which I revive the name Geospiza conirostris darwini Rothschild and Hartert. The latter authors (1899, p. 158) described this form as Geospiza darwini and noted its similarity in bill to propinqua. Later (1902, pp. 389-390) they considered it a subspecies of conirostris. One specimen of Geospiza magnirostris from Charles Island ex- amined in the collection of the U.S. National Museum appears doubtless to have been a straggler. Swarth (op. cit., p. 162) con- sidered it an abnormal example of fortis, but I cannot separate it from magnirostris. The three black males in the series of G. magnirostris collected by Darwin are all slightly larger than any collected since. Clearly a measurable evolutionary change has occurred since Darwin’s visit. This raises an issue in nomenclature which, while common in paleontology, is rare in ornithology, although liable to become more common in the future. Each species clearly has an extension in time as well as in space, and since modern magnirostris are of the same general type as those of Darwin’s day, it seems easier to retain the same name, even though there is no overlapping in measuremients. It is unfortunate that the island from which Darwin obtained these specimens is in dispute. The specimens were origi- nally attributed to either Charles or Chatham Island, in which case they represent an extinct form, as magnirostris is unknown from these islands today. I agree, however, with Swarth that it is more probable that the specimens came from James. Darwin undoubt- edly collected on James, as one of his G. scandens must have come from there. If magnirostris was as common on James at the time of Darwin’s visit as it is today, he could not possibly have over- looked it there. Darwin, however, also collected some smaller specimens of the magnirostris type which he described as G. strenua; these might have come from James. There is the addi- tional point that Darwin collected two specimens of Geospiza nebu- losa which I ascribe to an extinct form of the G. difficilis group. It is conceivable that the human inhabitants of Charles or Chatham Island directly or indirectly brought about the extermination of both Darwin’s G. magnirostris and G. nebulosa between the time of 10 CALIFORNIA ACADEMY OF SCIENCES [Oc. Papers his visit and that of the next collector; for at the time of Darwin’s visit Charles had but recently been colonized and Chatham had not yet been settled. Geospiza fortis Gould Plate 1, figures 3 and 4 This extremely variable species cannot be separated into island forms. Birds from the northern islands, north Albemarle, James, Jervis, Bindloe, and Abingdon, are smaller than those from the more southerly south Albemarle and Chatham, but the birds on Charles, the most southerly island of all, and also those on Inde- fatigable include all types between these forms. Fourteen specimens from Hood, one from Wenman (California Academy of Sciences), and one from Tower (Carnegie Museum) were presumably stragglers. Only the Academy expedition has ever collected this species on Hood, and all the males in this series are in immature plumage. Swarth (1931, pp. 152-154) gives a full list of the numerous synonyms of this species. The writer agrees with all of these ex- cepting Geospiza bauri, G. brevirostris, and G. nebulosa, which are discussed later. INTERMEDIATES BETWEEN Geospiza magnirostris AND Geospiza fortis Geospiza magnirostris and G. fortis differ from each other solely in size and in relative size of bill, overlapping in all measurements. Some specimens with mostly intermediate measurements can be identified if in one character they are typical of one or the other species. The length of the culmen is the most reliable guide. In Section V, I include a set of measurements of large fortis and small magnirostris, in order to illustrate the difficulties involved. In most cases, I have assigned the specimen to one or the other species, but in some instances where my decision is contrary to the opinion of other workers, [ do not claim complete certainty of my identifica- tions. Particularly difficult are three females and the type, a black male, of “Geospiza bauri’”’ from James Island, while two Bindloe Island specimens are similar. Swarth correctly identified the many specimens referred by later collectors to the form bauri as belong- No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 11 ing to the widely variable fortis. These James and Bindloe Island specimens, however, are more puzzling, since they greatly exceed the maximum normally attained by fortis on these islands, but are small for magnirostris. Possibly they are fortis which wandered from one of the southern islands, but they might be unusually small magnirostris, or hybrids. INTERMEDIATES BETWEEN Geospiza fortis AND Geospiza scandens Swarth synonymized Cactornis brevirostris Ridgway with Geo- spiza fortis. It is represented by a partially black male from Charles Island in the U.S. National Museum collection, while a female from Indefatigable in the Rothschild Collection (No. 516928; see also Rothschild and Hartert, 1899, p. 159) can pos- sibly be placed with it. These specimens are intermediate between fortis and scandens. Clearly Cactornis brevirostris is not a valid species, but whether the specimens are aberrant variants of either Geospiza fortis or G. scandens, or are hybrids between them, is uncertain. Another unsexed specimen from Indefatigable in the U.S. National Museum (No. 77756) is also intermediate between fortis and scandens, but is much nearer to scandens than is the type of Cactornis brevirostris. A fourth specimen, a black male from Charles Island, is also intermediate between these two species. For measurements and details, see Section V. Geospiza fuliginosa Gould Plate 2, figures 1 and 2 The rare Geospiza fuliginosa minor Rothschild and Hartert of Abingdon and Bindloe islands was described on the basis of its smaller wing, but overlapping with typical fuliginosa, however, is too great in the writer’s opinion to justify this. Swarth placed the four specimens taken on Wenman Island under minor. Although all have small wings, the male has an unusually large bill, and this bird and one of the females show slight plumage variations in the direction of Geospiza difficilis septentrionalis of Wenman Island. Perhaps they were stragglers, and conceivably hybrids, not neces- sarily first generation, with septentrionalis. 12 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS INTERMEDIATES BETWEEN Geospiza fortis AND Geospiza fuliginosa The differences between Geospiza fortis and G. fuliginosa are more clear-cut than are those between G. fortis and G. magnirostris. Birds intermediate in character in the Rothschild Collection were described by Ridgway as Geospiza harterti. The type of this form, from Chatham Island, is possibly an aberrant specimen of fuligi- nosa, but the bill is abnormally deep and the wing is long. It is well below the normal minimum size of fortis on Chatham Island, and does not suggest a hybrid between the two. Many specimens from the tiny islet of Daphne (between Inde- fatigable and James) conceivably represent an extremely small form of fortis, as Swarth suggested on the basis of specimens from Daphne in the collection of the California Academy of Sciences. It seems highly improbable, however, that fortis should be so very much smaller here than on either of the two neighboring islands only a few miles away. More probably, they are a hybrid popula- tion of G. fortis & G. fuliginosa. Birds from Crossman Island, off east Albemarle, are somewhat smaller than those from Daphne, and may also be of hybrid origin. Two others labeled “east Albe- marle” seem referable to the Crossman Island form, and perhaps were obtained there. These are all discussed in the section on measurements. A black male from Hood in the Rothschild Collection (No. 517663) has an unusually deep bill for fuliginosa, but, especially since fortis is normally absent from Hood, it is probably to be regarded as an aberrant specimen of fuliginosa. Two others labeled “harterti”’ from Hood Island are typical fuliginosa, the same apply- ing to two specimens from Gardner Island near Charles. The type of Geospiza dentirostris Gould was, I believe, correctly assigned by Swarth to fortis. In the British Museum there is a black male from Charles Island which was originally labeled fortis and later dentirostris, the measurements of which approach “harterti.” Also, there is a similar specimen in the Museum of Comparative Zoology (No. 134650). These I consider to be unusually small examples of fortis. An unsexed bird from Duncan Island in the British Museum (No. 99.9.-1.293) has an abnormally deep bill but is probably an aberrant specimen of fuliginosa. Hence, except on the tiny islets of Daphne and Crossman, intermediates between fortis and fuliginosa are quite rare. | No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 13 Geospiza difficilis superspecies Plate 2, figures 3 and 4 Swarth included here the species acutirostris Ridgway of Tower Island, difficilis Sharpe of Abingdon Island, debilirostris Ridgway of James and Indefatigable islands, and septentrionalis Rothschild and Hartert, the last separated into the typical race on Wenman Island and Geospiza s. nigrescens Swarth on Culpepper Island. Hellmayr (1938, pp. 132-134) lists all of these as subspecies of Geospiza difficilis, since they replace each other geographically and are quite similar in appearance. Geospiza acutirostris and G. diffi- cilis are not separable by plumage, and, although the former tends to have a smaller bill, measurements overlap so much that even sub- specific separation seems unjustified. The two forms of septentrio- nalis also overlap too much to justify separation. Accordingly, three subspecies of G. difficilis are admitted here, namely G. d. dif- ficilis of Tower and Abingdon islands, G. d. debilirostris of James and Indefatigable islands, and G. d. septentrionalis of Culpepper and Wenman islands. One black male debilirostris from south Albemarle in the Roth- schild Collection was possibly a straggler, as no others have been collected there, but it may well prove to be resident locally. High up on Narborough Island, Beck (cf. Swarth, 1931, p. 181) re- ported finding a nest of debilirostris, but no specimen was obtained. However, in the Stanford Collection there are two black males labeled fuliginosa which seem rather too large for this species and they have straight culmens. They may consequently be referable to a form of difficilis, although it is curious that they should more closely resemble difficilis (sensu stricto) of Tower and Abingdon islands than they do debilirostris of James and Indefatigable. Ad- ditional specimens are needed in order to settle this point. During our stay on Indefatigable we found no Geospiza debili- rostris. Either we overlooked it or it had disappeared. Swarth (MS) saw quite a number in 1932. At the same locality, Fortuna, he recorded Nesomimus as being extremely abundant. In 1939, we found the latter to be extremely scarce there near human habi- tations. Conceivably the clearing of the forest and other human activities of recent years have indirectly caused the disappearance _of Geospiza debilirostris and the decrease of Nesomimus. One specimen of G. fuliginosa from James Island (Rothschild 14 CALIFORNIA ACADEMY OF SCIENCES [Oc. ParEers Collection, No. 517369) has the straight culmen characteristic of difficilis; so also has a black male (Stanford Collection, No. 5251) and a female (University of Michigan, No. 88889) from Abing- don; but all these specimens are typical of fuliginosa in measure- ments, so have been assigned to this species. They are much too small for difficilis, but could conceivably be dwarf difficilis or of hybrid origin. Some Chatham Island specimens of fuliginosa have nearly straight culmens, and, since they also approach difficilis from Tower Island in size, might be difficult to distinguish if they occurred on the same island. A black male from James Island (Rothschild Collection, No. 517702) seems intermediate between Geospiza fortis and G. d. debilirostris, so that I cannot positively assign it. Geospiza nebulosa Gould The type of this species in the British Museum (No. 37.2.21.400) was collected by Darwin; hence its locality is uncertain, and there is no real evidence that it came from either Chatham or Charles Island, as was once thought. No later speci- mens have ever been assigned to this form, but another of Darwin’s specimens (No. 55.12.19.20) labeled “Cactornis scandens” seems to me to belong here, as it is extremely similar in size and shape of bill, wing length, and probably in color if allowance is made for fading. Swarth placed this second specimen with “Cac- tornis assimilis’ (see Geospiza scandens), and synonymized the type of nebulosa with fortis. I cannot agree with either of these conclusions. The specimens seem to me to fall in the difficilis group, agreeing particularly in the shape of the bill, including the almost straight culmen. Their dimensions, however, particularly as regards depth of bill, are too large for them to be assigned to any of the present forms of difficilis, and they also lack chestnut on the wing bar and undertail coverts. Either, as with Geospiza magni- rostris, measurable evolution has occurred since the time of Dar- Win’s visit, or they represent an extinct form of difficilis from Charles or Chatham Island, where this species is unknown today. Darwin’s collection also includes a typical specimen of debiliros- tris (originally labeled Geospiza strenua) which probably came from James Island, an occurrence which favors the second sugges- tion, although the first is not ruled out. If I am correct in relating G. nebulosa to the difficilis group, No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 15 then there is necessity for replacing the latter by the former as the specific name of the group. This seems inadvisable, however, since nebulosa is represented by only two specimens, and further is now extinct, so that its affinities cannot certainly be known. Geospiza scandens superspecies Plate 3, figures 1, 2, and 3 Swarth recognized the following forms of Geospiza scandens (Gould): abingdoni (Sclater and Salvin) of Abingdon Island, rothschildi (Heller and Snodgrass) of Bindloe Island, scandens (Gould) of James Island, and intermedia Ridgway of Indefati- gable, Duncan, Albemarle, Barrington, Chatham, and Charles is- lands. These names are correlated with measurable differences in general size and size of bill, but, as Swarth states, the differences are so slight that all the birds might be placed under one name. The latter procedure is adopted here, since in most cases less than 75 per cent of the specimens can be identified with certainty. The specimens from James Island can be separated readily from those from Bindloe Island, and there is almost no overlapping, hence they should bear separate names. However, the birds from Abing- don, Jervis, Indefatigable, etc., bridge the gap between the two. This would present no difficulties if James and Bindloe were at the two ends of a chain of linked forms, but this is not the case. James and Bindloe are adjacent islands, and the intermediate forms occur both to the north and to the south of them, making any consistent nomenclatorial treatment impossible. Under the circumstances, I consider it more practical not to use subspecific names, but to refer to the species followed by the name of the island, as adopted in other species where there is much overlapping between island forms. This procedure was adopted by Mayr (1932a, pp. 5-10) for Foulehaio c. carunculata which presents similar difficulties. As noted by Swarth (1931, p. 200), Chatham Island specimens of G. scandens tend to be shorter-billed than other “intermedia,” and, if the latter name is retained, the Chatham birds might be con- sidered as a separate form. Incidentally, their scarcity in collec- tions is because the plant Opuntia, to which scandens is closely restricted, does not occur where the collectors have worked on Chatham Island. We found the birds common locally, and not too easy to distinguish in the field from fortis; a difficulty not expe- rienced with the longer-billed scandens of Indefatigable Island. 16 CALIFORNIA ACADEMY OF SCIENCES [Oc. Papers The three specimens assigned to the scandens group collected on the voyage of the “Beagle” and now in the British Museum are the type, a black male with so small a bill that it almost certainly came from James, the specimen which I have assigned to Geospiza nebu- losa, and the type of Cactornis assimilis Gould. Swarth could not safely assign the last to any present form of scandens, but stated that it came nearest to Bindloe Island specimens, with which judg- ment I agree. There are three puzzling specimens, two in the Rothschild Col- lection (Nos. 517703, 517815) and one in the British Museum (No. 99.9.-1.358), two from James Island and one from Chatham Island, all in immature plumage. All were labeled scandens but their bills are so small that I at one time thought they were debili- rostris. This was wrong, as indicated both by their coloration and the proportions of their bills. I am not yet altogether satisfied that they are immatures of scandens, although this seems to be the most probable conclusion. Also placed in the scandens superspecies are Geospiza c. coni- rostris Ridgway of Hood Island, G. c. propinqua Ridgway of Tower Island, and G. c. darwini Rothschild and Hartert of Culpepper Island (for the last, see note under G. magnirostris). The female plumage, the relatively less deep bill as compared with other coni- rostris, and the song, show clearly that propinqua is closely re- lated to scandens. In conirostris, the females are typically much darker than in propinqua, but simply carry further the tendency found in the scandens group. The song is distinctive. The bill of the immature is not deep and shows a clear resemblance to that of scandens. The plumage of darwini is somewhat distinctive but shows resemblances to that of propinqua, while the bill is of the conirostris type. The writer agrees with Swarth in relating coni- rostris to scandens, which it replaces geographically. Stresemann (1936) suggested that conirostris was related to fortis. One speci- men of conirostris {rom Gardner Island near Charles Island in the Rothschild Collection was doubtless a straggler (see Rothschild and Hartert, 1899, p. 159). Platyspiza crassirostris (Gould) Plate 3, figure 4 Not separated into island forms. The type specimen has dis- appeared. The British Museum specimen, incorrectly listed by No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 17 Sharpe as the type (cf. Swarth, 1931, p. 209), is typical of Cama- rhynchus psittacula and Swarth suggests that it is an aberrant ex- ample of psittacula. Camarhynchus psittacula superspecies Plate 4, figures 1 and 2 This includes Camarhynchus psitiacula Gould of James, Inde- fatigable, Barrington, and Charles islands, C. habeli Sclater and Salvin of Abingdon and Bindloe islands, C. affinis Ridgway of Albemarle Island, and C. pauper Ridgway of Charles Island. Ridgway separated the form of psittacula on Charles Island as Camarhynchus townsendi, but Swarth provisionally synonymized the latter with the former, with which I fully agree as their meas- urements widely overlap. Both psittacula and pauper, two species of the same superspecies, occur on Charles Island, as is later dis- cussed in detail. Two specimens of psittacula from Chatham Island in the Roth- schild Collection were presumably stragglers. Gifford (1919) re- ported seeing this bird on Chatham. We saw none there. Some puzzling specimens from James, Jervis, Indefatigable, and Seymour islands were described by Ridgway as Camarhynchus incertus. Swarth referred them to affinis. In bill and plumage, they resemble affinis, not psittacula. Possibly all are stragglers of af- finis from Albemarle Island, but there are so many that affinis is perhaps establishing itself on these islands. Alternatively, psit- tacula may be more variable than supposed. It is possible that Camarhynchus psittacula and C. affinis meet on Duncan Island. Further collecting is much needed. The species here grouped with psittacula might better be clas- sified as subspecies of one species, but it seems inadvisable to make any alteration until the problem concerning affinis has been cleared up. Camarhynchus parvulus (Gould) Plate 4, figure 3 Camarhynchus parvulus salvini Ridgway of Chatham Island is larger than the typical race. Two specimens of C. p. parvulus from Wenman Island, and three from Abingdon, all in the California Academy of Sciences, were perhaps stragglers, but at least on Abingdon they may well be resident. 18 CALIFORNIA ACADEMY OF SCIENCES [Oc. Papers INTERMEDIATES BETWEEN Camarhynchus psittacula superspecies AND Camarhynchus parvulus Eight specimens from south Albemarle Island, six in the col- lection of the California Academy of Sciences (cf. Swarth, 1931, p- 232) and two at Stanford, seem intermediate between Cama- rhynchus affinis and C. parvulus, while two specimens from Charles are intermediate between pauper and parvulus. They might be ab- normal specimens of either species, or hybrids. Cactospiza pallida (Sclater and Salvin) Plate 4, figure 4 Swarth recognized the following races: pallida on James, Inde- fatigable, and Duncan islands, producta (Ridgway) on Albemarle, and striatipecta Swarth on Chatham. James Island specimens are gray, Indefatigable specimens olive, perhaps enough to justify re- vival of the name hypoleuca (Ridgway) for the James Island birds; but the Albemarle specimens present difficulties. Cactospiza p. producta was separated from C. pallida by its smaller bill, but their measurements overlap too much to justify this. North Albe- marle birds are gray, like pallida on James Island, while south Albemarle specimens are olivaceous, like those on Indefatigable. Only two races of pallida are admitted here, pallida and striati- pecta. One specimen (California Academy of Sciences Collection) from Charles Island is perhaps a straggler. Cactospiza heliobates (Snodgrass and Heller) Restricted to the mangrove belt on Albemarle and Narborough islands. Certhidea olivacea superspecies Plate 4, figure 5 This form was divided by Swarth into island species which, as he states, might equally have been called subspecies, as was done by Hellmayr (1938, pp. 142-145), and is followed here. The races of C. olivacea Gould are: becki Rothschild of Culpepper and Wenman islands, fusca Sclater and Salvin of Abingdon and Bind- loe islands, mentalis Ridgway of Tower Island, olivacea of James, Indefatigable, Albemarle, and Duncan islands, bifasciata Ridgway No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 19 of Barrington Island, luteola Ridgway of Chatham Island, cineras- cens Ridgway of Hood Island, and ridgwayi Rothschild and Har- tert of Charles Island. The Abingdon and Bindloe Island birds differ in wing length, but not sufficiently to justify separation. Certhidea olivacea on James, Indefatigable, and Albemarle islands differ almost as much from each other as do some of the forms separated by Swarth, so, as the latter states, the names salvini Ridgway and albemarlei Ridgway might possibly be revived for the birds on Indefatigable and Albemarle islands, respectively. Pinaroloxias inornata (Gould) Plate 4, figure 6 Found on Cocos Island. In bill and food habits, it comes closest to Certhidea, in plumage to Geospiza difficilis septentrionalis. SwarRTH’s NEw SPECIES The two specimens of Camarhynchus conjunctus from Charles Island, the one Camarhynchus aureus from Chatham Island, and the one Cactospiza giffordi from Indefatigable Island, all described by Swarth, are, in my opinion, merely variants, not valid species. The first and second seem intermediate between Camarhynchus and Certhidea, the third is either a diminutive Cactospiza, but with buff on the upper breast, or is intermediate between Cactospiza and Certhidea. While their characters suggest hybridization, field ob- servations make this seem unlikely, for the different genera are quite different in their habits, and hybridization between genera usually occurs only in birds which form an indefinite pair-bond. Hence their position is uncertain. On Indefatigable Island, the writer saw a bird identical with Cactospiza pallida in plumage, habits, and call notes, but similar in size to Cactospiza giffordi, although lacking the buff throat patch. Possibly it could be referred to giffordi. SECTION II. BREEDING BEHAVIOR Detailed studies of breeding behavior of the following nine species were made by W. H. Thompson and the writer: Geospiza magnirostris, fortis, fuliginosa, scandens; Platyspiza crassirostris; Camarhynchus psittacula, parvulus; Cactospiza pallida; and Cer- thidea olivacea. Each of us tended to concentrate on certain species, but each of us worked on all species at least sufficiently to check on the other’s main findings. In what follows, all species are treated together as their behavior patterns are extremely similar excepting where otherwise indicated. The Geospizinae are so indifferent to human presence that they can usually be observed from a distance of from ten to twenty yards without causing any disturbance. Owing to the variability in plum- age, bill, color, and song, it was often possible to recognize indi- vidual birds, hence, to make day to day studies of them. The early stages of the breeding behavior were worked out chiefly in the intermediate forest of Chatham, where breeding was just starting on our arrival. At this time observation was facilitated because the trees were not yet in leaf. Apart from the general problem of speciation in the Geospizinae, the observations are of interest in that they contribute additional information on the life histories of tropical birds. PatrR-FORMATION The breeding cycle follows a typical territorial passerine pat- tern. Outside of the breeding season many of the birds, if not all, are in flocks. Shortly before the beginning of the heavy rains in December or January, the mature birds begin to leave the flocks. In many cases, an unmated male was seen defending a territory, some days later acquiring a mate, which later nested with it in the territory. This seemed the normal method of pair-formation. In a few cases, the newly formed pair disappeared, and on several occasions members of a pair were seen to arrive together, the male claiming territory. Shift of territory shortly after pair-formation is not rare in other territorial birds. The possibilities of some pair- formation in the non-breeding flocks, and some birds remaining paired during the non-breeding season, are not excluded, particu- larly in Platyspiza crassirostris, where no flocks were seen (Gifford, [ 20] No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 21 1919, also notes this). Wandering pairs of P. crassirostris were common at the beginning of the breeding season, and an occasional wandering pair of Geospiza fortis was also seen. When a fortis or fuliginosa in female plumage enters the terri- tory of an unmated male, it is sometimes chased out, but may also elicit a sexual reaction. Hence the unmated male probably is able to distinguish potential mates from other birds, including males in immature plumage, which were indistinguishable to us. However, the factors involved are unknown. As in warblers and buntings (Howard, 1929), but not in the British robin (Lack, 1939), the interval between pair-formation and laying, which may last a few days or at least a month, is occu- pied by much courtship of the types described later. TERRITORY-AGGRESSIVE BEHAVIOR Each male or pair has a territory. As with most territorial pas- serine birds, it is usually impossible for a man to drive out the male in possession of the territory. As the observer approaches, the bird retreats, but when it reaches the edge of its territory it will not retreat farther, and eventually flies back past the observer into the center of the territory. The boundaries of territories were sharply defined where the birds came in contact with neighboring males. Here both males displayed aggressively at each other, and often each owner flew along his boundary with slow, descending, round-winged flight, singing hard. In all species except Certhidea, the male bird normally drove out vigorously all intruding individuals of its own species of both sexes. In Certhidea, the male normally drove out only trespassing males, the defending male showing excited movements but rarely attacking a strange female. The reason for this difference is not known. The females of Geospiza vigorously drove out other fe- males, but never attacked intruding males in black plumage. In two instances, the strange intruding male even went up to the nest without the owning female attacking. Males in immature plumage were thought to be abundant, but were indistinguishable to the observer from females, and probably to the female Geospiza too, since the latter drove out all birds in female plumage. When a male fortis, which was mostly in streaked plumage but showed par- tial black feathering, trespassed, the occupying female fortis started to attack it, but the male resisted instead of retreating, and the 22 CALIFORNIA ACADEMY OF SCIENCES [Oc. Papers female then desisted and fed quietly near it. The female’s mate shortly came up and drove the other male away. In a second case, after showing some aggressive behavior toward the strange male, the female actually received food from it as in courtship. In a third instance, the intruding male, in the absence of the owner, proceeded to visit the nest and display to the female, who did nothing. Females of Platyspiza, Camarhynchus, and Cactospiza possibly behave like those of Geospiza, since they were seen attacking in- truders of their own species in female plumage but did not attack distinctively plumaged male intruders. However, owing to the scarcity of distinctively plumaged males on Chatham, where most observations were made, the data are not really sufficient. One fe- male Platyspiza chased a black male Geospiza fortis from her nest. The female Certhidea did not usually show aggressive behavior toward intruders of either sex, but apparently chased them rarely. Away from the center of the territory, intruders of either sex feeding on the ground are sometimes tolerated in their territory by members of the established pair, which may even feed with them. This was noted once in Cactospiza pallida and several times in Geospiza fortis and in G. fuliginosa. Two incidents are of special interest. Near the edge of, but definitely inside of its own territory, a pair of fortis was feeding on the ground together with a strange fortis in female plumage. After some minutes, the pair flew to the center of the territory, and a moment later the intruder followed the pair (a social response) and was promptly driven out. On an- other occasion, a male fuliginosa was feeding on the ground ten yards inside the territory of the neighboring male, which fed peace- ably beside it. After an interval, the intruder stopped feeding and flew to a piece of lichen on a bush six feet off the ground. The owner promptly attacked and drove it out, singing loudly, and when it attempted to return drove it out again. The significance of lichen clumps in nest display is discussed later. By its action, the intrud- ing male became a sexual rival instead of a food companion. A parallel case involving song-posts is described for Anthus pratensis by Venables (1937, p. 77). Most territorial birds keep a strict lookout for intruders. This was not the case in the Geospizinae. Especially in Geospiza fortis, G. fuliginosa, and Camarhynchus parvulus, intruding males were sometimes in the territory for several minutes, and might even have No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 23 visited several display nests, before they were detected by the de- fending male, who might have been only ten yards away, but look- ing in another direction. When an intruder was seen, the defending male normally flew to it, at which the intruder usually flew off and was pursued out of the territory. If the intruder did not retreat, the owner normally flew around to perch directly in front of it, and if it still did not retreat, gripped its beak. The intruder soon retired. Song and threat postures also were used in attack, particularly between full- plumaged males, and especially between neighboring males on the edges of their territories. These remarks apply to all except Cer- thidea. When the occupying male saw an intruder in the territory, it usually uttered a special part of its song. An intruding male usually gave the same phrase in reply, at which the owner promptly attacked and drove it out. Silent birds, usually females, were sometimes, but not often, attacked. In Certhidea, too, the fights, particularly between neighboring males, were apt to be fiercer, and the birds not infrequently fell interlocked to the ground, peck- ing hard. The male normally confined his singing, attacking, and nest- building to his own territory. The pair did much of its feeding inside the territory, but also considerable outside of it. Sex-chases, male pursuing female, often took the members of the pair far outside of their own territory. These sex-chasing intruders par- ticularly excited territorial owners. FUNCTIONS OF SONG Male song serves the same functions as in other territorial spe- cies, that is advertisement to females and to neighboring males. It is loud in the unmated male, also in newly paired males. It declines conspicuously during incubation. It is especially loud during and just after attacks on intruders and sometimes when the male displays to the female. The song is delivered from perches in the territory, also, in all except Certhidea, during a special flight, slower with more rounded wings, usually descending from a higher branch to a lower. This song-flight is also common when going to or from the nest during display-building. In Cactospiza pallida, as in the other species, the song-flight is normally from one branch to another below the canopy of the forest trees. However, where this species occurs in the six- to ten-foot scrub in the higher hills, 24 CALIFORNIA ACADEMY OF SCIENCES [Oc. Papers the same movements result in an aerial song-flight above the low trees, which makes the bird conspicuous for a long distance. PosTURES—SECONDARY SEXUAL CHARACTERS The postures of the Geospizinae are often vigorous and excited but, like their songs, are unspecialized. Sometimes the wings were fully expanded and held out, the bird at times turning from one side to the other in this position, which occasionally involved turn- ing its back on the bird to which it was posturing. Sometimes the expanded wings were vibrated slowly up and down. Sometimes the wings were only partially expanded, somewhat lowered and quiv- ered rapidly. Sometimes the tail was spread laterally but not erected. These postures were extremely similar in all the species, excepting that in Certhidea wing-quivering was normal, but full ex- pansion with the motionless wings was seen only once. Loud song frequently accompanied the posturing. This posturing was used equally in aggressive behavior against intruders and in sexual behavior between the members of the pair. The sexes had similar postures, although males postured much more often than females. Usually birds have different attitudes for ag- gressive and sexual behavior, but this is not the case in the Geo- spizinae. Furthermore, fledglings begging food from adults used the same posture with quivering, dropped wings or with wings fully expanded and moved up and down. The postures exhibit no special plumage markings. The only striking color on the otherwise black male is seen on the under-tail coverts which are tipped with whitish. These are not displayed. When feeding on the ground, birds of both sexes are apt to spread and elevate the tail, which in the male reveals the under-tail coverts. The color of these coverts either was acquired incidentally, or would seem to have lost any function which it once possessed. The males of some forms of Certhidea olivacea show an orange patch on the throat and breast. The normal postures do not spe- cially display the breast, but twice we saw a male stretch its breast somewhat when another species came to its nest. At least in C. o. olivacea of Indefatigable and C. o. luteola of Chatham, the orange patch seems to have no function today. Males of Geospiza frequently breed in immature plumage, or in partially adult plumage. The behavior and postures of such individuals seemed identical with those of males in full plumage. No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 25 The same applied to males of Platyspiza, Camarhynchus, and Cacto- spiza, in which breeding in immature plumage is progressively more regular. NeEst-BUILDING In the Geospizinae, nest-building is closely linked with display, as occurs in some other birds which do not have enemies at the nest, as, for instance, the herons. The male often starts several nests, usually based on the remains of old ones from the previous year, which are to be seen everywhere in the forest. Early in the breeding season, building is sporadic, the male may pick up ma- terial then drop it and do something else, or may visit several nests one after the other. The male frequently sings when building. Nest- visiting and -building are greatly stimulated by the presence of a female, and their part in pair-formation was particularly well shown when a female Cactospiza pallida settled on the boundary between the territories of two unmated males. Both promptly sang violently, displayed, and made building movements at lichen tufts. The female left shortly. Many similar incidents were noted. When the female is near, the male may enter one nest after an- other, often singing and posturing in the entrance or going in and out as if to persuade her to enter. If the male does not yet have a nest of his own, he will often visit the nest of another species of Geospizinae. For example, a displaying male Camarhynchus par- vulus entered the nest where a female Cactospiza pallida was incu- bating and was chased out. A male Certhidea was often seen to visit the nest of a Geospiza fortis, or the reverse; the smaller species offered no resistance; the larger, if present at the time, chased away the intruder. Many similar happenings have been recorded in all the other species. During two days’ observation, one courting male Geospiza fuliginosa was seen to visit eight nests regularly, some of which were also visited by four other species, namely G. fortis, Platyspiza crassirostris, Camarhynchus parvulus, and Certhidea. When no nest of any kind is present in the territory, the male fre- quently visits lichen tufts, the most common nest material, and acts as if these were a nest; occasionally clumps of other vegetation have served, and in one case a strip of linen on a bush. The male frequently deserts the first nests. In later stages, the female takes part in building, and sometimes has been seen to build alternately in two of the male’s display nests. However, the pair 26 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS soon concentrates on one nest. Even then they might desert and start yet another nest, or more usually take over the partially built and deserted nest of another species. In some instances, the whole of the nest eventually used for breeding was built by the female. SEXUAL FLIGHTS As in the warblers and buntings (Howard, 1929), sexual flights are a prominent feature of courtship. The male pursues the female, often singing, and sometimes carrying building material. The wings may be spread so that they appear broader than usual. The writer agrees with Howard in interpreting these flights as incipient attempts of the male to copulate, the female taking flight before the male reaches her. This does not mean, however, as Howard implies, that the male is necessarily able to copulate. Initiatory movements, sometimes called symbolic, are frequent in the court- ship of many birds and sexual flights come in this category. In one case, the male was clearly ready to sex-chase but not to copulate, for when a female magnirostris invited the male to copulate he did not respond; the female then flew off and the male promptly com- menced sex-chasing. Not infrequently, sex-chases were initiated by the female. The sexual posture of the stationary male with flutter- ing wings may be regarded as an initiatory movement of sexual flight. COPULATION Copulation is normally initiated by the female assuming a posi- tion similar to that of many other passerine birds, motionless with bill sometimes pointing upward. A similar posture with bill up- ward was occasionally seen in the courting male. On two occasions, a male Certhidea attempted to mount a female which had not as- sumed the invitation posture, once when the female was temporarily entangled in lichen. CourTsHIP FEEDING As in many other species, the male feeds the female not only during incubation but also throughout courtship. In all species the food was first swallowed by the male, and then, almost immediately afterward, regurgitated to the female, often with much saliva. In Geospiza, berries and more rarely caterpillars were the chief food passed. In Certhidea and Cactospiza, insects were primarily used. No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 27 In one instance, a male and female Platyspiza crassirostris were seen to pass the same caterpillar from one to the other several times, but such an exchange was not otherwise observed. Often early in the breeding season no food was passed; the birds simply touched bills. Such billing is another example of an initiatory movement in courtship behavior. Normally the courtship feeding had no connection with other courtship, but in one instance a male Camarhynchus parvulus fed the female while copulating (W. H. Thompson). Conceivably the vertical bill of the female in the copulatory attitude is a relic of a former feeding habit which has now almost disappeared in the be- havior of the Geospizinae. For a general discussion of courtship feeding, see Lack (1940a). ORDER OF BREEDING CYCLE In the buntings, Howard (1929, pp. 1-27) described a definite sequence of breeding behavior. In the Geospizinae, some of the phases are variable in their order of occurrence. For instance, one male claimed a territory with song but had no mate and no nest interest, while a neighboring unmated male was building rapidly. Another male already had a mate, but neither sex showed any in- terest in building. Again, in some cases courtship feeding was seen within an hour or two of the formation of the pair, in another instance it was not observed before incubation started. INCUBATION AND FEEDING OF THE YOUNG Incubation was studied but slightly because almost all of the nests were out of reach. For all the species observed, four was the usual number of eggs; one pair of fuliginosa reared five young. In one magnirostris nest, the first, second, and third eggs were laid on consecutive days, while the time of laying of the fourth was not recorded. The young hatched, one on the 14th, one on the 15th, and two on the 16th days after the laying of the first egg, and they left the nest twelve days after the last one hatched. The female occasionally brooded the young for five days after the first young hatched. Only the female incubates. At intervals the male appears, calls her off the nest and feeds her. When some young magnirostris were newly hatched, the male not infrequently fed the female on the nest itself, but this was not observed during incubation. The 28 CALIFORNIA ACADEMY OF SCIENCES [Oc. ParEers female tended not to incubate during the heat of the day. The tem- perature of the eggs inside the nest was taken on two occasions: at 2:00 P.M. sun-time when the air temperature was 84° F. the eggs were at 98° F.; on the following day half an hour after noon with the air at 83° F. the eggs were at 96.5° F. In neither instance had the female been incubating for several hours. Clearly she had little need to. Both sexes feed the young in the nest. In all birds observed, the female fed the young more frequently than the male, coming approximately every 20 minutes with food during the early morn- ing and evening, at greater intervals during the heat of the day, while the male came every 30 or 40 minutes. During the first day or two, the male sometimes fed the female, who then fed the young. Feeding was by regurgitation. The food for the young in magni- rostris, fortis, and fuliginosa consisted primarily of sphingid cater- pillars, at that time extremely common in the coastal zone, and also berries. In the higher parts of the intermediate zone caterpillars were scarce and berries were the main diet. These would seem to provide inadequate nutrition for the young, but much saliva passed with them, and this perhaps contained nutriment. Certhidea fed its young primarily on small insects. Once the young leave the nest they are fed exclusively by the male. When two out of four young magnirostris had left a nest the female merely seemed bewildered by the food-begging of these two perched close to the nest, and she fed only the two inside. The posture of the young begging for food is with fluttering and ex- panded wings like that of courtship display. As soon as the fledg- lings leave the nest, they tend to posture at any moving bird, in- cluding any species of Geospizinae which happens to come near. Young fuliginosa have been seen to posture at an adult magniros- tris, also vice versa, and young scandens have been seen to posture at both. Evidently the external stimulus which releases food- begging is a simple one. The main factor in the parents’ location of the young is therefore supplied by the latter, probably correlated with the complete absence of enemies; the fledglings are much more conspicuous in their behavior than European passerine birds could afford to be. The female disappears when the young leave the nest. During the preceding few days there is sometimes a resumption of sexual activities between the members of the pair, chiefly sex-chasing and No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 29 rarely copulation. It is, therefore, possible that the female goes off and starts a new brood, but since she always moved immedi- ately away from the first brood this was not determined. Beebe (1924, pp. 260-261) recorded two parents feeding a fledgling and at the same time frequenting a nest with one egg, but this obser- vation does not fit into the breeding pattern as observed by us. NATURE OF ADULT SONGS The songs of all of the species are of primitive pattern, un- musical, with no complex phrases. The tone of the larger species, at its best, approaches that of Agelaius and Xanthocephalus, two North American icterids. Omitting Certhidea and Platyspiza, the typical songs of every species could be represented by some such words as “‘tchur tchur tchur tchur” (monosyllabic type) or “tchur-wee tchur-wee tchur- wee’ (disyllabic type). In addition, other phrasings occur in all of the species, such as: “tchur-lee-tee-tee,” “tee-chur,” and others. The phrases and number of notes, that is the “pattern” of the song, being so variable, the chief differences between the various species lie in the quality, and to a less extent in the time intervals between the notes. Hence, any attempt to describe the differences by means of human syllables breaks down hopelessly. With practice, it is possible for a field observer to identify correctly most of the adult songs heard, as indicated below: Geospiza magnirostris—Of a slower tempo than that of other forms of Geospiza and more melodious and forceful. Some indi- viduals were indistinguishable from fortis or scandens. Those on Tower Island seemed indistinguishable from those on Indefatigable. Geospiza fortis.—A generalized type of song, usually harsh, occasionally more melodious and usually stronger than in fuligi- nosa. Some individuals were indistinguishable from magnirostris, fuliginosa, scandens, and occasionally from Cactospiza, and one came extremely close to Platyspiza. Geospiza fuliginosa.—Typically like fortis, but weaker. This is the most generalized song of all the Geospizinae. While many individuals were distinguishable, some were not separable from fortis, scandens, or Camarhynchus parvulus. A disyllabic form of song was more common than the monosyllabic. — Geospiza difficilis difficilis——Although this species so closely resembles fuliginosa in appearance, its song is distinctive, consist- 30 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS ing of short and more feeble notes, often with a hissing note in the middle, so that in timbre it approaches the song of Certhidea. Geospiza difficilis debilirostris—(Not heard by us) Swarth (MS) states that the song is very distinct “like a tiny siren running down.” Geospiza scandens.—Typically a rapid succession of notes. In tone quality usually intermediate between fortis and fuliginosa, often, but by no means always, with more syllables in each phrase. Usually all syllables were of the same quality, but a “disyllabic” type also occurred. Some individuals were indistinguishable from magnirostris, fortis, fuliginosa, and Cactospiza. Geospiza conirostris propinqua.—The few individuals heard had a song consisting of a rapid succession of notes, indistinguish- able from the song of scandens on Indefatigable. Geospiza conirostris conirostris——Rather distinctive song, usu- ally fairly musical. Some individuals were perhaps indistinguish- able from fortis, scandens, and Platyspiza crassirostris. Platyspiza crassirostris——More distinctive than most. Several rather musical notes run into a grinding “churr,” this phrasing being remarkably constant and specific although in quality it varies slightly, particularly the vigor of the “churr.” As in other species, a “hiss” is sometimes added to the song phrase. Camarhynchus psittacula——Two main types of song, one soft, rather slow, disyllabic, and the second, a rapidly repeated succes- sion of harsh notes, similar to Cactospiza but less strong. A typical “churr” and a high-pitched, rapid succession of “see” notes are often added to the song phrase. Camarhynchus parvulus—Two main types of song, one a rap- idly repeated succession of similar harsh notes, the second, various combinations of two types of harsh notes succeeding each other rapidly, the phrasing and accenting being extremely variable. Some types were reminiscent of one call of the great tit (Parus major). The second type is usually harsher and shorter than the disyllabic type in Camarhynchus psittacula, but musical varieties occur. A “churr’” and a long-drawn “see” may be added to the song, but less commonly than in psittacula or Cactospiza pallida. Some individuals have been confused with psittacula or Geospiza fuliginosa. Near the summit of Indefatigable nearly all the par- vulus had musical songs, whereas in the intermediate zone they were normally harsher. Riya. > No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 31 Cactospiza pallida.—A rapid succession usually of seven to eight notes, often more, with either all of one type or of two. Typi- cally loud, and usually more musical than in other species, but harsher in some individuals. Phrasing and tone were extremely variable. A “churr” and long-drawn “see” were joined with the song much more frequently than in other species. Occasional indi- viduals have been confused with Geospiza fortis, G. scandens, Camarhynchus psittacula, and, in pattern but not in volume, with Camarhynchus parvulus and Certhidea. Certhidea olivacea.—A rapid succession of notes thinner than any other Geospizinae except Geospiza d. difficilis from Tower Island. In Certhidea olivacea luteola of Chatham Island, the main song is reminiscent of the European wren (Troglodytes troglo- dytes). In Certhidea olivacea on Indefatigable, there were typically far fewer notes in the phrase, which was a little louder and more harsh than in the Chatham birds, being not unlike the song of parvulus. One rapidly repeated succession of notes was common to the song in both islands. Sonc DIFFERENCES AND EVOLUTION As might be expected, the songs of the different genera are more distinctive than are those of closely related species, but even here considerable overlapping occurs. Points of special interest (see later discussions) are: the song of Geospiza conirostris propinqua seems indistinguishable from that of G. scandens of Indefatigable, one of several links between these forms. The song of G. c. coni- rostris is typically more differentiated, and bears no particular resemblance to that of scandens or propinqua. The song of G. d. difficilis from Tower Island is quite distinct from that of fuliginosa, a species which it resembles greatly in appearance, and from which it probably was evolved. Platyspiza crassirostris has the most spe- cialized and constant song of any form, with no particular resem- blance to the song of Camarhynchus, the genus in which it was usually placed until Ridgway and Swarth erected a new genus for it. Cactospiza pallida was formerly classified with Geospiza scan- dens, owing to their superficially similar bills. From its plumage, Swarth and others placed it next to Camarhynchus, and its song, as well as its feeding habits, strongly support this. In Certhidea the song is, like all its characters, more distinctive than that of the other Geospizinae, but shows obvious geospizid affinities. In general, if 32 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS song could be more objectively recorded, it might be a valuable guide to classification. The function of most bird song is advertisement by the unmated male to females in search of mates, and to rival males. The songs of closely related species are normally distinctive, which facilitates the female in finding a mate of her own species. The overlapping between the songs of the different species of Geospizinae is highly unusual among birds. Of special significance is the fact that the patterns of the songs overlap, the differences being mainly in qual- ity, and very possibly non-adaptive. Song may not be of funda- mental importance in keeping these species apart, although the possibility remains that, while the songs overlap to human ears, the birds detect differences. However, in a wide experience with European and other birds neither W. H. Thompson nor the writer have experienced similar difficulties in identification. JUVENILE SONG Juvenile song occurs in all species. It is sometimes like adult song, but tends to be more variable and so is less easily identified. Often “churr,” “hiss,” “zip” and other odd notes are introduced, the whole being roughly strung together and often uttered quietly, almost “conversationally” and directed at no special object. This song is frequently heard from males in streaked and partially streaked plumage while still in the non-breeding flocks, and has been heard from young which could not have been out of the nest for more than a month. Also some streaked male fortis acquiring their territories have begun with such song but much louder, grad- ually changing to an adult song in the course of several days. CaLL NOTES The calls of all the species are, like their songs, simple and generalized. A note which might be written “tchra” in some species, varying toward “keu”’ in others, is used in aggressive, sexual, and social behavior. The female sometimes repeats this note rapidly and in a higher pitch, especially under sexual excitement. W. H. Thompson describes a female “song” in Cactospiza pallida, but if we heard the same note, the writer considers that it was simply this excited call. All species also have a “hiss” and various clicking notes. The “hiss” is heard in fortis and fuliginosa particularly when the pair meet at the nest. No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 33 In Certhidea, the songs and calls seem of greater importance than in the other species. The pair constantly keeps in touch by calling, and an intruding male is also recognized by his voice. ATTACKS ON ALIEN SPECIES All of the species watched were, at times, seen to attack indi- viduals of most of the other species of Geospizinae. Occasionally aggressive individuals also attacked mockingbirds (Nesomimus), warblers (Dendroica petechia), and flycatchers (Myiarchus magni- rostris). As in other territorial birds, attacks on alien species were usually sporadic and rarely continued for long. Although aggressive behavior was often directed at other spe- cies, this was not observed for sexual behavior except that a mated female fortis once followed a singing unmated male fuliginosa, and the female’s mate, who was near by, ignored this. SPECIES RECOGNITION Frequently, upon seeing an individual of a strange species enter its territory, the male would fly down as if to attack, coming around in front as if to grip its bill, and then the whole behavior would collapse. This was observed in Geospiza magnirostris against GC. fortis; in fortis against fuliginosa and Cactospiza; in fuliginosa against fortis; and in Camarhynchus psittacula against parvulus. It happened so often that it seemed clear that the species recognized each other primarily by differences in the bills. It should be noted that the plumage of the species concerned is so similar that recog- nition from behind is difficult or impossible for the human observer. This behavior fits Tinbergen’s views (1939) on the importance of first and second reactions in social encounters in birds, the first reaction released by a very general signal, the second by a more precise and specific one. The importance of specific recognition through bill difference is discussed in a later section. FIELD EXPERIMENTS A caged male and then a caged female fortis were placed at different times in the territories of two mated pairs and one un- mated male fortis. The wild birds came down to the cage, showed mild interest, but soon left. A caged male and female fuliginosa were presented to four wild male fuliginosa with a similar negative result. This is in marked contrast to the behavior of the British 34 CALIFORNIA ACADEMY OF SCIENCES [Oc. ParEers robin (Lack, 1939) and chaffinch (Lack, 1941) which attacked such caged specimens. Experiments with mounted birds were more effective. Mounted specimens of a black male fuliginosa and of a fuliginosa in female plumage, a similar pair of fortis, and a Camarhynchus parvulus in female plumage were perched as naturally as possible near dif- ferent fuliginosa nests, one mount being presented at a time. Ex- periments on the British robin (Lack, 1939) showed that a bird sooner or later ceases to take notice of a mounted specimen. When a second specimen is presented to a bird, its reaction tends to be less intense, irrespective of the nature of the specimen. Hence an increase in the intensity of the reaction to the second specimen can safely be correlated with the nature of the specimen but a decrease cannot. These points were borne in mind in the following experi- ments. Experiments were performed on 16 male fuliginosa, 14 mated, two unmated, and on the 14 mated females, all building or courting birds. Behavior of male-—O{ 16 males presented with a black male mount, seven showed neither aggressive nor sexual behavior, ap- pearing either to be uninterested or slightly alarmed; two showed mild excitement; seven showed marked aggressive behavior, three of which also showed sexual behavior, in two cases mild, in the third more intense. When presented next with the female mount, one showed mild aggressive behavior, three attacked vigorously, three others showed strong sexual behavior, and the other nine took no special notice. The males always attacked by pecking at the bill of the mount. Males probably attacked male and female mounts about equally. The apparent decline in intensity of aggressive reaction to the female mounts is interpretable through the waning in the intensity of reaction with repetition, as already noted. The three males which reacted sexually to the male mount first attacked the specimen by standing on its back; from this they passed into definite copulatory actions, but two of them quickly desisted. One of the latter flew straight from the back of the male mount to copulate with his own female, and both were subsequently uninter- ested in all mounts, including the female mount. The third bird, after a vigorous attack, copulated vigorously with the male mount; then when presented with the female mount it again copulated vig- No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 35 orously; presented with a female fortis it continued to display sex- ually but without copulating; presented again with a female fuligi- nosa there was an immediate increase in sexual display; presented with the male fuliginosa mount again, it showed no interest; pre- sented with the female fuliginosa, it again displayed sexually. Hence, although this last bird copulated with a male mount, it re- sponded sexually in a greater degree to the female mount. This point could not be tested with the other two males since they so quickly lost interest in all mounts. Of the other two males which copulated vigorously with the female mount, one had shown no interest in the male mount, the other had attacked it vigorously, but showed no sexual response. The reaction of one male to a mounted female fortis has been described. The other two birds which reacted sexually to the female fuliginosa mount were also presented subsequently with a female fortis. One of these, like the bird already described, showed a mild sexual reaction, including an attempted copulation with the fortis mount, but this was not so intense as that delivered either before or subsequently to a mounted female fuliginosa. The other showed no reaction to the female fortis though subsequently it again reacted sexually to the female fuliginosa mount. A bird whose aggressive reaction to the male mount had been strong was also tested with a mounted male fortis and attacked it; the same bird a moment later did not attack a live male fortis which perched close to the nest. Behavior of female.—In nature the female does not attack black male intruders, hence it is not surprising to find that of the 14 females presented with a male mount, 13 showed no aggressive behavior, and the other delivered only two small pecks. No female attempted to court the male mount. When presented with the female mount, five females attacked vigorously. Three of these were subsequently presented with a female fortis mount, at which two promptly ceased to react aggres- sively, although one subsequently reacted strongly to a female fuliginosa; the other attacked the female fortis and subsequently attacked the female fuliginosa more vigorously. Of two birds presented with a mounted female Camarhynchus parvulus, one attacked quite as violently as it attacked a mount of its own species, although the same bird showed no aggressive be- havior to the female fortis mount. The other bird showed a more 36 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS intense reaction to its own mount than to the parvulus. The reac- tion of a third individual to a mounted parvulus was, as in its response to a mounted female Geospiza fuliginosa, negative, but it is of interest since a few moments later it attacked a live parvulus which happened to alight near by. Experiments with Geospiza fortis—Three males were tested with rather indefinite results. One pulled at the bill of the black male fortis mount, probably trying to feed it, and later mounted, probably an incipient copulatory action. One of the others postured excitedly, either in aggression or in courtship; the other attacked mildly. SUMMARY OF EXPERIMENTS ON Geospiza fuliginosa WITH MounTep Birps 1. Variability—Although the mounted specimens were always placed in similar situations, various birds reacted quite differently to them. Similar individual variation occurred with the British robin (Lack, op. cit.). 2. Sex discrimination.—Too much must not be argued from experiments with mounted birds, since the situation presented is unnatural, and it seems probable that a motionless bird is an im- portant factor in the external situation normally releasing copula- tory behavior, which may help to explain the action of the male fuliginosa which copulated with male mounts. Some sex discrimi- nation by plumage clearly exists, since two males which reacted sexually to the female mount showed no sexual behavior toward the male mount; also one male which showed some sexual behavior toward the male mount, showed more to the female mount. The two other males which copulated with the male mount subsequently lost interest in all mounts. In their experiments with various sex- ually dimorphic North American birds, Noble and Vogt (1935) found complete sex discrimination of mounted specimens except by juvenile birds. Hence sexual differences in plumage are per- haps less important in the lives of the Geospizinae than in most sexually dimorphic birds. It should be remembered that many males breed in plumage indistinguishable from that of the female. 3. Species discrimination.—Here also too much must not be expected from the experiments. Probably a foreign species would not normally give the signal for copulation. Also in two instances the wild bird’s reaction toward the mount of a foreign species was No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 37 different from its behavior toward a living individual of the same species a few moments later. The experiments, however, show that even under artificial con- ditions a wild fuliginosa does differentiate, to some extent at least, a member of its own species from fortis, even when this difference is reduced, for practical purposes, to a difference in bill, for the fortis and fuliginosa specimens were stuffed so that the difference in general size was small. The critical experiments occurred with three male and two fe- male fuliginosa. The three males (a) reacted sexually to a mounted female fuliginosa; (b) presented with a mounted female fortis, one did not respond, the other two reacted sexually but more mildly; (c) presented again with a mounted female fuliginosa, all three reacted sexually and more intensely than to the mounted female fortis, despite the fact that reactions to mounted specimens wane with repetition. The two females (a) vigorously attacked a mounted female fuliginosa; (b) presented with a mounted female foriis, one ceased to attack, the other attacked but not so strongly as be- fore; (c) presented again with a mounted female fuliginosa, both attacked it vigorously, the second bird more vigorously than before. The example of a third female is suggestive but not conclusive. The bird vigorously attacked a mounted female fuliginosa; when a mounted female fortis was substituted it ceased to attack; after this it lost interest in all mounts. To summarize, some sexual discrimination by plumage and some species discrimination by bill differences occurs in Geospiza fuliginosa, but neither is absolute, at least under the artificial con- ditions of the experiments. BREEDING BEHAVIOR AND EVOLUTION IN THE GEOSPIZINAE All the breeding habits of the different genera and species of Geospizinae are extremely similar. The acquiring and maintaining of territory, pair-formation, threat and sexual postures, display- building, nest-visiting, sexual flights, courtship feeding, the attitude in copulation, the position and structure of the nest, and the share of the sexes in incubation and the feeding of the young follow the same pattern. The songs and call notes are similar and are used in the same sorts of situations. This is all the more remarkable since the species of Geospiza are finch-like in general habits, Cactospiza climbs trees, and Certhidea is so like a warbler that it was at one 38 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS time classified with the North American warblers (Compsothly- pidae). There are slight differences between some of the species, particularly Certhidea, while Cactospiza seems more excitable than some of the others. In the Geospizinae, breeding habits have been far more conservative in evolution than feeding habits, which are described later. In general, patterns of breeding behavior seem a valuable guide to bird classification, although it is not true of all bird groups that the breeding habits are more conservative than the feeding habits. The gallinaceous birds, the birds of paradise (Para- diseidae) and the Icteridae, illustrate the opposite. The extremely generalized nature of display and song in the Geospizinae is of great interest. Perhaps correlated with this is the frequency with which male birds breed before they attain their full male plumage. It is possible that these facts are all correlated with a general decrease in the intensity of sexual selection in small island populations, but the nature of this correlation, if any, is obscure. SECTION III. ECOLOGY This section is based upon detailed work on the islands of Chat- ham and Indefatigable by W. H. Thompson and the writer, and brief visits by the writer to Hood on January 30 and to Tower on April 3 in 1939. Foop The differences between the bills of the various species of Geo- spizinae were formerly attributed to differences in diet. Snodgrass (1902, pp. 380-381) was the first to investigate this problem seri- ously and concluded that in the genus Geospiza “there is no corre- lation between the food and the size and shape of the bill.’’ The only difference he found was that “birds with small bills eat only small seeds; birds with large bills eat both small and large seeds.” This statement was based chiefly on a comparison of the small- billed fuliginosa with the larger-billed fortis, magnirostris, and conirostris. The present expedition did not make detailed stomach analyses, but instead concentrated on field observations of the food taken and especially on the manner of feeding. The results for each species are summarized, supplemented by references to Gifford (A919). Geospiza magnirostris.—Mainly native fruits and berries, such as those of Tournefortia, Maytenus, Cordia, and Croton; nectar from the flowers of Cordia lutea and others; the staminate flowers of Croton scouleri; the young green leaves of trees; various seeds; large green sphingid caterpillars and occasionally other larvae. Geospiza magnirostris feeds both in trees and upon the ground, the latter especially outside of the breeding season. Beebe (1924, p. 266) records it as feeding on ants. Swarth (MS) records magni- rostris digging in the sand on James Island. Geospiza fortis—Food and feeding habits similar to those of magnirostris; the same kinds of fruits, also the fallen fruits of manzanillo (Hippomane mancinella), flowers (chiefly tree flow- ers), leaf buds and young leaves, caterpillars, seeds, and occasional small insects. Beebe (supra cit.) states that ants are taken occa- sionally. Geospiza fuliginosa.—Has the most varied diet of any of the [ 39 ] 40 CALIFORNIA ACADEMY OF SCIENCES [Oc. Parers Geospizinae. Its food includes almost everything taken by fortis excepting the larger fruits and seeds, like those of the manzanillo. Grass seed, which is taken only occasionally by fortis, and not re- corded in magnirostris, forms an extremely important item of diet. Geospiza fuliginosa occasionally visits the flowers of Opuntia which was not observed in fortis, and also regularly eats green buds and young leaves, other flowers, small green caterpillars, and pecks about on the ground, probably taking small seeds and insects. Beebe (supra cit.) records that spiders are sometimes taken by members of this species. Gifford states that the bulk of the food consists of small seeds but also records pulp of fallen Opuntia blossoms, leaves, introduced fruits, carrion, refuse, and marine worms taken below high-tide mark. Geospiza difficilis debilirostris—Not observed by us; said to feed on the forest floor (Gifford, op. cit., p. 238). Geospiza difficilis septentrionalis—Harris (1899, pp. 89-91) and Drowne (1899, pp. 107-111) record this bird as walking about on the backs of nesting boobies (Sula dactylatra granti Rothschild) and picking off insects; also feeding on carrion. Gif- ford (op. cit., pp. 241-242) records that they feed chiefly on the ground where they scratch and dig, also on leaves, cactus, and, in one instance, on the blood of a shot bird. Geospiza scandens.—Karly in the breeding season this species is most commonly seen inserting its long bill into the flowers of Opuntia. It also takes local fruits as recorded under magnirostris, green sphingid caterpillars and other larvae, and not infrequently is seen eating ants. Grass seed is also taken and it sometimes takes scattered grain. Gifford (op. cit., pp. 239-241) records it as feed- ing on the soft pulp of cactus, and the introduced oranges. Geospiza conirostris conirostris from Hood.—Mostly seen pick- ing about on the ground, also taking young Acacia leaves and, like scandens, probing Opuntia flowers. Gifford (op. cit., pp. 225-226) records it as foraging on the rocks and beach. Geospiza conirostris propinqua from Tower.—Seen probing Opuntia flowers and eating Croton fruits. Platyspiza crasstrostris—Feeds mainly on blossoms (particu- larly those of trees), buds, and young leaves. At times comes to the ground, but there it eats young leaves of herbaceous plants, and not the foods which attracted fortis and fuliginosa. Fruits in- cluded those of Passiflora, Croton, Cordia, and others as in the Ree ——_ eee No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 41 larger forms of Geospiza. Large green sphingid caterpillars were taken occasionally, but the bird dealt with these in a clumsy man- ner, not gripping them with its feet, or trying to kill them before eating. Platyspiza crassirostris does great damage to the introduced fruit, Carica papaya . Camarhynchus psittacula.—Similar in feeding habits to C. par- vulus (see below), searching leaves for insects, excavating in branches, and taking the fruits of Cordia, etc. Gifford also records it feeding on heliotrope blossoms. Camarhynchus parvulus——C. p. parvulus from Indefatigable and C. p. salvini on Chatham have similar food habits. In habits, C. parvulus somewhat suggests a titmouse, examining twigs, bark, crannies, and on Indefatigable particularly, the leaf clusters of Scalesia, for insects. It examines the ground, turning over litter for seeds and insects, and also takes nectar from flowers, including the introduced tobacco, young buds and leaf centers, and green caterpillars. Camarhynchus parvulus also digs trenches in branches with its bill to get at larvae and beetles. It comes to grain, but usually carries it to a tree to eat it. Cactospiza pallida.—Almost exclusively insectivorous. In the coastal zone, it is found particularly on Opuntia, where it searches the crannies, etc., and excavates trenches for insects. In the inter- mediate and humid zones, it inspects leaf clusters, especially dead leaves, like Camarhynchus and also inspects and excavates wood for boring insects. On the ground, it turns over fallen leaves and also digs in the soil under rocks particularly for beetles. It climbs up and down the trunks and branches of trees like a nuthatch. In certain respects, it also resembles a woodpecker. When the latter has excavated in a branch for an insect, it then inserts its long tongue into the crack to withdraw the insect. Cactospiza pallida lacks the long tongue, but achieves the same result in a different way. Having excavated, it picks up a small twig, or in the coastal belt an Opuntia spine, one or two inches long, and holding it length- wise in its bill, inserts the twig into the crack, dropping it to seize the insect as it emerges. Sometimes the bird has been seen to reject one stick, if it was too short or too pliable, and it may break off another. Sometimes it carries the stick or spine about with it as it visits one tree after another, probing it into cracks as it goes. This remarkable habit, first recorded by Gifford (op. cit., pp. 253-257) and fully confirmed by W. H. Thompson and the writer, 42 CALIFORNIA ACADEMY OF SCIENCES [Oc. ParErs is one of the few recorded uses of tools in the animal kingdom outside of man. The nearest parallel in birds is the use of fruits for “‘bower-painting” by Ptilonorhynchus violaceus (Gilbert, 1939). The origin of this habit of Cactospiza is obscure. It is un- likely that it arose through the manipulation of nest material, as the latter is essentially pliable. It is probable that the spines of Opuntia, owing to their suitable lengths, were first used for probing by these birds, since Cactospiza frequents this plant in the coastal zone. Cactospiza heliobates——Reported by Snodgrass (1902, p. 367) to feed on insects. . Certhidea olivacea.—Feeds much like a warbler, searching the leaves and twigs, also the ground, for insects and making short aerial excursions for flying insects. It seems to feed almost exclu- sively on insects, but at times takes nectar from flowers and young green leaves. The female tends to feed closer to the ground than does the male. Gifford (op. cit., pp. 220-223) noted C. o. cine- rascens feeding on the rocks below high-tide line. Pinaroloxias inornata.—Probably feeds on insects in a manner similar to Certhidea. Gifford (op. cit., p. 242) reports, “This spe- cies combines the habits of a ground-feeding finch with those of a tree-feeding warbler.” SUMMARY OF Foops AND FEEDING HABITS The Geospizinae fall into four main groups with reference to food and feeding habits. (1) Members of the genus Geospiza feed mainly upon seeds on the ground, but also regularly eat flowers, young leaves and buds, fruits, and caterpillars when in season, also small insects occasionally. Geospiza scandens and conirostris also feed regularly on Opuntia flowers, as does fuliginosa occa- sionally. (2) Platyspiza feeds mainly on leaves, flowers, and fruits, occasionally taking caterpillars. (3) Camarhynchus and Cactospiza eat mainly insects picked off leaves or excavated from branches. Camarhynchus also eats most of the foods taken by Geospiza. (4) Certhidea feeds primarily on small insects like a warbler. There are marked variations in food as the different types be- come abundant. Thus, on our arrival at the beginning of the rainy season, almost all species were eating young leaves and buds, and the nectar from the tree flowers. Several weeks later, when the No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 43 young leaves and flowers were gone, almost all the Geospizinae were eating sphingid caterpillars, of which there was a great abun- dance. A little later most of the Geospizinae were in the bushes eating ripe fruits. Finally, just before our departure, the species of Geospiza, including G. scandens, left the trees and bushes and fed on the newly ripened grass seed. Our visit did not cover the non-breeding period. The heavy bills of Geospiza magnirostris, fortis, and fuliginosa suggest fruit- or seed-eating birds, but the slight differences between the diet of these three species cannot be the cause of their marked bill differences. The long bill of scandens seems adapted both to Opuntia flowers and to ants, but during part of the year it eats the same foods as the other forms of Geospiza. Geospiza scandens has a split tongue, in this respect resembling other pollen- or nectar- feeding birds, such as the hooded oriole (Icterus cucullatus) (in- formation from E. C. Kinsey). The split tongue is much less well developed in young scandens. It is slightly developed in adult fuliginosa, which also feeds to a small extent on Opuntia flowers. It is absent in fortis and magnirostris which were not seen eating Opuntia flowers. Geospiza c. conirostris which, like scandens, reg- ularly feeds on Opuntia flowers, has a bill less suitable for this purpose. The bill of Camarhynchus, primarily an insect-eater, seems sim- ilar to that of Platyspiza, which feeds chiefly upon leaves. The bill of the former is very unlike that of Cactospiza, which in feeding habits resembles Camarhynchus. The marked differences in bill between Camarhynchus psittacula and C. parvulus do not seem to be correlated with diet. The bill of Cactospiza is clearly adapted for excavating in wood, and the bill of Certhidea for catching small insects. Hence, the bill differences in the Geospizinae can, in some cases, be correlated with differences in feeding habits, but certainly not in many others. DRINK Probably the Geospizinae obtain much of their water from drops which collect on the vegetation. Geospiza magnirostris, fortis, and fuliginosa, but not the closely related scandens, visited wells and springs when present. Gifford (op. cit., pp. 239-241) also noted that scandens did not drink water on the ground. Perhaps it ob- tained sufficient water from Opuntia pads. 44 CALIFORNIA ACADEMY OF SCIENCES [Oc. Papers After a rain, several species were observed bathing, shaking drops over themselves from the wet vegetation. Gifford (op. cit., pp. 223-224) noted Certhidea olivacea cinerascens bathing in sea water. EFFECTS OF RAIN Members of the genus Geospiza seem to be singularly ill-adapted to withstand rain. During heavy downpours, they seek shelter under leaves and afterward are frequently seen in an extremely bedraggled condition, occasionally scarcely able to fly. In the coastal and intermediate zones, where the species of Geospiza breed, rain is normally restricted to about three months of the year, in the form of heavy showers with sunny intervals. Particu- larly since there are no enemies, this lack of adaptation to rain does not seem to matter. In the humid forests and in the open up- lands, the rainfall is heavier and more continuous. Possibly this is one factor influencing the restriction of G. magnirostris, fortis, and fuliginosa to the coastal and intermediate zones for breeding. The bedraggled appearance after rain was not noticed in Cactospiza and other forms which normally breed in the humid zone. NESTS The nest sites of all of the Geospizinae are similar. Indeed, as already noted, one species frequently takes over the nest of another. Geospiza magnirostris, fortis, and fuliginosa nest from three or four feet to 20 feet above the ground; scandens typically higher, rarely below 15 feet, often up to 30 feet; conirostris on Hood usually low, below eight feet; Platyspiza usually from 12 to 20 feet above the ground. Camarhynchus was usually found about 20 feet above the ground or higher, but perhaps nests much lower at times. Cacto- spiza typically nested 30 feet up or higher, but considerably lower in the uplands where only low trees and bushes were available. Certhidea nested in six-foot scrub and upward, occasionally to about 30 feet. The nests of scandens were found exclusively between two ter- minal pads of Opuntia. In the coastal zone, magnirostris, fortis, and fuliginosa also used such a situation more commonly than any other. Occasionally here, and regularly elsewhere, their nests were placed in clumps of fine, closely growing twigs of Acacia, May- tenus, and many other shrubs. No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 45 The nests are cup-shaped below with a large, domed roof and a side entrance. Those of the larger species, especially of Geospiza conirostris, tend to be larger than the nests of the smaller species, but there is considerable overlapping in size. In the intermediate and humid zones, a particularly abundant epiphytic lichen forms the bulk of the nest material, together with some twigs, grass, and cotton. The latter three predominate in nests on the coast. HABITATS The habitats for birds on the central Galapagos Islands can be roughly classified as follows: 1. Coastal zone, arid and characterized by the tree cacti (Opun- tia and Cereus), also Acacia and other shrubs and trees (Figs. 1 and 2). 2. Intermediate zone, at higher elevation and inland from the coast, where some coastal trees disappear and some trees of the third zone appear. 3. Humid zone, characterized particularly by two species of trees, Scalesia and Psidium, with many epiphytic ferns and orchids (Hie. 3). 4. Above the humid zone, the area of shrubs is replaced by open country containing ferns, club mosses, liverworts, mosses and grasses. On Indefatigable, ferns and liverworts predominate, to- gether with shrubs of Miconia robinsoniana (Fig. 4). On Chatham, man has undoubtedly altered conditions in the higher areas, where extensive grasslands occur, which presumably have now been much modified by the introduced cattle and horses. Around the settlement of Progreso, guava was introduced and has now spread in a continuous belt over the higher ground, forming moderately tall trees at lower altitudes, and disappearing as a scrub growth covered with moss on the highest ground. The high points on Chatham and Indefatigable range from 2,000 to 3,000 feet in altitude. The distribution of the different species of Geospizinae as re- gards their habitat may be summarized as follows: Geospiza magnirostris.—Breeds throughout the coastal and in- termediate zones; seen rarely on the edges of the humid zone, and then only feeding, not breeding; not abundant on Indefatigable. Geospiza fortis——Common throughout the coastal and inter- “mediate zones, excepting where the vegetation is less than about 46 CALIFORNIA ACADEMY OF SCIENCES [Oc. PaPErs ten feet in height or where it forms a dense forest; neither situation is common. Occasionally visits cultivated land in the humid zone in searching for food but does not breed there. On Chatham, a few individuals were found feeding in the grasslands on the top of the island; abundant on Chatham, much less so in the southern part of Indefatigable, but, judging from collections, it is evidently abun- dant in the northern part of this island. Geospiza fuliginosa.—Most common in the coastal zone, where it is extremely abundant, also breeds commonly in the intermediate zone. In the latter zone, it is found chiefly where the trees are fairly well spaced, being more particular in this respect than fortis. It regularly feeds in the humid zone and in the open grassy areas above the trees, and some individuals go through preliminary court- ship behavior there. However, breeding seems normally to be re- stricted to the coastal and intermediate zones. Geospiza difficilis —On Tower Island, frequents the coastal zone (no other zone is represented). On Abingdon Island, Gifford and others found it primarily in the humid zone, hence it does not overlap with fuliginosa on this island. Geospiza d. debilirostris of James and Indefatigable is also said to be found primarily in the humid zone, although some specimens have been collected on the coast. Swarth (MS) recorded this species, in 1932, as commonly frequenting the low bushes and forest floor in the humid zone on Indefatigable. As noted in Section I, we failed to find these birds in 1939. Geospiza scandens.—Restricted to localities where Opuntia oc- curs and in this environment usually the most abundant of all the Geospizinae. On Chatham, Opuntia is local in distribution and this accounts for the scarcity of specimens of scandens in collections from this island. The bird is common, however, on parts of the south coast. Like magnirostris and fortis, scandens occasionally feeds on the edges of the humid zone. On Charles Island, Gifford (op. cit., p. 239) reports that it regularly visits the orange groves when the fruits are ripe, but does not do so at other times. Platyspiza crassirostris—Found wherever there are tall trees, but scarce and probably not breeding in the coastal zone; most common (moderately abundant) in the intermediate zone, particu- larly in the taller and denser forest, also breeding in the humid zone, and extending up to the tree limit, although very scarce at the higher levels. It regularly visits the fruit trees of the planta- No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 47 tions, especially papaya. This bird is rare in the guava belt on Chatham. Camarhynchus psittacula——Observed by us only on Indefati- gable, where it was not at all common, in contrast to the other species of Geospizinae, excepting Geospiza d. debilirostris. Most psittacula were seen in the humid zone, some up to the tree limit, also breeding in the intermediate zone, and occasional birds were singing but perhaps not breeding in the coastal zone. Camarhynchus parvulus——Present but not abundant in the coastal zone, where at least some bred; abundant in the intermedi- ate and humid zones up to the tree limit. A certain number occurred in the guava belt on the high ground on Chatham. Camarhynchus pauper.—According to Snodgrass and Heller. (1904, p. 288), also Gifford (op. cit., p. 249), this bird breeds primarily above 1,000 feet on Charles; evidently abundant. Cactospiza pallida.—Predominantly a bird of the humid zone, moderately abundant up to the tree limit, common in the inter- mediate zone, also present, although much less numerous, in the coastal zone; probably does not breed there. In the coastal zone, it is chiefly found feeding on Opuntia, which is the basis of the name “Cactospiza.” This name, however, is inappropriate, since the bird mainly frequents the humid zone, where Opuntia is absent. On Chatham, it is one of the few Geospizinae regularly found throughout the guava belt of the high ground. It also occurs throughout the zone of open ground above the tree limit on Inde- fatigable, nesting in the clumps of Miconia robinsoniana. Swarth (MS) reports pallida in mangrove swamps at Conway Bay, Inde- fatigable. Cactospiza heliobates——All observers report that this species is confined to the mangrove swamps around Albemarle and Nar- borough. Certhidea olivacea-—Found particularly where there are low shrubs. It has the widest habitat range of all the Geospizinae, breeding commonly in the coastal, intermediate, and humid zones, the guava belt of Chatham, and the open bracken zone on the top of Indefatigable. Excepting where shrubs are scarce, as on some parts of the coasts, it is extremely abundant, often occurring in greater numbers than the sum total of all the other Geospizinae. Pinaroloxias inornata—Reported frequenting the forest trees on Cocos Island (Gifford, op. cit., pp. 242-243). 48 CALIFORNIA ACADEMY OF SCIENCES [Oc. Parers ta 4 oe “Ae f é. a ef " ar Figo An d Eonatal zone, Indefatigable Island. Photograph taken by Richard Leacock, December, 1938. 4 % ie ih od Fig. an ‘ nS hea : Ved RR SR 2.—A nest of Geospiza magnirostris in an Opuntia in the arid coastal region, Indefatigable Island. No. 21] LACK: VARIATION IN GALAPAGOS FINCHES Fig. 3—Humid forest, Indefatigable Island. Photograph taken by T. W. J. Taylor. 49 Fig. 4.—Top of Indefatigable Island showing fern and grassland in the foreground with low forest beyond. Photograph taken by T. W. J. Taylor. 50 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS DiscussION OF HABITATS Considerable overlapping occurs between the habitats of most of the Geospizinae. The species of the genus Geospiza, omitting some forms of difficilis, breed primarily in the coastal and inter- mediate zones, while Platyspiza, Camarhynchus, and Cactospiza breed mostly in the intermediate and humid zones, and Certhidea is common to all the zones. However, there is a wide area where all breed side by side. Geospiza scandens, unlike the other species of Geospiza, breeds only in Opuntia, but this does not differentiate it from magnirostris, fortis, and fuliginosa, since all of these also breed commonly in Opuntia. The only two known instances on these islands where habitat differences separate closely related species are: (1) G. d. difficilis (humid zone) and G. fuliginosa (coastal zone) on Abingdon; and (2) Cactospiza pallida (humid zone) and C. heliobates (coastal mangrove belt) on Albemarle and Narborough. Neither of these could be studied by us. Gauss (1939, p. 255) has expressed a common belief when he writes, “In the light of all this evidence one may claim that if two or more nearly related species live in the field in a stable associa- tion, these species certainly possess different ecological niches.” However, an exception is provided by Geospiza magnirostris and fortis. Careful field study failed to reveal any differences whatever in habitat, food, feeding habits, nest site, and breeding season be- tween these species. Furthermore they have similar plumages. They seem to differ solely in general size and in relative size of bills. Yet, normally at least, they do not interbreed. In addition, G. fortis and fuliginosa, and also Camarhynchus psittacula and parvulus have similar ecological requirements, although there are slight differences. Although others probably exist, G. magnirostris and fortis are exceptional examples of two closely related avian species having apparently identical habitats, food, feeding, and other ecological requirements. Dr. E. Mayr, however, informs me that two closely related species of Ptilinopus (fruit pigeon), representing two sep- arate colonizations, occur on the Marquesas Islands. Although not studied in the field, all known Piéilinopus are so similar in ecology that the two on the Marquesas are probably similar. In California, I observed Agelaius tricolor breeding in the identical habitat with A. phoeniceus, and having extremely similar feeding habits. How- No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 51 ever, there are two differences: first, tricolor is colonial and phoe- niceus territorial; second, while tricolor always nests in sites suit- able for phoeniceus, the reverse does not hold, phoeniceus being more generalized. In many other instances, two closely related species show considerable overlapping of habitat without inter- breeding, for example Parus major and P. caeruleus in Britain. See Lack (1940c) for a discussion of habitat distribution and speciation. To summarize, habitat differences would seem to have played a part in the speciation of Geospizinae in only a few cases. It is striking that as yet no form of Geospizinae has become adapted to the vacant ecological niches afforded by the extensive open grass and fern country above the tree limit on Chatham, Indefatigable, and Albemarle islands. INFLUENCE OF HABITAT ON GEOGRAPHICAL DISTRIBUTION The smaller and more barren of the Galapagos Islands do not possess intermediate or humid forests, which may account for the absence of Cactospiza pallida and possibly some other species on some of them. Again, forms of Geospiza difficilis are found in the humid zone on Abingdon and James islands but not on Bindloe, which lies between these two islands, but has no humid zone. Four- teen specimens of G. fortis were collected on Hood in 1905, which would seem to indicate an attempt at colonization. Apparently, however, the species has never become established, so here also a habitat factor may be involved. However, most species show sim- ilar ecological requirements. Thus when one of the species is absent from a particular island it is more likely due to its failure to reach the island in sufficient numbers to become established than to un- suitable ecological conditions. BREEDING SEASON The main breeding season of all species of Geospizinae is in the rainy season, beginning about mid-December and ending in early April. The rains clearly have considerable influence on breeding. Thus, in the intermediate zone of Chatham, breeding was well under way in late January, 1939, but had not started in the Chatham coastal zone or on the near-by island of Hood on January 30. The intermediate zone had received sufficient rain to make the vege- tation green about a month before the coastal zone. Swarth (MS) 52 CALIFORNIA ACADEMY OF SCIENCES [Oc. PapPErs found breeding continuing into June in 1932, a year in which the rains persisted later than usual. This late breeding is not recorded in years of more normal rainfall. In 1939, in the intermediate zone on Chatham Island, nest- building and courtship were well under way as soon as the rains began in mid-December. Rainfall was scanty during the next five weeks and the birds, excepting Cactospiza, did not advance further in their breeding activities. Possibly this delay in breeding was correlated with the delay in the rains, but considerably more data are needed in order to make certain of this. From the immature specimens in collections, it is clear that in certain years at least, some breeding takes place in August and September, and possibly in nearly all months of the year. No species is isolated from another by differences in breeding season. The rainy season occurs in the intermediate forest before it does on the coast. Correlated with this is the earlier appearance of green leaves. Consequently breeding commences about a month earlier in the intermediate zone than along the coast. Two species, Cactospiza pallida, found chiefly in the humid zone, and Geospiza scandens, found on the coast, start to breed about a month earlier than most other species in their respective habitats, but both con- tinue breeding for several more weeks along with the other species. Mott The adults normally molt immediately after the breeding season, that is, between April and June. The immature molts are uncertain and are discussed elsewhere. PREDATORS The natural predators of the Geospizinae are extremely few. The short-eared owl (Asio galapagoensis) has been recorded as eating Geospiza fuliginosa (Gifford, op. cit., p. 237) and G. fortis (Beebe, op. cit., p. 331). However, this owl is scarce, and can have only a slight effect upon the geospizid populations. Possibly the barn owl (T7'yto punctatissima) takes an occasional ground finch, but this species also is extremely scarce. The Galapagos hawk (Buteo galapagoensis) is probably harmless. The Geospizinae, therefore, have virtually no natural predatory enemies. Presum- ably correlated with this, they show extremely little fear of man or other introduced mammals. This tameness has been rather ex- No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 53 aggerated by some writers. It is not possible to pick the birds off the bushes, and it is extremely difficult, and usually impossible, to catch them in a hand net. Normally they allow one to approach to within six to ten feet provided sudden movements are avoided. Like other birds, the Geospizinae are somewhat wilder near their nests. They are not as tame as most of the other Galapagos land birds. The hawk, for instance, will frequently allow itself to be touched without moving away. The mockingbird (Nesomimus) will sometimes peck at one’s boots, and three individuals of the flycatcher (Myiarchus magnirostris) settled on our heads and bodies in an attempt to take our hair for nest material. The Geo- spizinae did not exhibit such lack of fear. The introduced black rats (Rattus rattus rattus), which are ex- tremely abundant on Chatham Island, probably take a heavy toll of birds’ eggs and young. The cats, which have recently been in- troduced on Indefatigable, may in time also become a serious men- ace to the birds. In one instance, without creating much disturb- ance, a cat was able to capture one of a group of Platyspiza crassi- rostris which were feeding on the ground on some fruit. The Geo- spizinae would thus seem to have lost the normal reactions of a small bird to predators, owing to the absence of the latter. How- ever, the Geospizinae occasionally attack a Galapagos hawk in the same way, although not so effectively, as small passerine birds in other countries attack hawks and owls. PoPuULATION DENSITY The terrain was so inaccessible on most of the islands that we found it impossible to take a census of the breeding populations. The writer has taken censuses of breeding birds in many different types of habitats in England. On the basis of this experience, he would estimate the breeding population in the intermediate forest on Chatham as about the same as that in rich British woodland, that is about 20 adult birds per acre. The different species were represented in approximately the following percentages: Geospiza fortis, 14; G. fuliginosa, 3; Platyspiza crassirostris, 11; Cama- rhynchus parvulus, 8; Cactospiza pallida, 4; Certhidea olivacea, 30; Nesomimus melanotis, 12; Myiarchus magnirostris, 8; Pyro- cephalus dubius, 6; and Dendroica petechia aureola, 2. This was probably the highest density of Geospizinae encoun- tered on the islands. In other parts of the coastal and intermediate 54 CALIFORNIA ACADEMY OF SCIENCES [Oc. PaPers forest, the breeding population typically varied from 10 to 20 adults per acre, and was much lower in areas of sparse vegetation correlated with more recent lava flows. The species composition varied in different places. Thus, Geospiza fuliginosa and G. scan- dens were usually the most common birds in the coastal region. The humid forest on Indefatigable probably supported about 10 adults per acre. Above the tree limit in the open country, the breeding density was much lower, probably similar to that of rich English heathland, that is between one and three birds per acre. The belt of taller guava on Chatham had an extremely low density of birds and few, if any, bred there. However, more were found where the guava became lower on the grassy uplands. On Hood, the density of small land birds was fairly high (about 10 adults per acre) around the coast, but was considerably sparser inland. On Tower, the density was much lower than encountered elsewhere, perhaps between three and seven adults per acre. Culpepper and Wenman are so small that, even if Geospizinae are abundant there, it is clear that the several endemic forms pe- culiar to these islands are represented by at most only a few thou- sand individuals alive at any one time, and the populations of the endemic forms on Tower cannot be much larger. This should be borne in mind in considering their evolution. Again, Daphne is only the top of a crater and Crossman a group of islets, so that the peculiar G. fortis X fuliginosa (?) hybrid form found there can be represented by at most only a few hundred individuals alive at any one time. SECTION IV. COLORATION ‘BLACK MALE PLUMAGE IN GEOSPIZA, PLATYSPIZA, AND CAMARHYNCHUS DESCRIPTION In all the species of Geospiza, the fully adult male plumage is black with the exception of the under-tail coverts, which are mar- gined with white or varying shades of buff. The juvenal and post- juvenal plumages resemble that of the female, being streaked gray- brown. It is possible that certain males attain a similarly streaked type of plumage following the postjuvenal plumage. From a study of the skins, it is clear that the black feathering first appears on the front of the head, then extends gradually down the body, the ab- domen being the last to become black. Males breed at all stages between the streaked and the fully black condition. In Platyspiza, the adult male is black to about halfway down the breast and back; the abdomen and lower back are colored as in the immature and female, that is, the “full” male plumage in Platyspiza corresponds to a transition stage of plumage in imma- ture individuals of Geospiza. The same is true of Camarhynchus, in which the black tends to be even more restricted than in Platy- spiza, although it often extends to the upper breast. Almost com- pletely black plumage is found very rarely in both Platyspiza (one collected specimen and one seen on Indefatigable) and Camarhyn- chus psittacula (one specimen collected). On the other hand, many individuals of both genera breed in streaked plumage or in an intermediate condition, the black feathering being restricted to the head, the front part of the head, or even to a small area around the bill. As in Geospiza, the black first appears anteriorly and grad- ually develops posteriorly. In Cactospiza pallida, the male is normally colored like the fe- male and immature, showing no black. However, out of hundreds of individuals seen on Indefatigable Island, one otherwise typical male had a black head. This shows that Cactospiza simply carries further the tendency to loss of distinctive male plumage shown by Camarhynchus, to which it is closely allied. : [55 ] 56 CALIFORNIA ACADEMY OF SCIENCES [Oc. PaPers SEASONAL AND ECOLOGICAL VARIATION On the islands of Chatham and Indefatigable, in 1939, the ma- jority, but not all, of the males of Geospiza in streaked and partly black plumage bred about a month later than did the males in fully black plumage. The coastal regions become suitable for breeding about a month after the intermediate forest, hence most males in streaked and partly black plumage bred near the coast because the territories in the intermediate forest were already held by fully black males. However, some males in streaked plumage bred early in the intermediate forest, and some fully black males bred late on the coast. | INTER-ISLAND VARIATION IN MALE PLUMAGE Text Table 3 is compiled from examination of males in various plumages in all of the collections that I have studied, obvious fledglings being omitted from the last total. This table is not an accurate guide to the proportions of each type breeding on the different islands, first, because collections made shortly after the breeding season include many males in streaked plumage which are not yet sexually mature, and second, because the proportions of each type which is breeding depend partly on the season. Thus, in 1939, a collection on Indefatigable in January would have con- tained mostly black males, and one in March mainly males in Table 3 Number of Males in Different Types of Plumage on the Various Islands (American and British Collections) = é ~~ w ay a ny a n © Tower 20 2 7 1 44 6 3 31 12 Abingdon 18 10 23 12 16 20 13 12 23 11 6 2 4 3 8 Bindloe 14 621 149 13 u 7 6 5 4 4 James 16 12 18 27 lo La ener t 33 7 4 16 9 20 win RP @ G nw Bw Jervis 2a) ft) <6 $215 Lt tart h 4 Daphne 23) 7) 1 1 1 S, Seymour 212 4 2512 8 21 20 Indefatigable 44 5 31 38 38 33 28 20 236 4 42 43 16 Ba Duncan 1 10 10 10 46 22 16 1 1 1 Narborough 1 15 N62 2? 5 1 a N, Albemarle 2 1 30 15 20 60 14 14 6 2 wy Ny NH w S, Albemarle 12 4 27.16 27 | 37 12 15 1 1 9 34S Barrington at 4 11] 2710 4 39). 122 2 2 Chatham 56 25 29 | 64 35 28 1 9 2 1 4 87 Hood + Gardner 8 29 7 4 102 57 Charles 1 fos 51 26 | 40 18 28 5945 6 Wy 1 3 See Ay) 6126 Cocos Ieland 84 18 23 *B-black plumage; P-partly black plumage; S-streaked plumage No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 57 streaked or partly black plumage, as was discussed under the pre: ceding heading. Nevertheless, Text Table 3 does give some indi- cation of the variation which exists. Text Table 3 is based upon all available collections and hence is much less influenced by a single, exceptional collection. Even making considerable allowance for chance variations, it is clear that Swarth (1931, 1934) is right in his general conclusion that there is inter-island variation in the number of males breeding in black, partly black, or streaked plumage. For Geospiza, the evi- dence is inconclusive, excepting that all collections and my field observations indicate an unusually high proportion of full-plum- aged males for G. magnirostris and G. d. difficilis on Tower. Es- pecially since Tower is a small island, this conclusion is reason- ably certain. In Platyspiza and Camarhynchus, there are such marked differences in the proportions of males in streaked and in partly black plumage collected on the different islands that this must reflect an actual difference in the populations concerned. To cite the most striking case, out of 87 male Camarhynchus par- vulus collected on Charles, 61 (or 70 per cent) showed black feathering, but out of 91 males collected on Chatham only 4 (or 4 per cent) showed black feathering. Field observations on Chatham confirmed this low proportion, although it is somewhat higher than the figures based upon specimens indicate. This may be due to the fact that a majority of the partly black males breed high up in the intermediate forest, while most collecting has been done near the coast. While inter-island variation undoubtedly exists in at least sev- eral species, Swarth appears to be wrong in his additional statement that on certain islands the proportion of males showing black is low in all species. As is discussed later, his Abingdon figures are completely misleading. On Charles, there is a low percentage of males showing black in both Platyspiza crassirostris and Cama- rhynchus pauper, but in C. parvulus it is unusually high. Chatham shows a low proportion of black males in both P. crassirostris and C. parvulus; this is confirmed by field observations. The figures in Text Table 3 for G. scandens on Chatham are not significant, since this form has not been collected on its breeding grounds (see Sec- tion I) ; however, we observed many black males there. Swarth (1931, 1934) concluded that there was an exceptionally low proportion of “‘fully-plumaged” males for all species on Ab- 58 CALIFORNIA ACADEMY OF SCIENCES [Oc. Papers ingdon Island. This was based on the California Academy of Sciences collection of 1905. As shown in Text Table 4, collections made in other years do not bear this out. Doubtless this difference is partly due to the seasonal variation already considered, but it is Table 4 Plumage of Male Geospizinae from Abingdon Island C.A.S. Collection Collections elsewhers Partly Partly Species Black black Streaked Others Black black Streaked Others Geospizs magnirostris 1 14 17 10 9 Geospiza fortis i 14. nu 16 6 Geospiza fuliginosa 1 2 8 12 10 14 Geospizea difficilis 1 n 6 2 Geospiza scandens 5 4 8 Platyspiza crassirostris 5 8 3 Camarhynehyg habeli t 9 2 so marked that one suspects that the percentage of males breeding in streaked plumage may be different in different years on the same island. AputLt Mo.Lts Adult Geospizinae, like most passerine birds, normally undergo a complete molt once a year, just after the breeding season, that is in May or June. This is fully substantiated both by the collec- tions and by our field observations. There is one possible complication. In the Rothschild Collec- tion are numerous specimens of Geospiza collected on south Albe- marle in December, 1900. In normal years this is the beginning of the rainy season. This collection includes both some fledglings and many males in new adult plumage. Perhaps there had been an abnormal breeding period (possibly correlated with an unusual rainy period), and the males in fresh plumage had molted just after breeding. Alternatively, they might be first-year birds mollt- ing into adult plumage for the first time. THe CHANGE FROM STREAKED TO BLACK PLUMAGE The situation is complicated and evidently different individuals of the same species on the same island may molt differently.* It * Observations made by Robert T. Orr at the California Academy of Sciences have shown much individual variation to occur with regard to the age at which black plum- age is attained in cage-reared geospizids. No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 939 is hoped that the captive birds brought to the California Academy of Sciences will aid in clarifying the sequence of the molts from streaked to full male plumage. Until then any conclusion is neces- sarily speculative. On Tower, the proportion of male G. magnirostris and G. dif- ficilis in fully black plumage is so high that one can be certain that a large proportion of the males must molt into fully black plumage before they are a year old. The same applies to Geospiza, Platy- spiza, and Camarhynchus on other islands where the proportion of fully plumaged males is high. On the other hand, on Chatham the proportion of partly black to streaked males in Camarhynchus parvulus salvini is so low that one can be certain that a large pro- portion of the males never acquire the black plumage. For various species on other islands conditions are doubtless intermediate, some individual males acquiring “full” plumage before they are a year old and others not. One can be certain that some individual male Platyspiza and Camarhynchus never acquire “full” plumage. However, it is not known whether all the individual male Geospizinae breeding in streaked plumage remain permanently in this type of plumage, or whether some of them molt into “full” plumage in the second year. The same problem is presented by the males breeding in “partial” black plumage. At least, in many cases, this plumage appears to have been acquired by a complete, not a partial, molt. In Platy- spiza and Camarhynchus the proportion of males in “partial” (black-headed) plumage is so high that one can be certain that some of them never acquire the “full,” black-breasted plumage. Thus, the “partial” plumage is a final stage so far as the individual in question is concerned. It is uncertain whether this applies to all the male Geospiza, Platyspiza, and Camarhynchus breeding in “partial” plumage, or whether some of these later molt into “full” plumage. In the exceptionally prolonged rainy season of 1932, Swarth (MS) collected birds in streaked plumage, with black bills, nor- mally an indication of breeding condition, well-developed gonads, but with incompletely ossified skulls, normally a sign of juvenility. Some specimens in the American Museum of Natural History, col- lected in 1935, show the same condition. This suggests that, as in certain tropical Ploceidae (Steinbacher, 1936), some individuals may breed when two or three months old. If this is so, such indi- 60 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS viduals would increase the numbers breeding in streaked plumage, but they certainly do not account for all cases, since most begin to breed before any birds of the same breeding season would be old enough (personal field observation). The discussion by Mayr (1933c) on plumage variation in Neo- lalage banksiana may be referred to; however, the situation in the Geospizinae seems to be more complex than the instances which Mayr discusses. HoRMONAL AND GENETIC CONTROL Recent work, notably by Witschi (1935) and his colleagues, has demonstrated that male secondary séxual plumage is controlled by various hormones (different in various species of birds), and in a few instances seems mainly under genetic control. With reference to this work, the Geospizinae would provide an extremely interest- ing problem. Presumably hormonal factors are involved in the change from immature to black plumage. Since in Platyspiza and Camarhynchus several stages exist in an apparently permanent form, a number of genetic factors would seem to be involved. Again the “partial” plumage, perhaps a transitional stage in Geo- spiza (and therefore under hormonal control), is certainly part of the inherited constitution of the species in Platyspiza and Cama- rhynchus. Two more points may be mentioned. As shown in Section V, males in streaked or “‘partial” plumage have, on the average, a smaller length of. wing than males in “full” plumage. Also, as has been already mentioned, males in streaked or “partial” plumage tend to breed later than do males in “full” plumage. In normal passerine birds both of these facts would indicate that the males in streaked and “partial”? plumage were one-year-old birds. This cannot be considered certain for the Geospizinae, hence the rela- tion of wing length and breeding season to black male plumage requires further investigation. EVOLUTIONARY SIGNIFICANCE All data appear to confirm Swarth’s (1931, 1934) contention that the black male plumage of the Geospizinae is in process of being lost, not acquired. Indeed it is inconceivable that so many genera, species, and island forms should be in process of acquiring black male plumage independently. It may also be noted that No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 61 among the Geospizinae the black male plumage is best developed in the generalized genus Geospiza, and is reduced or absent in the specialized genera, which latter have in other respects departed farther from the ancestral fringillid type. From the data in Text Table 3, it is clear that the disappearance of the black plumage is proceeding at a different rate in diflerent island populations of the same species. The field study provided no reason for the assump- tion that the black plumage had survival value to the species, and evidently the gradual disappearance of a now functionless char- acter is taking place. The Galapagos forms of Pyrocephalus, the vermilion flycatcher, also breed not infrequently in immature plumage, especially on Chatham; so perhaps does the martin, Progne modesta. Indeed this is a not infrequent tendency among insular land birds. Lowe (1923) notes it in Nesospiza on Tristan da Cunha; Murphy and Chapin (1929, pp. 20-22) find it in the form of Pyrrhula pyrrhula (normally a sexually dimorphic species) in the Azores, and there are numerous cases in Polynesia; see Murphy and Mathews (1928, p- 7) for Pomarea, and Mayr (1931-34) for Coracina, Pachy- cephala, Ptilinopus, Myiagra, Clytorhynchus, and Petroica. In some of the latter, namely, Coracina lineata, Pachycephala pec- toralis (P. p. feminina and P. p. xanthoprocta compared with the other races), Clytorhynchus nigrogularis, and Petroica multicolor, there is, as on the Galapagos, inter-island variation in the loss of male plumage, some races showing it and others not. Again Po- marea iphis and Myiagra vanikorensis resemble Geospiza in that males apparently sometimes breed in plumage intermediate be- tween the juvenal and fully adult condition. Loss of secondary sexual plumage is not confined to the birds of oceanic islands, as, for example, the instance of Pyrrhula in parts of Asia (Murphy and Chapin, Joc. cit.). The tendency, however, is certainly far more common on oceanic islands than on the mainland. The above brings up the question of what factor makes sexual selection of less importance on small islands. The function most commonly attributed to secondary sexual plumage characters is that they emphasize postures and displays, particularly sexual and threat displays (cf. Marshall, 1936, pp. 445-446). There seems to be no reason why this should be less important on small islands. For example, the Geospizinae display vigorously. Also on the Ga- lapagos Islands, it appears curious to see two male Pyrocephalus 62 CALIFORNIA ACADEMY OF SCIENCES [Oc. PaPERs indulging in identical threat display at each other, with the one possessing brilliant vermilion feathers which emphasize the pos- tures, the other being quite dull-colored, but apparently none the less effective. There is, however, another function that is sometimes attributed to male secondary sexual plumage and song in birds, particularly in closely related species in which the females are similar, namely that it enables females in search of mates to recognize readily males of their own species. Hybridization is, of course, at a selective dis- advantage since it is associated with a decrease in fertility. It is quite possible that the black plumage originally served this function in the ancestral Geospizinae, in which case the disappearance of the black plumage at the present time is readily understandable since, as discussed elsewhere, specific recognition now is effected primarily by bill differences, not by plumage. A similar explana- tion may well hold for the other known instances among the land birds of oceanic islands, since such species have normally been separated from all other species with which the females could possibly form pairs. Hence, the survival value associated with species recognition by plumage differences disappears. As a result of the recent advances in our knowledge of the functions of threat and courtship display, there has been perhaps a tendency to overlook this other function of secondary sexual plumage, namely in species recognition. The birds in which wild hybrids probably occur most frequently are the birds of paradise (Paradiseidae) (Stresemann, 1930), the hummingbirds (Tro- chilidae) (Stresemann, 1930, p. 14; Berlioz, 1927), the ducks (Anatidae) (Rothschild and Kinnear, 1929; Sibley, 1938), and certain gallinaceous birds, including the Phasianidae, Tetrao and Lyrurus (Poll, 1911). It seems rather more than coincidence that a list of the birds in which secondary sexual plumage is well de- veloped would include all of these groups. Hybridization in the wild state is probably more frequent in these birds, owing to the less definite bond between the pair, than in other birds in which both of the parents care for the young. The females of many of these species are similar, indicating that close relationships exist within the families; the striking differences between the males in part reflect the selective advantage of species recognition by the females. That is, owing to an unusually great tendency to hybridization, these birds exhibit just the opposite of the process found in the No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 63 land birds of oceanic islands. The peculiar display habits and the decreased selective value of protective coloration in the male, since he does not assist with the brood, also contribute to the great devel- opment of sexual ornamentation in these birds. These factors, however, do not account for the marked differences between the species. Sometimes, as for instance in Pomarea (Murphy, 1938, pp. 534-537), instead of the males of land birds of oceanic islands losing the secondary sexual plumage, the females acquire the male type of plumage. This tendency, while not found in the Galapagos, is, however, not uncommon in continental land birds (cf. Winter- bottom, 1929). Further data are needed to show whether or not this phenomenon is especially common on oceanic islands, and if it is in any way connected with the loss of male plumage in other species. Biack MALE PLUMAGE IN Pinaroloxias In Pinaroloxias inornata, as in the species of Geospiza, the “full” male plumage is black, and juvenal males are colored like the females. Unlike Geospiza, the black feathering comes in ir- regularly all over the body, and does not start at the head and pro- ceed posteriorly. It is not known if breeding occurs in immature plumage. Rurous UnpeEr-TAIL COVERTS The males of most species of Geospiza in fully adult plumage are entirely black except for white under-tail coverts, but those of G. difficilis often show rufous or occasionally buffy tips to these white feathers. The percentage showing this, and the degree of development, varies on different islands. Only two out of 17 black males on Tower Island and one out of 12 on Abingdon show any trace, and then but little. Of G. d. debilirostris on James, 16 out of 27 black males, and on Indefatigable, 13 out of 21 show some rufous tipping; when present, it is better developed in the birds of Indefatigable. The rufous under-tail coverts are best developed in G. d. septentrionalis from Culpepper and Wenman, where most of the males show it. Males of these forms in juvenal plumage also occasionally show the rufous under-tail coverts. In Pinaroloxias inornata, most males have white under-tail cov- erts, with buffy tips sometimes occurring. This links Pinaroloxias 64 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS with G. difficilis, since males of other forms of Geospiza normally show no buff. However, black males of G. d. septentrionalis, the form nearest to Pinaroloxias with reference to female plumage, which is discussed later, show far more rufous in the under-tail coverts than does Pinaroloxias. CHESTNUT THROAT PATCH Male Certhidea differ from other Geospizinae in possessing a chestnut patch on throat and breast. Traces of a similar patch of orange or chestnut, mixed with some black streaks, occurred in a male Geospiza fuliginosa in streaked plumage from Wenman Is- land, in a male G. d. debilirostris in streaked plumage from Inde- fatigable, and in a male Camarhynchus parvulus from Albemarle, while a male C. pauper from Charles showed a buffy patch. The chestnut throat patch also occurs in the one specimen of ““Cactospiza giffordi”’ and in two “Camarhynchus conjunctus.” Such occurrences link Certhidea with the other Geospizinae. Stresemann (1936) supposes “Cactospiza giffordv’ and “Cama- rhynchus conjunctus” to be hybrids of Certhidea and Cactospiza and Certhidea and Camarhynchus, respectively, but hybridization with Certhidea cannot possibly account for the other instances. Perhaps the genetic factor, or factors, upon which the male throat patch depends is present, at least sometimes, in other Geospizinae. Normally its presence would be completely obliterated by the de- velopment of black plumage. FEMALE PLUMAGE (omitting Certhidea) Unless otherwise stated, the descriptions in this section refer both to adult females and to immatures of both sexes, since these plumages seem indistinguishable. The plumage in Geospiza magnirostris, fortis, and fuliginosa (referred to hereinafter as the MFF series) is almost the same (Swarth, 1931, pp. 144, 154, 169), namely grayish-brown above and streaked below; these three species are used as the standard for comparison. While there is considerable individual variation, significant differences between populations on different islands seem to be absent. Geospiza fortis of Abingdon Island seems more streaked below than is normal for members of this species on other islands, but a few Chatham specimens are fully as streaked. One fortis and a few fuliginosa, several being from Chatham, are un- No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 65 streaked below and relatively pale above. Two fuliginosa from Bindloe have a darker ground color than any other specimens. Some from Abingdon and Bindloe are more streaked below than usual, but others are more normal and some are nearly unstreaked. Of four fuliginosa from Wenman, two have varied in the direction of G. d. septentrionalis, being rather buffy below with more olive tips to the back feathers and with a broader brown wing bar. The wing bar in the group varies from very pale buff to brown, with considerable individual variation. Possibly a greater percentage in magnirostris show the brown wing bar than in fortis and fuligi- nosa. Geospiza d. difficilis, debilirostris, and septentrionalis are gen- erally darker above and more streaked below than the MFF group (Swarth, supra cit., pp. 178-186). Geospiza difficilis tends to be darker above than the MFF group, but there is much overlapping. As in other characters, difficilis (sensu stricto), especially on Tower, is nearer to the MFF group than is debilirostris. Geospiza d. septentrionalis is as dark above as debilirostris, but with promi- nent olivaceous tips to the feathers, especially in the rump region. The under parts of Geospiza d. difficilis from Tower Island are intermediate, half being nearly as streaked as the darker fuliginosa, others resembling typical scandens, while two are darker than any fuliginosa. Geospiza d. difficilis on Abingdon is similar, but has a darker ground color. In debilirostris, these trends are carried further, but there is wide individual variation. Many are like the darkest, most streaked fortis, some definitely darker, and two are as dark as typical G. conirostris on Hood. Geospiza d. septentrionalis is dark and as streaked below as average debilirostris, but is often distinguishable by a buffy ground color. This species also tends to be dark-headed. In difficilis (sensu stricto), the wing bar is usually brown, more rarely buffy as in the MFF series, occasionally dull rufous, and in one specimen bright rufous. Of 32 specimens of debilirostris, 30 show a fairly bright rufous bar. All specimens of septentrionalis show some rufous on the wing, often more de- veloped than in any debilirostris. Hence the degree of rufous in the female’s wing bar in this group runs parallel to the rufous under-tail coverts of the male (q.v.). The upper parts of Geospiza scandens (sensu stricto) tend to be slightly darker and grayer, less brown than typical in the MFF group and grayer than in the difficilis group, lacking the olivaceous 66 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS tips, but scarcely distinguishable from difficilis. The under parts are distinctly darker and more streaked than the MFF group; it is streaked especially on the throat, where the markings sometimes run together to form a solid dark area. It is nearly as streaked as septentrionalis but does not have the bufly tinge of the latter. The wing bar is similar to the MFF group, but possibly fewer are pale. Birds from James, Seymour, Indefatigable, Charles, and Chatham islands seem alike. Two specimens from Albemarle are especially dark below, as dark as G. conirostris on Hood, but others are paler. Birds from Abingdon and Bindloe are darker both above and be- low, those from Abingdon especially below, those from Bindloe especially above. The darkest female plumage in the Geospizinae is exhibited by members of the species Geospiza conirostris conirostris. The upper parts are as dark as in scandens from Bindloe, but perhaps some- what browner. The under parts are usually darker than in scan- dens. The lightest are about as dark as the darkest scandens from Abingdon. The wing bar is often indistinct, otherwise as in scan- dens. The plumage of G. c. propinqua is indistinguishable from typical scandens from the central and southern islands, but is dis- tinguishable from the scandens on Abingdon and Bindloe, the two islands nearest to Tower. The broad, olivaceous-brown tips to the feathers of the upper parts of G. c. darwini distinguish it from the MFF group and propinqua. The under parts are as streaked as in typical scandens or propinqua and with a buffy tinge rarely seen in these forms. The wing bar in half the specimens is brown with a rufous tinge; in the rest it is brown or buff-brown. This form differs from the MFF group in the same respects as does septentrionalis, but less markedly than the latter. That the three forms of coni- rostris should have such different plumages suggests that they may have evolved from scandens stock independently. Platyspiza crassirostris is more olivaceous or brownish, less gray and less streaked above, more buffy below than in the MFF group. The wing bar is as variable as in the MFF group, often brown with a touch of rufous, usually pale and often inconspicuous. Slight differences apparent in series from different islands are per- haps due merely to differences in degree of wear of the feathers, but birds from Abingdon, also Chatham, perhaps have more pro- nounced olivaceous tips to the feathers of the upper parts, and as a result, Abingdon birds seem almost barred above. No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 67 The forms of the Camarhynchus psittacula superspecies are typi- cally paler and less streaked and nearly always more olivaceous than in the MFF group. Camarhynchus psittacula is gray-brown above and whitish below (Swarth, 1931, p. 216). The wing bar is normally pale buff or neutral. Camarhynchus habeli is usually grayer above, less olivaceous, and darker than psittacula; the under parts and wing bar are similar to psittacula. Camarhynchus pauper is darker with darker streaks, browner, less gray above than psit- tacula. The under parts are typically streaked, but in a few im- maculate. Two females show a buffy patch on throat and breast. Wide variation occurs in the wing bar, but it is more often brown than in psittacula. Camarhynchus affinis from Albemarle is inter- mediate between psittacula and pauper, and probably is nearer psittacula, from which it differs in all the tendencies shown by pauper; it intergrades with both. Camarhynchus “incertus” or Camarhynchus affinis from James, Indefatigable, and other islands (for taxonomic position see Section I) resembles affinis from Albemarle in the plumage of the upper parts, being browner not grayer; the under parts are intermediate, but none are as immacu- late as the extreme in psittacula. Camarhynchus parvulus parvulus resembles typical psittacula above. Some are streaked below and others are not. Birds from Albemarle are perhaps more streaked below than those from other islands. The wing bar is like psittacula. The upper parts of C. p. salvini are extremely similar to typical pauper, but the brighter birds show a yellow or greenish tinge. The under parts are yel- lower than parvulus and more individuals are streaked. “Camarhynchus conjunctus”’ is represented by two males per- haps in juvenal plumage. This is probably not a valid species. In bill and plumage, it seems intermediate between parvulus and Certhidea olivacea ridgwayi. “Camarhynchus aureus,” represented by one male possibly in juvenal plumage, is probably not a valid species. It seems inter- mediate between parvulus and Certhidea olivacea luteola. Cactospiza pallida, together with some forms in the Camarhyn- chus psittacula superspecies, is the palest of the Geospizinae. Specimens of Cactospiza p. pallida, from James Island, are very gray above, especially on the head, resembling the grayest Cama- rhynchus psittacula; those from Indefatigable are much more olive, being more olive than psittacula; those from Duncan are but 68 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS slightly less olive. Birds from south Albemarle (formerly sepa- rated as C. p. producta) also show an olivaceous tinge, less marked than in birds from Duncan; they closely resemble typical psittacula except that they are not noticeably barred. Specimens from north Albemarle (formerly called C. p. producta) are almost as gray as birds from James. The color differences between specimens from James and Indefatigable seem more pronounced than be- tween some races of Certhidea, perhaps enough to justify reviving the name hypoleuca (Ridgway). The specimens are normally im- maculate below, but a few show slight streaking on the breast. Birds from James are palest and: resemble the palest psittacula; birds from Indefatigable and Duncan are more buff than gray; those from south Albemarle intermediate; those from north Albe- marle similar to those of James. The wing bar shows as much indi- vidual variation as in most groups, varying from sandy to pale buff. It is palest in birds from James, and brownest in the more oliva- ceous forms, especially on Indefatigable and Chatham. Cactospiza p. striatipecta, from Chatham Island, is olivaceous above and often has traces of darker barring, especially on the head, in which it resembles Camarhynchus psittacula. It is nearest to the Duncan birds. However, it is more olivaceous than most psittacula, al- though overlapping occurs. Cactospiza p. striatipecta from Chat- ham is streaked below. Cactospiza heliobates is darker above than pallida and as dark as the MFF group, but with olive not pale gray tips to the feathers. The under parts are more streaked than in pallida, in this respect resembling Camarhynchus pauper, but not as heavily streaked as in the MFF group. The wing bar is incon- spicuous, usually brown. “‘Cactospiza giffordi,” represented by one male, perhaps in juvenal plumage, is probably not a valid species. The upper parts are darker and less olivaceous than typical pallida or typical Certhidea o. olivacea from Indefatigable. The under parts show a faint green-olive tinge, slightly distinguishable from typical pallida or C. o. olivacea. Traces of orange on the breast suggest Certhidea; otherwise it might have been considered simply a dwarf Cactospiza pallida. The miniature C. pallida seen on In- defatigable and perhaps referable to this form (see Section I) had plumage identical with pallida, and showed no trace of buff or orange on the throat. Pinaroloxias inornata from Cocos Island is similar to G. d. sep- tentrionalis. The upper parts are somewhat darker; the feathers No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 69 show similar olivaceous tips which are brighter and more exten- sive, especially in the rump region. The under parts have a bufly tinge. This species has a type of ventral streaking similar to sep- tentrionalis, but is less streaked on the breast, and does not show the dark throat, so appears paler. It has a well-developed rufous wing bar as in septentrionalis. In both male and female plumage, this species resembles G. d. septentrionalis, although in bill and feeding habits it resembles Certhidea. PLUMAGE VARIATIONS IN ADULT CERTHIDEA In the following discussion, the male and female plumages are not described separately because, in the various species of Cer- thidea, the plumage of the sexes is similar. The forms from the different islands are extremely similar in the plumage of the upper parts. All have a general brown-gray coloration, with some greenish-olive. Some are browner, some grayer, some paler, some more olive. Some individuals of each island form overlap in characters with at least one, and often with several other forms, the olive tinge being especially variable. The several forms differ in the percentage of individuals show- ing a rufous wing bar, but this is of little value as a diagnostic character and is difficult to classify, particularly since occasionally in the same specimen some wing coverts may be buff, some brown, and some rufous. For description of the species see Swarth (1931, pp. 250-267). Only one form, Certhidia o. olivacea, shows some indication of inter-island variation. The specimens from James Island are dull brownish-gray above, with some olive; those from Albemarle are similar but are slightly more olivaceous; those from Indefatigable are appreciably more greenish-olive, slightly paler than on Albe- marle, but with some overlapping. Eight specimens from Cross- man seem grayer and less olivaceous than Albemarle specimens, and look somewhat like ridgwayi, but all are too worn for a definite decision. The under parts are pale olive-buff. Many Indefatigable birds seem somewhat more olive-buff than the others; the Crossman specimens are too worn for comparison, but seem buff not olive, and they look not unlike becki. In specimens from James Island, the wing bar is bright rufous; about half show a rather dull rufous brown, the rest being brown-buff or inconspicuous. More than half 70 CALIFORNIA ACADEMY OF SCIENCES [Oc. Papers the Albemarle specimens show a rufous wing bar, but it is fre- quently a darker rufous-brown, and in more than one-third of the specimens the bar is buff or is absent. It would be valuable to know the immature plumages of the various forms of Certhidea, but the present series is too small for generalized descriptions, and possibly there are individual varia- tions. Swarth described the few that are known. The streaked im- mature plumage of C. o. ridgwayi is of special interest, as it ap- pears to link Certhidea with the other Geospizinae. The only secondary sexual plumage character is the chestnut or orange coloring on the throat and breast of the male. This is extremely well developed in Certhidea o. olivacea from James Is- land, less so in olivacea from Albemarle Island, and appreciably less in olivacea from Indefatigable and Duncan islands. Certhidea o. ridgwayi from Charles shows less development of this coloring, although many specimens still possess it. In C. o. becki, every adult male specimen shows definite traces but less intense in color and the area of the patch is small. Certhidea o. fusca, C. 0. mentalis, and C. o. luteola show traces only in some individuals. Indeed some writers have described the patch as being absent in these forms. Finally, in C. o. bifasciata and C. o. cinerascens there is only a slight trace in a very few individuals examined. Ten females, five from Albemarle, and one each from Narbor- ough, James, Jervis, Duncan, and Barrington islands, show some chestnut on the throat and breast. Such a number of specimens could not be due entirely to collectors’ error in sexing the birds. Occasionally females showing male coloration are found in other species of birds. This may sometimes be correlated with abnor- malities of the sex organs. Plumage differences are the main criterion for separating the various forms of Certhidea. Swarth gave different names to the forms on most of the islands, but united the birds on James, Inde- fatigable, Duncan, and Albemarle. In the latter, he seems to have been influenced, perhaps unduly, by one character, namely the chestnut throat patch of the male. When the other plumage char- acters are considered, it is perhaps justifiable to separate the birds from James, Indefatigable, and Albemarle. Thus, excepting for the male’s chestnut patch, C. 0. olivacea of Indefatigable more closely resembles luteola of Chatham than it does olivacea of James. An additional point, which only further collecting can de- No. 21] LACK: VARIATION IN GALAPAGOS FINCHES aL termine, is the possible distinctness of the Crossman specimens, the eight males examined being in too worn plumage to be of value. There are no diagnostic characters of plumage completely sepa- rating one form of Certhidea from the others; in almost every form some individuals overlap with some specimens of another form. It is curious that the birds of Culpepper and Wenman islands in the extreme north should come closest to those of Charles in the extreme south, while those from Abingdon and Bindloe in the north come closest to those from Hood in the south. Conceivably this might be due to evolution among the birds of the central islands leaving those on the extremities unchanged. A possible alternative is convergent evolution. Another curious instance is the resem- blance between the birds of Chatham and Indefatigable, whereas those on Barrington, which lies between these two islands, are dif- ferent. In view of these difficulties, it does not seem practicable to attempt to unite the forms on the different islands in one evolution- ary tree. DISCUSSION OF PLUMAGE VARIATIONS IN THE GEOSPIZINAE All the data uphold the view that the black male plumage of the Geospizinae is in the process of being lost. One can be less certain whether Certhidea is in the process of losing or acquiring the chestnut throat patch in the male. The apparent absence of any display correlated with this coloration makes it seem probable that it is being lost, not acquired, and this fits the general geospizid trend of loss of male secondary sexual characters. If the throat patch is being lost, one wonders how and when it was acquired, since Certhidea is certainly one of the Geospizinae, and hence pre- sumably evolved from an ancestral form with black male plumage. A possible explanation is that the genetic factor or factors for this throat patch were present in the ancestral Geospizinae, but were masked by the subsequent development of black feathering, a view independently suggested by the sporadic occurrence of the throat patch in male Geospiza and Camarhynchus, already mentioned. That black coloration may conceal an underlying pattern is shown by Pycraft (1925, p. 274) who figures an isabelline variety of the rook (Corvus frugilegus), revealing a color pattern which was pre- viously unsuspected, and which is normally concealed completely by the black plumage. 72 CALIFORNIA ACADEMY OF SCIENCES [Oc. ParErs In female plumage, the different species of Geospizinae and the different island forms of the same species differ mainly in the shades of brown, gray, olive, etc., of the upper parts; in the degrees of streaking of the under parts, and in the coloration of the wing bars. All recent work indicates that such differences are inherited. However, the differences do not seem to be adaptive. The various Galapagos Islands provide extremely similar environments, and the various species and island forms vary in a most haphazard way both with reference to the same species on different islands, and to different species on the same island. Thus, Cactospiza pallida is grayer on James and north Albemarle islands, more olive on Inde- fatigable and south Albemarle, and more streaked on Chatham. As compared with the form on James and Indefatigable, the Cama- rhynchus psittacula superspecies is grayer on Abingdon and Bind- loe, browner and somewhat streaked on Albemarle, while on Charles the form presumably derived from Albemarle is browner and more streaked, and the form derived from Indefatigable closely resembles typical psittacula. The best development of the rufous wing bar in the Geospizinae occurs in Certhidea o. olivacea on Albemarle and in Geospiza d. septentrionalis on Culpepper and Wenman. Geospiza d. difficilis on Abingdon and G. d. debilirostris frequent the humid zone, and differ from other forms of Geospiza in being darker and more streaked, the same trend being shown by G. scandens, which is typical of the arid coastal belt, while the other forms of Geospizinae characteristic of the humid belt, notably Cactospiza pallida, are much paler than Geospiza. Also the per- centage of male Platyspiza and Camarhynchus in partially black plumage is different on the various islands, with no general trends or island correlations, as already discussed. Many more examples of such haphazard variation could be cited. No species shows any established trend of color variation except for the possible radial distribution in Certhidea, already mentioned. In most cases, several species do not vary in the same direction on the same island. A possible exception is that, on Culpepper, Geospiza conirostris darwini and two of the four speci- mens of G. fuliginosa (all possibly stragglers and conceivably hybrids with G. d. septentrionalis) show a tendency toward a rufous-brown wing bar, buff on the under parts and olivaceous tips to the back feathers, similar to the way in which G. d. septen- trionalis differs from most other forms of Geospiza. Furthermore, No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 73 there is possibly a tendency for G. fortis and G. fuliginosa to be slightly darker and more streaked on Abingdon Island; this also applies to G. scandens there. Both Camarhynchus parvulus and Certhidea are greener on Chatham than elsewhere. Such instances, however, are rare, and in no instance are all of the forms of Geo- spizinae on the same island affected similarly. Hence, the varia- tions seem most probably to be due to chance. All present evidence is in favor of the supposition that in the Geospizinae the plumage differences between species, and between island forms of the same species, are not adaptive; in this there is agreement with the varia- tions shown in size of bill and wing (see full discussion later). In the Geospizinae, plumage characters seem to have been more conservative in evolution than bill characters. Probably bill differ- ences have been evolved more rapidly than plumage differences. In mainland birds, plumage characters are usually much more vari- able than are structural characters, but in general among insular birds structural variations seem to be relatively more pronounced, as is discussed later. Hence, among insular birds, plumage is some- times a more reliable guide to relationships than are structural characters; this is well shown among the Geospizinae by Cacto- spiza pallida and Geospiza scandens which were formerly united on account of bill characters in the genus “Cactornis.” Plumage characters indicate that pallida is much more closely related to Camarhynchus, and G. scandens to other forms of Geospiza than either is to the other. Habits and other characters fully support this conclusion. COLORATION OF THE BILL As reported by Swarth (1931), the bill in all Geospizinae of both sexes is normally dark in the breeding season, and pale out- side of the breeding season. A similar change occurs in certain Ploceidae, and Witschi (1935) has recently experimented on the hormones controlling this. From the large available series, it is clear that the black males of the species of Geospiza and the par- tially black “full-plumaged” males of the forms of Platyspiza and Camarhynchus develop the dark bill at the beginning of the breed- ing season earlier than do either adult females or young males. Occasionally males, and more often females, breed with a bill which is only partially dark. This occurs particularly in female 74 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS Certhidea, but was also observed in Geospiza magnirostris, fortis, fuliginosa, and scandens. The progressive changes in the coloration of the bill in the juvenile have not yet been worked out. Some young leave the nest with an entirely light-colored bill; in many it is dark above and yellowish below. Whether the latter changes to plain yellow or pinkish-yellow during the non-breeding season, like that of an adult, is not known. In this partially dark condition, the bill of the immature closely resembles that of adults who are changing from breeding to non-breeding condition. Hence, color of bill does not differentiate adult from immature specimens. On the other hand, several specimens collected near the end of the breeding season, with almost unossified skulls, and, therefore, presumably young hatched a month or two before, sometimes showed completely dark bills. It would appear as though none of the supposed character- istics of sexual maturity in the Geospizinae, that is black male plumage, dark bill, and ossified skull, are strictly correlated with the development of the sex organs. SECTION V. VARIATIONS IN BILL AND WING Some species of Geospizinae are so variable that to base conclu- sions on ten selected specimens, as did Swarth (1931), may be misleading. Actually, the whole group has been so extensively col- lected that it is extremely suitable for a general statistical study, and I therefore measured almost all of the specimens available in museum collections. As can be seen from the tables which follow, the majority of the species are adequately represented from most of the islands on which they occur. The actual measurements are, of course, far too extensive for publication, but since other workers may find them valuable, I am depositing copies with the California Academy of Sciences and with the British Museum, Natural His- tory. Modern statistical workers consider it important that when the actual measurements cannot be published, at least the frequency distributions should be. These also, however, are too bulky to pub- lish, so I am depositing copies with the two above-mentioned insti- tutions and also with the United States National Museum, Wash- ington, D.C., so that those interested may consult them. MeETHODs oF MEASUREMENT The Geospizinae vary mainly with reference to bill, so that for each species I measured the culmen and the depth of the bill. Orig- inally, I used the standard culmen measurement, that is from the tip of the bill to the place where the culmen joins the skull, but I found that, particularly in Geospiza magnirostris, Platyspiza cras- sirostris, and Certhidea olivacea, I was unable to estimate accu- rately the exact place where the culmen ended on the skull. Thus measurements of the same specimens taken a month or two later - sometimes gave different results. Dr. Alden H. Miller informed me that he experienced similar difficulty in his study of the genus Junco, and at his suggestion I measured the culmen from the an- terior part of the nostril opening to the tip of the bill. This is an extremely reliable measurement. It is unfortunate that I could not use the whole culmen, as, until this time, it has been the standard bill measurement used by taxonomists working on this group. Fur- thermore, in one or two species, notably G. magnirostris and G. fortis, there is a greater difference when the whole culmen is com- pared than when the culmen as measured from the nostril is used. [ 75 ] 76 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS Possibly, had I been more experienced in measuring birds, I could have taken the whole culmen measurement accurately, but after three months of intensive work I found it still unreliable. Wher- ever, in the present text, the word “culmen” is used it refers to the length of the culmen from the nostril, and not to the whole culmen, unless expressly stated. Swarth (1931, pp. 137-270) records a set of measurements of the whole culmen for each species. The depth of the bill was measured at the base, with the bill closed. It is sometimes difficult to be certain of the position of the base of the bill, but I found this measurement much less liable to personal variation than was that of the measurement of the whole culmen. In some specimens, the stuffing in the mouth prevented the bill from closing properly; such specimens were not measured. In Certhidea olivacea, the bill is so shallow that the degree of error is considerable; hence the measurement for the depth of the bill in this species is not of great value. In addition to these two bill measurements, I measured the length of the wing in the standard manner, that is from the carpal joint to the end of the longest primary in the flattened wing. The culmen (from nostril) and the depth of the bill were meas- ured with a pair of dial calipers, kindly loaned by the California Academy of Sciences, except for the specimens in the British Mu- seum, Natural History, for which I used calipers with a vernier scale. Both of these instruments recorded to 0.1 mm. which is a finer unit than the error in estimating the limits of the bill to be measured. The wing was measured with a millimeter rule, and recorded to the nearest millimeter. In the tables which follow, the number of the specimens measured is given first (N), then the mean (M), followed by the standard deviation (a). For calculat- ing the mean and standard deviation, the actual measurements were used, and they were not first grouped into broader divisions. In addition to these, I also include the range of each form, since, al- though the standard deviation gives a more accurate indication of the range of variation, the maximum and minimum specimens measured have often formed a standard part of taxonomic pro- cedure. Birds collected shortly after they had left the nest were excluded. The bills of such individuals were usually dark above and pink or yellow below. All specimens in streaked plumage with bills of this color collected between March (earlier for G. scandens and Cacto- No. 21] LACK: VARIATION IN GALAPAGOS FINCHES a spiza) and June were rejected, since most young leave the nest at this season. The color of the bill of the adult is changing from black to yellowish during the latter part of this period, so that the rejected specimens undoubtedly included a number of adults. How- ever, to be on the safe side all specimens with partly dark and partly pale bills at this season were excluded, except for males which already showed complete or partially black plumage, and which could safely be included as adults. The bill of an immature individual is probably fully grown, or very nearly so, two or three months after it has left the nest. The only immatures which may conceivably have been included among the adults measured were those that emerged from nests at an unusual time of year. This number is certainly small, and any which might have been imma- ture, as indicated by plumage or color of bill, were rejected, irre- spective of season. SizE VARIATIONS DuE To AGE Males of Geospiza spp. upon leaving the nest possess plumage essentially similar to that of the female; later the fully black plum- age is assumed. It is uncertain at what age the black plumage is acquired, and whether or not this varies individually. A great many males breed in streaked or partially black plumage, but it seems reasonable to assume that, on the whole, males in streaked plumage are younger than those in fully black plumage. The same may apply to partially black males in Platyspiza and Camarhyn- chus. It is well known in many birds that males several years old have a longer wing than first-year males. Hence for the wing measure- ment in the species of Geospiza and Pinaroloxias only fully black males were included. In Platyspiza and Camarhynchus, only males showing some black in the plumage were measured. That the older birds, as indicated by their plumage, have a longer wing is demon- strated in Text Table 5, where typical examples of each species are given. There is also the possibility that the older males have larger bills, although this has rarely been recorded in birds. Mayr (19342) cites this in a study of the hornbill (Rhyticeros plicatus). To test this, Text Tables 6 and 7 were prepared. Tables 6 and 7 show that there is a significant difference in the mean bill size for black males as compared with males in streaked 78 Species Islands Black Males No, Length o No, Length Geospiza magnirostris Abingdon 18 82,2 1.87 10 81.6 " i) James 16 83.9 2,12 12 81.3 = fortis Chatham 56 73.7 2.08 25°" 71.5 . Be Charles 102 72,1 2.47 48 (72.3 " fuliginosa Chatham 62 64,0 1,54 34 63,4 \- ms Charles 41 64,0 1,64 18 63.4 Pinaroloxias inornate Cocos 78 «68,1 1,25 17 67,1 Geoaspize d.septentrionalis Wenman © 2671.91.54 a " debilirostris James 320-7167 25 " scandens Indefatigable 40 72.6 1,11 ° o) Charles 59 71.2 1.54 ® conirostris Gardner (Hood) 38 78.9 2.43 Platyspiza crassirostris S, Albemarle 34 84,2 Camarhynchus habeli Bindloe Se aiiee: Ld pauper Charles 21 «70,4 . parvulus James 18 63.5 « © Charles 60 64,2 . p. salvini Chatham 65.3 Species Island Bleck Males No, Length o Geospiza magnirostria Abingdon 18 16.02 .654 . * James 16 15.79 «794 " fortis Indefatigable 29 11.99.7782 " . Chatham 56 12,32 536 3 c Charles 100 «11,62 836 Lo fuliginosa Indefatigable 33 8,46 371 ‘. ct Chatham 61 9.03 .517 0 ~ Charles 41 8.60 403 Pinaroloxias inornata Cocos 78 10,49 = 352 Geospiza d. septentrionalis Wenman 26 «10,74 465 ve " debdilirostria James 32 «10,24 370 " scandens Indefatigable 42 15.00 .721 . bl Charles 58 13,90 612 - conirostris Gardner (Hood) 38 14.77.9221 Platyspiza crassirostris S, Albemarle 5 ! Indefatigable c ~ Chatham Camarhynchus habeli Bindloe me ‘pauper Charles it parvulus James ) s Charles Species Island Black Males No, Length o Geospiza magnirostris Abingdon 16 19.99 _.668 "fortis Charles 100 12.33.97 / fuliginosa Chatham 57 8.23 406 * * Charles 31 8.30 .388 gp eh 22 8.36 _.267 y ‘scandens Charles 49 9.44 394 Platyspiza crassirostria CALIFORNIA ACADEMY OF SCIENCES Table 5 Variation in Wing Length with Respect to Sex and Age in the Geospizinae S, Albemarle Camarhynchus habeli Bindloe a pauper Charles = parvulus Charles Partly Black Males ov 2,88 1,42 2.22 3.00 1.45 1.89 0,83 1,87 1.55 1.54 1.54 1,52 Partly Black 4 Streaked Males No, Length 41 71.3 ll «(70.9 440 °«-71,1 45 1.4 52) i103: Table 6 Streaked Males o No, Length o 23 80.0 2,01 18 81.2 1.89 2B «69,6 2.60 26 70.4 «2.43 28 61.6 1.87 27 62.1 1.36 21 66.6 1.25 1.81 1.22 ‘ 1.97 2,07 2.34 23 82.7 1.98 ll 70.2) «1.17 59 69.7. 1.78 10 62,8 1,40 26 62.9 1.59 87 64.3 1.53 Variation in Length of Culmen with Respect to Age in Male Geospizinae Partly Black Males No, Length 10 16,17 12 15.99 390 «:11.93 25 12,19 50 11,81 28 B40 35 8.65 18 8.59 19 «10,22 34 =10,36 16 10,38 9 10,74 13. 10,58 al 9.06 18 6.93 54 1.33 Table 7 o +723 «605 +790 +639 334 «374 +708 435 +390 Partly Black & Streaked Males No, Length o 42 10.72 .397 ll 10,16 «=. 284 45 14,89 ,814 44° 13,85 629 34 14.50 1,052 Variation in Depth of Bill with Respect to Age in Male Geospizinse Partly Black Males No, Length 27° (12,18 11 10,60 18 8,96 34 7.48 oc Partly Black &.Streaked Males No, Length o 38 «8.23 £353 419.41 396 Females No, Length 29 78.7 21 80.0 64 69.6 97 «(68.8 108 61,0 87 61,2 43 «(65.4 3568.5 25 69.4 32 69.3 33 68.2 42 15.2 26 «80.8 14 68,1 62 68,2 14 60,0 38 «61,0 62,4 Streaked Males No 22 18 37 28 24 19 28 27 21 22 12 28 1 60 10 26 « Length 16,00 16,02 11.95 11.99 11.59 8.37 8.40 8.60 10.30 10,10 10,51 10.39 10,40 8.99 6.99 1.33 2395 227 247 269 335 388 129 «257 Streaked Males No, 23 26 26 24 47 Length 19.89 12,14 7.88 7.85 11,63 10,43 8.79 7.51 o 1,065 1,094 432 +296 548 +300 440 +305 No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 79 plumage in Geospiza fuliginosa on Chatham (in culmen only), for Platyspiza crassirostris on several islands, and for Camarhynchus psittacula on James and Indefatigable (depth of bill only). In each instance, the older males are the larger. For other species and island forms, the differences are either nonexistent or, when present, are extremely small, and, treated separately, are usually not statistically significant. However, when the series is considered as a whole, there is possibly a tendency for the older males to be slightly larger. In view of this slight difference, some workers might prefer to use only the measurements of older males when comparing different forms. On the other hand, if all the available males are used (excluding fledglings) many more specimens are available, so that this larger grouping is in some respects pref- erable, particularly since the difference in mean is so small (rarely more than 0.1 mm.). The position is complicated further since, particularly in a number of forms, some males never attain the black plumage. This percentage varies on the different islands, so that it is difficult to know whether all males, or only those show- ing black, are more “‘typical’” of the species (which is, after all, primarily a breeding unit). Accordingly, in the Main Tables I-X, pp. 142-151, all males are grouped together for bill measure- ments, but, following the Main Tables, the measurements for older males alone are given. In Geospiza magnirostris and G. fortis, there is one more difler- ence noticeable when the bills of males in black and streaked plum- ages are compared, namely that the black males seem to be less variable than the streaked males or than all of the males taken together. No such difference is apparent in the other species, and in G. magnirostris it seems to apply only to depth of bill. (See Text Table 8, p. 86, and compare the figures in the Main Tables for other species. ) In the species of Cactospiza, the immature males cannot be dis- tinguished from the older ones, since the plumages are similar. This also applies to females of all species, so that these represent birds of all ages. Sex DIFFERENCES In every species except one, the females have smaller bills than the males, both with reference to culmen and to depth of bill. This slight difference possibly has a similar genetic basis to that of such 80 CALIFORNIA ACADEMY OF SCIENCES {Oc. Papers sex differences as stature in man, discussed by Danforth (1939, pp. 344-345). Contrary to the other species, in Certhidea olivacea the female has a longer bill than the male. In this species, the female has different feeding habits from the male, foraging nearer the ground. Whether or not, however, the bill difference is corre- lated with feeding habits is unknown. (For details see the Main Tables, pp. 142-151). In variability of size of bill, male and fe- male show little difference in any species. In wing size, Text Table 5 indicates that in the forms of Geospiza the black males have a longer wing than the partly black males, which in turn average longer than .the streaked males, and these average longer than the females. The same applies to the partly black males as compared with streaked males and with the females in members of the genera Platyspiza and Camarhynchus. In all the species, including those of the genus Certhidea, the wings of the males average longer than those of the females. From the Main Tables, it would appear that in wing length the female is more variable than the male. This presumably results from the fact that wing length varies with age. Thus while the young males can, in most species, be separated by plumage, the females cannot, and must be grouped together. When the males are grouped together, as was necessary in Cactospiza, the standard deviation of the male wing is as large as in the female. SPECIMENS OF QUESTIONABLE IDENTITY As already noted in Section I, a number of specimens do not conform to any of the known species. A few seem to be aberrant specimens, while others are intermediate in characters between two described species. These specimens have not been included in the Main Tables, but are listed separately in Text Table 16. I have also included in this same list of questionable specimens a number which I have assigned to described species, since their allocation is partly a matter of personal opinion, and other authorities might differ from me. Obviously, the inclusion of abnormal specimens in the calculation of the mean and standard deviation in an island form will affect both of these figures, particularly the standard deviation, and especially if not many specimens are available. For Geospiza magnirostris and G. fortis, I have included in the tables not only the most questionable specimens but also a number of other specimens, in order to illustrate the similarity in measure- No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 81 ments of these two species. It is unfortunate that the most reliable character separating these two species is the length of the whole culmen, which, as stated elsewhere, I could not measure accurately. The shape of the culmen often gives a different appearance to the birds and makes identification of large specimens of G. fortis and small specimens of G. magnirostris easier than would appear from the table of measurements. DIFFERENCES BETWEEN ISLAND POPULATIONS OF THE SAME SPECIES 1. Differences in mean In the Main Tables I—X the population of each island has been treated separately, and, in addition, the populations from north and south Albemarle have been kept distinct as these two localities are far apart. Nearly all specimens labeled “north Albemarle” were collected at Tagus Cove, those labeled “south Albemarle” at Iguana Cove in the southwest and Vilamil in the southeast. For Geospiza fuliginosa some specimens are also available from cen- tral Albemarle, mostly labeled Elizabeth Bay or Perry Isthmus. It can be seen that in many species the island forms are slightly different. In a few instances the differences are sufficiently large for the forms to have received separate subspecific names. In many other cases, the degree of difference is too small to justify a nomen- clatural separation, although the differences are statistically sig- nificant. On the whole, as compared with other oceanic islands, the differences between island forms on the Galapagos are rather slight. From the summaries which follow, it will be seen that no species shows any completely regular trend of variation, that is from north to south, or in any other direction, or radially from any one island outward. In G. fortis, there is a tendency for increasing size in one direction, but it is not completely regular, and in G. scandens, the population of each island seems to be almost independent of the others. The relative independence of each island population is well shown on Bindloe, which lies between Abingdon and James. In three species, namely G. magnirostris, G. fuliginosa, and G. scan- dens, the populations on Abingdon and James are more like each other than like those on Bindloe. In the main, each island popu- lation seems to be evolving independently, but there is apparently 82 CALIFORNIA ACADEMY OF SCIENCES [Oc. PArErs a little inter-island wandering, which means that the populations of neighboring islands will show some resemblances. The data for the variations in the mean bill and wing sizes for each species are as follows: Geospiza magnirostris——The largest birds are from Tower Is- land in the northeast. Those on Wenman, in the far northwest, average smaller. Those on Abingdon are intermediate. Bindloe is only 13 miles from Abingdon, but in bill size magnirostris there averages nearly a millimeter smaller, although those from James, farther south, are closely similar to those from Abingdon in size. Jervis is a small island only five miles from James, but the birds there have a significantly smaller bill. Swarth’s contention that G. magnirostris tends to be larger in the north and smaller in the south is not altogether true, since the Wenman population in the extreme north is smaller than that on Tower, Abingdon, or James islands, and the Bindloe birds are smaller than those of James to the south. The apparent small size of birds from Indefatigable and south Albemarle (Main Table I) may not be altogether reli- able, since not enough specimens were available. It is also possible that they interbreed occasionally with G. fortis, and that some of these hybrid specimens, not necessarily first generation, have been included under G. magnirostris. Near the end of Main Table I are included the measurements of the specimens collected by Darwin, and it can be seen how much larger his “G. magnirostris” are than any specimens collected sev- enty years later. On the other hand, his “G. strenua” are like modern forms. Geospiza fortis——For this species Swarth claimed a trend of increasing size from north to south, but this is not altogether true. The largest birds are found on south Albemarle and Chatham, but on Charles, farther south, they average smaller, and this variable population includes some very small specimens. In bill size, birds from Abingdon and Duncan are smaller than those from other islands. Those from north Albemarle average about 1 mm. smaller in bill than those from south Albemarle. The forms on Daphne and Crossman are discussed later; they are thought to be of hybrid origin with G. fuliginosa. The form on Duncan is smaller than on either of the neighboring islands of Indefatigable and Albemarle, and it is conceivable that here also there has been some interbreed- ing with G. fuliginosa. In wing size, the birds on Abingdon and No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 83 Bindloe are distinctly smaller than those from other islands. Those on south Albemarle and Chatham are larger than elsewhere. Geospiza fuliginosa.—The bill variations in this species are not great. The form on Bindloe Island is of interest, since it is smaller than that on Abingdon to the north and James to the south. The wing variations are also slight, but, as in G. fortis, the birds on Abingdon and Bindloe are significantly smaller than elsewhere. Indeed, at one time they were separated as a subspecies, G. f. minor, but the overlapping is too great to justify this. Geospiza difficilis—tThe three subspecies of this species show fairly marked differences in bill and wing size. Furthermore, each subspecies occurs on two islands, and in each such pair the popu- lations on the two islands show significant differences in mean size, although not sufficient to justify subspecific separation. (See the Main Tables for details.) Geospiza scandens.—There is a large form on Abingdon. On Bindloe, the next island to the south, there is an even larger form. On James, the next most southerly island, the form is extremely small, and all specimens can be readily separated from those on Bindloe. On Jervis, only five miles from James, the mean is sig- nificantly larger than on James, and becomes larger still on Inde- fatigable, south Albemarle, and Barrington to the south. It is a little smaller again on Charles in the extreme south, while the form on Chatham to the southeast has a distinctly shorter bill, although it is no less deep, than those of the central islands. In other words, regular trends are almost absent in this species, and, although specimens from James and Bindloe are readily separable from each other and therefore could be named, the specimens from the other islands overlap with both and make consistent treatment im- possible. (See Section I and Frontispiece map.) Mayr (1940, pp. 262-263) records similar irregular size trends in the honey- eater (Foulehaio carunculata) on various Polynesian islands, which again make subspecific treatment impossible. Geospiza conirostris.—lIt is remarkable that the population on Gardner, less than a mile from Hood, should average nearly 1 mm. smaller in wing length and both bill measurements than the popu- lation on Hood. The other forms are considered later. Platyspiza crassirostris——This species shows less variation be- tween different islands than any other. Possibly the birds on Nar- borough and north Albemarle, few of which have been collected, 84 CALIFORNIA ACADEMY OF SCIENCES [Oc. Papers are slightly smaller than those elsewhere; apart from these there does not appear to be any significant variation in size of bill or wing. This stability is conceivably due to the fact that this species may wander more frequently from one island to another than do the other species. Thus, although it seems unlikely, the island pop- ulations may not be isolated from each other. Platyspiza crassi- rostris is also well separated from all the other species, and only one aberrant specimen, a dwarf bird from Narborough, has been examined. Camarhynchus psittacula.—The form habeli on Abingdon and Bindloe islands differs not only in-bill measurement, but also in shape of bill, from psittacula (sensu stricto) on James, Indefati- gable, and other islands (Swarth, 1931, pp. 219-222). Cama- rhynchus affinis on Albemarle is smaller than psittacula. The un- usual specimens found on the central islands, perhaps referable to affinis, have been discussed in Section I, as has the Duncan popula- tion. Camarhynchus pauper on Charles Island is discussed later. Camarhynchus parvulus——Birds from James and north Albe- marle have slightly smaller bills than those on Indefatigable, south Albemarle, and Charles islands, but the differences are not sufh- ciently great to justify subspecific separation. On Chatham there is a larger form, C. p. salvini. Cactospiza pallida.—The birds on James are significantly larger than those on islands to the south. Those from north Albemarle are larger than those from south Albemarle. Those from Chatham are particularly small, and have been separately named. Certhidea olivacea.—The inter-island variations in this species mainly concern plumage, but there are also slight variations in measurements. In particular, the forms on the central islands, James, Indefatigable, Duncan, and Albemarle, have smaller bills than those on the islands both to the north and to the south. The birds from Charles have appreciably deeper bills than those from other islands, a difference which is more apparent to the eye than the measurements indicate. There are also slight, but significant, wing variations, and here also the birds on the central islands tend to be smaller (see Main Tables). 2. Differences in variability The standard deviation gives a convenient measure of the degree of variability when comparing the island forms of the same species. No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 85 It can be seen from the Main Tables that in some species the island forms differ not only in mean, as already discussed, but also in variability in both bill and wing. The data presented in the Main Tables provide material for examining the view, first advanced by Darwin, namely that small populations are less variable than large populations. This view was recently tested statistically by Fisher (1937) for birds’ eggs, and was also developed at length on theo- retical grounds by Sewall Wright (1931, 1940). Environmental conditions in the coastal zone on the various Galapagos Islands would seem, so far as they affect the species of Geospiza and Cer- thidea, to be similar on the different islands. Hence, it is reasonable to assume that, for these species, the population on a small island will tend to be smaller than that on a large island. In the present investigation, the population must also be isolated from that of other islands because if there is regular mixing with a neighboring island, the populations of the two islands cannot be considered separately. Culpepper and Wenman are extremely small islands, separated from the main group, and a moderate distance from each other. Tower is small and isolated, although not to the same extent. Ab- ingdon, Bindloe, and Hood are moderately small and somewhat separated from the other islands. It should be noted that the pre- vailing wind is from the southeast, which means that the two east- erly islands of Tower and Hood are more isolated, for birds, than their positions on the map seem to indicate. The distances between the islands are given in Text Table 1. One can, therefore, compare the populations of these smaller, more isolated islands with the populations of the larger islands, together with those of the small islands situated near them. The populations of these larger islands probably mix to some extent with those of the neighboring ones. For this investigation, all the measurements given in the Main Tables can be used except those of the wings of females. The latter must be omitted, since, as mentioned earlier, it is not a homogene- ous sample, but includes birds of all ages. Hence, its variability depends upon the relative numbers of streaked and adult specimens included, which is unknown. Details are given in Text Table 8 for Geospiza magnirostris and G. fortis. I have given the values both for all males together and for those in full black plumage (see earlier discussion). ~ In Geospiza magnirostris, the birds on Wenman and Tower, the 86 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS two smallest and most isolated islands, have smaller standard devi- ations for all five measurements than have the birds on the larger islands. This is true for the male bill measurements whether all males or only those in fully black plumage are considered. Also, the populations on the moderately small and more or less isolated islands of Abingdon and Bindloe tend to have smaller standard deviations than those of the larger islands of James, Indefatigable, and Albemarle. As might be expected, the population on the small island of Jervis, which is close to James, has a similar variability to that of the large islands. Table 8 Variation in Standard Deviation of Culmen, Depth of Bill and Wing Length in Various Breeding Populations of Geospiza magnirostris and Geospiza fortis Geospiza magnirostris Culmen Depth of Bill Wing Islend Description All Males Black Males Females All Males Black Males Females Black Males Wenman Extremely small, isolated 636 650 2572 ~888 .707 332 1,09 Tower Small and isolated 2599 634 0592 1,127 786 2615 1,42 Abingdon Moderately small, isolated 2657 2654 633 941 668 ~ 783 1,87 Bindloe " : 770 768 2704. 954 891 +973 2,10 James Large, not isolated 702 2194 1,003 1,025 Bio 1,305 2eL2 Indefatigable " G ul ©7199 -265* «735 1.753 938 1,633 1,50* Jervis Small, not isolated 720 737 -659 1,067 1.148 nD 2.68 Abingdon Moderately small, isolated 511 442 oer! 661 615 541 2.06 Bindloe t " ul 466 2305 315 -410 2356 2473 1.27 James Large, not isolated +556 534 0599 714 2645 -781 1.75 Indefatigable " HW iw .809 . 782 -680 1,321 1,497 1,183 2.79 Narborough a " " 481 481 ~643* 516 516 1.85 N, Albemarle w " us 726 583 880 1,055 889 1,209 2,14 S, Albemarle " © " 831 827 -987 1,169 1.107 1.424 2.26 Seymour Small, not isolated . 780 429 662 ede 2552 2919 1.70 Duncan n " " 728 0575 2554 862 2615 +946 2.26 Chatham Large, moderately isolated ,668 536 765°) alevs 908 1,296 2,08 Charles n it " 883 836 -818 1,222 2917 963 2.47 * Insufficient number of specimens, Geospiza fortis is unfortunately not resident on Wenman, Tower, or Hood. The Hood birds were considered to be stragglers, hence they were not included. The populations on Abingdon and Bindloe tend to have smaller standard deviations in the bill measurements of both sexes than do those from the larger islands. This is not true with reference to the wing measurement of black males. Barrington No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 87 is a small island, moderately isolated, and here also the population has a small standard deviation. However, not many specimens were available. To judge from the numbers collected, G. fortis is con- siderably less numerous on James than on the islands to the south, so the James population should perhaps be considered as a mod- erately small one. Hence, Text Table 8 definitely supports the views advanced by the authors mentioned above. The standard error of the standard deviation is large, so that any one case taken singly is seldom sig- nificant. All of the small island forms, however, show the same trend in all measurements excepting the black male wing in G. fortis, and, taken together, the data seem significant. The difference is perhaps not as great as might have been expected, considering how extremely small some of the island populations must be as compared with others. However, it must be remembered that prob- ably no population is completely isolated, and an occasional bird wandering from one of the larger to one of the smaller islands and interbreeding there would at once make the small population more variable. The records of individuals which are known to have traveled from one island to another are given later in Text Table 14, and show that this occurrence is not uncommon. Geospiza fuliginosa and Certhidea olivacea are both widespread species which might have been expected to show the same phenom- enon of a lower variability on the small isolated islands. However, as can be seen from the Main Tables, no such tendency is apparent. It may be noted that, in general, these are less variable species than G. magnirostris and G. fortis. Here selection pressure is presum- ably higher, restricting random variation. Geospiza scandens and Platyspiza crassirostris also show no such tendency, but only a few specimens of G. scandens are available from the critical islands of Abingdon and Bindloe. Platyspiza crassirostris is a stable form everywhere. Camarhynchus psittacula cannot be considered in this connection because, judging from the numbers in collections, this species must be much more numerous on Abingdon and Bindloe than it is on James or Indefatigable. Thus, the populations on the two former islands are not necessarily smaller than those on the latter. The forms of G. difficilis are also difficult to compare, as they are rather distinct from each other, and it is uncertain whether septentrionalis of Culpepper and Wenman islands should be in- cluded in this species. Furthermore, owing to the restricted habitat 88 CALIFORNIA ACADEMY OF SCIENCES [Oc. PaPErs distributions of debilirostris on James and Indefatigable, and of difficilis on Abingdon, it is difficult to determine how their total numbers compare with each other. The other species of Geospizinae do not occur on the small isolated islands. CORRELATIONS BETWEEN BILL AND W1nG MEASUREMENTS Text Table 9 gives correlation coefficients for culmen and wing, and for culmen and depth of bill, in typical examples of all of the principal species of Geospizinae. As might be expected, there is, in many species, a high correlation between the length of the culmen and the depth of the bill. This is particularly noticeable in the larger-billed species of Geospiza. Indeed, there is a positive corre- lation for culmen and depth of bill in all species investigated ex- cepting Cactospiza pallida in which, strangely, there seems to be none. Certhidea olivacea and Pinaroloxias inornata were not in- vestigated in this connection, as they have slender bills, conse- quently error in measurement of depth of bill is relatively great. Table 9 Some Correlation Coefficients in Male Geospizinae Culmen: Wing Culmen: Depth of Bill Coefficient of Coefficient of Species Island No, Correlation No, Correlation Geospiza magnirostris all 109 +,418 " " Abingdon 46 +.633 Me fortis Charles 100 +.672 171 +.796 " " Chatham 56 +.401 102 +.683 a" L Abingdon 42 +.721 n fuliginosa Chatham 61 +,248 111 +869 x nN Charles 41 +,249 4: +324 x da, debilirostris James 32 +114 39 +515 " scandens Charles 58 +.210 88 +.431 " conirostris Hood 64 +491 80 +.749 Platyspiza crassirostris S, Albemarle 34 +.477 42 +.355 Camarhynchus pauper Charles 21 +361 66 +.476 " parvulus Charles 40 +, 130 54 +. 302 Cactospiza pallida S, Albemarle 53 +.049 48 +.093 Gerthidea olivacea Indefatigable 32 +.302 Pinaroloxias inornata Cocos 78 -.023 Note: For the culmen: wing correlation, only fully black males were used in the species of Geospiza and Pinaroloxias, and only males showing some black in Platyspiza and Camarhynchus, Mitel seed) Se eee aL OhYBpIZex SOE NT For the culmen: depth correlation, all males were used, No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 89 Correlation between length of culmen and length of wing is fairly high in a number of species, noticeably in the larger forms, but is much less marked in the smaller species. In addition, there is no correlation in Cactospiza pallida, or in Pinaroloxias inornata. The high correlation between bill and wing size in Geospiza magni- rostris, G. fortis, and G. conirostris, which are all extremely vari- able species, suggests that this variability may be due partly to general size factors. A. H. Miller (1941) did not get nearly such high positive correlation of these measurements in most species of Junco which he investigated. This is also true of my investiga- tion of introduced Passer domesticus in the United States (Lack, 1940d). DIFFERENCES IN MEAN OF BILL AND WING BETWEEN SPECIES Obviously, most species differ from each other in the mean size of bill and wing measurements, and these differences, apparent in the Main Tables, do not need discussion. However, Geospiza mag- nirostris, fortis, and fuliginosa are of special interest, since these three species have identical plumages and seem to differ from each other solely in general body size and in relative size of length and depth of bill. It is apparent from the Main Tables and the histo- grams shown in Figs. 5—26 that the gap between fuliginosa and fortis on the one hand, and fortis and magnirostris on the other, is much more marked on some islands, as for example on Abingdon and Bindloe, than it is on other islands, such as Indefatigable. In fact, on the latter island it is sometimes difficult to tell whether a particular specimen belongs to magnirostris or fortis, and it is also conceivable that there is a certain amount of hybridization. Fur- thermore, on some islands, fortis is extremely variable, while the gap between the species is narrow. Hence, on Indefatigable and some other islands, the smallest fortis are actually much closer in measurement to the largest fuliginosa than they are to the largest individuals of their own species. Similarly, the largest fortis are much closer in measurement to the smallest magnirostris than they are to the smallest birds of their own species. Since these species differ solely in these mensural characters, the species problem in this group is extremely interesting. It is hardly necessary to point out that fuliginosa, fortis, and magnirostris are three distinct species which normally do not inter- breed with each other, although they may do so occasionally. Dur- 90 CALIFORNIA ACADEMY OF SCIENCES [Oc. Papers ing our stay on Chatham and Indefatigable islands, we found no evidence for interbreeding, excepting that on Indefatigable we were sometimes in doubt when one or both birds of a pair were of such size that they might be either large fortis or small mag- nirostris. This is perhaps worth emphasizing since it has been suggested that magnirostris and fuliginosa are two homozygous stocks which regularly interbreed with each other, producing fortis as a heterozygous first generation. This is definitely not true. Geospiza magnirostris and G. fortis show a slight overlapping in measurements of culmen from nostril, depth of bill, and length of wing. Conceivably there is no overlapping in the length of the whole culmen, but here also they approach each other extremely closely. A definite statement on the degree of overlapping is im- possible since with some difficult specimens one cannot be abso- lutely certain of identification as to species. Inspection of the Main Tables I, II, Text Table 16 of questionable specimens, and the histograms for the limits of size for different island forms of mag- nirostris and fortis will show the complexity of the situation. Par- ticularly difficult are the four specimens formerly called CG. bauri obtained on James and two similar ones from Bindloe. Had these specimens occurred on Indefatigable, they would almost certainly be ascribed to fortis. On Bindloe and James, however, fortis is small and these specimens approach more nearly magnirostris, especially on Bindloe, where magnirostris is small. Hence their position is impossible to assign with certainty (see Section I, p. 10). The gap between fuliginosa and fortis is much more clearly defined than that between fortis and magnirostris, and one is rarely in doubt as to whether or not a specimen should be considered as a large fuliginosa or a small fortis. The measurements of a few specimens formerly named G. harterti are given in the table of ques- tionable specimens, and, in addition to these, the populations on the extremely small islands of Daphne and Crossman seem to be intermediate in character between fuliginosa and fortis. Swarth (1931, p. 163) considered the Daphne birds to be small fortis, but although the largest specimens resemble fortis on James, the aver- age for the population is much smaller than for fortis found any- where else. Some of them come very close to fuliginosa. The Cross- man population averages smaller than that of Daphne. The larger specimens resemble the Daphne birds, while the smaller birds seem No. 21] LACK: VARIATION IN GALAPAGOS FINCHES No. 25 . 24 Fig 0 23 22 21\- 20 19 |- 18 17 16 |- 14 12 aby 10 : 9 Ps eC] 8 on Novorororenens SND i senenrnnrenrenenennrine e] 5 eeeecseceetetatstatstatatetatstate! evetatatetaterereee eee eee, 4 roretereretereterereeeeetee, SRR KN] orerererererererere ee eee ere, 3 BSR ererereteereereeene ee 2 Recetecotecetecones SRR RRR AON OOOO OOO Meee, z DKK I KORO ta) S509 15 8.5 9.5 10.5 0.5 12.5 13.5 14,5 15.5 16 AL) 17.5 mm, No. 14 |= 13 |- 12)/- uj- 10 |- 8j- q\- 6l- 5\- 7 ‘- 4\- Seses SLD sosseennies & OOO aie stelctetetatatatetetctctetatetetets: Sececonececonecececectcerecececeeet 2 shetetatatetstatatetenctatatetetees = Nececeneceresectetce tee One Paramore, ROR KR RRR KI 1|- BOO ees ole LLL SSOCOCOCCCCCCCCCCCOCS 15 8.5 D5 Oe) S| ee 225 13.5| 3425 1525 16.5) 1765) mm, No. 17\- & 16 |- Fig. 7 15 j- 14/- 13 |- K LX KX b> Roe % Keg neato Bib 9554; 105 5 125 oe No. pul = Geospiza fuliginosa 10 2 : , 8 RQ _ ceospiza fortis 1 6 a Peay Geos>iza magnirostris 3 4 Geospiza fuliginosa x Geospiza fortis ? 0 ad 95 15 U5 ma "Geospiza bauri" Variation in length of culmen of males. Fig. 5, Abingdon; Fig. 6, Bindloe; Fig. 7, James; Fig. 8, Daphne. “ . » a : << “ (S) é CALIFORNIA ACADEMY OF SCIENCES Geospiza fuliginosa ortis Geospiza f ospiza irostris NS ase = ar La | Fig. 9 ales m in length of culmen of ASL rey Ke) Fig. 10, Crossman; Fig. 11, Chatham; Fig. 1 lation Var Y Zilli Vv 4 i 4 No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 93 Fig. 13 Crnvanau oo BRO RG RS EE 8 Ms Fig. 14 OH min ova ove BES GIB! Senos 65 - 2 . « js 21.5 m. Or rn ava coo 5 EIS GG Be STE Fig. 16 Geos=iza fulisinosea x Geosnize fortis 7 oOrnw Mn anna on | “Geospize beuri* ——————— 5 w5 jUS LS =, Variation in depth of bill of males. Fig. 13, Abingdon; Fig. 14, Bindloe; Fig. 15, James; Fig. 16, Daphne; Fig. 17, Crossman. PPR SSYPRUEGRABSSL EG ES EAE OPNnUrUarovbERGSGaSeS oe nGeGaSSSSeRnuyUussssses Bia | id FL a Ga | orev SF RERGESESSERARGAS CALIFORNIA ACADEMY OF SCIENCES °. 5 . Ie Variati patra ae of males. Fig. 18, Charles; g. 19, indemGeenle: [Oc. PAPERS Pe, No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 95 me Fig. 20 h a 1 Pee 57 58 59 60 61 62 63 64 65 66 67 6B 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 g5 86 87 88m. Fig. 21 opnuaule iS 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86m. 15 CORNY FU a3/0 87 88 89 m, 63 64 65 66 67 68 m, 64 65 66 67 68 69 70 71 m. No. 25 |= 5 Geospiza fuliginosa 24 |= i 23 |- 22)- Geospiza fortis 21|- 20 |= “A - Geospiza magnirostris 1 K 17/- 16 Geospiza fuliginosa x Geospiza fortis 15 |- 14 "Geospiza bauri" ae SSS 2 apne nado . 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87m, Variation in wing length of black males. Fig. 20, Abingdon; Fig. 21, Bindloe; Fig. 22, James; Fig. 23, Daphne; Fig. 24, Crossman; Fig. 25, Charles; Fig. 26, Indefatigable. 96 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS quite typical of fuliginosa. The most probable explanation is that these forms have originated from a cross between fuliginosa and fortis. One might expect hybridization on such tiny islands where wandering birds might find few or none of their own species avail- able. The histograms, Text Figs. 5-26, also show that on different islands the specific limits occur in a different place. For instance, specimens of male Geospiza with a culmen of 14 mm. are magni- rostris on Bindloe or Jervis, but fortis on Chatham or Charles, and on Indefatigable might be either. Also, males with culmen 10 mm. are fuliginosa on Chatham but fortis on Charles. As regards depth of bill, fuliginosa and fortis show no overlapping, and for magni- rostris and fortis there is much less overlapping in this measure- ment than for that of culmen from nostril. Males with a depth of bill of 16 mm. on Jervis and 16.5 mm. on Indefatigable are magni- rostris, whereas on Charles they are fortis. In wing length, black males with a wing of 64-66 mm. on Abingdon are fortis, but on all the southern islands they are fuliginosa. Black males with a wing of 79 or 80 mm. on Abingdon, Bindloe, or Jervis are magni- rostris, but on Charles are fortis. In all these instances, other char- acters make identification of the species certain, but in other cases such as G. bauri, identification is much less positive. Geospiza fuliginosa, fortis, and magnirostris differ from each other not only in actual size of wing and bill, but in relative length of bill as compared with length of wing, and in relative depth of bill as compared with its length. The larger species have relatively longer and deeper bills. This suggested that the differences in pro- portions of these parts in these three species might be due simply to allometric ratios between these parts and general body size (see Huxley, 1932). In this instance, one would expect similar allo- metric relations to hold within each species. To test this possibility Text Tables 10 and 11 were prepared. From these tables it will be seen that within each species there is, with increasing wing length, no increase in the proportion of length of bill to length of wing; likewise, with increasing length of bill, there is no increase in the ratio of the depth of bill to its length. In each species, regardless of the size of the individual, the ratio between culmen and wing and between depth of bill and culmen is approximately constant. There is no tendency whatever toward an allometric relation of these parts. No. 21] LACK: VARIATION IN GALAPAGOS FINCHES Table 10 Ratio of Culmen to Wing in Black Male Geospiza fuliginosa, fortis and magnirostris Culmen No, Wing Mean Range o Ratio Geospiza fuliginosa (Chatham, Charles, N, and S, Albemarle) 29 60 - 62 8.58 161-969 2427 SHEE) (61.8) 52 63 8,42 7.6-9.8 2534 134 68 64 8.52 7.4-10,2 2497 133 44 65 8.59 11-909 2525 «L352 26 66 8.77 7.9-10,1 2567 133 6 67 - 70 8.90 8.3-9.5 587 oe (67.7) Geospiza fortis (Chatham, Charles, N, and S, Albemarle) 9 66 - 68 10.89 10,1-11,8 2515 162 (67.4) 9 69 11,38 10,4-12,3 2645 165 27 70 11,35 10,7-12,1 0435 162 44 A: 11,58 10,1-13.9 ~655 163 38 72 11.59 10,7-13,2 3644 161 28 DB 11.74 10.3-13.5 726 161 17 74 12.33 11,.3-13.4 518 167 26 15 12,40 10,8-14,0 875 2165 12 76 12,78 12,1-13.6 483 168 18 717 - 80 12,88 11.9-14.1 614 165 (78.0) Geospiza magnirostris (all) 9 717 - 80 15,20 14,0-16,5 2770 193 (78.9) 13 re ijk 15.36 14,2-16.8 .670 190 13 82 15.41 14,0-16,0 2559 188 17 83 15.72 14,5-17,3 -680 189 13 84 16,05 14e7=1767 | 12055 191 13 " 85 16,41 15.9-17.2 355 189 14 86 15.96 15.1-17.2 -673 186 10 87 15,82 14,2-16.8 .798 182 1 88 - 90 16.39 15,1-17.4 113 185 (88.6) oh 90 - 94 18,43 18.0-18.9 2451 .202* (91.3) Geospiza fortis x fuliginosa? (Daphne) 12 65 - 67 10,33 9,2-11.2 665 156 (66.08) l 68 - 70 10,53 9.9-11.3 476 ole) (68.82) Geospiza fortis x fuliginosa? (Crossman) 8 63 - 67 9.36 8,0-11,2 1,049 143 (65.60) * Derwin's black males, 98 CALIFORNIA ACADEMY OF SCIENCES Table 11 [Oc. PAPERS Ratio of Depth of Bill to Culmen in Male Geospiza fuliginosa, fortis and magnirostris Culmen Depth of Bill No. Mean Range Mean Range o Geospiza fuliginosa (Chatham, Charles) 13 1.52 1.2-7.7 7.70 1.1-8.3 390 28 8.03 7.8-8.2 7.83 7.2-8.7 398 51 8.54 8.3-8.7 8.09 1.2-8.9 410 65 9.00 8.8-9.2 8.21 7.4-9.1 355 16 9.45 Jo9=F9al 8,40 7.4-9.1 489 9 9.93 9,8-10,2 8,31 8.0-8.6 215 Geospiza fortis (Chatham, Charles) 5 10,10 9,8-10.2 10.96 10,1-11.9 856 12 10,48 10,3-10.7 11.47 10,7-12.1 446 54 11,04 10,8-11.2 11,89 10,2-13.2 2567 55 11.44 11.3-0.7 12.25 10,8-13.6 612 72 12,02 11,8-12,2 12,80 11,1-15.3 .877 25 12,48 12,3-12.7 13.43 12,0-14.8 -750 33 12.93 12,8-13.2 14,44 12,1-16.0 1,032 10 13.48 13.3-13.7 14,32 13,1-15.4 787 6 14,03 13.8-14,2 14,85 13,6-16.0 1,035 Geospiza magnirostris (all) 11 13.97 13.8-14,2 17.34 15,8-19,4 9717 13 14,50 14,3-14,7 18,04 16.3-19.4 978 43 15.01 14,8-15.2 18.92 16,8-21.6 1.066 40 15.47 15.3-15.7 19,56 18,0-21.4 868 55 15.96 15,.8-16.2 20.23 18,2-22.3 ~963 32 16,43 16.3-16.7 20.85 18.9-22.5 831 19 16.96 16,8-17.2 20.94 19,5-22.9 2917 7 17,43 17.3-17.7 21,24 19.9=23.2 1.139 3* 18,43 18,0-18.9 23.67 23,6-23.7 058 Geospiza fortis x fuliginosa? (Daphne ) 10 10.04 9,8-10,2 10.07 8.6-10.6 383 6 10,60 10,3-10.7 10.93 9.7-11.9 0155 9 11.09 10,8-11.3 10.97 10,4-11.5 378 Geospiza fortis x fuliginosa? (Crossman) ll 9.43 8.4-11.2 9.55 8,310.4 802 * Darwin's black males, Ratio 1,02 1,01 No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 99 The tables therefore show that the ratio of culmen to wing, and of depth of bill to culmen, provide a criterion for distinguishing the three species. However, it must be noted that among the indi- viduals of each species, the variations in these ratios are large. For instance, on Charles the ratio of depth of bill to culmen in fortis varied between .92 and 1.32, that is from typical values for fuligi- nosa@ up to typical values for magnirostris. Similar considerations apply to the other species, as can be seen by inspecting the range in culmen for each particular wing length, and the range in depth of bill for each length of culmen, given in Text Tables 10 and 11. Hence, these ratios do not form a satisfactory method for identify- ing any particular specimen, but are only valuable when a series can be compared. I have also included in these tables the same ratios for the un- usual Daphne and Crossman populations and for the single male specimen of Geospiza bauri. It will be noted that the ratios for the Daphne and Crossman populations tend to be intermediate between those for fuliginosa and fortis. This provides further evidence for the hybrid origin of these forms. The specimen of G. bauri is inter- mediate between fortis and magnirostris and might be either. Camarhynchus species: in the species of Camarhynchus there is a problem similar to that presented by the three species of Geospiza described above. On most islands, there is a larger and a smaller species, which differ primarily in the mensural characters of bill and wing. Here also the larger species has a longer bill in propor- tion to wing length, and a deeper bill in proportion to its length. However, there are also small plumage differences. I have not given details, since there are too few specimens available. On Charles, a particularly interesting situation is presented, since on this island there are three species of Camarhynchus. The gap between the smallest, parvulus, and the middle form, pauper, is fairly well defined, although there are two specimens which can- not be assigned to either and might possibly be hybrids. A few other specimens show a slight overlapping either in wing length, or length of culmen from nostril, or depth of bill, but never in more than one of these characters at the same time. All can be safely assigned to one or the other species. Camarhynchus pauper and the larger psittacula show a wide overlapping in length of wing and length of culmen, but apparently there is no overlapping in depth of bill, although the two species come extremely close to each 100 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS other. However, few specimens of psittacula are available from Charles, and further collecting is needed before an extended dis- cussion would be justified. There seem to be only a few other instances in birds where closely related species differ markedly from each other in size of bill but in little else. This occurs in the two species of Nesospiza, the endemic finch genus of the Tristan da Cunha group (Lowe, 1923, pp. 519-523), also in the West Indian finch Oryzoborus and in the African weaver Pyrenestes (Chapin, 1924). Pyrenestes is particularly complicated since in each of two species two sizes of bill occur, and in a third species there are three sizes of bill. The relations between these forms are not properly understood. Fried- mann (1934) describes another possible instance in the hawk genus Chondrohierax. DIFFERENCES IN VARIABILITY BETWEEN SPECIES As may be noted from the Main Tables, the various species differ in variability, in some cases markedly, whether this is meas- ured by the standard deviation or by the coefficient of variation. Some typical examples are set out in Text Tables 12 and 13. Day and Fisher (1937, pp. 337-340) have recently criticized the use of the coefficient of variation, since there is no real reason for supposing that the variability should change in direct propor- tion to the mean. This is a valid criticism. However, as may be seen from the tables, the larger Geospizinae tend to be more vari- able than the smaller species. This suggests that there is some rela- tion between size and variability as measured by the standard de- viation, and that the latter is not an altogether fair means for com- parison. Hence, in Text Tables 12 and 13, I have given both standard deviation and coefficient of variation, although I agree that the relation between the degree of variability and the mean size is probably not a direct one. The tables suggest that dividing directly by the mean rather overcompensates for the increase in standard deviation due to increase in mean. A number of factors affect variability, and it seemed worth checking to see if the rule that the most abundant forms were the most variable, and the least common forms were the least variable, applied when species were compared. In Text Table 12, the species on Indefatigable have been listed in their order of abundance. The estimates of abundance are fairly reliable since they are based oe No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 101 Table 12 Relation between Variability and Abundance in Male Geospizinae Indefatigable Island Culmen Depth of Rill Wing Species Subjective estimate of abundance, Coefficient of Coefficient of Coefficient of (in order of abundance) expressed as percentage o Variability o Variability o Variability Certhidea olivacea 47 338 4,45 1.31 2,42 Geospiza fuliginosa 16 376 4.47 -380 4,65 acy] 2,15. fi scandens 12 - 767 5.13 433 4.44 1,11 1,53 n fortis 10 809 6.77 1,321 10.30 2.79 3.84 Camarhynchus parvulus 5 443 5.9% 324 4.43 0.63 0.99 Platyspiza crassirostris 3 282 2.70 338 2.75 1,20 1.39 Cactospiza pallida 3 376 3.11 2209 2,30 1.39 1.91 Geospiza magnirostris 2 2799 bees 6 L753 9.17 1.50 1,81 Camarhynchus psittacula i 2597 6.19 Geospiza d, debilirostris 1 320 Sey) cept 3.82 1.27 1.84 Tower Island Certhidea olivacea 43 - 206 2,52 0.95 1.76 Geospiza difficilis 43 2377 4,00 328 4.15 eri 2,08 " magnirostris 9 Ee) CeEy | abeilet) Saar 1,42 1,64 i conirostris 3 - 762 5.5L ~958 7.38 1.98 2,60 Table 13 Relation between Variability and Specialization in Male Geospizinae Culmen Depth of Bill Wing Degree of specialization of: Coefficient of Coefficient of Coefficient of Species Diet Bill Island Selected o Variability o Variability o Variability Geospiza fuliginosa extremely generalized very generalized Indefatigable «376 4,47 380 4,65 1.37 2.15 ® fortis generalized obviously for 2 809 6.77 1,321 10,30 2.79 3.84 grain-eating mw 4, difficilis ve generalized Tower 377 4,00 328 4,15 1,31 2.08 " ¢. conirostris x a ? Hood +938 6.08 1.104 6.91 2.37 2.97 . magnirostris . specialized ? Tower 3599 3.63 1,127 5.31 1,42 1,64 Camarhynchus parvulus ta) generslized Charles «297 4,05 284 3.78 1.52 2.37 . uper . . . 404 4,48 429 4.86 1.54 2.19 Geospiza scandens partly specialized specialized Indefatigable «767 5.13 433 4,44 1,11 1.53 Platyspiza crassirostris specialized generalized i‘ 282 2.70 338 2.75 1,20 1.39 Catospiza pallida - moderately S. Albemarle 432 3.84 400 4.93 1,49 2,08 specialized Certhides olivacea specialized Indefatigable -338 4.45 1,31 2,42 Pinaroloxias inornata n specialized Cocos 382 3.66 +250 4,05 1.25 1,84 Note: For comparison, male Passer d, domesticus, Berkeley, California: culmen: o .401, coefficient of variability 4,28 (91 measured); depth of bill: o .293, coefficient of variability 3,35 (88 measured); wing: o 1,46, coefficient of variability 1.90 (77 measured), on serial counts and general observations made during the expe- dition’s visit to the island. In the second part of the table, the same procedure has been adopted for Tower Island, which was also vis- ited by the expedition. It is seen that there is no obvious relation between variability and abundance on either island, and the same applies to Chatham, for which I did not make a table. This is true whether one uses standard deviation or coefficient of variation. 102 CALIFORNIA ACADEMY OF SCIENCES [Oc. Papers Clearly abundance is not the most important factor controlling variability when different species are compared. Another possibility is that the more specialized species, with reference to feeding habits or structure or both, are less variable than the generalized species. This is tested in Text Table 13. The degree of specialization of the bill has to be estimated subjectively, but I have based my estimates on comparisons with mainland forms. The table shows that there is less correlation between vari- ability and specialization than might perhaps have been expected. However, the four most specialized species are among the least variable, whether this is measured by standard deviation or by coefficient of variation. The most variable forms all have general- ized feeding habits. It has often been said that the Geospizinae are made up of ex- ceptionally variable species. From this study, however, it is clear that most species of Geospizinae are not exceptionally variable when compared with those mainland birds which have so far been investigated. Comparison may be made with Passer domesticus (Lack, 1940d), with Junco (Miller, 1941), and with Lanius (Mil- ler, 1931). There are, however, a few forms which are extremely variable. These are: (1) Geospiza magnirostris and G. fortis on most of the larger islands and on the small islands near large islands; (2) G. fuliginosa on Chatham; (3) the species of Geospiza on Daphne and Crossman; (4) G. conirostris. Apart from their high standard deviations, a further indication of exceptional vari- ability is that in the past several systematists named two forms on the same island from material which later proved to belong to one form. This occurred with G. fortis on Charles, Chatham, and other islands, G. fuliginosa on Chatham, and G. c. conirostris on Hood. As Swarth demonstrated, and my results fully confirm, each of these is a single but highly variable form, whose mensural char- acters lie on approximately normal curves with no trace of bi- modality, although the bill differences between extreme individuals of one form are comparable with specific differences in some main- land groups of birds. In addition to the dimensions, the shape of the bill is also extremely variable, and this applies particularly to the three forms of GC. conirostris. As already suggested, the high variability of G. magnirostris and G. fortis on most of the larger islands is probably a result of the relatively large size of the populations on these islands as com- No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 103 pared with those of the smaller islands. At the same time, associ- ated with generalized feeding habits, selection pressure as affecting the bill is probably not intense. There is the further possibility that magnirostris and fortis occasionally interbreed, since these two species possibly intergrade on Indefatigable and Albemarle. While such hybridization possibly occurs, it cannot be the cause of the high variabilities in question since: (a) Geospiza magnirostris is as variable on James and Jervis as it is on Indefatigable, but there are no grounds for assuming that it interbreeds with fortis on the two former islands; first, be- cause the gap between the two species is here clear-cut, and second, because fortis, which is rather uncommon on Indefatigable, is stable, and not especially variable, on James. (6) Geospiza fortis is slightly more variable on Charles than on Indefatigable or Albemarle, but one would expect hybridization with magnirostris to be much rarer on Charles than on Indefatigable and Albemarle, since magnirostris is not resident on Charles. One specimen of magnirostris is listed from Charles, but this must be a relatively rare occurrence. (c) Geospiza fortis from Charles is also exceptionally variable on the small side of the mean, and there is no reason for assuming that it interbreeds with fuliginosa. Hence, hybridization, if it oc- curs at all, is not the main cause of the high variability of magni- rostris and fortis on the larger islands. Geospiza fuliginosa is relatively stable excepting for the high variability in culmen (but not depth of bill) on Chatham. Possibly a smaller northerly and larger southerly form of fuliginosa have met and merged on Chatham, but this has not been proved. The forms of Geospiza on Daphne and Crossman, which are intermediate in characters between fortis and fuliginosa, are highly variable. The standard deviations are high when compared with those found in fortis, which has larger means, and are considerably higher than those found in fuliginosa. High variability is addi- tional support for the postulated hybrid origin of these forms. If they are of hybrid origin, however, some gene elimination has pre- sumably occurred later, since in neither form have specimens re- sembling larger fortis been collected. Another difficulty is that a few typical fuliginosa have been collected on Daphne. Further collecting might show that these grade into the postulated hybrid form, but there is a gap at present. 104 CALIFORNIA ACADEMY OF SCIENCES {Oc. PAPERS Geospiza conirostris presents special difficulties, since it occurs only on three small isolated islands. In such places, owing to the relatively rapid gene elimination, one would expect low, not high variability. In addition, it is the only species of the Geospizinae showing a markedly discontinuous distribution, since it occurs only on three outlying islands: Culpepper in the extreme northwest, Tower in the extreme northeast, and Hood in the extreme southeast of the group. This suggests that it may be an ancient species which has become extinct on the main islands in competition with later evolved forms. This does not account, however, for its high varia- bility. Another possibility is that it is of hybrid origin. Evidence summarized in Section I fully supports Swarth’s con- clusion that Geospiza conirostris is closely related to G. scandens. If it is of hybrid origin, then presumably it is between scandens and either magnirostris or fortis, since conirostris differs from scandens primarily in the depth of its bill. It may be noted that neither scan- dens nor fortis is resident on any of the islands where conirostris breeds; but magnirostris occurs with conirostris on Tower. Geo- spiza conirostris propinqua from Tower seems more like a scan- dens X fortis than a scandens X magnirostris cross, and further- more shows no intergradation with magnirostris on Tower. Hence, if it is a hybrid, it is presumably with fortis. Also, four specimens from Charles and Indefatigable, intermediate in character between fortis and scandens and possibly hybrids (“G. brevirostris’’), look somewhat like G. c. propinqua. As opposed to a hybrid origin for the Tower form, it should be noted that on Bindloe, the island nearest to Tower, the form of scandens has a deeper bill than else- where, and comes fairly close in appearance to small G. c. propin- qua. Hence, one might have been evolved from the other without any hybridization. In this case, either conirostris is an ancient form which later gave rise to the more specialized scandens through the Bindloe form, or alternatively scandens wandered to Tower and Hood where its specialized bill was of little value, and it reverted to a more generalized form. Geospiza conirostris conirostris of Hood throws little additional light on the problem. Neither magnirostris nor fortis is resident on the island. Geospiza magnirostris is also absent from the neigh- boring islands of Chatham and Charles, hence is less likely to have reached Hood than fortis, which is abundant on both. Actually, a number of fortis have been collected on Hood, and some of these No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 105 were indeed difficult to distinguish in bill from smaller G. c. coni- rostris. Thus if the latter form is of hybrid origin, it is presumably the result of a cross between scandens and fortis. In depth of bill, the larger specimens of G. c. conirostris far exceed the largest fortis, but this is not necessarily in opposition to a G. scandens X G. fortis origin, since one does not know what characters might appear when genes of the two species are combined. In addition, there has pre- sumably been evolution subsequent to the postulated hybrid origin. Geospiza conirostris darwini of Culpepper Island is the most variable of the three forms of conirostris, although restricted to one extremely small, isolated island. Smaller specimens look much like G. c. propinqua of Tower Island; larger specimens look more like magnirostris. Indeed Swarth classified them as such, although the shape of the bill places them as conirostris. On Wenman, the only neighboring island, a typical magnirostris occurs. It is sur- prising that these two islets, which otherwise have such similar land birds, should differ so markedly in this respect, excepting on the supposition that G. conirostris darwini is of hybrid origin be- tween magnirostris and some other species which has accidentally wandered to Culpepper. The most likely possibility is that it is a hybrid with G. c. propinqua, which it so closely resembles. In summary, G. c. darwini appears to be of hybrid origin be- tween magnirosiris and G. c. propinqua. All evidence supports this view. The Hood and Tower forms of G. conirostris may be of hybrid origin; if so, it is probably between scandens and fortis. There must have been, however, subsequent gene eliminations, since neither form produces typical scandens nor typical fortis. Alterna- tively, conirostris may be an ancient form from which scandens later evolved, the Bindloe form showing an intermediate stage. However, this does not explain the high variability of G. coniros- tris. Also, conirostris may have evolved from scandens, independ- ently on Hood and Tower islands, by acquiring more generalized feeding habits and bill, and increasing in variability. INTER-ISLAND WANDERING In Text Table 14 all known instances are summarized. Detec- tion of inter-island wandering is possible only when a specimen is collected at sea or on an island where that species or island race is not resident. Most species occur on nearly all islands, and most species are not divided into definitely recognizable island races. 106 CALIFORNIA ACADEMY OF SCIENCES [Oc. Pavers Table 14 Geospizinae which have Strayed from One Island to Another Museum Number Species * Sex Collected on: Presumably came from: U,S,N.M, 115905 Geospiza magnirostris Pd Charles Indefatigable or S, Albemarle C.A.S. 5465 " fortis S@ Wenman Albemarle (‘Too large for the form on Abingdon, Bindloe or James, ) c.M, 123796 a is BS ‘Tower Abingdon, Bindloe or James C.A.S, various C) “ 8 J SCO, 709 Hood and Gardner Charles or Chatham Rothschild Coll, 517705 a scandens soc James Bindloe C.A.S. 7191 4 " soc At sea 20 miles Charles or Indefatigable south of Indefatig- able Rothschild Coll, 517049 m ¢, conirostris 9 . Gardner near Charles Hood Rothschild Coll, 517986 Camarhynchus psittacula So Chatham Indefatigable or Barrington Rothschild Coll. 517985 W ul Q Chatham Indefatigable or Barrington C.A.S, 1904 i parvulus Q Wenman C.A.S, 71943 ia i 9 Wenman C.A.S. 7629 Cactospiza pallida toy Charles Indefatigable *B-black plumage; P-partly black plumage; S-streaked plumage Note: All the above specimens, except those of Geospiza scandens, were collected on islands where so much other collecting has been done that the species can safely be said not to be resident there, Of the Geospiza scandens, one was actually taken at sea and the unusual specimen from James was clearly not of the type resident on James, but was indistinguishable from Bindloe specimens, Some other specimens were perhaps stragglers, but the evidence is not conclusive: four Geospiza fuliginosa from Wenman, one Geospiza d, debilirostris from S, Albemarle, several Camarhynchus affinis (?) from James, Jervis and Seymour, and possibly two Camarhynchus parvulus from Abingdon, From their measurements, three other Geospiza scandens may also be wanderers, as each is well outside the normal limits to the size of bill on the island on which they occurred, Details are as follows: Stanford Coll, 4650, P °, from Seymour, measured 12,4, 10.2, 68, so perhaps came from James, Rothschild Coll, 517866, P GO, from Indefatigable, measured 13,0, 8.5, 68, so perhaps came from James, C.A,S. 7376, 9, from Jervis, measured 13,1, 11.2, 69, so possibly came from Abingdon or Bindloe, Gifford (1919, pp, 230-231) also gives four records of Geospiza fortis seen at sea at varying distances from the land, Hence, most inter-island wandering will escape detection. Despite this, there are a number of records showing that inter-island wan- dering must be not uncommon, an important conclusion in relation to various discussions in Sections V and VI. — SECTION VI. GENERAL EVOLUTIONARY PROBLEMS In the preceding five sections, the main facts concerning varia- tion among the Galapagos finches have been described. In this final section, some of the more theoretical aspects will be consid- ered. To summarize very briefly, the Geospizinae are almost unique among endemic land birds of oceanic islands in that there is more than one species on each island. The main genera show an adap- tive radiation. The species within the genera differ primarily in size and shape of bill, and some are extremely similar. Some show minor plumage differences, but the main plumage variation is a tendency to lose the male secondary sexual plumage, and, in this, closely related species resemble each other. A few forms are quite variable in size, but as a rule most of them are as stable as typical mainland birds. Most species are not segregated either by habitat or breeding season. Previous VIEws Since the time of Darwin, the Geospizinae have been the subject of considerable theoretical discussion. In recent years, several views have been put forward to account for their remarkable type of speciation. Lowe (1936, p. 320) follows Lotsy’s views on cer- tain plant groups in stating that the Geospizinae are “a swarm of hybridization segregates.” Rensch (1933, pp. 37-38) considers that in almost every instance the first step in the formation of a species in birds is through geographical isolation leading to the differentiation of a subspecies. With continued isolation, the segre- gated subspecies becomes so different that, if it eventually meets again the form from which it diverged, it no longer interbreeds. Rensch was evidently much puzzled to understand how the Geo- spizinae could fit into such a scheme, and although he does suggest how the species might have evolved through geographically iso- lated forms, he classifies them under “Nicht-geographische Typen der Artbildungen.” Stresemann (1936) is the only investigator to suggest that the evolution of the Geospizinae has not differed funda- mentally from that of other avian species, namely that they have evolved through geographically isolated forms, which later met again but kept segregated. { 107 ] 108 CALIFORNIA ACADEMY OF SCIENCES [Oc. Papers HisTORY OF THE ISLANDS Before proceeding further, the geological history of the Galapa- gos must be considered, and in particular the problem of land bridges. Darwin, Wallace, and other early workers concluded that the islands were oceanic; that is, that their fauna resulted from chance colonization over the sea, and did not arrive via a land bridge. Baur (1891) was the first to suggest that the Galapagos were once connected by land with America. (See Scharff, 1912, pp. 295-335, for a summary of this view.) Recently Swarth (1934) has summarized the arguments in favor of an oceanic origin. All the evidence is in favor of the supposition that the Galapagos are oceanic so far as birds are concerned. Otherwise, it is almost inconceivable that so few species of land birds should occur on the islands, especially in view of the many favorable habitats that exist there, varying from coastal thorn and Opuntia to humid forest and fern and grassy uplands. Only six passerine forms and one cuckoo are found, a number considerably smaller than that which has reached many of the Polynesian Islands situated a comparable dis- tance from the mainland. The latter lie in the trade wind belt, whereas the Galapagos are in an area of calms. This corroborates the oceanic origin of the birds, since it makes the Galapagos rela- tively less open to colonization by aerial forms. Furthermore, when compared with each other, the six passerine forms and the cuckoo show a very marked degree of difference from mainland forms. This seems obviously correlated in part with a different time of arrival on the islands, another argument in favor of an oceanic origin. Examples of the successive stages in differen- tiation among the land birds are: (1) the cuckoo (Coccyzus mela- coryphus) seems identical with a mainland form, and therefore presumably arrived recently; (2) the warbler (Dendroica petechia aureola), which also occurs on Cocos Island, and the martin (Progne m. modesta) are subspecies of mainland species; (3) the flycatcher (Myiarchus magnirostris) is a species of a mainland genus and is not obviously related to any mainland species. Hell- mayr (1927, p. 187) placed it in a genus by itself, Eribates, but this hardly seems justified. In appearance, call notes, and field habits it certainly comes very close to the ash-throated flycatcher (Myiarchus cinerascens) which I observed in California. (See also No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 109 Swarth, 1931, pp. 85-86.) (4) The Galapagos vermilion fly- catcher (Pyrocephalus) is less differentiated than Myiarchus magnirostris in that it is closely related to a mainland species, P. rubinus. In another way, however, evolution has proceeded further, in that three island forms have been separated on the Ga- lapagos. Hellmayr rightly considered, in my opinion, the Galapa- gos forms to be subspecies of Pyrocephalus rubinus. Swarth con- sidered them separate species, but they are obviously geographical forms of P. rwbinus and are not very different. (5) The Galapagos mockingbirds (Nesomimus) have been placed in a distinct genus. They are obviously related to the mainland genus Mimus, although there is no particular species with which they can be linked. Neso- mimus has developed into distinct forms on all the main Galapagos Islands, and in some cases these are so distinct as to merit specific rank. (6) Finally, the Geospizinae arrived so long ago that they have had time to develop into widely divergent genera and species, greatly different from the ancestral (? West Indian) stock from which they originated. Mayr (1940, p. 267) gives a similar table for the various stages in differentiation shown by the birds of Biak and Rennell islands (both oceanic islands). The main argument in favor of a former land bridge from the Galapagos to the mainland is the presence of several terrestrial reptiles. Representatives of each of these could have been carried to the islands during a hurricane or on driftwood. If they did cross to the islands on a land bridge, then it was in very early (? Meso- zoic) times, and this land bridge disappeared before the period when birds could make use of it. Except where they have been exterminated by man, all the main Galapagos reptiles are found on all the main islands, and so are rodents of the genera Oryzomys and Nesoryzomys. This strongly suggests that the main islands, possibly omitting Culpepper and Wenman, were connected with each other (although not necessarily with the mainland) at some time in the past, since otherwise it is almost inconceivable that each island should have been separately colonized. On Indefatigable, two species of Nesoryzomys (cf. Os- good, 1929, and Orr, 1938) and two species of the snake, Dromi- cus (Van Denburgh, 19126), are found, hence possibly there have been two separate occasions when Indefatigable was connected with other islands. This, however, does not apply to islands other than Indefatigable, since, on all the others that are inhabited by rodents 110 CALIFORNIA ACADEMY OF SCIENCES [Oc. Papers of either of the genera Oryzomys or Nesoryzomys, only one species per island occurs. The same is true of the snake, Dromicus. The gecko, Phyllodactylus (Van Denburgh, 1912a), and the lizard, Tropidurus (Van Denburgh and Slevin, 1913), are represented by only one species on each island, including Indefatigable, so there may be some other explanation for the two Indefatigable Nesoryzo- mys and Dromicus. The tortoise, Testudo, is represented by a sepa- rate form on each main island, and by five forms on Albemarle (Van Denburgh, 1914). There are five main mountains on Albe- marle; each mountain is populated by one of the five forms of tortoises present on the island. This may indicate fairly recent ele- vation of the land, uniting these mountains which were formerly islands. Dall and Oschner (1928) report both Pleistocene and Tertiary (probably Pliocene) fossil mollusca on raised beaches on various Galapagos Islands. From the evidence above, it is clear that insular conditions in the Galapagos have varied considerably in the past. Shepard and Beard (1938) have provided evidence for large submarine canyons off the coast of California and also in many other parts of the world, which are a further indication of big changes in sea level. The depths involved are more than sufficient to unite all of the Galapagos Islands, but there is no need to sup- pose that the Galapagos have been united with the mainland during or since the time at which the present avifauna colonized them. Since birds can fly across the sea, the Geospizinae are not of as much interest as the mammals and reptiles in studying the land connections between the islands. However, it is well to remember, in considering the evolution of the Geospizinae, that conditions of isolation in the past were probably very different from what they are today. COMPARISON OF THE DIFFERENT GALAPAGOS ISLANDS Conditions of isolation are clearly different on the various is- lands of the Galapagos. In Text Table 15, the size of each island and the degree of isolation are indicated. (For details of islands see Frontispiece map, and Text Table 1 in Introduction.) Also included in this table are the total number of resident species of Geospizinae and the total number of endemic forms on each island (for details see Text Table 2 in Section I). For the endemic forms, two divisions are made: forms found only on a particular island, and forms found on that island and the neighboring one. This is ae No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 111 necessary since in two instances a pair of islands are near each other but separated from the rest. These pairs are considered together at the end of the table. Table 15 Distribution of Geospizinae on Galapagos Islands Island Position Size Number of Number of Number of forms resident endemic found on only species forms one other island Culpepper Extremely isolated Extremely small 3 1 2 Wenman Extremely isolated Extremely small 3 10) 2 Tower Isolated Small 4 2 1 Hood Isolated Moderately small 3 2 - Abingdon Moderately isolated Moderate 8 (9)* - 3 et Bindloe Moderately isolated Moderate 1 - 24** Chatham Moderately isolated Large 7 3% = Charles Moderately isolated Moderately large 8 2 - Barrington Near central Moderately small 7 1 - James Central Large 10 & ab Indefatigable Central Large 10 - 1 Albemarle + Narborough Central west Very large 10 2 - Jervis Central Small 9 - - Duncan Central Small 9 - - Taken in pairs Culpepper + Wenman Extremely isolated Extremely small 4 3 - Abingdon + Bindloe Isolated Moderate 8 (9)* ok +* 1 a Camarhynchus parvulus is doubtfully resident on Abingdon. ** The partial differentiation of Geospiza scandens has been reckoned as +. Note: This table is based on the data in Table 2, Section I, The table shows that, as might be expected, the large central islands of Indefatigable, James, and Albemarle have a greater number of resident species than any others, while the small central islands of Jervis and Duncan have nearly as many. Furthermore, the percentage of endemic forms is smaller on these islands than on any others. At the other extreme, Culpepper, Wenman, Tower, and Hood show that the smallest and most isolated islands have the smallest number of resident species, but, at the same time, the largest percentage of endemic forms. Thus two out of the three 112 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS Table 16 Measurements of Spécimens of Questionable Identity Of or between Geospiza magnirostris and Geospiza fortis Museum Number Island *Sex Culmen Length Bill Depth Wing Length Here assigned to: B.M, 15.4.2.18 Bindloe Q 13.8 16.4 14 magnirostris Rothschild Coll, 516961 G fe) 13.4 15.4 13 unassigned-like Geospiza bauri BM, 85,12.14,293 m ? 13.0 15.7 14 " ) " Rothschild Coll, 17124 James Bo 13.8 16,1 81 " " " " " " bin AEE) Ud Q 13.7 15.3 78 LW Y - " " " 517123" 2 14,0 15.3 15 i! ud . M.C.2, 65747 +" 9 13.8 15.5 714 - ” a . C.A.S. 5337 Jervis Bo 14,0 15.8 19 magnirostris CLASS, 5481 Seymour so 13.3 % (1923 2B " 3 5995 y 9 14.9 17.6 3 * " 6164" Q 14,0 15.3 11 0 " 5223 Indefatigable B @ 13.8 16.6 19 fortis . 5336 C] BC tees 16,0 718 ld A.M.N.H, 292387 . Bo 13.5 15.4 16 a Rothschild Coll, 517126 " Po 12.9 15.4 15 3 Gass. 6780 " Po 12.6 16,2 15 " - 5550 y sc 14,3 16.3 15 magnirostris A.M.N.H. 443039 " so 14,0 16,3 716 “ B.M, 99.971.175 * SaCummetaeo 15.8 14 fortis Rothschild Coll, 516931 0 g 14,1 15.8 -- magnirostris (7?) ¥ “f 516929 2 2 13.9 15.7 15 = A.M.N,H, 292385 zy C) 13.7 15.3 11 fortis U,S.N.M. 189233 N, Albemarle P & 13.1 15.4 14 " Rothschild Coll, 517140 " so 13.3 15.6 5 Cc BLM, 99.9-1.156 a fo} ab] 16.7 15 magnirostris C.A.S. 5340 S, Albemarle B & 13.5 15.7 78 fortis Stanford Coll, 4355 i BC 12.7 15.7 18 Me " " 5150 " BC 13.0 15.6 716 J C.A.S. 5446 is Pc 13,1 15.8 15 ® " 5488 " Sc 14.5 16.5 16 magnirostris “ 5551 i SP 14,4 16.3 11 0 x 5508 "I SA 14,0 15.3 11 fortis Rothschild Coll, 517067 u so 13.0 15.3 74 " C.A.S, 6196 . 9 12.9 15.0 16 Ls " 6320 Q 9 13.2 15.2 19 " GeAuss 6036 S, Albemarle 9 13.9 15.8 14 fortis (7) " (new 1939) Chatham Q 12.9 16,2 718 ® Rothschild Coll, 517091 ” 9g 13.2 16.4 11 is C.A.S. 6011 us fo) 13.8 16,0 15 Note: Except for "Geospiza bauri" almost all the above specimens have been assigned by me to one or the other species - not always according to the label on the specimen, and I differ even from Swarth over C.A.S, 6196 and 6320, females from S, Albemarle, For those specimens assigned to magnirostris but within the range ot fortis, I have been especially influenced by the length of the whole culmen, (measurements not published, but available in the deposited measurements). * B-black plumage; P-partly black plumage; S-streaked plumge LACK: VARIATION IN GALAPAGOS FINCHES Table 16 (continued) Of or between Geospiza fortis and Geospiza fuliginosa Museum Number Island Rothschild Coll, 517667 Chatham M.C.Z. 134650 Charles BM, 55.12,14,297 " Rothschild Coll, 517663 Hood " i 517661 No locality, from date is E, Albemarle a a 517662 E, Albemarle BLM, 99.9-1.293 Duncan 55.12.19.176 ? Of or betwee Rothschild Coll, 517702 James Between Geospiza fortis and Geospiza scandens BLM, 37.2.21.400 2 BM. 55.12.19,20 ? C.A,S, 7047 Charles U.S.N.M, 115920 a Rothschild Coll, 516928 Indefatigable U.S.N.M, 17756 5 *Sex Culmen Length Bill Depth Wing Length Here assigned to: BC 9.6 10,2 69 BC 10.1 10,1 67 Bo 9.8 =: 68 BC 8.8 1G.'2 65 so 9.3 10.8 65 sc 9.7 10.7 67 ? 9.2 9.8 64 ? 10.6 10.7 68 unassigned (type Geospiza harterti) fortis uJ (labelled dentirostris) probably fuliginosa but not included in main tables intermediate, referable to Crossman form ?obtained there intermediate, referable to Crossman form ?obtained there fuliginosa fortis (type Geospiza dentirostris) n Geospiza fortis and Geospiza d. debilirostris BC 11.4 10,0 69 "Geospiza nebulosa" a, 10.9 10,1 72 10,2 7. BC 13.0 11,0 13 Po 13.2 11.6 69 - 12.8 13.0 68 ? 12.8 11.0 nl unassigned type-extinct form of Geospiza difficilis? type-extinct form of Geospiza difficilis? unassigned dt -type "Cactornis brevirostris" Unassigned specimens labelled juvenal Geospiza scandens Museum Number Island Rothschild Coll, 517703 James " " 517815 N, Chatham BLM, 99.9-1.335 James Rothschild Coll, 517705 James Stanford Coll, *Sex Culmen Length Bil] Depth Wing Length Here assigned to: shies 11.5 8.2 67 9 10.7 7.8 65 9 11.0 8.0 67 Large Geospiza scendens on James se 14,3 -- 72 Dwarf Platyspiza crassirostris 4469 E, Narborough S@ 8.6 9.8 71 unassigned stray Geospiza scandens from Bindloe (not included in main tables) Platyspiza cressirostris but not included in main tables * B-black plumage; P-partly black plumage; S-streaked plumzge 114 Apparent Camarhynchus affinis from the "wrong" islands ("Camarhynchus incertus") Rothschild Coll, 518053 James C.A.S. 8328 Rothschild Coll, 518058 Rothschild Coll, 518059 C.A.S. Bei) |G C.A.S. 8387 Jervis " 7848 Seymour Abnormal Camarhynchus psittacula from central islands not referable to Camarhynchus affinis M.C,Z, 65738 James BLM, 99.9.1.446 Rothschild Coll, 518005 Jervis C.A:S, 38009 Indefatigable Intermediate between Camarhynchus affinis and Camarhynchus parvulus Stanford Coll, 4310 S, Albemarle " " 4313 " C.A.S, 71815 my " 7883 " i! 7909 4 us 8346 rs ¢ 1969 Y 7988 " HK 71867 " fo +0 Qa 0 © 4 g a fe 7 640 4040 8.5 8.5 8.1 8.1 8.7 9.2 8.1 9.8 10,2 8.6 8.4 164 1.9 19 1.4 1.9 1.8 1.6 7.0 Table 16 (continued) 8.6 9.0 8.5 8.4 9.4 9.3 8.7 8.6 CALIFORNIA ACADEMY OF SCIENCES 70 67 67 67 63 63 64 63 64 64 64 64 67 [Oc. Papers Camarhynchus affinis? " " -type Camarhynchus incertus Camarhynchus affinis? " " " " " n probably juvenal Camarhynchus psittacule am ma Camarhynchus psittacule type"Camarhynchus compressirostris" unassigned, as juvenal unassigned Camarhynchus parvulus** Intermediate between Camarhynchus pauper and Camarhynchus parvulus Museum Number Rothschild Coll, 518136 GLAESe 8232 C.A.S, 7713 e 7714 . 8121 L 71522 *B-black plumage; P-partly black plumage; S-streaked plumage Island Charles Charles " Chatham fo (oy fos fo fo Indefatigable “ 7.8 8.6 Swarth's new species 7.85 7.8 8.0 10,4 8.7 7.6 5.4 509 6.2 6.4 65 66 59 60 59 65 Sex Culmen Length Bill Depth Wing Length Here assigned to: unassigned " Camarhynchus conjunctus nr " lid aureus Cactospiza giffordi “Considered intermediate by Swarth, but here considered a parvulus with unusually long wing No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 115 species on Hood are represented by endemic forms, as are two out of the four species on Tower (a third is found on only one other island), and, taking Culpepper and Wenman together, three out of the four resident species are represented by forms confined to these two islands. Charles and Chatham are two fairly large but moderately isolated islands and both have fewer species than have the central islands, although conditions of habitat, etc., seem quite as favorable. At the same time, they have a higher percentage of endemic forms. From this comparison it is clear that, although (see Text Table 14 in Section V) birds occasionally wander from one Galapagos Island to another, they find it difficult to reach the more isolated islands, so that the present geographical disposition of the islands must be an important factor in the recent evolution of the Geospizinae. From Text Table 15, Chatham would seem more isolated for birds than Abingdon, since Chatham possesses fewer species and a somewhat higher percentage of endemic forms. This is confirmed by evidence outside the Geospizinae, since both the mockingbird (Nesomimus) and the vermilion flycatcher (Pyrocephalus) are, as compared with the central islands, more differentiated on Chatham than on Abingdon. Conditions on the two islands seem fairly simi- lar, although Chatham is considerably larger. Both are about the same distance from the central group. Land connections do not offer adequate explanation for this since Abingdon is separated from the central islands by a greater depth of sea than is Chatham, so presumably became isolated before Chatham. The important factor is perhaps the direction of the prevailing trade wind, which is from the southeast, thus making flight of a bird from the central islands easier in the direction of Abingdon than in the direction of Chatham. Murphy (1938) correlates a number of Polynesian bird distributions with the direction of the prevailing wind, the correlation in these cases being considerably more marked than are those cited in the Galapagos. It may be noted that the trade wind does not blow from the South American mainland to the Galapagos, but re-forms some distance from the land, hence does not facilitate the dispersal of South American land birds to the Galapagos. EvoLuTionary Factors AFFECTING THE GEOSPIZINAE The most important factors affecting the evolution of the Geo- spizinae may be briefly reviewed as follows: (1) other species of 116 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS land birds on the islands are relatively scarce and there were prob- ably none there when the ancestral Geospizinae arrived, hence, while there is an abundance of varied food, the Geospizinae have no, or relatively few, food competitors; (2) as mentioned in Sec- tion III, the Galapagos finches have almost no predators; (3) owing to geographical conditions, they are divided into a number of partly, but not completely, isolated populations, some of which are of very small size. All the above-mentioned conditions favor rapid evolution. (1) There are many instances of adaptive radiation in animals which, through isolation, have found themselves with abundant and varied food and an almost complete absence of food competitors. (2) Worthington (1940) has demonstrated for the fish in East African lakes that rapid evolution and adaptive radiation are much more likely to occur in such an isolated population when predators are absent than when they are present. In lakes Albert and Rudolph, which possess large active predator fish, the Cichlidae have not evolved to nearly the same extent that they have in lakes Victoria, Kiogo, Edward, and George, in which predators are unimportant. The almost complete absence of predators of the Geospizinae on the Galapagos therefore appears significant. Owing to these factors, the intensity of selection pressure presumably has been less for the Geospizinae than for typical mainland birds, and it may be noted that Gifford (1919) and other collectors have commented on the high percentage of deformed and diseased specimens encountered. (3) A series of small and partially isolated populations forms the conditions which, as on theoretical grounds Sewall Wright (1931, 1940) concludes, are the most favorable for rapid evolution. EvoLuTIon oF IsLAND ForMsS There is no evidence in favor of the supposition that the differ- ences between the island forms of each species of Geospizinae are of adaptive significance. All the Galapagos Islands are very much alike, with similar foods, and yet sometimes the populations of a species on islands only a few miles apart are significantly different. For instance, Geospiza magnirostris is significantly smaller on Bindloe than on Abingdon, and on Jervis than on James. The strangely irregular variations in the bill size of G. scandens may also be recalled. Geospiza conirostris conirostris is difterent on two islands which are less than a mile apart. Such differences are ied No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 117 particularly difficult to explain on adaptive grounds in forms like G. fortis and G. conirostris, which show much more marked varia- tion between individuals on the same island than between the means of populations of different islands. Many other examples of appar- ently “pointless” island differences could be quoted. Similar re- marks apply to those species in which the island forms show minor differences in coloration, differences which cannot be correlated with differences in habitat. This is true particularly since, as dis- cussed in Section IV, different species do not vary in the same direction on the same island. This problem of the differences between island forms interested Darwin considerably, and he wrote of the Galapagos finches as follows: ... But how is it that many of the immigrants have been differently modified, though only in a small degree, in islands situated within sight of each other, having the same geological nature, the same height, climate, etc.? This long appeared to me a great difficulty; but it arises in chief part from the deeply-seated error of considering the physical conditions of a country as the most important; whereas it cannot be disputed that the nature of the other species with which each has to compete, is at least as important, and generally a far more important element of success... when in former times an immigrant first settled on one of the islands, or when it subsequently spread from one to another, it would undoubtedly be ex- posed to different conditions in the different islands, for it would have to compete with a different set of organisms.... If then it varied, natural selection would probably favour different varieties in the different islands. (Cf. Darwin, 1888, pp. 355-356.) There is no evidence in favor of Darwin’s suggestion. In fact, there is no evidence whatever, in any of the island forms of Geo- spizinae, that their differences have adaptive significance. In this they resemble most geographical forms which have been described in birds. There are some exceptions, such as various desert forms, whose coloration matches the desert, and which show changes paral- lel with alterations in the color of the desert. The work of Sewall Wright (1931, 1940) and others has demon- strated on theoretical grounds that, provided an isolated population is sufficiently small, it will after a time become different from a neighboring isolated population, although starting with the same gene complex, without the operation of natural selection, owing to the fact that the rate of gene elimination is quicker than the muta- tion rate. In each population, it is partly a matter of chance which 118 CALIFORNIA ACADEMY OF SCIENCES [Oc. Papers genes are eliminated, so that each population becomes stable at a different level. Given a sufficient length of time, such differences must arise provided the populations are small and isolated, and will arise irrespective of any differences between the environments of the populations. The theoretical views of Sewall Wright imply that the smallest and most isolated populations of Geospizinae will tend to be the most differentiated, and also the least variable. That they are the most differentiated is fully borne out by Text Table 15 and the re- lated discussion. That the small isolated populations are the least variable is demonstrated in Section V for Geospiza magnirostris and G. fortis, which are two of the most variable species of Geo- spizinae. However, the latter conclusion did not seem to hold true for the other species of Geospizinae. Hence isolation and size of population are major factors in the formation of island forms of Geospizinae. By themselves, however, they are not sufficient to account for the marked bill differences, since such marked bill dif- ferences are not normally found between small isolated populations of the same species of bird. Such a decreased intensity of selection pressure is perhaps to be expected in the Geospizinae because of their very generalized feeding habits, and an almost complete ab- sence of food competitors. It is also borne out by the wide range in individual variation in bill sometimes shown by one form on one island. Clearly it is where selection pressure is not severe that there is a greater opportunity for the type of nonadaptive evolution, suggested by Sewall Wright, to occur. Despite the first impression of their diversity, the Geospizinae actually show less differentiation into island forms than is typical in insular birds. Thus of the eleven well-distributed species four, namely Geospiza magnirostris, G. fortis, G. fuliginosa, and Platy- spiza crassirostris, are not sufficiently differentiated for me to rec- ognize any subspecies. Geospiza fortis and G. fuliginosa show slight wing differences on Abingdon and Bindloe. Two other spe- cies, Camarhynchus parvulus and Cactospiza pallida, are each rep- resented by only two subspecies, in each case the form on Chatham being separate from that on all the other islands. Geospiza scan- dens shows slightly more differentiation, although even here it is very incomplete. Geospiza difficilis, occurring on six islands, is divided into three subspecies. The other three species show marked differentiation, G. conirostris occurring on three islands in three No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 119 subspecies, Camarhynchus psittacula (superspecies) on eight main islands in four forms, and Certhidea olivacea on all the main is- lands in eight subspecies. The degree of subspecific differentiation in the Geospizinae is not so marked as in many land birds of Poly- nesian Islands (Mayr, 1931-33), although the various Galapagos Islands are separated from each other by distances similar to those separating the Polynesian ones. This suggests either that the pres- ent islands of the Galapagos are of relatively recent separation from each other, or that inter-island wandering is relatively much more common among several of the Geospizinae than among most small land birds of oceanic islands. While the “Sewall Wright effect” is probably the chief cause of subspecific differences between the members of the Geospizinae, another factor may be the random nature of colonization. For in- stance, the population of one island may be derived from a very small and not necessarily typical sample from another island. FUNCTION OF THE BILL DIFFERENCES BETWEEN THE SPECIES Like the island forms of one species, closely related species of Geospizinae differ primarily in bill, to a lesser extent in body size, and in some, but by no means all, instances in minor plumage characters. As summarized in Section III, there is no evidence that in closely related species the bill differences are related in any way to differences in food, feeding habits, or other differences in the ways of life of the species. In particular, Geospiza magniros- tris and G. fortis occupy identical habitats and seem to have iden- tical habits and food. This also seems true of the less-studied Camarhynchus psittacula and C. parvulus. Geospiza fortis and G. fuliginosa also show no significant differences. What then is the significance of the bill differences between closely related species? The field study suggests an answer. In courtship, the male’s most conspicuous action is to feed the female, this occasionally being reduced to billing. When one bird drives off a rival, it normally comes around in front to attack, and, if the intruder still remains, it usually grips the bill of the intruder. Hence the bill features prominently in both aggressive display and courtship, that is in the two items of behavior most closely con- nected with pair-formation. Moreover, when a member of one species flies into the territory of the male of another species, the 120 CALIFORNIA ACADEMY OF SCIENCES [Oc. Papers latter often flies down to it as if to attack, comes around in front, and then the whole behavior simply collapses, and nothing more happens. This strongly suggests that the different species recog- nize each other primarily by their bill differences, and so keep segregated. Experimental evidence, based upon the reaction of live birds to stuffed mounts (described in Section II), supports this view. It should be remembered that there are usually no plumage differences between closely related species. Lorenz (1937) has data for many birds to show that recognition of the object to which adult reactions (social, courtship, etc.) are later directed is acquired while the young are dependent on their parents. This might apply to the Geospizinae since the bill features so prominently in the feeding of the young. Fledgling members of the genus Geospiza beg for food from adults of several other species (see Section II), but the latter do not respond. Hence specific rec- ognition for subsequent breeding behavior reactions may be ac- quired by the juvenile at this stage, but the alternative, that such recognition is inherited, must not be overlooked. The bill differences between closely related species of Geospi- zinae therefore seem similar to numerous other recognition charac- ters in birds, the precise appearance of which is, as pointed out by Lorenz (1937), unimportant provided each character is distinct from that of other species. This view is adapted to the facts as applied to the Geospizinae much better than any other. In addi- tion to field evidence already quoted, it is worth noting that, whereas on Hood, and also on Culpepper, Geospiza conirostris has a deep bill, on Tower it is much less deep. On Tower this species occurs together with G. magnirostris, but its bill is sufficiently narrow to prevent confusion with the latter. On the other islands, where the bill of G. conirostris is so deep that confusion might occur, G. mag- nirostris is not resident. However, the recognition factor must not be overemphasized. First, the bill differences are not so clear-cut as are typical recognition characters in other birds. Second, it seems possible that the different species use other minute characters at close range; thus the bill difference is not the sole factor. At a distance, song is also used at times, but males do not always sing, and the songs of different species partly overlap. Third, there is the point that fortis on Indefatigable is so variable that, if bill dif- ferences were the sole means of recognition, one would hardly expect small fortis to recognize large individuals as its own species, a No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 121 when at the same time it recognizes large fuliginosa as a different species. A parallel difficulty might occur with the larger individ- uals in relation to small magnirostris, which they more closely resemble than they do small individuals of their own species. The frequency distributions and the bill measurements in fortis on Inde- fatigable and elsewhere form an approximately normal distribu- tion. This is evidence for believing that there is no selective mating by bill size within fortis, while at the same time fortis undoubtedly keeps separate, at least normally, from the other two species. On the northern islands this difficulty does not occur, fortis being well separated from the other two species. A possible explanation is that on south Albemarle and Indefatigable conditions may be in process of change at the present time. The whole situation is extremely complex. THE ORIGIN OF CLOSELY RELATED SPECIES I agree with Stresemann (1936) in the belief that the species of Geospizinae have evolved mainly through geographically segre- gated island forms. If two island forms become differentiated, and later one or a few individuals fly from one island to the other, there are two possibilities if these individuals persist and breed: (1) they will interbreed with the corresponding form on the other is- land; (2) they will keep segregated and breed among themselves. As can be seen from various contributions to “The New System- atics” (Huxley, 1940), modern geneticists seem agreed that isola- tion not only produces differentiation in structural characters, but also promotes intersterility, owing to changes in the gene complex. See particularly the discussion by Wright (1940); the following quotation from Muller (1940, p. 256) expresses this view: “Thus a long period of non-mixing of two groups is inevitably attended by the origination of actual immiscibility, i.e. genetic isolation.” Hence if the two forms have been isolated sufficiently long for some degree of intersterility to have appeared, any factors which prevent them from interbreeding will become intensified by natural selec- tion. Individuals of both forms in which this segregating factor was less well developed would tend to have fewer offspring, and eventu- ally, provided there was not too much interbreeding to begin with, two segregated species would be evolved. In Geospiza and Cama- rhynchus, the two genera in which closely related species differ + 122 CALIFORNIA ACADEMY OF SCIENCES [Oc. Papers primarily in size of bill, the island forms also differ primarily in size of bill and, as already mentioned, the species seem to recognize each other primarily by bill differences. Hence all conditions needed for the type of evolution postulated above are present. Those island forms which later met, but differentiated each other through bill differences, have remained segregated (presumably with subsequent intensification of the bill differences), while others which failed to differentiate each other have doubtless merged. Although this is not a completely satisfactory explanation, it is extremely difficult to see any other reason for the origin, and also for the persistence of the bill differences between these closely re- lated species with their similar food requirements. Speculation as to the origin of different forms is always danger- ous because it is difficult to check, and intervening links may have died out. In two cases, however, the origin of a segregated species from a geographically isolated form seems probable. Camarhyn- chus psittacula occurs on James, Indefatigable, and Charles. On Albemarle, it is replaced by the closely similar form C. affinis, which differs in its smaller bill, particularly in depth, and in its more streaked plumage. On Charles, there is another Camarhyn- chus species, pauper, which differs from psittacula primarily in the same ways as affinis, only more markedly. This strongly sug- gests that pauper of Charles is related to psittacula of the central islands through affinis of Albemarle, while, in more recent times, after the differentiation of these forms had occurred, psittacula has colonized Charles from Indefatigable. Hence on Charles psittacula (sensu stricto) has now met pauper, but the two forms have re- mained segregated. Geospiza d. difficilis (formerly called acutirostris) occurs on Tower Island. This species closely resembles fuliginosa, which is absent from Tower, differing only in its straight culmen, in its slightly longer bill, in a slight tendency for the female to be darker, and in its song. In all respects except the last, some specimens of the Chatham Island form of fuliginosa approach Tower difficilis very closely, much more so in bill than do fuliginosa of any other island. It is therefore not unreasonable to assume that the Tower difficilis originated from some large fuliginosa wandering from Chatham. Chatham is a long way to the south, but the prevailing wind is from this direction. If difficilis occurred only on Tower, it would almost certainly be classed as a form of fuliginosa, since the No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 123 latter species occurs on all the main islands excepting Tower (omit- ting Culpepper and possibly Wenman). However, on Abingdon both a small form of fuliginosa and also a form of G. d. difficilis occur. The latter is extremely similar to that on Tower, although averaging larger. Geospiza fuliginosa and G. d. difficilis keep dis- tinct on Abingdon and are also said to have habitat differences. Geospiza difficilis has also spread south on to James and Indefati- gable, as has G. d. debilirostris. Geospiza difficilis presumably evolved from G. fuliginosa via the Tower form. In the other species, evolution is not at a stage at which one can readily elucidate their origin, since most are widespread on all the main islands. One species, Cactospiza heliobates, has a very re- stricted distribution, but its connection with C. pallida is very ob- scure. The peculiar characteristics and distribution of Geospiza conirostris have been discussed in Section V. From studying Swarth’s work, one might suppose that, starting in the north as a small form, Geospiza fortis moved south, gradually becoming larger, and then, on south Albemarle, gave rise to a larger form which became segregated as G. magnirostris; the latter then moved north again, gradually becoming larger and thus increasing the difference between the two species. The trends in the two species, however, are not as regular as implied by Swarth (see discussion in Section V). In addition, too few specimens of magnirostris have been collected on the critical islands of Indefatigable and particu- larly Albemarle, for conclusions to be based on them. I regard such an evolution as possible, but unlikely. In summary, there seems no reason to believe that the species of Geospizinae have evolved in some quite unusual way. The facts are explicable on the view that they have evolved through geo- graphically isolated forms, which remained segregated when they met, that is that their method of speciation is the one which Rensch (1933) concludes is much the most common in birds. Rensch cites Parus major and P. minor, and Ticehurst (1938, pp. 11, 17, 21- 22) mentions Phylloscopus plumbeitarsus and P. viridanus, two examples of species which clearly have evolved via geographically isolated subspecies. Stresemann (1931), Salomonsen (1931), and Rensch (1934) refer to instances among closely related species of European birds which were isolated geographically for a period during the Ice Age and later met. Mayr (1940, p. 272) gives ex- amples from Larus and Pernis. 124 CALIFORNIA ACADEMY OF SCIENCES [Oc. Papers HYBRIDIZATION While two forms originally isolated geographically might re- main segregated when they later met, an alternative possibility is that they might hybridize. In some instances, this might result in a form sufficiently distinct to give rise to a new species. Lowe-( 1936) regards such hybridization as the main method of species formation in the Geospizinae. Positive evidence in favor of this is extremely scanty, but, as discussed in Section V, G. conirostris darwini of Culpepper is probably of hybrid origin between G. conirostris pro- pinqua and G. magnirostris, and the form of Geospiza on Daphne and Crossman is probably a hybrid between G. fortis and G. fuli- ginosa. Incidentally, if the latter case has been correctly inter- preted, it is an instance of two species occurring together over most of their range without interbreeding, but interbreeding in two isolated localities. Moreover, G. conirostris propinqua and G. mag- nirostris do not interbreed on Tower. Such instances seem rare in birds; Meise (19366) found that Passer domesticus and P. hispa- niolensis normally kept separate, but freely interbred in a few areas. This, however, was correlated with a habitat difference be- tween the two species, and only where this habitat difference dis- appeared did the two species interbreed. In other animal groups, the species of Patella discussed by Fischer-Piette (1938) seem somewhat similar, in that they keep distinct in some areas and merge in others. There are of course numerous examples among birds where species normally isolated geographically interbreed when they meet, but the species referred to above normally occur in the same geographical region. Some writers have greatly exaggerated the extent to which hy- bridization occurs among the Geospizinae. Beebe (1924) evidently considered it common, but had no real evidence. Swarth (1934) gives the impression that the boundaries between the species are relatively fluid, and even goes so far as to say that all the different Geospizinae might almost be classed in the same species. It is true that the species of Geospizinae are not quite so sharply separated from each other as typical mainland species in which there are normally no indeterminable specimens. Such indeterminable speci- mens of Geospizinae are, however, relatively rare; almost all speci- mens can be accurately identified, and, in three and one half months’ field work, we found no instances of hybridization even No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 125 though we made special effort to look for such. The species of Geo- spizinae are certainly segregated to nearly the same extent as main- land birds, and the apparent overlapping may be solely in char- acters and not due to interbreeding. Alternatively, the indeter- minate specimens may be the result of hybridization. Camarhyn- chus aureus and conjunctus are thought to be inter-generic hybrids, since they are intermediate in appearance between Certhidea oli- vacea and Camarhynchus parvulus. However, the field study showed that these two genera are so far apart that hybridization between them is most unlikely. It cannot, of course, be ruled out, especially when one remembers that mainland species do hybridize occasionally, as for example Carduelis carduelis and Chloris chloris (Meise, 1936a). If such hybridization occurred a number of times, either between two genera or two species, and the offspring were inter-fertile, one can see how a new species might be formed. There is no evidence, except that noted above, that this has actually oc- curred. It might, however, be difficult to detect, especially if one or both of the parent species had later become extinct. Where a species is only slightly differentiated on two islands and individuals from one island fly to the other, interbreeding will very probably occur, and, if this is regular, it will increase the variability of the island form, but will presumably have no other effect. The variable form of G. fuliginosa on Chatham is perhaps to be explained in this way, and possibly this also applies to some of the variable forms of G. fortis. More definite examples of racial hybridization are given by Mayr (1932b, 1932c) for species of Pachycephala in Polynesia. In particular, Pachycephala pectoralis whitneyi is a hybrid race between two others, dahli and bougain- villei, and has a high individual variation between the two parent races; being insular it can be separately named. Mayr (1938) gives another case in Megapodius. In brief, hybridization may have played a very minor part in the speciation of the Geospizinae, either through two forms, differ- entiated through long isolation, meeting and interbreeding to pro- duce a new form, or through hybridization on very small islands of species which normally keep distinct. There is no positive evi- dence for the former, and the latter occurs rarely. 126 CALIFORNIA ACADEMY OF SCIENCES [Oc. ParErs PARTIAL ISOLATION THE KEy FACTOR Each of the methods of speciation discussed is closely correlated with the existence of a number of islands which are both sufficiently isolated for forms to become differentiated on them, and at the same time sufficiently close together for occasional movements to occur from one island to another. If, as seems likely, there have been marked changes in sea level since the arrival of the Geo- spizinae on the Galapagos, it is not necessary to assume that move- ment between the islands has been possible the whole time. It may thus have been much easier at certain periods than at others. Some authors have stated that the exceptional evolution of the Geospizinae is correlated with their possessing a much greater tendency to vary than most birds, by which presumably is meant a much greater mutation rate. So little is known about mutation rates that such a generalization is unwise, and, as shown in Section V, most forms are not exceptionally variable. In any case, on the above views, it is the opportunities for isolation and subsequent re-meeting which have been primarily responsible for the evolution of the Geospizinae, not an exceptional variability. This view is strengthened by the existence of Pinaroloxias inornata, confined to Cocos Island. There is still only one species of Geospizinae on Cocos, presumably because there is only one island. Had varia- bility been the key factor, however, one would certainly have ex- pected to find as many Geospizinae on Cocos as on some of the Galapagos Islands, because the degree of difference from any Gala- pagos forms shown by Pinaroloxias indicates that it has been iso- lated from them for a long time. THE CLOSELY RELATED SPECIES The problem of the successful segregation of the closely similar species of Geospizinae still remains extremely difficult. Thus on most of the islands three species of Geospiza occur, and on Charles three species of Camarhynchus (two on the other islands), which so far as known show no significant differences with respect to habitat, food, feeding habits, breeding season, aggressive and sex- ual display, nests, or even (except for the Charles Camarhynchus) in plumage, while their songs overlap. The sole differences lie in size, and in the relative size of the bill. In these characters they often come extremely close to each other. Yet apart from a few No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 127 indeterminable specimens, all specimens can be identified. On some islands, the largest Geospiza fortis almost overlap with the smallest magnirostris, and are very different from the smallest fortis, which almost overlap with the largest G. fuliginosa. Again, on Charles Camarhynchus psittacula and C. pauper have a wide overlapping in all measurements except in depth of bill (there are also slight plumage differences). Hence among these species one sometimes finds two specimens extremely similar in all external characters, but still of different origin and genetic constitution. How such species keep distinct is extremely difficult to understand. That they normally do keep distinct, however, is shown by the fre- quency distributions for mensural characters of wing and bill, which lie on approximately normal curves for each species, while intermediate specimens are rare. Hybridization between species, if it occurs at all on the larger islands, certainly is rare, so pre- sumably there is a marked degree of intersterility between these closely related species, despite their similar appearance. ORIGIN OF THE LARGER UNITs Closely related species of Geospizinae are presumed to differ primarily in characters which are nonadaptive, except in so far as they function for species recognition, and have evolved from geo- graphical forms whose differences are primarily nonadaptive. Such specific differences have doubtless been intensified by selection in instances where they serve for recognition. Once specific segrega- tion, however, has been established and the two forms no longer interbreed, divergent evolution is possible, and adaptive differences often appear. Geospiza scandens illustrates an intermediate stage in this process, since its bill is undoubtedly adapted for feeding both on Opuntia flowers and on ants. Since Opuntia flowers are seasonal, during part of the year scandens feeds like other forms of Geospiza. The larger units, represented by the genera among the Geospizinae, clearly differ in adaptive, as well as nonadaptive, characters, as shown by the correlations between bill and feeding habits, discussed in Section III. (See especially Cactospiza and Certhidea.) Presumably such adaptive radiation has been greatly assisted by the absence of food competitors and of predators, as discussed earlier. These last two factors presumably account for the fact that, as stated by Sushkin (1929, p. 377), “In the insular 128 CALIFORNIA ACADEMY OF SCIENCES [Oc. Papers faunae the amount of adaptive radiation within narrow system- atical limits proves to be much larger than on the continents.” An EvoLuTIONARY TREE OF THE GEOSPIZINAE It is interesting to attempt to reconstruct evolutionary trees. In the Geospizinae there are probably many extinct intervening forms, hence all reference to relationships are tentative. Both Sushkin (1925) and Lowe (1936) consider the Geospizinae most closely related on anatomical grounds to the West Indian genus Euetheia (also called Tiaris). I find that Euetheia also resembles the Geo- spizinae in that it shows black or partially black plumage in the male, a seasonal darkening of the bill and long, dense, fluffy rump feathers, while Dr. Wetmore writes me that it builds a partial roof over the nest-cup. Separately, such characters are not of great sig- nificance, since all occur in other tropical passerine birds; together, they are at least suggestive of relationship. If Euetheia is closest to the ancestor of the Geospizinae, then, of the Geospizinae, Geospiza fuliginosa is probably closest to the prim- itive geospizine stock, for small G. fuliginosa closely resembles species of Euetheia in appearance. In addition, G. fuliginosa shows the following primitive or generalized geospizine features: (1) the adult male has black plumage (the evidence that this is the primi- tive condition in Geospizinae is discussed in Section IV); (2) it has the most generalized and varied feeding habits of any form, and eats all the types of food eaten by any of the Geospizinae, save that it does not excavate in wood or catch insects on the wing, which are clearly specialized habits; (3) it has the most generalized type of song of any species; (4) it breeds in the coastal and intermediate zones, which are the most primitive habitats; (5) its bill is the most generalized of any, considering the fringillid ancestry; (6) it is widely distributed in the islands. Geospiza fortis and, with it, magnirostris are closely related to fuliginosa. Geospiza scandens probably represents another line from fuliginosa, with the latter’s split tongue and habit of feeding on Opuntia flowers more developed. Geospiza conirostris probably developed from scandens, but the reverse is also possible, and there may have been hybridization (see discussion in Section V). The evolution of G. d. difficilis and debilirostris from fuliginosa has already been indicated. Geospiza d. septentrionalis of Culpepper and Wenman islands probably fits in here, as, in addition to close | No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 129 similarity in bill, it resembles debilirostris both in the dark female plumage and the bright rufous wing bar and undertail coverts. However, I am not completely convinced of this relationship. Pi- naroloxias inornata possibly came from near G. d. septentrionalis as it shows many similarities in plumage; if this is so its resem- blance to Certhidea in bill is due to convergence. Of the other genera, Certhidea is certainly the most distinct from fuliginosa; this applies especially to bill, plumage, feeding habits, and song. This suggests that it was the first to separate off from the main stock. That it is an old form is also indicated by its wide dis- tribution (it is the only species found on every island), and by the degree to which it has differentiated into island races. However, it has not given rise to more than one species. An alternative sugges- tion for Pinaroloxias is that it became separated from a Certhidea- like form after the latter had acquired feeding habits of the warbler pattern, with some correlated bill adaptation, but before it started to lose the black plumage in the male. The rufous wing bar, which suggests affinities with G. d. septentrionalis, is also found in some island forms of Certhidea. In the three remaining genera, the closest to the Geospiza fuli- ginosa stock is Camarhynchus; bill, feeding habits, and song all suggest this. In these respects C. parvulus seems more primitive than C. psittacula. The relationship of C. pauper to C. psittacula has been discussed. Cactospiza was probably derived from a com- mon ancestor with Camarhynchus; its song is more like that of Camarhynchus than that of any other Geospizinae; it carries fur- ther the tendency for loss of black male plumage, and it has similar feeding habits, in that it picks insects off the vegetation and exca- vates for them in wood, but the latter habit is more specialized in Cactospiza where there is an associated stick-probing habit, and the bill is adapted for excavating. Cactospiza heliobates seems more primitive than C. pallida in its shorter beak and more streaked plumage, but is of very restricted distribution. I have no views on the relationship of these two species. The remaining genus, Platyspiza, also seems closer to Cama- rhynchus than to any other genus, agreeing with it in the partial sup- pression of black male plumage, the habit of frequenting the trees rather than, like Geospiza, the ground, and, to some extent, in the shape of its bill. It has, however, quite different feeding habits, and a specialized song, so probably diverged at an early stage, and 130 CALIFORNIA ACADEMY OF SCIENCES [Oc. Papers certainly before Cactospiza became differentiated from Camarhyn- chus. Platyspiza is the only sharply separated geospizid form suf- ficiently distinct to warrant generic rank which has not evolved either into several species or into island races. The reason for this is unknown. COMPARISON WITH OTHER INSULAR SPECIES In most insular birds, there is only one form on each island, a principle illustrated by Nesomimus on the Galapagos. However, exceptions to this are not infrequent. Stresemann (1927-34) cites the two chaffinch species, Fringilla teydea and F. coelebs canarien- sis, of the Canary Islands, also three Zosterops species, Z. albogu- laris, Z. tenuirostris, and Z. lateralis norfolkensis, on Norfolk Is- land. These instances are due to two, and in the second case three, successive dispersals of the same mainland form to the islands. Similarly, Mayr (1933d, p. 323) found two species of Ptilinopus in the Marquesas, two Aplonis on Ponape, two Halcyon on Palau and Samoa, and, in a recent summary, Mayr (1940, p. 271) gives a number of other examples also due to two waves of colonization of the same species from outside. Hence they are not strictly com- parable with the situation on the Galapagos, where the different forms are presumed to have originated within the archipelago. However, a few cases, and two in particular, do seem comparable to the Geospizinae. Lowe (1923, pp. 519-523) finds two species of finch, Nesospiza wilkinsi and N. acunhae, on the same islands in the Tristan da Cunha group. There are only three islands, so this example is a Galapagos in miniature, and, interestingly enough, the two species differ primarily in size of bill, and also show a tend- ency to lose the full male plumage. At the other extreme are the famous Drepanididae of Hawaii (cf. Perkins, 1903, pp. 368-440), which, in the degree of their differentiation and specialization, have proceeded considerably further than the Geospizinae. The prin- ciples involved, however, seem the same, and, like the Geospizinae, they have spread into many ecological niches not normally occu- pied by Fringillidae; Sushkin (1929) considers that the Drepani- didae are a branch of the Fringillidae. I regard the Geospizinae as being at a stage intermediate between the condition of the Tristan da Cunha Nesospiza and of the Hawaiian Drepanididae. In only a few other instances, two closely related species of endemic land birds occur on the same island. Thus Perkins (1903) finds that No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 131 Phaeornis, a genus of Turdidae in Hawaii, is represented by one form on each main island, except on Kauai where two occur. Simi- larly, Mayr (1933a, pp. 18-20) finds two species of Mayrornis on Ongea Levu, both of which are thought to have originated within the Fiji group. On oceanic islands other groups of land animals sometimes show a type of evolution similar to that of the Geospizinae. In the insects of oceanic islands, Buxton (1938) notes the high proportion of endemics, the development of remarkable and peculiar types, and the extreme variability of certain species. Moreover, and here the parallel with the Geospizinae comes especially close, the members of a genus sometimes form a “complex” of species, with much indi- vidual variation, so that it is sometimes hard to arrive at specific determinations. He particularly cites the coleopteran genus Pro- terhinus, represented in Hawaii by 130 species when Perkins wrote in 1903. The condition found among the Geospizinae is rare among birds. Since groups of small islands are common in many parts of the world, it remains to inquire why the Galapagos, Hawaii, and, to a less extent, Tristan da Cunha, should be so exceptional. The most important factors would seem to be: first, conditions of extreme isolation of the archipelago so far as the birds are concerned; sec- ond, the respective birds must have arrived on the respective islands a very long time ago but yet without the arrival of other species forming food competitors or predators (this would presumably happen where colonization was very difficult). A third condition, suggested by the absence of any such phenomenon on single islands which have been extremely isolated for a very long time, is that for such evolution to occur there must be several islands a moderate distance apart. Indeed, that this type of evolution has occurred only where there are several islands in the group is additional strong evidence for Rensch’s view that the primary requirement for species formation in birds is geographical isolation with possibili- ties for re-meeting later. SUMMARY The expedition to the Galapagos Islands from mid-December, 1938, to early April, 1939, had three main fields of inquiry with regard to the endemic finches of the archipelago: (1) breeding behavior, (2) ecology, (3) hybridization. The first two were cov- ered adequately, but the breeding experiments were not successful. Thirty captive birds survived the trip from the Galapagos to San Francisco, California, and these were placed in aviaries at the California Academy of Sciences. The taxonomy of the group is extremely complicated, but Swarth (1931) did much to clarify most of the pre-existing confusion. The following modifications of Swarth’s systematic treatment of the Geospizinae are suggested. The birds from Culpepper Island placed under Geospiza magnirosiris are moved to Geospiza coni- rostris darwini Rothschild and Hartert. The type of Geospiza harterti Ridgway was correctly synonymized by Swarth with G. fuliginosa, but many specimens from Daphne and Crossman is- lands, originally referred to the same form, are not fuliginosa, nor, as considered by Swarth, are the Daphne birds fortis. They may be hybrid G. fortis & G. fuliginosa populations. Geospiza acuti- rostris Ridgway cannot be separated from G. difficilis Sharpe, even subspecifically, so the former becomes a synonym of the latter. Geospiza septentrionalis nigrescens Swarth cannot be separated from G. s. septentrionalis Rothschild and Hartert, and septentrio- nalis is considered a subspecies of G. difficilis. Two specimens from Narborough Island are possibly referable to an undescribed form of difficilis. Geospiza nebulosa Gould is not a synonym of G. fortis, but a form, now apparently extinct, related to CG. difficilis. The four races of Geospiza scandens (Gould) are merged, although the form on Bindloe Island is separable from that on James. Cacto- spiza pallida producta (Ridgway) is merged with C. p. pallida (Sclater and Salvin). The species of Certhidea are considered as subspecies of Certhidea olivacea Gould. The following-named forms are not considered valid species, but merely variants, or possibly hybrids, not safely assignable to any known species: Cactornis brevirostris Ridgway, Camarhynchus aureus Swarth, Camarhynchus conjunctus Swarth, and Cactospiza giffordi Swarth. A few specimens seem intermediate between Geo- [ 132] No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 133 spiza magnirostris and G. fortis, between G. fortis and G. fuligi- nosa, between Camarhynchus psittacula superspecies and C. par- vulus, and possibly between G. fuliginosa and G. difficilis, and cannot safely be assigned to either of the species concerned. Pair-formation in the Geospizinae follows the typical pattern in territorial songbirds. Postures are extremely generalized, similar in all species, and similar for aggression, sexual behavior, and food-begging by the young. Nest-building and nest-visiting are prominent in the sexual display, and one species frequently utilizes the nest of another. Courtship also includes sexual flights and male feeding female. Only the female incubates, the male feeding her near the nest. Both sexes feed the young in the nest, but only the male feeds the fledglings. The songs are generalized, and the songs of the different species overlap. In aggressive behavior many of the species appear to recognize each other primarily by bill differences. Experiments on Geospiza fuliginosa with mounted birds demonstrated some sexual discrimi- nation by plumage and some specific differentiation by bill, but neither was absolute. Breeding habits have been far more conserva- tive than food habits in the evolution of the Geospizinae. Provided Cactospiza is grouped with Camarhynchus, the main genera of Geospizinae have different feeding habits. The species within the genera feed largely on similar foods, Geospiza scandens being a partial exception. The bill is adapted to specialized feeding habits in Certhidea, Cactospiza, and Geospiza scandens, but many bill differences, especially those between closely related species, cannot be correlated with food differences. Cactospiza pallida uses spines and twigs for probing insects from crannies. This is one of the few known instances of use of tools by animals other than man. The nests of all species are similar, and are located in similar sites, excepting that only the species of Geospiza utilize the plant Opuntia. Habitat differences separate closely related species in two in- stances. Most closely related species frequent extremely similar habitats, and Geospiza magnirostris and G. fortis occupy identical ecological niches. In all species most of the breeding occurs during the rainy season. Geospiza scandens and Cactospiza pallida start nesting before the others. Censuses could not be taken. Estimates are given, varying from about 20 adults per acre in rich, intermediate forest to about two 134 CALIFORNIA ACADEMY OF SCIENCES [Oc. Papers per acre (or less) in the open country on the tops of the larger islands. Some well-defined species are represented by total popu- lations of only a few thousand individuals. Adult males of Geospiza are black, but they frequently breed in immature plumage. The fully adult plumage in species of Platy- spiza and Camarhynchus corresponds to a transitional stage in Geospiza, and in Cactospiza the black has been lost entirely, except for one individual examined which was black-headed. In Platy- spiza and Camarhynchus, the disappearance of black male plumage appears to be occurring at different rates on different islands. Loss of male secondary sexual plumage characters is not uncommon in insular forms; perhaps this is correlated with the disappearance of its function in specific recognition by the female. Differences in the female plumage of the species and island forms mainly concern the shades of brown, gray, or olive of the upper parts, the degree of streaking of the under parts, and the color of the wing bars. The differences between the island forms and species do not seem to be adaptive in character. In the Geospizinae, plumage characters have been much more conservative in evolution than the size and shape of the bill. In Certhidea, the northernmost form closely resembles the southern- most, and the next most northerly one resembles another southern form, while those on the central islands are more like each other. In the breeding season, bills of adult Geospizinae are dark; outside of the breeding season they are yellowish. Older males have longer wings than younger males, which in turn have longer wings than females. Males have larger bills than females, except in Certhidea, in which the female has a longer bill than the male. The island populations of each species are often slightly differ- entiated by mensural characters, although often not sufficient to justify subspecific recognition. Each island population seems to have evolved independently, with periodic disturbances through inter-island wandering, which appears to be uncommon. Geospiza conirostris conirostris is slightly differentiated on two islands less than a mile apart. Small isolated populations of Geospiza magni- rostris and G. fortis have a lower variability in measured charac- ters than large populations. The other species do not show this phenomenon. Correlation of bill and wing, and of culmen with depth of bill, is high in most species, especially the larger ones; although it may No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 135 be absent in a few species. Geospiza magnirostris, fortis, and fuliginosa seem structurally to differ from each other solely in gen- eral size and in relative size of bill. The specific limits sometimes occur at different values on different islands. On some islands, G. fortis is so variable that larger individuals are closer to magniros- tris and smaller ones to fuliginosa than they are to each other. The ratio of wing to culmen and the ratio of culmen to depth of bill show no allometric relation within each species, but have a typical average, with wide individual variation, for each species. A few intermediate specimens cannot safely be assigned. The forms on Daphne and Crossman are thought to be of hybrid origin between fortis and fuliginosa. The forms of Camarhynchus also differ pri- marily in size and in relative size of bill, a phenomenon which is rare in closely related species of birds. Most species of Geospizinae are not abnormally variable, but a few are. Geospiza conirostris darwini of Culpepper is thought to be of hybrid origin. Birds have not reached the Galapagos Islands via a land bridge, although the islands were probably connected with each other at some period in the past. The few land birds which have reached the Galapagos show very different degrees of differentiation from their mainland ancestors, probably correlated with differences in times of arrival. The small isolated islands have fewer resident species of the Geospizinae, and a higher proportion of endemic forms than have either the larger or the central islands. The most important factors in the evolution of the Geospizinae have probably been the almost complete absence of food competi- tors and predators, and the existence of several islands which pro- vided partial, but not complete, isolation for island forms. Differ- ences between island forms of the same species are considered non- adaptive, and due primarily to the “Sewall Wright effect,” while colonization by an atypical sample may be a subsidiary factor. Species probably originated mainly through geographically iso- lated races which later met and kept distinct, but some forms may be of hybrid origin. Differences between closely related species are nonadaptive except that bill characters serve in species recognition. The main genera show an adaptive radiation. The species of Geo- spizinae are not as sharply defined as in mainland birds, but they do not show the degree of overlapping or of hybridization some- times claimed for them, and there is no need to assume for them some quite exceptional method of evolution. LITERATURE CITED Baur, G. 1891a. On the origin of the Galapagos Islands, Amer. Nat., 25: 217-229. 18916. On the origin of the Galapagos Islands, Amer. Nat., 25: 307-326. 1891c. Chatham Island, Galapagos Archipelago, Aug. 28, 1891, Amer. Nat., 25: 902-907. BEEBE, W. 1924. Galapagos: world’s end. G. P. Putnam’s Sons, New York, pp. xxi+443, 422 text figs., 8 colored pls., 1 colored frontispiece. Bertroz, J. 1927. Remarques sur l’hybridation naturelle chez les oiseaux, Bull. Soc. Zool. France, 52: 393-403. Buxton, P. A. 1938. The formation of species among insects in Samoa and other oceanic islands, Proc. Linn. Soc. London, 150th session, 1938, pt. 4, pp. 264-267. Cuarin, J. P. 1924. Size-variation in Pyrenestes, a genus of weaver-finches, Bull. Amer. Mus. Nat. Hist., 49: 415-441, 11 figs. Dati, W. H., and W. H. OcHsNER 1928. Tertiary and Pleistocene Mollusca from the Galapagos Islands, Proc. Calif. Acad. Sci. (4), 17: 89-139, pls. 2-7, 5 text figs. DanrortH, C. H. 1939. Relation of genic and endocrine factors in sex (Jn: Sex and internal secretions. A survey of recent research, ed. E. Allen), The Williams & Wilkins Company, Baltimore, pp. 328-350, 1 pl., 2 text figs. Darwin, C. 1888. On the origin of species by means of natural selection, or the preservation of favored races in the struggle for life. John Murray, London, 6th ed., pp. xxi+458. Day, B., and R. A. FIsHER 1937. The comparison of variability in populations having unequal means. An ex- ample of the analysis of covariance with multiple dependent and independent variates, Ann. Eugenics, 7 : 333-348, 16 tbls. DrowneE, F. P. 1899. Notes from the diary of Mr. F. P. Drowne (Jn: A review of the ornithology of the Galapagos Islands. With notes on the Webster-Harris Expedition, by W. Rothschild and E. Hartert), Novitat. Zool., 6: 104-136, 10 illustr., 6 text figs. FIscHER-PIETTE, E. 1938. The concept of species and geographical isolation in the case of North Atlantic Patellas, Proc. Linn. Soc. London, 150th session, 1938, pt. 4, pp. 268-275. FisHer, R. A. 1937. The relation between variability and abundance shown by the measurements of the eggs of British nesting birds, Proc. Roy. Soc. London, ser. B, no. 826, 122: 1-26, 1 text fig. FRIEDMANN, H. 1934. The hawks of the genus Chondrohierax, Journ. Wash. Acad. Sci., 24: 310-318. Gauss, G. F. 1939. In: Analytical population studies in relation to general ecology, by T. Park, Amer. Midland Nat., 21: 235-255, 2 text figs. [ 136 J No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 137 Girrorp, E. W. 1919. Field notes on the land birds of the Galapagos Islands and of Cocos Island, Costa Rica. Expedition of the California Academy of Sciences to the Galapagos Islands, 1905-1906, XIII, Proc. Calif. Acad. Sci. (4), 2 (pt. 2) : 189-258. GitBert, P. A. 1939. The bower-painting habit of the satin bower-bird (Ptilonorhynchus violaceus) , Emu, 39: 18-22, pl. 7. Harris, C. M. 1899. Diary of Charles Miller Harris (In: A review of the ornithology of the Gala- pagos Islands. With notes on the Webster-Harris Expedition, by W. Rothschild and E. Hartert), Novitat. Zool., 6: 86-104, 2 text figs. HeEtimayr, C. E. 1927. Catalogue of birds of the Americas. Pt. V, Field Mus. Nat. Hist., zool. ser., 13: vi+-517. 1938. Catalogue of birds of the Americas. Pt. XI, Field Mus. Nat. Hist., zool. ser., 13: vi+662. Howarp, H. E. 1929. An introduction to the study of bird behaviour. Cambridge, The University Press, pp. xii+136, 10 pls., 1 frontispiece. Huxtey, J. S. 1932. Problems of relative growth. Lincoln MacVeagh, The Dial Press, New York, pp. xix-+276, 104 text figs., 1 frontispiece. 1940. Introductory: towards the new systematics (Jn: The new systematics, ed. J. S. Huxley), Oxford at the Clarendon Press, pp. 146. Lack, D. 1939. The behaviour of the robin, Proc. Zool. Soc. London, 109: 169-219, 1 pl., 5 text figs., 3 tbls. 1940a. Courtship feeding in birds, Auk, 57: 169-178. 19406. Evolution of the Galapagos finches, Nature, 146: 324, 1 text fig. 1940c. Habitat selection and speciation in birds, Brit. Birds, 34: 80-84. 1940d. Variation in the introduced English sparrow, Condor, 42: 239-241. 1941. Notes on territory, fighting, and display in the chaffinch, Brit. Birds, 34: 216— 219. Lorenz, K. Z. 1937. The companion in the bird’s world, Auk, 54: 245-273. Lowe, P. R. 1923. Notes on some land birds of the Tristan da Cunha group collected by the ‘Quest’ Expedition, Ibis (11), 5: 511-529, text figs. 8-11. 1930. Hybridisation in birds in its possible relation to the evolution of species, Bull. Brit; Orn: Cl., 50: 22-29. 1936. The finches of the Galapagos in relation to Darwin’s conception of species, Ibis (13), 6: 310-321, pls. III, IV, text figs. 1, 2. MarsHa.t, F. H. A. 1936. XI—The Croonian lecture: Sexual periodicity and the causes which determine it, Phil. Trans. Roy. Soc. London, ser. B, 226: 423-456. Mayr, E. 193la. A systematic list of the birds of Rennell Island with descriptions of new species and subspecies. Birds collected during the Whitney South Sea Expe- dition, XIII, Amer. Mus. Novit., no. 486, pp. 1-29. 19316. Notes on Halcyon chloris and some of its subspecies. Birds collected during the Whitney South Sea Expedition, XII, Amer. Mus. Novit., no. 469, pp. 1-10. 1931c. The birds of Malaita Island (British Solomon Islands). Birds collected dur- ing the Whitney South Sea Expedition, XVII, Amer. Mus. Novit., no. 504, pp. 1-26, 1 text fig. 138 CALIFORNIA ACADEMY OF SCIENCES [Oc. Papers 1932a. Notes on Meliphagidae from Polynesia and the Solomon Islands. Birds col- lected during the Whitney South Sea Expedition, XVIII, Amer. Mus. Novit., no. 516, pp. 1-30. 1932b. Notes on thickheads (Pachycephala) from Polynesia. Birds collected during the Whitney South Sea Expedition, XXI, Amer. Mus. Novit., no. 531, pp. 1-23. 1932c. Notes on thickheads (Pachycephala) from the Solomon Islands. Birds col- lected during the Whitney South Sea Expedition, XX, Amer. Mus. Novit., no. 522, pp. 1-22. 1933a. Notes on the genera Myiagra and Mayrornis. Birds collected during the Whitney South Sea Expedition, XXV, Amer. Mus. Novit., no. 651, pp. 1-20. 19336. Notes on Polynesian flycatchers and a revision of the genus Clytorhynchus Elliot. Birds collected during the Whitney South Sea Expedition, XXIV, Amer. Mus. Novit., no. 628, pp. 1-21. 1933c. Notes on the variation of immature and adult plumages in birds and a physi- ological explanation of abnormal plumages. Birds collected during the Whit- ney South Sea Expedition, XXVIII, Amer. Mus. Novit., no. 666, pp. 1-10. 1933d. Die Vogelwelt Polynesiens, Mitth. Zool. Mus. Berlin, 19: 306-323, 3 text figs. 1934a. Notes on some birds from New Britain, Bismark Archipelago. Birds collected during the Whitney South Sea Expedition, XXVIII, Amer. Mus. Novit., no. 709, pp. 1-15. 19346. Notes on the genus Petroica. Birds collected during the Whitney South Sea Expedition, XXIX, Amer. Mus. Novit., no. 714, pp. 1-19, 1 text fig. 1938. Notes on New Guinea birds, IV. Birds collected during the Whitney South Sea Expedition, XX XIX, Amer. Mus. Novit., no. 1006, pp. 1-16. 1940. Speciation phenomena in birds, Amer. Nat., 74: 249-278, 7 text figs. MeEtsE, W. 1936a. Uber Artentstehung durch Kreuzung in der Vogelwelt, Biol. Zentralbl., 56: 590-604, 1 text fig. 19366. Zur Systematik und Verbreitungsgeschichte der Haus- und Weidensperlinge, Passer domesticus (L.) und hispaniolensis (T.), Journ. f. Ornithologie, 84: 631-672, 3 text figs. Miter, A. H. 1931. Systematic revision and natural history of the American shrikes (Lanius), Univ. Calif. Publ. Zool., 38: 11-242, 65 text figs., 22 tbls. 1941. Speciation in the avian genus Junco, Univ. Calif. Publ. Zool., 44: 173-434, 33 text figs., 37 tbls. Mutter, H. J. 1940. Bearings of the ‘Drosophila’ work on systematics (In: The new systematics, ed. J. S. Huxley), Oxford at the Clarendon Press, pp. 185-268, 7 text figs. Murpny, R. C. 1936. Oceanic birds of South America. American Museum of Natural History, New York, 1: xxii+640, pls. 1-38-++5 colored pls., 1 colored frontispiece, text figs. 1-61. 2: 641-1245, pls. 39-72+9 colored pls., 1 colored frontispiece, text figs. 62-80. 1938. The need of insular exploration as illustrated by birds, Science, 88: 533-539, 2 text figs. Morpny, R. C., and J. P. CHApin 1929. A collection of birds from the Azores, Amer. Mus. Novit., no. 384, pp. 1-23, 1 text fig. Murpuy, R. C., and G. M. Matuews 1928. Notes on the genera Pomarea and Conopoderas. Birds collected during the Whitney South Sea Expedition, V, Amer. Mus. Novit., no. 337: pp. 1-18. No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 139 Nos te, G. K., and W. Voct 1935. An experimental study of sex recognition in birds, Auk, 52: 278-286, pls. 14, 115}; Orr, R. T. 1938. A new rodent of the genus Nesoryzomys from the Galapagos Islands, Proc. Calif. Acad. Sci. (4), 23: 303-306, pl. 25. Oscoop, W. H. 1929. A new rodent from the Galapagos Islands, Field Mus. Nat. Hist., zool. ser., 17: 21-24, Perkins, R. C. L. 1903. Vertebrata. Fauna Hawaiiensis, 1 (pt. 4) : 365-466, 1 text fig. Pott, H. 1911. Uber Vogelmischlinge, Verh. 5, Internat. Ornithologen-Kongresses, pp. 399- 468, 5 pls., 3 text figs. Pycrart, W. P. 1925. Camouflage in nature. Hutchinson & Co., London, pp. xiv-+280, 35 double plate illustr. in text, 1 colored frontispiece. Renscu, B. 1933. Zoologische Systematik- und Artbildungsprobleme, Verhandlungen der Deut- schen Zoologischen Gesellschaft, 35. Jahresversammlung in Koln vom 6 bis 8. Juni, 1933. Zoologischer Anzeiger, 6 Supplementband, pp. 19-83, 6 text figs., 5 tbls. 1934. Kurze Anweisung fiir Zoologisch-Systematische Studien, Akademische Ver- lagsgesellschaft, Leipzig, pp. 1-116, 22 text figs. Ripeway, R. 1897. Birds of the Galapagos Archipelago, Proc. U.S. Nat. Mus., 19: 459-670, pls. 56, 57, 7 text figs., map. Rortuscui.p, W., and H. HarTert 1899. A review of the ornithology of the Galapagos Islands. With notes on the Web- ster-Harris Expedition, Novitat. Zool., 6: 85-205, pls. V, VI, 13 illustr., 6 text figs. 1902. Further notes on the fauna of the Galapagos Islands, Novitat. Zool., 9: 373-418, pl. X. 5 text figs. RotuscHILp, W., and N. B. KInNEAR 1929. List of hybrid ducks, Bull. Brit. Orn. Cl., 49: 93-99. SALOMONSEN, F. 1931. Diluviale Isolation und Artenbildung, Proc. Seventh Internat. Ornithological Congress, pp. 413-438, 5 text figs. Scuarrr, R. 1912. Distribution and origin of life in America. The Macmillan Co., New York, pp. xvi+497, 21 maps. SHEPHARD, F. P., and C. N. Bearp 1938. Submarine canyons: distribution and longitudinal profiles, Geogr. Rev., 28: 439-451, 5 text figs., 2 tbls. Srptey, C. L. 1938. Hybrids of and with North American Anatidae, Congr. Orn. Int., 9: 327-335. Snoperass, R. E. 1902. The relation of the food to the size and shape of the bill in the Galapagos genus Geospiza, Auk, 19: 367-381, pls. 11-13, tbls. Snoperass, R. E., and E. HELLER 1904. Papers from the Hopkins-Stanford Galapagos Expedition, 1898-99, XVI, Birds, Proc. Wash. Acad. Sci., 5: 231-372, tbls. 140 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS STEINBACHER, J. 1936. Zur Frage der Geschlechtsreife von Kleinvogeln, Beitrage zur Fortpflanzungs- Biologie der Vogel, Jahrg. 12, Nr. 3, s. 139-144. STRESEMANN, E. 1930. Welche Paradiesvogelarten der Literatur sind hybriden Ursprungs? Novitat. Zool., 36: 6-15, 1 text fig. 1931. Sauropsida: Aves (Jn: Handbuch der Zoologie. Eine Naturgeschichte der Stimme des Tierreiches. Walter De Gruyter & Co., Berlin and Leipzig), Bogen 35 bis 41, Siebenter Band, Zweite Halfte, Sechste Lieferung, pp. 545-656, figs. 549-658. 1936. Zur Frage der Artbildung in der Gattung Geospiza, Orgaan der Club van Nederlandische Vogelkundigen, 9 (no. 1) : 13-21. SusHki, P. P. 1925. The evening grosbeak (Hesperiphona), the only American genus of a Palae- arctic group, Auk, 42: 256-261, 2 text figs. 1929. On some peculiarities of adaptive radiation presented by insular faunae, Verh. 6, Internat. Ornithologen-Kongresses, pp. 375-378. SwartH, H. S. 1929. A new bird family (Geospizidae) from the Galapagos Islands, Proc. Calif. Acad. Sci. (4), 18: 29-43, 1 text fig. 1931. The avifauna of the Galapagos Islands, Occas. Papers Calif. Acad. Sci., 18: 1-299, 57 text figs., tbls., 1 map. 1934. The bird fauna of the Galapagos Islands in relation to species formation, Biol. Rev., 9: 213-234, 2 text figs. TiceHursT, C. B. 1938. A systematic review of the genus Phylloscopus (willow-warblers or leaf-war- blers). London: Printed by order of the trustees of the British Museum, pp. viii+193, 2 pls., 7 maps. TINBERGEN, N. 1939. On the analysis of social organization among vertebrates, with special refer- ence to birds, Amer. Midland Nat., 21: 210-234, 2 text figs. Van DeEnzurcH, J. 1912a. The geckos of the Galapagos Archipelago. Expedition of the California Acad- emy of Sciences to the Galapagos Islands, 1905-1906, Proc. Calif. Acad. Sci. (4), 1: 405-430. 19126. The snakes of the Galapagos Islands. Expedition of the California Academy of Sciences to the Galapagos Islands, 1905-1906, Proc. Calif. Acad. Sci (4), 1: 323-374. 1914. The gigantic land tortoises of the Galapagos Archipelago. Expedition of the California Academy of Sciences to the Galapagos Islands, 1905-1906, Proc. Calif. Acad. Sci. (4), 2(pt. 1) : 203-374, pls. 12-124. Van DensurcH, J., and J. R. SLEvin 1913. The Galapagoan lizards of the genus Tropidurus; with notes on the iguanas of the genera Conolophus and Amblyrhynchus. Expedition of the California Academy of Sciences to the Galapagos Islands, 1905-1906, Proc. Calif. Acad. Sci. (4), 2(pt. 1) : 133-202, pls. 8-11. VENABLES, L. S. V. 1937. Bird distribution on Surrey Greensand heaths: the avifaunal-botanical correla- tion, Journ. Animal Ecol., 6: 73-85, pls. 2, 3, 2 maps. Wintersottom, J. M. 1929. Studies in sexual phenomena, VII. The transference of male secondary sexual display characters to the female, Journ. Genet., 21: 367-387, 6 tbls. No. 21] LACK: VARIATION IN GALAPAGOS FINCHES 141 Wirscui, E. 1935. Seasonal sex characters in birds and their hormonal control, Wilson Bull., 47: 177-188. WorrtuinctTon, E. B. 1940. Geographical differentiation in fresh waters with special reference to fish (In: The new systematics, ed. J. S. Huxley), Oxford at the Clarendon Press, pp. 287-302. Wricnt, S. 1931. Evolution in Mendelian populations, Genetics, 16: 97-159, 21 text figs. 1940. The statistical consequences of Mendelian heredity in relation to speciation (In: The new systematics, ed. J. S. Huxley), Oxford at the Clarendon Press, pp. 161-183, 7 text figs. MAIN TABLES gz°T (L8-S8) £°98 0°08 0°9L (6L-9L) S°LL ST°T (Og-gL) L°gl wS*2 (2g-eL) 6°LL S9°2 (6L-€L) 0°9L OT*2 (€Q-SL) 6°6L v0°2 (wg-SL) 0°08 Bv'2 (2g- (TS g=Lb°L) verk vor" €LS* (6°6-6°L) ov’g 9 OTr* zov’ = (6°Q-T°L) L6°L €E| eer" S0S° (g°g-S°L) S$o°g 9 gre" Lgeg* (6°g-6°9) L6°L Lv) Sge° ogg = (g°g-e°L) TL°L L 262° €gz° (v*g-v°L) vo°g ve] Lee? Ozv’ (6*g-z2°L) vg*L LE} oge? Lgc®’ = (v°g-6°9) 2l°L LT) 65¢° 09°L T pre’ (T’g-v°L) LL°L OL S2r* 9év° (9°-€*L) €0°g TT] 602° €6T° (S*L=L°9) vo"L 9T] S6¢° 26¢° (v°g-0°L) ¥5°L Te] Trr® 6gz° — (<*g-g°L) L6°L ¢ a” Ss ,UeWeInsBeR “ON te} seTuUe Ty TIFqa go y4deq (T°6-¢°L) (9°g-<°L) (¥°6-g°L) (T°6-T°L) (8°8-8°9) (0°6-6°L) (5°6-9°L) (0°6-L°L) (0°6-S°L) (v°g-7°L) (0°6-¢°L) (<°6-S°L) (0°6-S°L) (T°6=L°L) (0°6-S°L) (<°g-v°L) (2°g-L°9) ($°g-L°9) VTS v1 92°8 go"g 60°8 95°38 TE"8 228 £T°8 T6°L £T°g LT’ 22°98 org 61g 20° or’L oL*L 8S usmeiInseey seTen *on (€*0T-2°L) (2°6-g°L) (€£°6-6°L) (<°OT-T"L) (9°8-2°8) (£°6-T"8) (T°6-S°L) (¥°g-9°L) (7°6-0°L) (6°g-S°L) (¥°6-T°L) (v°6-9°L) (2°6-0°8) (6°g-g°L) (¢°6-6°L) (2°g-0°L) (L°g-<°L) (6°g-6°L) T's 6£°8 TS°8 ¥9°8 0S°g L9°8 v2"g 00°g £T°8 ¥0"g €5°g vg 9V°8 02° Tv'g Sv°3 98°L 20°8 €£°8 sj,ucweInsEeR soeTeuey ueutno jo 442uey BSOUTST [Ng BZyds0eD UT UOTIeTAeq PIBpPUe{S pue Ss4ucMOeINSsBe_ oWeIyXY pus eFerIOAY Lg cv S2 "On eTeu Teueanf , (v°6-L°L) (¥°6-0"8) (L*6-T°8) (2°oT-2°L) (9°6-2°8) (L°6-T"9) (9°6-6°L) (9°8-¥°L) (2°6-g°L) (s°g-9°L) (9°6-L°L) (£°6-S°L) (S°6-6°L) (0°6-£°8) (<°6-6°L) (T°6-9°L) (L°g-v°L) (T°6-S°L) 65°¢ 6S°8 gl" 6L°8 9L°8 98°8 (o) — "2 (69-65) 0°S9 9T/6Q°T (2L-L9) £°69 (7°6-9°8) vO°6 PTILLy® (<*0T-S°g) gz°6 STITLz* (6°g-0°g) 9£°g 9TISTv® (S°6-0°8) €5°g LT eTseueqTy “Ss g2°T (L9-79) 8°59 ¥ (TL-L9) 0°69 (g°g-L°g) SL°g 2 lzvL° (g°6-2°6) ge°6 «6S |Ievv* (0°6-T°Q) €5°8 4 199¥° (2°6-T'Q) 29°g 4 eTXBueQTY “HZ ¥ “OD g6°0 (99-99) 9°99 9 IST°T (69-L9) L°L9 ($°6-L°g) £0°6 9 |wLe* (T*oT-2°6) OL°6 vv |]9Lz° (L°g-6°L) Ov'g 9 |592° (g*e-€"g) OS°g eTreueqTyY “N (TL-L9) 0°69 2 (€lL-OL) $°TL (6°6-0°6) S¥°6 (€°0t-6°6) OT°OT 2 (0°6-2°g) 09°g 2 (2°6-6°8) S0°6 «2 yanorogzeN stutsse “OD gS°2 (TL-79) L°L9 9 (2L-2L) (o°TT-0°6) zo°oT 9 |6T9° (2°TT-9°6) SE°OT ¥ H166L° (v*0T-9°g) LE*6 9 |6TQ* (v*OT-0°6) OVS ¢ usoung ByUtsJe-B[Noey4yed 2 WN ea) O a ea) — O 7) 5 To*2 (¢l-99) 2°69 £T}90°2 (vL-OL)4S°2L (6°0T-6°6) vE*OT OT/Z9v° (T°TT-£°OT) EQ*OT <£ H}06z* (6°6-0°6) 9”°6 I 05 CE) Moderate size, robust, black or dark piceous, convex, glabrous above, Head finely and rather sparsely punctured, transverse impres- sion in front of eyes, clypeus acutely emarginate; antennae almost reaching hind margin of prothorax, third segment shorter than two 110 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS following united, clavate from seventh segment, tenth segment trans- verse. Prothorax almost two-fifths broader than long, apex broadly emarginate, base bisinuate, sides straight and parallel for basal half, convergent and feebly arcuate to apex, hind angles subacute and slightly extended backwards; dise very convex, finely and rather densely punctured and with head markedly alutaceous and opaque. Seutellum finely, sparsely punctured. Elytra very convex, aluta- eeous, two-fifths longer than broad and over two and a half times as long as prothorax, sides feebly arcuate, more rounded and con- vergent to apex; dise deeply and acutely suleate, with fine yet evi- dent and well spaced punctures, intervals broad, carinate and with summits finely punctured. Beneath, coarsely and shallowly punc- tured in front, rather coarsely and discretely punctured behind and longitudinally rugose. Males with front ventral segments concave, last ventral segment flattened apically and a small sparse patch of yellow pile on undersurface of middle and hind femora. Length 7 mm., breadth 3 mm. Boheman gives no particular island for his specimens. Blair lists one from James Island. The specimens in the California Academy of Sciences Collection which fit the description are five specimens from Indefatigable Island, collected May 5, 1932, by M. Willows, Jr., of the Templeton Crocker Expedition. These vary somewhat in size, most being a bit larger than the length given by Boheman, 514 mm. This opaque species with distinctly suleate elytra, punctured striae and earinate intervals should be confused with no others. In P. costatus, the pronotum is coarsely sculptured, with somewhat strigose punc- tures, and elytral intervals unequally elevated. In P. lugubris and those species which are associated with it, the elytral sulci are im- punctate. ; Pedonoeces batesoni Blair Pedonoeces batesoni Buatir, 1933, Ann. Mag. Nat. Hist., ser. 10, XI, p. 479. I have no specimens of this species before me. Pedonoeces duncani Van Dyke, new species Very similar to P. morio in shape and general appearance but some- what more elongate and less convex. Head coarsely and densely yet shallowly punctured and rugose, transverse impression in front of eyes vague, clypeal emargination acute. Prothorax two-fifths broader than long, apex broadly and somewhat shallowly emarginate, base markedly No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 111 bisinuate, sides straight and feebly divergent from base to beyond mid- dle and thence arcuate and narrowed to apex; dise coarsely, closely punctured with most punctures longitudinally anastomosing producing an irregular striate appearance. Elytra three-eighths longer than broad, acutely suleate with strial punctures fine but well impressed and close, the intervals of equal width to sulei and raised and eari- nate. Beneath coarsely punctate and rugose in front, more distinctly and discretely punctured behind. Holotype, a mature piceous specimen, 6 mm. long and 2.5 mm. wide, probably a male, a paratype of the same size but immature and two larger mature specimens, 7 mm. long by 3 mm. wide, presumably females, collected on Duncan Island, the first two August 14, 1906, and the others December 1-17, 1905, by F. X. Williams. This species belongs near P. morio as shown by its general appear- ance and elytral sculpturing but its more flattened upper surface and erossly punctured head and pronotum readily separate it. Pedonoeces costatus G. R. Waterhouse Plate V, figure 6 Pedonoeces costatus G. R. WATERHOUSE, 1845, Ann. Nat. Hist., XVI, p. 35. Pedonoeces costatus G. R. Waterhouse, LINELL, 1898, Proc. U. S. Nat. Mus., XT, pi 265. Pedonoeces costatus G. R. Waterhouse, MuTcHLeErR, 1925, Zoologica, V, no. 20, D. 237. Pedonoeces costatus G. R. Waterhouse, Buatr, 1933, Ann. Mag. Nat. Hist., ser. 10, XI, p. 479. Moderate in size, robust, black or piceous with antennae, mouth- parts and legs rufous, quite convex above, and glabrous. Head coarsely and closely punctured, more finely so in front and behind, aluta- ceous, transversely impressed in front of eyes, elypeus broadly emarginate; antennae about reaching posterior third of prothorax, third segment fully as long as two following united, clavate from eighth segment, ninth and tenth segments transverse. Prothorax about one-fifth broader than long, quite convex, apex broadly emarginate, base bisinuate, sides feebly arcuate from base to middle thence more gradually rounded and narrowed to apex, hind angles broad and acute and projecting backwards; dise irregularly longi- tudinally striate and punctate. Secutellum finely punctured. Ely- tra over one-fourth longer than broad, and fully twice as long as prothorax, somewhat spatulate in shape, humeri rounded, sides arcuate and well rounded to apex; dise very convex, sulci broad and deep, the strial punctures well rounded as well as well spaced 112 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS and well impressed, intervals narrowly carinate, the third, fifth, and seventh considerably more elevated than others, general surface alu- taceous. Beneath coarsely and densely punctured in front, more sparsely and more finely so towards apex of abdomen, and entire abdomen quite alutaceous. Males with front ventral segments broadly coneave, the last ventral segment broadly excavated, the second and third femora with dense patches of yellow pile beneath and the first less distinctly pubescent. The females with last ventral seg- ment deeply impressed. Lengths 6-7 mm., breadth 2.5-3 mm. James Island is the type locality. The California Academy of Sciences has numerous specimens from there, all collected by F. X. Williams on January 1-4, 1906. It is a well defined species because of its striate pronotum and elytra with alternately elevated costae and well defined strial punctures. It is only closely approached by the next species. Pedonoeces williamsi Van Dyke, new species Of moderate size, elliptical, robust, rufous (probably immature), very convex above and glabrous. Head very coarsely and closely and also eribrately punctured, transversely impressed in front of eyes, clypeus bilobed in front; antennae extending well back of middle of prothorax, third segment as long as two following united, eradually clavate outwards, ninth and tenth segments transverse. Prothorax fully two-fifths broader than long, apex broadly emar- ginate, base broadly lobed at middle, sides almost straight and feebly diverging from base to middle and thence areuate and gradu- ally narrowed towards apex, hind angles almost rectangular; dise very convex, feebly longitudinally impressed at middle with well marked fovea in front of scutellum and coarsely as well as closely and eribrately punctured with tendeney of punctures near middle to longi- tudinally anastomose. Scutellum with a few obscure punctures. Elytra one-fourth longer than broad and over twice as long as prothorax, sides feebly arcuate from base to posterior third, thence more rounded and rapidly narrowing to apex; dise very convex, sulei broad and deep, the strial punctures rather close and round and moderately impressed, intervals narrowly costate with the third and fifth and seventh some- what more elevated than the others, the general surface shining yet with sulei alutaceous. Beneath very coarsely and umbilicately pune- tured in front and rather coarsely and somewhat closely punetured behind and smooth and shining. Length 6 mm., breadth 2.5 mm. Holotype, a unique collected on Indefatigable Island, October 25- 28, 1905, by F. X. Williams. No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 113 This species somewhat resembles the preceding but its elliptical shape, very coarse punctuation on the upper surface and beneath and the sharp earinae of all intervals should separate it. The specimen ap- pears to be slightly immature to judge from the color. Pedonoeces lugubris (Boheman) Plate V, figure 8 Tessaromma lagubris BOHEMAN, 1858, Fregatten Eugenies Resa, I, p. 91. Tafl. I, fig. 5. Pedonoeces lugubris Boheman, Bratr, 1933, Ann. Mag. Nat. Hist., ser. 10, XI, pp. 479-480. Large to moderate size, robust, somewhat elongate, dull black, with depressed and minute hair seattered over the head and prothorax and to a lesser extent the elytra; antennae rufous and legs rufopiceous. Head alutaceous, rather finely and discretely punctured, transversely impressed in front of eyes, elypeus broadly emarginate in front; anten- nae extending beyond middle of prothorax, third segment as long as two following united, four other segments gradually enlarged to form a loose club with the eighth and ninth segments triangular and the tenth and eleventh segments transverse. Prothorax a little less than one-third broader than long, apex broadly emarginate with anterior angles obtusely projecting, base bisinuate with broad median lobe, hind angles acute and moderately projecting backwards, sides broadly and evenly arcuate from base to apex, feebly narrowed to front; dise moderately convex, very feebly longitudinally impressed at middle, alutaceous and finely punctured. Elytra somewhat spatulate, almost two-fifths longer than broad, twice as long as prothorax and dis- tinetly narrower at base than base of prothorax, sides feebly arcuate, eradually wider to beyond middle thence evenly rounded to apices, dise convex though somewhat flattened medially, sulci broad and deep, the strial punctures sparse and fine and also obscure, inter- vals narrowly carinate and equally elevated, and more or less evi- dently punctured at apices with minute punctures, the general surface dull and alutaceous as are the head and pronotum. Be- neath dull, coarsely, closely, umbilicately punctured and rugose in front, less coarsely punctured behind as far as last abdominal segment which is still more finely punctured. Males much smaller than females with front femora dilated and anterior tibiae broadly and triangularly dentate within at middle and with a few short spines in front of denta- tion, the abdomen longitudinally impressed at middle, and last seg- ment broadly and deeply impressed and rounded at apex ; females with 114 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS all tibiae normal, the abdomen not impressed in front and last ventral segment impressed only near apex and with latter somewhat blunt. Males, length 6.5 mm., breadth 2.5 mm., females 8-10 mm. length by an average of 3 mm. wide. The specimen so well illustrated by Boheman was evidently a large- sized female. The species was also listed by Boheman as from Panama but according to Blair was probably so listed in error, the real locality being one of the Galapagos Islands. Blair listed nine specimens from Eden Island, a small island near Indefatigable, and one from Inde- fatigable. The California Academy has two specimens from Inde- fatigable Island, including one fully as large as that figured by Bohe- man, one collected on Abingdon Island, September 8-23, 1906, and a large series, eighty specimens of average size, from various places in Albemarle Island: Villamil, Iguana Cove and Bank’s Bay and a series of nine much smaller specimens, 5 mm. in length, from Albemarle Island, collected in March and April, 1906. ‘With these smaller speci- mens I have associated one from Hood Island, collected September 24-30, 1905, not appreciably distinguishable. All of the above were collected by F. X. Williams except the largest female from Indefatigable which was collected May 4, 1932, by M. Willows, Jr., of the Templeton Crocker Expedition. This sooty black species is a very distinct one and apparently only approached by two species, possibly offshoots isolated each on one of the smaller islands, P. duncani on Dunean Island and P. opacus on Jervis Island. Pedonoeces opacus Van Dyke, new species Similar in general appearance to the above but both sexes are of about the same size as the males of P. lugubris, the differentiating features being that the punctuation of the head in P. opacus is coarse and much denser; that of the pronotum very much coarser, approxi- mate, and somewhat aciculate; the hind margin of prothorax less deeply emarginate near angles thus reducing the prominence of the middle lobe and the extent to which the angles themselves project backwards; the carinae of the elytra less sharply carinate and more evidently and irregularly crenulate at apices, and the general surface duller and in general less evidently alutaceous. In the males, the front femora are dilated and angularly arcuate along the outer mar- gin as in P. lugubris but in the front tibiae are normal; there is also often a tuft of pile on the posterior face of the middle femora near the base which seems to be absent in the other. Male, length 6 mm., breadth 2.5 mm., female, length 6.5 mm., breadth 3 mm. No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 115 Holotype male, allotype female and two paratypes, male and female, all collected on Jarvis Island, December 18-20, 1905. The coarser sculpturing of head and pronotum, less sinuation of hind margins of prothorax, less sharply defined elytral carinae, and generally duller appearance should place this species apart from P. lugubris. Key TO SPECIES OF GENUS PEDONOECES G. R. WATERHOUSE 1. Elytral intervals flat or feebly elevated at most, striae fine and shallow; sparsely pilose; rather small and with rudimentary wings...................... 2 — Elytral intervals well elevated, convex or carinate, striae deeply im- DRESSCUMOTEUSUICACCR sy sree Ati ce IA Slt A A! ee Rg ee eg 4 2. Pronotal punctures very fine and well separated, elytral striae quite dis- tinct and rather coarsely, closely punctured; general surface alutaceous and finely, sparsely pubescent........................ P. pubescens G. R. Waterhouse — Pronotal punctures rather coarse, deep and close, elytral striae finely impressed and rather finely punctured; general surface more uniformly IDINDES Cem iieere weer ele A ee So sae le A eh SS ns A 3 3. Strial punctures variable but those of intervals almost as evident as those COTES Tall Cnet ee ie et et es I dd he P. wenmani, new species — Strial punctures distinct and close, those of intervals very fine, elytra TORCNOHELES SHS MOM See ee ee es P. culpepperi, new species 4. Elytral intervals merely convex, not sharply carinate.......................-...--- 5 Eee livid imac eravell Smeivgl Ge MUG liye GA TL T AUC lee nese eee me eee 10 5. The intervals but moderately elevated and convex, striae and strial punc- tures fine, general surface more or less pilose under magn....................- 6 — The intervals well elevated but convex or blunt at apices, striae broad TG COD ep UM CEURE SR COANS CLs sec2 ios ona oo foes ob ace ce nec oe onepeteecetecns eee eee, See sees 8 6. Prothorax two-ninths broader than long, sides somewhat arcuate, strial punctures of elytra fine and interstrial punctures very fine but distinct, elytralvapexamonmiats ste NS Ne ee! R. blairi, new species —Mlviraleapex: prolonezed or caudate im’ Lema] erie coats ener 7 7. Prothorax one-third broader than long, pronotum densely punctured but with punctures not contiguous, the lateral margins very distinct, rather broad, strial punctures of elytra rather coarse. Hood Island__............. P. caudatus, new species — Prothorax two-fifths broader than long, pronotum densely punctured and punctures more or less contiguous, the lateral margins fine, strial punc- tures of elytra fine and shallowly impressed. Tower Island........................ P. apicalis, new species 8. Pronotum somewhat coarsely, closely punctured, strial punctures approxi- 116 10. 13. 14. 15. 16. il7t CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS mate, intervals very finely punctured, surface minutely pubescent............ or aoe A SERPENT SPL NES SSE Ne POPES oN) es, i CUE ET he 2 P. uniformis, new species Pronotum very finely, not closely punctured, strial punctures coarse and well spaced. surface slabrouse see 9 Prothorax one-third broader than long, punctures fine, surface dull and alutaceous, sides more or less straight and parallel behind; elytra with sides parallel in front, intervals rather unequally elevated and always well rounded at summits; male femora simple...................... P. bauri Linell Prothorax less than one-third broader than long, punctures fine but deep, surface shining, sides sinuate behind, elytra subspatulate, intervals equally elevated and subcarinate; male femora with patches of erect golden ype tere. ta eee a he ee ree ee eaten at P. galapagoensis, G. R. Waterhouse IDIyaceyl Sot ohisbaerAyy Toon AAWTES ecnsteceeene Li IDMAmeyL SOULE eaten e ne GMS GLE TON AV ELUNE 17 Somewhat flattened, narrow and elongate species, pronotum rather Coanseliyarversyy Gd ense liv. in TMG CULT este eee ici een ne eee 12 More or less elliptical and conspicuously convex species....................-.---- 13 Shining, pronotal punctures dense and sharply impressed, elytral in- tervals unequally-elevated= = P. spatulatus, new species Subopaque, pronotal punctures close but variolate, elytral intervals Equally elevate discs ee See ene P. barringtoni, new species Pronotum alutaceous and rather finely not closely punctured; elytral INTERVAISHequUallivyael | wate sina rier; oma teeceee eee ee eee een een nn 14 Pronotum coarsely sculptured, the punctures large and more or less ANASCOMOSING 32.32). 7 Aesth: Sat ee | ly oo ed ee ee 15 Strial punctures fine, sulci and intervals obtuse angulated_.................. Bp. Cente atte a te AP CBee CaM De IRR EPR RAMON ot Eres Py CBE IE LEENA Sal nt A eg P. batesoni Blair Strial punctures coarse, sulci broad and deep with narrow intervals... BP ts Re ite Ni: 4s ele A Or PS, Oe) lied MAB bese oP morio, (boneman) Elytral intervals equally and sharply elevated and carinate___................... yeas AAS ME RENN a 5 fi on PLE py 07 NO eee gh 2 P. duncani, new species Bylvtirall sini erviail sure qual lilcy7 ag eleven tee Clee eae ce 16 Pronotum somewhat strigose as result of longitudinal anastomosing of punctures, punctures of head discrete; abdominal punctures rather fine... Eee aes SNE A TGLN, St: aes, val UE ere SRE EA Wenig Noe ees P. costatus G. R. Waterhouse Pronotum and head very coarsely punctured, punctures somewhat con- fluent but pronotum strigose, abdominal punctures coarse and close........ By Rec UI ereece eaer RR Ie BNNs CO occ oh Me REA Beare Ue yom AE RU eR P. williamsi, new species Sooty, black species, pronotum very finely and closely punctured, hind angles somewhat acute, elytral costae sharply and smoothly defined, abdomen shaliowly punctured and rugose, females often quite large........ EE a ee Eo ae SAE P. lugubris (Boheman) Dull black, opaque, the pronotal punctures coarse, contiguous and longi- tudinally anastomosing so as to give a strigose appearance; abdomen somewhat discretely punctured.............00.0.00000000.-0.----- P. opacus, new species No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 117 Phaleria manicata Boheman Plate VI, figure 1 Phaleria manicata BonEMAN, 1858, Fregatten Eugenies Resa, I, p. 92. Phaleria manicata Boheman, C. WATERHOUSE, 1877, Proc. Zoo. Soc., V, p. 82. Phaleria manicata Boheman, LINELL, 1898, Proc. U. S. Nat. Mus., X XI, p. 266. Phaieria manicata Boheman, MurcH ier, 1925, Zoologica, V, no. 20, p. 237. Robust, ovate, somewhat convex, testaceous with variable black markings, upper surface smooth and alutaceous, and lateral margins of prothorax and elytra densely fimbriated with long yellowish pile. Head very finely, sparsely punctured and transversely impressed in front of eyes. Prothorax over one-third broader than long, apex broadly emarginate, base somewhat sinuate with feeble median lobe, sides al- most straight and parallel or feebly sinuate at basal half and mod- erately arcuate and narrowed forwards, the dise finely and sparsely punctured and with a short longitudinal impression in front of seutel- lum and small fovea between impression and lateral angles. Elytra one- fifth longer than wide and three times as long and one-fourth broader at middle than prothorax, sides feebly arcuate from base to posterior third and then gradually rounded to apex, dise with striae well im- pressed, intervals convex and finely and sparsely punctured. Be- neath unicolorous testaceous, smooth, finely alutaceous and very mi- nutely and sparsely punctured. The feet robust as usual in genus, femora punctured and setaceous, tibia broadly dilated outwardly, the apex lobed on outer side, coarsely punctured, more or less setose and densely set with short stubby spines. Length 7 mm., breadth 4 mm. This robust species is quite variable in color and color pattern, ranging from almost entirely testaceous above to almost entirely black, the lateral margins alone being testaceous. The usual color pattern is for the pronotum to be margined with black at apex and have a broad band at base; the elytra to have a broad transverse band near base, which is continued backwards two intervals wide, along the suture to near the middle where it becomes greatly expanded, often covering most of the disc, then narrowed to a sutural vitta as it approaches the apex. Boheman’s specimens are cited as coming from the Galapagos Islands without indicating any particular island. The California Academy of Sciences has a series of twenty-six more or less maculate specimens from Banks Bay, Albemarle Island, collected April 10-17, 1906, by F. X. Williams. This series I consider to be typical. Besides these, we have mounted a series of sixty-seven specimens of a prac- 118 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS tically immaculate phase from.a series of several hundred specimens from Hood Island, collected during January and February, 1906, by F, X. Williams and a set of twelve_smaller specimens with a small fuscous blotch at the middle of the elytra, four from Barrington Island, October 19-24, 1905, one from Charles Island, October 3, 1905, and six from Albemarle Island, March, 1906. These I consider but color phases of P. manicata. The genus is a wide spread one with repre- sentatives along the seacoasts of the continents of both the Old and New World. Gnathocerus cornutus (Fabricius) Trogosita cornuta Fasrictus, 1798, Supplem. Entomologica Syst., Hafnia, p. 51. Gnathocerus cornuta (Fabricius), THUNBERG, 1814, Beshr. p. t. nya ins. sl. Gnathocerus etc., Vetensk. Acad. Handl. p. 47. Gnathocerus cornutus (Fabricius), LINELL, 1898, Proc. U. S. Nat. Mus., XXI, p. 266. Gnathocerus cornutus (Fabricius), MuUTCHLER, 1925, Zoologica, V, no. 20, p. 237. This cosmopolitan grain beetle which has an extensive bibliography of which I have given only the essential items, was first collected in the Galapagos Islands on Albemarle Island by Dr. B. Baur and recorded by Linell. Two specimens were also taken on Albemarle Island by F. X. Williams, in March 1906, three on James Island, January 5, 1906, and seven at Villamil, 8. Albemarle, August 20, 1906. At the last location there is a permanent settlement so it is not surprising to find this cosmopolitan grain beetle there. Alphitobius laevigatus (Fabricius) Carabus laevigatus Faprictus, 1781, Species Insect., I, p. 304. Helops piceus Ouivigr, 1792, Encycl. Method. Insect., VII, p. 50. Alphitobius picipes STEPHENS, 1832, Ill. Brit. Mandib., V, p. 11. This beetle, now widely distributed throughout the world by ecom- merece, common in the Hawaiian Islands, was collected on Charles Island, May 23, 1906, by F. X. Williams. Eleven specimens were taken. Rhacius costipennis Blair Plate VI, figure 3 Rhacius costipennis Buatr, 1933, Ann. Mag. Nat. Hist., ser. 10, XI, p. 480. This species has been but recently described and will not be con- sidered further here. No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 119 The California Academy of Sciences possesses one specimen col- lected on Indefatigable Island, May 5, 1932, by M. Willows, Jr., of the Templeton Crocker 1932 Expedition. This specimen was care- fully compared with the type by Dr. Blair and labeled as the same. The undersurface is bright rufous, contrasting strongly with the upper surface. There are but three described species in the genus. Genus Prateus Le Conte Prateus LECONTE, 1862, Class. Col. N. Amer. (Smith. Contr.), p. 238; 1866, New Spec. N. Amer. Col. (Smith. Contr.), p. 131. Lorelus SHARP, 1876, Entom. Mo. Mag., 13, pp. 76-77. The type of Proteus is P. fusculus Le Conte, from the middle and southern states of the United States; the type of Lorelus Sharp is L. priscus Sharp from New Zealand. According to Champion in notes in the British Museum of Natural History Collection, they are con- generic. Besides P. fusculus Le Conte, originally described from Texas, the genus contains numerous species described under Lorelus from Central and South America, New Zealand, and there are in the British Museum of Natural History collection, many undescribed species from southern Asia, Africa, and elsewhere. The general facies of these small members of the Tenebrionidae is that of the larger Cryptophagidae. Prateus dentatus Van Dyke, new species Small, moderately elongate, more or less flattened, rufopiceous above, humeri lighter, the legs and underside rufocastaneous. Head rather coarsely and closely punctured, punctures almost contiguous in front, finely alutaceous, with feeble transverse impression in front of eyes; antennae almost reaching hind angles of prothorax, grad- ually and feebly clavate, third segment fully twice as long as second, fourth but little longer than broad, the fifth to tenth gradually more transverse and broader; eyes moderately convex, coarsely faceted, transverse, widest above and truncate posteriorly. Prothorax about a fourth wider than long, apex broadly emarginate and about one- fifth wider than base, base broadly lobed, the hind angles right and prominent; sides margined, feebly divergent from base almost to middle, thence arcuate and gradually rounded to apex and provided with four well marked denticles, two just behind middle, one-half way from middle to apex and the fourth between this and apex; the dise feebly convex, coarsely and closely punctured and with slight antescutellar impression and finely alutaceous. Scutellum moderate in size, transverse and feebly punctured. Elytra about twice as long 120 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS as broad and three times as long as prothorax, humeri rounded but almost rectangular, the sides gradually yet very feebly divergent from base to posterior third and thence evenly rounded to acute apices, the lateral margin well defined; dise somewhat flattened, coarsely and closely punctured, very sparsely and finely pilose, the individual fulvous hairs arising from the punctures, the humeral umbones prom- inent and without evidence of striae. Beneath coarsely and closely punctured in front and subopaque, more finely and distinetly pune- tured behind and shining, and with very sparse and minute fulvous pubescence. Length 4 mm., breadth 1.5 mm. Holotype, a unique collected on Indefatigable Island, July 20-24, 1906, by F. X. Williams. This small yet interesting species upon comparison with the ex- tensive series in the British Museum of Natural History Collection was found to differ from all by having the sides of the prothorax more broadly rounded, the margins dentate, and the third segment of the antennae about twice as long as the second. Family ANOBIIDAE The California Academy of Sciences Expedition secured but two species of this family. The St. George Expedition of 1891, collected four species, two at light and the others from the herbage and rotting wood. These were deseribed by Dr. Blair. Trichodesma denticollis Blair Trichodesma denticollis BLAtr, 1928, Ann. Mag. Nat. Hist., ser. 10, vol. i, p. 675. This species is not represented in the collection of the California Academy of Sciences. Thaptor galapagoensis Blair Thaptor galapagoensis Buatr, 1928, Ann. Mag. Nat. Hist., ser. 10, vol. i, p. 676. The California Academy of Sciences possesses one specimen col- leeted at Academy Bay, Indefatigable Island, in January, between 18-22, 1906, by F. X. Williams. Eupactus georgicus Blair Eupactus georgicus BuLatr, 1928, Ann. Mag. Nat. Hist., ser. 10, vol. i, p. 676. The California Academy has one specimen from Indefatigable Island, collected in November 17-19, 1905, by F. X. Williams that is possibly this species. No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 121 Kupactus alutaceus Blair Eupactus alutaceus Buarr, 1928, Ann. Mag. Nat. Hist., ser. 10, vol. i, p. 677. This species is not represented in the collection of the California Academy of Sciences. Family BOSTRICHIDAE Tetrapriocera longicornis (Olivier) Apate longicornis OLIVIER, 1795, Ent. IV, nr. 77, p. 15, t. 3, f. 18. Tetrapriocera schwarzi Horn, 1878, Proc. Am. Phil. Soc., XVII, p. 545, f. Tetrapriocera longicornis Olivier, LINELL, 1898, Proc. U. S. Nat. Mus., XXI, no. 1143, p. 256. Tetrapriocera longicornis Olivier, MuTcuier, 1925, Zoologica, V, no. 20, p. 237. Linell makes the following remarks concerning this species: ‘‘One example taken on Indefatigable Island by the Albatross Expedition in 1888. The species is distributed from southern Florida and West Indies to Central and South Amerieca.’’ Mutehler in 1925 refers to this note but adds no new information. In the California Academy of Sciences Collection, is one specimen collected on Dunean Island, on December 1, 1906, by F. X. Williams. This specimen, I checked with a specimen of Telrapriocera tridens Lesne (not Fabricius), a synonym of I. longicornis (Olivier), and found it to agree perfectly. It is also in agreement with three specimens labeled ‘‘longicornis’’ which we have from Florida. Micrapate scabratus (Hrichson) Rhizopertha scabrata Ericuson, 1847, Wiegm. Arch. fur. Naturg., XIII, 1, p. 87. Bostrychulus scabratus (Erichson), Lesnr, 1898, Ann. Soc. Ent. Fr., LXVII, pp. 596, 612, f. 221. Micrapate scabrata (Erichson), Lesnr, 1938, Junk and Schenkling Coleopt. Cat., pars. 161, p. 46. A single specimen from Chatham Island, collected May 23-29, 1906, by F. X. Williams, was compared in the British Museum of Na- tural History, with a specimen determined by Lesne as Micrapate scrabatus (Erichson) and found to agree perfectly. This is a wide- spread species, being listed from Peru, Bolivia, and Chile. Amphicerus cornutus galapaganus [esne Amphicerus cornutus galapaganus LESNE, 1910, Bull. Mus. Paris, pp. 184-186. Apate species G. R. WATERHOUSE, 1845, Ann. Nat. Hist., XVI, p. 36. 122 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS Amphicerus punctipennis LeConte, LINELL, 1898, Proc. U. S. Nat. Mus., XXI, no. 1143, p. 256. Amphicerus cornutus galapaganum Lesne, MUTCHLER, 1925, Zoologica, V, no. PADS Ty) CE This insect was first cited by G. R. Waterhouse in 1845, from Dar- win species. Later these same specimens were studied by Lesne and pronounced a subspecies. The species ‘‘cornutus’’ is widely distributed throughout North and South America. The weak subspecies is sup- posedly confined to the Galapagos Islands. In the material collected in the Islands during 1905-1906 by the California Academy Expedition, there is a series of seventy specimens, collected on the following islands: Hood, February, 1906; Abingdon, September, 1906; Wenman, September 24, 1906; Duncan, December, 1905; Albemarle, April 30, 1906; and Indefatigable, January 11-22, 1906. The specimens from Hood and Indefatigable are the largest. The series is quite uniform as to sculpturing and general appearance. When compared with typical specimens Amphicerus cornutus (Pallas) of which we have one hundred and twenty-eight from various places in southern California, Arizona, Texas, Utah, Florida, Lower California, western Mexico, and Hawaii, members of the subspecies appear to be smoother, more shining, with the granulations over the basal portion of the pronotum less pronounced and the elytral sculpturing less rugose and the punctuation less coarse and better spaced. Family SCARABAEIDAE This important family is but poorly represented in the Galapagos Archipelago, probably because of the general aridity of much of the region. Copris lugubris Boheman Copris lugubris BOHEMAN, 1858, Fregatten Eugenies Resa, I, p. 42. Copris lugubris Boheman, C. WATERHOUSE, 1877, Proc. Zoo. Soc. Lond., V, p. 82. Copris lugubris Boheman, LINELL, 1898, Proc. U. S. Nat. Mus., XXI, no. 1143, p. 258. Copris lugubris Boheman, MutTcuHuEr, 1925, Zoologica, V, no. 20, p. 237. This species has not been collected subsequent to the voyage of the ELugemeé in 1858. It may possibly never have been taken on the Gala- pagos Islands, for many of the species cited by Boheman were no doubt given erroneous localities through the carelessness of the collectors as has been frequently pointed out. No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 123 Ataenius arrowi Hinton Ataenius arrowi HINTON, 1936, Ann. Mag. Nat. Hist., ser. 10, 17, pp. 414-416, f. 1-4. Ataenius cribrithoraz Buatr, 1933, Ann. Mag. Nat. Hist., ser. 10, XI, p. 476. The California Academy of Sciences has eleven specimens of this species: two from Tagus Cove, Albemarle Island, Mareh 22, 1905, F. X. Williams, collector; two from Indefatigable Island, May 5, 1932, M. Willows, Jr., of the Templeton Crocker Expedition, collector ; seven from Abingdon Island, September 18-23, 1906. This species is very close to A. cribrothorax. One of my Albemarle specimens has evident though reduced rugae at the sides of the head in front. Ataenius scutellaris Harold This beetle which is found in Mexico and the West Indies and ex- tends well into South America, is characterized in the main by having the head coarsely punctured towards the base and gradually more finely so forwards and with the punctures fading out medially towards the front; with the pronotum convex and rather coarsely and densely punctured towards the sides and more finely towards the center; the seutellum suleate and with a median longitudinal carina; and the elytra quite convex, with the humeral angles rectangular and toothed, the striae deeply impressed or suleate, and somewhat obscurely pune- tured and the intervals greatly elevated, with an irregular row of pune- tures on either side and the median portion more or less ecarinated especially towards the sides. The California Academy of Sciences has a series of fifty-two speci- mens, fifteen collected on Charles Island, April 30, 1906, twenty-four from Abingdon Island, September 8-23, 1906, one from Albemarle Island, April 24-27, 1906, and twelve from Chatham Island, January 1906, all by F. X. Williams. Trox suberosus Fabricius Trox suberosus FAasrictus, 1775, Systema Entom., p. 31. Trox suberosus Fabricius, MurcHuEr, 1925, Zoologica, V, no. 20, pp. 229, 238. This rather large and common species of eastern North America which has been recorded from Central and South America as well as ‘the West Indies and the Cape. Verde Islands, was first reported from the Galapagos Islands by Mutchler, his single specimen having been taken on Tower Island, April 28, 1923, by the Harrison Williams Gala- pagos Expedition of the New York Zoological Society. The California 124 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS Academy of Sciences has twenty-one specimens: seven collected on James Island in December, 1905; one from Albemarle Island, in April 1906; three from Charles Island in April, 1906; two from Chatham Island, February, 1906; one from Hood Island, April 22, 1932, by M. Willows, Jr., of the Templeton Crocker Expedition of 1932; one from Abingdon Island, September 18-238, 1906; and five from Villamil, S. Albemarle Island, March 4-14, 1906, as well as two from Cowley Mountain, Albemarle, August 9-13, 1905, the majority collected by F. X. Williams. This species has well developed wings, hence the wide distribution in the Islands could be accounted for. Trox seymourensis Mutchler Plate VI, figure 8 Trox seymourensis MUTCHLER, 1925, Zoologica, V, no. 20, pp. 229-230, 238. In the California Academy of Sciences collection there are fifty-six specimens, most of which were collected at Villamil, S. Albemarle, August 20, 1906, by F. X. Williams, but there is one simply labeled ‘* Albemarle Island, April 1906,’’ another ‘‘Cowley Mt., Albemarle Island, July, 1906,’’ and a well-worn elytron from Abingdon Island, picked up September 18-23, 1906. This species has well developed wings. The abundance of specimens from Villamil can be accounted for by the fact that there was a settlement there and some slaughtering carried on. Our Cowley Mountain specimen was carefully compared with Mutchler’s type, both by Mr. Mutchler and myself, and found to be typical. Trox galapagoensis Van Dyke, new species Plate VI, figure 7 Oblong, black, more or less covered with an earthen-colored in- dument. Head feebly convex, prolonged in front to an obtuse angle, slightly depressed at apex, sides also strongly angulate, the margins fimbriated and surface irregularly studded with short setae. Prothorax two-fifths broader than long, apex broadly lobed at middle and with the angles prominent and projecting well forward, the base broadly arcuate, sides lobed, feebly emarginate in front of right-angled hind angles, suddenly constricted at base and margined with short setae; the dise longitudinally and shallowly suleate at middle with an obtuse and sinuous ridge on either side, an elongate tubercle near hind angles, and a deep and oblique impression in front of this, the surface covered with a dense earthy colored indument through which project short and scattered setae. Secutellum one-third longer than broad and broadly No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 125 suleate in front. Elytra about one-seventh broader than long, with sides broadly arcuate from base to apex and margined with short widely spaced setae; the disc evenly convex, with ten well elevated intervals, the odd the most prominent, and each surmounted with a series of small tubercles from the apex of which projects a short seta; the striae with a series of large, deeply impressed pit like punctures with small tubercle-like elevated anterior and posterior margins. The legs much as in T. seymourensis but the front tibia with lateral spines more equal in size and closer together. The true wings appear to be little more than one-half normal size therefore not functional. Length 10 mm., breadth 6.5 mm. Holotype, a unique, collected Culpepper Island, in September 1905, by F. X. Williams. This species is of about the same size as 7. seymourensis but differs by having a very different type of elytral sculpturing as indicated by the more regular elevation of the elytral intervals surmounted by small crater-like tubercules and the regular arrangement of the strial punctures, in 7. seymourensts, striae and strial punctures appear to be absent. Neoryctes Arrow Neoryctes Arrow, 1908, Trans. Ent.-Soc. Lond., p. 342. Parapseudoryctes MUTCHLER, 1925, Zoologica, V, no. 20, pp. 237-238. Pseudoryctes LINELL, 1898, Proc. U. S. Nat. Mus., XXI, no. 1143, p. 258. Oryctes Illiger, G. R. WATERHOUSE, 1845, Ann. Nat. Hist., XVI, pp. 26-27. The genus Neoryctes was proposed by Arrow to replace Pseu- doryctes Linell, the name of which was found to be preoccupied. Mutchler, not seeing the citation, later on also proposed a new name, Parapseudoryctes. Neoryctes of course has priority. Linell, however, was the first to note that the species ‘‘galapagoensis’’ of G. R. Water- house was very different from those associated under the Old World genus Oryctes, the genus to which Waterhouse assigned it, so erected the genus Pseudoryctes to receive it. Neoryctes galapagoensis (G. R. Waterhouse) Plate VI, figure 5 Oryctes galapagoensis G. R. WATERHOUSE, 1845, Ann. Nat. Hist., XVI, pp. 26-27. Oryctes galapagoensis G. R. Waterhouse, C. WATERHOUSE, 1877, Proc. Zoo. Soc., V, p. 82. Oryctes galapagoensis G. R. Waterhouse, Howarp, 1889, Proc. U. S. Nat. Mus., NNO da Da Lo. Pseudoryctes galapagoensis (G. R. Waterhouse), LINELL, 1898, Proc. U. S. Nat. Mus., XXI, no. 1143, p. 250. 126 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS Neoryctes galapagoensis (G. R. Waterhouse) Arrow, 1908, Trans. Ent. Soc. Lond., p. 342. Parapseudoryctes galapagoensis (G. R. Waterhouse), MuTcHuer, 1925, Zoolo- gica, V, no. 20, pp. 237-238. Mr. Waterhouse does not indicate from what island the species that he described came from. Linell states that the Albatross Expedition in 1888 collected one female on Chatham Island, the female that he undoubtedly used for drawing up his generic deseription, and that Dr. G. Baur obtained six males on the same island. No other specimens seem to have been collected between that date and 1938 when Mutchler separated out his species, N. linelli, from the known specimens. In 1905-1906, F. X. Williams collected fourteen specimens including both males and females, which run to N. galapagoensis, these came from the following localities: Chatham Island, July, 1906; Charles Island, March, 1906; and Cowley Mountain, Albemarle Island, July, 1906. These specimens range in size from 15 mm. to 17 mm. in length. The wings, though present, are much reduced in size and nonfune- tional. Besides these are four large females, averaging 28 mm. from Chatham Island, July, 1906, that seem to fit Mutchler’s species, N. linelli. Neoryctes linelli Mutchler Neoryctes linelli MuTCHLER, 1938, Am. Mus. Novitates, no. 981, pp. 10-11, figs. 10, 11 and 12. The California Academy of Sciences has four specimens from Chatham Island, collected in July, 1906, by F. X. Williams. Neoryctes sp. ? Two specimens of the same size and general appearance as N. gala- pagoensis were collected on Indefatigable Island, one on May 2, 1932 by M. Willows, Jr., of the Templeton Crocker Expedition, the other on March 1906, by F. X. Williams. These are much smoother, more shin- ing, in general more finely punctured, with the prothorax wider, and with the sides more broadly rounded. These I have set aside for further study. They may be new or merely a variety of N. galapagoensis. Family PASSALIDAE Passalus interruptus Linnaeus Passalus interruptus LINNAEUS, 1754, Mus. Adolph. Frieder; p. 82. Neleus tlascala PERCHERON, 1835, Monogr. des Passales, etc., p. 35, t. 3, f. 5. No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 127 Neleus tlascala Percheron, LINELL, 1898, Proc. U. S. Nat. Mus., X XI, no. 1143, p. 257. Neleus tlascala Percheron, MuTcHLER, 1925, Zoologica, V, no. 20, p. 238. Passalus interruptus Linnaeus is a rather large species which ranges from Texas to Argentine. Under the name Neleus tlascala Percheron, a synonym of the above, Linell lists a specimen as having been taken by the Albatross Expedition in 1891, on Charles Island. No specimens have been taken since. Numerous specimens collected on Cocos Island, by F. X. Williams, are much smaller than P. interruptus, and have been proven to be Popilius lenzi Kuwert. Family CERAMBYCIDAE This family, rich in South American species, has numerous repre- sentatives in the Galapagos Archipelago, some of which are also to be found on the South American mainland but most are restricted to the islands. Parandra galapagoensis Van Dyke, new species Male: rather large, elongate, subeylindrical, smooth and shining, dark rufous with legs and undersurface of afterbody a lighter color; the greater part of the head including most of the mandibles, eyes, and sides of head more or less black, the pronotum generally with an almost complete and fine marginal black line, likewise the apices of femora, tibiae, and tarsal segments black or at least dark in color. Head transverse, front feebly convex, sparsely punctured, a well marked triangular impression at middle of front margin and the post- ocular region coarsely punctured and rugose ; mandibles large, arcuate, as long as head, notched at apices, inner margin with a blunt tooth one-fourth distance back from apex and three or four smaller yet evi- dent blunt teeth near middle, the entire surface also rather finely and sparsely punctured though more densely so towards apices; antennae almost reaching hind angles of prothorax, segments 4-6 triangular and as long as broad, 7-10 also triangular but longer than broad and eleventh fusiform and about three times as long as broad; eyes mod- erately prominent, evenly arcuate in front, suddenly constricted be- hind causing them to stand out prominently ; submentum very coarsely punctured with punctures more or less anastomosing, the front margin smooth and flattened and anterior angles prominent and smooth except for a series of small punctures. Prothorax three-eighths wider than long at middle, narrower than head across eyes, apex broadly emarei- 128 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS nate, base feebly arcuate; sides, from obtusely rounded hind angles, divergent and oblique almost to middle, then straight and parallel to apex, the marginal bead broader posteriorly ; dise flattened above and minutely, sparsely punctured. Scutellum cordiform and smooth except for a few small punctures. Elytra almost twice as long as broad, 15 mm. to 8 mm., three times as long as prothorax, the sides almost straight and parallel to posterior third thence arcuate and narrowed and sud- denly rounded to apices; dise smooth and practically impunctate, the few very minute punctures only observed under considerable magnifica- tion. Beneath, the base of head and prothorax finely, sparsely punc- tured, the meso- and metasternum finely punctured and finely pilose and the abdomen rather coarsely, shallowly punctured and dull espe- cially towards sides and apex. Wings large and fully developed. Length 28 mm., breadth 8 mm. Female: Similar to male as far as color and general appearance go but with head smaller, frontal punctures finer and sparser, the post- ocular punctures also sparse and well separated ; the mandibles short, two-thirds length of head, feebly arcuate. Emarginate at apex with pronounced tooth at inner part of notch, a vague tooth at middle of inner surface and two pronounced teeth, united at base near base of inner margin; the antennae somewhat shorter and with basal segments more transverse; and the submentum rather coarsely, sparsely punc- tured. Prothorax four-thirteenths broader than long, with sides rather evenly arcuate and feebly narrowed forwards. Length 23 mm., breadth 7 mm. Holotype male, James Island, January 1906; allotype female, James Island, December, 1905; and several designated paratypes from a series of seventy-six specimens from James Island. There are also two rather small specimens from Indefatigable Island, January 11, 1906, and four somewhat larger females from San Tomas, Albemarle Island at 1200 feet altitude, collected September, 1906. Mr. F. X. Wil- liams who collected all specimens tells me that he chopped all specimens out of rotten logs, on the mountain tops. This species was compared with all Central and South American species in the British Museum of Natural History collection and not found to agree with any. It runs to P. brachyderes in Lameere’s (1902) key, but differs from a cotype of the same in the British Museum in that the upper surface is very smooth, rather finely and sparsely punc- tate, and has long curved mandibles in the male in contrast to the al- most straight mandibles in P. brachyderes which are without teeth along the inner border; by having large and confluent punctures be- hind the eyes in contrast to moderately coarse and well spaced ones; No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 129 by the antennae having the segments 46 as long as broad, 7-10 longer than broad and the eleventh at least three times as long as broad, the segments in brachyderes 4-8 being longer than as broad, 9-10 as long as broad, and the eleventh about twice as long as broad; the undersurface of the head coarsely and confluently punctured. In P. galapagoensis most of the undersurface posteriorly is quite smooth whereas it is well punctured anteriorly in the other. All the Lameere specimens of P. brachyderes come from Mexico. Stenodontes molarius (Bates) Mallodon molarium BATES, 1879, Biol. Centr. Amer., Col. V, p. 9. Mallodon molarium Bates, Howarp, 1889, Proc. U. S. Nat. Mus., XII, no. 771, DL oie Mallodon molarium Bates, LINELL, 1898, Proc. U. S. Nat. Mus., XII, no. 1143, D259. Stenodontes molarius (Bates), MutcHuirrR, 1925, Zoologica, V, no. 20, p. 238. As mentioned by Linell: ‘‘The Albatross expedition in 1888 col- lected on Charles, Chatham, and Auxican islands seventeen examples of this large Prionid, which is distributed from Lower California through Mexico and Central America to Panama. The species is amply winged.’’ The California Academy of Sciences has a series of twenty-four specimens from the Galapagos Islands, including six females, and from the following localities: Chatham Island, four, January, 1906, three, July 1906; Indefatigable Island, eight on January 2, 1906, and three in July, 1906; and Villamil, south Albemarle Island, six on March 22, 1906; all collected by F. X. Williams. The Villamil specimens are all considerably smaller than those from the other islands which are large, apparently of normal size. The Academy’s specimens, especially the large males, are characterized by having the mandibles curved, the inner face concave and the front of the head very coarsely, rugosely, and approximately punctured. Mexican specimens, with which they have been compared, generally have the mandibles straighter and the punctuation of the head less rugose and less approximate. The island specimens thus appear to be a variety or weak subspecies. A second species described by Mutchler, appears to me to be some- what pathological, especially as regards the mandibles which are so rotated as to have the inner face turned upwards. It also comes from Indefatigable Island from whence Mutchler received fourteen speci- mens and the California Academy eleven specimens of S. molarius. 130 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS Stenodontus (Mallodon) galapagoensis Mutchler Stenodontus (Mallodon) galapagoensis MUTCHLER, 1925, Zoologica, V, no. 20, pp. 11-12, plate, fig. 6. The California Academy of Sciences does not possess specimens of this species. Strongylaspis kraepelini Lameere Plate VII, figure 5 Strongylaspis kraepelini LAMEERE, 1903, Mem. Soc. Ent. Belg., XI, p. 28 (Re- vision Prionides) ; 1919, Gen. Insect. (Wytsman), fase. 72, p. 25, pl. 2, f. 4. Large, robust, subeylindrieal, dark reddish brown. Head of mod- erate size, slightly more than one-half width of prothorax, coarsely punctate-rugose, punctures somewhat finer behind eyes, larger pune- tures finely setiferous, narrowly longitudinally suleate at middle, the clypeus triangular and flattened and depressed in front and in common with base of mandibles clothed with rather long golden pile, genae prominent and dentate, extending somewhat forward ; mandibles short but robust and with well marked tooth near apex; antennae reaching the middle of elytra or a little beyond, more robust in male, first seg- ment robust. Clavate and with large punctures, generally well sepa- rated but often confluent; second segment small, as broad as long in female, slightly longer than broad in male, third segment equal to fourth plus one-half of the fifth in female, and in male equal to the fourth plus a third of fifth, the terminal segments distinetly striate ; eyes large and coarsely granular. Prothorax less than twice as wide as long, apex broadly yet feebly lobed at middle, front angies prominent, base broadly arcuate, sides with prominent and acute spine near base, sometimes dentate, then feebly arcuate and convergent forward to front angles and with margins irregularly serrate; dise convex, in female with general surface moderately coarsely and closely punctured and asperate, a well marked median groove extending from base two-thirds forward, the front portion move flattened and smooth and the smooth area connected laterally with a raised arcuate callosity which broadens out apically in the form of tubereules some distance before the front angles, the male with the punctures much coarser and more irregular, the median groove hardly evident, but the median smooth area he- coming an irregular transverse callosity, the median basal area also somewhat elevated and smooth and the anterior areuate line sharply elevated. Seutellum transverse and acutely granulate. Elytra less than two and one-half times as long as broad, almost five times as No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 131 long as prothorax, the dise rugose and punctate, with two well de- fined carinae besides broader and obtuse lateral carinae, the basal area granular and the sutural apices feebly dentate, the males with sculpturing more pronounced than in females. Length 30 mm., breadth 11 mm. Three female specimens of this fine prionid were collected on James Island, during December, 1905, by F. X. Williams. One of these was compared with specimens from the type material kindly sent to the British Museum for my use by Hans Gebien of the Hamburg Museum. I also have a male from the type locality, Ecuador. According to La- meere, this species belongs in the true subgenus Strongylaspis, with the third antennal segment longer than the fourth and one-half of the fifth segment combined, the antennae not attaining the posterior third of the elytra in the female nor the extremity in the male. He considers it as an admirable transitional form between Chiasmetes limae Gierin- Méneville and Strongylaspis of Central America. The Galapagos Island record for this species is its first record outside of the type locality record from Guyaquil, Ecuador. Achryson galapagoensis Linell Plate VII, figure 6 Achryson galapagoensis LINELL, 1898, Proc. U. S. Nat. Mus., XXI, no. 1143, p. 259. Achryson galapagoensis Linell, MurcuHLer, 1925, Zoologica, V, no. 20, p. 238. The California Academy of Sciences collection contains eighteen specimens, ten collected on Chatham Island, February 23, 1905, two on same island in July, 1906, and three collected on South Seymour Island, July, 1906, by F. X. Williams, and a small male collected on Charles Island, May 15, 1932, by M. Willows, Jr., of the Templeton Crocker Expedition of 1932. The series shows considerable variation in size and in the size and arrangement of markings, the Charles Island specimen not only being quite small but with the dark markings more evident than usual. This species which is quite distinct from the well known and widely distributed Achryson surinamum (Linnaeus), found throughout east- ern North America as well as much of South America, has hitherto been considered to be restricted to the Galapagos Islands. This is, how- ever, not the case for I found in the British Museum of Natural His- tory collection, fourteen specimens from the following localities: two from Eton, Peru; five from Chile; one from Quito, Ecuador ; one from Colombia; and five without locality labels; all of the above could not be separated from the Galapagos Islands specimens. A Peruvian 132 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS specimen was absolutely identical with the specimen that I took abroad for purposes of comparison. This South American lot bore the manu- seript name of A. lineolatum Chevrolat. It thus appears that this spe- cies is widely distributed throughout western South America as well as found in the Galapagos Islands. Eburia lanigera Linell Eburia lanigera LINELL, 1898, Proc. U. S. Nat. Mus., XXI, no. 11438, p. 259. Eburia lanigera Linell, MuTCHLER, 1925, Zoologica, V, no. 20, p. 238. Eburia lanigera Linell, BLAtr, 1933, Ann. Mag. Nat. Hist., ser. 10, XI, pp. 480-481. The St. George Expedition of 1924 obtained two specimens of this species, one from James Island and one from Eden Island. Dr. Blair after reporting these, comments as follows: ‘This appears to me to be only a variety of the Central-American and West-Indian E. stigma Oliv. Linell made no comparison with this species nor gave any indication of its position in the genus. Both Mr. Bateson’s specimens have much shorter spines on both femora and elytra than normal FE. stigma.’’ The California Academy of Sciences possesses nineteen specimens of this species, collected by F. X. Williams from the following locali- ties: three from Gardner Island near Hood, January, 1906; five from Chatham Island, February, 1906; two from Albemarle Island, one col- lected in December, 1905, the other April, 1906; one from Hood Island, November, 1905; four from Jervis Island, December, 1905; and four from Dunean Island, December, 1905. The entire lot are fairly uniform as to character. It is thus apparent that this beetle is the most common and widely distributed species of Eburia on the Islands. Eburia proletaria Erichson Eburia proletaria Ericuson, 1847, Arch. fiir Naturgesch., XIII, i, p. 140. Eburia proletaria Erichson, Buarr, 1933, Ann. Mag. Nat. Hist., ser. 10, XI, p. 481. Eburia bauri LINELL, 1898, Proc. U. S. Nat. Mus., XXI, no. 1143, p. 260. Eburia bauri Linell, Mutcuurr, 1925, Zoologica, V, no. 20, p. 238. Elongate, subeylindrical, reddish brown, rather densely clothed with appressed grayish yellow pile and scattered hairs, especially evi- dent on antennae, legs and upper surface. Head somewhat flattened between eyes; antennae long, in the male reaching four segments be- yond apex of elytra, scape robust, subearinate, two-thirds length of third, smooth and darker at apex, the third to eleventh filiform. Pro- No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 133 thorax slightly broader than long, somewhat more narrowed in front than behind, sides arcuate with prominent acute tooth at middle and blunt tubercle nearer apex than median tooth; dise convex, with sparse granules, a faint median carina behind middle and a prominent, some- what acute tubercle on either side of middle slightly in front of the center; the discal tubercles, the lateral teeth and tubercles and front margin black. Scutellum transverse and more densely pilose than elytra. Elytra less than three times as long as broad and less than three times as long as prothorax, convex, emarginate and bispinose at apex, the sutural spines the shorter; dise with granules rather numerous towards base, very sparse and irregular apically, a pair of light yellow elongate and elevated ivory spots at base of each elytron, each pair midway between humerus and scutellum and the inner spots about twice length of outer, and another pair of similar spots at about the middle and midway between suture and sides of each elytron, with the outer spots about three times length of inner, and in addition a black area posterior to anterior spots and more or less surrounding the pos- terior spots. Legs slender, middle and posterior femora bispinose at apices with inner spines about three times as long as short outer ones, the posterior femora extending slightly beyond apex of elytra. Length 23 mm., breadth 6 mm. The California Academy of Sciences has four specimens collected by F. X. Williams, including a typical male from Chatham Island, collected in February, 1906, which was carefully compared with speci- mens in the British Museum of Natural History. This served for the basis of the description given above. A second specimen, a large male, 28 mm. in length, was likewise collected on Chatham Island, February 25, 1906. This specimen was also lighter in color, more ochraceous, and without black markings of any sort. In general appearance, it agreed with many of the lighter specimens in the British Museum. A third specimen was from James Island, in December, 1905, this agrees per- feetly with the first-mentioned specimen. The fourth specimen was reared from wood collected in the Islands by F. X. Williams. It emerged in the rooms of the Department of Entomology, July 21, 1914. Later on it was sent to Washington and compared with the type of Eburia bauri Linell by W. S. Fisher and pronounced to be the same species. On the strength of this I have suppressed Eburia bauri Linell as a species. Dr. Blair mentions one specimen of F. proletaria as having been taken by C. L. Collenette of the St. George Expedition at Tagus Cove, Albemarle Island, which agrees in all essentials with examples from Peru. I examined this specimen ‘and also noted that in the British 134 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS Museum there is a large series of E. proletaria from many parts of western South America such as Guyaquil in Eeuador, Colombia, Peru, Bolivia, and Chile. When compared with FH. lanigera, E. proletaria stands out as larger and of a more rufous color, the other being domi- nantly piceous with gray pile, with the discal and anterior lateral tubercles of the prothorax more prominent and the apical spines of the elytra and the femoral spines more unequal. Eburia amabilis Boheman Eburia amabilis BOHEMAN, 1858, Fregatten Eugenie Resa, I, p. 150. Eburia amabilis Boheman, C. WATERHOUSE, 1877, Proc. Zoo. Soc., VI, p. 82. Eburia amabilis Boheman, How arp, Proc. U. S. Nat. Mus., XII, p. 192. This species has apparently never been taken since the voyage of the Hugenie or at least recognized since then. The California Academy of Sciences has no specimen of this species. Compsa apicalis Blair Compsa apicalis Buarr, 1933, Anns. Mag. Nat. Hist., ser. 10, XI, p. 481. No specimens of this species are in the California Academy of Sciences collection. Desmiphora hirticollis Olivier Desmiphora hirticollis OLiviER, 1795, Ent. IV, 68, p. 11, t. 4, f. 37. Desmiphora mexicana THOMSON, 1860, Class. Ceramb., p. 75. Desmiphora mexicana Thomson, Batss, 1886, Biol. Centr.-Amer., V, p. 116. Desmiphora hirticollis Olivier, BLatr, 1933, Ann. Mag. Nat. Hist., ser. 10, vol. XI, p. 482. This species, one specimen of which was taken by the St. George Expedition of 1924, on James Island at light (C. L. Collenette), and reported by Blair, is a common and widely distributed tropical-Ameri- can species occurring from Mexico and the West Indies to the Argentine. The California Academy of Sciences has no specimens from the Galapagos Islands. Estola galapagoensis Blair Plate VII, figure 4 Estola galapagoensis BuLAtr, 1933, Ann. Mag. Nat. Hist., ser. 10, XI, p. 482. The three specimens in the California Academy of Sciences col- lection were collected on Albemarle Island, one on December, 1905, the No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 135 others at San Tomas, altitude 1200 feet, Albemarle Island September, 1906, both by F. X. Williams. The December specimen was checked with the type and labeled as a paratype by Dr. Blair. Estola cribrata Blair Estola cribrata Buair, 1933, Ann. Mag. Nat. Hist., ser. 10, XI, p. 483. This species is lacking in the collection of the California Academy of Sciences. Estola insularis Blair Estola insularis BLAtR, 1933, Ann. Mag. Nat. Hist., ser. 10, XI, p. 483. The California Academy of Sciences possesses two specimens of this species, agreeing perfectly with the description and collected on Indetatigable Island, January 11, 1906, and James Island, January 5, 1906, both by F. X. Williams. Estola duncani Van Dyke, new species Rather small, elongate, yet closely knit, reddish brown, sparsely clothed with closely appressed rufous pubescence with here and there a few tufts of lighter color to give it a slightly maculated appearance, and with short, much inclined black setae somewhat uniformly though sparsely dispersed over the surface of the elytra. Head rather wide, somewhat flattened, coarsely but not closely punctured, with a narrow, median impunctate longitudinal line between the eyes; pubescence short and depressed and with a few short setae about the eyes and on the occiput ; antennae robust, annulated, the base of each segment pale. reaching to posterior third of elytra, third segment slightly shorter than fourth, the following gradually shorter, with a few hairs fringing the under surface. Prothorax transverse, convex above, with short lateral tubereules situated just behind the middle, dise with coarse and rather closely placed punctures from each of which arises an inclined seta, the intervening areas clothed with the maculated rufous and gray pile. Elytra three-sevenths longer than wide and three and a half times as long as prothorax, gradually narrowed from shoulders, with pune- tures coarser and less closely placed than on pronotum yet numerous and with the setae arising from them somewhat larger than on pronotum and black. Beneath with vestiture much as on upper surface but the punctures more widely spaced and giving the surface a spotted ap- pearance. Length 10 mm., breadth 4 mm. 136 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS Holotype, a unique, collected on Duncan Island, during Decem- ber, 1905, by F. X. Williams. This species is apparently somewhat similar to a number of un- named South American species, specimens of which I examined in the British Museum of Natural History Collection, yet different from any. Acanthoderes galapagoensis Linell Plate VII, figure 2 Acanthoderes galapagoensis LINELL, 1898, Proc. U. S. Nat. Mus., X XI, no. 1143, p. 261. Acanthoderes galapagoensis Linell, MurTcHLeER, 1925, Zoologica, V, no. 20, p. 238. This conspicuous cerambycid is apparently widely distributed throughout the Archipelago. Mr. F. X. Williams collected one speci- men on March 2, 1906, on Albemarle Island, a second at Villamil, Albemarle Island, August 20, 1906, and seven specimens in January, 1906, on James Island. Both sexes are represented. Leptostylus galapagoensis Van Dyke, new species Plate VII, figure 1 Rather small, robust, a reddish brown color which is mostly con- cealed above the short dense, closely appressed scalelike, chalky white pubescence, uniform except for a small brown patch of pubescence near suture and slightly in advance of apex and a smaller patch near elytral margins on a level with the hind femora, the punctuation of both pronotum and elytra conspicuous because of the punctures not being covered by the dense pubescence. Head with posterior margin and a sharply defined median longitudinal line extending from occupit to elypeus denuded of pile; eyes strongly reniform, well separated above and coarsely faceted; antennae long and delicate, extending four and a half segments beyond elytral apex in males, third segment shorter than fourth and slightly more robust, the following segments gradually shorter. Prothorax considerably more than a third broader than long with a short acute tubercle on lateral margin, slightly behind the middle, the sides feebly arcuate and convergent forwards from spine and straight and parallel to base behind spine; disc transversely impressed behind apex and in front of base, with two low tubercles, one on either side of middle, a short denuded median line between middle and base and rather densely and somewhat coarsely punctured, the punctures conspicuous because uncovered by pile. Seutellum un- clothed medially. Elytra over two-sevenths longer than wide, broad No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 137 at base with prominent humeri, gradually narrower behind; dise with prominence at middle of each elytron near base and a feeble ridge extending from humeri obliquely inwards almost to apex, the rest of surface densely and finely pubescent except for the uncovered con- spicuous punctures which are more or less regularly scattered over the surface, the punctures separated from each other by from two to several times their diameter. Beneath densely, uniformly clothed with fine white pile, the punctuation concealed. Legs stout. Length 10 mm., breadth 4.75 mm. Holotype male and several designated paratypes from a series of eighteen specimens collected at Villamil, South Albemarle Island, March 14, 1906, by F. X. Williams. Mr. Williams also collected one specimen on James Island, March 2, 1906, and one specimen on Jervis Island, December 18, 1905. This attractive chalky-colored species should be readily recognized. It is placed in Leptostylus for it agrees with the members of that genus in having the sides of the prothorax angulate behind the middle, the scape more or less cylindrical, antennae without cilia, broad meso- sternum and rather short hind metatarsus but differs from typical species by the fact that the sides of the prothorax are acutely tuber- culate, not bluntly tuberculate, and the mesosternum provided with two short posterior tubercles. Females are needed to see whether they have a pronounced exposed ovipositor or not. Almost no South American species of the genus are known though the group is well represented in Mexico. This and the following species also show certain relationships with Bate’s Atrypanuis and associates like Trypanidias, more strictly southern genera, but lack the elongated lower lobe of the eyes of the former and the long ovipositor of the latter. In Atrypanius we have a genus which strongly suggests our species because of the presence of the posterior lateral tubercles on the mesosternum but the eyes are very different. Leptostylus williamsi Van Dyke, new species Quite small, reddish brown, rather densely clothed with short and closely appressed scalelike pile which is in the main white; marked with a small transverse black marking at shoulders, a bare irregular and somewhat triangular patch at sides behind the middle, which reaches the outer margin but not quite to the suture, two small spots subapically, one along suture, the second near side margin, and a series of five or six black dots extending along the suture 138 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS behind the middle, the small pronotal punctures somewhat evident because unclothed but the elytral punctures less conspicuous as more or less concealed by the vestiture. Head with impressed median longitudinal line unclothed; eyes strongly reniform, narrowly sep- arated above and coarsely faceted; antennae long and delicate, reaching about five segments beyond apex of elytra, scape some- what ‘sinuous and subeylindrical outwardly and feebly fuscous towards apex, the third segment shorter than the fourth, the following gradually shorter. Prothorax less than twice as wide as long, the lateral tubercles very acute, almost spiniform and placed as usual slightly back of middle, sides barely arcuate, almost straight and convergent forwards from the spines, straight and parallel behind; dise trans- versely impressed back of apex and in front of base, with a small tubercle on each side of middle and rather densely punctured; the punctuation evident because not covered by the pile. Seutellum bare at middle, tufted laterally. Elytra about twice as long as broad and almost five times as long as prothorax, widest at base, humeral angles prominent, gradually narrowed posteriorly, apices of elytra feebly and obliquely truneate; dise with elongate tubercle near middle of each elytron close to base and a feeble ridge reaching from humeri, obliquely inwards almost to apex, the surface densely pubescent but with seat- tered punctures more or less conspicuous. Beneath densely uniformly pilose, the pile finer and not sealelike, punctuation concealed ; the two small tubercles at hind margin of mesosternum hardly visible. Legs stout, the clavate portion of hind femora and tibiae fuscous. Length 6 mm., breadth 2.25 mm. Holotype, collected on James Island, March 2, 1906, by F. X. Williams; a paratype collected at Academy Bay, Indefatigable Island, March 24, 1925, by the Templeton Crocker Expedition ; and a specimen since badly injured by Anthrenus, collected on Jervis Island, Decem- ber 18, 1905, by F. X. Williams. The Indefatigable Island specimen is slightly smaller than the holotype and with the color pattern more definite. The Jervis Island specimen was larger than either of the others and with the markings rather inconspicuous. This little species differs from the preceding by being in general considerably smaller and proportionally narrower, with a decided maculate color pattern and by having the dorsal punctuation of both prothorax and elytra finer and less evident. It is distinctly congenerie with the preceding, in fact very closely related, having the same acute lateral tubercles to prothorax, the small tubercles on hind margin of mesosternum somewhat conspicuous and the dorsal punctuation un- covered by the vestiture and therefore quite evident. No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 139 Taeniotes hayi (Mutchler) Monochammus hayi Mutcuier, 1938, Am. Mus. Novitates, no. 981, p. 13, pl. eee Monochammus cocoensis MuTCHLER, 1938, Am. Mus. Novitates, no. 981, p. 13. [Dr. Van Dyke retained this species in Monochammus. The present generic assignment and synonymy conforms to Dillon and Dillon (1941, p. 17).]—Ep. Family CHRYSOMELIDAE This family is very poorly represented on the Galapagos Islands. There are, however, several species that are definite representatives of its fauna while a number of others that have been attributed to it, chiefly well-known Central and South American species which have no doubt been collected by careless collectors and wrongly attributed to the Islands. This is no doubt the case with the following: Doryphora guerini Stal, var., Diabrotica ventricosa Jacoby, and Physonota alu- tacea Boheman, species collected by members of the St. George Ex- pedition of 1924, and listed by Blair as from the Islands. I believe that the above were all collected on the mainland, later taken to the Islands and mixed with local material. Those definitely known to occur on the Islands are the following: Metachroma labrale Blair Metachroma labrale Buatr, 1933, Ann. Mag. Nat. Hist., ser. 10, XI, p. 484. The California Academy of Sciences has one specimen of this species, collected on Indefatigable Island, in October, 1905, which was compared with Blair’s type, and sixteen specimens collected at Villamil, South Albemarle on March 4-14, 1906. Most of the specimens have a ereenish gloss to the upper surface. The species is fully winged. Diabrotica limbata C. Waterhouse Diabrotica limbata C. WarTERHOUSE, 1877, Proc. Zoo. Soc. Lond., pp. 81-82. Diabrotica limbata C. Waterhouse, Mutcruer, 1925, Zoologica, V, no. 20, p. 238. This species has not been collected by any of those who have visited the Islands since Darwin’s time. Docema Charles Waterhouse Docema C. WATsRHOUSE, 1877, Proc. Zoo. Soc. Lond., p. 80. 140 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS Docema galapagoensis (G. R. Waterhouse) Haltica galapagoensis G. R. WATERHOUSE, 1845, Ann. Nat. Hist., XVI, pp. 39-40. Docema galapagoensis (G. R. Waterhouse), C. WATERHOUSE, 1877, Proc. Zoo. Soc. Lond., p. 81. F Haltica galapagoensis G. R. Waterhouse, LINELL, 1898, Proc. U. S. Nat. Mus., XXI, p. 262. Docema galapagoensis (G. R. Waterhouse), MuTcHLeEr, 1925, Zoologica, V, no. 20; p. 238. The California Academy of Sciences has a series of fifty-five speci- mens: one from Charles Island, May 17, 1932, the type locality; five specimens from Jervis Island, June 6, 1932; and three specimens from Tagus Cove, Albemarle Island, March 23, 1932; all collected by M. Willows, Jr., of the Templeton Crocker Expedition of 1932; and thirty-seven specimens from Albemarle Island, April 24-26, 1906, collected by F. X. Williams. They are all black above and agree in every way with the description except that there is a variation in the eolor of the legs. Docema darwini Mutchler Docema darwini MUTCHLER, 1925, Zoologica, V, no. 20, pp. 230, 238. A specimen collected by F. X. Williams for the California Academy of Sciences was collected on Chatham Island, in February, 1906. It has been carefully compared with one of Mutchler’s paratypes. According to Williams notes, it was taken on a heliotrope-like plant on various islands, also on some Laguminosae. This species is larger than the preceding and aeneous, not black in color. Longitarsus lunatus Charles Waterhouse Longitarsus lunatus C. WATERHOUSE, 1877, Proc. Zoo. Soc. Lond., p. 81. Longitarsus lunatus C. Waterhouse, MutcHuiEr, 1925, Zoologica, V, 20, p. 238. The California Academy of Sciences has no specimens in its collection. Longitarsus galapagoensis Van Dyke, new species Oval, quite convex above; piceous as to body, head, seutellum, fourth and following antenal segments, first and second pair of legs, and the tibia and tarsi of hind pair; the prothorax variable from en- tirely piceous to almost entirely castaneous. The holotype is castaneous with a V-shaped median and oblique lateral piceous spot, the elytra castaneous with sutural area and lateral margins somewhat piceous, No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 141 the general surface smooth and shining. Head smooth, eyes rather coarsely faceted, third antennal segment shorter than second or fourth, fifth longer than fourth, the following gradually more robust. Pro- thorax distinctly transverse, very finely, diffusely punctured. Elytra elongate oval, with prominent though rounded humeri; finely and rather sparsely punctured; wings fully developed. Length slightly over 2 mm., breadth somewhat less than a millimeter. Holotype and numerous designated paratypes from a series of nine- teen specimens collected by F. X. Williams on Charles Island, May 15, 1906. There is also one specimen collected in February, 1906 on Chatham Island by Mr. Williams; also four specimens collected by M. Willows, Jr., of the Templeton Crocker Expedition of 1932, two on Charles Island, April 25, 1932, and two on Indefatigable Island, May 5 and May 6, 19382. This small and very distinctly marked species was compared with the type of L. lunatus, which is in the British Museum, and found to differ not only in color pattern, but in being larger, having the pro- thorax distinctly transverse while hardly wider than long in the other, in having the elytra broad at the shoulders and with them well de- veloped while ‘in the other elytra are narrowed at the shoulders and with them practically obliterated, also much broader across the middle. The wings are well developed in L. galapagoensis while apparently atrophied in ZL. lunatus. Family BRUCHIDAE But two species of this family so far have been collected in the Galapagos Islands. They were both described by K. G. Blair. Spermophagus galapagoensis Blair Sphermophagus galapagoensis BuAtr, 1928, Ann. Mag. Nat. Hist., ser. 10, I, pp. 678-679. The California Academy of Sciences has two specimens of this species which have been carefully compared with the type. They were collected by M. Willows, Jr., of the Templeton Crocker Expedition of 1932, on Hood Island, April 20, 1932. Bruchus fuscomaculatus Blair Bruchus fuscomaculatus Buarr, 1928, Ann. Mag. Nat. Hist., ser. 10, I, pp. 679- 680. No specimens of this species appear to have been taken except by the St. George Expedition of 1924. It is not represented in the California Academy of Sciences collection. 142 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS Family ANTHRIBIDAE Ormiscus G. R. Waterhouse Ormiscus G. R. WATERHOUSE, 1845, Ann. Nat. Hist., XVI, p. 37. Ormiscus variegatus G. R. Waterhouse Ormiscus variegatus G. R. WATERHOUSE, 1845, Ann. Nat. Hist., XVI, pp. 37-38. Ormiscus variegatus G. R. Waterhouse, C. WATERHOUSE, 1877, Proc. Zoo. Soc., V, p. 82. Ormiscus variegatus G. R. Waterhouse, LINELL, 1898, Proc. U. S. Nat. Mus., XXI, no. 1148, p. 268. Ormiscus variegatus G. R. Waterhouse, MuTcH Er, 1925, Zoologica, V, no. 20, p. 238. In the California Academy of Sciences there are three specimens of what is apparently this species. One is lighter in color than the type, somewhat immature, I would say, agreeing more with var. B. than with the type. It was collected on Gardner Island, near Hood Island, April 22, 1932, by M. W. Willows, Jr., of the Templeton Crocker Expedition of 1932. The second specimen was collected at Villamil, Albemarle Island, March 4-14, 1906, by F. X. Williams. The third specimen was collected on Abingdon Island, September 18-23, 1906, by F. X. Williams. The species has ample wings. Family CURCULIONIDAE Quite a number of weevils belonging to various subfamilies and tribes have been taken at different times on the Islands. Most of these have been collected in limited numbers too, which is not surprising considering that many of them are quite small. Judging from this I believe that there are many species still undiscovered in the Islands. Amphideritus cuneiformis (G. R. Waterhouse) Plate VII, figure 9 Otiorhynchus cuneiformis G. R. WATERHOUSE, 1845, Ann. Nat. Hist., XVI, p. 38. : Otiorhynchus cuneiformis G. R. Waterhouse, C. WATERHOUSE, 1877, Proc. Zoo. Soc., p. 82. Otiorhynchus cuneiformis G. R. Waterhouse, LInELL, 1898, Proc. U. S. Nat. Mus., XXI, no. 1143, p. 267. Otiorhynchus cuneiformis G. R. Waterhouse, Mutcuier, 1925, Zoologica, 20, p. 238. No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 143 The original type of this species which should be in the British Museum, appears to have been lost. I looked for it very carefully while studying there, also had the assistance of the keepers, but without result. The California Academy of Sciences has two specimens col- lected by F. X. Williams on Chatham Island, one during July, 1906, the other and larger one, January 24-30, 1906; both agree absolutely with the Waterhouse description. These are the only specimens, I believe, that have been collected since Darwin’s time, but these are sufficient to settle some points that have long been in doubt. They, of course, do not belong to the Old World genus Otiorhynchus but to the genus Amphideritus Schwarz, a well known South American genus, which is sufficiently characterized by the scrobes being lateral, broad behind, the prothorax without postocular lobes, the rostrum rather short and broad; the scape of the antennae rather long, passing the eyes; the elytra slightly broader than the prothorax; the seutellum visible between the elytra at base; and the front coxae contiguous but with a small postcoxal process. The first specimen had but recently emerged, for the deciduous pieces of the mandibles are still attached, broad at base, and sickle-shaped. This specimen is the smaller, 6 mm. long by 2.75 mm. wide. The second specimen, a somewhat worn indi- vidual with much of the sealing removed from both pronotum and dise of elytra, is larger, 8 mm. long by 4 mm. wide. In this second indi- vidual, as a result of the removal of much of the scaly covering, the pronotum is shown to be rather broadly, longitudinally impressed at middle and somewhat coarsely, irregularly punctured. The elytral striae in this second individual are rather well impressed as well as coarsely punctured and the intervals elevated and convex. Pantomorus Schonherr Pantomorus ScHONHERR, 1840, Genera et Species Curculionidum, V, 2, Paris, p. 942. Aramigus Horn, 1876, Proc. Amer. Phil. Soc., XV, p. 93. This rather large genus which is so well represented in Mexico and Central America, differs from its parent stock, Nawpactus Schonherr, which is dominant throughout much of South America, only by being wingless. The genus is a very polymorphic one. The typical forms as represented by the Mexican genotype, P. albosignatus Boheman, and the common and widely distributed species, P. godmani Crotch, now well established in western North America, are of rather moderate size and with a cylindrical prothorax. The species from the Galapagos Islands, for there are several rather closely related or slightly divergent 144 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS forms, are in general somewhat larger and have the prothorax more or less spherical like the Mexican species, P. albicans Champion. Cham- pion (1911) in his treatment of Mexican species, has divided them into two groups: 1. Males without mucro on the inner edge of extremity of middle tibia. 2. Males with mucro on the inner edge of extremity of middle tibia. To the first group, the more typical species belongs P. albosignatus Boheman, while to the second group belong such species as P. albicans Champion and those to be found in the Galapagos Islands. In the Galapagos Islands, the genus is widely distributed and according to F. X. Williams the species in the adult stage are to be found on various plants, but preferably on the croton. In the various islands they have had a tendency to develop into more or less closely allied races and species. In the material at hand, I have a sufficient number of specimens to be able to separate about six species. A speci- men from Hood Island is too rubbed to enable it to be properly char- acterized. All of the species in the Islands are quite variable as to size and color pattern, the usual coloration being gray or light brown pro- duced by a rather closely appressed pile, with erect or suberect setae or long hair projecting above and a certain number of silvery white seales concentrated to form spots or maculations or placed in denser formation along the sides. In the majority of specimens the true seales are lacking while in others they are very evident, the most complete sealy pattern being shown in a large female specimen of typical P. galapagoensis, trom Chatham Island. In this specimen, there are numerous elongate scales along the sides of the prothorax, placed vertically ; somewhat more robust scales densely placed along the sides of the elytra, arranged obliquely; a series of elongate, more or less elliptical scales arranged in spots as follows: a small one at the base of the elytra near its middle and a series of three arranged in the manner of a broken lunule extending from the humerus to the suture near its center, a subapical macule, and a series of somewhat broader seales closely applied to the side pieces of the meso- and meta- sternum. The males are in general narrower than the females, less robust, with the legs longer and the front tibiae more arcuate towards the apex, the antennae apparently also longer, the prothorax more spherical with the base of the elytra more evidently elevated and carinate towards the suture, the abdomen also more flattened or broadly impressed towards the base. No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 145 Pantomorus galapagoensis Linell Plate VII, figure 7 Pantomorus galapagoensis LINELL, 1898, Proc. U. S. Nat. Mus., XXI, no. 1143, p. 268. Pantomorus galapagoensis Linell, MurcHuier, 1925, Zoologica, V, no. 20, p. 238. Pantomorus galapagoensis Linell, BLatr, 1933, Ann. Mag. Nat. Hist., ser. 10, XI, p. 485. Rather large and robust, piceous to rufopiceous, in fresh specimens rather densely clothed with a closely appressed pile of mixed light and dark brown or gray hairs, often with a silvery gloss, in addition with rather long, erect, fine hairs rather uniformly scattered over the elytra and in many specimens with patches of elongate silvery white scales arranged as three white spots of an interrupted lunule extending from the humerus to the suture near the middle, also along the side pieces of the meso- and metathorax. Head feebly convex, alutaceous, finely punctured and rugose and somewhat aciculate; rostrum broad, one and a half times as long as broad, flattened above, rugose and punctate, strigose at sides and with a sharply impressed longitudinal groove extending from the middle of frons to above the level of insertion of antennae; mandibles prominent with the supports to the deciduous pieces conspicuously projecting; eyes prominent, very convex, and obliquely set; antennae long, fully reaching to hind margin of pro- thorax, second funicular segment one-third longer than first. Prothorax one-fourth broader than long in female and but little broader than long in males, apex transverse, base feebly arcuate in females, dis- tinetly sinuate at sides and with well defined median lobe in males, also rather finely margined in both sexes, sides broadly rounded, con- stricted in front and behind; dise very convex, alutaceous, fairly to coarsely rugose, sparsely punctured, sometimes aciculate and generally with a well impressed longitudinal line at middle. Elytra cordate, about two-sevenths longer than broad and three times as long as prothorax, convex, striae feebly impressed but strial punctures coarse but close and deeply impressed, intervals flat or feebly convex and generally al- most twice as wide as striae and derm alutaceous and finely rugose. Legs long, front tibiae arcuate apically and coarsely serrate on inner margin in both sexes, middle tibia mucronate within at apex. Be- neath with pile somewhat less dense than above. Length 10-14 mm., breadth 4.5-6 mm. The males are in general narrower, with more spherical prothorax, somewhat coneave abdomen, and with legs apparently longer. Linell was in error in stating that the males had the larger thorax. Type, no. 1327, U.S.N.M. Linell had before him one male and four 146 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS females from Chatham Island (three collected by the Albatross Ex- pedition in 1888 and two by Dr. G. Baur). One of these specimens, a paratype sent to the British Museum, I used for comparison with some of my specimens. The California Academy of Sciences has thirty-three specimens, thirty-two collected January 24-30, 1906, and February 21-24, 1906, by F. X. Williams on Chatham Island; also one collected April 15, 1932, by M. Willows, Jr., of the Templeton Crocker Expedi- tion of 1932. This species, judged by the specimens which we have, is in general the largest and most robust of the Galapagos Islands species, also when fresh the most densely pilose and with the longest and finest erect hair. The color of the pile varies with the specimens, some being almost uniformly brown while others are varicolored, and in a limited number the maculation of silvery seales are very conspicuous. Pantomorus blairi Van Dyke, new species Of moderate size, less robust than preceding species, piceous with appressed pile rather short and sparse, a unicolored brown or gray and not appreciably concealing the derm beneath, the erect hairs rather stiff and of moderate length, and well dispersed over the surface of the elytra. Head somewhat coarsely, closely punctured and rugose with the usual sharply impressed longitudinal line and generally the strigae, the rostrum and antennae with similar proportions to the preceding but the eyes a bit less prominent, less convex. Prothorax with propor- tions of P. galapagoensis but with dise rather finely and irregularly punctured and generally quite rugose, smoother in females and base generally more transverse in males. Elytra with base quite transverse, sides but slightly arcuate and convergent backwards and apex less acute. As a result of the sparse pubescence the derm of the dise is quite exposed, showing the strial punctures to advantage, which ap- pear to be less coarse and closer together than they are in P. gala- pagoensis. Beneath with sparse pubescence. Length of holotype male 10 mm., breadth 5 mm., of allotype female, length 13 mm., breadth 7 mm. Three of the paratypes are much smaller. Holotype male, allotype female, and two paratypes will be returned to the British Museum of Natural History. They are from a series of six specimens, kindly loaned to me for purposes of study by Dr. K. G, Blair. Two of the specimens will be retained. Three specimens are de- pauperized, much smaller than the others. All specimens were collected on James Island, February 20-22, 1925, by G. Bateson: This species is in general slightly smaller and darker than P. gala- pagoensis, with head more distinctly punctured, the eyes less convex, No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 147 pronotum more definitely loosely punctured, the afterbody more par- allel sided, the appressed pubescence sparser and the erect hairs more rigid. The usual color of the pile is of a uniform brown, gray in one or two, and in but one specimen are there a few silvery scales forming a small macule behind the base of the elytra though the seales of the meso- and metapleurae are quite evident in all specimens as usual. Pantomorus crockeri Van Dyke, new species Small, piceous with legs somewhat rufous; body sparsely clothed with fine, rather short pile which is not closely appressed and which has the individual hairs somewhat curled and irregularly directed, the erect hairs of moderate length, stiff and rather oblique, scales absent except on the metapleurae where they are sparse and more hairlike than usual. Head coarsely punctured and rugose; eyes moderately prominent but widely separated by at least three times their own diameter ; beak short and broad, as broad as long. Prothorax somewhat broader than long, coarsely and irregularly punctured and very rugose, with granules very conspicuous, and a well defined median longitudinal impression. Elytra nearly two-fifths longer than broad and three-fifths longer than prothorax, very convex, the declivity more precipitous than usual, the striae moderately impressed, strial punctures coarse and close, intervals feebly convex and about as wide as striae, the sur- face minutely punctured. Beneath rather sparsely clothed with fine hair. Length 7 mm., breadth 3.5 mm. Holotype, a unique collected on Tower Island, March 25, 1935, by the Templeton Crocker Expedition of 1935, and deposited in the collection of the California Academy of Sciences. This very distinct and stubby species has as its most characteristic features, its shortness and small size, its short rostrum, widely separated eyes, very granular pronotum, convex elytra, and irregularly inclined pubeseence. I have named it after Templeton Crocker, a generous patron, who has added much to our knowledge of the Galapagos Islands. Pantomorus caroli Van Dyke, new species Of moderate size, piceous or rufopiceous ; body above rather densely clothed with closely appressed, regularly dispersed fulvous or gray pile, concealing to a great extent the derm beneath; and in addition provided with very short, stiff semierect hairs, irregularly dispersed over the elytra though most evident on the elytral declivity ; and with scales generally absent above, only a few specimens showing them usually obliquely within and behind the humeri, but the scales on meso- and metapleurae present as usual but limited in number. Head 148 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS alutaceous, rather densely, finely punctured, often strigose at sides, flattened in front; eyes prominent, rather widely separated; rostrum broad but almost a third longer than broad, similarly sculptured to rest of head but with strigae more evident and with the usual sharply and deeply impressed longitudinal median groove running back to the occiput. Prothorax at least 5 mm. broader in female than male and barely broader than long in males, apex transverse or feebly emarginate at middle, base broadly feebly arcuate at middle in female, with a pronounced lobe at middle and strongly sinuate each side in male, sides broadly arcuate, somewhat constricted at base and apex in female, more definitely so in male and with the anterior and posterior margins also better defined; dise quite convex, finely diffusedly pune- tured, rugose and often distinctly alutaceous especially in males, and generally with a more or less observable median longitudinal impres- sion. Elytra about a third longer than broad and three times as long as prothorax, somewhat transverse at base and with humeral angles well defined though rounded at apices especially in males, sides feebly arcuate and widening to posterior third thence more broadly rounded and convergent to apex; dise convex, striae not or feebly impressed, the strial punctures coarse, their own diameter apart, and deeply im- pressed, intervals generally flat and much wider than striae, general surface finely rugose and with minute punctures here and there. Be- neath rather sparsely pubescent, legs as in related species. Length 6-13 mm., breadth 2.5-5.5 mm., the smaller measurements from depau- perized specimens. Holotype, aliotype and several designated paratypes from a series of twenty-seven specimens in the California Academy of Sciences col- lection, all collected by F. X. Williams on Charles Island, October 3-15, 1905. There is also a single specimen before me, loaned by the British Museum which was collected by G. Bateson of the St. George Expedition, March 1925. The specimen is from Charles Island and is much denuded. There is considerable variation in size between indi- viduals, several being quite small; there is also variation to a slight extent in color of pile, in degree of strial impression, and sculpturing of pronotum, most being quite rugose while two large females have the surface much smoother and shining. The males all have the prothorax quite spherical. The main diagnostic characters of this species are the somewhat dense appressed pubescence, the short, dispersed yet evident erect hairs or setae, the (in general) finely rugose pronotum, and the fair size of the normal individuals. It appears to be most closely related to P. galapagoensis, differing primarily in having less dense pile, shorter No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 149 and sparser setae, and more coarsely sculptured pronotum; also the elytra are more transverse at base and the humeri more sharply angulate. Pantomorus conwayensis Mutchler Pantomorus conwayensis MuUTCHLER, 1938, Am. Mus. Novitates, no. 981, pp. 15-16. Of moderate size, rufopiceous; sparsely clothed with short, gray, closely appressed pile which does not appreciably conceal the derm beneath; with a few short, curved setae, very obliquely set, hardly projecting above normal pile and only evident on elytral declivity ; and in a limited number of cases with macules of elliptical shape and silvery scales, placed one on either side of the seutellum, a series of from three to four extending obliquely inwards from the humerus towards the suture and one on the sixth interval a third of the dis- tance from the apex as well as forming a marginal patch along the outer side of the meso- and metasternal pleurites. Head alutaceous in front, rather densely punctured with broad, shallow, somewhat acicu- late punctures, with the usual deep median longitudinal impression. Elytra almost a third longer than broad and about twice as long as prothorax, quite transverse at base, with well defined subangular humeri, the sides feebly arcuate and gradually expanded to posterior third; dise with striae slightly impressed, strial punctures coarse, close together and deeply impressed, the intervals as wide or slightly wider than striae, and flattened or feebly convex depending upon the degree to which the striae are impressed. Beneath sparsely pilose. Holotype male and allotype, in the American Museum of Natural History. In deseribing this species, Mutchler had before him, thirty-three specimens, collected either by the Williams’ Galapagos Expedition of 1923 or by the Crocker Expedition of 1935. All of these were collected in Conway Bay, Indefatigable Island. The California Academy of Sciences has the following specimens: one small paratype from Conway Bay, received from the American Museum of Natural History ; two specimens collected March 24, 1925, at Academy Bay, Indefatigable Island, by the Templeton Crocker Expedi- tion of 1925; two specimens collected May 6—7, 1932, by M. Willows, Jr., of the Templeton Crocker Expedition of 1932, and one collected October 25-28, 1905, by F. X. Williams, all labeled Indefatigable Island. Besides these are six other specimens, all badly rubbed, which were collected May 2 or 6, 1932, by M. Willows, Jr., of the Templeton 150 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS Crocker Expedition of 1932, on Indefatigable Island. One smal] speci- men loaned by the British Museum, which was collected by G. Bateson of the St. George Expedition of February 16, 1925, is before me. This species is best separated by its sparse pubescence and short, curved setae which are mainly confined to the elytral declivity. It is rather closely related to the species that follows: Pantomoris williamsi Van Dyke, new species Very much like P. conwayensis in general appearance and in most essential features, having in particular the short sparse pubescence which does not appreciably conceal the derm. It differs, however, in having the erect setae which are rather short stiff hairs, quite evident, generally dispersed over the elytra and but little inclined, and in having the silvery scales which are quite elongate and hairlike, not elliptical as in the other species, assembled in various conspicuous macules in all of the specimens before me. These macules are arranged as follows: two somewhat vague ones at the base of the elytra, a series of three arranged in the form of an interrupted lunule from the humerus towards the suture, and a series of from two to three ar- ranged transversely across the posterior third of the elytra, the outer- most, the largest, placed on the sixth and seventh intervals. Length 9 mm., breadth 4.5 mm. Holotype female and two paratype females, the first collected on Albemarle Island, April 15, 1906, by F. X. Williams, one at Cebes Settlement, Albemarle Island, April 24, 1906, collected by F. X. Wil- liams, and the other at Iguano Cove, Albemarle Island, March 17-21, 1906, also by F. X. Williams. I also have before me three specimens from Banks Bay, Albemarle Island, April 10-17, 1906, collected by F, X. Williams, and two specimens on loan from the British Museum, collected in 1925, by G. Bateson, on Albemarle Island. These will also be considered as paratypes. Key To GALAPAGOS ISLANDS SPECIES OF PANTOMORUS SCHONHERR 1. Erect hairs or setae rather long, fully as long as intervals are wide, quite dense and mone or lesshumitormlly, USPC Se see 2 — Hrect hairs or setae somewhat short, generally sparse and as a rule more evidentrom- ely tral GEC yiaty sce 2e ecco cc ce8e Specter ese cence eet ea rece een 4 2. The closely appressed pile of the elytra generally unicolored and sparse, allowing the brown derm to be readily seen beneath, true scales rarely — The closely appressed pile of the elytra generally varicolored, and dense, No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 151 concealing the derm to a great extent, scales more or less evident (Chat- LENE OSY LUGS ANG 19) a en NaS ae cp ae ach A ko OYE R. galapagoensis Linell 3. Rostrum much longer than broad; pronotum moderately, not coarsely rugose; pile of elytra light brown or gray and regularly inclined (James SS Tern 1p) berate ree ee a Sa nn ne Ree a ee P. blairi Van Dyke — Rostrum litile longer than broad; pronotum coarsely rugose with evident granules; pile of elytra gray with individual setae much curved or curled and not regularly inclined (Tower Island) .................. P. crockeri Van Dyke 4. The appressed pile of elytra very short, fine and sparse, not concealing to any degree the brown derm beneath; macules of silvery scales gen- erally present on (elytra and very evident. -2 2.cc. seen 5 — The appressed pile of elytra rather coarse and dense, concealing the derm to an appreciable extent, scaly macules rarely indicated on elytra, erect setae short and sparse on disc, longer, denser and more inclined on dechivibye (Charles pisland))e.: ts. 2 see eete eg cee P. caroli Van Dyke 5. Short, erect setae sparsely dispersed over the elytra (Indefatigable AVS air C1) eee ene nn ee en ae Bye OPE a ee P. conwayensis Mutchler — Setae hardly observable on dise of elytra, short and much inclined on elytral declivity, pile abundant (Albemarle Island)..P. williamsi Van Dyke Gerstaeckeria galapagoensis Van Dyke, new species Plate VII, figure 8 Medium sized, robust, black with rufous antennae; very densely clothed above with short, broad, varicolored scales, the majority dark brown, giving the basic color, the others varying from light brown to white, the latter generally assembled so as to give the distinctive color pattern. These scales are disposed as follows: about the eyes; on the pronotum, in the form of a short longitudinal line in front of the seutellum and a few spots on the dise, generally a pair near the center and several at the sides; and on the elytra as a lunate patch extending obliquely inwards from the humerus, and a slightly arched transverse patch, widest at center and sides, placed at the summit of the elytral declivity and extending from fourth interval on one elytron to the fourth on the opposite elytron. The sealing on the underside and legs is somewhat less dense than above. Head very coarsely, densely pune- tured above, the punctures extending on to the sides of the rostrum but becoming finer and more widely spaced toward the apex with each puncture on head and base of rostrum supporting a scale which is rather broad on occiput and narrower and more upright between the eyes, also in general a lighter brown or dirty white about the eyes; eyes separated in front by slightly more than the diameter of a single eye, a small fovea between; the rostrum about as long as head, feebly arched, carinated above, constricted in front of eyes and feebly nar- 152 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS rowed about middle. Prothorax about one-fifth broader than long, base transverse, apex over one-fourth narrower than base, broadly arcuate, the lobe covering base of head, sides arcuate from base forward to beyond middle, thence narrowed and feebly constricted before apex, dise convex, densely and coarsely punctured, each puncture supporting a seale. Elytra more a fifth longer than broad and about two and a half times as long as prothorax, without posthumeral process, the sides evenly arcuate from base to posterior third then gradually narrower and somewhat sinuate to the somewhat extended apex; dise very con- vex, the declivity almost vertical, striae broad and well impressed, the strial punctures very coarse, separated by one-half their own diameter, intervals as wide as striae, convex and finely and rather densely punc- tured, all punctures supporting a scale as elsewhere above and all intervals equally sealy. Beneath rather coarsely, densely punctured. Legs with femora unarmed, clothed with light and dark seales, which are generally arranged in a somewhat annulated manner on the outer face, the third tarsal segment very broad, lobed, much wider than second. Length, without rostrum, 7-8 mm., breadth 3.5-4 mm. Holotype and designated paratypes from a series of forty-six speci- mens, collected by F. X. Williams on Abingdon Island, September 18-23, 1906. This species would probably fall in the subgenus Copun- tiaphila Pierce and somewhere near G. cruciata Champion accord- ing to Pierce’s (1889) key. Besides the rather large Abingdon Island series of specimens of Gerstaeckeria in the collection of the California Academy of Sciences, there are specimens from several of the other islands of the Galapagos Archipelago. These all possess the basic characters and general scale color pattern indicated in G. galapagoensis though those from certain islands seem to possess in addition, and in spite of their variability, certain definite peculiarities which I think entitles them to a name as subspecies, though nothing more. Gerstaeckeria galapagoensis barringtonensis Van Dyke, new subspecies This form apparently differs from the more typical form by being in general slightly smaller; by having the prothorax barely broader than long, the sides well rounded and with the greatest breadth in front of the middle and almost straight behind and convergent to base ; and by having the dise of both pronotum and elytra as seen from the side, more convex, the humeral angles more rounded, and the strial No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 153 punctures of the elytra more sharply defined. Length 6 mm., breadth 3mm. Holotype and several designated paratypes from a series of thirteen specimens, collected on Barrington Island, July 4-10, 1906, by F. X. Williams. Gerstaeckeria galapagoensis hoodensis Van Dyke, new subspecies This form is about the same size as the preceding and also differs from typical G. galapagoensis in having the prothorax but slightly broader than long, with sides but feebly arcuate; the elytra with the sutural interval depressed, the others very convex, and the strial pune- tures sharply defined. Holotype and four paratypes collected on Hood Island, by F. X. Williams, on the following dates: two in January, 1906, two on June 23-30, 1906, and one on February 1-14, 1906. Gerstaeckeria galapagoensis seymourensis Van Dyke, new subspecies The specimens of this subspecies are all, unfortunately, almost com- pletely denuded of scales. They are the largest of any of the races, have the prothorax one-sixth broader than long, with the sides well rounded and broadest at middle, the dise somewhat flattened, elytra elliptical, one-third longer than broad, and with the dise quite flattened as seen from the side but not when viewed from behind, and the in- tervals equally elevated. The size and elongated body chiefly distinguish it. Length 9 mm., breadth 4 mm. Holotype and three paratypes, three collected on South Seymour Island, November 22, 1905, and one July, 1906, by F. X. Williams. I have also associated with these a single specimen collected on Inde- fatigable Island, October 25-38, 1905, by F. X. Williams. It cannot be distinguished from the others. All of the members of the genus Gerstaeckeri are cactus feeding. We have numerous species in the southern part of the United States, chiefly in the semiarid Southwest, in Mexico and the West Indies, I know of none that have been deseribed from South America though I feel that they must be found somewhere along the more desert parts of the West Coast. They have no functional wings. Geraes batesoni Blair Geraeus batesoni Buatir, 1933, Ann. Mag. Nat. Hist., ser. 10, IX, pp. 485-486. No specimens have been taken by any of the Academy’s expeditions 154 CALIFORNIA ACADEMY OF SCIENCES [Oc. Papers nor by any expeditions previous to the St. George Expedition which furnished Blair’s specimen. Lembodes subcostatus Van Dyke, new species Plate VII, figure 3 Small, elongate and subdepressed, black opaque, more or less cov- ered with a chalky indumentum which gives the insect a clouded gray and brown appearance, and with stubby seales arranged in series or tufts over the upper surface and more or less regularly arranged on the legs. Head partly concealed by overhanging pronotum and with a few stubby erect scales placed between the eyes and on base of rostrum ; antennae rufous. Prothorax almost a third longer than broad, apex in the form of a lobe, notched at middle and overhanging the head, base transverse, sides almost straight or very feebly arcuate to beyond middle where they are suddenly narrowed as they merge with the mar- gins of the shovel-like apical lobe; dise rather irregularly convex be- hind, more flattened in front with the erect, brown, stubby scales seat- tered over the base, arranged in the form of four tufts one on either border near the middle and one on either side of the middle of dise, and in a denser series in arched formation just back of apical margin, as well as a few seattered behind this series. Elytra about twice as long as wide and twice as long as prothorax, base transverse with humeral angles extending slightly forward, sides almost straight and feebly diverging to beyond the middle, thence arcuate, narrowed and sinuate at apical fourth and continued on to the rather broad subtruneate apex which is feebly notched at suture; the dise convex, somewhat flattened suturally and with a series of three feeble longitudinal costae on either side which are surmounted and more definitely outlined by series of the suberect stubby brown and gray scales, the costae placed one along the side margin, one somewhat above and one midway between this and suture, this last diverging from the suture as it passes backwards, a few scales also placed along the suture. Beneath with surface con- cealed by gray indumentum, in addition to a few scales scattered over second ventral and arranged in transverse rows on the third and fourth ventral segments. Legs gray, somewhat annulated with black on outer surface and rather densely set with the more or less erect scales which are larger and spoon shaped on the tibiae. Length 4.5 mm., breadth 1.75 mm. Holotype and two paratypes, collected as follows, the holotype from Duncan Island, November 1, 1905, a second specimen, also from Dun- ean Island, January 1-17, 1905, and the third from Iguano Cove, Albe- marle Island, March 17-21, 1906, all by F. X. Williams. No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 155 Five other species have been described previously: L. solitarius Boheman, the type species from the West Indies; Z. ulwlu Chevrolat from Santo Domingo; L. truz Champion from Guatemala; L. furci- collis Chevrolat from Colombia; and L. albo-signatus Chevrolat from Chile. The species Lembodes subcostatus has been compared with the type of L. trux and found to be very different, being larger, less parallel, and with a lobed not truncate prothorax, as well as differing otherwise, and it apparently does not agree with the descriptions of the other species. Anchonus galapagoensis G. R. Waterhouse Anchonus galapagoensis G. R. WATERHOUSE, 1845, Ann. Nat. Hist., XVI, p. 39. Anchonus galapagoensis G. R. Waterhouse, C. WATERHOUSE, 1877, Proc. Zoo. Soc., V, p. 82. Anchonus galapagoensis G. R. Waterhouse, LINELL, 1898, Proc. U. S. Nat. Mus., XXI, no. 11438, p. 268. Anchonus galapagoensis G. R. Waterhouse, MutTcuier, 1925, Zoologica, V, no. 20, p. 238. Anchonus galapagoensis G. R. Waterhouse, Biair, 1933, Ann. Mag. Nat. Hist., ser. 10, XI, p. 485. Outside of the specimens collected by Charles Darwin and de- seribed as above by G. R. Waterhouse, there are only two specimens that I know of that have been collected since in the Archipelago, one listed by Blair from the type locality, James Island, and one in the California Academy of Sciences collection, collected on James Island between December 22, 1905, and January 5, 1906, by F. X. Williams; this last specimen has been carefully compared with the type. The genus Anchonus Schonherr is a large one, widely distributed through- out the warmer parts of the New World, being found in Mexico, Cen- tral and South America, and the West Indies. One specimen has even been taken in the South Pacific. Dryotribus mimeticus Horn Dryotribus mimeticus Horn, 1873, Proc. Am. Phil. Soc., XIII, p. 433. Dryotribus mimeticus Horn, Buair, 1933, Ann. Mag. Nat. Hist., ser. 10, XI, p. 486. Thalattodora insignis Perkins, 1900, Fauna Hawaiiensis, vol. 2, p. 146. Three specimens of this species were taken on Narborough Island by C. L. Collenette of the St. George Expedition. Blair states that this widespread species is represented in the British Museum of Natural 156 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS History collection by specimens from Florida, the West Indies, China, western Australia, and Hawaii. I have seen specimens also from the Marquesas Islands. It is a strand inhabiting species so no doubt owes its wide distribution to this fact. No specimens of this species were collected in the Galapagos Islands by the California Academy of Sciences expedition. Macrancylus gracilis Van Dyke, new species Small, linear and subeylindrical, smooth and shining, rufous with base of head piceous, head parallel sided or feebly convergent forwards to beyond eyes, occiput smooth with a feeble transverse impression demarking it in front, the frons alutaceous and rather coarsely and regularly punctured, the punctures separated by at least their own width; eyes small, lateral, much flattened, hardly projecting beyond side margin; rostrum about as long as head proper, slightly narrower at base than head, with straight sides feebly convergent forwards, the supper surface with punctures continuous with those of head but eradually finer; antennae rufous but with club very light colored, a yellowish white. Prothorax at least twice as long as broad, sides rounded at base, straight and gradually convergent forwards; dise alutaceous with punctures somewhat coarser than on head, rather regularly placed and about their own diameter apart. Elytra over three times as long as broad and one-third longer than prothorax, as broad at base as base of prothorax, with base transverse, the humeri rounded, sides straight and parallel until near apex where they become rounded; dise striatopunctate, the striae feebly im- pressed, punctures rather coarse and close together, almost con- tiguous in places, the intervals flat, as wide as striae and crenulate as a result of being indented by punctures. Beneath alutaceous, with punctures rather regularly yet widely distributed, the after- body piceous. Length 3 mm., breadth 5 mm. Holotype, a specimen collected on Abingdon Island, in September, 1906, by F. X. Wilhams. Nine other specimens from Abingdon Island collected in September 18-23, 1906, by F. X. Wililams have been designated as paratypes. This species has been carefully compared with specimens of M. linearis Le Conte, from Florida and the West Indies and with M. im- migrans (Perkins) from Hawaii, the two other species in the genus, and found to be more narrowed, shining, and in general more graceful than either, with the head including rostrum a bit narrower, the eyes less prominent but otherwise with about the same proportions; the No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 157 puncture somewhat finer (in M. linearis), coarse and close while in M. immigrans they are rather coarse but not close; the prothorax more narrowed than in either of the others as is also the general body, alu- taceous, with punctures finer and better separated ; the elytra smoother, intervals finer and more flattened, in others more or less convex, strial punctures also much finer and as a result the striae themselves less deeply impressed. The general color is also lighter, a clearer red, not rufopiceous as in the others. The distribution of these three species is somewhat suggestive of that of the species previously mentioned. Neopentarthrum Mutchler Neopentarthrum MutTCHLER, 1925, Zoologica, V, no. 20, p. 231. The California Academy of Sciences possesses a paratype of Neopentarthrum towerensis Mutehler, kindly donated by Mr. Mutcehler, also a single specimen collected on Tower Island (Darwin Bay), June 16, 1932, by M. Willows, Jr., of the Templeton Crocker Expedition of 1932, which has been compared with the type. In addition the Acad- emy possesses specimens of several other species, differing greatly from the above, from other islands. These will be deseribed later on. Neopentarthrum towerensis Mutchler Neopentarthrum towerensis MUTCHLER, 1925, Zoologica, V, no. 20, pp. 231-232, fig. 45. Neopentarthrum towerensis Mutchler, Biatr, 1933, Ann. Mag. Nat. Hist., ser. 10, XI, p. 468. Besides the above, Blair lists one specimen from Franklin Lake, as having been taken by the St. George Expedition. The California Academy of Sciences specimens have been mentioned previously. Neopentarthrum cunicollis Van Dyke, new species Of the same size and general proportions as NV. towerensis but with the prothorax well rounded and broadest close to the base, the sides almost straight and convergent to apex with hardly a perceptible in- terruption by the feeble post-apical constriction, thus forming a wedge- shaped body; the elytra with striae rather deeply and sharply im- pressed on the disc, the punctures clear cut, close together, and to a great extent confined to striae, indenting the intervals but little thus the latter are more regular, less crenulate throughout. The punctua- tion of the pronotum and underside and other characters are practically the same as in N. towerensis. 158 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS Holotype, a unique collected on Duncan Island, December 1, 1905, by F. X. Williams. With this species I have associated another speci- men, not differing morphologically as far as I can see except in having the elytral striae less sharply impressed and the punctures a bit larger thus nicking the intervals to a greater extent. This specimen which was collected near Iguano Cove, Albemarle Island, May 21, 1932, by M. Willows, Jr., of the Templeton Crocker Expedition of 1932, I would consider but a variety of N. cuncollis. Neopentarthrum mutchleri Van Dyke, new species Of about the same size and proportions as N. towerensis, piceous or slightly rufopiceous beneath with legs and antennae rufous. Head alu- taceous and with fine sparse punctures; the rostrum slightly longer and with basal portion narrower ; the eyes more flattened, not protrud- ing at all beyond side margins. Prothorax feebly broader than long, with sides evenly and well rounded, very narrowly constricted at base and more broadly so at apex; dise as regularly but slightly more finely punctured, the post apical transverse impression distinct sharply de- marking a collar; elytra about twice as long as broad, the sides feebly but regularly arcuate from transverse base to posterior fourth where regularly rounded to apex; the dise with striae not defined but the strial punctures distinct and serially arranged while the finer inter- strial punctures are only observable here and there. Beneath with the few punctures finer than in N. towerensis. Hoiotype, a specimen collected on Abingdon Island, September 18-23, 1906, by F. X. Williams. Three other specimens from Abingdon Island, collected at the same time as the above, have been designated as paratypes. A fourth specimen, which I have associated, is from Indefatigable Island, collected October 25-26, 1905, by F. X. Williams. It is somewhat narrower and generally less robust than the Abingdon specimens but with identical sculpturing. I cannot see that this speci- men is anything more than a variety. The Abingdon specimens, I am naming after my good friend, A. J. Mutchler, former curator of insects in the American Museum of Natural History. Neopentarthrum glabrum Van Dyke, new species Somewhat narrower than NV. towerensis, with head and pronotum distinetly alutaceous, the elytra elliptical and practically smooth, the sculpturing obscure at most and of a black color, with antennae and legs rufous. Head feebly, sparsely punctured behind, more coarsely and closely so in front and on base of rostrum, the latter about as long No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 159 as head and broad, especially in front; the eyes feebly projecting be- yond side margin of head. Prothorax with sides broadly rounded in type, less so in paratype, narrowly constricted at base and broadly narrowed and sinuate towards apex; the dise with punctures rather fine and well spaced, the anterior transverse impression vague, not forming a well defined collar though the apex is distinctly narrowed. Elytra elliptical, twice as long as broad, sides evenly arcuate from transverse base to posterior fourth, then more definitely rounded to apex; the disc moderately convex, smooth and somewhat shining, the striae obliterated and strial punctures likewise except those near the suture which are small and feebly indicated. Beneath smooth in front, with a few fine and sparse punctures on metasternum and somewhat coarser and more numerous punctures on last three abdominal seg- ments. Length 3 mm., breadth 1 mm. Holotype and one paratype, the first collected on Hood Island, January, 1906, and the second on Abingdon Island, September 18-23, 1906, both by F. X. Willams. This comparatively smooth species is most closely related to V. mutchleri but. readily separated by its black color, pronounced elliptical elytra, and practical absence of elytral sculpturing. In spite of one of the specimens coming from Abingdon Island, the home of N. mutchlert, it agrees absolutely with the type of N. glabrum and not with the former. Key To SPECIES OF NEOPENTARTHRUM MUTCHLER 1. Elytra with sides straight and parallel in basal two-thirds, the elytral striae sharply impressed and regularly punctured................2.2-2.2--.------- 2 — Elytra somewhat elliptical, that is with sides more or less arcuate throughout, striae either vague or not impressed...............222...--22:00-2:--2---- 3 2. Prothorax widest at about the middle..................... N. towerensis Mutchler — Prothorax widest near the base.................-.......--.. N. cunicollis, new species 3. Elytral striae vaguely impressed but strial punctures distinct though shallowly impressed and somewhat regularly arranged........................----- N. mutchleri, new species — Elytra smooth without striae and with only a vague puncture here and AEG The eae eta eye teh be ee AA, ced Meodcn edad N. glabrum, new species Family PLATYPODIDAE Platypus santacruzensis Mutchler Platypus santacruzensis Murcuter, 1925, Zoologica, V, no. 20, pp. 232-233, fig. 24. Recently while going over some duplicate material, I found what is presumably a second specimen collected on South Albemarle Island, 160 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS August 20, 1906, by F. X. Williams. This gives the California Academy of Sciences a representative of this species. Family SCOLYTIDAE Pycnarthrum insulare Blair Pycnarthrum insulare Buair, 1933, Ann. Mag. Nat. Hist., ser. 10, XI, p. 487. Besides the four paratypes cited by Blair which are in the Cali- fornia Academy of Sciences collection and which were collected: the Tower Island specimens, September 14, 1905, and the Hood Island specimens, January, 1906, both sets by F. X. Williams, the Academy also possesses five more specimens from Tower Island and five more from Hood Island, same data as above, as well as a series of fifty-five specimens collected on Albemarle Island, March 4-14, 1906, by F. X. Williams. These insects are stated by Dr. Williams to breed in man- grove seeds. REFERENCES Buarr, K. G. 1928. Coleoptera (Heteromera, Teredilia, Malacodermata, and Bruchidae) from the Galapagos Islands, collected on the St. George Expedi- tion, 1924. Annals and Magazine of Natural History, ser. 10, vol. I, pp. 671-680. 1933. Further Coleoptera from the Galapagos Archipelago. Annals and Magazine of Natural History, ser. 10, vol. XI, pp. 471—487. BLAISDELL, FRANK EH. 1932. Studies in the tenebrionid tribe Scaurini. A monographic revision of the Eulabes (Coleoptera). Transactions of the American En- tomological Society, vol. LVIII, pp. 35-101. BoHEMAN, C. H. 1858-59. Coleoptera. In Kongliga Svenska Fregatten Eugenies Resa etc. Stockholm, pp. 1-217. CHAMPION, G. C. 1894-97. Blateridae. In Biologia Centrali-Americana. Insecta, Coleoptera, vol. III, pt. 1, pp. 258-556. 1911. Pantomorus. Jn Biologia Centrali-Americana. Insecta, Coleoptera, vol. IV, pt. 3, pp. 152-166. DARWIN, CHARLES 1839. Journal and remarks. Jn Fitzroy, Narrative of the surveying voy- ages of His Majesty’s ships Adventure and Beagle, between the years 1826 and 1836, describing their examination of the southern No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 161 shores of South America and the Beagle’s cirumnavigation of the globe. Vol. 3, pp. xiv-+615. H. Colburn, London. 1860. The origin of species by means of natural selection, or the preserva- tion of favoured races in the struggle of life. 432 pages. D. Apple- ton and Co., New York. DILLON, LAWRENCE S. AND ELIZABETH S. DILLON 1941. The tribe Monochamini in the Western Hemisphere (Coleoptera: Cerambycidae). Reading Public Museum and Art Gallery Scien- tific Publications, no. 1, pp. 1-135, pls. 1-5. FAUVEL, ALBERT 1889. Liste des coléoptéres communes 4 l'Europe et 4 l’Amérique du Nord. Revue d’Entomologie (Caen), vol. 8, pp. 92-174. HINTON, HOWARD E. 1936. New species of Ataenius allied to A. cribrithorar Bates (Coleoptera Scarabaeidae). The Annals and Magazine of Natural History, 10th series, vol. 17, pp. 413-428. Hope, F. W. 1837. Descriptions of some species of Carabidae collected by Charles Darwin, Esq., in his late voyage. Transactions of the Ento- mological Society of London, vol. II, pp. 128-131. HorRN, WALTHER 1915. Coleoptera Adephaga, Fam. Carabidae, Subfam. Cicindelinae. Genera Insectorum, fase. 82c, pp. 209-486, pls. 16-23. HowarpD, LELAND O. 1889. Scientific results of explorations by steamer Albatross, no. 5. An- notated catalogue of the insects collected in 1887-88. Proceedings of the United States National Museum, vol. 12, pp. 185-216. LAMEERE, AUGUSTE 1902. Revision des Prionides. Annales de la Société Entomologique de Belgique, vol. XLVI, pp. 59-111. LESNE, P. 1910. Notes sur les Coléoptéres Térédiles. Les Bostrychides des Iles Gala- pagos. Bulletin du Muséum National d’Histoire Naturelle, vol. 16, pp. 183-186. LINELL, MARTIN L. 1898. On the coleopterous insects of Galapagos Islands. Proceedings of the United States National Museum, no. 11438, pp. 249-268. MuTCHLER, ANDREW J. 1925. Coleoptera of the Williams Galapagos Expedition. Zoologica, vol. V, pp. 219-240. 1938. Coleoptera of the Galapagos Islands. American Museum Novitates, no. 981, pp. 1-19. 162 CALIFORNIA ACADEMY OF SCIENCES [Oc. PAPERS PIERCE, W. DWIGHT 1912. Systematic notes and descriptions of some weevils of economic or biological importance. Proceedings of the United States National Museum, vol. 42, pp. 155-170. REGIMBART, M. 1907. Essai monogravhique de la famille des Gyrinidae, 3e supplément. Annales de la Société Entomologique de France, vol. 76, pp. 137-245. SHARP, DAVID 1891. Trogositidae: Jn Biologica Centrali-Americana. Insects, Coleoptera, vol. II, pt. 1, pp. 388—437. SLEVIN, JOSEPH R. 1931. Log of the schooner “Academy” on a voyage of scientific research to the Galapagos Islands, 1905-1906. Occasional Papers of the California Academy of Sciences, no. XVII, pp. 1-162, pls. 1-16. WALLACE, ALFRED RUSSELL 1876. The geographical distribution of animals. Vol. 1, xxi+503; vol. 2, 607 pp. Harper and Brothers, New York. 1895. Island life. 2nd edition. xx-+563. Macmillan and Company, London. WATERHOUSE, C. 1877. Coleoptera. In Gunther, Account of the zoological collection made during the visit of the H.M.S. “Petrel’ to the Galapagos Islands. VII. Proceedings of the Zoological Society of London, no. 5, pp. 77-82. WATERHOUSE, GEORGE R. 1845. Descriptions of coleopterous insects collected by Charles Darwin, Esq., in the Galapagos Islands. Annals of Natural History, vol. XVI, pp. 19-41. WILLIAMS, F. X. 1907. |Report of the expedition of the California Academy of Sciences to the Galapagos Islands, stating that 150 species of Coleoptera were taken.] Entomological News. vol. XVIII, pp. 260-261. PLATE 1 . Calosoma howardi Linell . Cicindela galapagoensis W. Horn Calosoma darwinia, new species Right wing of Calosoma howardi Linell Right wing of Calosoma darwinia, new species . Calosoma galapageium Hope . Calosoma linelli Mutchler [ 164 ] [VAN DYKE] Pate 1 OC. PAPERS CALIF. ACAD. SCI., No. 22 E| 3! | 2 —— 4, CONDO OF WwW DF H i=) PLATE 2 . Scarites galapagoensis Linell . Selenophorus obscuricornis (G. R. Waterhouse) . Scarites williamsi, new species . Feronia duncani, new species . Feronia galapagoensis G. R. Waterhouse . Feronia calathoides G. R. Waterhouse . Agonum darwini, new species . Prothorax of Agonum darwini, new species . Prothorax of Agonum chathami, new species . Agonum albemarli, new species [ 166 ] 2 [VAN DYKE] Prater No. 22 OC. PAPERS CALIF. ACAD. SCI., eS Ne Oe Oe Stomion Stomion Stomion Stomion Stomion Stomion Stomion Stomion PLATE 3 galapagoensis G. R. Waterhouse helopoides G. R. Waterhouse cribricollis, new species longulum, new species laevigatum G. R. Waterhouse linelli Blair longicornis, new species rugosum, new species [ 168 ] [VAN DYKE] PLATE 3 OC. PAPERS CALIF. ACAD. SCI., No. 22 aNanarnrrwn kr PLATE 4 . Ammophorus galapagoensis G. R. Waterhouse . Ammophorus galapagoensis subpunctatus, new subspecies . Ammophorus galapagoensis laevis, new subspecies . Ammophorus cooksoni C. Waterhouse . Ammophorus obscurus G. R. Waterhouse . Ammophorus bifoveatus G. R. Waterhouse . Ammophorus abingdoni, new species . Ammophorus insularis Boheman [ 170 ] [VAN DYKE] Ptate 4 oc. PAPERS CALIF. ACAD. SCI., No. 22 © 00 Aa TP wD . Pedonoeces . Pedonoeces . Pedonoeces . Pedonoeces Pedonoeces Pedonoeces . Pedonoeces . Pedonoeces . Pedonoeces PLATE 5 wenmani, new species caudatus, new species bauri Linell galapagoensis G. R. Waterhouse spatulatus, new species costatus G. R. Waterhouse pubescens G. R. Waterhouse lugubris (Boheman) barringtoni, new species [172 | [VAN DYKE] PLaTeE 5 OC. PAPERS CALIF. ACAD. SCI., No. 22 aAaNAantrrwnr rr PLATE 6 . Phaleria manicata Boheman . Pelonium longfieldi Blair . Rhacius costipennis Blair Conoderus galapagoensis, new species . Neoryctes galapagoensis (G. R. Waterhouse) . Chrysobothris williamsi, new species . Trox galapagoensis, new species . Elytra of Trox seymourensis Mutchler | 174 ] [VAN DYKE] PLATE 6 SCI., No. 22 OC. PAPERS CALIF. ACAD. PLATE 7 . Leptostylus galapagoensis, new species . Acanthoderes galapagoensis Linell . Lembodes subcostatus, new species Estola galapagoensis Blair . Strongylaspis krapelini Lameere Acryson galapagoensis Linell Pantomorus galapagoensis Linell . Gerstaeckeria galapagoensis, new species . Amphideritus cuneiformis (G. R. Waterhouse) [176 ] [VAN DYKE] Pate 7 OC. PAPERS CALIF. ACAD. SCI., No. 22 ue] yas & phew es 7) - ti s aah Hh ae ae de hid A 7 ee ' vw , No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 179 abingdoni, Ammophorus, 96 Ablechrus, 40 Acanthoderes, 136 Achryson, 131 Aeolus, 54 aequatorialis, Bledius, 37 Agonum, 25 albemarli, Agonum, 27 Allecula, 67 Alleculidae, 66 Alloxacis, 43 Alphitobius, 118 alutaceous, Eupactus, 121 amabilis, Eburia, 134 Ammophorus, 83 Amphicerus, 121 Amphideritus, 142 Anchastus, 52 Anchonus, 155 Anobiidae, 120an antennatus, Ammophorus, 90 Anthribidae, 142 apicalis, Pedonoeces, 105 apicalis, Compsa, 134 arrowi, Ataenius, 5, 123 Ataenius, 5, 123 bauri, Pedonoeces, 106 barringtonensis, Gerstaeckeria galapagoensis, 152 barringtoni, Ammophorus bifoveatus, 96 barringtoni, Pedonoeces, 109 basillaris, Thermonectes, 32 batesoni, Geraes, 154 batesoni, Ochthebius, 35 batesoni, Pedonoeces, 110 batesoni, Physorhinus, 51 batesoni, Saprinus, 39 becki, Feronia galapagoensis, 23 beebei, Tachys, 16 Bembidon, 14 bifoveatus, Ammophorus, 95 bisigma, Psyllobora, 66 bitoma, 64 blairi, Ammophorus, 89 blairi, Feronia, 19 blairi, Pantomorus, 146 blairi, Pedonoeces, 103 blairi, Physorhinus, 50 Bledius, 37 Bostrichidae, 121 Bruchidae, 141 Bruchus, 141 Buprestidae, 56 calathoides, Feronia, 16 Calocladon, 40 Callosomlar 355.001) 00 Cantharidae, 40 Carabidae, 7 Carcinops, 38 INDEX carnivorus, Dermestes, 59 caroli, Pantomorus, 147 caudatus, Pedonoeces, 104 Cerambycidae, 127 chathami, Agonum, 26 chathami, Physorhinus, 51 Chauliognathus, 40 Chrysobothris, 56 Chrysomelidae, 139 Cicindela, 4, 6 Cicindelidae, 6 Cissites, 45 Cleridae, 41 Coccinellidae, 65 collenettei, Alloxacis, 43 Colydiidae, 64 Compsa, 134 Conoderus, 45 conwayensis, Pantomorus, 149 cooksoni, Ammophorus, 91 Copelatus, 31 Copris, 122 Coptostethus, 56 cornutus galapaganus, Amphicerus, 121 cornutus, Gnathocerus, 118 costatus, Pedonoeces, 11 costopennis, Rhacius, 118 Creophilus, 37 cribrata, Estola, 134 cribricollis, Stomion, 75 crockeri, Pantomorus, 147 Cteisa, 69 Cucujidae, 62 culpepperi, Pedonoeces, 102 cuneaticollis, Mastogenius, 58 cuneiformis, Amphideritus, 142 cunicollis, Neopentarthrum, 157 Curculionidae, 142 darwini, Agonum, 25 darwini, Docema, 140 darwini, Galapagodacnum, 37 darwini, Calosoma, 10 Dendrophilinae, 38 dentatus, Prateus, 119 denticollis, Ammophorus, 94 denticollis, Trichodesma, 120 Dermestes, 59 Dermestidae, 59 Desmiphora, 134 Diabrotica, 139 dichroa, Physorhinus, 47 Docema, 140 Dryotribus, 155 Dytiscidae, 31 duncani, Estola, 134 duncani, Feronia, 21 duncani, Pedonoeces, 110 Eburia, 132 Elateridae, 45 180 VAN DYKE: GALAPAGOS COLEOPTERA Enochrus, 36 equatoriale, Bembidion, 15 Eretes, 34 Estola, 134 Eupactus, 120 exarata, Bitoma, 64 Feronia, 16 (Key, 24) flavipes, Ablechrus, 40 fuscatus, Aeolus, 54 fuscipes, Parepitragus, 82 fuscomaculatus, Bruchus, 141 galapageium, Calosoma, 3, 5, 11 Galapagodacnum, 37 galapagoensis, Acanthoderes, 136 galapagoensis, Achryson, 131 galapagoensis, Allecula, 67 galapagoensis, Ammophorus, 85 galapagoensis, Anchastus, 53 galapagoensis, Anchonus, 155 galapagoensis, Bembidion, 14 galapagoensis, Carcinops, 38 galapagoensis, Cincindela, 4, 6 galapagoensis, Conoderus, 45 galapagoensis, Copelatus, 31 galapagoensis, Docema, 140 galapagoensis, Estola, 134 galapagoensis, Feronia, 22 galapagoensis, Gerstaeckeria, 151 galapagoensis, Grammophorus, 55 galapagoensis, Gyrinus, 34 galapagoensis, Leptostylus, 136 galapagoensis, Lobopoda, 67 galapagoensis, Longitarsus, 140 galapagoensis, Mastogenius, 57 galapagoensis, Mordellistena, 44 galapagoensis, Neoryctes, 125 galapagoensis, Oxacis, 42 galapagoensis, Pantomorus, 144 galapagoensis, Parandra, 127 galapagoensis, Pedonoeces, 107 galapagoensis, Scarites, 13 galapagoensis, Selenophorus, 28 galapagoensis, Spermophagus, 141 galapagoensis, Stenodontes (Mallodon), 130 galapagoensis, Stomion, 71 galapagoensis, Temnochila, 60 galapagoensis, Thaptor, 120 galapagoensis, Thermonectes basillaris, 33 galapagoensis, Trox, 124 georgicus, Eupactus, 120 Geraes, 154 Gerstaeckeria, 151 glabrum, Neepentarthrum, 158. Gnathocerus, 118 Grammophorus, 55 gracilis, Macrancylus, 156 Gyrinidae, 34 Gyrinus, 34 hayi, Taeniotes, 139 helopoides, Stomion, 73 Heterocrepidius, 54 hirticollis, Desmiphora, 134 Histeridae, 38 hoodensis, Gerstaeckeria galapagoensis, 153 hoodi, Alloxacis, 43 hoodi, Physorhinus, 49 howardi, Calosoma, 7 Hydrophilidae, 35 insulare, Pycnarthrum, 160 insularis, Allecula, 68 insularis, Ammephorus, 87 insularis, Estola, 134 insularis, Feronia, 18 insularis, Stelidota, 61 interruptus, Passalus, 126 kraepelini, Strongylaspis, 13@ labrale, Metachroma, 139 laevigatum Stomion, 77 laevigatus, Alphitobius, 118 laevis, Ammophorus galapagoensis, 86 Langurlidae, 64 lanigera, Eburia, 132 lateralis, Tropisternus, 35 Lembodes, 154 Leptostylus, 136 limbata, Diabrotica, 139 linelli, Calosoma, 12 linelli, Neoryctes, 126 linelli, Stomion, 78 Lobopoda, 67 longifieldae, Pelonium, 41 longicornis, Stomion, 79 longicornis, Tetrapriocera, 121 Longitarsus, 140 longulum, Stomion, 76 lugubris, Copris, 122 lugubris, Pedonoeces, 113 lunatus, Longitarsus, 140 Lycidae, 40 Macrancylus, 156 maculata, Cissites, 45 maculatus, Dermestes, 59 manicata, Phaleria, 117 Mastogenius, 57 Meloidae, 45 Melyridae, 40 Metachroma, 139 Micrapate, 121 mimeticus, Dryotribus, 155 modestior, Saprinus, 39 molarius, Stenodontes, 129 Monotomidae, 63 Mordellidae, 44 Mordellistena, 44 morio, Pedonoeces, 109 mutchleri, Feronia, 20 mutchleri, Neopentarthrum, 158 [Oc. PAPERS No. 22] VAN DYKE: GALAPAGOS COLEOPTERA 181 Necrobia, 41 scutellaris, Ataenius, 123 Neopentarthrum, 157 (Key, 159) Selenophorus, 28 (Key, 31) Neoryctes, 125 _ seymourensis, Alloxacis, 43 Nitidulidae, 61 seymourensis, Gerstaeckeria ‘ galapagoensis, 153 obesum, Stomion, 74 seymourensis, Trox, 124 obscuricornis, Selenophorus, 30 signatus, Rhantus, 32 obscurus, Ammophorus, 93 simplex, Ammophorus, 87 obscurus, Enochrus, 36 spatulatus, Pedonoeces, 108 Ochthebius, 35 Spermaphagus, 141 Oedemeridae, 41 Staphylinidae, 37 opacus, Pedonoeces, 114 Steliodota, 61 Ormiscus, 142 Stenodontes, 129 Ostomidae, 60 sticticus, Eretes, 34 Oxacis, 42 Stomion, 69 (Key, 81) Pantomorus, 143 (Key, 150) subeosraiue, ‘heinbodes, 6d Becca a? suberosus, Trox, 123 Paeenlidan’ 126 subpunctatus, Ammophorus galapagoensis, 86 Passalus, 9126 sulphureus, Chauliegnathus, 40 peninoides, Cteisa, 69 Pedonoeces, 99 (Key, 115) i Tachys, 16 ee FE Taeniotes, 139 Physorhimus, 47 (Key, 52) Temnochila, 60 Physorhini, 46 Tenebrionidae, 69 pilosa, Oxacis, 42 Tenebroides, 61 Plates, 164 tenellus, Carcinops, 39 Platypodidae, 159 testaceum, Calocladon, 40 Platypus, 159 Tetrapriocera, 121 Prateus, 119 Thaptor, 120 proletaria, Eburia, 132 Thermonectes, 32 é Psyllobora, 66 towerensis, Neopentarthrum, 157 Trichodesma, 120 Tropisternus, 35 Trox, 124 punctipennis, Stomion galapagoensis, 72 puberulus, Heterocrepidius, 54 pubescens, Pedonoeces, 102 EcgiaNn ae, ali uniformis, Pedonoeces, 105 variegatus, Ormiscus, 142 quirsfeldi, Physorhinus, 47 yillosus, Creophilus, 37 References, 160 vonhageni, Cincindela, 7 Rhacius, 118 Rhantus, 32 waterhousei, Enochrus, 36 ruficeps, Physorhinus, 48 waterhousei, Feronia, 17 rufipes, Necrobia, 41 wenmani, Pedonoeces, 100 rugosum, Stomion, 80 wenmani, Selenephorus, 29 williamsi, Anchastus, 53 santacruzensis, Platypus, 159 williamsi, Chrysobothris, 56 Saprininae, 39 williamsi, Coptostethus, 56 Saprinus, 39 williamsi, Feronia, 21 scabratus, Micrapate, 121 williamsi, Leptostylus, 137 Scarabaeidae, 122 williamsi, Pantomoris, 150 Scarites, 13 williamsi, Pedenoeces, 112 Scolytidae, 160 williamsi, Scarites, 13 my - ae * » sey a TONS tye Rp dee be ' a gh F aa | a , GAY aie *. wee hi h es th a 1 hy TEMES pes i iat é if i) i Agen! iy : hey \ aa ei Sn a Hh agi ae at iy +) Nig ae i , ne z pa i'h4 SMITHSONIAN INSTITUTION LIBRARIES MINEO 9088 01302 6307 | , ’ % ry i nit pint ue ; HY 7 iit i ; i i t a fi strate al vee Wy an } iff 2 Harti