Taw egortay. Fava : Saiwije Ne Rewye 5h ph, whe ipietintenites, Ped shete tI Wy gt og Footie Tasgenet eh Loe if if yi uM aan tial i 1 hy HS De BULLETIN OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE, IN CAMBRIDGE VOL. 107 CAMBRIDGE, MASS., U.S. A. 1952-1953 Tue Cosmos Prsss, INc. CAMBRIDGE, Mass., U.S. A. CONTENTS 1.—TuHe AMPHIBIANS OF THE SOLOMON ISLANDS. By Walter C. Brown. (8 plates.) August, 1952 2—REvVISION OF THE ANT GENUS SERRASTRUMA. By William L. Brown, Jr. August, 1952 3.—THE CaraBip BEETLES oF NEw Gutnka. Part 2. The Agonini. By P. J. Darlington, Jr. (4 plates.) August, 1952 : : 4—REVISION OF THE SPECIES CURRENTLY REFERRED TO ALEPOCEPHALUS, HALISAURICEPS, BATHYTROCTES AND BAJACALIFORNIA, WITH INTRODUCTION oF Two NEw GENERA. By A. E. Parr. September, 1952 . 5.—GEOGRAPHIC VARIATION IN THE RED-EYED TOWHEE OF THE EASTERN UNITED StTaTEs. By Us: C. Dickinson, Jr. October, 1952 6.—CYATHASPID FISHES FROM THE VERNON SHALE OF New York. By R. H. Flower and R. W ten Smith. (8 plates.) October, 1952 7—NoTES ON SOME PETRELS OF THE NortH Pactiric. By Oliver L. Austin, Jr. November, 1952 8.—AIRPLANE OBSERVATIONS OF Homine PicEons. By Donald R. Griffin. December, 1952 9—Tuer APSEUDID CHELIFERA OF THE EASTERN TROPICAL AND NortH TEMPERATE Paciric OckAN. By Robert James Menzies. February, 1953 .10—On THE EARTHWORMS OF THE ARNOLD ARBORETUM, Boston. By G. E. Gates. February, 1953 .11—A New Fossit TortToisE FROM THE THOMAS FaRM Miocene OF Fiorina. By Ernest Williams. February 1953 PAGE 1 65 87 497 535 ms kin Bete i ta wiranenhss te 1h ais Yael "Ree ac F kan ae sort i Wve ive. s 4 aa cvanatta’ et fh Vides f He t bi rena itt) yee eee ee ak Pape iE ah i Me 4 i/ 7. ii % peas } peas ‘a Rat an caine 126 T ae ATL. yy Rae ities hi ‘ae Lu 1D it enunasl A. “ 4! bi Ra il SO ie ee Bi i Ae, Wi DPA ipl wa at VA? hyd Y Sh Tie Misc, sid pny ARS Re, ret a e erep ii) bt : [eigen 7 ; me sane ts it sii wie io UT. ti Ce ee Von cs rae neil ei ae Hate fii ANY Rae Tae | aie rat BAGUER I Aesth % et Pia Mies) A EAS Ou Cisse pak een is) evra ea i late Moda! hb Ge 190), Sie IaiA et Sepa t/ iy } ; ae Be a Anas nents: on vast eee Len HAs gay Mel te Prt autho ALS Pea grained readrasif an 4 "3 6 nn q i i? f one)! 0 vy 7 oy i VV 5 LU aa a ' te wore * Naat | Bi ‘ , ve oy x , rae *) \ : ’ Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vot. 107, No. 1 THE AMPHIBIANS OF THE SOLOMON ISLANDS By WALTER C. Brown Wit Erent PLaTEs CAMBRIDGE, MASS., U. S. A. PRINTED FOR THE MUSEUM August, 1952 PUBLICATIONS ISSUED BY OR IN CONNECTION WITH THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE BULLETIN (octavo) 1863 — The current volume is Vol. 107. BrEviora (octavo) 1952 — No. 4 is current. Memorrs (quarto) 1864-1938 — Publication was terminated with Vol. 55. JOHNSONIA (quarto) 1941 — A publication of the Department of Mollusks. Vol. 2, no. 30 is current. OccAsIONAL PAPERS OF THE DEPARTMENT OF MOLLUSKS (octavo) 1945 — Vol. 1, no. 16 is current. PROCEEDINGS OF THE NEw ENGLAND ZOOLOGICAL CLUB (octavo) 1899- 1948 — Published in connection with the Museum. Publication terminated with Vol. 24. These publications issued at irregular intervals in numbers which may be purchased separately. Prices and lists may be obtained on application to the Director of the Museum of Comparative Zodlogy, Cambridge 38, Massachusetts. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE VoL. 107, No. 1 TEE AMPHIBIANS OF THE SOLOMON ISLANDS By WaLTerR C. Brown Wits Erexut PLatTEs CAMBRIDGE, MASS., U. S. A. PRINTED FOR THE MUSEUM August, 1952 No. 1. — The Amphibians of the Solomon Islands! By WALTER C. Brown CONTENTS Page LINER ODUCTION cooked seein tare ieee eS ae ona Weg Si OWA PSD RR rhs 4 Ncknowledomientse Mn. c2 tne i: SpE A Mase meatier et hoe ey eae roy 5 Gazetteer of Solomons collecting stations ......................... 6 lhistromnomenclaturaluchangespe mmr seer a ice) Seine ie 7 List of amphibians of the Solomon Islands with an index to species... 7 List of extralimital amphibians discussed in the text................ 8 Notes on the geology and geography of the Solomon Islands......... 8 Nature and distribution of the amphibian fauna................... 11 DIsTRIBUTIONAL CHART SHOWING KNOWN RANGES OF THE SPECIES WITH- LN GRU SOLOMONS cyte cases x sites sian one attra eee ae Same TNR onl 14 Treatment of data in the systematic section....................... 16 SMG REMATCH IS CUSSION Wise a:i alah e mete inne mus pany Aah tae tieaman tama |My ate SINe es 17 Sy AEO MLA TATITES St CPP en Ae AIR a eT VAR a Tee oC acim NRT RRA Soe Les JSLUOROVADLO VeVi rae eaiats an Ta Bias Re MArS Coo cee aE Tn Gin ar Mee Aue OA OU Cena 17 GFOMUS RB Usf Ole B so oa spire Siey gale ON eA deed a be parnrlt i Ae Re Mi Re 17 iy LIGA Apacer ice paracise ts ae asiveap wales NB EtOH a GIO See ee ee eRe Aone 18 (ens PEL Lars cas cess of oR apa Races ea a ee a ee 18 EAI AC Mprt arte ier te ont a ciaier ceo ey ne, eo naa Cs cannes phic ennny ae ara aie oe Aca 23 GreMUSEES CL GCI UY LOM eS\t eras reeds Getter econ SCART 3 ORE INY Rn eee eR 23 GenusiCenatovatrachusea ete til ee ce ee ee oe et ree 28 (GemuUsr@onnuper coe ese atn eee eG ie aaa OE We eer RT AN 31 GenuseDiscodeles: aie ie May 228 NORE PAUSE Se MEAOR ike 35 Conse alm atona pian acetates veh cosa ie ee ee ae neds 44 Genusy Rat mantis: (Aen SL Ae ee ee EY Oa Se LORE OE emia ae 45 (CRT oa aN as, oes ys. SAN ae ee MRE ys et 55 AUB LO GRATE EDV? phrwasorany trey i ein ete eer Mayne skal alisun We ogee OM cei le pag ps 61 IE PUTACIEESS are waite Venere tect etal as Sarat pag EM AN a pA city looney a NE ios 1-8 1 Submitted in partial fulfillment of the requirement for the degree of Doctor of Philosophy in the Department of Biology, Stanford University, April, 1950. 4 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY INTRODUCTION The first general account of the herpetofauna of the Solomon Islands was that of Boulenger (1886). In that account he gave detailed descriptions of the nine species of amphibians and the nineteen species of reptiles then known to inhabit this archipelago. At that time all of these amphibians and five species of the reptiles were known only from the Solomons. In the second general study of this herpetofauna (Kinghorn, 1928) the number of species of amphibians listed was still nine while the number of reptiles had increased to thirty-five. The amphibian list differed, however, from that of Boulenger in that two of the species on his list had been placed in synonymy and two species which had been discovered in the intervening period were included. Since 1928, six species and subspecies of amphibians, including those described by the present writer, have been recognized from the Solo- mons; two earlier described species have been revived; and one species from the Americas has been introduced. Thus, eighteen species and subspecies are now known. ‘tug This growing list of species, which are recorded in numerous, scattered papers, as well as the recent collections made by various men at the time of the occupation of these islands by the United States military forces, make timely a preliminary revision and summary of our knowledge of this part of the fauna of this zoogeographically im- portant archipelago. A total of 1044 amphibian specimens from the Solomons have been examined during the course of the present study! In some instances the available material has shown the existence of subspecific popula- tions within the Archipelago. In other instances suspected populations are represented by inadequate samples and no prediction as to their racial distinctness is made at this time. The number of new forms discovered in recent collections, in correla- tion with our relatively limited knowledge of the interior portions of many of the islands, suggests that additional new species may well remain to be discovered. Certainly the opposite conjecture (Boulenger, 1888a and Barbour, 1921) that the herpetofauna of the Solomons was probably well known has not been substantiated. This study, as well as the author’s work with the herpetofaunas of other Pacific island groups, has shown the great difficulty of determin- ing phylogenetic affinities and the most probable distributional paths. There is a great need of generic and family revisions in both the 1 Highty tour of these are not listed in this paper because they are to be reported by other workers. BROWN: AMPHIBIANS OF THE SOLOMON ISLANDS D amphibian and reptilian faunas which would include all the forms known to inhabit any part of this island region and also their Austra- lian and Asiatic relatives. Acknowledgments An expression of thanks is due especially to Dr. George S. Myers of the Natural History Museum at Stanford University and Mr. Arthur Loveridge of the Museum of Comparative Zoology at Harvard College who have assisted most generously with both time and knowl- edge. I also wish to thank Dr. H. W. Parker of the British Museum for making available to me unpublished distributional tables of the herpetofauna of these islands. I am also deeply indebted to the following persons who have permitted me unlimited use of the Solomon Islands collections deposited at the institutions with which they are associated: Mr. Charles M. Bogert of the American Museum of Natural History, New York; Dr. Doris M. Cochran of the United States National Museum, Washington, D. C.; Mr. Joseph R. Slevin of the California Academy of Sciences, San Francisco; Dr. Lawrence M. Klauber of San Diego, California; Dr. Robert C. Stebbins of the Museum of Vertebrate Zoology, University of California, Berkeley. Thanks are due Dr. Robert Mertens of Senckenberg Museum, Germany; Mr. Karl P. Schmidt and Mr. Clifford H. Pope of the Chicago Natural History Museum and Dr. Vasco M. Tanner of Brigham Young University for the loan of critical material whenever it was needed, even though their Solomon Islands collections were being studied by themselves or other workers at this time. I am also indebted to Mr. Walter L. Necker of Chicago who was so kind as to give his assent to the use of some of his material deposited in the National Museum. Finally, I wish to acknowledge my indebtedness to Miss Virginia Field for some of the drawings of Batrachylodes and Hyla thesaurensis, Mr. William Theiss for the full-figure drawings of Batrachylodes trossulus and Platymantis myersi and especially to Miss Jean Allred for the preparation of all the other drawings illustrating the text and to Mrs. Charles 5S. Richards and my wife, Jeanette S. Brown, who patiently typed and corrected the manuscript. The field work which provided the collections directly used in this study was conducted by: William M. Mann for the Museum of Com- parative Zoology, 1916; J. A. Kusche for the California Academy of Sciences, 1921; Rollo H. Beck for the American Museum of Natural History, 1920-1928; Karl P. Schmidt for the Chicago Natural History Museum, 1929; Maurice Willows, Jr. for the California Academy of 6 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Sciences, 1933; Lowell Adams for the Museum of Vertebrate Zoology, 1944; D. Eldon Beck for Brigham Young University, 1944-1945; John Chattin for the Museum of Vertebrate Zoology, 1944; J. A. Gray for the Museum of Vertebrate Zoology, 1944; J. P. Heath for the Natural History Museum of Stanford University, 1944; L. W. Jarcho for the Museum of Comparative Zoology, 1943-1944; D. H. Johnson for the United States National Museum, 1944; Walter L. Necker for the United States National Museum, 1944; R. C. Pendleton for Brigham Young University, 1944; Ernest Reimschissel for Brigham Young University, 1944; Charles G. Sibley for the Museum of Verte- brate Zoology, 1944. Gazetteer of Solomons Collecting Stations In the following list of localities, whether small, generally unnamed islands on the ordinarily available maps or known collecting stations on the larger islands, latitude and longitude as well as synonyms occurring in the literature are given. Latitudes and longitudes were determined largely from tables in the United States Navy Department Gazetteer, H. O. Pub. No. 881, July, 1944. Locality Remarks Lat. S. Long. E. Arnavon Id. (Isabel group, off northwest coast) (2264 als Sn O Ata (see Atta) Atta, Malaita Id. (cove area on northeast coast) Sd MlG0s504 Auki, Malaita Id. (village on west coast) 8°47’ 160°43’ Banika Id. (2nd largest in Russell group) 95054 eel Soman Bio Id. (San Cristobal group, off north - coast) TOMO. Seen Fauro (Faro) Id. (Bougainville group, off south coast) 7°47’ 158°37’ Florida Id. (Nggela group, between Guadalcanal and Malaita Ids.) 9°05’ 160°16’ Fulakora Pt., Isabel Id. (east coast near southern end) So2 1 Na Oeil Gatukai Id. (New Georgia group, southeast end) 8°47’ 158°12’ Gela Id. (not positively known, perhaps in Florida group) Gizo Id. (New Georgia group, northwestend) 8°05’ 156°49’ Ganongga Id. (see Ronongo Id.) Kirigi River area, Florida Id. 9°08’ 160°16’ Kolombangara Id. (New Georgia group, northwestend) 8°00’ 157°05’ Mono Id. (Bougainville group, south end) (2D DS OmS On Munda, New Georgia (southwest coastal area of New Georgia) Salo ats 7ellon BROWN: AMPHIBIANS OF THE SOLOMON ISLANDS U Locality Remarks Lat. S. Long. E. Narovo Id. (New Georgia group, southwestern end) 8°16’ =156°31’ Neggela Id. (see Florida Id.) Puruata Id. (Bougainville group, off western coast) Oe 5 yells 5205s Rendova Id. (New Georgia group, off south- western coast) Srl aiic20% Ronongo Id. (New Georgia group, northwest end) 8°03’ 156°35’ Roviana Lagoon area, New Georgia Id. Sel Gmmmaliaiielyr Rubiana Lagoon area = Roviana Lagoon area Russell Id. (between New Georgia and Guadal- canal) 9°04’ = 115 9°12” Santa Ana Id. (San Cristobal group, southeast end) 10°50’ 162°28’ Santa Ysabel Id. (see Isabel Id.) Simbo Id. (see Narovo Id.) Shortland Id. (Bougainville group, southern end) (gOS Gmelo nea se Stirling Id. (Bougainville group, southern end) ee SO OO Tenaru River (Guadalcanal Id., north coast) 9°25’ 160°07’ Tertere area (Guadalcanal Id., north coast) OS2 dv LGOeTAY, Torokina Pt. area (Bougainville Id., west coast) O22 lob sOly Treasury Id. (see Mono Id.) Tulagi Id. (small island in Florida group) 9°06’ 160°09’ Ugi Id. (San Cristobal group off northeast coast) 10°14’ =161°44’ Vangunu Id. (New Georgia group off southeast end) 8°39’ —158°00’ Vella Lavella Id. (New Georgia group, northwestend) 7°43’ 156°40’ Yandina, Pavuvuld. (Russell group) DLO ae a9 els List of Nomenclatural Changes Platymantis webert = Platymantis papuensis weberi Rana bufoniformis = Discodeles bufoniformis Rana guppyt = Discodeles guppyt Rana opisthodon = Discodeles opisthodon Rana krefftit = Rana papua krefftu List of Amphibians Known From the Solomon Islands BUFONIDAE Page BO TROP ACIP (ULNA) 4'6 |G BAG tyne dodo udon eh oebedbooonobe Be 17 HYLIDAE i lomiiteamaOuleneeemewary sony ict terete cece i ckale Gin eet 19 EM LOMNESGUUETISTS MAC LELS ER AU ice ateeets he le ewer et ce oleae che ie aes ci 20 8 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY RANIDAE Page Batrachylodes trossulus) Brown and Myers)...). 2.» > 45 eee 24 iBatrachylodeswertebralzsyBoulenger: 944 445450 eee ooo eee 26 Ceratobatrachusiquentherd Boulenger.4.-4 ease eee ae eee 29 Cornufersquppys Boulengert: Nicosia. eee oo ea ae ee eee 32 CornujermeckermBrowm and Myers...) 4 o> sees eee 2s oe eee 34 Discodeles bujoniformis (Boulenger))-\s-24 4-222 450 6256 oe eee 37 Discodelesigup pyc (Boulenger)== +). 0. eee eee ae ee eee 39 Wiscodeles opisthodon,(Boulenger)...-5--4ee- ces. oe eee 42 Palmatorappia solomonis (Sternteld). = yase 4+ ee 22a eee 44 Piatymantts aculeodactylus Sp nova jee see se ee 46 iPlarjmantis myersiplOw le seer oer are on ane eee 48 iRiatymantis) papuenstsywebert Schmidt. ar (see aoe eee eo ilatymantisssolomonts (Boulenger) ae n eeeee eee eee 53 Rana (Hylarana) papua krefftii Boulenger.....................-.-- 56 Rana (Hylarana) papua novaebritanniae Werner............--..+--- 58 List of Extralimital Amphibians Discussed in the Text Connufer vitiensis (Girard) PS85a tee. a ee ee ere eee eee 12,32 Discodeles bufoniformis cognatus (Hediger), 1934.................... 36 Discodeles'ventrzcosus (Vogt), LOI 2 Saale ae) ee eee 36 Platymantis beaufort: (van) Kampen), 1913582) 2250). eee 12, 48 Platymantis cheesmanae)| Parker, MiO4Ok Seeeniieys rentals ae teeta 12, 46 Platymantis corrugatus (A. Dumeértl)) 1853 2732s eee 50 Platymantis papuensis papuensis Meyer, 1874.................5..5 50 Platymantis viranus (Av Dumeénil) 18535255228. 22 os eee 12, 48 Rane papua papua Lesson, 1830) eee ces = 2 oi. aye) alte eee 12 Notes on the Geology and Geography of the Solomon Islands A study of the relationship of any element of the fauna of these islands to corresponding elements in surrounding island areas necessi- tates some consideration of the general features of their geology and geography. This is necessary in order to determine the possible bearing of these factors on the distributional and evolutionary patterns which systematic considerations suggest. Chubb (1934, pp. 289-302) calls attention to the fact that the andesitic zone which borders the Pacific shores of the American and Asiatic continents includes a roughly quadrangular, submerged area extending far into the Pacific Ocean from southeastern Asia. Its northeastern border extends from the island of Honshu through the Caroline Islands to a point just northeast of the Fijis; the shorter BROWN: AMPHIBIANS OF THE SOLOMON ISLANDS 9 southeastern border extends southwestward to a point near the eastern shore of South Island, New Zealand; the long southwestern border lies between Australia and New Guinea and along the southern shores of the lesser Sundas, Java and Sumatra. This area is presumed to represent a continental extension subjected to extensive fracture and eruptive emergence during Mesozoic and Tertiary times. Many of the island groups within these limits tend to lie in northwest-southeast lines or arcs in the form of ridges but partly raised above the level of the ocean. These marginal, exposed ridges, according to one hypothe- sis, have probably never been directly connected to the adjacent continental mass. The fluctuations in the level of the now submerged portions of these ridges and the interlying basins can be presumed only on the indirect evidence of floral and faunal relationships and the direct evidence buried beneath the ocean floor (see also Myers, 1950). The Solomons Archipelago forms a double chain some 600 miles in length (exclusive of the Santa Cruz group). This extends in a west- northwest to east-southeast line and apparently forms a continuation _ of the ridge system occupied by New Ireland and the Admiralty Islands. There are seven major groups of islands in the Solomons: the Bougainville, Choiseul, Isabel and Malaita groups on the north chain; the New Georgia, Guadalcanal and San Cristobal on the south. These do not represent two strictly parallel ridges, however. Bougainville, Choiseul, Isabel and Guadalcanal groups rest on one continuous undersea plateau which is submerged to a depth of not more than 600 fathoms; the other groups, New Georgia, Malaita and San Cristobal, are separated from this plateau and from each other by basins of greater depth, 1000 to 2000 fathoms (Lever, 1937, p. 272). The larger islands, as Guadalcanal and Bougainville, are marked by a series of volcanic peaks, generally inactive in very recent history, which may attain heights of 8,000 to 10,000 feet. Old voleanic rocks are now covered by sedimentary rocks generally only at lower altitudes, having been denuded at heights above 600 to 1000 feet. There is evidence, however, based on these remaining sedimentary rocks, that they were formed, in part at least, at depths of 12,000 feet or more beneath the surface of the ocean. This, if true, attests to the great orogenic changes which have taken place in this region. The distances which now separate these major island groups within the Archipelago as well as that between the most northern group and the Bismarck Islands are not great. This in correlation with certain features of climate and ocean currents may have some bearing on the distribution of some elements of the fauna. These distances are given in the following table to the nearest mile (-+ or —) as determined from 10 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY the U. S. Navy hydrographic charts, 1945 edition (Nos. 5593, 5912, 2896, 2926, 5967, and 2920). The breadth of the present ocean basins between the larger islands of the seven primary groups ranges from about 25 to 100 miles. It should New Guinea Rooke (Umboi) 29— Rooke New Britain 12+ New Guinea Manus (Admiralties) 175 Manus Mussau 150 New Britain New Ireland 20+ New Ireland Nissan Island 70— Nissan Buka 42— New Ireland Buka 105 Buka Bougainville 2or3 Bougainville Shortland = ar Bougainville Fauro 9— Shortland Mono 17+ Mono Vella Lavella 60+ Bougainville Choiseul 30+ Vella Lavella Kolombangara 15— Kolombangara Arundel Tes Arundel New Georgia 2- Choiseul Rob Roy ies Rob Roy Wagina 5+ Wagina Arnavon == Arnavon Gag 12— Wagina Gagi 26+ Gagi Barola ee Barola Isabel i] 25 Isabel New Georgia 65+ Isabel Malaita 50— Isabel Florida 35— Isabel Guadalcanal 50— Florida Guadalcanal 15-— New Georgia Vangunu il == Vangunu Gatukai 5- Gatukai Payvuvu 60+ Gatukai Guadalcanal 100 Pavuvu Guadalcanal 32+ Florida Malaita 25+ Guadalcanal Malaita 32+ Guadalcanal San Cristobal So Guadalcanal Rennell 105 San Cristobal Rennell 100 San Cristobal Malaita 40— BROWN: AMPHIBIANS OF THE SOLOMON ISLANDS 11 be noted, however, that the greater distances are in all instances re- duced by the presence of small islands at intermediate points. These small islands might well function as stepping stones in the process of faunal dispersal by flotsam methods. The humid, tropical climate and heavy rainfall support a rich covering of vegetation, mainly tropical forest. There are, however, extensive grassy areas on some of the larger islands, Guadalcanal, for example (Guppy, 1887b, p. 25). Short, often large, and generally rapid streams occur on the islands but freshwater lakes are few. Nature and Distribution of the Amphibian Fauna The amphibian fauna of the Solomon Islands is zoogeographically very interesting because of the highly endemic nature of its ranid component and the fact that this Archipelago is, with the Fijis and New Zealand, a Pacific amphibian outpost. Three genera, Batrachy- lodes, Ceratobatrachus and Palmatorappia, are known only from the Solomons. Discodeles has been recorded for the Bismarcks and the Admiralties outside of the Solomons. Cornufer and Platymantis have more extended ranges. Both are represented by distinct species in the Fiji Islands to the east (see Brown and Myers, 1949a), while Platy- mantis ranges west to Borneo and Cornufer as far as Burma. Eso occur also in the Philippines. All of these genera (Batrachylodes possibly exe pied) are apparently closely related, as held by Noble (1931, pp. 522-524). They are possibly descended from a single, more primitive a stock. How- ever, further work both from the anatomical and embryological ap- proaches is needed in order to outline with greater assurance the probable lines of evolution within the group and to reconstruct the theoretical prototype or prototypes. The close relationship between Ceratobatrachus, Cornufer, Discodeles and Platymantis is shown not only in such skeletal features as the broadly forked omosternal style and the large, broad nasals, but also in certain reproductive modifications. The eggs of species within these genera, as far as I have been able to observe (fourteen of twenty- nine recognized species), are relatively large and unpigmented. This suggests that the larvae of all the species may complete their develop- ment within the egg capsule as is, indeed, known for Cornufer guenthert and Discodeles opisthodon. Maturing, ovarian eggs of Palmatorappia are unpigmented. In Batrachylodes the omosternal style is unforked and it may possibly be derived from a hylaranid stock as Noble (1931, p. 521) 12 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY proposes for Micrixalus and Simomantis. However, eggs of B. trossulus are relatively large and unpigmented and a modification of developmental habits similar to that of the genera Cornufer and Discodeles is also suggested for this genus. The Solomons representatives of Platymantis and Cornufer, the only two genera having ranges known to extend much beyond the Archi- pelago, appear to have their closest affinities with Fijian and New Guinean species. In the first-mentioned genus P. solomonis is probably most closely related to P. vitianus of the Fijis, both are very large, rather smooth forms; P. aculeodactylus is closely related to P. chees- manae of New Guinea; P. myerst may have its closest affinities with P. beauforti of New Guinea; and P. papuensis weber is regarded as but subspecifically differentiated from P. papuensis papuensis of New Guinea. In the second genus C. guppyi is related to C. vitiensis of the Fijis but is a larger form, while C. neckeri is a very distinct species and appears not to be related closely to other known forms. The presumably phylogenetically older hylids, so numerous in Australia and New Guinea, are represented by only two known species. One of these, Hyla thesaurensis, is also recorded from New Guinea! Rana (Hylarana) is also known from only two representa- tives and both are regarded as subspecifically related to the common New Guinean form, Rana (Hylarana) papua papua. The members of these two genera, as far as known from the Solomons, lay small, typically pigmented eggs. This suggests, as is known for Hyla thesau- rensis, that the eggs are laid in permanent or semi-permanent bodies of water and that the larvae undergo a period of development and metamorphosis after hatching. Thus the amphibian fauna of the Solomons appears to be comprised of two elements. The older element is a rather closely related, possibly diphyletic, endemic group of highly specialized ranid frogs, the members of Batrachylodes, Ceratobatrachus, Cornufer, Discodeles, Palmatorappia, and Platymantis (in part). The more recent element appears to include members of Hyla, Rana (Hylarana) and two (possibly three) species of Platymantis. The older ranid element has its present center of abundance in the Solomons Archipelago and possibly the Bismarcks. Only two of the genera, as already noted, have ranges much beyond this region. Two hypothetical explanations are suggested for the more extended ranges of Platymantis and Cornufer: (1) that they represent the older, more widely dispersed, possibly ancestral genera as held by Noble (1931, 1The possible subspecific status of this New Guinean population remains for future in- vestigation. BROWN: AMPHIBIANS OF THE SOLOMON ISLANDS 13 pp. 521-524) or (2) that they have simply been the more successful in spreading through an island range. The specialization of large, unpigmented eggs, and probably general practice of direct development, parallels that of certain plethodontids, brevicipitids, and leptodactylids. Such a specialization, it may be hypothesized, could have enabled these frogs to maintain themselves near their present center in the Solomon Islands at some time in the past, under conditions when permanent or semipermanent bodies of still or slow moving water may have been less readily available as breeding sites. On the other hand, the Hylas and Hylaranas, which on the basis of their closer affinities with the New Guinean forms would appear to be later arrivals, may have been able only to invade or at least maintain themselves in the Solomons when suitable breeding sites became more generally available. To this group of later arrivals would also belong the two (or three) species of Platymantis, which are very closely related to the Papuan species. Geological evidence, as far as known, would in general be consistent with this hypothesis of a possible change in the physical nature of the islands in relatively recent times. There has been a general uplifting of many of the islands to a height of sometimes hundreds of feet (Guppy, 1887b, pp. 125-136). Their general profile suggests that prior to this they may have been even more steep and rugged, mountainous islands with more limited marginal lowland areas. When the amphibian faunas of different island groups within the Solomons are compared, in contrast to comparisons of the fauna of the Archipelago as a whole with those of the Bismarcks and New Guinea, a striking general homogeneity is observed. This is modified in two ways, however: (1) the northern group of islands, Bougainville and Choiseul, appear to have a richer amphibian fauna and San Cristobal at the extreme south the poorest; (2) inter-island subspecific popula- tions are known for some species and when more extensive collections become available may be substantiated for others. The more limited fauna of the San Cristobal group suggests that the southern part of the chain has perhaps always been more completely isolated and lends support to the hypothesis that the island groups (or perhaps the four primary, submarine plateaus) have always been more or less separated by barriers such as marine basins which have acted as distributional filters. At the same time, the two factors, (1) a homogeneous, highly endemic fauna and (2) a fauna apparently derived from at least two invasions widely separated in time, suggest a probably greater isolation of the Archipelago as a whole at some intermediate period in the past and also possibly closer connections between at least some of the island groups within the Archipelago. DISTRIBUTIONAL CHART OF AMPHII Bougainville Isabel New Geo Group Group Grout Bougainville Puruata Fauro (Faro) Mono (Treasury) Stirling Choiseul Arnayon Vella Lavella Ronongo (Ganongga) | Solomon Islands Buka | Shortland (Alu) | | | | | Marwan (Oi b-\ | | Bufo m. marinus Hyla lutea ©) = av Hyla thesaurensis =|© On = = Os = Batrachylodes trossulus oa F Ty Batrachylodes + vertebralis = oO fe) Fi AW ay Ty Ceratobatrachus guentheri = OQ. = Q “© sO = - - OQ = = aT Cornufer guppyi Oo - O Oo - = -|O - Cornufer neckeri = i, Discodeles bufoniformis oF O ©), = - © -< = a ay Discodeles guppyi Oo - O O O Ae TT Discodeles opisthodon = - O -!0 Oo ae Palmatorappia solomonis ©. = Platymantis aculeodactylus = =) Platymantis myersi = Platymantis : papuensis weberi - ‘Onn I Platymantis solomonis O Oe = ©) Rana papua krefftii Rana papua cfs novaebritanniae = i Type locality. Specimens identified in present study. Reported in the literature but not examined in the present study. Based on records in the literature of species regarded as synonyms. Reported to be in the collections of the British Museum (Natural History). Specimens in the British Museum which are here referred to novaebritanniae are catalogued as krefftii. However, on the basis of the present study these islands are occupied by the subspecies novaebritanniae. = «Ol 4 HE SOLOMONS ISLANDS Santa Ana New Georgia Russell Guadaleanal San Cristobal Group Group Group Group | BS t a is as) o = eI E a S Ey 3 ° q g a a Zz iS ° dz! i} 3 3 ae = = cS a a 3S S Eaa=| = 3S = 3 Ps 2 = 8 a 3 a) 5 O38 S ad i) > ae Ee 3 S oO = 3 ie ins} jan] i} 3 n = = cs ° a a 3 a RB Ea | = 2 DB 5 a 3 5 5 a = S ort ° om = © o— a 3 3 a 3 =) rs AS) = 8 op a Ss Ss 2 | (eI > oO a=) Ay ea S na & a = na > a = = © sO = eee = - ae Mone se ar ae = = = O Q = O * ei sveraieae 2 * * * aaa - = - -|O - GO a0) SO = O =) a A : = = O aoe Palins aml i polar. x (fae ia F = O = = = * Preemie mre atin cto * ay eta = O — = » 2 @ ie Maer s|saicralbo.S iecas [ag | easel aly * fenton tease pean * O =~ = — oneal a OA Ra | St Ee ee + + I. A specimen from Fauro Island in the British Museum, referred by Boulenger (1887) to Cornufer dorsalis, is re- garded as an error and is not recorded here. II. Specimens of Hyla thesaurensis and Ceratobatrachus guentheri in the British Museum from Gela Island are not shown in the chart. The exact location of this island is unknown to me, but it is suspected that it may be a small island perhaps in the Florida group. III. Two specimens, reported as Cornufer corrugatus by Sternfeld (1920 [1921]) from Buka Island, may be either Platymantis papuensis weberi, P. solomonis or possibly some other species and hence are not shown on the chart. 16 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY The probability of occasional population interchanges between islands by rafting is supported by the observations of Guppy (1887b, pp. 125-144). Commenting on the rapid denudation of the higher parts of the islands, he describes the frequent torrential flooding of the streams and the extensive muddying of the sea by this debris for a distance of at least a third of a mile from shore following the frequent heavy tropical rains. He also comments on the extensiveness of the pumice drift reaching the shores of the Solomons following the voleanic eruption at Blanche Bay, New Britain, May 1878, borne there by the eastward-moving ocean currents. Further geological information concerning past land fluctuations in this general region of the Pacific basin area may in the future help to clarify our understanding of some rather puzzling distributions and may point the way to a better understanding of phylogenetic rela- tionships. Treatment of Data in the Systematic Section The descriptions of the species which are discussed in the systematic part are sufficiently detailed, it is hoped, to enable workers to avoid some of the errors in identification to which overly brief descriptions have frequently lent themselves in the past. Whenever sufficiently large series of mature or near-mature specimens were available from any one island, such series were the only ones used in determining the means and their standard errors of measurements or body proportions given in the descriptions. This was done to facilitate comparisons by those working with series from other islands where possible island races may exist. Measurements were made as follows: length of head, along the side from the tip of the snout to the posterior edge of the tympanum unless otherwise stated; breadth of head, at the angle of the jaws; diameter of the eye, along the anteroposterior axis; length of tibia, in the flesh. Body proportions used were: (1) head width/length from snout to vent, head width/length of tibia, length of tibia/length from snout to vent, diameter of eye/head width, diameter of tympanum/head width, and diameter of tympanum / diameter of eye. The mean and its standard error are given in each case. When samples of populations were compared to determine the possible significance of differences in the means of measurements or body proportions, Student’s ‘‘t”’ was calculated taking into considera- tion the small size of the samples. For the sake of brevity the following abbreviations have been used throughout the text with reference to collections of various institutions: BROWN: AMPHIBIANS OF THE SOLOMON ISLANDS WZ American Museum of Natural History (A.M.N.H.), California Academy of Sciences (C.A.S.), Museum of Comparative Zodlogy, Harvard (M.C.Z.), Museum of Vertebrate Zoology, University of California (M.V.Z.), San Diego Society of Natural History (S.D.S5. N.H.), Natural History Museum, Stanford University (S.N.H.M.), United States National Museum (U.S.N.M.). SYSTEMATIC DISCUSSION Key to the families of amphibians in the Solomon Islands 1. The halves of the pectoral girdle fused ventrally........ Ranidae The halves of the pectoral girdle overlapping ventrally......... 2 2. Intercalary cartilages present; terminal phalanges claw-shaped. . . 1 i Gilet Se a at ce AAI Rea Rea A MPR a iN Stayed Hylidae Intercalary cartilages absent; terminal phalanges straight or nearly SOMUMOtClaw=shaped))ehanw i... ee. dnt eens avant Bufonidae BUFONIDAE The only representative of this family known from the Solomon Islands, Bufo marinus marinus (Linné), was introduced, presumably early in 1940 (Lever, 1945, p. 1). Genus BUFO Laurenti BuFo MARINUS MARINUs! (Linné) Rana marina Linné, 1758, Systemae Naturae (ed. 10), p. 211: America. Bufo marinus Lever, 1945, p. 1. 26 larvae (M.C.Z. 26008-9) Guadalcanal Id. (L. W. Jarcho) 1944. 2 (U.S.N.M. 119624-5) Lunga area, Guadalcanal Id. (D. H. Johnson) 1944. 6 larvae (A.M.N.H. 51958) Guadalcanal Id. 12 (M.V.Z. 39648-59) Banika Id. (J. A. Gray) 1944. 2 (M.V.Z. 40731-32) Guadalcanal Id. (Lowell Adams) 1944. 1 (S.D.S.N.H. 18019) Guadalcanal Id. Description. Head broader than long; snout rounded; canthus rostralis forming a prominent crest continuous with the preorbital and supraciliary crests; postorbital crest reaching a point almost as far ventral as the center of the tympanum; supratympanic crest extending (Lt tee specimens have been compared with the description of Bufo marinus paracnemis utz). 18 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY from the postorbital portion of the supraciliary crest to the anterior border of the parotid gland; loreal region nearly vertical, tuberculate; interorbital region concave with scattered, moderate warts; parotid glands large, heavily pitted, reaching posterior to the axillary region; upper eyelids strongly rugose; roof of mouth with a transverse ridge just posterior to the choanae, strongly projecting laterally but low near the mid-point; tongue broadly oval without a median notch posteriorly. Forelimb well developed; first finger much longer than the second; subarticular tubercles moderately large, rounded, less broad than the subtending digit; inner metacarpal tubercle narrow; outer broadly oval; hindlimb relatively short, heels failing to meet when limbs are folded at right angles to the body; length of the tibia somewhat less than, or equal to, the breadth ef the head; toes about 1/3 webbed, the distal 3 phalanges of the fourth toe being free; subarticular tubercles small, round to oval, narrower than the subtending digit; inner metatarsal tubercle prominent, narrowly oval or elliptical; outer smaller, less protrudent; dorsum and lateral surfaces marked by rows of warts, those of the dorsum larger; venter granulate, granules with brown-tipped centers except in the region of the hindlimbs; limbs pro- fusely tuberculate or warty on upper surfaces. Color of the Marine Toad, as in its native habitat, the tropical New World, is highly variable. The dorsum (in preservative) shows various shades of brown or gray. The top of the head, middorsal line and lateral surfaces are generally lighter as are the larger warts and tubercles. The warts and tubercles are, however, brown-tipped. The venter is grayish or yellowish, somewhat marbled with darker gray or brown. It may be noted that tadpoles were collected on Guadalcanal both in June and November. Range. Definitely known from Guadalcanal, Malaita and Banika as well as perhaps other islands in the Russell group. This species has also been introduced into the Admiralties; U.S.N.M. 121854-5 are from Manus Island and 121660-2 are from Los Negros. HYLIDAE Genus HYLA Laurenti Parker (1939, p. 2) regarded H. lutea as definitely distinct from H. thesaurensis whereas van Kampen (1923, p. 50) was somewhat doubtful following Barbour’s relegating it to the synonymy of the latter (Barbour 1921, p. 93). Parker noted among other characters BROWN: AMPHIBIANS OF THE SOLOMON ISLANDS 19 the more extensive webbing of the fingers, the larger size and the longer, more pointed snout of H. lutea. That the snout is longer is not borne out by the present series, its ratio to head width or total length being approximately the same for both species. However, the distinct- ness of H. lutea from H. thesaurensis in other respects is amply evi- denced. The following key serves to separate these two species: Head moderately depressed; disk of third finger generally smaller than tympa- num; disks of fingers (except the inner one) much broader than the sub- tending digit which is not bordered by a wide flange of skin. . . . thesaurensis Head much depressed; disk of third finger generally larger than tympanum; disks of fingers not or scarcely broader than the subtending digit which is pond eredubyadawidentlange otis kines sel) eise ee) ets pes eevee tense etrel test lutea Hywa LuTea Boulenger Riga ties) Hyla lutea Boulenger, 1887, Proc. Zool. Soc. London, 1887, p. 337, pl. xxviii, fig. 4; Fauro Id., Solomon Ids. (Type in British Museum). Hyla thesaurensis (part), Burt and Burt, 1932, p. 488. Cornufer guppyt (part), Burt and Burt, 1932, p. 489. 4 (A.M.N.H. 34273, 35342, 35344-5) Bougainville Id. (Whitney Exped.). 3 (A.M.N.H. 34636-7, 35387) Choiseul Id. (Whitney Exped.). 1 (S.N.H.M. 9346) Bougainville Id. (Exch. Amer. Mus.). Description. Head slightly less broad to as broad as long, its breadth about 1/3 the length from snout to vent; snout round-pointed; eye moderate, its diameter about 1/3 the breadth of the head; tympanum round, its diameter 1/2 or slightly more than that of the eye; loreal region strongly oblique, concave; canthus rostralis rounded, rather indistinct; interorbital space broader than upper eyelid; vomerine teeth in two transverse patches between the choanae and almost in contact medially; tongue broadly oval and but feebly indented at the mid-point of the posterior margin. Forelimb well developed; finger tips strongly depressed with large disks; disks broader than long (except for the inner finger), but scarcely broader than the subtending digits as measured to include the flanges of skin on the lateral margins; fingers more extensively webbed than in H. thesaurensis, the third finger being webbed to the subdistal tubercle on the outside and to a point between this and the basal tubercle on the inside, the second almost to the distal tubercle on the 20 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY outside; subarticular tubercles small, low, transversely elongate; metacarpal tubercles indistinct; hindlimb long, length of tibia about 3/5 the length from snout to vent; disks of toes smaller than those of fingers; subarticular tubercles small, more strongly protrudent than those of hands; inner metatarsal tubercle narrow elliptical, its length less than its distance from the basal tubercle of the inner toe; outer absent; toes webbed almost to the disks except for the fourth; skin smooth except for the flat granules of the lower surfaces of the thighs and abdomen. Color (in preservative) of the dorsum is quite uniformly grayish or yellowish tan for these eight specimens; somewhat lighter on the venter. Measurements o&'(S.N.H.M. 9346) 9 (A.M.N.H. 34273) SNOUtHCONVeMi tear ater sane 54 mm. 65.5 mm. ~ Snoutelem cttw rer ter Ors 11 § Head length (to base of skull) 1g % 19 ef Eveadgoneddit hee eee Wer ss 22 e Hiyerdiameterse qrice ace ae @ GH. * Tympanum diameter....... oo Sedo) ie Aibiaulengthpocpareeee eee 50m 39 a Range. Known from Bougainville, Mono and Choiseul Islands. HyLa THESAURENSIS Peters (Riss vii stale) Hyla thesaurensis Peters, 1877, Monatsb. Akad. Wiss. Berlin, p. 421: Treasury Id., Solomon Ids. (Type in Berlin). Hyla macrops Boulenger, 1883, Ann. Mag. Nat. Hist. (5), xii, p. 164: Treasury Id., Solomon Ids. (Type in British Museum). Hyla solomonis Vogt, 1912, Sitzb. Ges. Naturf. Freunde Berlin, p. 10: Bou- gainville Id., Solomon Ids. (Type probably in Berlin). Hyla thesaurensis (part), Burt and Burt, 1932, p. 488. 1 (M. c Z. 7373) Fulakora, Isabel Id. (W. M. Mann) 1916. DB (( 7374-75) Auki, Malaita Id. i sf 31C 8 3(6=8) hulacield: a 10( ‘“ 7379-88) Atta, Malaita Id. fs “ 3( “ 7390, 7892-93) Isabel Id. se is 4( ‘* 7401-4) Yandina, Pavuvu Id. e ch 1( ‘“ 7405) Rubiana Lagoon area, 66 66 New Georgia Id. 1( ‘ 9374) German Solomon Ids. (Exch. Berlin Mus.) 1922. 1( ‘ 26051) Guadalcanal Id. (L. W. Jarcho) 1943. Dies 26052-3) uf if i 1948. BROWN: AMPHIBIANS OF THE SOLOMON ISLANDS Pall 1 (S.N.H.M. 8390) Tetere area, Guadalcanal Id. (J. P. Heath) 1944. 2 (C.A.S. 49952-3) Guadalcanal Ite (J. A. Kusche) 1920. 6( ‘ 49956-61) oe uF 1921. 1( “ 654665) Malaita Id. (W. M. Mann) 1916. Gs Gece 2100) i ‘* (Crocker Exped.) 1933. 24 (U.S.N.M. 119560-5, 119567-75, 119756-64) Torokina Pt., Bou- gainville Id. (W. L. Necker and D. H. Johnson). 38 juvs. and larvae (U.S.N.M. 119765) Torokina Pt., Bougainville Id. (W. L. Necker and D. H. Johnson). 1 (U.S.N.M. 119566) Puruata Id. (W.L. Necker and D.H. Jolnacem). 1 ( o 119721) Mono Id. ie 2 (M.V.Z. 40733-4) Malumba River area, Guadalcanal Id. (Lowell Adams) 1944. 3 (M.V.Z. 44189-90, 44223) lower Lunga River area, Guadalcanal Id. (J. Chattin) 1944. 2 (S.D.S.N.H. 18057-8) Russell Ids. 4 (A.M.N.H. 34320, 35327, 35339, 35346) Bougainville Id. (Whitney Exped.) 1930. 1 (A.M.N.H. 35423) Mono Id. (Whitney Exped.) 1930. 1 ( és 39998) Central Malaita Id. (Whitney Exped.) 1930. 72 oo, 2 Y, Juvs., and larvae (A.M.N.H. 51959-75) Guadalcanal Id. 1944. 8a, 2 9 (A.M.N.H. 52178, 52176-79) Guadalcanal Id. 1945. Hyla thesaurensis is highly variable as to color and may prove to be constituted of at least two geographical races or be undergoing such differentiation. Only from the southwestern islands have I observed occasional specimens heavily mottled with dark-brown on the throat and chin. However, the majority of the specimens from these islands possesses the uniformly light venter and lower lip. Therefore I have chosen to regard H. thesaurensis as a highly variable species until larger series are available from many more of the islands. Description. Head slightly less broad to as broad as long, its breadth about 1/3 the length from snout to vent, depressed but less so than for H. lutea; snout rounded, its length less than 1/2 the breadth of the head (45.99% + .692 for 24 specimens from the Bougainville group); eye moderate, its diameter 1/3 to about 2/5 the breadth of the head (38.18% + .647 for 24 specimens); tympanum large, its diameter generally about 1/2 that of the eye (49.83% + 1.44 for 24 specimens); interorbital space broader than the upper eyelid; loreal region slightly oblique, shallowly concave; canthus rostralis distinct, slightly rounded; vomerine teeth in two short, narrowly separated, transverse or slightly oblique patches between the choanae or their posterior borders; tongue broad with a shallow notch at the mid-point of the posterior margin. 22 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Forelimb well developed; fingers with large, rounded or transversely elliptical disks at the tips; disks (except for inner finger) much broader than the subtending digit which is bordered at most by only a narrow flange of skin on the lateral margins; disk of third finger much smaller than, or equal to, the tympanum; fingers with slight or moderate webs, reaching the basal tubercle on the inside and about halfway between the basal and subdistal tubercles on the outside of the third finger, and the subdistal tubercle on the outside of the second; subarticular tubercles moderately protrudent (except the distal ones), transversely elliptical; inner metacarpal tubercle narrow elliptical; middle and outer shorter; hindlimb long, tibia length about 1/2 to 3/5 the length from snout to vent (58.42% + .514 for 24 specimens); disks of toes somewhat smaller than those of fingers; subarticular tubercles moder- ate, round or transversely elongate, smaller than those of hands; inner metatarsal tubercle narrow elliptical, short, its length less than its distance from the subarticular tubercle of the inner toe; outer small, round; toes webbed almost to the disks except for the fourth; skin smooth on the dorsum or occasionally with a few, scattered granula- tions; venter with moderate granules posterior to the pectoral region and sometimes on the chin as well; lower proximal region of the thighs granular. Color (in preservative) of the dorsum highly variable, being light- grayish to olive-brown, often with darker mottling; the three white lines along the back, as originally described by Peters (1877, p. 421) and figured by Boulenger (1886, fig. 4), present or absent, or broken, more commonly present in juveniles but also in some adults; venter white, yellowish or tan, throat and chin sometimes mottled with brown as observed for specimens from Guadalcanal Island. Mature ovarian eggs are small and pigmented at the animal pole region. Measurements o'(U.S.N.M. 119569) 9(U.S.N.M. 119721). Snout tOnventr eeu eeie ore 40.5 mm. 49 mm. Snoutilencthie seg Gees if y Head length (to posterior edge of tympanum) 222 5242 13:5 “8 17 rf Headtbreadth ss i245 aac NAYS) 9 17 sf Hye diameter 224 haesa- ce BeEspctsy 6.25 “ Tympanum diameter....... Pei) =) 1 3 oe Aibiavlengthersaese sree 23 oh Ppeiige i Range. (see distributional chart) Since a description of the larval stage of this species has apparently not been recorded in the literature the following notes are given, based on a series of larvae and transforming individuals (U.S.N.M. 119765). BROWN: AMPHIBIANS OF THE SOLOMON ISLANDS 23 One of the younger larvae with the forelimbs rather well developed but still beneath the operculum, measures 58 mm. in total length; vent to tip of tail 38 mm.; width of body 12 mm. (preserved condition) ; nostril to tip of snout 3 mm.; eye to nostril 3 mm.; interorbital space 6 mm.; width of mouth 4 mm.; greatest breadth of tail 9 mm. The mouth is antero-ventral in position. The lips are papillate except for the median third of the upper lip; the upper lip possesses two rows of horny teeth, the inner divided at the midline. The lower lip has three rows, the inner one also medianly divided. The edges of the jaws have a dark horny covering. The eyes are dorsolateral in position. The spiracle forms a narrow slit-like opening (1 mm. in length) slightly ventral to a line connecting the corner of the mouth and the groin on the left side and about equidistant from these two points of reference. The dorsal and ventral fins of the tail are of approximately equal width with the tip rounded, and heavily pigmented (dark grayish- brown in preservative). RANIDAE The following key will serve to distinguish the genera of the Ranidae known to occur in the Solomon Islands: 1. Omosternal style entire or with a very small notch at the base......... 2 Omosternal style with a broad fork at the base....................... 3 2. Vomerine teeth present; toes with webs................ Rana (Hylarana) Vomerine teeth absent; toes without webs................. Batrachylodes 3. Vomerine teeth absent; fingers with webs................ Palmatorappia Vomerine teeth present; fingers without webs........................ 4 4. Terminal phalanges broadly T-shaped; digital disks large; toes moderately RHE] 0) BEC Kenn ah Cede ey meee Eyam bt 6 eA a ete ar re Me eee NAPIER Cornufer Merminaljphalanges bluntly,rowmded eye yan a1. yee ee ee sea 5. Digital disks small to large; toes moderately to almost fully webbed Discodeles Digital disksismall; toes) without) webs . 3/40) 44) SONA e aa ya ie 6 6. Odontoids absent from lower jaw.................2.ece00- Platymantis Odontoids present on lower jaw.............1....00005 Ceratobatrachus Genus BATRACHYLODES Boulenger This genus was originally set up by Boulenger (1887, p. 337) on the basis of a single female specimen of B. vertebralis from Fauro Island. Neither at that time nor in his later writings did he mention the possible relationship of Batrachylodes to other ranid genera. Noble (1931, p. 524), apparently without having seen a specimen, held it to 24 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY be a diminutive relative of Cornufer lacking vomerine teeth. Deckert (1938, p. 181), who also was unable to examine a specimen of the genus, follows Noble. This proposed derivation from Cornufer, however, is not borne out when one examines closely the pectoral girdle. The omosternum, though broad at the base, is unforked. On the basis of this evidence, although Batrachylodes may have been derived from the same ranid stock as Platymantis and Cornufer, the undivided omosternum makes possible its origin in the Hylarana group as Noble (1931, p. 521) proposes for Micrixalus and Staurots. Eggs are large and unpigmented as observed for Batrachylodes trossulus. ; The following key serves to distinguish the two known species: Disks of fingers and toes little dilated; breadth of disk of third finger about halfdiameter oighymipanum):-. 29 eeieeo te See eae ee eee trossulus Disks of fingers and toes broadly dilated; breadth of disk of third finger equal to or greater than diameter of tympanum.....................vertebralis BATRACHYLODES TROSSULUS Brown and Myers (RIS Ge ies Sake Sasticasl)) Rana solomonis (part), Burt and Burt, 1932, p. 491. Batrachylodes trossulus Brown and Myers, 1949b, Jour. Wash. Acad. Sci., vol. 39, no. 11, pp. 379-80; Torokina, Bougainville Id. (Type in United States National Museum). 7 (U.S.N.M. 119577 holotype, 119586-88, 119787-89) Torokina Pt., Bougainville Id. (W. L. Necker). 1 (A.M.N.H. 35425) Choiseul Id. (Whitney Exped.). As no additional specimens of this small frog have been found in collections examined since its description was published in 1949, nothing more is known of its variability, and only a brief restatement of that description is given here. Description. Head relatively narrow, its breadth about 1/3 the length of the body, 2/3 to 3/5 the length of the tibia; snout rather pointed, strongly projecting beyond the tip of the lower jaw, its length about 1/3 to 1/2 the breadth of the head; eye large, its diameter about equal to or slightly greater than the length of the snout; diameter of tympanum slightly more or less than 1/2 the diameter of the eye; loreal region nearly vertical; canthus rostralis somewhat rounded; vomerine teeth absent; tongue oval without a distinct notch posteriorly. BROWN: AMPHIBIANS OF THE SOLOMON ISLANDS 25 Forelimb moderately developed; fingers slender with small round . disks at their tips, much smaller than in B. vertebralis; the inferior portion of the disk separated from the superior by a nearly terminal crescentic groove; diameter of the disk of the third finger about 1/2 that of the tympanum; terminal phalanx a narrow “T”’; subarticular tubercles poorly or moderately developed; inner metacarpal tubercle broadly oval, moderate; middle one similar in size and development; outer one smaller; hindlimb rather well developed; length of the tibia about half the length of the body; heel reaching the eye; tips of toes expanded into large disks, larger than those of the fingers and slightly broader than long; subarticular tubercles moderately developed; inner metatarsal tubercle oval, elongate, its length about equal to its distance from the end of the outer toe; outer metatarsal tubercle moderate, round; toes without webs; skin smooth. Color (in preservative) very uniform for the entire series from Bougainville Island; dorsum dusky gray suffused with pale reddish, most prominent on the hindlimbs; sides of head and body dark reddish-brown to black, bordered above by a light edge and continuous with the dark blotch on anteroventral surface of the forelimb; axillary surface of forelimb with an uneven-edged, dark slate or blackish band; as is also the anterior border of the thigh; anterior surface of lower leg with a more or less broken series of dark blotches or dashes; anal region and proximal portion of posterior surface of thighs blackish; under, surface of head, throat and pectoral region dark reddish-brown with scattered light flecks; white nuptial tubercles onthe chin andsome- times throat of males, belly and under surface of thighs whitish flecked with dark reddish-brown. Ovarian eggs are unpigmented, large. Measurements o'(U.S.N.M. 119577) 9 (U.S.N.M. 119586) SIX WO) WING 6956 cuedue 20.25 mm. 19 mm. Snoutdlength!: J2na.. 1.50 3 as DIO Head length (to posterior edge of tympanum)..... a i 6.25 lead ibreadthyn a seieen GH 8 Op & Byexdiamieter aja. sicily. .05 Length of tibia Length of tibia —————_ 52.34 + .53 50.67 =.89 1.67 2.506 37 <.02 Length snout to vent 52 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Both populations exhibit a pronounced sexual dimorphism when snout to vent lengths for males and females are compared; the difference between the means in either case is highly significant. The males are much smaller than the females. There is no significant difference in body proportions, except perhaps for the length of the tibia in relation to the snout to vent length, between New Guinea and Solomons populations. Description. Head slightly less broad than long; its breadth generally less than 2/5 the length of the body (88.48% + .20 for 19 specimens from Isabel Island), about 3/4 the length of the tibia (76.14% + .70 for 19 specimens); snout round-pointed, relatively more pointed than for P. solomonis; nostril nearer tip of snout than eye; eye moderately large, its diameter about 1/3 the breadth of the head (33.71% + .364 for 19 specimens); interorbital space less than the breadth of the upper eyelid; tympanum large, its diameter generally about 1/5 the breadth of the head (22.56% + .625 for 19 specimens); vomerine teeth in two short, rather widely separated, oblique patches generally posterior to the choanae; tongue moderate, oval with a rather wide, deep notch at the mid-point of the free, posterior margin. Forelimb well developed; first finger longer than the second; finger tips slightly dilated, usually somewhat depressed with a shallow but generally distinct, terminal, crescentic groove separating the inferior and superior parts of the disk at the distal and lateral margins; sub- articular tubercles other than the basal ones large, round, about as broad as the subtending digit, strongly protrudent and tending to be pointed distally; basal ones smaller, round, weakly protrudent; inner metacarpal tubercle broadly elliptical, its length about equal to its distance from the distal tubercle of the first finger; middle one shorter, broadly oval; outer much shorter, oval or elliptical; hindlimb moderate, length of tibia about 1/2 the length of the body (50.67% + .39 for 19 specimens); tips of toes moderately dilated, depressed, the inferior pad of the disk separated from the superior part by a deep, terminal, crescentic groove; subarticular tubercles round to broadly oval, strongly protrudent and pointed distally; solar area prominently tuberculate; inner metatarsal tubercle moderately broad, elliptical, its length about equal to its distance from the distal end of the tubercle of the first toe; outer moderate, round, strongly protrudent; a low tarsal fold extending proximally from the base of the inner metatarsal tubercle; toes without webs; dorsum with short to moderate, narrow folds generally displaying a pattern of about eight longitudinal rows anteriorly and diminishing posteriorly; rugose with small tubercles between the folds of the dorsum, the upper eyelids, loreal areas, upper lateral surfaces and the upper proximal surfaces of the hindlimbs; BROWN: AMPHIBIANS OF THE SOLOMON ISLANDS 53 lower abdomen and posterior thighs with moderate flat granules; under surface of head of males generally finely granulate. Color (in preservative) of the dorsum light-brown to dark reddish- brown, often with darker blotches and dark interorbital bar, frequently with a wide purplish-gray dorsolateral band on each side, occasionally with bright red markings in association with some of the folds and tubercles on dorsum and limbs; margins of lips with more or less distinct, irregular transverse, dark bars; loreal region sometimes very light; upper surface of thighs and tibial region with wide, dark trans- verse bars or blotches; venter uniformly light or more or less powdered or marbled with brown anterior to the forelimbs. Ovarian eggs are large and unpigmented. Measurements o'(M.C.Z. 7562) 2 (M.C.Z. 7560) SMOURLORVIEMG ete see sn eine ater: 40 mm. 56 mm. SHOuUHemacheae ee eee ee Tai OMe Head length (to posterior edge of (HPO OMNI) 3S 5 Bla oa go 0 cb ood oon IG: 22a JSlGAVG | crReHYO LO Geena 5 & Shee eve aiciara oe 1G Pilea Hiverciametensis 4 oak csi users Digi Sains ALympanum diameter. 4s. e222. -\- Ad ANN diloiaslene-thy ame ec. alee rte Weg ant ON Be Range. (See distributional chart). PLATYMANTIS SOLOMONIS (Boulenger) (2G tl, thers Zig 2A, Mp ste, hp BL Ai itiexs 11) Cornufer solomonis Boulenger, 1884, Proc. Zool. Soc. London, p. 212: Fauro, Treasury and Shortland Ids. (Type in British Museum). Platymantis solomonis, Boulenger, 1918b, p. 372. Platymantis solomonis (part), Barbour, 1921, p. 96. Rana solomonis (part), van Kampen, 1923, pp. 191-192. Rana solomonis (part), Burt and Burt, 1932, p. 491. 1 (M. C: Z. 3499) Stirling Id. (Exch. Brit. Mus.). a ( 7444, 7554, 7561) Isabel Id. (W. M. Mann) 1916. 2G ie 497, 7500) Rulaen Td: fy 1916. Cae 7581) Malaita Id. ne 1916. 2( “ 7585-86) New Georgia Id. <$ 1916. 6( “ 26085-89 —1 uncat.) Stirling Id. (L. W. Jarcho) 1945. 1( ‘ 72080) New Georgia Id. (Crocker Exped.) 1933. 2 (M.V.Z. 44948-49) Munda, New Georgia Id. (C. G. Sibley) 1944. 1 (S.N.H.M. 8393) Torokina Pt., Bougainville Id. (J. P. Heath) 1943. 54 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY 12 (U.S.N.M. 119590-93, 119776-81) Torokina Pt., Bougainville Id. (W. L. Necker and D. H. Johnson). 1 (U.S.N.M. 119594) Puruata Id. 18 (A.M.N.H. 34257, 35259-60, 35306, 35308-11, 35314-15, 35317, 35320-26) Bougainville Id. (Whitney Exped.). 1 (A.M.N.H. 22856) Vangunu Id. (Whitney Exped.). iL ¢ HY 35386) Arnavon Id. “ oh 1 (( ty 35389) Choiseul Id. a 1G se 35391) Mono Id. fe re AX te 39997) Malaita Id. oN 3 Boulenger (1884, p. 212) based his description on specimens from the small islands near the southern end of Bougainville, noting such characteristics as the large head and eyes, tips of fingers swollen rather than dilated, interorbital distance less than breadth of upper eyelid, and dorsal surfaces slightly rugose with short longitudinal folds. His figure (1886, pl. xi, fig. 2) well illustrates this large eyed, moderately or but slightly rugose species which is widely distributed in the Solomons. This very large species may have its closest affinities with P. vitianus. Subspecific populations are very strongly suggested when series from different island groups are carefully examined. However, since the only samples available containing five specimens or more, of either sex are those from Bougainville, Stirling and New Georgia Islands, no well defined geographical limits can be determined for the suggested sub- species. Consequently no predictions are made at this time. Description. Head about as broad as long; its breadth almost 2/5 the length from snout to vent (37.77% + .434 for 11 specimens from Bougainville Island), 3/4 to 4/5 the length of the tibia (77.19% + .908 for 11 specimens); snout round or round-pointed, only slightly pro- trudent beyond the lower jaw; nostril nearer tip of snout than eye; eye large, its diameter more than 1/3 the breadth of the head (37.37% + 555 for 11 specimens), tympanum round, large, its diameter about 1/5 to 1/4 the breadth of the head; vomerine teeth in two strongly protrudent, oblique or transverse patches with the crests posterior to the choanae and the outer angle of the bases extending outward beyond the sagittal plane of the inner border of the choana or either side, the distance separating the patches slightly more or less than the length of either; tongue rather broadly oval with a narrow to broad cleft at the mid-point of the posterior free margin. Forelimb well developed, first finger longer than the second; tips of fingers swollen and not or scarcely dilated, without a groove separating inferior and superior portions; subarticular tubercles large, distal ones about as broad as the subtending digit; round or somewhat oval, BROWN: AMPHIBIANS OF THE SOLOMON ISLANDS 55 strongly protrudent; inner metacarpal tubercle broadly elliptical, its length about equal to its distance from the subarticular tubercle of the first finger; middle one broadly oval, shorter; outer narrower; hindlimb moderately long; length of tibia about 1/2 the length from snout to vent (49.79% + .513 for 11 specimens); tips of toes dilated into small disks, the inferior portion separated from the superior by a crescentic, terminal groove; subarticular tubercles oval, smaller than on hands, strongly protrudent and pointed distally; solar area strongly tubercu- late; inner metatarsal tubercle narrow elliptical, its length equal to its distance from the distal end of the tubercle of the first toe; outer round, strongly protrudent; toes without webs; dorsum only moder- ately rugose with scattered, short, relatively broad folds or tubercles, very nearly smooth for larger, older females; lateral surfaces granulate or with some tubercles; venter granulate posteriorly and on the inner surface of the thighs. Color (in preservative) of the dorsum grayish-brown to reddish- brown or blackish-brown often with darker blotches; limbs often lighter; thighs and tibial regions with more or less distinct, wide, transverse bars or blotches; margin of lips with more or less distinct, broad, dark transverse bars; lower lateral surfaces lighter, powdered or marbled with darker shades; venter white, grayish or pale-brown posteriorly, generally powdered or marbled with brown or reddish-gray anterior to the forelimbs. Ovarian eggs are large and unpigmented. Measurements o'(U.S.N.M. 1193892) ¢(U.S.N.M. 119591 ) SMOUt CO NVeEnt wets. eee hc) 49 mm. 66 mm. SHOCOTOIG Jey a¥ea n OW a uele etn ele nee Py eam! 9 us 11 ie Head length (to posterior edge Ofstympanunn) - 92 sD 20 : 26 is Head breadth....... ELAN G8 Seay PAB BSNS Hyerdiameter ian. hisias sks U te 9 ‘t Tympanum diameter........ A Reg regi Ay uses Misiiniaglen oth. i Septal cen ue PAD a) oe a Range. (See distributional chart). Genus RANA Linné The genus Rana is represented in the Solomon Islands by two known populations of the subgenus Hylarana. Rana papua krefftii occupies only the extreme southwestern islands of the San Cristobal group. The subspecies represented in the other groups, so far as known, is here referred to Rana papua novaebritanniae, although subspecific 56 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY identity of the Bismarcks and northern Solomons populations may be shown to be in error when a large number of specimens from the former locality become available for comparison. Some of the specimens in the limited collections from Guadalcanal and Malaita which I have been able to examine are somewhat inter- mediate between these two populations. However, since they agree in most characteristics with R. papua novaebritanniae, I have assigned them here until such time as more material is available from these island groups as well as the interlying Choiseul and Isabel groups. These two subspecies may be distinguished as follows: Tympanum large, its diameter more than 1/4, generally more nearly 1/3 the breadth of the head; disks of toes small, oval; snout generally rounded, little protrudent; venter whitish or light, not or little mottled with blotches OUOTO WA Arey eras aie a eee beeen set Ie 9 er ees papua novaebritanniae ‘Tympanum moderate, its diameter generally less than 1/4 the breadth of the head; disks of toes moderate, somewhat pointed; snout moderately pointed, generally protrudent; venter usually mottled with large brown blotches. . . papua kreffti RANA PAPUA KREFFTII Boulenger GL Gans, 1) Hylarana erythraea (part), Giinther, 1858, p. 73. Rana kreffidzi Boulenger, 1882, Cat. Batr. Sal. Brit. Mus., p. 64, pl. ii, fig. 2: San Cristobal Id. (Type in British Museum). Rana (Hylarana) krefftii (part), Boulenger, 1918a, pp. 239-41. Rana krefftii (part), Barbour, 1921, pp. 97-98. Rana kreffti (part), van Kampen, 1923, pp. 206-07. Rana krefftii (part), Kinghorn, 1928, pp. 125-26. 1 (M.C.Z. 2517) Solomon Ids. (W. M. Mann) 1916. 46( ‘ 7482-34, 7505-47) Bio Id. es ie 1( ‘ 7440) San Cristobal Id. - a 1( “ 7447) Santa Ana Id. of f 4 (U.S.N.M. 63402-05) Bio Id. ss a 1 (C.A.S. 54666) San Cristobal Id. * ss Boulenger (1882, p. 64) described R. papua krefftii on the basis of two specimens, one from San Cristobal Island and one with the more general locality designation of Solomon Islands. In view of the fact that well differentiated subspecies occupy different geographical areas within the Solomon Islands, I designate San Cristobal as the type locality of R. papua krefftit. Description. Head less broad than long, its breadth about 1/3 the BROWN: AMPHIBIANS OF THE SOLOMON ISLANDS 57/ length from snout to vent (84.75% + .34 for 5 specimens from Bio and San Cristobal), about 2/3 the length of the tibia (66.48% + 1.44 for 5 specimens); snout pointed, generally strongly protrudent beyond the lower jaw; nostril nearer tip of snout than eye; eye moderate, its diameter 1/3 to about 2/5 the breadth of the head (36.8% + 1.136 for 5 specimens) ; tympanum moderate, its diameter generally less than 1/4 the breadth of the head (22.02% + .277 for 5 specimens); inter- orbital space less than, or equal to, the breadth of the upper eyelid; loreal region nearly vertical, concave; canthus rostralis angular; vomerine teeth in two oblique patches between the choanae, the dis- tance separating them generally less than the length of either; tongue oval with a moderate to broad, rounded notch at the mid-point of the posterior, free margin. Forelimb well developed; fingers long, slender, the first longer than the second; tips dilated into somewhat pointed disks nearly as large as those of the toes, a rather angular groove (frequently incomplete at the vertex) separating the distally narrower, inferior pad from the superior portion; subarticular tubercles round or broadly oval, gener- ally more protrudent distally, nearly as broad as the subtending digit (except the basal ones which are low and somewhat narrow elliptical) ; inner metatarsal tubercle broadly elliptical, its length less than its distance from the distal end of the distal tubercle of the inner finger; middle one shorter, broadly oval, outer narrow elliptical; fingers without webs; males with a prominent humeral gland; hindlimb long; length of tibia about 1 2/5 to slightly more than 1 1/2 times the breadth of the head; toes long; tips dilated into moderate, somewhat pointed disks, the inferior pad separated from the superior portion by a deep, distally complete, somewhat angular groove; subarticular tubercles small, narrow elliptical, its length less than its distance from the subarticular tubercle of the inner toe; outer one small, round; toes webbed to the distal tubercle or almost the disk except the fourth which is webbed only to the penultimate or occasionally the distal tubercle on the inside; outer metatarsals separated to base; a distinct outer metatarsal fold present; skin of the dorsum generally rather smooth; upper surfaces of the limbs with very fine longitudinal folds; lateral surfaces nearly smooth or moderately granulate; venter nearly smooth or finely granulate posteriorly; inferior and inner surfaces of the thighs granulate at least proximally. Color (in preservative) of dorsum reddish-brown, occasionally olive- brown, but generally lighter than for R. papua novaebritanniae; lateral surfaces of head and upper lateral surfaces of body dark reddish-brown to almost blackish-brown, bordered ventrally by a narrow light band beginning anteriorly at the angle of the jaws; lower lateral surfaces 58 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY mottled with brown; margin of lips generally suffused with brown; upper surface of limbs lighter to same color as the dorsum, thigh and lower leg barred or irregularly blotched with dark-brown; venter, es- pecially on the head and throat, heavily blotched with brown, or occasionally only retaining the dark-brown on the ventral margin of the lower jaw and the pectoral blotches. Ovarian eggs small and heavily pigmented at the animal pole region. Measurements &'(M.C.Z. 7533) 9 (C.A.S. 54666) SHNOUWNH WO WE G4 oo ce oacsooooboesc 52 Tam 82. mm. Snroudplen'c thee eerie eer 8 e iS alias Head length (to posterior edge of EYAL ATUUNTMN) Mees eel le sols aera 20 ie 32 of Eeadtorestdithw ee as nemeei iit 17 ss 29 Ee Biv erdiamehereii sive ac ina tntenait one ee OP atten IO 7/5) Tympanum diameter). ....5-....... 4.5“ GLOW Te Aiiaslen ablater yy seine sey). ela tees ae 2! Cis 41 c Range: (See distributional chart). RANA PAPUA NOVAEBRITANNIAE Werner (Pl. 6, fig. 4) Rana novaebritanniae Werner, 1894, Zool. Anz., 17, p. 155: New Britain. (Type in Berlin). Rana (Hylarana) krefftii (part), Boulenees 1920, pp. 186-88. Rana papua, Sternfeld, 1920 (1921), p. 483. Rana krefftii (part), Barbour, 1921, pp. 97-98. Rana kreffti (part), van Kampen, 1923, pp. 206-07. Rana krefftii (part), Kinghorn, 1928, pp. 125-26. Rana krefftii, Schmidt, 1932, p. 180. Rana kreffti (part), Hediger, 1934, pp. 451, 486. Rana krefftzi, Slevin, 1934, p. 184. 2 (M.C.Z. 7442-43) Isabel Id. (W. M. Mann) 1916. 2( ‘“ 7445-46) Tulagi Id. e as (2)1 (‘°° (97448) Santa Cruz Td:! os HY 15 (U.S.N.M. 119611-23, 119774-75) Torokina, Bougainville Id. (W. L. Necker). 2 (M.V.Z. 44192, 44222) Guadalcanal Id.- (J. Chattin) 1944. 5 (C.A.S. 72076-79, 72166) Malaita Id. (Crocker Exped.) 1933. Description. Head less broad than long, its breadth about 1/3 the length from snout to vent (33.67% + .318 for 15 specimens from Bougainville Island) and about 2/3 the length of the tibia (67.65% + .984 for 15 specimens); snout round-pointed, little protrudent; nostril 1 Exact locality for this specimen is in doubt; it is almost certainly some other island. BROWN: AMPHIBIANS OF THE SOLOMON ISLANDS 59 nearer tip of snout than eye; eye large, its diameter about 2/5 to nearly 1/2 the breadth of the head (40.74% = .759 for 15 specimens) ; tympanum large, round, its diameter 1/4 to 1/3 the breadth of the head (31.93% + .617 for 15 specimens); interorbital distance less than the breadth of the upper eyelid; loreal region nearly vertical, somewhat concave; canthus rostralis rather angular; vomerine teeth in two short, oblique patches between the choanae, the distance separating them less than the length of either; tongue generally narrow-oval with a deep, narrow, rounded notch at the mid-point of the free posterior margin. Forelimb well developed; fingers long, the third finger longer than the snout, first longer than the second; tips of fingers dilated into moderate, oval or slightly pointed disks about the same size as those of toes, with a prominent terminal, semicircular groove separating the inferior pad from the superior portion; subarticular tubercles large, oval; inner metacarpal tubercle elliptical, about twice as broad as long, its length equal to its distance from the distal end of the tubercle of the inner finger; middle one shorter, almost as broad as long; outer short and narrow; fingers without webs; male generally with humeral gland present; hindlimb long; heels moderately overlapping when the hindlimbs are placed at right angles to the body; heel of the appressed limb reaching the loreal region; length of tibia about 1/2 the length from snout to vent (49.97% = .577 for 15 specimens); tips of toes dilated into moderate, oval disks, the inferior pad less broad than the superior portion and separated from it by a semicircular groove as in the fingers; subarticular tubercles small, oval, more strongly pro- tuberant distally; inner metatarsal tubercle elliptical-elongate, 2 to 21% times as long as broad, shorter than its distance from the distal end of the subarticular tubercle of the inner toe, outer moderate, round; toes webbed to a point between the distal tubercle and the disk except for the fourth and the third on the-inside where the web reaches the penultimate tubercle or beyond; skin of the dorsum almost smooth or finely granulate posteriorly; venter smooth; proximal region of the posterior surface of the thighs and anal area granulate. Color (in preservative) of the dorsum dark-reddish or grayish-brown, raised tubercles, where present, darker; upper lateral surfaces of head and body blackish-brown, bordered dorsally by the moderately narrow, light dorsolateral fold; white or dusky gray coloration of upper lip continuous with the narrow, light stripe along the upper lateral surface from the angle of the jaw to the groin; lower lip white or dusky; venter white or occasionally powdered with brown or gray anteriorly; upper surface of limbs grayish or brown; lighter than body; thigh and lower leg with broad, dark transverse bars; posterior thighs mottled with 60 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY blackish-brown; a large brown blotch on the outer inferior surface of the upper arm; lower thighs whitish or somewhat mottled with grayish-brown; under surface of hands and feet grayish-brown. Ovarian eggs are small and darkly pigmented in the animal pole te) region. Measurements ot (U.S.N.M. 119611) @ (U.S.N.M. 119616) SHAME WO AMG Aigo eo ob bow ooo euR 50 mm. 51 mm. SHAH MADER, 6 Goo Gop ooo soho ode 8 a Si25u Head length (to posterior edge of UNaNar VALENS sg aoa we guy oa OG 20 ‘ 20 us laleayol lorrctachdls 4g Gs gio ace a obloes IRS 5), SO 17 Bye digmeter 470 sain Gdns ctl 7 *f 7 ys Tympanum diameter........... 5 ‘it 5 Wy bilo iewlemo:blaiewe eer jarccuca note 25 ss 26 by Range. (See distributional chart). BROWN: AMPHIBIANS OF THE SOLOMON ISLANDS 61 BIBLIOGRAPHY Aut, Ernst 1927. Ueber neue oder seltene Froschlurche aus dem Zoologischen Museum Berlin. Sitz.-Ber. Ges. Naturf. Freunde, Berlin, Jahrg. 1926, pp. 111-117. BarBour, THOMAS 1912. 1921. A Contribution to the Zoégeography of the East Indian Islands. Mem. Mus. Comp. Zodl., 44, pp. 1-205, pls. i-vili. Reptiles and Amphibians from the British Solomon Islands. Proc. New England Zool. Club, 7, pp. 91-122, pls. ii-iv. BouLENGER, G. A. 1882. Catalogue of the Batrachia Salientia s. Ecaudata in the Collection of the British Museum. (ed. 2, London), pp. xvi+503, text figs., pls. i-xxx. 1883. Descriptions of New Species of Reptiles and Batrachians in the British Museum. Ann. Mag. Nat. Hist. (5), 12, pp. 161-167, pl. v. 1884. Diagnosis of New Reptiles and Batrachians from the Solomon Islands, collected and presented to the British Museum by H. B. Guppy, Esq. Proc. Zool. Soc. London, 1884, pp. 210-213. 1886. On the Reptiles and Batrachians of the Solomon Islands. Trans. Zool. Soc. London, 12, pp. 35-62, pls. vii—xiii. 1887. Second Contribution to the Herpetology of the Solomon Islands. Proc. Zool]. Soc. London, 1887, pp. 333-338, pl. xxviii. 1888a. Third Contribution to the Herpetology of the Solomon Islands. Proc. Zool. Soc. London, 1888, pp. 89-90. 1888b. Note on the Classification of the Ranidae. Proc. Zool. Soc. London 1888, pp. 204-206. 1890. Fourth Contribution to the Herpetology of the Solomon Islands. Proc. Zool. Soc. London, 1890, pp. 30-31, pl. ii. 1910. Les Batraciens et Principalement Ceux d’Europe. (Paris), pp. 1-305, 55 figs. 1918a. On the Papuan, Melanesian, and North-Australian Species of the Genus Rana. Ann. Mag. Nat. Hist. (9), 1, pp. 236-242. 1918b. Remarks on the Batrachian Genera Cornufer Tschudi, Platymantis Ginther, Simomantis, g.n., and Staurois Cope. Ann. Mag. Nat. Hist. (9), 1, pp. 372-375. 1920. A Monograph of the South Asian, Papuan, Melanesian, and Australian Frogs of the Genus Rana. Rec. Indian Mus., 20, pp. 1-226. Brown, W. C. 1949. A New Frog of the Genus Platymantis from the Solomon Islands. Amer. Mus. Novit., no. 1387, pp. 1-3. Brown, W. C., and Mynurs, G. 8. 1949a. A New Frog of the Genus Cornufer from the Solomon Islands, 62 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY with Notes on the Endemic Nature of the Fijian Frog Fauna. Amer. Mus. Novit., no. 1418, pp. 1-10. 1949b. A New Frog of the Genus Batrachylodes from the Solomon Islands. Jour. Washington Acad. Sci., 39, no. 11, pp. 379-380. Burt, C. E., and Burt, M. D. 1932. Herpetological Results of the Whitney South Sea Expedition. VI. Bull. Amer. Mus. Nat. Hist., 63, pp. 461-597, 38 figs. Cuuss, L. J. 1934. The Structure of the Pacific Basin. Geol. Mag., 71, pp. 289-302, 4 figs. DaruineTon, P. J., JR. 1948. The Geographical Distribution of Cold-blooded Vertebrates. Quart. Rev. Biol., 23, no. 1, pp. 1-26; no. 2, pp. 105-128. DEcKERT, Kurt 1938. Beitrage zur Osteologie und Systematik ranider Froschlurche. Sitz.-Ber. Ges. Natur. Freunde, Berlin, Jahrg. 1938, pp. 127-184, 49 figs. Dumérit, A. 1853. Mémoire sur les Batraciens Anoures, de la Famille des Hylaeformes ou Rainettes. Ann. Sci. Nat. (Paris), (3), 19, pp. 135-179. GrrarRD, CHARLES 1853. Descriptions of New Species of Reptiles, Collected by the U. S. Exploring Expedition, under the Command of Capt. Charles Wilkes, U.S.N., 2nd Part — including Species of Batrachians, Exotic to North America. Proc. Acad. Nat. Sci. Philadelphia, 6, pp. 420-424. GUNTHER, ALBERT 1858. Catalogue of the Batrachia Salientia in the British Museum. (London), pp. i-xvi+160, pls. i-xii. Guppy, H. B. 1887a. The Solomon Islands and their Natives. (Swan Sonnenschein, Lowrey and Co., London), pp. xvi+384. 1887b. The Solomon Islands, their Geology, General Features, and Suita- bility for Colonization. (Swan Sonnenschein, Lowrey and Co., London), pp. vii+152. Hepicer, HEIN1 1934. Beitrag zur Herpetologie und Zoogeographie Neu Britanniens. Zool. Jahrb., Abt. Syst., 65, pp. 441-582, 6 figs. Kampen, P. N. van 1923. The Amphibia of the Indo-Australian Archipelago. (Leiden), pp. xii +304, 29 figs. BROWN: AMPHIBIANS OF THE SOLOMON ISLANDS 63 KinGuorn, J. R. 1928. Herpetology of the Solomon Islands. Rec. Australian Mus., 16, pp. 123-178, pls. xili-xv, 35 figs. Lesson, R. P. 1830. ‘‘Zoologie’’, in M.L.I. Duperry, Voyage autour du Monde... sur la Corvette de sa Majesté, La Coquille, pendant les Années 1822, 1823, 1824 et 1825. (Paris), pp. iv-+743, and Atlas of 157 pls. Lever, R. J. A. W. 1937. The Geology of the British Solomon Islands Protectorate. Geol. Mag., 74, pp. 271-277, 1 map. 1945. The Giant Toad in the Solomon Islands. Agric. Jour. Fiji, 16, XO, Bi 05 Ile LovERIDGE, ARTHUR 1948. New Guinean Reptiles and Amphibians in the Museum of Com- parative Zoédlogy and the United States National Museum. Bull. Mus. Comp. Zodl., 101, pp. 305-430. MERTENS, ROBERT 1929. Zur Synonymie der Froschgattung Batrachylodes Boulenger. Zool. Anz., 80, pp. 266-268. Meyer, A. B. 1874. Uber die von ihm auf Neu-Guinea und den Inseln Jobi, Mysore und Mafoor im Jahre 1873 gesammelten Amphibien. Monatsb. - Akad. Wiss. Berlin, pp. 128-140. Myers, G. 8. 1950. Ability of Amphibians to Cross Sea Barriers, with Especial Refer- ence to Pacific Zoogeography. Proc. 7th Pacific Sci. Congress, New Zealand (in press). NIEDEN, F. 1923. Das Tierreich. Anura I. Subordo Aglossa und Phaneroglossa; Sectio 1 Arcifera. No. 46, pp. xxxii +584, 380 figs. (Berlin und Leipzig). NoBLgE, G. K. 1931. Biology of the Amphibia. (McGraw-Hill Book Co., New York), xiii +577 pp., 174 figs. Parker, H. W. 1939. Reptiles and Amphibians from Bougainville, Solomon Islands. Bull. Mus. Hist. nat. Belgique, 15, no. 60, pp. 1-5, 1 fig. 1940. Undescribed Anatomical Structures and New Species of Reptiles and Amphibians. Ann. Mag. Nat. Hist. (11), 5, pp. 257-274, figs. 1-3. 64 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Prrers, W. 1877. Herpetologische Notizen, II, Bemerkungen tiber neue oder wenigen bekannte Amphibien. Monatsb. Akad. Wiss. Berlin, pp. 415-423, 1 pl. Scumipt, K. P. 1932. Reptiles and Amphibians from the Solomon Islands. Field Mus. Nat. Hist., zool. ser., 18, pp. 175-190. SLEVIN, J. R. 1934. Templeton Crocker Expedition to Western Polynesian and Mela- nesian Islands, 1933, No. 15, Notes on the Reptiles and Amphibi- ans, with the Description of a New Species of Sea-snake. Proc. Cal. Acad. Sci., (4), 21, pp. 183-188. STERNFELD, ROBERT 1921. Zur Tiergeographie Papuasiens und der pazifischen Inselwelt. Abhand. Senckenberg. Naturf. Ges., 36 (for 1920), pp. 375-436, Dleexxcae Unirep States Navy DEPARTMENT 1944. Gazetteer (No. 1), Solomon Islands, Bismarck Archipelago and Islands off the Southeastern End of New Guinea. Hydrographic Office Pub. No. 881, pp. iv+188. Voat, THEODORE 1912. Beitrag zur Reptilien- und Amphibienfauna der Siidseeinseln. Sitz.-Ber. Ges. Naturf. Freunde, Berlin, Jahrg. 1912, pp. 1-13. WERNER, FRANZ 1894. Uber einige Novitiiten der herpetologischen Sammlung des Wiener zoolog. vergl. anatom. Instituts. Zool. Anz., 17, pp. 155-157. 1900. Die Reptilien- und Batrachierfauna des Bismarck-Archipels. Mitt. Zool. Mus. Berlin, 1, Heft 4, pp. 1-132, 46 figs. Manuscript received for publication December 3, 1951. PLATE 1 Batrachylodes vertebralis Boulenger, pectoral girdle (ventral view) Discodeles bufoniformis (Boulenger), pectoral girdle (ventral view) Platymantis papuensis weberi Schmidt, pectoral girdle (ventral Platymantis solomonis (Boulenger), pectoral girdle (ventral view) Palmatorappia solomonis (Sternfeld), pectoral girdle (ventral view) Cornufer guppyt Boulenger, pectoral girdle (ventral view) BULL. MUS. COMP. ZOOL. Brown. AMPHIBIANS OF SOLOMON ISLANDS. Plate 1 Figure 1 Figure 3 Figure 2 Figure 5 Figure 4 Figure 6 PLATE 2 Platymantis solomonis (Boulenger), terminal phalanx Batrachylodes vertebralis Boulenger, terminal phalanx Cornufer guppy Boulenger, terminal phalanx Discodeles bufoniformis (Boulenger), terminal phalanx Ceratobatrachus quenthert Boulenger, terminal phalanx BULL. MUS. COMP. ZOOL. Brown. AmpuiBiAns or SOLOMON Istanps. PLATE 2 Figure 2 Figure 4 Figure 5 PLATE 8 Wig. 1. Hyla thesaurensis Peters, head of larva Vig. 2. Hyla thesaurensis Peters, inferior view of hand BULL. MUS. COMP. ZOOL. Brown. AMPHIBIANS OF SOLOMON IsLANDs. PLate3 Figure | PLATE 4 Fig. 1. Platymantis solomonis (Boulenger), inferior view of hand Fig. 2. Hyla lutea Boulenger, inferior view of hand Fig. 3. Platymantis myersi Brown, inferior view of hand Fig. 4. Platymantis papuensis weberi Schmidt, inferior view of hand BULL. MUS. COMP. ZOOL. Brown. AMPHIBIANS OF SOLOMON ISLANDS. PLATE 4 Figure 3 Fig. Fig. Fig. Fig. po PISS PLATE 5 Platymantis aculeodactylus Brown, inferior view of hand Ceratobatrachus guenthert Boulenger, inferior view of hand Palmatorappia solomonis (Sternfeld), inferior view of hand Cornufer guppy? Boulenger, inferior view of hand BULL MUS. COMP. ZOOL. Brown. AMPHIBIANS OF SOLOMON ISLANDS. PLATE 5 Figure 2 Figure Figure 4 Fig. Fig. Fig. Fig. nN Re hm gS PLATE 6 Rana papua krefftii Boulenger, inferior view of hand Batrachylodes vertebralis Boulenger, inferior view of hand Batrachylodes trossulus Brown and Myers, inferior view of hand Rana papua novaebritanniae Werner, inferior view of hand BULL. MUS. COMP. ZOOL. PLATE 7 Fig. 1. Discodeles opisthodon (Boulenger), ventral view of foot. Fig. 2. Discodeles opisthodon (Boulenger), ventral view of hand Fig. 3. Discodeles bufoniformis (Boulenger), ventral view of hand BULL. MUS. COMP. ZOOL. Brown. AMPHIBIANS OF SOLOMON ISLANDs. PLATE 7 as 290° CO Sog Figure 2 Dp yas) SYo) Figure 3 PLATE 8 Fig. 1. Batrachylodes trossulus Brown and Myers Fig. 2. Platymantis myersi Brown BULL MUS COMP. ZOOL. Brown. AmrHiBiANs of SoLomon IsLtanns. PLaTe 8 I Ve S SS v Users Ee ari Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vou. 107, No. 2 REVISION OF THE ANT GENUS SERRASTRUMA By Wiuu1am L. Brown, JR. CAMBRIDGE, MASS., U. S. A. rok N aD SO at Hab MES yw Mi August, 1952 PUBLICATIONS ISSUED BY OR IN CONNECTION WITH THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE BULLETIN (octavo) 1863 — The current volume is Vol. 107. BREVIORA (octavo) 1952 — No. 4 is current. Menmotrs (quarto) 1864-1938 — Publication was terminated with Vol. 55. JOHNSONIA (quarto) 1941 — A publication of the Department of Mollusks. Vol. 2, no. 30 is current. OccaSIONAL PAPERS OF THE DEPARTMENT OF Mo.uusks (octavo) 1945 — Vol. 1, no. 16 is current. PROCEEDINGS OF THE NEW ENGLAND ZOOLOGICAL CLUB (octavo) 1899- 1948 — Published in connection with the Museum. Publication terminated with Vol. 24. These publications issued at irregular intervals in numbers which may be purchased separately. Prices and lists may be obtained on application to the Director of the Museum of Comparative Zodlogy, Cambridge 38, Massachusetts. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vou. 107, No. 2 REVISION OF THE ANT GENUS SERRASTRUMA By Wi.uiAm L. Brown, Jr. CAMBRIDGE, MASS., U.S. A. Paw N teh ELO dine ove WS Vi August, 1952 2 i iy i 4 Fin No. 2. — Revision of the Ant Genus Serrastruma By WittrAM L. Brown, JR. INTRODUCTION When, in my division of the large and heterogeneous group which had long been known as Strumigenys, I demonstrated the essential differences between the ‘‘long-mandibulate” or sensu stricto species and the shorter-jawed forms, the latter were mostly placed in the genus Smithistruma, which was divided in turn into four subgenera. An African group of species which seemed distinctive was assigned the subgeneric name, then new, of Serrastrwma (Brown, 1948, loc. cit. infra). At that time, the possibility still remained that previously- described but obscure short-mandibulate species might be annectant between these two groups (Smithistruma sensu stricto and Serrastruma). By 1949 (Brown, loc. cit. infra) all the described species but one or two were well enough known to permit the clean separation of the latter group as a genus in its own right. Subsequent investigation of strumigenite genera has amply confirmed this separation. Serrastruma stands as a very compact and homogeneous group, distinct from al] other related genera both in the essential plan of mandibular structure and in general habitus. The present study was initiated with the inclusion of 26 specific, subspecific and varietal names, none of which had been seriously challenged previously. This number is now reduced by synonymy to seven reasonably distinct species and two species inquirendae. It appears doubtful that many more new species will be recognized in this group during the strictly morphological period of investigation, and synonymy will probably still further reduce the number when certain types are made available for study. MATERIAL STUDIED AND ACKNOWLEDGEMENTS Specimens seen during the present investigation include types of most of the previously described forms, which I have fortunately been able to borrow or gain through exchange with various European collections. In addition, previously unstudied material in abundance has arrived from several sources. The largest single group of material is that collected during two extensive tours in Africa by Dr. Neal A. Weber of Swarthmore College and the American Museum of Natural History. Dr. Weber’s material probably equals by itself the entire array of material seen by all previous authors taken together. 68 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Future additions of material in Serrastruma collections will probably lead to the further reduction of the number of species rather than to proliferation. Further collection of females and males in association with workers is highly desirable, for these castes, when known, may dissolve certain doubts remaining after the present treatment. The females seem to show particular promise as a systematic aid. I owe thanks to the following individuals and institutions for their invaluable aid in the matter of types, etc. Dr. H. Bischoff, Zoologisches Museum der Universitat, Berlin, camera, lucida sketches and measurements of reticulata types. Dr. Charles Ferriére, Museum d’Histoire Naturelle, Geneva, loans and sketches of Forel types. Dr. George Arnold, National Museum of Southern Rhodesia. M. Pierre Basilewsky, Musée du Congo Belge, Tervuren, Forel and Santschi types. Sig. Mario Consani, Florence, Italy, Santschi and Menozzi types. Father J. van Boven, Roermond, Netherlands, Wasmann types. Dr. Delfa Guiglia, Museo Civico di Storia Naturale, Genoa, Emery types. Dr. J. C. Faure, University of Pretoria. Mr. H. St. J. K. Donisthorpe, British Museum (Natural History), Donisthorpe types. Dr. M. R. Smith, United States National Museum. Dr. E. S. Ross, California Academy of Sciences. Special thanks are also due Dr. Weber and Dr. J. C. Bequaert for aid both in the matter of material and in other means essential to the success of this paper. Other than the collections indicated above, the only important depository is the Santschi Collection in the Natural History Museum at Basle. Unfortunately, it has not been possible to view material from this latter source. Through exchange, the first important collections of Serrastruma have been built up in the United States at the Museum of Comparative Zoology, Harvard University, at the American Museum of Natural History, and at the U. S. National Museum. Other collections have also been augmented. GEOGRAPHICAL DISTRIBUTION The natural range of Serrastrwma embraces the whole of the Ethiopian Region, except for the most arid portions. Certain species, like Jujae and serrula, appear to be restricted to rain- and gallery-forest, in which they are widely distributed. S. bequaerti prefers the cool, humid montane forests of central parts of the continent, while S. stmoni BROWN: REVISION OF ANT GENUS SERRASTRUMA 69 has an extraordinary distribution, ranging from the dry belt just south of the Sahara to the Cape region, but avoiding most rain-forest areas. We have too little data concerning the other species at present, but lotti and mayneit appear to be moist-forest inhabitants, while allwaud? seems to range widely in several vegetational zones, but avoids or is rare in the Congo forests. The probable youth and dominance of the genus is reflected in the collection rate, for to date more records of Serrastruma captures are known from Africa than for all other dacetine genera combined. Smithistruma is weak in species and numbers of colonies as compared to other faunas, and this is also true to a lesser extent of Strumigenys. The latter genus has developed a special group (group rogeri Emery) with high dominance, able to withstand competition from most dacetine genera, but it still seems subordinate to Serrastrwma within its natural range. In discussing competition, it is assumed that Serrastruma, like related genera, feeds on collembolans captured by stalking. No confirmation of this assumption has yet come forth, but it still seems very likely to be correct in view of the basic tendency to collembolan predation in all dacetine genera so far studied. The colonies are reported from much the same situations as are chosen by Smithistrwma— in rotten logs, moss, under bark, under stones, in leaf litter, etc. Nest series I have studied, accompanied sometimes by collectors’ remarks on their completeness, indicate that up to 300 workers and 5 queens are not uncommon colony-popula- tions. Outside continental Africa, flourishing populations of Serrastruma are known from the islands in the Gulf of Guinea and from Mauritius. One species, ludovici, occurs on Madagascar, but we know nothing about the abundance of the genus on this island. The Mauritius populations (S. simoni and S. allwaud?) certainly, and the Madagascar record possibly, represent recent introductions through human com- merce from an African source. The possibility of “tramping’’ is demonstrated by the recent removal of a colony, containing males and workers at least, of S. lujae from plants arriving at Honolulu in U. 8. Plant Quarantine. The origin of this shipment was the Belgian Congo. Probably Serrastruma species have been introduced elsewhere outside Africa in the tropics, but no further records have yet turned up. The tramping ability of this genus gives another hint of its dominance under varied conditions. 70 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY SERRASTRUMA Brown Smithistruma (Serrastruma) Brown, 1948, Trans. Amer. Ent. Soc., 74: 107-108. Strumigenys of authors, part, nec Fred. Smith. Strumigenys (Cephaloxys) of authors, part, nec Fred. Smith. Strumigenys (Trichoscapa) of authors, part, nec Emery, 1869. Serrastruma Brown, 1949, Mushi, 20: 6, 7. Genotype. Strumagenys simont Emery, by designation of Brown, 1948, as subgenotype. Gross External Morphology and Generic Characters Worker. Resembling in size, color, sculpture and most structural features the same caste of Smithistruma Brown. Head together with the closed mandibles almost perfectly and rather broadly cuneiform seen in dorsal view, the sides nearly straight, evenly converging anteriorly; posterior occipital border feebly to moderately and rather broadly excised. Head seen from the side subcuneiform, rather thick posteriorly and tapering anteriorly, dorsum convex; antennal scrobes long and broad, shallow to moderately deep. Eyes small, placed just dorsad of the ventral scrobe margins posterior to the cephalic mid- length, exposed to direct dorsal view. Antennae as in other short-mandibulate strumigenite genera, with similar segmentation; scapes slender, or at most very feebly incrassate. Clypeus with a broadly triangular disc having a transverse, more or less arcuate to approximately straight ‘false margin” which bears the principal hairs of fringing pilosity if present. Anterior to and de- pressed below the level of the false margin is a translucent median lobe or apron with rounded free margin, the latter normally covering or fitting between the bases of the mandibles. This lobe has fre- quently been misinterpreted as an intermandibular space or other structure. The mandibles are narrowly triangular or subtriangular, their exposed length slightly to very much longer than the clypeus; seen from the side, they are weakly arched. The armament consists of very fine serial denticulation, directly opposable throughout and occupying the entire inner (apical or masticatory) margins; at least 30 denticula- tions on each mandible. Most specimens show clear development of a small but stout apical and one to three minute subapical teeth, these acute, alternating with and projecting slightly beyond the even level of the denticulation as seen at high magnifications. The basalmost denticle is slightly larger and more rounded than the succeeding series, and in certain species (Group A, see below), the five or six most basal denticulae are distinctly larger, coarser and more acute than the succeeding ones and are sometimes alternate with extremely BROWN: REVISION OF ANT GENUS SERRASTRUMA a minute (intercalary) denticulae. The majority of the denticulation appears to represent secondary erosion of a very long, narrow basal lamella such as is found in Smithistruma alberti (Forel). Jf this homology is correct, it becomes apparent at once that the diminutive acute subapical teeth (alternating with lower denticles) are strictly comparable with the similarly alternating teeth of the apical series as seen in Smithistruma of groups alberti and capitata (Fred. Smith). In comparison with these groups of the related genus, one can find many reasons for assuming that they represent the primitive stock from which Serrastruma arose. I presently consider this assumption to be correct. Neither Smithistruma group is now found in Africa, but it appears probable that their ancestral stock once inhabited the Ethiopian Region. At any rate, the form and armature of the Serra- struma mandible is presently very distinctive in its modifications. Labral lobes small and inconspicuous, but projecting and conical, much as in Smithistruma. The palpal segmentation (1, 1) is as in Smithistruma, and the maxillary palpi are similarly reduced to minute vestiges. Alitrunk with a robust and convex promesonotum, the promeso- notal suture weakly indicated or obsolete; metanotal groove strong and deeply impressed; propodeal dorsum rising from the groove and usually more or less convex, at least anteriorly. Propodeal teeth subtriangular, acute, laterally compressed, small to fairly large (obtuse and more or less vestigial in S. bequaerti), continued below by a narrow to cariniform infradental lamella on each side of the propo- deal declivity. Petiole with slender peduncle and distinct, raised, more or less dorsally-rounded node. Postpetiole transversely elliptical, convex dorsally, always distinctly broader than the petiolar node. Spongi- form appendages vestigial to only moderately well-developed, espe- cially on the petiole. Contrary to the observations of former authors, however, I have found all species to possess at least some remnants of appendages on the postpetiole. The midventral strip of the petiole is most often weak or obsolete. Gaster as in related genera of the Strumigeniti; not markedly de- pressed; anterodorsal spongiform margin feeble or absent, but the usual basal costulae present, variably distinct; sting developed. Sculpture paralleling that of Smuthistruma. Head, mesonotum, propodeal dorsum, and usually the petiole finely and densely reticulate punctate and opaque. Gaster, except for basal costulae, smooth and shining. Pronotum and postpetiole varying in sculpture with the species, as do the sides of the alitrunk. Mandibles smooth; clypeus extremely finely punctulate-granulose, opaque, as are also the scapes and legs to a greater or lesser degree. 72 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Pilosity basically as in Smithistruma. Ground hairs short, reclinate or subreclinate, fine to spatulate or spoon-shaped. Fringing hairs of scapes and false clypeal margin usually larger and more conspicuous; ground hairs of promesonotal dorsum small and usually inconspicuous. Specialized erect hairs more or less stiff, fine or narrowly spatulate, oar-shaped, or clavate, those on the nodes and gaster usually larger and heavier than those on the head; 2 to 8 on head, situated well back on occiput; mesonotum with one or two pairs, usually situated on dorsolateral tubercles; nodes and gaster with a more numerous, but still limited set each. In addition, the lateral border of each occipital lobe bears a fine, outstanding flagellate or subflagellate hair, while each humerus bears a long flagellate or spatulate hair. Color yellow to dark ferruginous. Female. Similar to worker, with the usual full sexual characters. In Serrastruma, this caste is usually larger and darker in color relative to the accompanying workers than in related strumigenite genera. Male. Known only for a few species. Mandibles small, but oppos- able; triangular, the inner margins weakly produced as broadly rounded lamellae; apices acute (S. lujae). Otherwise paralleling the male of Smithistruma, especially in dark color, sculpture, etc. Specific Characters, Variability, ete. Serrastruma betrays its relative youth as a genus and differs from most other dacetine genera in the decided variability of some of its species and in the concomitant slightness of specific distinctions. Because of these qualities, it is a “difficult” genus from the taxono- mist’s point of view. Most of the past taxonomic confusion in the group, however, cannot be blamed entirely upon variation or slightness of specific distinctions. This confusion was due partly to the fact that Strumigenys, in the old, portmanteau sense, has effectively concealed, one might say “swallowed up”, the group, so that efforts to distinguish the species have been dissipated in comparison with a much larger number of short-mandibulate forms actually having only a distant relationship to Serrastruma. Thus, the two species ludovici Forel and reticulata Stitz were effectively lost into Strumigenys sensu stricto when subsequent classifiers interpreted too literally the uncannily misleading original descriptions. Camera lucida drawings of types enable me to place these two forms quite certainly in Serrastruma. Beyond this, the extensive synonymy must be laid to pure careless- ness in construction and subsequent interpretation of descriptions. Apparently, when authors were describing a ‘‘new”’ form, reference was had almost entirely to previous descriptions instead of to types or = BROWN: REVISION OF ANT GENUS SERRASTRUMA 13 reliably identified material, even though the latter might be readily available, or even in the specialist’s own collection. That many of the older descriptions were seriously in error is now clear. The lack of critical appraisal of descriptions in the past is amazingly general, as one can determine from reference to the discussion of the synonymy of S. simont Emery, below. Unfortunately, the poor systematic work visted upon Serrastruma has also been widespread among other, much larger formicid groups that will not be so easy to untangle. A con- clusion to be drawn from Serrastruma is that, though a given group of ant species may show variation and blurring of interspecific characters, these “difficult” qualities are seldom of the degree of seriousness indi- eated by the usually excessive number of synonyms and infraspecific variants named in the group. Thus, in the case of S. s?monz, Santschi repeatedly described forms the types of which, under present direct comparison, fail utterly to show the differences cited in the original descriptions. In my own attempt to sort Serrastruma material into species, a very careful search of the worker caste was made in an attempt to recognize features of consistent value in separating species. Few were found, and this fact is reflected in the very large new synonymy. However, one seemingly constant and most useful character concerns the mandibular dentition, falling into two sorts as follows: Group A In alluaudi, lotti and ludovici, the mandibles are long (MI 40 or more), relatively slender and evenly tapered toward their apices; basal quarter or fifth of the apical margins each with the denticulae suddenly and decidedly larger, coarser, more irregular and more acute than those following distally. This coarse basal series usually numbers 5-6 units on each mandible; each unit may alternate with an indefinite minute denticle. Group B In lwjae, serrula, maynei, bequaerti and simoni the mandibles are a bit more robust and are usually under MI 40, rarely slightly more. The apices are somewhat more blunted in dorsal view. The denticu- lation is fine (lujae, maynei, bequaerti) or extremely fine (serrula, simoni), and the denticulation toward the base of the apical border is not or only very slightly and gradually coarsened, regular and not acute; the basalmost denticle may, however, be slightly enlarged and sublamelliform. 74 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY In doubtful cases, the denticulation is best examined by placing the mandibles in dark contrast over a brightly lighted white background. The differences are so plainly apparent that, if magnification is at the necessary 60-80 diameters, contrast lighting should rarely be necessary. Other characters of use in distinguishing species concern the sculp- ture, especially of the pronotum and nodes of the pedicel; relative head width; pilosity; development of propodeal teeth and lamellar or spongiform appendages, etc. Measurements and proportions given are considered the bare minimum essentials, here cited for the first time. Former statements of “‘total length” have been highly inaccurate and are not at all comparable, even in the consecutive descriptions of one author. Abbreviations: HL— Maximum measurable length of head from dorsal view, mandibles excluded. Measurement is made from the center of the true anterior clypeal margin to a line connecting the posterior occipital extremities. CI — Cephalic index: maximum measurable idl of head expressed as a percentage of the head length, or head width/HL * 100. MI — Mandibulo-cephalic index: “exposed length” of mandibles expressed as a percentage of the head length. Measurement is made in the fully closed condition from the center of the true anterior clypeal margin to the extreme apex of the most advanced mandible. Exposed mandibular length/ HL ~* 100. These measurements and proportions are standard in my works on the Dacetini. SYSTEMATIC TREATMENT BY SPECIES In the synonymies below, only essential! references are cited, and those appearing subsequent to about 1920. Full synonymies are cited in: Emery, C., 1922, in Wytsman’s Genera Insectorum, Fase. 174, pp. 320, 324 (as Strumigenys). Wheeler, W. M., 1922, Bull. Amer. Mus. Nat. Hist., 45: 918-920, 1034 (as Strumigenys). An early key to African Strwmigenys is of interest only as an illus- tration of the early confusion of the taxonomy of the group: Santschi, 1913, Bull. Soc. Ent. France, pp. 257-259. The best habitus figures of Serrastruma (though portraying the mandibular dentition as a somewhat inaccurate convention) are those BROWN: REVISION OF ANT GENUS SERRASTRUMA 5 of Eidmann, for which references are given in the synonymy of S. lujae, below. Cited with the descriptions below are all type localities, whether or not types have been examined. All other records represent specimens actually seen during the course of this work. SERRASTRUMA ALLUAUDI (Santschi) new combination Strumigenys (Trichoscapa) allwaudi Santschi, 1910, Ann. Soc. Ent. France, 79: 360, worker, female; original description. Strumigenys oleae Wasmann, 1918, Ent. Mitt., Berlin, Us 142, Pl. 2, figs. 9, 10, worker. NEW SYNONYMY. Sieunigenis (Cephaloxys) raymondi Donisthorpe, 1945, Ann. Mag. Nat. Hist., (11) 12: 779, part., worker. 1946, Jbid., 13: 32; part. NEW SYNONYMY. Worker. HL 0.55-0.60 mm., CI 75-78, MI 40-44. Mandibles of Group A form. Postpetiole swollen and very convex, its surface weakly to distinctly sculptured, more or less opaque. Sides of alitrunk with at least some punctulate-reticulate sculpture; not completely smooth and shining as in s¢monz, and the propodeal teeth slightly less well- developed than in simoni. Otherwise, this species is very much like simoni and likely to be confused with that species (gq. v.). Pilosity variable, especially the shorter ground hairs. At least part of the variation, both in erect and ground pilosity, seems to be due to differing amounts of adherent foreign matter, possibly a hardened secretion, which makes the hairs appear thicker at their apices when abundant. Color variable as in simoni; vertex often infuscated. I have not seen types of alluaudi, but specimens determined by Dr. Arnold and stemming from Natal are considered near-typical or typical. Dr. Arnold was in close contact with Santschi for many years, and his specimens agree closely with Santschi’s characterization. A rothkirchi type from the Wasmann Collection, kindly sent by Father van Boven, agrees closely with the Natal specimens and with a cotype of raymondi sent by Mr. Donisthorpe. Mr. Donisthorpe has disagreed (in litt.) with my synonymy of raymondi, and a subsequent visit to the British Museum may possibly explain his disagreement. The specimen labelled “type” in the British Museum is actually a specimen referable to s7moni, while those labelled ‘‘cotypes” (= para- types in American usage) are divided between simoni and alluaudi specimens. Evidently both species are well established on Mauritius, and Mr. Donisthorpe has confused them under a new name. Speci- mens taken by Weber at Kampala, Uganda are the most atypical I 76 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY have seen, but even here the difference is largely one of pilosity and is not considered taxonomically significant. S. alluaudi is widely spread in Africa, and seems to have been carried about a greal deal by human commerce. The “rothkirchv”’ specimens from Mt. Kamerun may, as Dr. Bequaert has suggested, have been taken in or near the former German horticultural experiment station. During the period of German administration, plants were brought here from all parts of Africa, and might easily have harbored migrant Serrastruma. Otherwise, allwaudi seems to have its main natural range in eastern Africa. The Mauritius records must be put to tramp mi- gration, while the Madagascan Judovici, quite possibly identical with alluaudi, may have come from Africa in the same way. Alluaudi has not yet been taken in the Congo rain-forests, and in this respect it resembles distributionally the related s7monz. However, further in- vestigation may show that Jott?, a smaller but similar species which ranges into the eastern Congo Basin, is only an extreme variant of alluaudi. The male and female of allwaudi have not been studied by me. Type locality. Grotto of Tanga, ‘“Kulumuzi,’ German East Africa (C. Alluaud). Types are probably in the Paris Museum and the Santschi Collection. Localities for material examined. Kampala, Uganda (N. A. Weber, series no. 1503). Eshowe, Natal (G. Arnold). Soppo, 730 m., Kame- runberg (von Rothkirch); cotype of rothkircht. Cocotte Mt., Mauritius (R. Mamet); cotype of raymondi. Other raymondi cotypes from Mauritius briefly examined in the British Museum, several series. SERRASTRUMA LOTTI (Weber) new combination Strumigenys (Cephalorys) escherichi subsp. lotii Weber, 1943, Bull. Mus. Comp. Zool. Harvard, 93: 327, 378-379, Pl. 15, fig. 13, worker, female, biology; original description. Worker. HL 0.46-0.49 mm., CI 80-84, MI 42-46. Female con- siderably larger than the worker and darker in color, but not large enough to fit readily as the female of allwaudi, considering the workers of the latter species. Cotypes have been compared directly with addi- tional specimens from the localities cited below, and agreement is close. This form is essentially a smaller and lighter-colored version of allwaudi with finer pilosity. The form and denticulation of the mandibles in the two species are quite similar, although those of lotti seem slightly longer relatively. The sculpture of the pronctum is similar and equally variable in both species. Future collections may show that intergrades occur and that synonymy is indicated, but the present material does BROWN: REVISION OF ANT GENUS SERRASTRUMA i not support this. The species calypso (species inquirendae section below) is poorly described, and may quite possibly be the same as lotti, over which it has nomenclatorial precedence. Errors in Weber’s description should be noted. S. lott: is not at- tachable to “‘escherichi”’ because of the dentitional differences. The figure shows the shape of the scape quite wrongly, and depicts the funiculus as seven-segmented, whereas the funiculus, as in all Ser- rastruma, is really five-segmented. This species has been taken in the eastern Congo and in the adjacent gallery-forests of the Sudan; it is probably widespread in eastern Africa. Type locality. Lotti Forest, west slope of Imatong Mts., Anglo- Egyptian Sudan (N. A. Weber, series no. 1451). Cotypes are in Dr. Weber’s collection and in the Museum of Comparative Zoology, Harvard University, as well as in other institutions. Several workers and a dealate female have been examined from the type series. Material examined from other localities. Belgian Congo: Stanleyville (A. Collart). Beni to Irumu (N. A. Weber, no. 2124). Anglo-Egyptian Sudan: Kagelu, Equatoria (Weber, no. 1284). SERRASTRUMA MAYNEI (Forel) new combination Strumigenys maynei Forel, 1916, Rev. Suisse Zool., 24: 427, worker, female, male; original description. Strumigenys maynei var. latiuscula Forel, 1916, [bid., p. 428, worker. NEW SYNONYMY. Worker. HL 0.52-0.56 mm., CI 83-87, MI 36-40. Color yellowish- to medium ferruginous. Mandibular dentition of Group B, and closely resembling that of lujae. Readily distinguished by means of the sharp, definite, fine and close longitudinal costulation (or striation) of the pronotum, by the broad head, with pronounced, thin lamelliform borders along the upper margins of the antennal scrobes, and by the small but conspicuous, broadly spoon-shaped hairs of the cephalic and promesonotal ground pilosity. The differences in head width and color between the types ef maynez and its variety latiwscula presently before me are perceptible, but scarcely significant for taxonomic purposes. The very much greater variation shown by different series of the re- lated lujae confirms for me the opinion that latiwscula should never have originally received nomenclatorial distinction. Type locality. Stanleyville, Belgian Congo (Kohl); three cotypes examined from the Forel Collection, one of which has been retained as an exchange in the Museum of Comparative Zoology. 78 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY A single cotype worker of var. latiuscula has been received in ex- change from the Musée du Congo Belge; the type locality for this form is Kala, Belgian Congo (R. Mayné). No other material has been seen. SERRASTRUMA LUJAE (Forel) new combination Strumigenys lujae Forel, 1902, in Wasmann, Allg. Zeitschr. Ent., 7: 51, nota, Pl. 1, fig. 1, worker; original description. Strumigenys reticulata Stitz, 1910, Mitt. Zool. Mus. Berlin, 5: 141, worker. NEW SYNONYMY. Strumigenys (Cephaloxys) glanduscula Santschi (‘new var.”’), 1919, Rev. Zool. Air., 7: 88, worker. Bequaert, 1925, Ibid., 8: 146, biology. NEW SYNONYMY. Strumigenys (Cephaloxys) gerardi Santschi, 1923, Rev. Zool. Afr., 11: 287-288, worker. NEW SYNONYMY. Strumigenys (Cephaloxys) aequalis Menozzi, 1942, Zool. Anz., 140: 177-178, worker, female; original description. Eidmann, 1943, Mitt. H. Goring Akad. Deutsch. Forstwiss., 1: 262, fig. 25, worker, biology; 1944, Zool. Jahrb. Syst., 76: 457-458, figs. 15, 16, worker, biology. NEW SYNONYMY. Worker. HL 0.52-0.71 mm., CI 77-83, MI 37-42. Distinguished from all other species except serrula and bequaerti (q. v.) by virtue of its Group B mandibles, evenly rounded, densely reticulo-punctulate pronotum, its fine erect and ground pilosity of the head, and its spongiform appendages, which are strongly reduced, especially the mid-ventral petiolar strip; the latter appears as a low, blunt, non- spongiform, carina-like vestige. Funicular segments II and III varying from slightly to considerably longer than thick. Postpetiole with weak appendages; dorsal surface usually reticulate-punctulate and opaque, but occasionally having the sculpture effaced and the surface nearly completely smooth, shining. Median carinula of the pronotum absent or feeble. Color somewhat variable, but usually lighter or darker yellowish-ferruginous. Female larger than workers from the same nest, and usually a bit darker. Lateral occipital hairs borne on very low, weakly convex lamellae or carinae which do not occur, or at least are not readily apparent, in the workers. The males vary in sculptural and other details, even within one nest series. The color is variable, but usually is predominantly dark casta- neous or brownish-black, with the alitrunk lighter. The volsellae- laciniae are not markedly different from those of many Smithistruma species; these will be figured in another paper. The extensive synonymy proposed here for lujae, unlike that of simoni, is at least partly a reflection of the considerable variation BROWN: REVISION OF ANT GENUS SERRASTRUMA 79 shown by different nest series. Forel originally described the petiole and postpetiole as without spongiform appendages, although a type received from the Wasmann Collection shows a fine posterodorsal collar on the petiole and small but distinct ventral postpetiolar ap- pendages of spongiform consistency. This misdeseription threw later authors off the track rather badly, as Santschi’s hesitant description of glanduscula clearly shows. A cotype of glanduscula in the Museum of Comparative Zoology (on loan from the Musée du Congo Belge) com- pares well with the lujae cotype, but is slightly smaller. Stitz described reticulata so poorly that all subsequent workers were led to include it among the species of Strumigenys sensu stricto. A camera lucida sketch and measurements of a type, kindly furnished by Dr. Bischoff, show that this species is only a small variant of lujae like those taken in the Honolulu Plant Quarantine. Beyond rectification of these obvious errors, the treatment of lujae becomes more subject to theoretical considerations. The series I have actually seen represent 17 separate nests from nearly as many locali- ties. While each nest series is relatively homogeneous in size and proportions, certain of the series appear very different, especially in size, when directly compared. The various series may be arranged into a completely intergradient row, each broadly overlapping the next in all variable characters, and connecting the extremes. Such slight differences as occur in proportions, mandibular denticulation, pilosity, development of propodeal teeth and pronotal carinula, etc. seem to be correlated with overall size, and one cannot escape the impression that the very noticeable size difference, would, if known only from the inspection of a few series, form the most striking means of dis- tinguishing taxonomic entities. Nevertheless, the intergradation is solidly established in the present material, and this alone will serve to negate any separation in the absence of evidence for geographical apportionment. Large specimens convene well with a type of gerard, medium ones with types of /ujae and glanduscula, and the smallest ones with reticu- lata and the specimens from Hawaiian Plant Quarantine. The largest specimens I have seen (Sad Tomé) are larger than the gerardi types, while the smallest (Burunga and Honolulu) are smaller than cotypes of aequalis examined through the courtesy of Signor Consani. The essential relationship (and non-distinctness) of all these series seems to me more striking than the known variable “‘specific characters,” and I. have emphasized the former quality in the synonymy. The variability is such that future investigation may well show that serrula and even bequaerti are to be included. S. lujae is clearly a sylvicolous species, and the most commonly- collected form of Serrastruma in the equatorial forest belt and adjacent 80 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY gallery-forests. Its presence on the islands in the Gulf of Guinea and in plant shipments at Honolulu indicates with what ease it can spread to new areas. The nests are made in rotten wood or in the soil or soil cover. I have examined series from nests containing up to 300 workers and four or five queens plus numerous winged forms of both sexes, males predominating. Males are often present as fully-pigmented free imagoes while the majority of the females are still in the pupal stage. Large nests are often infested with an apparent myrmecophile- a small ferruginous staphylinid beetle. Type locality. Morumballe, on the Zambesi River, Portuguese East Africa (E. Luja). Types are in the Forel and Wasmann Collections; I have seen one worker from the latter source. Additional localities for material seen. Belgian Congo: South of Watsa, Ituri Forest. West side of Ruwenzori (N. A. Weber, nos. 2139, 2112). Burunga (J. C. Bequaert). Yambuya (J. C. Bequaert); cotype of glanduscula. Manyema (Gerard); cotypes of gerardi. Precise source unknown, via U.S. Plant Quarantine, Honolulu, Hawaii; workers and males, in plants. Cameroon: Gross Batanga (G. Schwab). Sad Tomé Island: Makambrera, 4000 feet. Roca Zampalma, 2500 feet (B. Malkin); three colonies and a stray worker. Fernando Po Island: Concepcion (H. Eidmann); cotypes of aequalis. Uganda: Fort Portal (N. A. Weber, nos. 2095, 2102, 2103); several collections, including females. SERRASTRUMA BEQUAERTI (Santschi) new combination Strumigenys (Xephaloxys) [sic] bequaerti Santschi, 1923, Rev. Zool. Afr., 11: 286-287, worker; original description. Strumigenys (Cephaloxys) bequaerti Bequaert, 1925, Rev. Zool. Afr., 13: 146, biology. Worker. HL 0.58-0.61 mm., CI 73-78, MI 35-37. Quite similar to medium-sized specimens of lujae, but with the head narrower on the average. Second and third funicular segments long and cylindrical, nearly or quite twice as long as thick.The chief distinction of bequaerti lies with its vestigial propodeal teeth; these are extremely reduced, little more than pronounced angles, obtuse or subrectangular in profile. Other characters as in lujae. This form may eventually prove to be a montane subspecies or even a synonym of lujae. Dr. Bequaert says (loc. cit.) of the type collection (in translation): “The nest of this ant was found at about 2200 meters altitude in the humid montane forest of the Butagu Valley, on the west side of Ruwenzori. It was situated in the humid and strongly shaded soil.” Additional specimens, which convene well with the cotypes at my BROWN: REVISION OF ANT GENUS SERRASTRUMA 81 disposal, were taken by F. Meneghetti in the similar cool montane forest (Mau Forest) of the Kenya Colony. This series, which contains winged females, reached me through the courtesy of Signor Consani, who has indicated his desire to prepare the description of this winged caste. Cotypes are in the Musée du Congo Belge, Museum of Comparative Zoology, Consani Collection, and presumably inthe Santschi Collection. SERRASTRUMA SERRULA (Santschi) new combination Strumigenys lujae var. serrula Santschi, 1909, Ann. Soc. Ent. France, 78: 390- 391, worker; original description. Sirumigenys (Cephaloxys) serrula Santschi, 1923, Rev. Zool. Afr., 12: 288, fig. 4c, worker. Sirumigenys (Trichoscapa) concolor Santschi, 1914, Boll. Lab. Zool. Portici, 8: 375, worker. NEW SYNONYMY. Strumigenys (Cephaloxys) uelensis Santschi, 1923, Rev. Zool. Afr., 12: 289-290, fig. 4b, worker. NEW SYNONYMY. Worker. HL 0.45-0.52 mm., CI 80-89, MI 34-39. The distinctions between the worker of this species and that of small lujae series is not very satisfactory. The smaller size, shorter mandibles, and usually broader head will serve to distinguish most serrula specimens. Even the smallest /ujae specimens will not show these dimensions and pro- portions all within the serrula rangeandsimultaneously. Many serrula specimens from French Equatorial Africa and from West Africa (con- color cotype) show a well-developed lamelliform margin along the dorsal scrobal borders. On each side at the point of greatest lateral expansion of the occipital borders, these margins end rather suddenly at a gentle depression, from which arises the lateral flagelliform hair. Specimens from the Belgian Congo and Uganda often lack the lateral occipital depression or show it in very weak form, while others from these localities have it distinctly developed. Females from all localities seen so far all have the depressions distinctly developed. This character will serve to distinguish the females and most workers of serrula from lujae, in which the female, it will be remembered, has the lateral occipital hair arising from a low, convex ridge. Both female castes of serrula seen in full-face view tend to have the external mandibular margins very feebly convex to straight; in /ujae the same margins are straight to feebly concave. A serrula female from south of Watsa in the Belgian Congo (Weber, no. 2139) shows a total (synthetic aggre- gate) length, mandibles included, 2.47 mm., HL 0.52 mm., CI 87, MI 35. Other females from various localities showed scant variation away from these values. The difference in size from the lujae female is considerable. 82 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY In sculpture, the worker is like that of Jujae, but some specimens have the punctulation of the pronotum overlain with feebly-suggested longitudinal striation or costulation. The hairs of the ground pilosity are usually slightly broader than those of Jujae, and the punctulation of the upper occipital region appears relatively coarser. The node of the petiole is not so high or so steep anteriorly as in /ujae. The den- ticulation of the mandibles appears to be finer than in small lujae specimens, and a magnification of 80-100 diameters is needed to resolve the separate units with any clarity. Color varying from pale yellow to medium ferruginous. In spite of all the characters cited, the species remains doubtfully distinct from Jujae. In the series collected by Dr. Weber, specimens of Jujae and serrula were on at least three occasions (nos. 2095, 2103 and 2139) presumably taken together or in close proximity. Dr. Weber’s field notes, not presently available, will be of considerable interest when published. Were it not for the seemingly clearcut differences in the female caste, I should be tempted to combine the small and large forms under the name /ujae. A concolor type sent by Sig. Consani, mentioned above, seems to fall within the limits of (continuous) variation of the abundant series here referred to serrula, and nothing in Santschi’s descriptions will serve to force separation of the two. The form uelensis is separated on the basis of the supposedly straight anterior clypeal margin. I have seen no specimens from the type series, but review of Santschi’s de- scription and figure convinces me that he overlooked the anterior clypeal lobe or apron; his anterior border is actually the false border. Santschi was unsure of the distinctness of this form himself, and it seems certain that we owe the name to his hasty examination of the types. Type locality. Brazzaville, Congo (Weiss). Type not seen. Additional material seen. Abundant material collected by Dr. Weber at virtually all of the localities in the Belgian Congo and Uganda from which he also secured S. /ujae, also a series collected by him at Haut Mbomu in French Equatorial Africa. A cotype of concolor Santschi: type locality, Aburi, Gold Coast (Silvestri). SERRASTRUMA SIMONI (Emery) new combination Strumigenys simoni Emery, 1895, Ann. Soc. Ent. France, 63: 42, Pl. 2, fig. 21, worker; original description. Strumigenys escherichi Forel, 1910, Zool. Jahrb. Syst., 29: 261, worker. NEW SYNONYMY. Strumigenys cognata Santschi, 1910, Ann. Soc. Ent. France, 79: 362, worker. NEW SYNONYMY. BROWN: REVISION OF ANT GENUS SERRASTRUMA 83 Strumigenys (Trichoscapa) cognata st. beerorum Santschi, 1913, Bull. Soc. Ent. France, p. 259, worker. NEW SYNONYMY. Strumigenys (Trichoscapa) biconvera Santschi, 1913, Bull. Soc. Ent. France, p. 258, worker. NEW SYNONYMY. Strumigenys eschericht subsp. limbata Forel, 1913, Deutsche Ent. Zeitschr., p. 222, worker. NEW SYNONYMY. Strumigenys escherichi var. cliens Forel, 1913, Rev. Zool. Afr., 2: 317, worker. NEW SYNONYMY. Strumigenys (Trichoscapa) eschericht subsp. cognaia var. obscuriventris Santschi, 1914, Boll. Lab. Zool. Portici, 8: 375, worker. Preoccupied name, nec Wheeler 1908. Strumigenys (Trichoscapa) escherichi subsp. cognata var. fusciventris Santschi, 1915, Ann. Soc. Ent. France, 84: 261; nom. pro obscuriwentris. NEW SYNONYMY. Strumigenys (Trichoscapa) alluaudi subsp. nigeriensis Santschi, 1914, Boll. Lab. Zool. Portici, 8: 376, worker. NEW SYNONYMY. Strumigenys (Cephalorys) raymondi Donisthorpe, part. (see allwaudz). Worker. HL 0.52-0.60 mm., CI 75-83, MI 35-42. Mandibular denticulation of Group B, extremely fine and regular, the basal denticles not or virtually imperceptibly enlarged. Dr. Arnold has pointed out (in litt.) that the posterolateral mesonotal angles are better developed than in alluaudi and that the postpetiole is much less strongly convex and smaller. Otherwise, except for the striking den- titional difference and the completely smooth and shining sides of the alitrunk in s?moni, the two species are very similar (see alluaudi, above). In simoni, as in alluaudi, the pronotum is more or less shining, with a median carinula and weak, well-spaced oblique costulation on each side of the dorsum. The dorsal surface of the petiole is punctulate, but that of the postpetiole is completely smooth and shining when clean. Spongiform appendages rather well-developed on the petiole, in- cluding the midventral strip, and on the postpetiole. The propodeal teeth are the largest in the genus, and like those of the other species, are sharply upturned. The female is conspicuously larger and darker than the worker, but otherwise similar. Male not seen. S. stmoni is very widely distributed south of the Sahara. It ranges from Makapan in the Transvaal and Angola to Eritrea and French Guinea. Dr. Bequaert, who is quite familiar with African vegetation, has checked the localities for this ant very carefully with me. From the data, it appears that s¢moni avoids the regions of true rain- and gallery-forest quite consistently, although crossing the political boundaries of the Belgian Congo at several places. This form is at home in the more open forest and savannah covering such a large part of the continent. Colonies are found in rotten wood, under stones, and in the soil cover. This is perhaps the most familiar, and certainly the most-named, species of Serrastruma. 84 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY The extensive synonymy is truly surprising, for this species shows little variation compared to lujae, serrula or alluaudi, even though its range is wider. The confusion began with Emery’s original description, which states unequivocally that the basal gastric costulae are lacking. Later authors all utilized this as a difference when describing forms which were obviously close. In reply to my specific request, Dr. Delfa Guiglia has kindly examined the type material in the Emery Col- lection. She states that the gastric dorsum is basally costulate, and her accompanying sketch shows quite clearly the costulation, which differs in no way from that of the forms next considered. With due allowances for Emery’s draughtsmanship, which often erred in small details, the original characterization fits, and fits only the specimens before me. I have examined type specimens of all the species mentioned in the synonymy except biconvexa, boerorum and limbata. Of the last two species, I have seen specimens, presumably authentic because de- termined by Dr. Arnold and collected at or near the type localities. The limbata specimen may be from the original series. Dr. Arnold states in litt. that he has been able to find no differences between limbata and boerorum, and he has had material from Forel and Santschi. My material is outstanding in its uniformity, and cannot be dis- tinguished from simont by any satisfactory character. The descriptions of Forel and Santschi frequently differ in important details from the corresponding types, and one cannot but wonder at the thoroughness of observation leading to such publication. In 1910 and 1913 Forel and Santschi crossed in description, but neither attempted an investi- gation they must have known could scarcely have avoided unearthing synonymy; the result was some minor juggling of subspecific names and further confusion. In each case, both authors apparently com- pared their specimens against the previous descriptions, and notagainst authentic specimens. But the descriptions, which they themselves had so largely written, misled them still further. Santschi’s description of the homonym-synonym obscuriventris on the same page with the synonym nigeriensis remains inexplicable, especially since the types of the forms are as alike as two ants can be. These were collected by Silvestri at the same Nigerian locality (types sent by Sig. Consani). In lieu of a continued detailed statement of the reasons for the new synonymy, [| can only point to the lack of reliable differences among the types seen by me, and which have been directly and painstakingly compared in all directions. The species biconvera Santschi, of which I have seen no authentic material, appears to be based on small, dark specimens of s¢mont. If differences other than these questionable ones occur in the types, Santschi has not mentioned them in his publications BROWN: REVISION OF ANT GENUS SERRASTRUMA 85 or proved them by his figures. In the types of the synonyms and the abundant additional material I have seen, the chief variation, and that slight, is in relative length of the mandibles and in depth of color; no significant geographical apportionment of these characters can be de- tected. The coloration ranges from medium to deep ferruginous. Type locality. Makapan, Transvaal (Simon). Localities for other material examined. Southern Rhodesia: Victoria Falls. Bulawayo (G. Arnold). Zululand: Sordwana. Richards Bay; specimens from 3 colonies, including females (J. C. Faure). Belgian Congo: Elisabethville (J. C. Bequaert); 4 cotypes of var. cliens. Stanleyville (Kohl). Nigeria: Olokomeji (F. Silvestri) ; specimens from type series of nigeriensis and obscuriventris ex Consani Collection. Angola: Cucala, Benguela (J. Cruchet); cotype, Musée du Congo Belge, of cognata. Eritrea: Ghinda (K. Escherich); cotypes of esche- richt, Forel Collection and American Museum of Natural History. French Guinea: Kindia (F. Silvestri); det. cognata by Santschi. Mauritius: Several series (raymondi types in part.), British Museum. SPECIES INQUIRENDAE SERRASTRUMA LUDOVICI (Forel) new combination Strumigenys ludovici Forel, 1904, Ann. Mus. Acad. Sci. S. Petersbourg, 8: 369, worker, original description. This species has been considered by all previous authors to belong to Strumigenys sensu stricto. A camera lucida sketch graciously pre- pared for me by Dr. Ferriére from the type (presently designated lectotype) in the Forel Collection shows conclusively that ludovici is a member of the alluaudi group of Serrastrwma, for the mandibles are nearly or quite half the length of the head and have the basal five or six denticulae much coarser than the succeeding ones. In fact, ludovici can scarcely be differentiated from allwaudi, and may well prove to be a senior synonym of the latter when types can be directly compared. Type locality. Southern Madagascar (Sikora). SERRASTRUMA CALYPSO (Santschi) new combination Strumigenys (Cephaloxys) calypso Santschi, 1923, Rev. Zool. Afr., 12: 288-289, fig. 4a, worker, original description. From Santschi’s description and figure, this species seems virtually indistinguishable from Weber’s lott?. An examination of the mandib- 86 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY ular armature is needed to show whether or not the two are syn- onymous; in the latter case, the name calypso would take precedence. The type material of calypso is presumably in the Basle Museum. Dr. E. Handschin has written that he cannot lend material from the Santschi Collection because of catalogue difficulties. Type locality. Ouha, British East Africa (Meyer). Key to the sufficiently well known species of Serrastruma, based on the worker caste 1. Basal 5 or 6 denticulae of inner mandibular margin distinctly and suddenly larger, coarser, more acute and more irregular than those succeeding Gistally. sie ME cians Sinise sh Hates, Pn ens AEE one ee 2 Basal 5 or 6 denticulae of inner mandibular margin not or only slightly and gradually enlarged, quite regular and with blunt apices (the first, basal- mostidenticle mayaberenlanged)) ramen 0) casei ca eer eee eee 3 . HL 0.55 mm. or more; body robust; ground pilosity of head variable, but usually more or less broadly spatulate; postpetiole swollen. and broadly CONVENE at ea Pate a teatie rear PaO ae tone ct nen am alluaudi (Santschi) HL less than 0.55 mm., usually less than 0.50 mm.; body slender; ground pilosity of head inconspicuous, narrowly spatulate; postpetiole small and NAPTO Wy ICON MEX el TMU MUSEO a Vee, A cashec ee Les, ean ce aD lottt (Weber) 3. Propodeal teeth reduced to obtuse or rectangular vestiges............... bequaerti (Santschi) Propodeal teeth large or small, but always strongly acute.............. 4 4. Pronotum with oblique or longitudinal costulate or striate sculpture, its punctulationvabsentior secondary. > 4-5... 2) aa ie eee eee 5 Pronotum densely reticulate-punctulate like the mesonotum; costulation or striation absent or feeble and secondary to the punctulation.......... 6 5. Costulation of pronotum fine, sharp, and close, longitudinal in direction (appearing like dense striation); ventral petiolar appendage vestigial, scarcely spongiform; cephalic ground pilosity broadly spatulate or spoon- shaped, very conspicuous, but short................... maynet (Forel) Costulation of pronotum loose and indefinite, largely oblique, with broad, usually shining interspaces; ventral appendage of petiole developed and spongiform; cephalic ground pilosity narrowly spatulate, not conspicuous simoni (Emery) 6. Larger form (HL 0.52-0.71 mm.) with narrower head (CI 77-83) and rela- tively longer mandibles (MI 37-42); funicular segments II and III always lomperiGhamibroadica teu inet de sa pAunacte en crete vx gr Hace Rare lujae (Forel) Smaller form (HL 0.45-0.52 mm.) with broader head (CI 81-89) and rela- tively shorter mandibles (MI 34-39); segments II and III of funiculus not or just barely longer than broadio... 9.440.422: serrula (Santschi) bo Refer to species inquirendae section above for discussion of ludowecr (Forel) and calypso (Santschi). Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vou. 107, No. 3 & THE CARABID BEETLES OF NEW GUINEA PART 2. THE AGONINI By P. J. DARLINGTON, JR. With Four Puates CAMBRIDGE, MASS., U. S. A. PRINTED FOR THE MUSEUM August, 1952 PUBLICATIONS ISSUED BY OR IN CONNECTION WITH THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE BULLETIN (octavo) 1863 — The current volume is Vol. 107. BREVIORA (octavo) 1952 — No. 4 is current. Memorrs (quarto) 1864-1938 — Publication was terminated with Vol. 55. JOHNSONIA (quarto) 1941 — A publication of the Department of Mollusks. Vol. 2, no. 30 is current. OccASIONAL PAPERS OF THE DEPARTMENT OF MoLuusks (octavo) 1945 — Vol. 1, no. 16 is current. PROCEEDINGS OF THE NEw ENGLAND ZOOLOGICAL CLUB (octavo) 1899- 1948 — Published in connection with the Museum. Publication terminated with Vol. 24. These publications issued at irregular intervals in numbers which may be purchased separately. Prices and lists may be obtained on application to the Director of the Museum of Comparative Zodlogy, Cambridge 38, Massachusetts. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vo. 107, No. 3 THE CARABID BEETLES OF NEW GUINEA PART 2. THE AGONINI By P. J. DARLINGTON, JR. With Four Piates CAMBRIDGE, MASS., U. S. A. PRINTED HOR THE MUSH UM August, 1952 2S Feng’ iy ie . No. 3. — The Carabid Beetles of New Guinea! Part 2. The Agonini*® By P. J. DaRLineTon, JR. CONTENTS PaaE Maat ra clUnc GLO TIR eett ever oain tren tna otal ott taal mon cinco ecvem eles gasbouanie cs marl citer aie 90 Purpose; sources and disposition of material; acknowledgements...... 90 TBO CaliGiesee eters fats Aous rain. coseia chien. Wak cen las meg eae eee clay at ee ReRN LON oc 91 Methods smeasuremrents euCay kia os etee Aetecse: eect autem eilere tease 92 Structuressandicharactensins sso eens ee ey eer ere rie Ciclrel 94 dlreplnriley Ac OMimihse tile ints ees cts aes Sieve eas oe etary Mi ede nese tar eles 94 BOI, Cole, WMG ROTM NIK, .bodubnodhadorsqennsosscoceeobeoacs 95 TRUCE lp vs ca te ee Seat ty Me On eR ae aa tes a Wtel CR SPREE cis een mae 95 Brot nora xi tree reenel pe ols ctiale Late ene SR Coan eR Rent. Fe NEEDLES sch 97 LO RARE ete BG oid eee aS taeer einem eEN DS aa Fae th en eaten. aia en era ale 98 MMC Ra WAT TS We mre. cea cpa hd seer toa Oe ao emery 100 WO WereSlLta CO users tate cede nits, orca aerate eetee et aan Mani gat 100 LEASES) ies GRU I cea Saeed eet ne Pe EE RA orc eee mh seie eer Beara eentoune 101 Secondanyasextalechanaccersi rear teat retiree since eee react acl 103 Males. con nlaiornin-oraninties woite belgian @omee ones com Selo Selva coin ce is uc 103 Genera; nature of the New Guinean agonine fauna................. 105 HOUT Me Meee Ae ee tae a peek eck Stses al ene aster at cotey aaa ec PAT ew anes o ite¥ 107 Loss of wings and setae; wing-and-seta formulae................. 107 Reducihonkolaeyesstarsalllobeseis 6 seca: cee as rae sere ner 108 Adaptive nature of evolutionary trends........-............:... 109 Parallelismeandiconvercencesenn sameeren here ere tee 110 Rolewmecorraphicallisolatlomy yess ci athe ny et eileen 111 VAG ENCES OLSMAULATIOMN Ns paysite tlt ead sate ical nea eunuch eut yen ara 112 PRAKOMOMMICIBECIOM: ciara Naas hala ii soa ee eae alc wind atueee oe CN MON 113 “GRreiloxeycie easy ob bah tal witeete arent an au tet sare teed We) CEN cy Pes Baty CoS GE Ieee cae 113 Keyato cenera of Agonimiton News Guinea epee oe ae else 114 PAEUEUILUSR RR eer ae A ies a SM es Ae NU Es Stk ap MSHA gt ord 116 RORSAGOTULIUP EN te nC MER MR NICs Ue MeL hol esau ot lar. sit eee 120 FETT DIC TIC STR meat Mire EME My ae Rasy ee ner meee ae Feo ates ken lelees 122 DUCT ANOMCUS Hah = SEs UR REN ETS EASA ae eo enna sens CEN eel gloat BNR ek 124 US GTOSLETIUITIG Gs saa ieN ee ieee PSN ass one Ba ee A SUG = DN TRA Mees See es hae nee Eas 125 INI OSLO (OITA ICN hoe OR See OnE IS oars CNTR CIS NM IE os ree Ee Ee PE 127 Col pods Raptr tty. cicen Nea eet eh ate mie tfe: «cc bey hhhs GER CAML SFL eat 158 SEU CGH DOD GOP e aA S Ole one Onda Rar eee Oc Oe eae pide eek tree nete vere 173 ALTE ee GSS ele eee a et Eero ete dee Binet ey oe 176 J (RIGHT DLAI Cem eee Shes Dies GIS SG ol eh Ss cachchoN OA Moh Oo NED ene eeU ce eoes Slay kee eo 181 ANGST OULED dice ott enSca le ahd ea ereco ehOU Sockae BNC Gua ROLE Darien e oo aE ACO aon ah REC REEC 185 1 Work aided by grant of a fellowship by the John Simon Guggenheim Memorial Foundation, 1947-48; see page 91. 2 The first part of this series is not yet published. See text for explanation. 3 Manuscript received for publication December 20, 1951. 90 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY WE crulagonaarn «<4, Weis niwis'« MS HE ae POs alee RTS) noe 213 POLAMAGONUIM 6.6 cise we wenn dein oe) dated Srelol Sere tear le ge eae enna 221 GASHAGONUIMN qezpeN i. Uayeyyrsuly # Buaqjayes ee “dx SW MOUS aly = "pul spue}4a4 JIN . ae ‘ST fyyeaumply Xow De Pi “en os < OREN ¥ y =) 8 PETS eb x & of GSS oy RS ‘S] oableM 94 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY These ratios must be determined by measurements; they cannot be estimated satisfactorily by eye. It should be remembered, too, that the ratios vary somewhat in any given species and slight variations should not be given too much importance. In each case the specimens on which detailed measurements have been made and of which the proportions are given in the description are indicated in a special paragraph headed “Measured specimens’. However, my statements of measurements of total length and width (see next paragraph) show the extremes of the entire available series of each species. The measurements of total length represent specimens in fully ex- tended natural position and include the elytral denticles or spines if present. In the case of specimens not actually fully extended I have measured separately, with a micro-ruler in the microscope, the elytra (from above, with the anterior part of the elytral disc level), the prothorax, and the head, and have added the fractional measurements together. The resulting total lengths are more accurate and more uniform and therefore more significant than those given by most earlier writers on Carabidae. Measurements of width have been made across the closed elytra at widest point. When, as is often the case, the elytra are slightly separated, the width of separation has been subtracted from the measurement. When the elytra are widely separated or warped, width has been given only approximately. Some further notes on methods are included in the following paragraphs. Structures and characters: the tribe Agonini. The tribe Agonini (Platynini or Anchomenini) is a well-recognized group, although different authors have set different limits to it. The New Guinean forms of the group are diverse, but all of them seem to be true members of the tribe in a fairly strict sense. They are distinguished from Pterostichini primarily by having the outer elytral margin not inter- rupted by an inner, subapical elytral plica. This interruption is absent in all Agonini, so far as I know!. However, though present and very distinct in most Pterostichini, the interruption is indistinct or absent in at least a few of those of the Australian Region. In doubtful cases the form of the parameres of the male copulatory organs js useful in separating the tribes. In most or perhaps all Pterostichini the left and right parameres are very unequal in size and shape; in all genera of Agonini that occur in New Guinea the parameres, though by no means equal, are much less differentiated (see Figs. 20-66). 1 Andrewes (Trans. Ent. Soc. London, Vol. 78, 1930, p. 40; and Jour. Federated Malay Museums Vol. 16, 1931, p. 451) mentions the presence of an internal elytral plica in Aparupa and Idiastes, both of which I consider true Agonini, but in these cases the plica does not interrupt the elytral margin as it almost always does in Pterostichini. DARLINGTON: CARABID BEETLES OF NEW GUINEA 95 The following discussion is not a definition of the tribe Agonini but an analysis of some structures and characters of the New Guinean members of the tribe. The generic and specific descriptions of this paper will be modeled on the order of this discussion. Characters normal for a given genus will usually not be repeated under the species. Form, color, microsculpture. The Agonini of New Guinea vary in size from 4.0 (Arhytinus granum n. sp.) to 23.0 millimeters (Colpodes rex n. sp.). They vary in form from Bembidion-, Agonum-, or Platynus- like to broadly oval or subquadrate, and in convexity from strongly depressed to very convex or with inflated elytra. A strikingly fusiform shape has been evolved apparently independently in certain (but not all) species of several different genera (Maculagonum, Iridagonum, Altagonum, and Fortagonum), and a subfusiform or Amara-like shape has been evolved in several additional stocks. The color is usually black or brown; less often at least partly, especially on elytra, purple, blue, or green (Euplenes, several Colpodes, Altagonum cheesmani and regiscapha, and Fortagonum limum); and rarely the elytra are blue or green with yellow blotches (Huplenes lactus) or red tips (Euplenes apicalis and Colpodes laetus). In Maculagonum the elytra, though not metallic, are mottled or blotched with dark and pale. A majority of the species of New Guinean agonines are not iridescent, but a good many scattered through a number of different genera are faintly so at least on elytra in strong light, and a few (especially the species of Iridagonum) are strongly iridescent. The upper surface varies in sculpture but is rarely coarsely or extensively punctate. Reticulate microsculpture is almost always present (absent or nearly so only in Lathagonum and in Gastragonum laevisculptum) but varies in distinctness and in size, shape, and depth of meshes. In the descriptions the micro- sculpture is said to be normal when the meshes are visible in good light at a magnification of 54, and when they are isodiametric on head, moderately transverse on pronotum, and equally or more transverse on elytra. Many departures from this normal pattern are described under different species below. Generally speaking, so far as New Guinean Agonini go, characters derived from the microsculpture are of no more than specific and sometimes of less value. In several cases distinct differences in microsculpture separate geographical subspecies which are otherwise indistinguishable or nearly so. The head in different New Guinean Agonini varies from rather short to more or less elongate, and from slightly less than half as wide to fully as wide as the prothorax. The mandibles are only moderately long and rather strongly curved except in certain species of Fortagonum (forceps and cychriceps) in which they are much longer, more slender, and straighter than usual. The eyes are often large and prominent, 96 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY but also often reduced in both size and prominence, and rarely (Laeva- gonum citum, Fortagonum cychriceps) so small and flat as scarcely to break the outline of the sides of the head. In several independent cases the eyes, though much reduced in size, are still more or less abruptly prominent, sometimes abnormally and strikingly so. This is the case in one species of Notagonum (reversior), one of Iridagonum (subfusum), one of Maculagonum (setipox), one of Gastragonum (laevi- sculptum), all four species of [diagonum, some Nebriagonum, the single known species of Montagonum, and some Fortagonum, especially Forta- gonum limum, in which the eyes are not only abnormally abruptly prominent but also set off above from the front of the head by deep channels. More or less similar, small but very abrupt ‘‘popped”’ eyes occur in some other Oriental Agonini, including the genotype of Colpodes (brunneus Macl.), some other species of Colpodes (aeneipennis Dej., sjostedti Andr., and latus Louwerens), and all three species of the Himalayan genus Aparupa Andr. The eyes are not equally abrupt in all of the New Guinean forms listed above, but in each case they are much more so than in closely related forms. This peculiar modification of the eyes may be an adaptation, but it seems more likely that it is merely a secondary mutational effect which occasionally, but not usually, accompanies reduction of eyes, and which is obviously not necessarily of much phylogenetic significance. The usual two supra- ocular setae are present above each eye in most New Guinean Agonini. The posterior pair of the setae is about between the posterior edges of the eyes when the latter are normally large and prominent, but more or less behind the level of the posterior edges of the eyes when the latter are much reduced. This is a useful taxonomic character in some cases. The anterior supraocular setae have been lost in three probably independent cases: in one species of Iridagonum (subfusum), one of Nebriagonum (percephalum), and the entire genus Fortagonuwm; and in Fortagonum bufo the posterior as well as the anterior supraocular setae have been lost. It should be added that, in the case of Nebriagonum percephalum, although the anterior supraocular setae are absent in eleven specimens, the right anterior seta (but not the left one) is present in the twelfth specimen. Of course in the case of these and other fixed setae the setae themselves are often broken off, but their position is shown by strong punctures. The punctures as well as the setae are absent in the cases listed above. The antennae of different New Guinean Agonini vary in length and thickness of segments, but are always more or less normal in structure, with dense pubescence beginning near the base of the fourth segment. I have not taken taxonomic characters from the antennae and have usually not men- tioned them in the descriptions. The neck in some cases is and in DARLINGTON: CARABID BEETLES OF NEW GUINEA 97 others is not impressed above. The front is usually more or less evenly convex with a pair of variable but usually slight anterior impressions. I have called this condition normal; variations from it are noted in the descriptions. The mentum is usually toothed, but the tooth is absent in Arhytinus and is either absent or broken off in one species of Colpodes (sinuicauda, represented by two specimens, both of which lack a mentum tooth). When present, the mentum tooth is triangular, varying in exact form in different species, with the apex pointed, rounded, more or less truncate, or slightly emarginate. Some of this variation occurs within single species as well as between species. In general I have not found the form of the mentum tooth useful in taxonomy of New Guinean Agonini, and I have usually not mentioned it in the descriptions. This is true also of the other mouth parts. In general I have not found them useful in the present study, and I have therefore omitted them from the descriptions. The prothorax in different New Guinean Agonini varies greatly in form, proportions, and other details. No general discussion of the details is necessary except in the case of the lateral pronotal setae. There are normally two setae on each side of the pronotum, on or near the lateral margin, one (the anterior-lateral seta) near or before the middle of the prothoracic length, the other (the posterior-lateral seta) at or near the base. Both pairs (anterior and posterior) of these setae are uniformly present in the first eight genera (through Plicagonwm) here treated, in all the species of Gastragonum and Idiagonum, and also in Maculagonum setipox and most specimens of Nebriagonum cephalum, although all the other species of the two last-named genera have lost one or both pairs of the setae. One or both pairs are absent also in all New Guinean Agonini not named above. A few have lost the posterior- lateral setae but not the anterior-lateral ones (Lithagonum, some speci- mens of Nebriagonum cephalum, and both known specimens of Laeva- gonum subcitum). More often the anterior but not the posterior pair has been lost ([ridagonum, most Altagonum, most Maculagonum, Montagonum). And in the remaining cases both pairs of the setae are lost (Altagonum nudicolle and fatuum, Potamagonum, most Nebri- agonum, most Laevagonum, all Fortagonum). The preceding lists sug- gest what is certainly the case, that presence or absence of one or both pairs of lateral pronotal setae is not necessarily of itself an im- portant taxonomic character. In fact in Nebriagonwm cephalum the posterior-lateral setae are present or absent, or sometimes present on one side only, in different specimens of the type series. Nevertheless, in many other cases presence or absence of these setae does give a useful “tag” for the identification of species or genera. These setae, like the supraocular ones, are often broken off, but then their former 98 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY presence is shown by characteristic punctures which are absent when the setae have failed to develop. A formula for indicating presence or absence of the standard supraocular and lateral pronotal setae is de- scribed below, under the heading ‘Evolution’. The disc of the pro- notum is described as normal when it is moderately convex and has a more or less impressed, more or less abbreviated, fine, median longi- tudinal line, and two less sharply defined but distinct transverse im- pressions near the apex and base respectively. The actual apex and base sometimes are and sometimes are not margined by fine impressed lines. The elytra of different New Guinean Agonini differ greatly in form and proportions, and also in many details which yield good taxonomic characters. The elytral disc is in some cases regularly convex and in others more or less impressed before the middle. This character is surprisingly constant in some species and is useful especially in super- ficial recognition of certain Notagonum. The bases of the elytra are usually entirely margined to or nearly to the scutellum, but the margin is incomplete inwardly in a few cases (Tarsagonum, Colpodes acuti- cauda, Idiagonum, Fortagonum cychriceps). At the humeri the marginal line is in some cases rounded and in others angulate, the angles being obtuse, right, or acute in different cases; generally speaking the angles are rounded or obtuse in species with normally formed elytra, more nearly right or acute in those with elytra oval or the whole body oval. The lateral margins or gutters of the elytra vary in width in different species. The lateral margins posteriorly, just before the subapical sinuations, are usually rounded, but in a few cases are abruptly angu- late or even with very short spines (Notagonum externum, Colpodes saphyrinus sloanei, Colpodes antedens). The subapical sinuations are usually moderate or strong but in some cases slight or absent. The form of the subapical sinuations, or their absence, is usually a specific rather than a generic character, although it is more or less constant in some small genera. The apices themselves vary extremely in form (conjointly or independently rounded, subtruncate, angulate, or slightly produced), and are dentate or spined in a number of taxo- nomically diverse species. Well developed spines occur in Tarsagonum latipes; Notagonum subrufum and spinulum; Colpodes violaceus, saphy- rinus sloanei, and antedens; Altagonum tutum, cheesmani, and some specimens of scapha; and shorter spines occur in some other species, including some specimens of Potamagonum diaphanum and Nebri- agonum arboreum. Spines seem to have been evolved independently in each of the species named. The apical elytral spines vary in position in different species; they may be opposite the ends of the fourth, third, second, or sutural intervals. The sutural angles of the elytra are only DARLINGTON: CARABID BEETLES OF NEW GUINEA 99 rarely actually spined but they are very often denticulate, the denticles being in some cases prominent and in others inconspicuous. Sometimes the denticles vary in development or are even present or absent in single populations. Also, they sometimes vary with angle of view. When the denticles are only slightly developed they are likely to ap- pear distinct if seen from in front of the perpendicular, indistinct if seen from farther back. Among the Agonini of New Guinea the presence or absence of these denticles is often useful to distinguish species, but rarely genera. The character is obviously one which should be used with discretion. The striation of the elytra is entire or nearly so, and the sutural striae are long, in all New Guinean Agonini. The striae are usually impunctate or nearly so, rarely coarsely punctate. In some cases the edges of the striae are slightly irregular although no distinct punctation can be distinguished. In these cases I have described the striae as ‘‘not distinctly punctate” or “indistinctly punctulate’”’. The striae vary in depth, and the intervals, therefore, in convexity. The outer intervals, especially the eighth and ninth and sometimes also the seventh, are in some cases variously modified toward apex. In most cases they end or fuse without note- worthy modification other than a moderate narrowing of the eighth and widening of the ninth interval toward apex. In these cases I have described these intervals as ‘not much modified toward apex”. In other cases their form is described in detail. The eighth and to a less extent the seventh intervals are sometimes much compressed toward apex (Colpodes bennigseni), or longitudinally impressed or sulcate (Tarsagonum latipes, Colpodes acuticauda, all species of Iridagonum except quadripunctellum, Altagonum pallinox, A. sphodrum ete., and some specimens of Fortagonum fortellum). The ninth or normal sub- marginal interval, which is usually somewhat widened (and more or less interrupted) toward the apex, is sometimes abnormally narrow and convex, or longitudinally impressed. In certain cases the marginal channel itself is more or less modified especially posteriorly, being sometimes abnormally wide and flat (Fortagonum cychriceps) or with its inner part swollen and forming posteriorly a more or less distinct extra or tenth interval. Such a partial or fragmentary tenth interval is characteristic of the entire genus Idiagonum and is more or less developed in several species of Fortagonum. Fortagonum fortellum is remarkable for the variation of its outer elytral intervals, which are sometimes not much modified toward apex, sometimes strongly longi- tudinally impressed; and in this species a short but well defined fragment of an extra tenth interval is present in some specimens but not in others (these variations are all in one series from a single small strip of mountain forest). The third elytral interval normally has 100 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY three dorsal punctures, each bearing a rather inconspicuous seta. The punctures are usually more or less evenly spaced along the length of the elytron, about 14 from base, near middle, and about 14 from apex; the first is usually near the third stria, the others, near the second one. When the punctures are arranged in about this way, the third interval is said in the following descriptions to be normally 3-punctate. Many departures from the normal condition, involving differences in position of, or loss of one, two, or all of the punctures, are noted in the de- scriptions of different species. In some cases loss of one or more punctures is a constant character useful in defining species. For example in the first species-group of Altagonum and in Iridagonum some species have the third elytral interval always 3-punctate, others, always 2-punctate, with the anterior puncture missing. In other cases, however, the occurrence of punctures on the third interval is extremely variable in single species (see descriptions of Nebriagonum cephalum and Fortagonum fortellum). The inner wings vary in New Guinean Agonini, being in some cases fully developed, in others vestigial. I am well aware that such vari- ation is often of little importance among Carabidae. However it so happens that in New Guinea there is only one agonine genus in which the wings are variable. This is Gastragonum, which includes species which, so far as my material goes, are fully winged, others which are vestigially winged, and one (terrestre) which is dimorphic. In every other genus here treated the wings are either uniformly full or uni- formly vestigial in all my New Guinean material. Under these circumstances the state of the wings becomes a very useful taxonomic character. When the wings are vestigial, the metepisterna are often more or less shortened. This character has been overstressed in the past, especially in classifications of Colpodes. I shall use it very little here. The lower surface is usually impunctate or virtually so in New Guinean Agonini, except of course for the presence of certain “‘fixed”’ setigerous punctures, but in a few cases the lower surface is more or less extensively punctate especially at the sides. The abdomen usually lacks pubescence other than fixed setae, but in about ten separate cases scattered pubescence is present either localized on some part of the abdomen (usually near middle of base) or over much or all of it. The cases are: (1) Arhytinus, in which scattered pubescence seems to be confined to the last ventral segment of the female only; (2) Nota- gonum angustellum, subnigrum, and vile, in which the abdomen is extensively pubescent, though not equally so in all the species named; (3) Notagonum externum, with a very little fine pubescence near middle of abdomen; (4) Notagonwm sinuum and vaporum, in which abdominal DARLINGTON: CARABID BEETLES OF NEW GUINEA 101 pubescence is again extensive; (5) Notagonum subimpressum, in which the pubescence is slight and mostly near the middle; (6) Colpodes acuticauda, with a little fine and irregular pubescence; (7) Lithagonum, with abdomen rather variably and sparsely pubescent; (8) Altagonwm pubinox, pallinox, noctellum, and planinox; (9) Altagonum sphodrum and postsuleatum; (10) Gastragonum laevisculptum. 1 think that pubescence has appeared on the abdomen independently in most or all of these ten cases. The prosternal process in New Guinean Agonini is as a rule simple: not margined nor tuberculate at tip, not setose, and with the posterior declivity not strongly compressed, though sometimes moderately so. Exceptions to this rule are noted in the descriptions. The principal exceptions are that the tip of the prosternal process is margined or tuberculate in Tarsagonum latipes and Fortagonum bufo, and setose in all four species of Idiagonum. The mesepimera are usually very narrowly triangular, but are somewhat wider, with outer margin about one-half as long as the anterior one, in Euplenes. This is corre- lated with and probably a result of evolution of a rather broad and depressed body-form in Euplenes. The metepisterna vary with the state of the wings, and have already been mentioned in that connection. The legs are more or less normally formed in all New Guinean Agonini but vary in many details, of which I shall mention only those found useful in taxonomy in the present paper. It is likely that many other details, including the presence or absence of certain setae on the femora and the clothing of the lower surface of the tarsi, may eventu- ally prove of great taxonomic value, but they may perhaps be more profitably studied in some connection other than the present one, preferably in the course of a world-wide classification of agonine genera. The following discussion is, for practical reasons, limited to the hind tibiae and tarsi and applies to the Agonini of New Guinea only. The hind tibiae are deeply sulcate along their extreme outer edges in Tarsagonum only. The hind tarsi, though variable, are usually more or less slender, but are somewhat wider than usual and symmetrical in Euplenes, and wide and asymmetrical in Tarsagonum. The first three or four segements of the hind tarsi are usually more or less grooved on each side above. The presence or absence and the depth of the grooves can be used in taxonomy, but the characters are difficult both to see and to describe accurately, and I have therefore usually avoided them here. They have been much over-used in the past. The form of the fourth segment of the hind tarsi is variable and is im- portant in agonine taxonomy, though not so important as it has been thought to be. This segment may be simply emarginate at apex or with short or moderate or long apical lobes below (see Figs. 16-19). When lobes are present, the outer one is usually longer than the inner, 102 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY but the two are nearly equal in Euplenes. In observing the shape of the fourth hind-tarsal segment, great care should be taken not to mistake a middle tarsus for a hind one and to see both sides of the fourth segment clearly so as not to overlook a lobe that may be hidden below the base of the fifth segment. There is a definite correlation between the form of the fourth hind-tarsal segment and the habitats or habits of different agonines. The segment tends to be simply emarginate in ground-living species, lobed in arboreal species and in those that live beside running water, but there are many exceptions to this general rule. As might be expected in view of the apparently adaptive nature of its modifications, the fourth hind-tarsal segment is not really fundamental in agonine taxonomy. The shape of this segment does not, or at least not usually, define large and diverse genera such as Colpodes, although it is more or less constant in many smaller genera. In the New Guinean genera, the fourth hind-tarsal segment is emarginate in Arhytinus; strongly but very asymmetrically lobed in Tarsagonum; strongly and symmetrically lobed in Euplenes; asymmetrically lobed in Dicranoncus; emarginate in Lorostemma; variable (emarginate in several stocks and more or less lobed in several others) in Notagonwm; variable also in Colpodes, briefly lobed in Plicagonum, rather briefly lobed in Lithagonum; emarginate in Irida- gonum; usually simply emarginate in Altagonwm but with short lobes in four species which are probably not directly related to each other (caducum, cheesmant, scapha, nudicolle); emarginate in Maculagonum; briefly lobed in Potamagonum; variable in Gastragonum; emarginate in Idiagonum; variable in Nebriagonum; emarginate in Laevagonum; emarginate in Montagonum; and variable in Fortagonum. I give this list in full to emphasize the variability of this character so far as New Guinean agonines are concerned. The fifth segment of the hind tarsi sometimes does and sometimes does not have a row of conspicuous “accessory setae’ on each side of its lower edge. The presence or absence of these setae is an important taxonomic character, but it again is not so important as has been thought. The setae are absent or only rudimentary in most New Guinean agonines but are more or less obvious in a few: Euplenes, (Dicranoncus), two Colpodes which are probably not directly related to each other (s. sloanez, antedens), Potamagonum, and some Nebriagonum (see notes under this genus). Their somewhat erratic occurrence and the fact that they are not equally developed in the different forms listed show that presence of obvious accessory setae on the fifth hind-tarsal segment is not neces- sarily a character of full generic value. This conclusion is reinforced by another fact. Although only a few of the New Guinean species of the tribe have the accessory setae well or even moderately well de- DARLINGTON: CARABID BEETLES OF NEW GUINEA 103 veloped, very many of the species, perhaps even most or all of them, have the setae present in a rudimentary form, as very short, fine hairs barely visible in clean specimens at 54 with good light. I do not know whether these minute hairs are vestiges indicating the presence of longer setae in the ancestral stocks from which most or all existing New Guinean agonines have been derived, or whether they are parts of minute sensory organs or other structures normally present in Agonini, from which larger setae may be evolved. In either case the presence of minute hairs in so many forms lessens the significance of the occurrence of more or less enlarged setae in a few forms. The tarsal claws are simple in all the agonines now known from New Guinea, but each claw has an acute tooth below at base in Dicranoncus, of which one species may occur on the island. The sole of the first four hind-tarsal segments is variably clothed or margined with hairs or bristles in different Agonini, and the variations will probably be useful in taxonomy, but I have not attempted to use them in the present paper, except as an aid to distinguishing Lorostemma from Notagonum. The secondary sexual characters of New Guinean Agonini are in general those which are normal for the tribe in other parts of the world: the front tarsi are slightly dilated in the male, with the first three segments biseriately squammulose beneath; and the last ventral segment bears one seta on each side in the male, two on each side in the female. The modification of the front tarsi of the male occurs in all New Guinean Agonini so far as my material goes. However there are a few exceptions to the normal in occurrence of ventral setae: in Notagonum altum the female usually has only one seta on each side of the last ventral segment, like the male; in Colpodes rex, the male has one, the female usually three such setae on each side; and in Litha- gonum the male has usually two or more (not one) and the female usually four or more (not two) setae on each side. Two other, minor external sexual differences appear in single genera of New Guinean agonines. In Arhytinus the last ventral segment is glabrous in the male (except for the conspicuous setae mentioned above), sparsely and inconspicuously pubescent in the female. And in Maculagonum the last ventral segment is moderately or strongly emarginate at apex in the male, entire in the female. This character appears also in a less marked form in a few other species of New Guinean agonines, but in most cases the last ventral is entire or nearly so in both sexes. The male copulatory organs of some Agonini present good characters which define major groups within the tribe. In certain genera which do not occur in New Guinea (Calathus, Sphodrus, etc.) the right para- mere is distinctively long and slender, not relatively short and spatulate 104 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY as it isin Agonum (cf. Jeannel, Faune de France, Coléoptéres Carabiques, Part 2, 1942). However, I have found no such obvious group character among New Guinean agonines, all of which have copulatory organs of the general Agonwm type, with the right paramere more or less smaller than the left one, both being simply oval or spatulate. There are many small variations in the form of the apex of the middle lobe which, after adequate study of series to determine individual variation (which is sometimes considerable), might be used to define species, and there are also differences in the armament of the internal sac which, in different New Guinean agonines, may bear a conspicuous hook (Tarsa- gonum latipes), or a few or many small spines (Arhytinus major, Notagonum reversior, Colpodes antedens, and others), or thickened areas (Plicagonum rugiceps, ete.), or (most commonly) no conspicuous arma- ment but inconspicuous areas of minute bristles or granules. There are undoubtedly specific and generic characters to be found in these differences, although they are not always so important as they seem at first glance. However, I have not tried to use them here. To study them properly would require too much time in proportion to the probable results so far as the classification of the Agonini of New Guinea alone is concerned. Other characters are so numerous and so good among New Guinean agonines that the genitalic ones are hardly necessary. The basic genitalic characters of different groups of Agonini should be worked out as part of a study of the phylogeny and classi- fication of the tribe for the whole world. Some day I hope to make such a study. In the meantime, I do not know which of the many details of the agonine copulatory apparatus should be stressed in drawings and I do not know from what angle the drawings had best be made, so the labor of drawing every species now might be wasted. However, for the information of specialists, I have figured (Figs. 20-66) the male copulatory organs of at least one species of each genus treated, of genotypes of all new genera except Maculagonwm (of which another species is figured), of all new species of which only one male is known, and of a few other noteworthy species. It is to be hoped that specialists will not make a habit of describing new forms of Agonini based only on small genitalic differences and especially not on differences which appear between my figures and their specimens. The figures have been made carefully, with a camera lucida, but they are not perfect. Carabid copulatory organs are very difficult to figure definitively. Some details are entirely without sig- nificance. For example, the dorsal (convex) profile of the middle lobe varies with the position of the movable internal sac. Many other details seem different from different points of view. This is especially true of the parameres, which are often not flat but curved or warped DARLINGTON: CARABID BEETLES OF NEW GUINEA 105 so that their outlines vary with angle of view; and whether or not the basolateral forks of the parameres are visible depends on angle of view. The armament of the internal sac is even more difficult to show accu- rately, especially when (as is usually the case) the sac is not everted. Even when small genitalic differences do exist between specimens, their significance is often doubtful, for the male copulatory organs do vary in some species. Almost more than any other parts of the body, the male copulatory organs of Carabidae should be studied and their variations understood before they are used in taxonomy. After the two preceding paragraphs were written, I discovered that the position of the internal sac of the male copulatory organs varies with the way the specimens are killed. If they are killed dry, by fumes of acetic ether for example, the sac is usually fully retracted within the middle lobe. In this position, the sacs of different specimens can be compared in detail. If the insects are killed in alcohol, however, as my New Guinean ones were, the sac is often partly but not completely everted, and more so in some specimens than in others. Under these conditions it is difficult or impossible to compare the sacs of different specimens properly. This is a very serious disadvantage of alcohol- killed material. It could probably be overcome by dry-killing the insects and then preserving them in alcohol. Measurements of total length and of width, which are given at the beginning of the generic descriptions and at the end of the specific ones, have already been discussed above. Genera; nature of the New Guinean agonine fauna. So far as the Agonini of New Guinea are concerned, there are no characters which, of themselves, are always of generic value. The preceding discussion should have made this fact clear. The basic criterion which I have tried to use for genera is actual relationship, as shown by a sum of characters. In practice I have usually treated as genera groups of species which seem to be closely related among themselves but much less closely related to any other species and all of which share at least two distinctive characters, one of which may be a distinctive form or appearance. In some cases I have given weight to continuity of variation; that is, I have included in one genus species which are very unlike if the differences between them are bridged by a more or less continuous series of intermediate species. A good example of this is discussed in the notes under the genus Nebriagonum. In a few cases I have made monotypic genera for single species which are very strongly characterized. Use of these practical criteria has resulted in recognition of nineteen genera of Agonini in New Guinea, including one (Dicranoncus) which is not actually recorded from the island but probably occurs there. 106 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Four of the nineteen are previously known, small, natural genera which are primarily Oriental in distribution. These are Arhytinus, Euplenes, Dicranoncus, and Lorostemma. Three others are genera of convenience, each containing a number of species not necessarily all related to each other but not sufficiently characterized to be set apart in separate small genera now. One of these genera of convenience is the well known Colpodes, here used in a somewhat restricted sense. The other two I am calling Notagonwm (southern Agonwm) and Alta- gonum (alticoline Agonum) respectively. The remaining twelve genera are new, small, apparently natural groups all of which are probably confined to New Guinea (one or two reach also adjacent small islands) and most of which are confined to the mountains of the island. It is obvious that in general this fauna consists of three elements. First are several stocks which are shared with and in most cases probably derived from the Orient, including not only the four small genera first named above but also several stocks of Colpodes and possibly some of Notagonum. Then comes a mass of species (most of those of the composite genera Notagonum and Altagonwm and some Colpodes) which, though not strongly enough characterized to form separate small genera now, represent stocks which are more or less endemic to New Guinea and which are apparently in the early stages of differentiation and radiation there. Finally there are the twelve smaller, endemic genera, most of which are probably derived directly from the composite groups but which are more differentiated and some of which have radiated on New Guinea or even on single mountain ranges of the island. This whole picture is clearly one of accumulation of a rather limited number of stocks apparently partly (but perhaps not wholly) from the Oriental Region, of preliminary differentiation and adaptation to different lowland and mountain habitats, and finally of evolution and radiation of a considerable number of more distinct endemic groups which I have called genera. As to when the different stocks reached New Guinea, about all that can be said is that they have probably arrived at different times. Some probably came long ago (but not necessarily all at the same time) but whether they arrived in the Tertiary or before it can hardly be said in the absence of a fossil record. Some others have probably come more or less recently. Some of them are still only specifically or subspecifically different from Oriental forms, and in two cases (Colpodes laetus and C. bennigsent) I have found no significant differences between Oriental and New Guinean specimens. Two facts of general interest arise from this brief discussion of the New Guinean agonine fauna. The first is that the whole fauna of nineteen genera and one hundred and twenty-one species and sub- DAWLINGTON! CAHATID BEWTLNS OF MIUW GUENIA 1()7 species is derived from comparatively lew mucestors which live come, at least partly from the Onent, at different tines, some of then probably very recently, The other fact of general iiterest is that the mountain agoutnes — and Avon ive the most numerous Carabidne on the higher mountains of New Guinen — seem to lave evolved more or loss in place, wnd at least in most cases to have been derived fron stocks which stil occur wt lower altitudes in New Guines, Tn other words, the mountain agonine finns is in its orving an endemic one, not reliet one, Myolution: lows of wings and aca; wingeandeacta formula, Covtain definite evolutionary trends ave distinguishable among New Guineun Agonimn, especially one toward loss of wings and of certain setae especially inimountain environments, Dhiave discussed the adaptations of mountain Carabidae elaewhere (Meological Monographs, Vol, V4, 144, pp, 37-61), and have suggested Cp. 49 of the paper cited) the terms -- winged (+-w) Cplusewinged) and «winged (=w) Grins winged) for Carabidae with and without fully developed wings, mud ewinged (ew) (plus. or minusewinged) for dimorphiewly winged formes; and -’s and —’s con be used also to indicnte presence or absence of the most important standard setae and dorsal elytra punctures of Ayoninis. the two pate of supraccular setae, two pais of literal prothoracic setae, wid three setigerous punetures of the third elytra) interval, By arranging the sine in the order just given, formulas con be made up for the state of the wings and setae in different cases, Vor example in Notagowum ond other genera in which the wings wre always fully developed and all the standard setae and punctures me always present, the formula ia few, dete) teetey eet, In Attagowwm, in which the wings are always fully developed, both pairs of supraoculin setae always present, the anteriordlateral pronotal setae always aheent wid the posterior-lateral ones usually present but sometines absent, and the punctures of the third elytral interval uiiilly present but oll sometimes abwent, the formula is -w, tek, condition in the genus; the parentheses indicate that exceptions occur, The ultimate stage in lose of wings and setae is reached in one species A Vortagonum (ufo) in which the formula is - ooo Ge aeRO 5 . Much of abdomen including most of last segment pubescent; (small, slender, depressed, with very long antennae) (p. 133)...... angustellum Abdomen much less extensively pubescent.........-....--.2)52--= +8 4 . Depressed; black, legs blackish; lateral margins of prothorax narrower, not distinctly translucent (olS4)len) 2) ae eee subnigrum Less depressed; brownish-piceous, legs yellow or brown, rarely blackish; lateral margins of prothorax wider (but still rather narrow), more or less translucent: (p.185).. 2. ke eee ie ee vile . Eyes normal, with posterior supraocular setae about between their posterior margins; prothorax wider than long, not punctate across base (DS ABE) diss cosh crete Roe Aes CME aS ey ane reversum Eyes smaller but abruptly prominent, with posterior supraocular setae well behind line of their posterior margins; prothorax as long as wide (by measurement), extensively punctate across base (p. 137) . .reversior . Apices of elytra at least in part very finely subserrate or granulate, though sometimes only faintly so (piceous, Agonum-like, with 4th hind-tarsal segment simply emarginate) (inserted here because of similarity to some INORG CIOUTO)). (GOs PAD) 5.6 oo boo ooobedea se ob so (Lorostemma informalis) Apices of elytra not subserrate or granulate (form, appearance, and form of 4th hind-tarsal segment) variable). <2 45.224. 220% - see ae a 10. 11. 12. 13. 16. DARLINGTON: CARABID BEETLES OF NEW GUINEA WB . Lateral margin of elytron with an abrupt, sharply defined angle before XS SULOBONGAY UAE HAKONA (Gos MIB) ob s's Gos eoceoneyvuedoucceose externum Lateral margin of elytron not abruptly angulate before the subapical SITUA UL OI retest tsk Sey Ce Rar Wa Rte Se Oa aU 8 . Apex of each elytron more or less rounded, not distinctly denticulate, not AlomDjowhy HiewIENe, ion GOIN bsp boousbodusoobebcoovostoscuoneede 9 Apex of each elytron denticulate at suture, and/or abruptly angulate or Spinedeaboutroppositearduntenyalenii eee eerie cri: 14 . Abdomen extensively pubescent; 9th (submarginal) interval of elytron near apex rather narrow and strongly convex...................-- 10 Abdomen not pubescent, except for fixed setae; 9th interval of elytron MEADE By oes wwoloPianNel meray sikNhs Seo coscceLonscaceoeocscocooosuD. 11 Sides of prothorax strongly sinuate, with posterior angles right or nearly so; basal foveae of pronotum shallow, almost flat, much roughened (Gompli3 9D) Wate rice s eee s cement at Ecorse oa ad cial ane a a a sinuum Sides of prothorax only moderately sinuate, with posterior angles a little obtuse; basal foveae of pronotum rather deep, less roughened (p. 140) vaporum Small (4.8-6.5 mm.) lowland species; 4th hind-tarsal segment simply GUTTER OTE EY os cattle SoC aM eee eRe Eo anne es eae aitape (11a) Elytron with pale margin confined to lateral gutter; average size smiallensl(42S—o soya) a (9591451) poee eine nes ee eae (aitape s. s.) (11b) Elytron with pale margin including several outer intervals; average size larger (5.5-6:0 mm:) (p. 142).5...5...... (subsp. sansapor) Larger (7.1-8.6 mm.) mountain species; 4th hind-tarsal segment lobed. .12 Form very convex, the pronotum gibbous and with exceptionally narrow margins; (elytra lightly striate, inner striae often almost obliterated ATHLETE OL se (Os AD) Mier teat uaroyt cee a Ieee cesar Hovde aan eee gibbum Form less convex, the pronotum only normally convex and with margins Onlygnormal lysnannowiree weet ae eek ae ee nee ee 13 Elytra more than usually narrowed toward base, deeply striate; micro- sculpture of elytra coarser, the meshes distinct at 54X.......... altum (13a) Base of prothorax relatively wider (base/apex 1.15 to 1.22) (p. 144) (altum s. s.) (13b) Base of prothorax relatively narrower (base/apex 1.03 to 1.13) (Gos a ee ioe piece hls te 2 ih ep ae aD a (subsp. zbele) Elytra not unusually narrowed basally, lightly striate; microsculpture of elytra very fine, the meshes not visible at 54X (p. 148)........... sigt . Apex of each elytron with a denticle at sutural angle but with no other denticle, no well-defined angulation, no spine..................... 15 Apex of elytron with a second denticle, or with a well-defined angulation or spine about opposite end of 3rd interval....................... 18 . Subapical sinuation of elytral margin rather weak; (elytral striae usually distinctly punctulate) (see also Couplet 20) (p. 145)....... margaritum Subapical sinuation of elytral margin strong; (elytral striae not distinctly PUNCtUIALeRE Mr nee ee re ee eGR MINS eo PO Ee ee 16 Basal foveae of pronotum irregularly punctate; 4th hind-tarsal segment SUMP lyeMAh ena ve reeves er cles Arent ohn a Aaah etme subpunctum Wf 18. iLO: 20. 21. 24. BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY (16a) Smaller (6.5-7.4 mm.); elytral microsculpture coarser, the meshes very distinct at O420 (pAl4G)inaoe eerie: (subpunctum s. s.) (16b) Larger (7.7-8.5 mm.); elytral microsculpture finer, the meshes barely, visibleyatio4x@\(pal4 (ie 2 eee (subsp. capitis) Basal foveae of pronotum not or indistinctly punctate; 4th hind-tarsal segment lobedins sce ie cg Solway cee lk Oe a iliz/ Elytra not distinctly impressed before middle, their outer margins con- trastingly palewu a Minin Vale Ciaeacn fob ets cert kat iene en dentellum (17a) Smaller (7.3-8.5 mm.); eyes relatively larger and more prominent; etc (see description) i(ps 14a) ee eee ee (dentellum s. s.) (17b) Larger (8.4-9.7 mm.); eyes relatively smaller and less prominent; etc, (see description) (py 149) eee see eee (subsp. chimbu) Elytra broadly impressed across disc 14 or 26 from base, their outer margins not contrastingly pale (p. 149)................ subimpressum Apex of each elytron bi-denticulate or bi-angulate, the outer denticle or angle (about opposite end of 3rd interval) only about as prominent as thejinner (Suturall) ones sess eee cree eee ree 19 Apex of each elytron strongly angulate or spined about opposite 3rd interval, the angulation or spine much more prominent than the sutural denticles if amiy?é).). 0 acgk ee mya Hay Ses an a aie 22 Small (5.6-6.6 mm.); elytra not distinctly iridescent........... paludum (19a) Microsculpture of elytra coarser, the meshes very distinct at 54X (ore 10) We arene ene NEL Aes ae ert meee Ae unt ners ols Sale (paludum s. 8.) (19b) Microsculpture of elytra finer, the meshes barely or not visible at YD, @E Gov! Cot) Pane mage OM omN eH i wie Mee bS Weems 8 Gn (subsp. velum) Larger (7.0-9.0 mm.); elytra more or less iridescent................- 20 Legs yellowish; elytral striae usually punctulate; (see Couplet 15)....... margaritum in part Legs dark; elytral striae not distinctly punctulate................... 21 Slightly broader, with prothorax wider (slightly more than 14 wider than long) and head relatively narrower (.68 width prothorax), but with very prominent eyes; front of prothorax only normally emarginate (p. 151) malkint Slightly less broad, with prothorax narrower (slightly less than 14 wider than long) and head relatively wider (.71 to .74 width prothorax), though with less prominent eyes; front of prothorax more deeply emarginate (with anterior angles more advanced) than usual (p. 152)...... tridior Elytra strongly angulate at apex opposite ends of 8rd intervals, but not SPOUT C5 coset 55.5) ov eoke 15 eR Sa Laon aU Ste OAR er 23 Elytra spined at apex about opposite ends of 8rd intervals. .......... 25 . Broad, relatively small-headed (head/prothorax .66 & .69); lateral margins of prothorax scarcely translucent (p. 153)................. addendum More slender, relatively larger-headed (head/prothorax .77 to .80); lateral margins of prothorax contrastingly translucent................... 24 Prothorax narrower (width/length 1.30 in specimens measured) and with narrower base (base/apex 1.18 & 1.23 in specimens measured); elytra stronglypridescemt (pi V5) ie oe east awh cars ere angulum DARLINGTON: CARABID BEETLES OF NEW GUINEA 13333 — Prothorax wider (width/length 1.41 & 1.42) and with wider base (base/apex 1.33 & 1.36); elytra only faintly iridescent (p. 155)........ subangulum DH, LioRGl: lolevolke, NontenlejXNKOIE Glad codec euacnadsdduodsoebopegsoe 26 — Slender, rufescent, mountain species (p. 156)................. subrufum 26. Larger (7.3-8.4 mm.); elytral striae lightly impressed; 4th hind-tarsal Segment with very: short lobes (pom ein cis eae eee ee: spinulum — Smaller (6.8 mm.); elytral striae deeply impressed; 4th hind-tarsal segment without distinct lobes; (see also description) (p. 158)... .. subspinulum NoTAGONUM ANGUSTELLUM 0. sp. Description. With characters of genus as described above. Form of large, slender, flattened Bembidion; brownish-piceous above, scarcely paler below, appendages testaceous or brownish-testaceous, outer margins of prothorax and elytra only slightly translucent; surface moderately shining, not or faintly iridescent; microsculpture normal, light. Head .89 & .86 width prothorax; eyes large, prominent, with posterior supraocular setae just behind line of their posterior edges. Prothorax subcordate; width/length 1.34 & 1.40; base/apex 1.12 & 1.16; sides less arcuate than usual, sometimes faintly angulate at anterior marginal setae, rather broadly sinuate before basal angles; latter approximately right, well defined; lateral margins rather narrow; basal foveae flat, very shallow (but sometimes a little impressed), with surface somewhat irregular but not distinctly punctate; anterior marginal line faint or interrupted at middle, posterior one vague. Elytra rather narrow, subparallel, depressed; subapical sinuations moderate; apices more or less independently rounded, often vaguely angulate at suture, and rarely subdenticulate there; striae moderately deep, impunctate or faintly punctulate; intervals slightly convex, 8th narrowed and strongly convex apically, 9th wide apically, partly or completely interrupted by ocellate punctures, but remaining portions convex. Lower surface with at most a few vague punctures, except that abdomen, including nearly the whole of its last segment, has extensive pubescence rising from fine punctures. Legs: 4th hind-tarsal segment lobed, outer lobe longer than inner. Male copulatory organs: Fig. 25. Measurements: length 5.8-7.4; width 2.0-2.7 mm. Types. Holotype & (M.C.Z. No. 28,593), and 16 paratypes from Dobodura, Papua, Mar.—July 1944 (Darlington). Additional para- types as follows: Papua: 7, Milne Bay, Dec. 1943 (Darlington), and 1, same locality, Oct. 20, 1943 (W. B. Jones, Alabama Mus. Nat. Hist.). N-E. N. G.: 14, Nadzab, July 1944 (Darlington); 16, Chimbu Valley, Bismarck Range, 5,000-7,000 ft., Oct. 1944 (Darlington). Neth. N. G.: 33, vicinity of Hollandia (actually S. foothills of Cyclops 134 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Mts.), July—Sept. 1944 (Darlington); 1, Mt. Lina, Cyclops Mts., 3,500 ft., Mar. 1936 (Cheesman); 1, Mt. Cyclops, 3,500 ft., Mar. 1936 (Cheesman); 4, Sigi Camp, Snow Mts., 1,500 m. (about 4,875 ft.), Feb. 1939 (Toxopeus); 1, Rattan Camp, Snow Mts., 1,150 m. (about 3,750 ft.), Feb._Mar. 1939 (Toxopeus). Other material. Six specimens from Sigi Camp, Snow Mts., 1,500 m. (about 4,875 ft.), Feb. 1939 (Toxopeus); and 2, Araucaria Camp, Snow Mts., 800 m. (about 2,600 ft.), Mar. 1939 (Toxopeus). These specimens have the ventral pubescence of angustellum but approach subnigrum (below) in color and (especially the Araucaria Camp speci- mens) in appearance. I should consider them a subspecies of angus- tellum except that more typical specimens of the latter occur in the Snow Mts., including Sigi Camp, too. Measured specimens. The & holotype and 1 @ paratype from Dobodura. Notes. This species is easily recognized by its small size, slender and depressed form, long antennae, and extensively pubescent. abdomen. The last character is repeated in Notagonum sinuuwm and vaporum which, however, are differently shaped, much wider species, with rela- tively narrower heads. Specimens of angustellum from different lo- calities vary a little in size and appearance, but I cannot define recognizable subspecies. My specimens were all taken among stones by rapid streams. At Dobodura, angustellum occurred only beside small streams in forest. Along larger, opener streams it was replaced by Notagonum subnigrum and Lithagonum annulicorne, which also live among stones or in stone-and-gravel bars. NOTAGONUM SUBNIGRUM 0. Sp. Description. With characters of genus as described above. Form of a very large, slender, slightly flattened Bembidion; black or piceous above and below, legs and first 4 antennal segments blackish, outer antennal segments browner, lateral margins of prothorax and elytra not translucent; surface moderately shining, not iridescent; micro- sculpture normal. Head .90 & .88 width prothorax; eyes large, promi- nent, with posterior supraocular setae about between their posterior margins. Prothorax quadrate-subcordate; width/length 1.36 & 1.41; base/apex 1.10 & 1.15; sides less arcuate than usual, strongly sinuate about 1/6 or 1/7 of length before basal angles; latter approximately right, very well defined; lateral margins narrow; basal foveae very shallow, flat, with surface somewhat irregular but not distinctly punctate; anterior marginal line widely interrupted at middle, posterior one vague. Elytra rather narrow but with sides a little more arcuate DARLINGTON: CARABID BEETLES OF NEW GUINEA 135 than in angustellum, subdepressed; lateral margins narrow; subapical sinuations moderate; apices independently more or less rounded to suture, not denticulate; striae moderately impressed, the outer ones especially more or less distinctly punctulate; intervals nearly flat or slightly convex, 8th and 9th not much modified toward apex. Lower surface nearly impunctate except that abdomen has a little scattered pubescence (much less than in angustellum) rising from fine punctures. Legs: 4th hind-tarsal segment lobed, outer lobe longer than inner. Measurements: length 7.5-8.3; width 2.8-3.1 mm. Types. Holotype @ (M.C.Z. No. 28,594) and 28 paratypes from Dobodura, Papua, Mar.-—July 1944 (Darlington); and 1 o& paratype from Kokoda, Papua, 1,200 ft., Sept. 1933 (Cheesman). My specimens were taken in cobble-stone-and-gravel bars and in other cover by fairly large streams. Measured specimens. The o& holotype and 1 2 peratype from Dobodura. Notes. This is sufficiently distinguished from other species in the key, above. There is no sign of intergradation between this species and angustellum at Dobodura, but some specimens of angustellum from the Snow Mts. of Netherlands New Guinea are somewhat subnigrum- like at least superficially, as already noted. NOTAGONUM VILE N. sp. Description. With characters of genus as described above. Form of a very large Bembidion (Peryphus) or small Agonum (Europhilus) ; brownish-piceous, legs yellowish or brownish, antennae brownish, outer margins of prothorax and elytra somewhat paler or translucent; surface moderately shining, not or faintly iridescent; microsculpture normal, light. Head .86 & .89 width prothorax; eyes large, prominent, with posterior supraocular setae about between their posterior edges. Prothorax more or less subcordate; width/length 1.35 & 1.35; base/apex 1.26 & 1.23; sides moderately arcuate, then moderately sinuate a little before posterior angles; latter right or slightly obtuse, very little blunted; lateral margins rather narrow; basal foveae somewhat vari- able, flat to moderately impressed, roughened but not punctate; an- terior marginal line variable, light or interrupted at middle, posterior one vague. Elytra slightly shorter than in preceding species, of about average outline and convexity, with disc sometimes faintly impresse- before middle; lateral margins normal; subapical sinuations moderated apices independently more or less rounded to suture, rarely sub; denticulate at sutural angles; striae moderately impressed, not or 136 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY faintly punctulate; intervals flat or slightly convex, 8th and 9th not much modified toward apex. Lower surface nearly impunctate, but abdomen with a little scattered pubescence (much less than in angus- tellum) chiefly along median area. Legs: 4th hind-tarsal segment lobed, outer lobe longer than inner. Measurements: length 6.7-8.2; width 2./-3.3 mm. Types. Holotype & (M.C.Z. No. 28,595) and 26 paratypes from Dobodura, Papua, Mar.—July 1944 (Darlington); and 4 paratypes from Oro Bay (near Dobodura), Dec. 1943-Jan. 1944 (Darlington). Other material. Papua: 9, Milne Bay, Dec. 1943 (Darlington), and 1, same locality, Oct. 20, 1943 (W. B. Jones, Alabama Mus. Nat. Hist.); 3, Palmer River at Black River, June & July 1936 (Archbold Exped., American Mus.). N-E. N. G.: 17, Lae, Oct. 1944 (Darling- ton); 15, Nadzab, July 1944 (Darlington); 9, Surprise Creek, Morobe Dist., Sept. & Oct. (Stevens, M.C.Z.); 62, Chimbu Valley, Bismarck Range, 5,000-7,500 ft., Oct. 1944 (Darlington). Neth. N. G.: 31, vicinity of Hollandia (actually S. foothills of Cyclops Mts.), July—Sept. 1944 (Darlington); 1, Cyclops Mts., 3,500 ft., Mar. 1936 (Cheesman), and 1 Cyclops Mts. without further locality (Cheesman); 2, Bewani Mts., Humboldt Bay Dist., 400 m. (about 1,300 fet.), July 1937 (W. Stiiber, British Mus.); 6, Araucaria Camp, Snow Mts., 800 m. (about 2,600 ft.), Mar. 1939 (Toxopeus); 2, Sigi Camp, Snow Mts., 1,500 m. (about 4,875 ft.), Feb. 19 & 25, 1939 (Toxopeus); 2, Sansapor (Vogelkop), Aug. 1944 (Darlington). Also New Britain: 28, Cape Gloucester, Jan.—Feb. 1943 (Darlington). Measured specimens. Holotype @ and 1 2 paratype from Dobodura. Notes. See key (above) for distinguishing characters of the species. It is very common, in my experience, in grass and other cover beside large streams in more or less open country. The species varies both individually and geographically. I cannot now define useful subspecies, but it is likely that they will be recognized in the future. In the meantime, and for this reason, I have limited the actual type series to specimens from a restricted area. NOTAGONUM REVERSUM N. sp. Description. With characters of genus as described above. Form of preceding (vile) but a little more slender and convex; piceous-black, lower surface, legs, and antennal bases brownish-piceous, lateral margins of prothorax and elytra not noticeably paler; surface moder- ately shining, not distinctly iridescent; microsculpture normal, light. Head .87 & .85 width prothorax; eyes moderately large and prominent, with posterior supraocular setae about between their posterior margins. DARLINGTON: CARABID BEETLES OF NEW GUINEA 137 Prothorax quadrate-subcordate, wider than long; width/length 1.18 & 1.18; base/apex 1.25 & 1.23; sides moderately arcuate anteriorly, strongly sinuate 14 or more of length before base; posterior angles more or less acute, very well defined; lateral margins narrow; basal foveae shallow, flat or vaguely linear, not punctate and not much roughened; anterior marginal line more or less interrupted at middle, posterior one usually lightly impressed. Elytra rather elongate and convex; marginal line distinctly but obtusely angulate at humeri; lateral margins rather narrow; subapical sinuations rather weak; apices irregularly almost conjointly rounded, slightly produced, vaguely angulate (not denticulate) near sutural angles; striae moderately im- pressed, more or less punctulate; intervals moderately convex, 8th and 9th not much modified toward apex. Lower surface not distinctly punctate; abdomen not pubescent. Legs: 4th hind-tarsal segment lobed, outer lobe longer than inner. Measurements: length 7.3-8.6; width 2.7-3.3 mm. Types. Holotype @ (M.C.Z. No. 28,596) and 11 paratypes all from Chimbu Valley, Bismarck Range, N-E. N. G., 5,000-7,500 ft., Oct. 1944 (Darlington), taken along streams in open country. Measured specimens. The & holotype and 1 @ paratype. Notes. This species is distinguished from all others of the genus except the following (reversior, q.v.) by the form of the prothorax, with sides sinuate at least 14 of the prothoracic length before the base. NoTAGONUM REVERSIOR N. sp. Description. With characters of genus as described above, except eyes abnormal (see below). Form of preceding (reverswm), rather slender and convex; brownish-piceous, slightly paler below, legs and antennae brownish-yellow, lateral margins of prothorax and elytra not distinctly paler; surface moderately shining, not iridescent; micro- sculpture normal but light and restricted. Head .93 & .95 width prothorax; eyes relatively small but abruptly prominent, with posterior supraocular setae well behind line of their posterior edges. Prothorax subquadrate, relatively narrow anteriorly; width/length 1.00 & .99; base/apex 1.44 & 1.35; sides slightly, more or less irregularly arcuate anteriorly, rather strongly sinuate about 14 of length before base; basal angles acute; lateral margins very narrow; basal foveae only moderately impressed but entire basal area strongly depressed in the type and paratype but not in the third specimen; entire base of prothorax irregularly punctate in all specimens; anterior marginal line entire or nearly so, posterior one entire. Elytra suboval in the type 138 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY and paratype, longer in the third specimen; broadly and irregularly impressed about 14 from base; marginal line moderately angulate at humeri; outer margins rather narrow; subapical sinuations absent; apices rather narrowly rounded, not denticulate in the type and para- type but strongly denticulate (at sutural angles) in the third specimen; striae moderately impressed, faintly or not punctulate; intervals slightly convex, 8th and 9th not much modified toward apex. Lower surface with sides of sterna more or less punctate; abdomen not pu- bescent. Legs: 4th hind-tarsal segment lobed, outer lobe longer than inner. Male copulatory organs: Fig. 26. Measurements: length 8.6; width 3.3 mm. (third specimen 9.3 by 3.4 mm.). Types. Holotype @ (Leiden Mus.) and 1 9 paratype (M. C.Z. No. 28,597) from Moss Forest Camp, Snow Mts., Neth. N. G., 2,800 m. (about 9,100 ft.), Oct. 9-Nov. 5, 1938 @iosinen) Other material. A third specimen with the same data as the types, but differing from them in several rather striking details. Measured specimens. The o& @ types. Notes. This species is probably related to reversum (above), but differs from it in several characters including form of eyes, which are unique in Notagonwm. However, more or less similar “popped” eyes occur in several other, unrelated groups of Agonini, as already noted in the introduction. The differences in form of base of prothorax and in form and apices of elytra between the types and the third specimen mentioned above are so great that I should consider the third specimen a different species, if it came from a different locality. However, since all the specimens are from one locality, I think it likely that they all represent one strikingly dimorphic species. NOTAGONUM EXTERNUM DN. sp. Description. With characters of genus as described above. Form of Agonum s. s. (Fig. 2); piceous-black, lower surface only slightly paler, antennae and legs more or less dark-brown, lateral margins of prothorax and elytra more or less pale-translucent; surface moderately shining, not iridescent; microsculpture normal. Head .81 & .76 width prothorax (head still narrower in some specimens); eyes large, promi- nent, with posterior supraocular setae between their posterior edges. Prothorax subcordate; width/length 1.47 & 1.49; base/apex 1.30 & 1.25; sides normally arcuate, moderately or slightly sinuate before posterior angles; latter obtuse or nearly right, blunted or narrowly rounded; lateral margins moderate; basal foveae deep, not sharply defined, micro-reticulate but not punctate; anterior marginal line faint or interrupted at middle, posterior one faint. Elytra of normal outline DARLINGTON: CARABID BEETLES OF NEW GUINEA 139 and convexity; disc vaguely, broadly impressed about 14 from base; external margins moderate, each ending in a right or obtuse, well defined angle at the end of the lateral gutter; apices broadly emarginate from outer angle to opposite 3rd interval, then truncate to sutural angles; latter denticulate; striae deep, impunctate; intervals moder- ately convex, 8th and 9th not much modified toward apex. Lower surface almost impunctate; abdomen usually with a very little fine pubescence near middle. Legs: 4th hind-tarsal segment with a moder- ate outer and shorter inner lobe (Fig. 18). Male copulatory organs as figured (Fig. 27). Measurements: length 6.2-7.0; width 2.4-2.8 mm. Types. Holotype & (M.C.Z. No. 28,598) and 30 paratypes from Dobodura, Papua, Mar.—July 1944 (Darlington). Other material. Papua: 23, Milne Bay, Dec. 1943 (Darlington); 1, Port Moresby, Oct. 1944 (Darlington). Neth. N. G.: 6, Sansapor, on Vogelkop, Aug. 1944 (Darlington). Measured specimens. The o holotype and 1 92 paratype. Notes. This species is distinguished from all other members of the genus by the well-formed outer angles of the elytra. It apparently ranges over the whole of New Guinea, but I cannot divide it into subspecies. I know that the species occurs on the ground in wet places, but since I did not distinguish it in the field, I cannot define its habitat more exactly. NOTAGONUM SINUUM DN. sp. Description. With characters of genus as described above. Form of Agonum (Platynus); piceous, legs and basal segments of antennae brownish-piceous, outer antennal segments paler brown, lateral margin of prothorax moderately translucent, of elytra scarcely so; surface moderately shining, not distinctly iridescent; microsculpture normal but less transverse than usual on elytra. Head .80 & .78 width pro- thorax; eyes moderately large and prominent, with posterior supra- ocular setae about between their posterior edges (but see aberrant specimen described below). Prothorax subcordate; width/length 1.38 & 1.41; base/apex 1.24 & 1.14; sides rather strongly arcuate anteriorly, strongly sinuate about 1/6 of length before base; posterior angles right or slightly obtuse, well defined; lateral margins moderate; basal foveae shallow, almost flat, much roughened, almost punctate; anterior marginal line more or less interrupted at middle, basal one vague or absent. Elytra of average outline and convexity; lateral margins normal; subapical sinuations moderate; apices more or less conjointly rounded, not distinctly denticulate; striae rather deep, not or faintly punctulate; intervals convex, 8th moderately narrowed toward apex 140 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY 9th toward apex less wide and much more convex than usual. Lower surface at sides more or less subrugose or subpunctate; abdomen extensively pubescent. Legs: 4th hind-tarsal segment strongly lobed, outer lobe longer than inner. Measurements: length 6.6-7.5; width 2.5-3.0 mm. Types. Holotype & (M.C.Z. No. 28,599) and 7 paratypes from Chimbu Valley, Bismarck Range, N-E. N. G., 5,000-7,500 ft., Oct. 1944 (Darlington). Additional paratypes from Neth. N. G.: 2, Mt. Cyclops, 3,500 ft., Mar. 1936 (Cheesman); 2, Araucaria Camp, Snow Mts., 800 m. (about 2,600 ft.), Mar. 1939 (Toxopeus). Other material. One 9, Sigi Camp, Snow Mts., 1,500 m. (about 4,875 ft.), Feb. 19, 1939 (Toxopeus). In this specimen the eyes, though of nearly normal shape, are reduced in size, so that the posterior supra- ocular setae are behind the line of their posterior edges. In all other characters (form, etc., sinuation of sides of prothorax, form of 9th elytral interval, pubescence of abdomen, form of 4th hind-tarsal segment, and form of mentum tooth, which is more or less emarginate at tip in this species) this specimen appears to be sinuwm, of which I tentatively consider it a variant. Measured specimens. The o& holotype and 1 Q paratype from Chimbu Valley. Notes. The extensive abdominal pubescence, plus the general form and strong sinuation of the sides of the prothorax and also the unusual convexity of the 9th elytral interval toward apex, make this an easily recognized and strongly characterized species. It is probably closely related only to the following (vaporum), q.v. NOTAGONUM VAPORUM N. sp. Description. With characters of genus as described above. Form of Agonum (Platynus); piceous, legs and especially antennae browner, outer margins of prothorax moderately translucent, of elytra less so; surface moderately shining; microsculpture normal but less transverse than usual on elytra. Head .81 & .81 width prothorax; eyes moderately large and prominent, with posterior supraocular setae about between their posterior edges. Prothorax subcordate; width/length 1.41 & 1.42; base/apex 1.17 & 1.15; sides irregularly arcuate anteriorly, moderately sinuate near base; basal angles a little obtuse (nearly right), well de- fined; lateral margins rather wide, a little reflexed and elevated toward base; basal foveae deep, moderately roughened; anterior marginal line more or less interrupted at middle, posterior one vague at middle. Elytra of average outline and convexity; disc vaguely impressed about basal 14; lateral margins normal; subapical sinuations rather slight; DARLINGTON: CARABID BEETLES OF NEW GUINEA 14] apices irregularly rounded, vaguely angulate (not denticulate) near suture; striae rather deep, not or faintly punctulate; intervals slightly convex, 8th moderately narrowed toward apex, 9th unusually narrow and convex toward apex. Lower surface nearly impunctate; abdomen extensively pubescent. Legs: 4th hind-tarsal segment lobed, outer lobe longer than inner. Male copulatory organs: Fig.28. Measurements length 6.5; width 2.5 mm. Types. Holotype &@ (Leiden Mus.) and1 92 paratype (M.C.Z. No, 28,600) from Mist Camp, Snow Mts., Neth. N. G., 1,800 m. (about 5,850 ft.), Jan. 1939 (Toxopeus). Measured specimens. The types. Notes. The general form, relatively narrow and convex apical part of the 9th (submarginal) elytral intervals, pubescent abdomen, and other details indicate that this species is closely related to the preceding (stnwum), but the prothorax of the present species is much less sinuate at sides, with wider and posteriorly much more elevated lateral margins and much deeper basal foveae. NoTAGONUM AITAPE N. sp. Description. With characters of genus as described above. Form of broad but very small Agonum s. s.; piceous, legs and antennae dark- brown, lateral margins of prothorax and elytra more or less conspicu- ously yellow; surface moderately shining, slightly iridescent especially on elytra; microsculpture nearly normal, but very fine (scarcely visible at 54x) on elytra. Head .70 & .71 width prothorax; eyes large and prominent, with posterior supraocular setae about between their posterior edges. Prothorax rather large; width/length 1.52 & 1.55; base/apex 1.26 & 1.17; sides arcuate for most of length, slightly sinuate near base; basal angles obtuse and slightly blunted; lateral margins rather wide but only slightly reflexed; basal foveae not very deep, a little roughened but not punctate; anterior marginal line vague at middle, posterior one indistinct. Elytra of about normal outline and convexity; lateral margins rather wide (in group); subapical sinuations rather slight; apices rather irregularly independently rounded, more or less subangulate (not denticulate) near suture striae deep, not distinctly punctate; intervals more or less convex, 8th and 9th not much modified toward apex. Lower surface impunctate; abdomen not pubescent. Legs: 4th hind-tarsal segment emarginate at apex, not lobed. Measurements: length 4.8-5.6; width 2.0-2.4 mm. Types. Holotype @ (M.C.Z. No. 28,601) and 10 paratypes all from Aitape, N-E. N. G., Aug. 1944 (Darlington), taken in floodwater in forested or recently forested country. 142 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Other material. One, vicinity of Hollandia, Neth. N. G., July- Sept. 1944 (Darlington). Measured specimens. The o& holotype and 1 @ paratype from Aitape. Notes. This species should be easily recognized by its small size, rather broad form, simple elytral apices, yellow lateral margins, and simply emarginate 4th hind-tarsal segment. NoTAGONUM AITAPE SANSAPOR N. subsp. Description. Apparently same as typical aitape (of which see de- scription, above) in structure, but a little larger, and differing in color, the elytra having broad yellowish-brown lateral margins and the legs and antennae being paler. The pale elytral margins reach inwardly about to the 6th intervals, but are not sharply limited, grading into the piceous color of the elytral dise. Proportions: head/prothorax 71 & .70; width/length of prothorax 1.45 & 1.50; base/apex of pro- thorax 1.18 & 1.22. Measurements: length 5.5-6.5; width 2.2-2.7 mm. Types. Holotype o& (M.C.Z. No. 28,602) and 21 paratypes all from Sansapor (Vogelkop), Neth. N. G., Aug. 1944 (Darlington), taken in wet places on the ground in forested country. Measured specimens. The &@ holotype and 1 92 paratype. Notes. Sufficiently compared with typical actape above. NoOTAGONUM GIBBUM N. sp. Description. With characters of genus as described above. Form of a very large Mecyclothorax (much more convex than normal Agonwm); piceous-black, elytra faintly brassy in some lights, appendages reddish- yellow, outer antennal segments browner, outer margins of prothorax and elytra slightly or not paler; surface shining, not or only faintly iridescent; microsculpture finer than usual, not distinct at 54x, but surface especially of elytra with silky texture. Head .78 & .77 width prothorax; eyes moderately large and prominent, with posterior supra- ocular setae about between their posterior edges. Prothorax transverse- rounded, much more convex than usual; width/length 1.47 & 1.52; base/apex 1.13 & 1.17; sides arcuate for most of length, not or only slightly sinuate before basal angles; latter obtuse, blunted; lateral margins usually very narrow, but somewhat variable; basal foveae rather small, shallow, not well defined, somewhat roughened or vaguely punctate; anterior and posterior marginal lines both rather vague, more or less interrupted at middle. Elytra of about normal outline but much more convex than usual; disc of each elytron slightly impressed DARLINGTON: CARABID BEETLES OF NEW GUINEA 148 about 26 from base; basal margin rounded at humeri or at most vaguely subangulate (as usual in genus); lateral margins very narrow; subapical sinuations slight; apices conjointly or slightly independently rounded, not denticulate; inner discal striae very fine, usually almost obsolete anteriorly, but somewhat variable; outer striae and inner ones toward apex moderately impressed; striae vaguely punctulate; inter- vals flat or slightly convex, 8th and 9th not much modified toward apex. Lower surface impunctate; abdomen not pubescent. Legs: 4th hind-tarsal segment lobed, outer lobe longer than inner. Male copu- latory organs: Fig. 29. Measurements: length 7.1-7.8; width 2.6- 3.2 mm. Types. Holotype &@ (Leiden Mus.) and 12 paratypes (some in M.C.Z. No. 28,603) all from Sigi Camp, Snow Mts., Neth. N. G., 1,500 m. (about 4,875 ft.), Feb. 1939 (Toxopeus). Measured specuomens. The o holotype and 1 @ paratype. Notes. The very convex form, narrow prothoracic and elytral margins, and fineness of elytral striae on anterior part of disc distin- guish this species and give it a remarkable similarity to a very large Mecyclothorax. In most other ways the species is very close to the following (sigi), from which it may have been derived. NOTAGONUM SIGI n. sp. Description. With characters of genus as described above. Form of Agonum (Platynus); piceous, appendages yellowish, outer margins of prothorax slightly translucent, of elytra scarcely so; surface moderately shining, not or faintly iridescent; microsculpture normal except that of elytra too fine to see at 54, but elytral surface slightly silky. Head .78 & .78 width prothorax; eyes moderately large and prominent, with posterior supraocular setae about between their posterior edges. Prothorax subcordate; width/length 1.89 & 1.41; base/apex 1.21 & 1.27; sides moderately arcuate, moderately sinuate a little before basal angles; latter slightly obtuse, only slightly blunted; lateral margins average; basal foveae average, not sharply defined, slightly roughened; anterior marginal line fine but entire, posterior one light or vague. Elytra of normal outline and convexity; lateral margins rather narrow; subapical sinuations rather slight; apices more or less independently rounded to suture, where subangulate but not distinctly denticulate; striae moderately impressed (7th sometimes finer toward base), not or vaguely punctulate; intervals only slightly convex, 8th a little narrowed toward apex, 9th widened and nearly flat toward apex (i.e. 8th and 9th intervals not much modified). Lower surface with at most a little rather vague punctation; abdomen not pubescent. Legs: 144 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY 4th hind-tarsal segment lobed, outer lobe longer than inner. Measure- ments: length 7.7-7.9; width 2.9-3.0 mm. Types. Holotype @ (Leiden Mus.) and 3 paratypes (1 co in M.C.Z., No. 28,604) all from Sigi Camp, Snow Mts., Neth. N. G., 1,500 m. (about 4,875 ft.), Feb. 27 & 25, 1939, or (1 @ only) 1,600 m. (about 5,200 ft.), Dec. 1938 (Toxopeus). Measured specimens. The o& holotype and 1 @ paratype. Notes. This species is superficially rather like vaporum of the same general region, but differs in such important characters as the form of the 9th elytral interval (wide and nearly flat toward apex instead of narrow and very convex) and the non-pubescent abdomen. I have already suggested the possibility that this comparatively normally formed species may be ancestral to the preceding very convex one (gibbum). NOTAGONUM ALTUM N. sp. Description. With characters of genus as described above, except 9 with usually only 1 seta each side last ventral segment. Form of Agonum (Platynus); piceous, appendages brown, lateral margins of prothorax and elytra only slightly paler; microsculpture normal. Head .79 & .79 width prothorax (sometimes a little wider) ; eyes moder- ately large and prominent, with posterior supraocular setae about be- tween their posterior edges. Prothorax subcordate; width/length 1.48 & 1.44; base/apex 1.16 & 1.21; sides moderately arcuate anteriorly, moderately sinuate about 14 of length before base; basal angles more or less obtuse, blunted; lateral margins moderate; basal foveae moder- ate, roughened or subpunctate; anterior marginal line entire or nearly so but often light at middle, posterior one vague at middle. Elytra a little more than usually narrowed toward humeri and a little more convex than usual; lateral margins rather narrow; subapical sinuations slight; apices independently rounded; striae rather deep, not or vaguely punctulate; intervals moderately convex, 8th and 9th not much modi- fied toward apex. Lower surface nearly impunctate; abdomen not pu- bescent. Legs: 4th hind-tarsal segment lobed, outer lobe slightly longer than inner. Secondary sexual characters normal except 2 usually with only 1 (not 2) seta each side last ventral segment (at least one @ is asymmetrical, with 2 setae on one side, 1 on other). Measurements: length 7.7-8.6; width 2.9-3.6 mm. Types. Holotype & (M.C.Z. No. 28,605) and 25 paratypes all from Chimbu Valley, Bismarck Range, N-E. N. G., some (including type) from the forested zone between 7,000 & 10,000 ft., others from open country between 5,000 & 7,500 ft., but all actually taken under cover beside the Chim River, Oct. 1944 (Darlington). DARLINGTON: CARABID BEETLES OF NEW GUINEA 145 Measured specumens. The & holotype and 1 @ paratype. Notes. The distinguishing characters of this species are sufficiently given in the key. NOTAGONUM ALTUM IBELE N. subsp. Description. Similar to typical altwm (of which see description, above) except in proportions of prothorax, which is relatively slightly narrower in zbele, with base especially narrower: in zbele the ratio base/apex of prothorax is 1.03 (co type), 1.09 (2), and 1.13 (second o'); in 6 measured specimens of typical altum this ratio is 1.15 to 1.22. Other proportions of ibele (o'o' Q ) are head/prothorax .80, .79, .84; width/length prothorax 1.40, 1.41, 1.38. Thesingle 9 of ibele has only 1 seta each side last ventral as usual in altum. Measurements: length 8.8-9.2; width 3.1-3.4 mm. (a little larger but relatively more slender than typical altum). Types. Holotype &@ (Leiden Mus.) and 2 paratypes (o”, M.C.Z. No. 28,606; 9, Buitenzorg Mus.) all from Iebele (Ibele) Camp, Snow Mts., Neth. N. G., 2,250 m. (about 7,325 ft.), Nov.—Dec. 1938 (Toxopeus). Measured specimens. As indicated above. Notes. Sufficiently compared with typical altwm above. NoTAGONUM MARGARITUM N. sp. Description. With characters of genus as described above. Form between Agonum s. s. and Platynus; piceous-black with pearly lustre, appendages brownish-yellow, lateral margins of prothorax and elytra moderately pale-translucent (elytra the less so); surface moderately shining, moderately (not strongly) iridescent especially on elytra; microsculpture apparently normal but almost too fine to see at 54X. Head .74 & .74 width prothorax; eyes large and prominent, with posterior supraocular setae about between their posterior edges. Prothorax somewhat transverse; width/length 1.42 & 1.40; base/apex 1.26 & 1.29; sides arcuate for most of length, slightly or not distinctly sinuate before base; posterior angles obtuse, blunted or narrowly rounded; lateral margins moderate; basal foveae rather deep, scarcely roughened, not punctate; anterior marginal line fine but entire and distinct at middle, posterior one less distinct. Elytra of about normal outline and convexity; lateral margins moderate; subapical sinuations slight or moderate; apices independently rounded (sometimes subangu- late opposite 3rd intervals), each with a strong denticle at suture; striae moderately impressed, usually more or less distinctly punctulate; 146 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY ntervals a little convex, 8th and 9th not much modified toward apex. Lower surface impunctate or slightly or vaguely punctate at sides of sterna; abdomen not pubescent. Legs: 4th hind-tarsal segment briefly lobed, outer lobe slightly longer than inner. Measurements: length 7.0-9.0; width 2.7-3.3 mm. Types. Holotype o& (M.C.Z. No. 28,607) and 10 paratypes from Nadzab, N-E. N. G., July 1944 (Darlington). Also the following additional paratypes: Papua: 1, Mafulu, 4,000 ft., Dec. 1933 (Chees- man). N-E. N. G.: 1, Finschhafen, Apr. 20, 1944 (E. S. Ross, Cali- fornia Acad.); 13, Chimbu Valley, Bismarck Range, 5,000-7,500 ft., Oct. 1944 (Darlington). Neth. N. G.: 3, vicinity of Hollandia, July- Sept. 1944 (Darlington); 1, Sabron, Cyclops Mts., 930 ft., May-June 1936 (Cheesman); 12, Cyclops Mts. (including Mt. Cyclops and Mt. Lina), 3,400(or 3,500)-4,500 ft., Mar. 1939 (Cheesman); 1, Rattan Camp, Snow Mts., 1,150 m. (about 3,750 ft.), Feb._Mar. 1939 (Toxo- peus); 2, Baliem Camp, Snow Mts., 1,600 & 1,700 m. (about 5,200 & 5,025 ft.), Dec. & Nov. 16-27, 1938 (Toxopeus). Measured specumens. The o& holotype and 1 9 paratype from Nadzab. Notes. The comparatively weak subapical sinuations of the elytra, usually distinctly iridescent surface, and frequently (but not always) punctulate elytral striae distinguish this species from other similar ones. In doubtful cases and in the absence of comparative material the following details should aid in distinguishing this from other species with denticulate but otherwise unarmed elytral apices: 4th hind-tarsal segment lobed (not simply emarginate as in suwbpunctum); anterior transverse marginal line of pronotum entire and distinct at middle (not vague or interrupted at middle as in dentellwm); and elytra not transversely impressed before middle (as in subimpressum). NoOTAGONUM SUBPUNCTUM 0. sp. Description. With characters of genus as described above. Form of Agonum s. s.; black or piceous, appendages dark-brown, outer margins of prothorax and elytra slightly translucent; surface moderately shining, not distinctly iridescent; microsculpture of pronotal foveae less distinct than usual, of elytra more distinct, coarse, and less trans- verse than usual. Head .69 & .67 width prothorax; eyes large and prominent, with posterior supraocular setae about between their posterior edges. Prothorax rather large; width/length 1.45 & 1.47; base/apex 1.24 & 1.29; sides arcuate anteriorly, nearly straight and converging posteriorly, slightly or not sinuate before base; posterior angles obtuse, blunted; lateral margins a little wider than usual and DARLINGTON: CARABID BEETLES OF NEW GUINEA 147 more elevated near base; basal foveae rather deep, usually irregularly punctate, with some punctures usually also before foveae near pronotal margins; anterior marginal line entire, basal one vague. Elytra of about normal outline and convexity; outer margins moderately wide; sub- apical sinuations strong; apices independently rounded, bluntly denti- culate at suture; striae deep, impunctate; intervals somewhat convex, Sth narrowed and very convex toward apex, 9th only slightly widened and moderately convex toward apex. Lower surface impunctate or nearly so; abdomen not pubescent. Legs: 4th hind-tarsal segment simple emarginate, not lobed. Measurements: length 6.5-7.4; width 2.4-2.7 mm. Types. Holotype &* (M.C.Z. No. 28,608) and 16 paratypes all from Dobodura, Papua, Mar.—July 1944 (Darlington). Measured specumens. The o& holotype and 1 Q paratype. Notes. This species is instantly distinguishable from superficially rather similar ones (margaritum, dentellum, subimpressum) by the simply emarginate rather than lobed 4th hind-tarsal segment. The unusually heavy micro-reticulation of the elytral surface should aid in recognition of the typical form of this species, but not of the following subspecies. NoTAGONUM SUBPUNCTUM CAPITIS n. subsp. Description. Nearly the same as typical subpunctum (of which see description, above) in form, proportions, and most detailed characters, including simply emarginate 4th hind-tarsal segment (Fig. 19), but larger, with elytra much more finely micro-reticulate, the meshes barely visible at 54X. Proportions of measured specimens: head/prothorax .69 & .69; width/length prothorax 1.46 & 1.43; base/apex prothorax 1.20 & 1.24. Measurements: length 7.7-8.5; width 2.8-3.2 mm. Types. Holotype ®@ (M.C.Z. No. 28,609) and 10 paratypes all from Sansapor (Vogelkop), Neth. N. G., Aug. 1944 (Darlington), taken in wet places in forested country. Other material. Neth. N. G.: 1, Maffin Bay, Aug. 1944 (Darlington) 1, Hollandia, June 10, 1945 (from K. M. Fender, M.C.Z.). Measured specimens. The o& holotype and 1 @ paratype. Notes. Sufficiently compared with typical subpunctum above and in the key to species of Notagonum. NOoTAGONUM DENTELLUM N. sp. Description. With characters of genus as described above. Form of rather broad Agonum s. s.; piceous-black, appendages yellow or brownish-yellow, lateral margins of prothorax and elytra yellow; 148 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY surface moderately shining, slightly iridescent; microsculpture normal. Head .74 & .73 width prothorax; eyes large and prominent, with posterior supraocular setae about between their posterior edges. Pro- thorax moderately transverse; width/length 1.43 & 1.47; base/apex 1.26 & 1.21; sides arcuate for much of length, then moderately (some- what variably) sinuate near base; posterior angles obtuse, blunted; lateral margins rather wide; basal foveae moderately deep, slightly roughened; anterior and posterior marginal lines irregular, faint, usu- ally incomplete. Elytra of about normal outline and convexity, not or very little impressed on disc; lateral margins slightly wider than usual; subapical sinuations rather strong; apices typically narrowly sub- truncate (but exact form somewhat variable), usually denticulate at suture; striae moderately deep, not or slightly punctulate; intervals slightly convex, 8th and 9th not much modified toward apex. Lower surface impunctate; abdomen not pubescent. Legs: 4th hind-tarsal segment lobed, outer lobe longer than inner. Male copulatory organs: Fig. 30. Measurements (Dobodura series): length 7.3-8.5; width 2.8- 3.4 mm. (some smaller specimens from other localities). Types. Holotype & (M.C.Z. No. 28,610) and 11 paratypes from Dobodura, Papua, Mar.—July 1944 (Darlington). Additional para- types as follows: N-E. N. G.: 1, Lae, Oct. 1944 (Darlington); 5, Nadzab, July 1944 (Darlington); 12, Aitape, Aug. 1944 (Darlington). Neth. N. G.: 10, vicinity of Hollandia, July—Sept. 1944 (Darlington); 3, Maffin Bay, Aug. 1944 (Darlington). Other material. Papua: 4, Dobodura, Mar.—July 1944 (Darlington) (more slender than typical specimens, with elytra somewhat impressed at or slightly before middle). N-E. N. G.: 2, Nadzab, July 1944 (Darlington) (slender, elytra slightly impressed and not-denticulate) ; 1, Surprise Creek, Morobe Dist., Oct. 4 (Stevens, M.C.Z.) (broad, sides of prothorax strongly sinuate). Neth. N. G.: 1, vicinity of Hollandia, July-Sept. 1944 (Darlington) (more slender and with elytral striae more strongly punctulate than usual); 1, Mt. Cyclops, 3,500 ft., Mar. 1936 (Cheesman) (large, with margins and appendages darker than usual); 1, Araucaria Camp, Snow Mts., 800 m. (about 2,600 ft.), Mar. 1939 (Toxopeus) (rather small and with sides of prothorax more sinuate than usual); and 1, Camp 1, Mt. Nok, Waigeo Is., 2,500 ft., May 1938 (Cheesman) (rather small, with head relatively large and prothorax smaller and with sides more sinuate than usual, and with elytral apices squarely truncate). More material is needed to clarify the status of these forms. A form of this species occurs also at Cape Gloucester, New Britain (Darlington); and the following subspecies is, so far as known, localized on the Bismarck Range, N-E. N. G. Measured specimens. The & holotype and 1 9 paratype from Dobodura. DARLINGTON: CARABID BEETLES OF NEW GUINEA 149 Notes. Typical specimens of this species are identifiable by charac- ters given in the key to species of Notagonum. The species is, however, a variable one, as the notes given under “Other material” suggest. This is one of the species in which the elytral denticles are variable, being acute, blunt, or faint in different individuals even in the type series, and completely absent in a few aberrant individuals. NoTAGONUM DENTELLUM CHIMBU N. subsp. Description. Generally similar to typical dentellum (of which see description, above) but larger; eyes slightly less prominent; prothorax relatively smaller, with sides usually more sinuate and basal angles usually more distinct. These differences are such that, although they give the insect a somewhat different appearance, they change its proportions very little. The proportions of the measured specimens are head/prothorax .74 & .73; width/length prothorax 1.43 & 1.45; base/apex prothorax 1.23 & 1.15. The elytra of chimbu have slightly stronger subapical sinuations than in typical dentellum and are a little more coarsely striate, and the intervals are flatter toward apex. Measurements: length 8.4-9.7; width 3.2-3.8 mm. Types. Holotype &@ (M.C.Z. No. 28,611) and 17 paratypes all from Chimbu Valley, Bismarck Range, N-E. N. G., 5,000-7,500 ft., Oct. 1944 (Darlington). Measured specimens. The o holotype and 1 9 paratype. Notes. Sufficiently compared with typical dentellum above. NOTAGONUM SUBIMPRESSUM ND. sp. Description. With characters of genus as described above. Form of Agonum s. s.; piceous-black, appendages brownish-piceous, outer antennal segments paler brown, lateral margins of prothorax only slightly translucent, of elytra even less so; surface moderately shining, not or slightly iridescent; microsculpture normal, but light and re- stricted on head and prothorax. Head .74 & .73 width prothorax; eyes large and prominent, with posterior supraocular setae about between their posterior edges. Prothorax of average size; width/length 1.44 & 1.50; base/apex 1.39 & 1.36; sides arcuate for much of length, nearly straight and converging and usually slightly sinuate toward base; basal angles obtuse, blunted or narrowly rounded; lateral margins average; basal foveae rather deep, slightly roughened, sometimes vaguely punctate; anterior marginal line entire or nearly so, posterior one entire or vague at middle. Elytra of average outline and convexity, but rather strongly impressed across dise about 14 from base; lateral 150 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY margins normal; subapical sinuations rather strong; apices inde- pendently rounded or sometimes vaguely angulate about opposite 3rd interval, rather inconspicuously denticulate at suture; striae moder- ately impressed, not or vaguely punctulate; intervals slightly convex, 8th and 9th not much modified toward apex. Lower surface nearly impunctate; abdomen with a very little fine pubescence near middle of some segments. Legs: 4th hind-tarsal segment lobed, outer lobe longer than inner. Measurements: length 8.1-8.7; width 3.1-3.5 mm. Types. Holotype & (M.C.Z. No. 28,612) and 22 paratypes from Dobodura, Papua, Mar.—July 1944 (Darlington), taken in a variety of damp places. Additional paratypes as follows: Papua: 1, Oro Bay, Dec. 1943 (Darlington); 1, Kokoda, 1,200 ft., Sept. 1933 (Cheesman) ; 10, Milne Bay, Dec. 1943 (Darlington). Neth. N. G.: 18, Hollandia, July—Sept. 1944 (Darlington), and 1, same locality, Apr. 1945 (Malkin, U.S.N.M.); 1, Araucaria Camp, Snow Mts., 800 m. (about 2,600 it.), Mar. 1939 (Toxopeus); 7, Wasian, Sept. 1939 (Wind, M.C.Z.). Other material. Six, Cape Gloucester, New Britain, Jan.—Feb. 1943 (Darlington). Measured specimens. The o holotype and 1 2 paratype from Dobodura. Notes. Several other species of Notagonum have the elytral disc more or less impressed before the middle, but no other so much as this, which has almost a sway-backed appearance. This character, together with the rather dark legs and antennal bases, makes this species easy to recognize even superficially. Other distinguishing characters of the species are given in the key to species of Notagonwm. NoOTAGONUM PALUDUM N. sp. Description. With characters of genus as described above. Form of broad Agonum s. s., but small; piceous-black, appendages brownish- piceous, lateral margins of prothorax and elytra brownish-translucent; surface moderately shining, not distinctly iridescent; microsculpture neatly normal but that of pronotum very light, that of elytra more distinct. Head .67 & .65 width prothorax; eyes moderately large and prominent, with posterior supraocular setae a, trifle behind line of their posterior edges. Prothoraz relatively large; width/length 1.45 & 1.45; base/apex 1.29 & 1.28; sides arcuate for much of length, straight and converging and sometimes slightly sinuate before base; postericr angles obtuse, slightly blunted; lateral margins rather wide; basal foveae moderately deep, scarcely roughened; anterior marginal line entire, posterior one lighter or vague. Elytra of about normal outline and convexity, each somewhat impressed before middle; lateral margins DARLINGTON: CARABID BEETLES OF NEW GUINEA 151 rather wide (in genus); subapical sinuations rather strong; apices each bi-angulate or bi-denticulate (the angles or denticles about opposite 3rd and sutural intervals), with apex between angles or denticles emarginate; striae moderately impressed, not or faintly punctulate; intervals only slightly convex, more so laterally and apically (as usual), Sth and 9th not much modified toward apex. Lower surface impunctate; abdomen not pubescent. Legs: 4th hind-tarsal segment scarcely more than emarginate, but with very small inconspicuous lobes, the outer longer than inner. Measurements: length 5.6-6.5; width 2.2-2.5 mm. Types. Holotype & (M.C.Z. No. 28,613) and 27 paratypes all from Dobodura, Papua, Mar.—July 1944 (Darlington), taken among wet dead leaves by pools in forest. Measured specimens. The o& holotype and 1 2 paratype. Notes. This distinct little species should be easily recognized by characters given in the key to species of Notagonum. NoOTAGONUM PALUDUM VELUM n. subsp. Description. Generally similar to typical paludum (of which see description, above); differing only slightly in size (averaging slightly larger), proportions, and most other details (elytral striae slightly deeper, etc.); but easily distinguished by elytral microsculpture. In typical paludum the micro-reticulations on the elytra are coarser than usual and are easily seen in all specimens at 54%, but in the present new subspecies the elytral reticulations are so fine as to be barely or not visible at the same magnification. Proportions of measured speci- mens: head/prothorax .64 & .64; width/length prothorax 1.54 & 1.61; base/apex prothorax 1.24 & 1.26. Measurements: length 6.3-6.6; width about 2.5 mm. Types. Holotype & (M.C.Z. No. 28,614) and 4 paratypes (all 2 9 ) all from Aitape, N-E. N. G., Aug. 1944 (Darlington), taken in a flood in forested or formerly forested country. Measured specimens. The o holotype and 1 @ paratype. Notes. Sufficiently compared with typical paludum above. NOTAGONUM MALKINI 0. sp. Description. With characters of genus as described above. Form of rather broad Agonum s. s.; piceous-black, legs and antennal bases brownish-piceous, outer segments of antennae brown, lateral margins of prothorax slightly paler or translucent, lateral margins of elytra scarcely paler; surface moderately shining, moderately iridescent; microsculpture normal, light. Head .68 & .68 width prothorax; eyes 152 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY large and very prominent, with posterior supraocular setae about be- tween their posterior edges. Prothorax rather large and rather trans- verse; width/length 1.54 & 1.55; base/apex 1.35 & 1.38; sides arcuate for much of length, then straight and converging or very slightly sinuate to posterior angles; latter obtuse, slightly blunted; lateral margins wide but not sharply set off from disc; basal foveae rather wide, moderately deep, only slightly roughened; anterior marginal line entire or slightly interrupted at middle, posterior one vague. Elytra rather broad, of about normal outline and convexity, not distinctly impressed on disc; lateral margins slightly wider than usual; subapical sinuations rather strong; apices each minutely angulate about opposite 3rd interval (which turns somewhat toward suture), then emarginate to denticulate sutural angles; outer angle or denticle slightly more prominent than sutural denticle; striae moderately impressed, not distinctly punctulate; intervals slightly convex, Sth narrowed and much more convex toward apex, 9th somewhat widened and somewhat flattened toward apex (but these intervals still not very much more modified than usual). Lower surface nearly impunctate; abdomen not pubescent. Legs: 4th hind-tarsal segment emarginate, scarcely lobed. Male copulatory organs: Fig. 31. Measurements: length 7.6-8.2; width 3.0 or slightly more mm. Types. Holotype & (M.C.Z. No. 28,615) from vicinity of Hollandia, Neth. N. G., July-Sept. 1944 (Darlington), and 1 2 paratype from the same locality, Apr. 1945 (Borys Malkin, U.S.N.M.). Measured specimens. The types. Notes. Sufficiently distinguished from other bi-denticulate species in the key to species of Notagonum. I have considered whether the two specimens described above can be bi-denticulate individuals of a normally uni-denticulate species, but they seem not to be. They differ from subpunctum not only in form of elytral apices but also in lack of distinct punctation in the pronotal foveae and in other ways; and they differ from dentellwm not only in form of elytral apices but also in lack of conspicuous pale lateral elytral margins, in lack of distinct lobes on the 4th hind-tarsal segment, and in other ways. They are distinguished from bi-denticulate individuals of margaritum in the key. NOTAGONUM IRIDIOR N. sp. Description. With characters of genus as described above. Form of Agonum s. s.; piceous-black, legs and antennal bases slightly more reddish-piceous, outer antennal segments browner, lateral margins of prothorax and elytra somewhat translucent; surface moderately shining, elytra more iridescent than in any other species of genus except DARLINGTON: CARABID BEETLES CF NEW GUINEA 53} perhaps angulum (below); microsculpture normal but finer than usual. Head .71 & .74 width prothorax; eyes large and prominent, with posterior supraocular setae about between their posterior edges. Pro- thorax rather transverse-subcordate, with apex more deeply emarginate than usual; width/length 1.48 & 1.44; base/apex 1.28 & 1.18; sides rounded for most of length, sinuate near posterior angles; latter obtuse (partly because of rounding of sides of base), blunted; lateral margins rather wide, flatter than usual anteriorly; basal foveae rather deep, only slightly roughened, vaguely punctate; anterior marginal line entire or nearly so, posterior one vague. Elytra of about normal outline and convexity, not or faintly impressed on disc; lateral margins moder- ate; subapical sinuations rather weak; apices each angulate opposite 3rd interval, then emarginate to sutural angle; latter denticulate; striae less impressed than usual, not distinctly punctulate; intervals nearly flat, Sth and 9th not much modified toward apex. Lower surface slightly punctate at sides of sterna; abdomen not pubescent. Legs: 4th hind-tarsal segment emarginate, not lobed. Measurements: length 8.0-9.0; width 3.0-3.5 mm. Types. Holotype 2 (M.C.Z. No. 28,616) and 3 paratypes (all 2 2) all from Wasian, Neth. N. G., Sept. 1939 (R. G. Wind). Measured specimens. The 2 holotype and 1 @ paratype. Notes. The characters for recognition of this species are the iri- descence of the elytra and the rather deep emargination of the front of the prothorax. Of the less obvious characters, the form of the 4th hind-tarsal segment (simply emarginate) is noteworthy, though re- peated in a few other species of the genus. NoTAGONUM ADDENDUM DN. sp. Description. With characters of genus as described above. Form of rather broad Agonum s. s.; brownish-black, appendages brown, lateral margins of prothorax and elytra moderately translucent; surface moderately shining, not iridescent; microsculpture normal. Head .66 & .69 width prothorax; eyes large and prominent, with posterior supraocular setae a little before line of their posterior edges. Prothorax rather large and wide; width/length 1.52 & 1.53; base/apex 1.28 & 1.32; sides rather strongly arcuate for much of length, nearly straight and strongly converging and sometimes slightly sinuate before very obtuse but distinguishable basal angles; lateral margins rather wide especially toward base, moderately reflexed; basal foveae average, roughened but not distinctly punctate; anterior marginal line entire, posterior one more or less entire but vague at middle. Elytra rather broad, of normal outline and convexity, not distinctly impressed on disc; lateral margins 154 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY rather wide (in genus); subapical sinuations moderate; apices each strongly angulate about opposite 3rd interval (this angle more promi- nent than the sutural one), then obliquely subtruncate or slightly emarginate to slightly denticulate sutural angle; striae moderately impressed, not distinctly punctulate; intervals nearly flat or slightly convex, 8th and 9th not much modified toward apex. Lower surface virtually impunctate; abdomen not pubescent. Legs: 4th hind-tarsal segment deeply emarginate, with small outer but scarcely any inner lobe. Measurements: length 6.8-7.6; width 2.8-3.0 mm. Types. Holotype @ (M.C.Z. No. 28,617) and 3 oc’ paratypes from Dobodura, Papua, Mar.—July 1944 (Darlington); and 2 para- types from Mt. Nok, Waigeo Is., Neth. N. G., 2,500 ft., Apr. & May 1938 (Cheesman). Measured specimens. The o& holotype and 1 o paratype from Dobodura. Notes. This species is sufficiently distinguished from others in the key to species of Notagonum, above. In form and in shape of elytral apices it is so much like Altagonum vallicola n. sp. (below) that I at first thought it might be a form of that species which had retained the anterior pronotal and anterior dorsal elytral setigerous punctures, but the present new species differs from vallicola not only in possessing these setae and punctures but also in being less black in color, with more translucent lateral prothoracic and elytral margins, and in having small but distinct outer lobes on the 4th hind-tarsal segments. NoTAGONUM ANGULUM DN. sp. Description. With characters of genus as described above. Form of rather slender Agonum or Platynus; piceous-black, browner-piceous below, appendages yellow or brownish-yellow, lateral margins of pro- thorax rather strongly pale-translucent, of elytra scarcely so; surface moderately shining and (especially elytra) moderately iridescent; microsculpture probably nearly normal but very light and restricted on pronotum and too fine to distinguish in detail at 54 on elytra. Head .77 & .80 width prothorax; eyes moderately large and prominent, with posterior supraocular setae about between their posterior edges. Prothorax: width/length 1.30 & 1.30; base/apex 1.18 & 1.23; sides arcuate anteriorly, nearly straight and converging posteriorly, usually slightly sinuate near posterior angles; latter obtuse, blunted or narrowly rounded; lateral margins moderate; basal foveae moderate, only slightly roughened; anterior marginal line entire, posterior one inter- rupted or vague at middle. Elytra rather long but otherwise of normal outline and convexity, not or slightly impressed on disc; lateral margins DARLINGTON: CARABID BEETLES OF NEW GUINEA ays) normal; subapical sinuations moderate; apices each strongly, about rectangularly angulate (and sometimes subdenticulate) opposite 2nd or 3rd interval, then emarginate to more or less denticulate sutural angle; striae moderately deep, not distinctly punctulate; intervals slightly or moderately convex, 8th and 9th not much modified toward apex. Lower surface impunctate or nearly so; abdomen not pubescent. Legs: 4th hind-tarsal segment lobed, outer lobe longer than inner. Measurements: length 8.2-9.7; width 2.9-3.4 mm. Types. Holotype &@ (Leiden Mus.) and 16 paratypes from Rattan Camp, Snow Mts., Neth. N. G., 1,150 & 1,200 m. (about 3,750 & 3,900 ft.), Feb._Mar. 1939 (Toxopeus); and 6 paratypes from Sigi Camp, Snow Mts., 1,500 m. (about 4,875 ft.), Feb. 1939 (Toxopeus). (Some paratypes in M.C.Z., No. 28,618). Other material. One teneral, Sigi Camp (as above); 1, Chimbu Valley, Bismarck Range, N-E. N. G., 5,000-7,500 ft., Oct. 1944 (Darlington); 1, Mafulu, Papua, 4,000 ft., Dec. 1933 (Cheesman). Measured specimens. The o holotype and 1 @ paratype from Rattan Camp. Notes. This fine species is easily known by its rather slender form (compared with related species), pale-translucent prothoracic margins, iridescent elytra, and strongly angulate elytral apices. NOTAGONUM SUBANGULUM DN. sp. Description. With characters of genus as described above. Form of rather slender Agonwm s. s.; brownish-piceous (perhaps sometimes darker), appendages brownish-yellow, lateral margins of prothorax rather strongly translucent, of elytra much less so; surface moderately shining, only faintly iridescent; microsculpture normal, a little more distinct than in angulum. Head .77 & .77 width prothorax; eyes large and prominent, with posterior supraocular setae about between their posterior edges. Prothorax somewhat transverse; width/length 1.42 & 1.41; base/apex 1.36 & 1.33; sides arcuate anteriorly, converging and straight or faintly sinuate posteriorly; posterior angles obtuse, blunted or narrowly rounded; lateral margins moderate; basal foveae moderate, scarcely roughened; anterior marginal line entire, posterior one almost so. Elytra a little longer than usual but otherwise normal in outline and convexity; lateral margins average; subapical sinuations moderate; apices each strongly but a little obtusely angulate about opposite 3rd interval, then oblique forward to minutely or vaguely denticulate sutural angle; striae moderately impressed, not distinctly punctulate; intervals nearly flat or slightly convex, 8th and 9th not much modified toward apex. Lower surface not distinctly punctate; abdomen not 156 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY pubescent. Legs: 4th hind-tarsal segment lobed, outer lobe longer than inner. Measurements: length 7.7-8.2; width 2.7-3.0 mm. Types. Holotype &@ (Leiden Mus.) and 1 o paratype (M.C.Z. No. 28,619) both from Sigi Camp, Snow Mts., Neth. N. G., the type at 1,600 m. (about 5,200 ft.), Feb. 20, 1939, and the paratype at 1,500 m. (about 4,875 ft.), Feb. 1939 (both, Toxopeus). Measured specimens. 'The types. Notes. This species resembles the preceding (angulwm) in a general way but differs in several details, especially in having a wider pro- thorax with relatively wider base (cf. ratios given in descriptions) and in having much less iridescent elytra. NoTAGONUM SUBRUFUM 0. sp. Description. With characters of genus as described above, except hind-tarsal sole with a nearly regular row of bristles each side, with middle of sole broadly bare. Form of slender Agonum (Platynus) ; rather dark rufous (elytra a little darker than head and prothorax), appendages yellow, lateral margins of prothorax and elytra moderately translucent; surface moderately shining, not iridescent; microsculpture faint on head, otherwise normal. Head .82 & .79 width prothorax; eyes moderately large and prominent, with posterior supraocular setae just behind line of their posterior edges. Prothorax rather narrow, subcordate; width/length 1.20 & 1.23; base/apex 1.12 & 1.11; sides rather weakly arcuate for much of length, then straight and converging posteriorly, and slightly or moderately sinuate before base; posterior angles obtuse (partly because sides of base obliquely rounded), slightly blunted; lateral margins average; basal foveae rather deep, slightly roughened, and basal area between foveae slightly depressed and roughened; anterior marginal line entire or slightly interrupted at middle, posterior one rather vague. Elytra rather narrow but otherwise of about normal outline and convexity; lateral margins rather narrow; subapical sinuations absent (except as margins turn onto spines); apices each with a spine opposite 3rd interval (the spine about as long as width of 11% discal elytral intervals), with sutural angles broadly rounded, not denticulate; striae rather deep, not distinctly punctulate; intervals moderately convex, 8th and 9th not much modified toward apex. Lower surface impunctate; abdomen not pubescent. Legs: 4th hind-tarsal segment emarginate, very briefly (scarcely) lobed, outer lobe very slightly longer than inner; sole of hind tarsus as described above. Male copulatory organs: Fig. 32. Measurements: length 7.2- 7.4; width 2.3 mm. Types. Holotype o (Leiden Mus.) and 1 @ paratype (M.C.Z. No. DARLINGTON: CARABID BEETLES OF NEW GUINEA 7 28,620) both from Rattan Camp, Snow Mts., Neth. N. G., 1,200 m. (about 3,900 ft.), Feb.-Mar. 1939 (Toxopeus). Measured specumens. 'The types. Notes. This fine little species is not closely related to any other known to me. It is sufficiently characterized in the key to species of Notagonum. It is probably not related to Lorostemma, although the hind tarsi are similarly clothed below. NOTAGONUM SPINULUM N. sp. Description. With characters of genus as described above. Form of broad Agonum s. s.; black, appendages blackish, outer segments of antennae brown, lateral margins of prothorax and elytra not paler; surface moderately shining, not distinctly iridescent; microsculpture normal, light. Head .64 & .63 width prothorax; eyes large and promi- nent, with posterior supraocular setae between their posterior edges. Prothorax large and wide; width/length 1.52 & 1.59; base/apex 1.40 & 1.39; sides arcuate for most of length, usually slightly sinuate before base; posterior angles obtuse, more or less rounded; lateral margins relatively wide but less reflexed or elevated than usual; basal foveae wide, only moderately deep, only slightly roughened; anterior marginal line entire or nearly so, posterior one vague or interrupted at middle. Elytra broader than usual but otherwise of normal outline and con- vexity; lateral margins rather wide (in genus); subapical sinuations rather strong; apices each with a spine about opposite 3rd interval (spines about as long as width of 11% discal elytral intervals), then emarginate to denticulate sutural angle; striae rather lightly impressed, not distinctly punctulate; intervals flat on disc, slightly convex laterally and apically, 8th and 9th not much modified toward apex. Lower surface impunctate or nearly so; abdomen not pubescent. Legs: 4th hind-tarsal segment emarginate and with very short lobes below, outer lobe a little longer than inner. Male copulatory organs: Fig. 33. Measurements: length 7.3-8.4; width 2.8-3.3 mm. Types. Holotype o& (M.C.Z. No. 28,621) and 25 paratypes all from Dobodura, Papua, Mar.—July 1944 (Darlington), taken among dead leaves and in leaf mold on the ground in heavy rainforest. Measured specimens. The & holotype and 1 9 paratype. Notes. The relatively broad form, black color, and spined elytra distinguish this species from all other Notagonum except the following one (subspinulum), q.v. 158 BULLETIN: MUSEUM: OF COMPARATIVE ZOOLOGY NoTAGONUM SUBSPINULUM N. sp. Description. With characters of genus as described above. Form of rather broad Agonum s. s.; black, appendages brownish, lateral margins of prothorax only slightly translucent, of elytra not distinctly so; surface moderately shining, not distinctly iridescent; microsculpture normal. Head .67 & .68 width prothorax; eyes large and prominent, with posterior supraocular setae about between their posterior edges. Prothorax moderately large; width/length 1.43 & 1.48; base/apex 1.33 & 1.35; sides arcuate anteriorly, nearly straight and converging posteriorly, slightly sinuate before base; posterior angles obtuse and finely blunted but somewhat better defined than in spinulum; lateral margins moderately wide, more reflexed than in spinulum; basal foveae rather deep, not much roughened; anterior marginal line entire, pos- terior one vague or interrupted at middle. Elytra of about normal outline and convexity, usually with disc a little impressed about 14 from base; lateral margins average; subapical sinuations rather strong; apices each with a very short spine or long tooth about opposite 3rd interval (spines about as long as width of 1 discal elytral interval), then emarginate to finely denticulate sutural angle; striae rather deep, not punctate; intervals moderately convex, 8th and 9th not much modified toward apex. Lower surface impunctate; abdomen not pu- bescent. Legs: 4th hind-tarsal segment emarginate, not lobed beneath. Measurements: length 6.8; width 2.5 mm. Types. Holotype & (M.C.Z. No. 28,622) and 1 o paratype both from Dobodura, Papua, Mar.—July 1944 (Darlington). Measured specimens. ‘The types. Notes. This species is generally similar to and probably related to the preceding (spinulum). It is from the same locality; I do not know whether it is from the same ecological habitat. It differs from spinulum in being smaller, less wide, with prothorax proportionately narrower (cf. proportions given in descriptions) and with less wide and more reflexed lateral margins and somewhat better defined posterior angles, elytra much more deeply striate and with shorter apical spines, and 4th hind-tarsal segment without the short lobes which are present in spinulum. i Genus COLPODES Macl. Macleay 1825, Annulosa Javanica, p. 17. Csiki 1936, Coleop. Cat., Carabidae, Harpalinae 5, p. 745 (see for synonyms and additional references). Jeannel 1948, Coléoptéres Carabiques de la Région Malgache, Part 2, pp. 514, 515, 516. DARLINGTON: CARABID BEETLES OF NEW GUINEA 159 Diagnosis. See key to genera. Description. None required here. Genus as here restricted has inner wings always fully developed; seta-formula ++, ++, (+) (4+)+; secondary sexual characters normal except in rex (which see); male copulatory organs of several species figured (Figs. 34-40). Genotype. C. brunneus Macl. (op. cit., p. 17, Pl. 1, Fig. 3) of Java. It should be noted that the species figured as brunneus by Jeannel (op. cit., p. 514, Fig. 235a) is not that species and is not even closely related to it. True brunneus is apparently still known only from the single type, and is a peculiar, large, brown form, with small but abruptly prominent eyes and denticulate sutural angles of elytra. I have myself examined the type briefly at the British Museum, and I am indebted to Mr. E. B. Britton for additional notes on it. Generic distribution. As here limited the genus is widely distributed from the Orient to northern Australia. Very many species of other tropical areas are now assigned to the genus, but it remains to be seen whether or not they are really congeneric with the Oriental forms. Notes. It is obvious that the mass of diverse species now included in Colpodes should be broken up into genera or subgenera, and I have taken some steps toward breaking it up here, but only so far as the New Guinean forms are concerned. The eleven New Guinean species that I am leaving in Colpodes still present such a diversity of characters that I feel sure a further division will be necessary. It may well prove that none of the eleven is really congeneric with the genotype. Jeannel (loc. cit.) has offered a table of the principal Oriental genera of Anchomenini (Agonini) which is concerned largely with components of the old Colpodes. I have already discussed this table, above, in notes under the tribe Agonini. It is superficial and otherwise unsatis- factory. In my opinion Colpodes will not be finally, satisfactorily divided until a large number of forms from many different regions are carefully studied and compared. This is something I should like to do some day, but it is far too big a task to undertake here. Colpodes violaceus is a common lowland species. The other New Guinean species here assigned to the genus seem particularly charac- teristic of the lower and middle mountain slopes. In habits, they, or at least the ones that I know in life, are more arboreal or subarboreal than the species of Notagonum. Key to the Species of Colpodes of New Guinea 1. Head relatively short, with short mandibles; prothorax wide, 14 or more wider than long, with base 44 or more wider than apex; size small, about 10 mm. or less; (form broad; color purple; each elytron with an apical spine about opposite 2nd interval) (p. 160)................. violaceus 160 BULLETIN! MUSEUM OF COMPARATIVE ZOOLOGY — Head longer, with relatively longer mandibles; prothorax narrower, with relatively narrower base; size larger, except in small specimens of OAT ene MnO eon LENS a oboe, OC CREE tl aee ero pROIDAT alo Dadiloo Uv ayia eae tas 2. Wach elytron spined at outer angle (outside the subapical sinuation) as well as at sutural angle; (ength about 131% to 15 mm.; color blue or DUNG) (OAUGL) cremate warty, «cu Cuma poratnge nnn cine trnt saphyrinus sloanet Outer angles of elytra not spined, rarely angulate, usually not defined. .3 3. Pach elytron with a spine at sutural angle, the spine longer than width of a discal elytral interval; (length 11.6-13.8 mm.; form slender, depressed; color rufopiceous; elytra very deeply striate) (p. 162)........... helluo Dlytra either not spined or with spines not at sutural angles (but latter foponkoquhantes GkcvauebUEMNCD Goo ua poawouhBoudunooooveoqoonn PRATT Alo ur, A 4, Color atileast partly green, blue, or purplish. 7; . (0. os cams «einen 5 Color DIG Ck: OF PICCOUB Ws oie kysjei wietereceuaraidyein octais, sho) «ap cielo) aaa te OR REE Re 8 5. Strikingly bicolored: red, elytra blue or greenish with red apices; length AboOUtOy6=135) mime (Ds LGB). wie, cscecveke eteus en seneus oie caren yen renee laetus INKoys, wiahokss LoKerovloy flo IEW WTA o Gaus uo Goddonobuouso cou Ldoououns 6 6. Elytra not spined (but denticulate at sutural angles) and .with outer intervals not compressed toward apex; length 18-17 mm. (p. 164).... habilis Elytra edther with outer intervals compressed toward apex or with short apical spines about opposite 38rd intervals; size still larger. .......... 7 ~I Plytra with outer intervals (especially 8th) compressed to narrow carinae toward apex; apices not spined (but denticulate at sutural angles); longth, LS=2i amie. s GD) cite vet ainownnvnnkegaraan wn ot Unban in ane bennigsent Elytra with outer intervals not compressed toward apex; each elytron with a short apical spine about opposite 3rd interval; length 19-23 mm. CDS RGT)ieeciece Galrate a lhedete Weta gleitlats: ete eh toa pec Khas thr se is cnc aa rex 8. Blytra with outer angles (outside subapical sinuations) well defined, finely sub-rectangular (elytral apices also spined about opposite 8rd intervals) (a fpal (12) aie rence iui sbi rete teortdh, croc tennis OcimkG cea G Os antedens Blytra: withouter angles not defined. «0. sc. 1.) Jaa. mn vetinne encanta 9 9, BPlytra with basal margin incomplete, only 1 dorsal puncture (the posterior one) present on 8rd interval, and with apices conspicuously angulate about opposite 2nd intervals, with points of angles usually slightly DFOGUCEA (Ds: LILO) wrcisvs descentavinds ausnefole the: vy) heavy forest. Measured specimens. The o& from Kokoda, and 1 9 from Dobodura, Notes. Colpodes saphyrinus and closely related forms at least some of which are perhaps to be considered subspecies (together constituting Jeannel’s genus Nestocolpodes, which may be recognizable, though not by the characters given by Jeannel) occur in different areas from the Indo-Chinese Subregion of the Orient through most of the Indo- Australian Archipelago, at least to the Philippines and New Guinea. The New Guinean form, sloane?, is best distinguished from the other members of the group by presence of a short spine rather than a mere tooth at the outer angle of each elytron. COLPODES HELLUO N. sp. Description. Form of a large, slender, flattened Platynus or of a slender helluonine; piceous or rufo-piceous, appendages a little paler, lateral margins of prothorax and elytra only slightly translucent; surface moderately shining, not iridescent; microsculpture scarcely visible on head and disc of pronotum, more distinct and isodiametric or only slightly transverse in basal foveae and along base and sides of pronotum, still more distinct and only slightly transverse on elytra. Head .82 & .81 width prothorax; eyes moderate in size and prominence, with posterior supraocular setae about between their posterior edges; front smooth, with anterior impressions slight; antennae long, normally formed; mentum tooth triangular with vaguely emarginate apex. Prothorax more or less cordate; width/length 1.28 & 1.30; base/apex DARLINGTON: CARABID BEETLES OF NEW GUINEA 163 1.04 & 1.17; sides arcuate anteriorly, strongly sinuate well before basal angles; latter nearly right, but slightly blunted; lateral margins mod- erately wide and rather strongly reflexed; basal foveae deep, slightly wrinkled but not distinctly punctate; dise normal, impunctate; an- terior marginal line distinct and entire, posterior one less well defined but more or less entire. Elytra long, almost parallel, depressed; basal margin entire, slightly angulate at humeri; lateral margins moderate; subapical sinuations rather strong; apices bluntly subangulate about opposite 3rd striae, then briefly subtruncate, then strongly spined at sutural angles, the spines a little longer than width of a discal elytral interval; striae very deep, entire, vaguely punctulate; intervals con- vex, 8th and 9th not much modified toward apex, 3rd with usual 3 dorsal punctures, the anterior one a little farther back than usual. Lower surface impunctate or nearly so; abdomen not pubescent, prosternal process normal, simple. Legs: hind tarsi slender, with first 4 segments sulcate each side above; 4th hind-tarsal segment shallowly emarginate, not lobed; 5th hind-tarsal segment without obvious acces- sory setae below. Male copulatory organs: Fig. 35. Measurements: length 11.6-13.8; width 3.8-4.5 mm. Types. Holotype & (Leiden Mus.) and 1 2 paratype (M.C.Z., No. 28,623) both from Rattan Camp, Snow Mts., Neth. N. G., 1,150 m. (about 3,750 ft.), Feb.-Mar. 1939 (Toxopeus). Measured specimens. The types. Notes. Characters for the identification of this species are given in the key, above. In appearance it is unlike any other species of Agonini known to me. COLPODES LAETUS (Er.) Anchomenus laetus Erichson 1834 (1835), Nov. Act. Akad. Caesareae Leo- poldino-Carolinae Germanicae Naturae Curiosorum 16, Suppl. p. 222, Jk BOs LO Ae Colpodes laetus Andrewes 1930, Cat. Indian Carabidae, p. 123 (see for synonymy and additional references). Description (significant characters only). Form rather Platynus-like; color red, with elytra bright green or blue and with red apices. Head .76 & .73 width prothorax. Prothorax: width/length 1.31 & 1.30; base/apex 1.25 & 1.22. Elytra with outer intervals scarcely modified toward apex. Legs: hind tarsi with first 4 segments sulcate each side above; 4th hind-tarsal segment with a moderate outer and shorter inner lobe; 5th hind-tarsal segment without obvious accessory setae. Male copulatory organs: Fig. 36. Although it is strikingly colored, the species notably lacks striking or unusual structural characters. Measurements: length about 914-13; width about 3144-444 mm 164 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Type(s). From Luzon in the Philippines; in Zoological Mus., Berlin (seen by Andrewes, 1927). Occurrence in New Guinea. Papua: 4, Dobodura, Mar.—July 1944 (Darlington). Neth. N. G.: 1, Hollandia, May 1945 (B. Malkin, U.S.N.M.); 2, Bewani Mts., 400 m. (about 1,300 ft.), July 1937 (W. Stiiber, British Mus.); 2, Toem (Maffin Bay), Mar. 1945 (D. B. Vogtman, U.S.N.M.); 5, “Neth. New Guinea” without further locality, “jungle vege.”, 225 ft., Oct. 10, 1944 (T. Aarons), Nov. 10, 1944 (T. Aarons), and Dec.—Feb. 1945 (H. A. Levy) (all in American Mus.). The Dobodura specimens were all taken at light. Measured specimens. One o' and 1 9 from Dobodura. Notes. I have seen this species also from Luzon, Leyte, and Min- danao in the Philippines; Celebes (Andrewes Coll., British Mus.); Bougainville (M.C.Z. & U.S.N.M.) and Kulambangra (British Mus.) in the Solomons; and Espiritu Santo, New Hebrides (American Mus. and California Acad.). The coloration of this species is like that of Euplenes apicalis (above), but the Colpodes is of course much larger (about 914-13 mm.), with different generic characters (see key to genera, above). It is nearly matched in color also by Colpodes felix Andr., known from Buru and the Philippines, but fel’x is much smaller and broader, with different technical characters, probably related to C. ruficeps. C. laetus is a somewhat variable species, but I am not able to divide it into distinct subspecies. In the nine Philippine specimens now before me the elytra are bright green (except of course for the red apices). In most specimens from New Guinea, the Solomons, and the New Hebrides the main elytral color is blue rather than green, but some greenish specimens occur in the New Guinean series. Variation in proportions and in shape of prothorax seems to be more individual than geographical. In some New Guinean specimens the elytra are distinctly denticulate at sutural angles (as usual in the Philippines) but in other New Guinean specimens there is almost no trace of the denticles. Both forms occur in my short series from Dobodura. COLPODES HABILIS SI. Sloane 1907, Deutsche Ent. Zeits., pp. 178 & 179. Andrewes 1930, Treubia 7, Suppl., pp. 333 & 338. van Emden 1937, Stettiner Ent. Zeit. 98, p. 34. Description (Significant characters only). A large (about 13-17 mm.) Colpodes with greenish, bluish, or purplish elytra which are not spined but denticulate at sutural angles, with outer intervals not much modi- fied toward apex. Head .72 & .73 width prothorax. Prothorax: width/ DARLINGTON: CARABID BEETLES OF NEW GUINEA 165 length 1.34 & 1.35; base/apex 1.16 & 1.22. Legs: hind tarsi slender, with first 4 segments deeply sulcate each side above; 4th hind-tarsal segment with a moderate outer and shorter inner lobe; 5th hind-tarsal segment without obvious accessory setae, but sometimes (perhaps always) with vestigial ones. Male copulatory organs: Fig. 37. Measure- ments: length about 13-17; width about 414-6 mm. Type. From Sattelberg, N-E. N. G.; presumably in Deutsches Entomologisches Institut, Berlin-Dahlem, Germany. Occurrence in New Guinea. Papua: 1, Mafulu, 4,000 ft., Dec. 1933 (Cheesman); 1, Mondo, 5,000 ft., Jan.Feb. 1934 (Cheesman); 2, Palmer River at Black River, June 7-14, 1936 (Archbold Exp., Ameri- can Mus.); 1, Mt. Mabiom, July 15, 1936 (Archbold Exp., American Mus.). N-E. N. G.: 1, Sattelberg (topotype) (British Mus.). Neth. N. G.: 2, Hollandia, 250 ft., Nov. 11, 1944 & May 1945 (H. Hoog- straal, M.C.Z.); 1, same locality, 300-600 m. (about 975-1,950 ft.), Jan. 1938 (W. Stiiber, British Mus.); 3, Bewani Mts., 400 m. (about 1,300 ft.), July & Sept. 1937 (W. Stiiber, British Mus.); 1, Pukusam Dist. (W. of Tami River), June 1937 (W. Stiiber, British Mus.); 1, Mt. Cyclops, 3,500 ft., Mar. 1936 (Cheesman); 21, Araucaria Camp, Snow Mts., 800 m. (about 2,600 ft.), Mar. 1939 (Toxopeus); 24, Rattan Camp, Snow Mts., 1,150 & 1,200 m. (about 3,750 & 3,900 ft.), Feb.- Mar. 1939 (Toxopeus); 1, Sigi Camp, Snow Mts., 1,500 m. (about 4,875 ft.), Feb. 1939 (Toxopeus); 2, Mist Camp, Snow Mts., 1,800 m. (about 5,850 ft.), Jan. & Feb. 3, 1939 (Toxopeus); 10, Mt. Baduri, Japen Is., 1,000 ft., Aug. 1938 (Cheesman); 1, Wasian, Sept. 1939 (R. G. Wind, M.C.Z.); 2, Mt. Nok, Waigeo Is., 2,500 ft., May 1938 (Cheesman). Measured specimens. One pair (co Q ) from Rattan Camp. Notes. This species is recorded also from Buru (Andrewes 1930); I have seen specimens from Cape Gloucester, New Britain (Darling- ton), and from Guadalcanar Is., Solomons (J. A. Kusche, Bishop Mus.); and it is recorded also from Vanikoro, Santa Cruz Islands (van Emden 1937). COLPODES BENNIGSENI SI. Sloane 1907, Deutsche Ent. Zeits., pp. 177 & 179. Colpodes louwerensi Andrewes (new synonym). Andrewes 1937, Bull. Ann. Soc. Ent. Belgique 77, pp. 39 & 41. Louwerens 1949, Tijd. v. Ent. 90, p. 45. Description (significant characters only). Very large (about 18-21 mm.); color black, elytra with strong purple or green reflections which vary somewhat in different individuals and more in different lights, 166 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY appendages dark. Head .82 & .84 (.81) width prothorax. Prothorax rather narrow and narrowly margined; width/length 1.17 & 1.19 (1.19); base/apex 1.09 & 1.11 (1.08) if base measured between pos- terior-lateral setae, but base actually slightly narrower than apex if measured between apparent (but rounded) basal angles, which are behind the setae; sides moderately arcuate anteriorly, faintly or not distinctly angulate at anterior-lateral setae, nearly straight and rather strongly converging in more than posterior half; anterior and posterior marginal lines entire. Elytra rather ample, with weak subapical sinu- ations, strongly rounded lobes or blunt angulations at apex about opposite 3rd intervals, and denticles at sutural angles; 3rd interval with 3 dorsal punctures but anterior one a little farther forward and others farther backward than normal, the posterior one being very far back, on the declivity; 7th interval toward apex much narrowed but somewhat variable, 8th toward apex compressed to a narrow costa, and 9th narrow and much interrupted by ocellate foveae. Legs: hind tarsi with first 4 segments very deeply impressed on each side (and so 3-carinate) above; 4th hind-tarsal segment with a moderate outer and shorter inner lobe; 5th hind-tarsal segment without obvious accessory setae but with minute vestigial ones. Measurements: length about 18-21; width about 614-734 mm. Types. That of bennigseni from Sattelberg, N-E. N. G., should be in the Deutsches Entomologisches Institut, Berlin-Dahlem, Germany. That of louwerensi, from Bali, is in the Andrewes Collection in the British Museum, where I have seen it. Occurrence in New Guinea. Papua: 1, Mafulu, 4,000 ft., Dec. 1933 (Cheesman); 1, Mondo, 5,000 ft., Jan._Feb. 1934 (Cheesman). Neth. N. G.: 2, Humboldt Bay Dist., 1937 (W. Stiiber, British Mus.); 1, Bewani Mts., Sept. 1937 (W. Stiiber, British Mus.); 1, Pukusam Dist. (W. of Tami River), June 1937 (W. Stiiber, British Mus.); 1, Hollandia, Jan.—Mar. 1939 (Toxopeus); 12, Araucaria Camp, Snow Mts., 800 m. (about 2,600 ft.), Mar. 1939 (Toxopeus); 58, Rattan Camp, Snow Mts., 1,150 & 1,200 m. (about 3,750 & 3,900 ft.), Feb.—Mar. 1939 (Toxopeus) ; 1, Mist Camp, Snow Mts., 1,800 m. (about 5,850 ft.), Jan. 1939 (Toxopeus) ; 4, Mt. Baduri, Japen Is., 1,000 ft., Aug. 1938 (Cheesman). Measured specimens. One pair (oi 2) from Rattan Camp, and in parentheses a Q cotype of lowwerensi from Bali. Notes. Outside of New Guinea this species is known from Bali (types of lowwerenst); Java (Louwerens 1949); Celebes (co? @ received from Louwerens); and Malaita, Solomon Islands (1 9, American Mus.). I have carefully compared a © cotype of louwerenst from Bali with New Guinean specimens of bennigseni and find no differences that seem to be of even subspecific value. DARLINGTON: CARABID BEETLES OF NEW GUINEA 167 COLPODES REX N. sp. Description. Very large; near average form and convexity for genus; black, pronotum and elytra green or blackish-green or purple- green (varying a little in different individuals but more at different angles), appendages dark; lateral margins of pronotum and elytra invaded by metallic color, not translucent; surface rather shining, not iridescent; microsculpture indistinct on head (but latter with a very little fine, sparse punctulation), light and transverse on pronotum, very fine and transverse on elytra. Head .78 & .77 width prothorax; eyes large and prominent, with posterior supraocular setae a little in front of line of posterior edges of eyes; front smooth, with anterior impressions rather shallow and irregular; antennae not very long (in genus), normally formed; mentum tooth triangular. Prothorax sub- hexagonal; width/length 1.31 & 1.24; base/apex 1.05 & .99 if base measured across posterior-lateral setae, but slightly less than .90 if base measured across apparent posterior angles, which are behind the setae; sides slightly arcuate and converging forward anteriorly, sub- angulate before middle (at anterior-lateral setae), then nearly straight and strongly converging backward, and sometimes slightly sinuate before posterior angles; latter obtuse, narrowly rounded; lateral mar- gins rather narrow especially posteriorly, moderately reflexed; basal foveae not distinct from posterior ends of lateral margins, impunctate; dise with light median and deeper transverse impressions, impunctate; anterior and posterior marginal lines entire, well marked. Elytra long, moderately wide, convex, with sides subparallel, only faintly arcuate for most of length; basal margin entire, rounded at humeri; lateral margins narrow; subapical sinuations slight or absent; apices each with a short spine about opposite 3rd stria, then emarginate to den- ticulate sutural angle; striae fine and light, very finely punctulate Ist deeper at base and apex, 5th in a depression toward base; intervals flat, scarcely modified toward apex except that Sth tends to overhang 9th at outer posterior curve of elytron; 3rd interval almost normally 3-punctate except posterior puncture farther back than usual, behind apical 14. Lower surface impunctate; abdomen not pubescent; pros- ternal process simple. Legs: posterior tibiae not sulcate along outer edges; posterior tarsi with first 4 segments above lightly but dis- tinctly sulcate on outer side but less distinctly or not suleate on inner side especially in o; 4th hind-tarsal segment rather deeply emarginate but hardly lobed, inner and outer apical angles nearly equal; 5th hind- tarsal segment without obvious accessory setae. Secondary sexual characters normal except @ usually with 3 (instead of 2) setae each side last ventral segment. Measurements: length 19-23; width 6.5- 7.7 mm. 168 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Types. Holotype o (British Mus.) and 10 paratypes (British Mus., & M.C.Z., No. 28,624) from Bewani Mts., Humboldt Bay Dist., Neth. N. G., Sept. (some paratypes July) 1937 (W. Stiiber). Addi- tional paratypes as follows: Neth. N. G.: 1, Pukusam Dist. (W. of Tami R.), June 1937 (W. Stiiber, British Mus.); 1, Mt. Cyclops, 3,500 ft., Mar. 1936 (Cheesman); 1, Araucaria Camp, Snow Mts., 800 m. (about 2,600 ft.), Mar. 1939 (Toxopeus); 18, Rattan Camp, Snow Mts., 1,150 & 1,200 m. (about 3,750 & 3,900 ft.), Feb.—Mar. 1939 (Toxopeus); 2, Sigi Camp, Snow Mts., 1,500 m. (about 4,875 ft.), Feb. 22, 1939 (Toxopeus); 20, Mist Camp, Snow Mts., 1,800 m. (about 5,850 ft.), Jan. 1939 (Toxopeus); 1, Top Camp, Snow Mts., 2,100 m. (about 6,825 ft.), Jan. 22, 1939 (Toxopeus); 2, Mt. Baduri, Japen Is., 1,000 ft., Aug. 1938 (Cheesman); 3, Mt. Nok, Waigeo Is., 2,500 ft., May 1938 (Cheesman). N-E. N. G.: 1, Wamoro (not located on map) (British Mus., marked ‘‘Colpodes sp. nov.?” by Andrewes). Papua: 1, Mafulu, 4,000 ft., Dec. 1933 (Cheesman); 4, Fly R. 5 mi. below Palmer R., May 14-31, 1936 (Archbold Exp., American Mus.); 1, Palmer R. at Black R., June 15-22, 1936 (Archbold Exp., American Mus.); 1, Mt. Mabiom, July 15, 1936 (Archbold Exp., American Mus.). The records suggest that this is a species of the lower and middle mountain slopes, and that it commonly extends to higher altitudes than habilis or bennigseni. Measured specimens. The o holotype and 1 @ paratype from the Bewani Mts. Notes. This magnificent species is apparently related to the pre- ceding (bennigseni), which it resembles in large size and striking ap- pearance and in at least two significant technical characters: the posi- tion of the apparent posterior angles of the prothorax (behind the posterior-lateral setae) and the position of the posterior punctures of the 3rd elytral intervals (unusually far back, though not so far back as in bennigseni). However rex differs from bennigsent not only in such specific characters as presence of metallic color on pronotum as well as elytra, relatively narrower head, more hexagonal prothorax, finer elytral striae, and presence of apical elytral spines, but in certain other characters which are surprising if the species are really closely related. For example, the outer elytral intervals especially near apex are strongly compressed in bennigseni but not in rex; and the hind tarsi are strongly sulcate on both sides above in bennigseni but only lightly so or sometimes not distinctly sulcate at all on the inner side in rex. DARLINGTON: CARABID BEETLES OF NEW GUINEA 169 COLPODES ANTEDENS N. sp. Description. Form almost of Agonum (s. s.) but larger, more elongate, with relatively longer head and spined elytra; black, ap- pendages dark; lateral margins of prothorax and elytra vaguely or not translucent; surface moderately shining, not iridescent; micro- sculpture of head faint, nearly isodiametric; of pronotum faint; of elytra more distinct, of rather large, only slightly transverse meshes. Head .70 width prothorax; eyes moderately large and prominent; genae short, oblique, not prominent; posterior supraocular setae about between posterior edges of eyes; front smooth, impressed on each side anteriorly; antennae rather long and slender, normally formed; mentum tooth narrowly triangular with slightly blunted apex. Pro- thorax: width/length 1.40; base/apex 1.21; sides evenly arcuate for most of length, almost straight and converging before basal angles; latter obtuse but only slightly blunted; lateral margins rather wide and rather strongly reflexed; basal foveae deep, vaguely roughened or subpunctate; disc with usual impressions, impunctate; anterior and posterior marginal lines entire. Elytra rather long, normally convex, with sides only slightly arcuate for most of length; basal margin entire, obtusely angulate at humeri; lateral margins moderate; outer angles distinct, almost right; subapical sinuations emarginate; apices spined opposite ends of 3rd intervals, then obliquely emarginate to slightly obtuse (nearly right) sutural angles; spines slightly longer than width of 1 discal interval; striae moderately deep, finely punctulate; intervals convex, 8th and 9th not much modified toward apex; 3rd normally 3- punctate. Lower surface impunctate; abdomen not pubescent; pros- ternal process simple. Legs: hind tibiae not sulcate along outer edges; hind tarsi slender, lightly sulcate each side above; 4th hind-tarsal segment deeply emarginate, outer and inner angles prominent, about equal; 5th hind-tarsal segment with short and inconspicuous but dis- tinct accessory setae below. Male copulatory organs as figured (Fig. 38). Measurements: length 16.6; width about 6 mm. Type. Holotype & (British Mus.) from Mt. Lina, Cyclops Mts., Neth. N. G., 3,500 ft., Mar. 1936 (Cheesman); unique. Measured specimen. The type. Notes. This species should be easily recognized by characters given in the key to species of Colpodes, above. It resembles the following ones (acuticauda, sinuicauda, simplicicauda) in being rather large and black, but I am not sure it is really related to any of them, nor am J sure that the latter are related among themselves. It (antedens) 1s notable not only for having the outer angles of the elytra well defined, but also for possessing distinct, though very small, accessory setae on the 5th hind-tarsal segments. 170 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY COLPODES ACUTICAUDA N. sp. Description. Large; black, appendages black or piceous; surface rather dull, faintly silky but not iridescent; microsculpture fine but distinct, isodiametric on head, only slightly transverse on pronotum and elytra; upper surface also faintly, sparsely punctulate. Head .76 & .77 width prothorax; eyes smaller than usual in genus but prominent; genae angulately prominent in profile; posterior supraocular setae about between posterior edges of eyes; vertex transversly impressed; front a little irregular, with very slight anterior impressions; antennae of moderate length, normally formed; mentum tooth triangular, with apex rounded. Prothorax: width/length 1.38 & 1.44; base/apex 1.07 & 1.06; sides arcuate for most of length, variably sinuate before some- what obtuse but well defined posterior angles; lateralmarginsmoderate, rather strongly reflexed especially posteriorly; basal foveae moderate or rather shallow, not punctate; disc less convex than usual, im- punctate; anterior and posterior marginal lines vague or obsolete. Elytra of about normal proportions, slightly more convex than usual; sides subparallel for much of length; basal margin incomplete, ob- literated inwardly from bases of 5th or 4th striae; lateral margins rather narrow; outer angles not defined; subapical sinuations almost absent; apices strongly angulate opposite 2nd intervals, with angles usually slightly produced into acute denticles; sutural angles not de- fined (unless the aforementioned angulations are sutural angles dis- placed outward); striae rather deep, impunctate; intervals moderately convex, 8th narrowed and longitudinally impressed toward apex, 9th widened and irregular toward apex; 3rd interval with only 1 dorsal puncture, far back, at top of declivity. Lower surface impunctate; abdomen with a little fine pubescence irregularly distributed ; prosternal process simple. Legs: posterior tibiae not sulcate along outer edges; posterior tarsi slender, sulcate each side above; 4th hind-tarsal segment deeply emarginate, outer angle forming a very short lobe, slightly longer than inner one; 5th hind-tarsal segment without obvious ac- cessory setae below. Male copulatory organs: Fig. 39. Measurements: length 17.6-18.3; width about 6 or slightly more mm. Types. Holotype 2 (British Mus.) and 1 @ paratype (M.C.Z., No. 28,625) from Mt. Tafa, Papua, 8,500 ft., Mar. 1934 (Cheesman) ; and 1 o paratype from Top Camp, Snow Mts., Neth. N. G., 2,100 m. (about 6,825 ft.), Jan. 22, 1939 (Toxopeus). Measured specimens. The 2 holotype and o& paratype. Notes. This very distinct new species has several characters (form of eyes and genae, reduction of basal margin of elytra, reduction of dorsal punctures of elytra, dullness of surface) which suggest that it may be near the ancestral stock of Idiagonwm (new genus described DARLINGTON: CARABID BEETLES OF NEW GUINEA 17] below), but it is still much more of a Colpodes than an Idiagonum, with eyes still only a little modified, one dorsal elytral puncture still remaining, no added 10th elytral interval, no setae on prosternal process, and fully developed inner wings. COLPODES SINUICAUDA N.. sp. Description. Rather large; black, appendages dark, lateral margins of prothorax slightly and of elytra scarcely translucent; surface moder- ately shining, not iridescent; microsculpture rather light, isodiametric on head, transverse on pronotum, very fine and transverse on elytra. Head .79 & .79 width prothorax; eyes much smaller and somewhat less prominent than usual in genus; genae long, oblique, nearly straight but with a slight sinuation (in profile) posteriorly; posterior supra- ocular setae a little behind line of posterior edges of eyes; front nearly smooth, a little irregular and slightly impressed each side anteriorly; antennae moderately long, normally formed; mentum tooth absent in both specimens. Prothorax: width/length 1.33 & 1.43; base/apex 1.06 & 1.06; sides moderately or strongly arcuate for much of length, strongly sinuate well before slightly acute basal angles; lateral margins moderate, moderately reflexed; basal foveae average, impunctate; disc normal, impunctate; anterior and posterior marginal lines present but vague. FHlytra of normal proportions, normally convex, with sides slightly arcuate; basal margin entire, faintly angulate at humeri; lateral margins moderately wide; outer angles not defined; subapical sinuations strong; apices strongly rounded or subangulate about oppo- site 3rd intervals, then subtruncate to more or less plainly denticulate sutural angles; striae well impressed, not distinctly punctulate; inter- vals slightly convex, 8th and 9th not much modified toward apex; 3rd interval normally 3-punctate. Lower surface nearly impunctate, but abdomen with a little fine scattered punctation (apparently with- out pubescence) especially on segments 3 & 4; prosternal process simple. Legs: posterior tibiae not sulcate along outer edges; posterior tarsi apparently stouter in o than in Q, not distinctly sulcate above; 4th hind-tarsal segment with rather long outer and shorter inner lobe; 5th hind-tarsal segment without obvious accessory setae below. Male copulatory organs: Fig. 40. Measurements: length 13.6-14.8; width 4.8-5.3 mm. Types. Holotype &@ (Leiden Mus.) from Sigi Camp, Snow Mts., Neth. N. G., 1,500 m. (about 4,875 ft.), Feb. 24, 1939 (Toxopeus); and 1 2 paratype (M.C.Z., No. 28,626) from Lower Mist Camp, Snow Mts., 1,700 m. (about 5,525 ft.), Jan. 17, 1939 (Toxopeus). Measured specimens. The types. 2 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Notes. Besides the key characters, which are distinctive enough, this species possesses one characteristic which, so far as I know, is unique in Colpodes even in the broadest sense: absence of the usual mentum tooth. However, I do not want to stress this too much. The tooth may possibly be broken off, although that this should have happened in the same way in both specimens seems unlikely. COLPODES SIMPLICICAUDA N. sp. Description. Very large; black, appendages dark, lateral margins of prothorax and elytra vaguely or not translucent; surface moderately shining, not iridescent; microsculpture indistinct on head, light and moderately transverse on pronotum, heavier but only moderately transverse on elytra; surface of head, pronotum, and elytra also with very fine, sparse, inconspicuous punctulation. Head .81 & .79 width prothorax; eyes much smaller and less prominent than usual in genus; genae long, oblique, slightly sinuate in profile (transversely swollen in side view) posteriorly; posterior supraocular setae slightly behind line of posterior edges of eyes; head rather deeply transversely impressed posteriorly; front slightly, irregularly wrinkled and very slightly impressed each side anteriorly; antennae moderately long, normally formed; mentum tooth triangular with apex blunted or vaguely emarginate. Prothorax: width/length 1.31 & 1.37; base/apex .96 & 1.02; sides moderately or weakly arcuate for much of length, broadly or moderately sinuate before nearly right, well formed basal angles; lateral margins moderately wide, normally refiexed; basal foveae normal, impunctate; disc normal, slightly and superficially trans- versely wrinkled, impunctate (except for fine, sparse punctulation mentioned above); anterior and posterior marginal lines entire, well impressed. Elytra rather long, of normal width and convexity, nearly parallel-sided; basal margin entire, not distinctly angulate at humeri; lateral margins rather narrow; outer angles not defined; subapical sinuations weak; apices nearly conjointly rounded, with sutural angles slightly dehiscent, not denticulate; striae moderately im- pressed, not distinctly punctate; intervals slightly convex, 8th rather strongly narrowed and more convex toward apex and 9th slightly widened toward apex (but these intervals still not very much modified toward apex); 3rd interval normally 3-punctate, but posterior punc- ture rather far back, at top of declivity. Lower surface impunctate; abdomen not pubescent; prosternal process normal. Legs: posterior tibiae not or vaguely suleate along outer edges; posterior tarsi rather lightly suleate each side above (inner sulcus sometimes almost obliterated, especially in co”); 4th hind-tarsal segment with rather long DARLINGTON: CARABID BEETLES OF NEW GUINEA 173 outer and shorter inner lobe; 5th hind-tarsal segment without obvious accessory setae below. Measurements: length about 17-21; width 5.7-7.0 mm. Types. Holotype o (Leiden Mus.) and 10 paratype (M.C.Z., No. 28,627) from Ibele (Iebele) Camp, Snow Mts., Neth. N. G., 2,250 m. (about 7,325 ft.), Nov.-Dec. 1938 (Toxopeus); 5 paratypes (all 2 9) from Mist Camp, Snow Mts., 1,800 m. (about 5,850 ft.), various dates in Jan. 1939 (Toxopeus); 12 paratype from Sigi Camp, Snow Mts., 1,500 m. (about 4,875 ft.), Feb. 17, 1939 (Toxopeus); and 1@_ para- type from Saiko (Bubu R., Upper Waria R.), N-E. N. G., 5,500- 6,000 ft., Sept.-Oct. 1936 (F. Shaw-Mayer, British Mus.). Measured specuomens. The o& holotype, and 19 paratype from Mist Camp. Notes. This species should be easily identifiable by characters given in the key to species of Colpodes, above. It lacks the conspicuous or unusual structural peculiarities which define, in different ways, the three preceding species. PLICAGONUM new genus Diagnosis. See key to genera. — Description. Rather large (14.3-18.5 mm.), somewhat Platynus- like (Fig. 3), brownish forms with fully developed wings and all usual supraocular and pronotal setae; surface above virtually impunctate except for fine punctulation on head and pronotum; microsculpture absent on head, almost absent on pronotum, rather fine but deeply impressed and not or only slightly transverse on elytra. Head only moderately elongate, with large prominent eyes; posterior supraocular setae at or in front of line of posterior edges of eyes; front strongly but irregularly longitudinally wrinkled between eyes; anterior frontal im- pressions slight; antennae rather slender, normal in structure; mentum tooth triangular with apex more or less (irregularly) blunted. Pro- thorax normal. Elytra margined at base; fully striate; 3rd interval normally 3-punctate except anterior puncture missing on one or both sides in several individuals: 8th & 9th intervals not much modified toward apex; subapical sinuations weak; apices variable as described in species. Inner wings fully developed. Lower surface virtually im- punctate; abdomen not pubescent; prosternal process normal. Legs normally formed; hind tibiae not sulcate along outer edges; hind tarsi slender, sulcate each side above (but sulci, especially inner one, some- times faint); 4th hind-tarsal segment deeply emarginate, with short lobes, outer one scarcely longer than inner; 5th hind-tarsal segment without obvious accessory setae; claws simple; sole clothed with 174 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY numerous setae not in regular rows but with middle of sole narrowly bare. Secondary sexual characters normal. Male copulatory organs as figured (Figs. 41 & 42). Genotype. Plicagonum fulvwm n. sp. (below). Generic distribution. Mountains of New Guinea. Notes. This genus, though not very different from Colpodes even as here restricted and perhaps actually of only subgeneric value, is distinguished especially by the longitudinal wrinkling of the head, and is a natural group which has undergone more or less radiation at rather high altitudes in New Guinea. I here describe only two forms of the genus, but my few scattered specimens of rugifrons are so variable as to suggest that this species will be found to break up into distinct subspecies on different mountain ranges. Key to the Species of Plicagonum Elytral apices spined or at least acutely angulate about opposite 3rd intervals (Goya 7: anenaeen sear ee een i Pr IRCA EES enG c'o 60.0.0 « rugifrons Elytral apices neither spined nor acutely angulate opposite 8rd intervals (sie (5) ereanrs Sebo ene rece orange Dat UGMN OUICe o:5'9'5-5 0 0 fuluum PLICAGONUM RUGIFRONS Nn. Sp. Description. With characters of genus as described above. Brownish- piceous, elytra usually (not always) browner with piceous apices and sometimes faintly striped (with striae piceous and middles of intervals brown), appendages brown, lateral margins of prothorax moderately translucent. Head .72 & .73 width prothorax. Prothorax: width/ length 1.40 & 1.40; base/apex 1.22 & 1.20; sides variably arcuate anteriorly, nearly straight and converging and sometimes faintly sinuate posteriorly; basal angles moderately to very obtuse; lateral margins rather wide but only slightly reflexed (a little more so toward base); basal foveae rather small but deep, sometimes slightly wrinkled but not punctate; disc with very light median longitudinal line and deep anterior and posterior transverse impressions, impunctate; anterior and posterior marginal lines entire, deeply impressed. Elytra rather ample, more convex than usual in Colpodes, nearly parallel- sided (slightly narrowed anteriorly); basal margin rounded or faintly angulate at humeri; lateral margins narrow; apices each with an acute tooth or short spine opposite 3rd interval, then emarginate to sutural angle; latter vaguely or distinctly denticulate; striae not deeply impressed, not distinctly punctulate; intervals nearly flat to slightly convex. Male copulatory organs: Fig. 41. Measurements: length about 17.5-18.5; width 6.1-6.4 mm. DARLINGTON: CARABID BEETLES OF NEW GUINEA 175 Types. Holotype & (Leiden Mus.) and 2 2 @ paratypes (1 in M.C.Z., No. 28,628) all from Top Camp, Snow Mts., Neth. N. G., 2,100 m. (about 6,825 ft.), Jan. 22 (type), Jan. 26, & Feb. 2, 1939 (Toxopeus). Other material. One 2, Mt. Misim, Morobe Dist., N-E. N. G. (Stevens, M.C.Z.); and 19, Mt. Tafa, Papua, 8,500 ft., Mar. 1934 (Cheesman). Measured specimens. The o holotype and 12 paratype. Notes. This species is defined and discussed in the key and under the generic description, above. PLICAGONUM FULVUM N. sp. Description. With characters of genus as described above. Form as figured (Fig. 3); brownish- or reddish-piceous, elytra typically rather pale-brown with darker apices and sometimes also with stripes along striae slightly darker (Moss Forest Camp specimens have elytra less contrastingly pale); appendages light-brown; lateral margins of pro- thorax broadly translucent. Head .76 & .79 width prothorax; front in types a little less wrinkled especially at middle than in rugzfrons (but strongly wrinkled in Moss Forest Camp specimens of fulvwm). Prothorax: width/length 1.35 & 1.88; base/apex 1.14 & 1.22; sides moderately arcuate anteriorly, nearly straight and converging and usually slightly sinuate posteriorly; otherwise as in rugifrons. Elytra as in rugifrons except apices only more or less rounded-prominent about opposite 3rd intervals, then obliquely subtruncate or slightly emarginate to vaguely or not denticulate sutural angles. Male copula- tory organs as figured (Fig. 42). Measurements: length (types) 14.3- 15.6; width (types) 5.1-5.8 mm. (Moss Forest Camp specimens about 18.5 mm. long). Types. Holotype o (Leiden Mus.) and 9 paratypes (some in M.C.Z., No. 28,629) from Top Camp, Snow Mts., Neth. N. G., 2,100 m. (about 6,825 ft.), Jan. 20-Feb. 8 (holotype Jan. 22), 1939 (Toxopeus). Other material. Two, referred to in description above, from Moss Forest Camp, Snow Mts., 2,800 m. (about 9,100 ft.), Oct. 9-Nov. 5, 1938 (Toxopeus). Measured specimens. The o& holotype and 12 paratype. Notes. Sufficiently characterized in the key to species, above. The Moss Forest specimens may represent a different subspecies or species, but more material is needed to prove it. 176 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY LITHAGONUM new genus Diagnosis. Generally similar to Notagonum (of which see descrip- tion) except in the following characters. Posterior-lateral setae of pronotum absent. Much of upper surface of head and pronotum, elytral striae, and much of lower surface rather closely and more or less coarsely punctate; upper surface without reticulate micro- sculpture except (in some forms only) at sides and apices of elytra. Male usually with 2 or more (not 1) and @ usually with 4 or more (not 2) setae each side last ventral segment. Description. See that of single species, below. Genotype. Lithagonum annulicorne dilutior n. subsp., below. Generic distribution. Much or all of New Guinea in suitable habitats. Notes. The single species of this genus, with five subspecies, is more fully described below. It is perhaps derived from Notagonum. It is somewhat similar superficially to Notagonwm subnigrum Darl., which it somewhat resembles in habits too, but in my experience Lithagonum is even more strictly confined to cobble-stone and similar areas on the banks and bars of rather large streams. Different populations of Lithagonum are probably more or less isolated from each other by this habitat restriction, and this may account for the existence of the several distinct subspecies here described. LITHAGONUM ANNULICORNE (Maindr.) Colpodes annulicornis Maindron 1908, Nova Guinea 5, p. 297. Ibid. 1908, Bull. Soc. Ent. France, p. 185. Description (species as a whole). Form (Fig. 4) rather Platynus-like, but head relatively large, prothorax rather small-cordate, and elytra rather wide and convex; black, appendages brown or piceous; upper surface extensively punctate, but polished between the punctures, without reticulate microsculpture except sometimes at sides and apices of elytra. Head: relative widths given under subspecies; eyes rather large and prominent; both pairs of supraocular setae present, posterior ones about between posterior edges of eyes; front normally convex, with rather deep anterior impressions; surface rather closely punctate especially across base and at sides, with middle of front much less or not punctate; antennae rather long and slender, normal in structural details; mentum tooth triangular, more or bless blunted at apex. Prothorax narrowly cordate; proportions given under subspecies; sides arcuate anteriorly, strongly sinuate well before right or acute basal angles; lateral margins narrow, with usual anterior-lateral setae about DARLINGTON: CARABID BEETLES OF NEW GUINEA N77 26 from apex but no posterior-lateral ones; basal foveae small, shallow, sometimes more or less linear, rugosely punctate, and areas between and before them also more or less closely punctate; dise with median area less closely or not punctate, with usual median and slight trans- verse impressions; basal and apical marginal lines poorly defined or absent. Elytra rather broad and convex, varying in length in different subspecies; sides subparallel (often slightly arcuate) at middle; humeri broadly rounded; base margined, margin usually slightly angulate at humeri; lateral margins moderate; subapical sinuations moderate, broad, each ending in a prominent angulation, tooth, or short spine about opposite 4th interval, then subtruncate or more or less deeply emarginate to sutural angle; latter usually (not always) more or less denticulate; striae deep, entire except some very slightly abbreviated at base, more or less coarsely punctate; intervals convex, 8th and 9th not much modified toward apex; 3rd interval normally 3-punctate. Inner wings fully developed. Lower surface with all sterna and base of abdomen closely and more or less coarsely punctate; abdomen variable (slightly to extensively), rather sparsely pubescent; prosternal process simple. Legs: hind tibiae not sulcate along outer edges; hind tarsi slender, lightly or not distinctly suleate each side above; 4th hind-tarsal segment deeply emarginate, with short outer lobe and subequal or slightly shorter inner one; 5th hind-tarsal segment without obvious accessory setae; claws simple. Secondary sexual characters normal except o with usually 2, 2 with usually about 4 setae each side last ventral segment. Male copulatory organs as figured (for subsp. dilutior, Fig. 43). Measurements: see under subspecies. Types. From “Tana Mera’, “Jamur supérieur’, and “‘Moso’, Neth. N. G.; 4 examples in all, probably now in Paris Mus. I here restrict the type locality to the first one named, commonly spelled Tanahmerah (Bay), on the north coast at the west end of the Cyclops Mts. (Moso is just east of the Tami River, which is not far east of the Cyclops Mts. ‘“Jamur supérieur” is probably far to the west, in the vicinity of Lake Jamur.) Occurrence in New Guinea. See under subspecies. Measwred specimens. See under subspecies. Notes. It should be noted that, although the exact form of the elytral apices varies in different subspecies, it varies somewhat also in individuals from single localities. It should not be used as a subspecific character except after examination of series. The habitat of this species has been described under the genus, above. After the preceding discussion of types was written, I have examined one apparent cotype of annulicorne, sent by the Paris Museum. Un- fortunately the locality of this specimen is not clear, but it appears to 178 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY be “B. Jamoer’’, and the date of collection is 5.VIII.1903. This may be the cotype originally recorded from ‘‘Jamur supérieur’. In its characters (elytra with 9th intervals not reticulate and apices with relatively well developed spines opposite 3rd intervals and relatively short sutural denticles) this specimen agrees better with politior than with any other form here recognized, but it may well be a different, undescribed subspecies. However this may be, I see no reason now to change the arrangement of subspecies proposed below. Key to Subspecies of Lithagonum annulicorne (Maindr.) 1. Legs brown; (at least 9th and usually also 8th elytral intervals with distinct reticulate microsculpture); (Papua); (p. 178)... 222.) ee. ee dilutior —' Whegs darker, piceous.’. 0) 52... 0. bons aes Sek oa cee ne eee 2 2. No reticulate microsculpture (or rarely faint traces of it) even on 9th intervals atemiddle of length (N-EeINaG) oe. . - ee eee 3 — At least 9th intervals of elytra (and often also 8th ete.) with distinct reticulate microsculpture; (Neth: (N2G))2 5 oe. a. aoe eee 4 3. Stouter; apices of elytra with relatively prominent denticles or short spines; (Markham! Vialleyiic&=vics))\(os 179) eee a eee eee politior — More slender; apices of elytra with less prominent denticles; (Bismarck Range) a(n FO) as ocak careers. tack ate pan cae ara ee eae bismarckense 4, Stouter; elytral apices with outer denticles more prominent; elytral striae coarsely punctate; (Cyclops Mts. & vic.) (p. 180)...... annulicorne s. 8. — More slender and depressed; elytral apices with outer denticles usually less prominent (but variable); elytral striae more finely punctate; (vic. of Snow? Mits!)\(p2 180) se Fa. ee ed ee baliem LITHAGONUM ANNULICORNE DILUTIOR n. subsp. Description. Form as figured (Fig. 4). Smaller than other subspecies. Legs brown (not blackish). Head .90 & .87 width prothorax. Pro- thorax: width/length 1.20 & 1.22; base/apex 1.03 & 1.03. Elytra rela- tively short (in species) and usually a little more oval (with more arcuate sides) than in other subspecies; apices variable, usually com- paratively weakly emarginate between denticles, and latter usually comparatively small, with outer ones often no more prominent than sutural ones (but in some individuals outer ones are slightly or much more prominent than sutural ones); striae moderately coarsely punc- tate; at least 9th, almost always Sth, and sometimes less distinctly 7th etc. intervals with somewhat irregular, somewhat transverse re- ticulate microsculpture (microsculpture almost lacking on 8th though distinct on 9th intervals in one individual from Kokoda). Male copu- latory organs as shown in Fig. 43. Measurements: length 8.0-9.1; width 3.1-3.5 mm. DARLINGTON: CARABID BEETLES OF NEW GUINEA 179 Types. Holotype & (M.C.Z. No. 28,630) and 21 paratypes from Dobodura, Papua, Mar.—July 1944 (Darlington). Additional para- types from Papua as follows: 4, (near) Oro Bay, Dec. 1943 (Darling- ton); 31, Kokoda, 1,200 ft., Aug. 1933 (Cheesman). These three lo- calities are within about 40 miles of each other, all on north-flowing drainage systems. Measured specimens. The & holotype and 1 @2 paratype from Dobodura. LITHAGONUM ANNULICORNE POLITIOR n. subsp. Description. Legs blackish. Head .83 & .84 width prothorax. Pro- thorax: width/length 1.18 & 1.17; base/apex 1.06 & .98. Elytra rela- tively short, slightly suboval (with sides slightly, variably arcuate); apices rather strongly emarginate between denticles, and outer denticles usually (not always) much more prominent than sutural ones; striae moderately coarsely punctate; all intervals including 8th and 9th without reticulate microsculpture at least at middle of length (1 speci- men from Nadzab has 9th intervals faintly reticulate). Measurements: length 9.1-10.2; width 3.4-4.0 mm. Types. Holotype & (M.C.Z. No. 28,631) and 20 paratypes from Nadzab, lower Markham Valley, N-E. N. G., July 1944 (Darlington); and 4 paratypes from Lae, also in the lower Markham Valley, Oct. 1944 (Darlington). Other material. Eighteen, Morobe Dist., N-E. N. G. (some spe- cifically from Surprise Creek, Sept. 28 & Oct. 7) (Stevens, M.C.Z.) The Morobe Dist. is not far south of the Markham Valley. Measured specimens. The o holotype and 1 @ paratype from. Nadzab. | LITHAGONUM ANNULICORNE BISMARCKENSE N. subsp. Description. Legs blackish. Head .90 & .86 width prothorax. Pro- thorax: width/length 1.20 & 1.18; base/apex 1.01 & 1.02. Elytra slightly more elongate, less oval, and less convex than in preceding (politior); apices comparatively weakly emarginate between denticles, and latter less prominent than in politior; striae comparatively less coarsely punctate; all intervals including Sth and 9th without reticu- late microsculpture at least at middle of length (except 9th intervals faintly reticulate in 1 specimen). Measurements: length 9.3-10.1; width 3.5-3.8 mm. Types. Holotype o (M.C.Z. No. 28,632) and 7 paratypes from 180 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Chimbu Valley, Bismarck Range, N-E. N. G., 5,000-7,500 ft. (probably all actually from near 5,000 ft.), Oct. 1944 (Darlington). Measured specimens. The o& holotype and 1 @ paratype. LITHAGONUM ANNULICORNE ANNULICORNE (Maindr.) For references, see species as a whole, above. Description. Legs blackish. Head .86 & .87 width prothorax. Prothorax: width/length 1.20 & 1.17; base/apex 1.00 & 1.01. Elytra relatively short, convex, subquadrate (with sides at most very slightly arcuate at middle); apices rather deeply emarginate between denticles; outer denticles large, almost short spines, much more prominent than sutural denticles; striae coarsely punctate; 9th and (much less distinctly) 8th intervals with reticulate microsculpture. Measure- ments: length 9.6-9.8; width about 3.7 mm. Types. As indicated under species as a whole, where type locality is restricted to ““Tana Mera” (Tanahmerah Bay), at western end of Cyclops Mts., Neth. N. G. Occurrence in New Guinea. I have seen only 3 2 Q referable to typical annulicorne, all taken not far from the type locality, at Sabron, Cyclops Mts., 930 ft., Apr. 1936 (Cheesman). Measured specimens. Two of the 3 2 9 recorded above. Notes. The present description covers only the subspecific charac- ters of typical annulicorne. The species as a whole is more fully described above. LITHAGONUM ANNULICORNE BALIEM 0. subsp. Description. Legs blackish. Head .90 & .94 width prothorax. Prothorax: width/length 1.29 & 1.23; base/apex 1.01 & .98. Elytra relatively longer than in typical annulicorne and less convex; apices usually less emarginate between denticles, and outer denticles smaller and less prominent than in annulicorne s. s.; striae less coarsely punctate than in other subspecies; reticulate microsculpture always distinct on 9th, usually so on 8th, and often faintly visible on discal intervals too. Measurements: length 9.3-10.0; width about 3.3-3.7 mm. Types. Holotype o& (Leiden Mus.) and 12 paratypes (some in M.C.Z., No. 28,633) all from Baliem Camp, Snow Mts., Neth. N. G., 1,700 m. (about 5,525 ft.), Nov. 16-27, 1938 (Toxopeus). Measured specimens. The &@ holotype and 1 @ paratype. DARLINGTON: CARABID BEETLES OF NEW GUINEA 181 IRIDAGONUM new genus Diagnosis. Small or medium sized (8.2-13.4 mm.); Agonum-like or rarely fusiform; wing-and-seta formula +w, (+)+, —+, (+)++; elytra always strongly iridescent, characteristically toothed at sutural angles but not otherwise armed, usually with 7th and 8th intervals longitudinally impressed toward apex (not so in quadripunctellum) ; sides of body below always more or less strongly punctate. Description. Agonum-like or fusiform; color including that of appendages always dark; upper surface always strongly iridescent especially on elytra, impunctate except sometimes punctate in basal foveae of pronotum; microsculpture light or very fine, not distinguish- able at 54X except sometimes on front of head. Head only moderately elongate; eyes moderate in size, varying in prominence; both pairs supraocular setae present except anterior pair absent in subfusum; posterior pair between or behind line of posterior edges of eyes; antennae of average or less than average length, normally formed; frontal impressions slight; mentum tooth triangular, more or less blunted at tip. Prothorax with anterior-lateral setae always absent, posterior-lateral ones always present; otherwise differing in details in different species. Elytra margined at base, the margin always strongly (obtusely to about rectangularly) angulate at humeri; subapical sinuations always slight or absent; sutural angles always with denticles or short spines but apices not otherwise armed; striae normal; 7th and 8th intervals and sometimes others at least toward apex deeply longitudinally impressed and so appearing doubled (except not so in quadripunctellum). Inner wings fully developed. Lower surface with sides of sterna extensively and more or less closely but irregularly punctate; abdomen not pubescent; prosternal process normal. Legs normally formed, as in Notagonum, but 4th hind-tarsal segment always simply emarginate, not distinctly lobed, and 5th hind-tarsal segment always with fine, short, but distinct accessory setae; claws simple. Secondary sexual characters normal. Male copula- tory organs as figured (Figs. 44 & 45). Genotype. Iridagonum quadripunctum n. sp. (below). Generic distribution. Widely distributed in New Guinea; as yet unknown elsewhere. Notes. The species of this new genus form a natural group apparently endemic to New Guinea, perhaps derived from Notagonum. Two of the species (quadripunctum and sexpunctum) are unusually variable in form and proportions, but after careful study I can distinguish no more than the four species keyed out below. I have taken only one species (quadripunctum) of this genus in its natural habitat, among 182 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY dead leaves on the ground in heavy rainforest. This is probably the habitat of the other species too. The types of quadripunctellum were found in flood-debris in forested or formerly forested country. Most of the specimens of quadripunctum and sexpunctum collected by Cheesman and Toxopeus were apparently caught in light traps, for the specimens are liberally sprinkled with scales of Lepidoptera. Key to the Species of Iridagonum 1. Third interval of elytron 2-punctate (anterior puncture missing); (form rather Agonwm-like) .0.2..< 00, «i> | Tee ss 2-.+ eee 9 se eee 2 = Third interval’ d=punctate. (5. .6 . ci eos sore eo deigge cic re cae ee eee 3 2. Larger (9.3-13.4 mm.); elytral intervals 7 & 8 deeply longitudinally im- pressed at least toward apex (p. 182).................. quadripunctum — Smaller (8.2-9.0 mm.); elytral intervals not or only slightly impressed (DAES) es aren S20 ee ASE) eat es 0a Lee tn quadripunctellum 3. Form Agonum-like; both pairs of supraocular setae present (p. 184)...... sexpunctum — Form fusiform; anterior supraocular setae missing (and eyes smaller and Mmore;abrupt than) usual) (pa 135) eer eee ae eee subfusum IRIDAGONUM QUADRIPUNCTUM N. sp. Description. With characters of genus as described above. Form of broad Agonum s. s. (Fig. 5). Head .67, .67, & .69 width prothorax; eyes normally formed, of moderate size and prominence; both pairs supraocular setae present, posterior ones about between posterior edges of eyes. Prothorax moderately narrowed behind and somewhat more so in front; width/length 1.40, 1.33, & 1.36; base/apex 1.28, 1.25, & 1.26; anterior angles only moderately prominent; sides rather broadly arcuate anteriorly, nearly straight (or slightly arcuate or slightly sinuate in some individuals) and moderately converging posteriorly; posterior angles obtuse, narrowly rounded; lateral margins moderate, wider posteriorly, not sharply defined, only moderately reflexed; basal foveae rather broad and shallow, variably (usually rather closely) punctate; disc with median line moderately impressed, transverse impressions slight, anterior marginal line light and variable, posterior one still lighter, often absent. Elytra rather short and broad, somewhat variable in form, slightly and variably impressed before middle; basal margin strongly but obtusely angulate or subangulate at humeri; striae moderately impressed, faintly or not punctulate; 7th interval apically, Sth for much of length, (and 9th somewhat variably) longitudinally impressed; 3rd interval 2-punctate (anterior puncture missing); sutural angles with short, slightly out-curving spines in DARLINGTON: CARABID BEETLES OF NEW GUINEA 183 large individuals but with only strong denticles in smaller individuals. Male copulatory organs as figured (Fig. 44). Measurements: length 9.3-13.4; width 3.8-5.0 mm. Types. Holotype large & (M.C.Z. No. 28,634) and 10 paratypes (including entire size-range of species) from Dobodura, Papua, Mar.—July 1944 (Darlington). Also the following additional para- types: Papua: 1, Kokoda, 1,200 ft., Aug. 1933 (Cheesman); 1, Mafulu, 4,000 ft., Dec. 1933 (Cheesman); 2, Palmer River at Black River, June 7-14 & 15-22, 1936 (Archbold Exp., American Mus.). N-E. N. G.: 1, Sattelberg (British Mus.); 3, Aitape, Aug. 1944 (Darlington). Neth. N. G.: 1, Dorey (‘‘Dory’’) (Vogelkop) (British Mus.); 1, mountain slope above Bernhard Camp, Snow Mts., 100 m. (about 325 ft.), Apr. 1939 (Toxopeus); 1, Rattan Camp, Snow Mts., 1,150 m. (about 3,750 ft.), Feb.—Mar. 1939 (Toxopeus). Other material. Four, Waigeo Is., Neth. N. G., (Mt. Nok, 2,500 ft., May 1938, Cheesman). Measured specimens. Large & holotype and large 9 and small @ paratypes from Dobodura. Notes. Although this is in some ways an excessively variable species, the characters used to define it in the key (above) appear to hold without exception. TRIDAGONUM QUADRIPUNCTELLUM 0. sp. Description. With characters of genus as described above, except outer elytral intervals not or only slightly impressed. Form of average Agonum s.s. Head .68 & .68 width prothorax; eyes normally formed, moderately large and prominent; both pairs of supraocular setae present, posterior ones about between posterior edges of eyes. Pro- thorax moderately narrowed in front and behind; width/length 1.32 & 1.32; base/apex 1.24 & 1.22; anterior angles only normally prominent; sides broadly arcuate, straighter (converging) and usually slightly sinuate toward base; basal angles obtuse, almost rounded-out; lateral margins moderate, broader toward base, only slightly reflexed; basal foveae broad, shallow, not sharply defined, punctate, the punctation extending (more sparsely) onto sides of pronotum before foveae; disc about as in quadripunctum. Elytra of average width and convexity, not impressed on disc; basal margin distinctly but obtusely angulate at humeri; apical-sutural denticles small but distinct; striae moderately impressed, faintly or not punctulate; Sth interval usually vaguely impressed in part only, 7th usually not impressed; 3rd interval 2- punctate (anterior puncture missing). Measurements: length 8.2-9.0; width 3.0-3.5 mm. 184 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Types. Holotype o& (M.C.Z. No. 28,635) and 7 paratypes all from Aitape, N-E. N. G., Aug. 1944 (Darlington). Measured specimens. The o holotype and 1 @ paratype. Notes. Sufficiently defined in the key, above. ITRIDAGONUM SEXPUNCTUM nN. sp. Description. With characters of genus as described above. Form of Agonum s. s. (more slender than quadripunctum), but somewhat vari- able. Head .61, .62, & .65 width prothorax; eyes normally formed, only moderately large and prominent; both pairs supraocular setae present, posterior ones about between posterior edges of eyes. Pro- thorax longer than in quadripunctum, rather strongly narrowed in front, less so behind; width/length 1.27, 1.30, 1.18; base/apex 1.40, 1.41, & 1.50; anterior angles more prominent than in quadripunctwm (so that prothorax seems longer than figures suggest), acute but with apices narrowly rounded; sides broadly (variably) arcuate for much of length, more or less straight and slightly converging and often slightly and broadly sinuate before base; basal angles somewhat obtuse but well defined, only a little blunted; lateral margins moderate, not sharply defined, wider and a little more reflexed toward base; basal foveae shallow, more or less lightly punctate; dise as in quadripunctum. Elytra relatively long (in genus), with sides nearly parallel at middle; dise usually very slightly impressed about 14 from base; basal margin strongly, about rectangularly angulate at humeri; denticles at sutural angles smaller than in quadripunctum; striae lightly impressed, not or only faintly punctulate; outer elytral intervals longitudinally im- pressed about as in quadripunctum; 3rd interval 3-punctate, the punctures about normally placed. Measurements: length 10.2-12.2; width 3.6-4.4 mm. Types. Holotype @ (British Mus.) and 10 paratypes (some in M.C.Z. No. 28,636) from Mt. Cyclops, Cyclops Mts., Neth. N. G., 3,900 ft., Mar. 1936 (Cheesman). Also the following additional para- types from Neth. N. G.: 6, Mt. Lina, Cyclops Mts., 3,500-4,500 ft., Mar. 1936 (Cheesman); 1, simply Cyclops Mts., 3,400-4,500 ft., Mar. 1936 (Cheesman); 1, Araucaria C amp, Snow Mts., 800 m. (about 2,600 ft.), Mar. 1939 (Toxopeus); 6, Rattan Camp, Snow Mts., 1,150 & 1,200 m. (about 3,750 & 3,900 ft.), Feb.—Mar. 1939 (Toxopeus); 1, Sigi Camp, Snow Mts., 1,500 m. (about 4,875 ft.), Feb. 24, 1939 (Toxopeus); 1, Mist Camp, Snow Mts., 1,800 m. (about 5,850 ft.), Jan. 1939 (Toxopeus). Other material. One 9, Chimbu Valley, Bismarck Range, N-E. N. G., 5,000-7,500 ft., Oct. 1944 (Darlington). This probably repre- DARLINGTON: CARABID BEETLES OF NEW GUINEA 185 sents at least a distinct subspecies, but I do not care to describe it without more material because of the great individual variation of some species of this genus. Measured specimens. A large o’, large 9, and smaller and narrower 3", all from the Cyclops Mts. Notes. This variable species seems to be constant in at least the characters given in the key to species, above. IRIDAGONUM SUBFUSUM n. sp. Description. With characters of genus as described above. Broadly subfusiform, more attenuate in front than behind. Head .63 width prothorax; eyes a little smaller but more abruptly prominent than usual; anterior supraocular setae absent, posterior ones somewhat be- hind line of posterior edges of eyes. Prothorax strongly narrowed in front, only slightly so behind; width/length 1.21; base/apex 1.47; anterior angles prominent and acute except slightly rounded at tips; sides slightly arcuate at middle, nearly straight anteriorly and pos- teriorly, slightly sinuate near base; basal angles somewhat obtuse but well defined, only slightly blunted; lateral margins as usual in genus; basal foveae slightly deeper than usual, only vaguely punctate; disc as in other species; anterior marginal line almost entire, posterior one absent. Elytra rather wide, more convex than usual, scarcely impressed before middle; basal margin about rectangular at humeri; sutural angles with rudimentary denticles; striae deeply impressed, not punctate; outer intervals more lightly and briefly longitudinally impressed than in quadripunctum and sexpunctum; 3rd interval 3-punctate. Male copulatory organs: Fig. 45. Measurements: length 12.4; width about 5.0 mm. Type. Holotype & (Leiden Mus.) from Ibele (Iebele) Camp, Snow Mts., Neth. N. G., 2,250 m. (about 7,325 ft.), Nov.—Dec. 1938 (Toxopeus) ; unique. Measured specumen. The type. Notes. In form, modification of eyes, and loss of anterior supraocular setae this interesting species parallels certain Fortagonum (bufo etc.) (see below) which also occur at high altitudes on the Snow Mts., but there is probably no direct relationship. ALTAGONUM new genus Diagnosis. Small or medium-sized (5.3-13.7 mm.); very variable in form, often Agonwm- or Calathus-like, sometimes Europhilus- or Sphodrus-like or fusiform; variable also in color, usually black or 186 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY brown, rarely with elytra green (cheesmanz) or purplish (regiscapha), elytra sometimes broadly margined with pale but never mottled or blotched, sometimes (but not usually) more or less iridescent; wing- and-seta formula +-w, +-+, — (++), (+) (+) (4+); see also key to genera. Description. Size, form, and color as described above; upper surface usually impunctate (sometimes punctate in basal foveae of pronotum); microsculpture variable. Head more variable in form than in Nota- gonum, often a little more elongate; eyes variable, usually at least moderately large and prominent, but sometimes reduced in size or prominence; both pairs supraocular setae always present, posterior pair usually about between posterior edges of eyes, but relatively farther back when eyes reduced; head otherwise about as in Notagonum. Prothorax variable in form; anterior-lateral setae always absent, pos- terior-lateral ones usually present but absent in nudicolle and fatuwm; ‘prothorax otherwise essentially as in Notagonwm; further details as described for species below. Elytra variable in form; basal margin entire, rounded or variably angulate at humeri; apices variable; striae normal; intervals usually not much modified toward apex, but outer ones sometimes (pallinox, sphodrum, ete.) longitudinally suleate at least apically; 3rd interval usually normally 3-punctate, sometimes 2-punctate with anterior puncture absent, rarely impunctate (fatwwm only). Inner wings always fully developed. Lower surface usually impunctate or nearly so; abdomen usually not pubescent, but with moderate or extensive pubescence in a few cases (pubinox, pallinox, noctellum, & planinox; sphodrum & postsulcatum); prosternal process simple. Legs including tarsi as in Notagonwm, but 4th hind-tarsal segment usually simply emarginate, and if lobed (caducum, cheesmani, scapha, nudicolle) lobes very short, but with outer lobe still usually slightly longer than inner one; 5th hind-tarsal segment without obvious accessory setae but always or almost always with minute, very incon- spicuous, perhaps vestigial ones (as in at least some Notagonum); claws simple. Secondary sexual characters normal. Male copulatory organs as figured (Figs. 46-51). Genotype. Altagonum caducum n. sp. (below). Generic distribution. Numerous in New Guinea, chiefly in the mountains; further range not determined. Notes. Like Notagonum, this is a genus of convenience, distinguished from Notagonum by only one constant detail, absence of the anterior- lateral pronotal setae. This character is not of itself of generic value, and the group which it defines is, as I have indicated, not entirely a natural one. However its recognition is useful. The group is a tran- sitional one which bridges the gap between the more generalized low- DARLINGTON: CARABID BEETLES OF NEW GUINEA 187 land Agonini of New Guinea, especially Notagonum, and several well-marked groups of mountain species, here called genera, charac- terized by further loss of setae, loss of wings, and in other ways. Altagonum is intermediate in altitudinal distribution as well as in structure. The majority of Notagonum occur at low altitudes, although there are also some which occur high in the mountains. But of Alta- gonum only one species (vallicola) seems to occur regularly at lowest altitudes and only one other (grossulwm) was found even rarely at Dobodura. Most species of the genus are chiefly or entirely confined to middle and high altitudes. Their structure is correlated with their distribution. In several ways (loss of setae, modification of eyes in some cases, partial or complete loss of well-developed lobes of the 4th hind-tarsal segments) they show the beginnings of the effect of the mountain environment on what were obviously originally normal Agonini. These same changes, and eventually also loss of wings and other changes, have occurred among mountain Carabidae, especially Agonini, in many other parts of the world, as I have mentioned in the introduction to the present paper. The few species of Altagonwm that I have myself collected in sufficient numbers to be sure of their habitats (vallicola, sphodrum, nudicolle) all occur on the ground in heavy forest, and are not associated with running or standing water. This is probably the habitat of most species of the genus, although some may depart from it. One species evidently referable to this genus has been previously described, but I do not recognize it in the material before me. It is Altagonum papuense (Sl.) Platynus papuensis Sloane 1890, Records Australian Mus. 1, p. 103. Colpodes papuensis Sloane 1907, Deutsche Ent. Zeits., p. 179. Type. From St. Joseph’s River District, on the south coast of Papua about opposite Yule Is.; probably in the Australian Mus. at Sydney, Australia. Occurrence in New Guinea. Known only from the type. Notes. This is a rather large (9 by 3144 mm.), black, Agonum-like form, with 3rd elytral interval 2-punctate (anterior puncture absent). It may be a form of the species here called vallicola, but if so it is probably not identical with any of the subspecies here described: it is larger than typical vallicola, probably broader than subspecies huonis, and apparently without the iridescence of subspecies subvividum. Its locality (the south coast of New Guinea) is consistent with its being a fourth subspecies of the group. But Sloane, in his careful description of papuense, mentions no angulations of the elytral apices, which 188 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY suggests that he had before him a distinct species which I have not seen. There is no use in guessing further about this now. For the present it seems better not to assign the name at all than to risk assigning it wrongly. The species is not included in the following key. Key to the Species of Altagonum of New Guinea . Posterior-lateral pronotal setae present... --.....2...2..45-5 so eee 2 Posterior-lateral pronotal setae absent. ......-.....:2..5...8:5s0e: 23 . Plain black or brown, Agonum-like forms, not fusiform; prothorax normally formed, at least moderately narrowed behind, with posterior angles obtuse: orrounded she wa. 68 obey ee sie ee 3 Notas above invonelonmore details). 56 oh ee eee ne 14 . Third elytral interval 2-punctate (anterior puncture absent).......... 4 Thirdelytralinterval’3-punctate> 71.44 4. is oe ee eee 6 . Black, lateral margins of prothorax and elytra not much paler, legs dark; elytralapices strongly angulatelse = s. aoec ae ace cee ee eee vallicola (4a) Smaller (6.8-7.9 mm.); not distinctly iridescent (p. 190).......... (vallicola s. s.) (4b) Larger (8.3-9.0 mm.); more slender; faintly iridescent (p. 190)..... (subsp. huonis) (4c) Still larger (9.1-9.7 mm.); broad as typical form; more strongly iri- descent (op. 191) ean ael vale ai ey Nea oie ne, pane (subsp. swbvividum) — Brownish-black, lateral margins of prothorax and elytra yellowish- translucent, legs yellowish-brown; elytral apices not angulate........ 5 5. Broader, prothoracic width/length 1.47-1.54; length 5.3-6.7 mm........ grossulum (5a) Reticulate microsculpture (at 54) faint or almost absent on disc of pronotum, distinct and moderately transverse on elytra; latter My vPOKECI MH Gos WO). cee acccoccgovssccccgucde (grossulum s. 8.) (5b) Microsculpture distinct on pronotum, indistinguishable (at 54) on elytra; latter not distinctly iridescent (p.193)..... (subsp. reticolle) (5c) Microsculpture of pronotum light and variable, of elytra indistinct (at 54); elytra and sometimes also pronotum iridescent (p. 193) (subsp. intensum) — Narrower, prothoracic width/length 1.31-1.40; length 6.4-7.5 mm. (p. 194) grossuloides 6. Basal margin of elytra obtusely angulate at humeri; length 6.5-10.6 mm. (if less than 9.0 mm., abdomen pubescent)...................-+--- 7 — Basal margin of elytra rectangular or nearly so at humeri (points of angles sometimes slightly blunted); length 6.6-7.9 mm. (abdomen not pu- IbeSCeNt)) mi heepe ested cients ices cap petvgthegeteis urekoneueioy acai Ga ORe nNOS Skee a mae 13 . Abdomen not. pubescent; length 9.0-10.6 mm........................ 8 Abdomen pubescent at least near middle basally; length 6.5-9.0 mm... .10 = Wength,9:0-9o mame (py 94) (Raye eie a eer. sont oe eae eee eae nox Larger, relatively a little broader; (see also descriptions).............. 9 . Lateral margins of prothorax not obviously pale (p. 195)........ magnox wm |p » | oo | oloalan 23. DARLINGTON: CARABID BEETLES OF NEW GUINEA 189 Lateral margins of prothorax yellowish-translucent; (see also description) (Gee SIGLOD EC CARR a ale aa acne Re Shane Mr Co Wn, bin LALLA jJapenox mlbeng ches: 1—9) Opmamne (y= 9G) get age age eoatees ay ee ieee: pubinox Sizermallerss(seeralsondescriptions) 2-45 7 rece eis oe iat . Outer elytral intervals longitudinally impressed at least toward apex (oe LAUD) eee iy Ba enone tees pC email ene crane Hist, Suter: Sau raat pallinox Outer elytral intervals not longitudinally impressed................. 12 . Basal foveae of pronotum moderately impressed (p. 197)...... noctellum Basal foveae of pronotum scarcely impressed, flat (p. 198)...... planinox . Broad Agonum-like; prothoracic width/length 1.29-1.34 (p. 198). .dilutipes More slender Hurophilus-like; prothoracic width/length 1.15 & 1.18 (Gos L99) tye Peo Pek i came at neu Rpe cents Kane ath 8 europhilum mE lytramotbrosdlysmareinedawithipalen esse eater ae cen 15 Bly traybroadlyamarginedswithepaleneee seiko eras ene 22 . Apex of each elytron drawn out into a single, short, acute spine nearly in line of sutural interval, the spines slightly dehiscent; (see also description (Gore ZO) Bremer tye tod, acm ana esa eee Naaru oi ore ei eile tat tutum Elytra with apices simple, or armed differently from above.......... 16 weNiot strikingly tusiformerelytra notumetallice ta. 4 4er se see cis oe oe 17 Fusiform; elytra sometimes (not always) green or purplish........... 20 . Outer elytral intervals not much modified toward apex (p. 202). .caducum Elytral intervals 7, 8, & 9 longitudinally impressed at least toward apex .18 . Prothorax subquadrate, relatively narrow, width/length 1.18 & 1.20 in measured specimens; base/apex 1.28 & 1.20; (abdomen pubescent) (ays PAYED We ice chetcees Sanne ou pemenen eter oti RE NEUEN Se VE CREO ICE sphodrum Prothorax wider, more narrowed in front, width/length 1.30-1.44, base/ Ue Xan (eA es cy en Mae Memes ctivagnn Say gt TM eb ea lon cin se eaytarcmee eh pas 19 . Apices of elytra unarmed or with at most small sutural denticles; abdomen pubescent at least near middle basally (p. 205).......... postsulcatum Apex of each elytron with 2 prominent denticles; abdomen not pubescent (RFA cn asta nent ce ESR af eatoars a rae. Bt aie year MeN SAG misim . Elytra green, (spined); eyes very prominent, head .80 width prothorax (Gos ANIL) Ne et tel ofr ic Capea te oe ee as eae emer CAREY Se aries Me cheesmant Elytra black or purplish, (spined or angulate at apices); eyes less prominent, head? Golomless widthyprothoraxseen ation ee 21 . Elytra black; posterior-lateral pronotal setae on flat surface of margins well pinefromled ges) (o-p20S) kn iersive ala erate lord nec r ee scapha Elytra purplish; posterior-lateral pronotal setae on edges of margins (GaP DD) ei ne See ee eee ema PD yO OALT At cs Pg eee Sra regiscapha . Broader; prothoracic width/length 1.48 & 1.44 in measured specimens; dark discal color of elytra not extending in a narrow sutural stripe toward APEXa (POZO MLE faded nL ee: 2) AE TaR A RR CER hate Os aca latilimbus More slender; prothoracic width/length 1.32 & 1.34; dark discal color of elytra extending along sutural intervals toward (not to) apex (p. 211) paralimbus Form rather broad Calathus-like; 3rd elytral interval 3-punctate (p. 211) nudicolle Form slender; 3rd elytral interval impunctate (p. 212)........... fatuum 190 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY ALTAGONUM VALLICOLA N. sp. Description. With characters of genus as described above. Form of broad Agonum s. s.; black, appendages brownish-black, lateral margins of prothorax and elytra only slightly translucent; upper surface im- punctate except vaguely punctate in basal foveae and on lateral margins of pronotum, moderately shining, not distinctly iridescent; microsculpture lightly impressed, normal (isodiametric on head, trans- verse on pronotum, more transverse on elytra). Head .61 & .66 width prothorax; eyes large and prominent, with posterior supraocular setae about between their posterior edges. Prothorax large and wide; width/ length 1.55 & 1.44; base/apex 1.43 & 1.42; sides strongly arcuate for most of length, nearly straight and converging and sometimes slightly sinuate before basal angles; latter obtuse, moderately rounded; lateral margins rather wide, wider and more reflexed toward base; basal foveae normal, vaguely punctate; disc normal; anterior and posterior marginal lines entire or nearly so. Elytra moderately wide, normal in outline and convexity, not or faintly impressed before middle; basal — margin usually vaguely subangulate at humeri; lateral margins rather wide; subapical sinuations moderate; apices prominently angulate about opposite 3rd intervals, these angulations more prominent than sutural ones; apices then obliquely subtruncate to finely denticulate sutural angles; striae variably, usually rather lightly impressed, not or faintly punctulate; intervals nearly flat or slightly convex, 8th and 9th not much modified toward apex; 3rd interval 2-punctate (anterior puncture absent). Lower surface virtually impunctate; abdomen not pubescent. Legs: 4th hind-tarsal segment emarginate, not lobed. Male copulatory organs: Fig. 46. Measurements: length 6.8-7.9; width 2.8-3.1 mm. Types. Holotype & (M.C.Z. No. 28,637) and 25 paratypes all from Dobodura, Papua, Mar.—July 1944 (Darlington); taken among dead leaves on the ground in heavy rainforest in company with superficially similar Notagonum spinulum. Measured specimens. The o holotype and 1 9 paratype. Notes. Sufficiently discussed and defined under the genus and in the key to species, above. Perhaps it should be added that small specimens of this species are extremely similar to Notagonum subspinulum except that they lack the anterior-lateral pronotal setae and have the elytral apices armed with only acute teeth rather than short spines. ALTAGONUM VALLICOLA HUONIS n. subsp. Description. Similar to typical vallicola in all details except a little larger, obviously more elongate, and with faint iridescence on elytra DARLINGTON: CARABID BEETLES OF NEW GUINEA 191 in strong light. Head .67 & .69 width prothorax. Prothorax: width/ length 1.39 & 1.39; base/apex 1.46 & 1.47; sides less strongly rounded than in typical vallicola, and basal angles correspondingly more dis- tinct, though still obtusely rounded. Elytra more slender and elongate, with basal margin a little more distinctly angulate at humeri. Measure- ments: length 8.3-9.0; width 3.1-3.4 mm. Types. Holotype # (M.C.Z. No. 28,638) from (hills north of) Nadzab, N-E. N. G., July 1944 (Darlington); and 1 & paratype from Sattelberg, N-E. N. G. (British Mus., ex Coll. G. Hauser). Measured specimens. ‘The types. Notes. Sufficiently compared with typical vallicola above. ALTAGONUM VALLICOLA SUBVIVIDUM n. subsp. Description. Similar to typical vallicola in all details except larger, with elytra a little more distinctly (but variably) impressed before middle, a little more deeply striate, and more or less strongly iridescent. Head .63 & .64 width prothorax. Prothorax: width/length 1.41 & 1.42; base/apex 1.46 & 1.55; sides rather strongly rounded. Elytra less elongate than in huonis, a little more deeply striate than usual in typical vallicola or huonis, and rather strongly iridescent, with elytral microsculpture not distinguishable at 54 and probably too fine and transverse to see. Measurements: length (types) 9.1-9.7; width (types) 3.4-3.7 mm. Types. Holotype @ (British Mus.) and 4 paratypes (2 in M.C.Z. No. 28,639) all from Mt. Baduri, Japen Is., Neth. N. G., 1,000 ft., Aug. 1988 (Cheesman). Other material. All from Neth. N. G., as follows: 1, Hollandia, July—Sept. 1944 (Darlington); 4, Mt. Sabron, Cyclops Mts., 2,000 ft., June & July 1936 (Cheesman); 2, Rattan Camp, Snow Mts., 1,150 & 1,200 m. (about 3,750 & 3,900 ft.), Feb.—Mar. 1939 (Toxopeus) ; and 1, “New Guinea” (H. E. Andrewes Coll., British Mus., labeled “Colpodes sp.”). These specimens, though referable to this subspecies, are a little smaller and/or less strongly iridescent than the types. Measured specimens. The o holotype and 1 @ paratype. Notes. Sufficiently compared with typical vallicola and subspecies huonis above. ALTAGONUM GROSSULUM N. sp. Description. With characters of genus as described above. Form of small but very broad Agonuwm s. s.; brownish-black, appendages yellowish-brown, lateral margins of prothorax and elytra yellowish- 192 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY translucent; upper surface virtually impunctate, moderately shining, not iridescent; microsculpture distinct and isodiametric on head, almost absent or faint on dise of pronotum, distinct and moderately transverse on elytra. Head .61 & .63 width prothorax; eyes large and prominent, with posterior supraocular setae slightly before line of their posterior edges. Prothorax relatively large and wide; width/length 1.54 & 1.50; base/apex 1.31 & 1.48; sides nearly evenly, rather strongly arcuate for entire or nearly entire length; basal angles obtuse, moder- ately rounded; lateral margins rather wide, but not much more reflexed toward base; basal foveae rather broad and shallow, vaguely or not distinctly punctate; disc normal, with anterior marginal line entire, posterior one more or less entire but sometimes vague at middle. Elytra broad, otherwise of normal outline and convexity, distinctly impressed before middle; basal margin obtusely angulate or blunted at humeri; lateral margins rather wide (in genus); subapical sinuations moderate or slight; apices nearly simple, somewhat irregularly rounded, with sutural angles not or sometimes vaguely denticulate; striae well impressed, impunctate; intervals more or less convex, 8th and 9th not much modified toward apex, 3rd 2-punctate (anterior puncture absent). Lower surface virtually impunctate; abdomen not pubescent. Legs: 4th hind-tarsal segment simply emarginate, not lobed. Measurements: length 5.4-6.5; width 2.3-2.8 mm. Types. Holotype & (M.C.Z. No. 28,640) and 2 oo" paratypes from Dobodura, Papua, Mar.—July 1944 (Darlington); and additional para- types as follows: Papua: 1, Kokoda, 1,200 ft., Aug. 1933 (Cheesman) ; 2, Mafulu, 4,000 ft., Dec. 1933 & Jan. 1934 (Cheesman). N-E. N. G.: 1, Nadzab (“E. fork Ngafir Cr. 1,000-3,000 ft. native trail”), July 13, 1944 (K. V. Krombein, U.S.N.M.). Other material. Two additional subspecies described below; and 1 specimen from Mt. Nok, Waigeo Is., Neth. N. G., 2,500 ft., May 1938 (Cheesman), which does not fit into any of the subspecies here recognized. Measured specimens. The @ holotype and 1 o paratype from Dobodura. Notes. This species is distinguished from all related forms known to me by its small size and relatively broad but still Agonum-like form, plus details given in the key to species. The present, typical subspecies is distinguished from the others described below by having the disc of the pronotum more shining, without or with only faint traces of reticu- late microsculpture, and the elytra with comparatively coarse and obvious microsculpture (most obvious in the Kokoda and Dobodura specimens, finer but still visible in the Mafulu and Nadzab ones) and no or only very slight iridescence. I do not know the ecological habitat DARLINGTON: CARABID BEETLES OF NEW GUINEA 193 of this species. Although it does occur, rarely, near sea level at Dobo- dura, it (as represented by the following subspecies) seems to be much commoner in foothills and lower mountains. ALTAGONUM GROSSULUM RETICOLLE n. subsp. Description. Size, form, and structural details almost as in typical grossulum, but microsculpture of upper surface different, as follows: on head, about same; on disc of pronotum, meshes distinct though somewhat variably so, moderately transverse (not faint or absent as in typical grossulum); on elytra, so fine as not to be visible at 54X (not distinct as in typical grossulum), but in spite of fineness of micro- sculpture of elytra, latter not or only faintly iridescent even in strong light. Head .65 & .63 width prothorax. Prothorax: width/length 1.51 & 1.51; base/apex 1.32 & 1.38. Measurements: length 5.7-6.7; width 2.4—2.8 mm. Types. Holotype &® (British Mus.) and 4 paratypes (2 in M.C.Z. No. 28,641) from Sabron, Cyclops Mts., Neth. N. G., 2,000 ft., June & July 1936 (Cheesman); 2 paratypes from Mt. Lina, Cyclops Mts., 3,500-4,500 ft., Mar. 1936 (Cheesman). Measured specimens. The o holotype and 1 2 paratype from Sabron. Notes. Sufficiently compared with typical grossulwm above. ALTAGONUM GROSSULUM INTENSUM n. subsp. Description. Size, form, and structural details almost as in typical grossulum and subspecies reticolle, but pronotum sometimes and elytra always rather strongly iridescent; microsculpture of head as in pre- ceding forms, of prothorax light and variable, of elytra not distinctly visible at 54X but probably very fine and transverse. Head .66 & .62 width prothorax. Prothorax: width/length 1.48 & 1.51; base/apex 1.41 & 1.47. Measurements: length 5.3-6.7; width 2.3-2.9 mm. Types. Holotype & (British Mus.) and 18 paratypes (some in M.C.Z. No. 28,642) from Mt. Baduri, Japen Is., Neth. N. G., 1,000 ft., Aug. 1938 (Cheesman). Other material. Nine, Rattan Camp, Snow Mts., Neth. N. G., 1,150 m. (about 3,750 ft.), Feb._Mar. 1939 (Toxopeus); and 1, Lower Mist Camp, Snow Mts., 1,700 m. (about 5,525 ft.), Jan. 17, 1939 (Toxopeus). These specimens average a little larger and somewhat less strongly iridescent than the types. Measured specimens. The o holotype and 1 9 paratype. Notes. Sufficiently compared with other subspecies above and in the key to species of Altagonum. 194 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY ALTAGONUM GROSSULOIDES N. sp. Description. Extremely similar to A. grossulum, and answering to the description of the latter (see above) in every detail not mentioned below, but more slender, with prothorax especially narrower. Head .67 & .64 width prothorax. Prothorax: width/length 1.31 & 1.39 (5 other specimens 1.34 to 1.40); base/apex 1.50 & 1.48; sides much more weakly arcuate than in grossulum and a little more narrowly margined. Elytra as in grossulum. Measurements: 6.4-7.5; width 2.6-3.0 mm. Types. Holotype o (Leiden Mus.) and 8 paratypes (some in M.C.Z. No. 28,643) from Mist Camp, Snow Mts., Neth. N. G., 1,800 m. (about 5,850 ft.), Jan. (1 paratype Feb. 3) 1939 (Toxopeus). Also the following paratypes, all from Neth. N. G.: 3, Rattan Camp, Snow Mts., 1,150 m. (about 3,750 ft.), Feb._Mar. 1939 (Toxopeus); 3, Top Camp, Snow Mts., 2,100 m. (about 6,825 ft.), Jan. 27, 1939 (Toxo- peus); 4, Mt. Lina, Cyclops Mts., 3,500-4,500 ft. (2 specifically from 4,500), Mar. 1936 (Cheesman); 1, Cyclops Mts. without more exact locality, 3,400-4,500 ft., Mar. 1936 (Cheesman); and 2, Mt. Baduri, Japen Is., 1,000 ft., Aug. 1938 (Cheesman). Measured specimens. The & holotype and1 9 paratype from Mist Camp. Notes. If this were geographically separated from grossulum, I might consider it merely a subspecies of the latter, but since the two occur in the same areas they are presumably distinct species. There is no overlapping of prothoracic ratios: width/length of prothorax in 9 measured specimens of grossulwm (including all subspecies) is 1.47 to 1.54; in 7 of grossuloides, 1.31 to 1.40. Grossuloides tends to be larger than grossulum and to occur at higher altitudes, but these are not absolute differences. ALTAGONUM NOX DN. sp. Description. With characters of genus as described above. Form of rather large Agonum s. s.; black, appendages only slightly paler, lateral margins of prothorax and elytra scarcely translucent; upper surface impunctate except sometimes more or less vaguely punctate in and near pronotal foveae, moderately shining, elytra slightly or faintly iridescent; microsculpture fine but apparently normal in form. Head .64 & .64 width prothorax; eyes moderately large and prominent, with posterior supraocular setae about between their posterior edges. Pro- thorax rather large and wide; width/length 1.45 & 1.43; base/apex 1.47 & 1.40; sides moderately arcuate for much of length, nearly straight and moderately converging and sometimes slightly sinuate toward base; basal angles obtuse but well defined, only slightly blunted; DARLINGTON: CARABID BEETLES OF NEW GUINEA 195 lateral margins rather wide especially toward base, moderately re- flexed; basal foveae rather deep, not or vaguely punctate; disc normal; anterior and posterior marginal lines entire, the posterior one more lightly impressed. Elytra of normal outline and convexity, not or vaguely impressed before middle; basal margin obtusely angulate at humeri; lateral margins average; subapical sinuations rather weak; apices rather narrowly rounded to slightly or not distinctly denticulate sutural angles; striae moderately impressed, not punctate; intervals somewhat convex, 8th and 9th not much modified toward apex, 3rd 3-punctate with punctures normally placed. Lower surface virtually impunctate; abdomen not pubescent. Legs: 4th hind-tarsal segment rather deeply emarginate but without produced lobes. Measurements: length 9.0-9.5; width 3.3-3.5 mm. Types. Holotype o& (Leiden Mus.) and 5 paratypes (2 in M.C.Z. No. 28,644) all from Mist Camp, Snow Mts., Neth. N. G., 1,800 m. (about 5,850 ft.), Jan. (type Jan. 18, 1 paratype Jan. 4) and Feb. 10, 1939 (Toxopeus). Measured specimens. The o holotype and 1 @ paratype. Notes. This species, which is of course differentiated from others in the key to species, above, will be taken as the standard of comparison for several more or less similar, following forms. ALTAGONUM MAGNOX N. sp. Description. Very close to nox and answering to the same description (see above) in all details except those noted below. Larger and a little broader; basal foveae and margins of pronotum before them a little more distinctly punctate; elytra a little more iridescent. Head .64 & .60 width prothorax. Prothoraz a little more narrowed in front and less narrowed behind, with sides less converging toward base; width/ length 1.38 & 1.45; base/apex 1.56 & 1.59. Other characters as in now. Male copulatory organs: Fig. 47. Measurements: length 10.6; width 4.2 mm. (both specimens). Types. Holotype o (Leiden Mus.) from Rattan Camp, Snow Mts., Neth. N. G., 1,200 m. (about 3,900 ft.), Feb._Mar. 1939 (Toxopeus) ; and 1 9 paratype (M.C.Z. No. 28,645) from Mist Camp, Snow Mts., 1,800 m. (about 5,850 ft.), Feb. 3, 19389 (Toxopeus). Measured specimens. ‘The types. Notes. The only characters distinguishing this species from nox to which I can give exact expression are the larger size and relatively greater ratio of base/apex of prothorax. The difference in form of prothorax is obvious to the eye. 196 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY ALTAGONUM JAPENOX N.. sp. Description. Again very close to nox, and answering to the de- scription of the latter (see above) in all details except those noted below. Larger and a little wider than now; legs browner, lateral margins of prothorax yellowish-translucent; elytra more obviously iridescent. Head 58 & .58 width prothorax. Prothorax relatively a little wider; width/length 1.51 & 1.51; base/apex 1.50 & 1.50. Measurements: length 9.9-10.2; width about 3.9 mm. Types. Holotype 2 (British Mus.) and 1 9 paratype (M.C.Z. No. 28,646) both from Mt. Baduri, Japen Is., Neth. N. G., 1,000 ft., Aug. 1938 (Cheesman). Measured specimens. ‘The types. Notes. This species is perhaps even closer to magnox (above) than to now, the base-species with which I have compared it. The principal characters distinguishing japenoa from magnow are the obviously pale- translucent prothoracic margins and the relatively slightly narrower -head and wider prothorax as shown by the ratios given. It remains to be seen to what extent these differences will hold in series. Japenox may prove to be a subspecies of magnox, but I am not sufficiently sure of relationships in this group of Al/tagonum to make it a subspecies now. ALTAGONUM PUBINOX N. Sp. Description. Again close to nox, and answering to the deseription of the latter (see above) in all details except those noted below. A little smaller and much narrower than now, with elytra obviously im- pressed before middle and at most only faintly iridescent. Head .70 & .70 width prothorax. Prothorax much narrower than in now; width/ length 1.27 & 1.27; base/apex 1.38 & 1.42; sides rather weakly arcuate for most or all of length, moderately converging and sometimes nearly straight or even faintly sinuate posteriorly; basal angles more obtuse and more blunted than in now primarily because sides of base (in pubinox) more rounded-oblique; lateral margins narrower than in noe. Lower surface as in nox except abdomen with a little pubescence near middle of at least basal segments. Measurements: length 8.1-9.0; width 3.0-3.4 mm. Types. Holotype &@ (Leiden Mus.) and 7 paratypes (some in M.C.Z. No. 28,647) from Sigi Camp, Snow Mts., Neth. N. G., 1,500 m. (about 4,875 ft.), Feb. (type & 3 paratypes specifically Feb. 19, 2 paratypes Feb. 25) 1989 (Toxopeus); and the following additional paratypes from the Snow Mts.: 1, Lower Mist Camp, 1,700 m. (about 5,525 ft.), Jan. 17, 1939 (Toxopeus); 1, Mist Camp, 1,800 m. (about 5,850 ft.), Jan. 1939 (Toxopeus). land DARLINGTON: CARABID BEETLES OF NEW GUINEA 197 Measured specimens. The & holotype and 1 2 paratype from Sigi Camp. Notes. Although this species differs from nox (and from magnox and japenox) by presence of a little pubescence near middle of at least the basal ventral abdominal segments, other characters are so similar that I think the relationship to now is probably close. ALTAGONUM NOCTELLUM 0. sp. Description. Again close to nox, and answering to the description of the latter (see above) in all details except those noted below. Much smaller, somewhat narrower, and a little more brownish (less deep black) than nov; elytra slightly, sometimes indistinctly impressed before middle, at most faintly iridescent. Head .77 & .76 width pro- thorax. Prothorax relatively much narrower than in now; width/length 1.24 & 1.29; base/apex 1.34 & 1.35; sides weakly arcuate anteriorly, straight or slightly sinuate and only moderately converging toward base; basal angles much more obtuse or rounded than in now, chiefly because sides of base rather strongly oblique; lateral margins slightly narrower and less reflexed than in now, and basal foveae shallower, but still moderately impressed. Lower surface as in nox except abdomen with a little pubescence near middle of at least basal segments. Measurements: length 6.8-7.9; width 2.5-2.9 mm. Types. Holotype @ (British Mus.) and 2 paratypes (co? in M.C.Z. No. 28,648) from Cyclops Mts., Neth. N. G., 3,400-4,500 ft., Mar. 1936 (Cheesman); 3 paratypes from Mt. Lina, Cyclops Mts., 3,500- 4,500 ft., Mar. 1936 (Cheesman); and 2 paratypes from Rattan Camp, Snow Mts., 1,200 m. (about 3,900 ft.), Feb.—Mar. 1939 (Toxopeus). Other material from Neth. N. G. as follows: 1, Sabron, Cyclops Mts., 2,000 ft., May 1936 (eyes abnormally prominent) (Cheesman) ; 1, Camp Nok, Waigeo Is., 2,500 ft., Apr. 1938 (Cheesman). Measured specimens. The & holotype and 1 Q paratype from the Cyclops Mts. Notes. Although I have compared this with what I have taken as the base-species of this group (now), noctellum is actually closer to pubinox, which it resembles in ventral pubescence. It is in fact very close to pubinow, differing most obviously in smaller size, with head relatively slightly wider, pronotum a little less convex, and elytra less obviously impressed before the middle. It evidently occurs at slightly lower altitudes than pubinor. The two Snow Mts. specimens of noctellum are not intermediates but show all the characters of the led species, including small size (both specimens under 7 mm.). 198 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY ALTAGONUM PLANINOX N. sp. Description. A member of the nox group but with form more of Europhilus than of Agonum s. s. Answering technical description of nox (see above) in all details except those noted below. Much smaller, more slender, with flatter pronotum than nox; more brownish-black, appendages yellowish-brown, lateral margins of prothorax narrowly but distinctly yellowish-translucent; elytra slightly impressed before middle, at most faintly iridescent. Head .80 & .78 width prothorax; eyes a little more prominent than usual in group. Prothoraz relatively small, subquadrate; width/length 1.36 & 1.38; base/apex 1.37 & 1.39; sides weakly arcuate, appearing subparallel, rather weakly converging and slightly or not sinuate basally; basal angles obtuse, blunted or rather narrowly rounded; lateral margins narrow, very little reflexed; basal foveae scarcely impressed, flat; disc flatter than in nox. Elytra with basal margin rounded or vaguely angulate at humeri; lateral margins narrow; outer intervals as in noz, not much modified toward apex. Lower surface: abdomen with some pubescence near middle of all segments. Measurements: length 6.5-7.5; width 2.6-2.9 mm. Types. Holotype & (British Mus.) and 5 paratypes (some in M.C.Z. No. 28,649) from Mt. Lina, Cyclops Mts., Neth. N. G., 3,500-4,500 ft., Mar. 1936 (Cheesman); 2 paratypes, Cyclops Mts. without further locality, 3,400-4,500 ft., Mar. 1936 (Cheesman); and 1 paratype, Rattan Camp, Snow Mts., 1,150 m. (about 3,750 ft.), Feb.—Mar. 1939 (Toxopeus). Other material. One specimen, Mafulu, Papua, 4,000 ft., Jan. 1934 (Cheesman). Measured specimens. The & holotype and 1 9 paratype from Mt. Lina. Notes. Set beside nox this species seems very different indeed, but there is a practically complete transitional series beginning in fact with magnox (and japenox) through nox, pubinox, and noctellum to the present species, and extending perhaps even to the following ones (dilutipes and europhilum), which may be transitional toward more fusiform species. The present species (planinox) differs from the pre- ceding ones of the series in having paler legs and antennal bases, and a flatter pronotum with scarcely impressed basal foveae. ALTAGONUM DILUTIPES n. sp. Description. With characters of genus as described above. Broad Agonum- or Stenolophus-like; brownish-black, appendages yellowish- brown, lateral margins of prothorax and elytra yellowish-translucent ; DARLINGTON: CARABID BEETLES OF NEW GUINEA 199 upper surface impunctate except vaguely punctate in pronotal foveae, moderately shining, elytra iridescent; microsculpture about normal except indistinct (presumably very fine and transverse) on elytra. Head .71 & .65 width prothorax; eyes moderately large and prominent, with posterior supraocular setae about between their posterior edges. Prothorax of moderate size; width/length 1.29 & 1.34; base/apex 1.24 & 1.28; sides moderately arcuate for most or all of length, sometimes nearly straight (and converging) toward base; basal angles broadly rounded, very obtuse or not at all defined; lateral margins moderate, wider basally, moderately reflexed; basal foveae average, vaguely punctate; disc normal; anterior and posterior marginal lines entire except posterior one sometimes vague near middle. Elytra rather broad, of normal outline and convexity, not or only faintly impressed before middle; basal margin strongly, almost rectangularly angulate at humeri (but points of angulations more or less blunted); lateral margins rather wide (in genus); subapical sinuations weak; apices rather narrowly rounded to subdenticulate sutural angles; striae moderately impressed, not punctate; intervals somewhat convex, 8th and 9th not much modified toward apex, 3rd normally 3-punctate. Lower surface impunctate; abdomen not pubescent. Legs: 4th hind- tarsal segment simply emarginate, not lobed. Measurements: length 6.6-7.8; width 2.5-3.0 mm. Types. Holotype & (British Mus.) and 3 paratypes (1 in M.C.Z. No. 28,650) from Mt. Lina, Cyclops Mts., Neth. N. G., 3,500-4,500 ft., Mar. 1936 (Cheesman); 2 paratypes, Cyclops Mts. without further locality, 3,500 & 3,400-4,500 ft., Mar. 1936 (Cheesman); and 2 para- types, Rattan Camp, Snow Mts., 1,150 & 1,200 m. (about 3,750 & 3,900 ft.), Feb.—_Mar. 1939 Gineenens) Measured specimens. The & holotype, and 1 9° penning from the Cyclops Mts. Notes. This species is sufficiently distinguished from others in the key to species, above. ALTAGONUM EUROPHILUM N. Sp. Description. With characters of genus as described above. Form of Agonum subgenus Europhilus; brownish-black, appendages some- what paler, lateral margins of prothorax and elytra only slightly translucent; upper surface virtually impunctate, moderately shining, elytra not or faintly iridescent; microsculpture about normal. Head .67 & .70 width prothorax; eyes rather large but only a little prominent, with posterior supraocular setae about between their posterior edges. Prothorax rather small, subquadrate, appearing almost as long as wide 200 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY but width/length actually 1.15 & 1.18; base/apex about 1.41 & 1.32; sides slightly arcuate for most of length, nearly straight and slightly converging basally; posterior angles very obtuse, almost rounded out, partly because sides of base rounded-oblique; lateral margins narrow anteriorly, wider basally, slightly reflexed; posterior-lateral setae set a little in from edges of margins near posterior angles; basal foveae slightly impressed, not distinctly punctate; disc weakly convex, with middle line and transverse impressions only slightly impressed; an- terior and posterior marginal lines nearly entire, but light and more or less interrupted at middle. Elytra long-oval, normally convex, only faintly impressed before middle; basal margin almost rectangular at humeri; lateral margins rather narrow; subapical sinuations almost obliterated, sides of elytra curving in almost evenly to subdenticulate sutural angles; striae moderately impressed, impunctate; intervals moderately convex, 8th and 9th not much modified toward apex, 3rd normally 3-punctate. Lower surface impunctate; abdomen not pu- bescent. Legs: 4th hind-tarsal segment simply emarginate. Measure- ments: length 7.8-7.9; width 2.8-2.9 mm. , Types. Holotype o& (Leiden Mus.) from Mist Camp, Snow Mts., Neth. N. G., 1,800 m. (about 5,850 ft.), Jan. 1939 (Toxopeus); and 1 & paratype (M.C.Z. No. 28,651) from Top Camp, Snow Mts., 2,100 m. (about 6,825 ft.), Jan. 26, 1939 (Toxopeus). Measured specimens. The types. Notes. This species may be related to both dilutipes and planinox (above), and through them to the nox group of Altagonum, but it differs in so many ways that I have given it a full description. The following is a list of the more important differences between the present species and planinox: head relatively narrower, with much less promi- nent eyes; prothorax longer, with posterior-lateral setae set in from edges of margins, and with less impressed anterior and posterior marginal lines; elytra more oval, with humeral margins much more strongly angulate, and with subapical sinuations almost obliterated; abdomen not pubescent. In the face of this list of differences it is obviously unwise to assume a real relationship between this species and planinox. The relationship with dilutipes is more probable. The present species has the form of a slender Dicranoncus, but the tarsal claws are not toothed. ALTAGONUM PALLINOX N. sp. Description. With characters of genus as described above. Form of a large, rather slender, somewhat depressed Bembidion or Europhilus; brown, head slightly darker, abdomen with irregular pale areas, ap- DARLINGTON: CARABID BEETLES OF NEW GUINEA 201 pendages yellowish, lateral marginal gutters of prothorax and elytra narrowly yellowish; surface only moderately shining, not iridescent; microsculpture normal. Head .86 & .85 width prothorax; eyes large, prominent, with posterior supraocular setae about between their posterior edges. Prothorax somewhat transversely subquadrate or subcordate; width/length 1.34 & 1.38; base/apex 1.27 & 1.24; sides less arcuate than usual, moderately sinuate about 1% of length before base; basal angles slightly obtuse but well defined; lateral margins rather narrow; basal foveae poorly defined, only slightly impressed, slightly roughened but not punctate; anterior marginal line faint or interrupted at middle, posterior one vague. Elytra rather narrow, subparallel, a little less convex than usual, with disc broadly, vaguely impressed about 24 from base; basal margin rounded or faintly angu- late at humeri; subapical sinuations moderate; apices moderately rounded, distinctly but finely and bluntly subdenticulate at sutural angles; striae rather deep, not distinctly punctulate; intervals convex, 8th and 9th and usually 7th longitudinally impressed or sulcate toward apex, 3rd normally 3-punctate. Lower surface virtually impunctate; abdomen with a little pubescence near middle near base. Legs: 4th hind-tarsal segment simply emarginate. Measurements: length 6.8- 7.6; width 2.4-2.8 mm. é Types. Holotype & (British Mus.) and 4 paratypes (2 in M.C.Z. No. 28,688) from Cyclops Mts., Neth. N. G., 3,400-4,500 ft., Mar. 1936 (Cheesman). Additional paratypes as follows: Neth. N. G.: 1, Mt. Lina, Cyclops Mts., 3,500-4,500 ft., Mar. 1936 (Cheesman); 1, Cyclops Mts. without more exact data (Cheesman). Papua: 1, Mondo, 5,000 ft., Jan._Feb. 1934 (Cheesman). Measured specimens. The & holotype and 1 Q paratype from the Cyclops Mts. Notes. This is a very distinct species, distinguishable by characters given in the key. I have associated it with the nox group, but I am not sure there is a close relationship. The species has a remarkable superficial similarity to the Notagonwm of the angustellum group, but it differs from them not only in absence of the anterior-lateral pronotal setae but also in impression of the outer elytral intervals toward apex and in form (simply emarginate) of the 4th hind-tarsal segment. ALTAGONUM TUTUM DN. sp. Description. With characters of genus as described above. Form nearly that of small, rather slender Agonum s. s.; piceous-black, legs brownish-piceous, antennae browner, lateral margins of prothorax vaguely translucent; upper surface nearly impunctate, moderately 202 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY shining, not iridescent; microsculpture normal, light on head and pronotum. Head .82 width prothorax; eyes large, moderately promi- nent, with posterior supraocular setae about between their posterior edges; neck-constriction vague. Prothorax rather small, subquadrate, widest about middle, rather strongly narrowed in front; width/length 1.23; base/apex 1.44; anterior angles scarcely advanced; sides moder- ately arcuate for much of length, rather broadly sinuate before base; basal angles approximately right but narrowly rounded; lateral margins rather narrow, moderately reflexed posteriorly; posterior- lateral setae on edges of margins at basal angles; basal foveae normal, vaguely punctate; disc normal; anterior and posterior marginal lines entire. Elytra normal in outline, a little more convex than usual, not impressed on disc; basal margin only vaguely angulate at humeri; lateral margins rather narrow; subapical sinuations moderate, each leading onto a short apical spine almost at sutural angle; striae well impressed, impunctate; intervals moderately convex, 8th and 9th not much modified toward apex, 3rd normally 3-punctate. Lower surface virtually impunctate; abdomen not pubescent. Legs: 4th hind-tarsal segment simply emarginate. Measurements: length 7.7; width 2.8 mm. Type. Holotype 2 (British Mus.) from Mt. Nok, Waigeo Is., Neth. N. G., 2,500 ft., May 1938 (Cheesman); unique. Measured specimen. 'The type. Notes. This is an inconspicuous but safely characterized species, placed in the key to species (above) but not closely related to any other species known to me. It may have originated independently from Notagonum by development of sutural spines and loss of the anterior-lateral pronotal setae. ALTAGONUM CADUCUM D. sp. Description. With characters of genus as described above. Form nearly of elongate Calathus (Fig. 6); brown to brownish-black, ap- pendages brown, lateral margins of prothorax broadly and plainly, of elytra not distinctly yellowish-translucent; upper surface impunctate, moderately shining, not iridescent; microsculpture light and isodia- metric on head, very light (sometimes scarcely visible) and transverse on disc of pronotum, distinct and transverse on elytra. Head .76, .74, & .73 width prothorax; eyes large and moderately prominent, with posterior supraocular setae about between their posterior edges; neck- constriction distinct but not deep. Prothorax strongly narrowed in front, only slightly so behind; width/length 1.33, 1.30, & 1.32; base/ apex 1.46, 1.53, & 1.53; sides rather weakly arcuate for much of length, broadly but slightly sinuate before rectangular, scarcely blunted pos- DARLINGTON: CARABID BEETLES OF NEW GUINEA 203 terior angles; lateral margins rather wide especially toward base, moderately reflexed; posterior-lateral setae on edges of margins almost at basal angles; basal foveae rather deep, not or at most vaguely punctate; disc normal; anterior and posterior marginal lines entire. Elytra rather ample, of average outline and convexity, distinctly im- pressed before middle; basal margin obtusely angulate or subangulate at humeri; lateral margins narrow; subapical sinuations moderate or rather weak; apices each rounded-prominent about opposite 3rd inter- val, minutely denticulate or subdenticulate at sutural angle; striae moderately impressed, impunctate; intervals slightly convex, 8th and 9th not much modified toward apex, 3rd 3-punctate, punctures normally placed except posterior one often not so far back as usual (exact position variable in both Mt. Misim and Snow Mts. series). Lower surface virtually impunctate; abdomen not pubescent. Legs: 4th hind-tarsal segment with a short outer and still shorter inner lobe. Male copulatory organs: Fig. 48. Measurements: length 9.4-11.0; width about 3.5-4.1 mm. Types. Holotype & (M.C.Z. No. 28,652) and 8 paratypes from Mt. Misim, Morobe Dist., N-E. N. G. (Stevens). Additional para- types from Neth. N. G. as follows: 2, Cyclops Mts., 3,400-4,500 ft., Mar. 1936 (Cheesman); and from the Snow Mts.: 8, Sigi Camp, 1,500 m. (about 4,875 ft.), Feb. 1939; 1, Lower Mist Camp, 1,700 m. (about 5,525 ft.), Jan. 17, 1939; 30, 1,800 m. (about 5,850 ft.), Dec. 30, 1938 & Jan. 1939; 2, Top Camp, 2,100 m. (about 6,825 ft.), Jan. 25 & 26, 1939; 1, Ibele (Iebele) Camp, 2,250 m. (about 7,325 ft.), Nov. 1938 (all Snow Mts. specimens collected by Toxopeus). Measured specimens. Holotype o', 1 9 paratype from Mt. Misim, & 1 o& paratype from Mist Camp, Snow Mts. Notes. Although not a striking species, this is a very distinct one apparently not closely related to any other known to me. It may lead toward sphodrum ete. (below), but the outer elytral intervals are not impressed as they are in the sphodrum group; the abdomen is not pubescent as it is in sphodrum and its closest ally, postsulcatum; and the 4th hind-tarsal segment is different, briefly lobed externally in caducum, but simply emarginate in the sphodrum group. The series from Mt. Misim and that from the Snow Mts. agree almost perfectly in size, proportions, and structural details. The Snow Mts. specimens are a little darker, but the difference is slight and inconstant and may be due at least partly to methods of preservation: the Mt. Misim specimens, in alcohol; the Snow Mts. ones, apparently dry. 204 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY ALTAGONUM SPHODRUM DN. sp. Description. With characters of genus as described above. Form of a rather slender, large-headed sphodrine; piceous, appendages brownish, lateral margins of prothorax and elytra scarcely translucent; upper surface virtually impunctate, only moderately shining, not iridescent; microsculpture normal. Head large, appearing almost wide as pro- thorax but actually .82 & .81 as wide; eyes reduced, only moderately prominent; genae about as long as eyes, oblique, straight or slightly convex in profile; posterior supraocular setae behind line of posterior edges of eyes; neck-constriction distinct, moderately impressed; front irregularly convex, with moderate anterior impressions. Prothorax subquadrate; width/length 1.18 & 1.20; base/apex 1.28 & 1.20; sides weakly arcuate anteriorly, broadly sinuate well before basal angles; latter right-acute, accurately formed; lateral margins rather wide (in genus), moderately reflexed (outer edges more so than margins as wholes); posterior-lateral setae on edges of margins almost at basal angles; basalfoveaerather shallow and poorly defined, a little roughened but not punctate; disc rather flat; middle line and transverse im- pressions poorly impressed; anterior and posterior marginal lines vari- able, not well impressed. Elytra rather narrow and long, otherwise about normal in form and convexity, not or only faintly impressed before middle; basal margin rather strongly but a little obtusely angu- late at humeri; lateral margins rather narrow; subapical sinuations very slight; apices independently rounded, with suturalangles variable, sometimes rounded, sometimes obtuse and vaguely denticulate; striae moderately impressed, impunctate; intervals slightly convex, 7th and Sth toward apex and 9th for much of length more or less impressed longitudinally, 3rd usually normally 3-punctate (but only 2-punctate on left elytron of type). Lower surface roughened but virtually im- punctate, except abdomen with fine punctation and extensive but inconspicuous pubescence. Legs: 4th hind-tarsal segment simply emarginate. not lobed but with outer apical angle slightly more prominent than inner one; claws simple (not modified as in many true sphodrines). Male copulatory organs: Fig. 49. Measurements: length 10.5-12.0; width 3.4-4.1 (types only). Types. Holotype o' (M.C.Z. No. 28,653) and 2 2 9 paratypes from Mt. Wilhelm, Bismarck Range, N-E. N. G., 7,000-10,000 ft., Oct. 1944 (Darlington), under cover on the ground in heavy forest. Other material. Besides the types I have a fourth (2) specimen from the same locality which is larger (13.1 by 4.7 mm.), with some- what more prominent eyes and relatively wider prothorax (width/ length 1.29). This specimen is probably referable to sphodrum, but I have not included it in the type series. DARLINGTON: CARABID BEETLES OF NEW GUINEA 205 Measured specimens. The & holotype and 1 2 paratype. Notes. This species is very distinct from any of the preceding ones in form and other characters. It has, however, two apparent relatives, treated below. In describing them IJ have taken sphodrwm as the base- species for comparison. ALTAGONUM POSTSULCATUM hn. Sp. Description. With characters of genus as described above. Ap- parently closely related to the preceding (sphodrum) and answering to the same description (see above) in all details except those noted below. Form more Calathus-like than sphodrine; color about as in sphodrum except elytra more shining or even faintly iridescent in strong light; microsculpture same except finer and more transverse on elytra. Head relatively narrower than in sphodrum, .77 and .76 width pro- thorax, but with larger and more prominent eyes and short, oblique genae; posterior supraocular setae only slightly if at all behind line of posterior edges of eyes. Prothorax much more narrowed in front; width/length 1.30 & 1.38; base/apex 1.40 & 1.48; sides weakly arcuate, straight and slightly or scarcely converging and usually slightly sinuate toward base; basal angles right or slightly obtuse; disc normal; anterior and posterior marginal lines entire. Elytra slightly broader than in sphodrum, sometimes more distinctly impressed before middle; apices usually vaguely or distinctly denticulate; outer intervals impressed about as in sphodrum. Lower surface less roughened than in sphodrum and abdomen less extensively pubescent, but still plainly so especially near middle basally. Legs: 4th hind-tarsal segment emarginate. Measurements: length 9.0-9.9; width 3.3-3.7 mm. Types. Holotype o (Leiden Mus.) and 9 paratypes (some in M.C.Z. No. 28,654) from Mist Camp, Snow Mts., Neth. N. G., 1,800 m. (about 5,850 ft.), Jan. (some paratypes Dec. 30, 1938, & Jan. 7 & 9) 1939 (Toxopeus); and the following additional paratypes all from the Snow Mts.; 5, Sigi Camp, 1,500 m. (about 4,875 ft.), Feb. (including Feb. 19 & 25) 1939; 1, mountain slope above Bernhard Camp, 1,700 m. (about 5,525 ft.), Jan. 7, 1939; 1, Top Camp, 2,100 m. (about 6,825 ft.), Feb. 8, 1939; and 4, Ibele (Iebele) Camp, 2,250 m. (about 7,325 ft.), Nov. & Nov.—Dec. 1939 (all specimens collected by Toxopeus). Other material. One 2 from Mt. Wilhelm, Bismarck Range, N-E. N. G., 7,000-10,000 ft., Oct. 1944 (Darlington), in forest. This speci- men differs slightly from the types in form and has the elytra a little less shining, with less strongly transverse microsculpture, but these and other differences are so small that I do not care to make them the basis of even a new subspecies without more material. The proportions 206 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY of this specimen are: head/prothorax .73; prothoracic width/length 1.36; base/apex 1.38. Its length is 10.0 mm. Measured specimens. Holotype o& and 1 @ paratype from Mist Camp. Notes. This species is adequately distinguished from sphodrum and others in the key to species of Altagonum, above. ALTAGONUM MISIM 0. sp. Description. With characters of genus as described above. Probably related to sphodrum; differing from latter in many details, but described partly by comparison in order to stress both similarities and differences. Form only vaguely sphodrine, more Colpodes-like; color, and sculpture of upper surface about as in sphodrum. Head smaller but with much larger eyes; .79 & .73 width prothorax; eyes large, prominent, with posterior supraocular setae before line of their posterior edges; neck- constriction scarcely indicated; front normally convex, with rather slight anterior impressions. Prothorax wider and relatively more narrowed in front than in sphodrum; width/length 1.36 & 1.44; base/ apex 1.44 & 1.47; sides moderately arcuate for much of length, more or less strongly sinuate near base; basal angles right; lateral margins moderate, wider basally, moderately reflexed; posterior-lateral setae as in sphodrum; basal foveae deeper than in sphodrum, indistinctly punctate; disc normal; anterior and posterior marginal lines entire, well impressed. Elytra long and rather ample, a little less convex than usual, slightly impressed before middle; basal margin obtusely angulate or subangulate at humeri; lateral margins rather narrow; subapical sinuations moderate; apices somewhat produced, each with a short tooth or abrupt angulation about opposite 3rd interval and a stronger tooth (almost a short spine) at sutural angle; striae and intervals about as in sphodrum, with outer intervals impressed in about same way. Lower surface virtually impunctate; abdomen not pubescent. Legs: Ath hind-tarsal segment simply emarginate, as in sphodrum. Measure- ments: length 12.4-13.7; width 4.2-5.0 mm. Types. Holotype 2 (M.C.Z. No. 28,655) and 1 @ paratype from Mt. Misim, Morobe Dist., N-E. N. G., the type specifically from 6,400 ft., Mar. (Stevens). Measured specimens. The types. Notes. This species probably belongs near sphodrum and _ post- sulcatum (above) because of the impressed outer elytral intervals and form of prothorax, which agrees closely with that of postsulcatuim though not of sphodrum itself, but misim differs from both these species in having elytra armed at apex and abdomen not pubescent, as well DARLINGTON: CARABID BEETLES OF NEW GUINEA 207 as in other less significant details: larger eyes, virtually no neck- constriction, etc. ALTAGONUM CHEESMANI N. sp. Description. With characters of genus as described above. Rather large, rather slender, subfusiform; black, elytra (except suture) green, appendages yellowish- or brownish-red, lateral margins of prothorax yellowish-translucent; upper surface impunctate except slightly punctate in basal foveae and marginal gutters of pronotum, moderately shining, not iridescent; microsculpture of head and pronotal disc not visible at 54, of elytra distinct and transverse. Head .80 width pronotum; eyes large and very prominent, with posterior supraocular setae slightly before line of their posterior edges; neck-constriction slight and not well defined; front slightly convex, with moderate anterior impressions. Prothorax widest about middle, strongly narrowed in front, only slightly so behind; width/length 1.27; base/ apex 1.69; anterior angles not at all advanced, so front of prothorax almost evenly truncate; sides moderately rounded for much of length, slightly sinuate toward base; basal angles right-obtuse, blunted; lateral margins rather wide but not well defined, slightly reflexed anteriorly, rather strongly so posteriorly; posterior-lateral setae on edges of margins at posterior angles; basal foveae deep, slightly punctate; disc normal, with light middle line and rather deep transverse impressions; anterior and posterior marginal lines fine but entire. Elytra long, subparallel for much of length, more convex than usual, not impressed on disc; basal margin obtusely angulate at humeri; lateral margins narrow; subapical sinuations moderate, broad, each leading onto a short, strong spine opposite 3rd interval; apices inside of spines strongly emarginate; sutural angles denticulate; striae light, punctulate; inter- vals flat, 8th and 9th not much modified toward apex, 3rd normally 3-punctate. Lower surface distinctly punctate only at sides of meso- sternum; abdomen not pubescent. Legs: 4th hind-tarsal segment rather deeply emarginate, almost lobed, but outer angle or lobe not obviously longer than inner. Measurements: length 13.1; width 4.3 mm. Type. Holotype 2 (British Mus.) from Mafulu, Papua, 4,000 ft., Jan. 1934 (Cheesman); unique. Measured specimen. The type. Notes. This species, although placed among others in the key to species of Altagonwm (above), is quite unlike any other known to me. It may be independently derived from Colpodes. It is more fusiform than any Colpodes known from New Guinea, however, and of course it lacks the anterior-lateral pronotal setae. It is a fine species, and I 208 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY take great pleasure in naming it for Miss L. Evelyn Cheesman, who collected it and so many other interesting Carabidae in New Guinea. ALTAGONUM SCAPHA N. sp. Description. With characters of genus as described above. Rather small, slender, fusiform; brownish-black, legs only slightly paler, antennae yellowish, sides and sometimes base of prothorax narrowly, vaguely paler; upper surface impunctate, moderately shining, not distinctly iridescent; microsculpture normal but light. Head narrow, .65, .62, & .63 width prothorax; eyes large but less convex than usual, not very prominent, with posterior supraocular setae about between their posterior edges; neck constriction vague; front normal. Prothorax much narrowed in front, scarcely at all so behind; width/length 1.24, 1.28, & 1.29; base/apex 1.67, 1.80, & 1.70; anterior angles scarcely advanced; sides moderately arcuate anteriorly, nearly straight pos- teriorly, not distinctly sinuate; basal angles almost right except narrowly rounded; lateral margins narrow, scarcely reflexed; posterior- lateral setae on flat surface a little in from edges of margins near basal angles; basal foveae slight, almost obsolete; disc only weakly convex, with median line and transverse impressions weak; anterior and pos- terior marginal lines light or interrupted at middle. Elytra elongate- oval, more pointed behind and more convex than usual; basal margin about rectangular at humeri; lateral margins narrow; subapical sinu- ations variable, present only when elytra spined; apices each with a short spine or acute angulation opposite 3rd interval, then oblique forward and more or less emarginate to more or less strongly denti- culate sutural angles; striae rather lightly impressed, not punctate; intervals flat or slightly convex, 8th and 9th not much modified toward apex, 3rd 2-punctate (anterior puncture absent on both sides in all specimens). Lower surface impunctate; abdomen not pubescent; prosternal process simple or vaguely margined at apex. Legs: hind tarsi very slender; 4th hind-tarsal segment emarginate or with very short lobes, with outer angle always a little longer than inner. Measure- _ ments (types): length 7.2-8.0; width 2.6-2.9 mm. Types. Holotype o' (Leiden Mus.) from Rattan Camp, Snow Mts., Neth. N. G., 1,150 m. (about 3,750 ft.), Feb——Mar. 1939 (Toxopeus) ; 1 & paratype (M.C.Z. No. 28,656) from Top Camp, Snow Mts., 2,100 m. (about 6,825 ft.), Jan. 22, 1939 (Toxopeus); 1 Q paratype (British Mus.) from Camp Nok, Waigeo Is., 2,500 ft., Apr. 1938 (Cheesman). Other material. The type of scapha has the elytra spined; the paratype, acutely angulate. The @ paratype from Waigeo Is. has DARLINGTON: CARABID BEETLES OF NEW GUINEA 209 them spined and matches the type in all other significant non-sexual characters. I have examined also 2 additional, larger specimens from the Snow Mts. In one, a o from Ibele (Iebele) Camp, 2,250 m. (about 7,325 ft.), Nov. 1938 (Toxopeus), the apical angulations of the elytra are slightly obtuse rather than acute, less prominent even than in the o paratype, and the proportions are head .61 width prothorax, pro- thoracic width/length 1.15 and base/apex 1.81, and length 9.3, width 3.1 mm. In this o’ and in the 2 oc" types the apex of the aedeagus is recurved or barbed below, the exact form being a little different in each specimen, but within the possible range of variation of a species. The final specimen is a 2 from Mist Camp, 1,800 m. (about 5,850 ft.), Jan. 1939 (Toxopeus). It has the elytra spined as in the type, but it differs slightly in proportions and is much larger: head .57 width pro- thorax, prothoracic width/length 1.26, base/apex 1.73, length 11.38, width 4.0 mm. The significance of these variations cannot be de- termined without more material. Measured specimens. ‘The 3 types. Notes. This species is sufficiently compared with others in the key to species of Altagonum, above. ALTAGONUM REGISCAPHA N. sp. Description. With characters of genus as described above. Similar to scapha, to the description of which (see above) it answers in all details except those noted below. Broader than scapha; elytra incon- spicuously purple with greenish reflections, legs dark reddish; micro- sculpture of head and pronotum so light as to be scarcely visible. Head with eyes somewhat more prominent than in scapha, .65 width prothorax. Prothorax relatively wider but otherwise similarly formed; width/length 1.36; base/apex 1.78; basal angles slightly more obtuse; lateral margins wider; posterior-lateral setae virtually on (not well inside of) edges of margins at basal angles; basal foveae more distinct, broad, but still shallow and poorly defined. Elytra slightly broader and less pointed behind; subapical sinuations virtually absent; apices rectangular about opposite 3rd intervals, then oblique forward to slightly denticulate sutural angles; 3rd interval 2-punctate as in scapha. Legs: 4th hind-tarsal segment simply emarginate, not lobed. Measure- ments: length 9.1; width 3.4 mm. Type. Holotype 9 (Leiden Mus.) from Mist Camp, Snow Mts., Neth. N. G., 1,800 m. (about 5,850 ft.), Jan. 1939 (Toxopeus); unique. Measured specimen. 'The type. Notes. This species is very similar to scapha, but whether it is an actual relative or a convergent form I do not know. The difference 210 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY in position of the posterior-lateral setae and the somewhat different form of the 4th hind-tarsal segment suggest that the relationship may not be very close. ALTAGONUM LATILIMBUS DN. sp. Description. With characters of genus as described above. Form of rather broad Calathus; brownish-black, broadly margined with yellow (yellow covers actual margins of prothorax, margins and about 3 outer intervals in basal half of elytra, and a still wider zone posteriorly, including elytral apices; boundary of dark discal area fairly regular, dark color not forming a narrow sutural stripe toward apex), sides and apex of abdomen narrowly margined or spotted with yellow, appen- dages yellow; upper surface impunctate, only moderately shining, not distinctly iridescent; microsculpture normal. Head small, .57 & .56 width prothorax; eyes moderately large and prominent, with posterior supraocular setae about between their posterior edges; neck not con- stricted above; front weakly convex, with weak anterior impressions. Prothorax rather broad, much narrowed in front, scarcely or not narrowed behind; width/length 1.48 & 1.44; base/apex 1.61 & 1.59; anterior angles broadly and rather strongly advanced, subacute except narrowly rounded; sides weakly arcuate for much of length, almost straight and sometimes very slightly sinuate posteriorly; basal angles subrectangular except very narrowly rounded; lateral margins narrow anteriorly, wide and flat posteriorly; posterior-lateral setae on flat surfaces of margins about equally distant from inner and outer edges and base, or nearer inner edges; basal foveae almost obsolete, not distinet from flattened margins, not or at most vaguely punctate; disc moderately convex, with middle line and transverse impressions slight; anterior marginal line entire, posterior one faint or interrupted at middle. Elytra nearly normal in outline and only slightly more convex than usual, with disc not or faintly impressed; basal margin almost rectangular at humeri (but points of angles blunted); lateral margins average; subapical sinuations very slight; apices simple, with sutural angles obtuse, not or faintly denticulate; striae rather lightly impressed, impunctate; intervals nearly flat or slightly convex, 8th and 9th not much modified toward apex, 3rd normally 3-punctate. Lower surface impunctate; abdomen not pubescent. Legs: 4th hind-tarsal segment emarginate, not lobed. Measurements: length 8.5-8.9; width 3.4-3.5 mm. Types. Holotype o (Leiden Mus.) and 2 oo paratypes (1 in M.C.Z. No. 28,657) from Mist Camp, Snow Mts., Neth. N. G., 1,800 m. (about 5,850 ft.), Jan. (the paratypes Jan. 9 & 11) 1939; DARLINGTON: CARABID BEETLES OF NEW GUINEA 211 1 & paratype from Sigi Camp, Snow Mts., 1,500 m. (about 4,875 ft.), Feb. 25, 1939; and 1 co paratype and hind-body of another specimen from Top Camp, Snow Mts., 2,100 m. (about 6,825 ft.), Jan. 27 & Feb. 8, 1939; (all specimens collected by Toxopeus). Measured specimens. The & holotype and 1 & paratype from Mist Camp. Notes. So far as I know, this species has only one close relative, described below. It may be more distantly related to scapha (above), which also is fusiform, with posterior pronotal setae somewhat removed from the edges of the margins, but latilimbus is very different from scapha in coloration, strongly advanced anterior prothoracic angles, simple elytral apices, and other details. ALTAGONUM PARALIMBUS N. sp. Description. Very close to latilimbus, to the description of which (see above) it answers in all details not noted below. Slightly narrower than latilimbus; yellow margins of elytra narrower anteriorly (including only about 2 intervals), broad but irregularly limited posteriorly, with dark discal color extending along sutural intervals toward (but not to) apex; surface a little more shining than in latilimbus. Head .57 & .57 width prothorax; eyes slightly less prominent than in Jatilimbus. Prothorax: width/length 1.32 & 1.34; base/apex 1.51 & 1.48. Elytra with disc a little more distinctly impressed slightly before middle than in latilimbus. Male copulatory organs: Fig. 50. Measurements: length 9.9-10.2; width 3.7-3.8 mm. Types. Holotype o' (M.C.Z. No. 28,658) from Mt. Misim, Morobe Dist., N-E. N. G. (Stevens); and 1 @ paratype (British Mus.) from Mt. Tafa, Papua, 8,500 ft., Mar. 1934 (Cheesman). Measured specumens. The types. Notes. As compared with latilimbus, this differs chiefly in its more slender form and more irregular boundary of dark and light areas toward apex of elytra. ALTAGONUM NUDICOLLE N. sp. Description. With characters of genus as described above. Form of rather broad Calathus; dark-brown or brownish-black, legs not much paler, antennae a little redder, lateral margins of prothorax slightly reddish-translucent; surface not very shining, not iridescent; micro- sculpture very distinct, normal except meshes less transverse than usual on elytra. Head .67 & .64 width prothorax; eyes moderately large and prominent, with posterior supraocular setae about between 212 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY their posterior edges; neck slightly impressed above; front somewhat irregularly convex, with slight anterior impressions. Prothorax widest near middle, strongly narrowed anteriorly, scarcely if at all so pos- teriorly; width/length 1.33 & 1.35; base/apex 1.65 & 1.64; anterior angles moderately advanced, would be right or slightly acute except narrowly rounded; sides arcuate for much of length, then straight or slightly and broadly sinuate before base; posterior angles approxi- mately right, scarcely blunted; lateral margins rather wide especially toward base but only moderately reflexed; both pairs (posterior as well as anterior) lateral setae absent; basal foveae slight, scarcely distinet from ends of lateral margins, impunctate; disc less convex than usual; anterior marginal line entire, posterior one usually entire but some- times vague at middle. Elytra rather broad and ample, nearly normal in outline but more convex than usual; basal margin rectangular at humeri; lateral margins narrow; subapical sinuations slight or nearly absent; apices rather narrowly rounded; sutural angles obtuse or rounded; not denticulate; striae lghtly impressed, not punctate; intervals flat or nearly so, more or less vaguely longitudinally impressed at extreme apices, but 9th not or only vaguely impressed, 3rd normally 3-punctate. Lower surface impunctate; abdomen not pubescent. Legs: 4th hind-tarsal segment with short outer and still shorter inner lobe. Male copulatory organs: Fig. 51. Measurements: length 11.2-12.3; width 4.3-5.1 mm. Types. Holotype o (M.C.Z. No. 28,659) and 67 paratypes all from Mt. Wilhelm, Bismarck Range, N-E. N. G., 7,000-10,000 ft., Oct. 1949 (Darlington), in and under various ground-cover in forest. Measured specimens. The o& holotype and 1 Q paratype. Notes. The slight impression of the elytral intervals and some other details suggest that this species may be related to sphodrum and postsulcatum (above), but it is much broader, without abdominal pubescence, and has lost the posterior- as well as the anterior-lateral pronotal setae. ALTAGONUM FATUUM N. sp. Description. With characters of genus as described above. Form nearly that of slender Platynus, with slender appendages; brownish- black, elytra browner, appendages yellow, lateral margins of prothorax and elytra yellow-translucent, epipleurae yellow, sides and apex of abdomen broadly yellow; upper surface virtually impunctate, moder- ately shining, not iridescent; microsculpture normal, distinet. Head rather long, .74 & .72 width prothorax; eyes slightly reduced, only moderately prominent; genae about half as long as eyes, oblique; DARLINGTON: CARABID BEETLES OF NEW GUINEA 213 posterior supraocular setae slightly behind line of posterior edges of eyes; neck slightly impressed above; front normal. Prothorax rather narrow, subquadrate, moderately narrowed anteriorly, only slightly so posteriorly; width/length 1.09 & 1.14; base/apex 1.40 & 1.43; anterior angles scarcely advanced; sides rather weakly arcuate for much of length, more or less straight and sometimes faintly sinuate toward base; posterior angles rather narrowly rounded; lateral margins rather narrow, only a little wider toward base, moderately reflexed; both pairs lateral pronotal setae absent; basal foveae normal, a little irregular at bottom but not distinctly punctate; disc normal; anterior and posterior marginal lines entire. Hlytra long but otherwise of nearly normal outline, slightly more convex than usual, with disc not im- pressed; basal margin distinctly but somewhat obtusely angulate at humeri; lateral margins narrow; subapical sinuations faint or absent; apices moderately rounded to suture; sutural angles not or faintly denticulate; striae moderately impressed, not punctate; intervals nearly flat or slightly convex, 8th and 9th not much modified toward apex, 3rd impunctate. Lower surface impunctate; abdomen not pu- bescent. Legs: 4th hind-tarsal segment emarginate, not lobed. Measurements: length 10.0-10.3; width 3.6-3.9 mm. Types. Holotype o (British Mus.) and 2 (&@) paratypes (c” in M.C.Z. No. 28,660) all from Mt. ‘Tafa, Papua, 8,500 ft., Mar. 1934 (Cheesman). Measured specimens. The & holotype and 1 @ paratype. Notes. This species, although placed in the key to species of Alta- gonum (above), is very different from any other known to me. MACULAGONUM new genus Diagnosis. Rather small (5.7 to 8.9 mm), either narrow Agonum- like or more or less fusiform; elytra always mottled or blotched with dark and pale; wing-and-seta formula +w, ++, (—)+, —(+)+; last ventral abdominal segment of o rather deeply notched at middle of apex; genus otherwise within range of variation of Altagonum (above). Description. Form and color as indicated above; upper surface impunctate (except in setipox), more or less shining, not iridescent; microsculpture variable. Head moderate or small, usually rather short (in tribe), sometimes slightly elongate; neck more or less impressed above except in scaphipou; eyes variable, either large, or slightly re- duced in size and prominence, or small but abruptly prominent; both pairs supraocular setae present; antennae normal; mentum tooth triangular. Prothorax somewhat variable, more or less strongly 214 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY narrowed in front (least so in pow), with anterior angles not advanced, more or less rounded out; anterior-lateral setae absent (except in setipox), posterior-lateral ones present on or just within basal angles. Elytra long and/or ample, more convex than usual; dise not distinctly impressed; basal margin entire, variably angulate at humeri; apices variable but never spined nor abruptly angulate; intervals not much modified toward apex (except 8th more convex than usual toward apex in scaphipox); 3rd interval 2- or 1-punctate (anterior and sometimes middle puncture absent). Inner wings full. Lower surface impunctate; abdomen not pubescent; prosternal process simple. Legs: hind tibiae not suleate along outer edges; 4th hind-tarsal segment emarginate, not lobed; 5th hind-tarsal segment without obvious accessory setae; claws simple. Secondary sexual characters normal except last ventral ab- dominal segment more or less deeply notched at middle of apex in o’, though not in 9. Male copulatory organs as figured (Fig. 52). Genotype. Maculagonum pox n. sp. (below). Generic distribution. Mountains of New Guinea. Notes. Of the six species of this new genus, I have seen males of only four, but the four include the extremes of generic variation in most other characters. All these males have the last ventral segment more deeply and abruptly emarginate than males of any Altagonwm that I have been able to examine (as a rule the male last ventral is entire in Altagonum, but it is slightly emarginate in scapha and some- times in vallicola and perhaps in other species). I think that this character will probably be found to hold for males of all species of Maculagonum. The mottling or blotching of the elytra is the most obvious character of this genus. It may be an adaptation to life in grass, tending to conceal the insects in finely divided light and shade. At least the only specimen of the genus that I have collected (the type of altipox) was found in a tussock on a grassy slope above timber line on Mt. Wilhelm. Some of the specimens of Maculagonum collected by Cheesman and Toxopeus were probably taken in light traps, for they have seales of Lepidoptera stuck to them. Key to the Species of Maculagonuim 1. Lateral margins of prothorax moderately wide; length 8.2 to 8.9 mm... .2 — Lateral margins of prothorax very narrow; length 5.7 to 7.7 mm........ 3 2. Elytra finely mottled with dark and pale (p. 215).............25.. 20; pox — Elytra dark with a large common pale blotch before middle and another about) 4cfrom\apexi (4 216)» aye Sih nae gts eee plagipox 3. Anterior-lateral as well as posterior-lateral pronotal setae present; eyes small butrabnormalliy, prominent (27) seen eee eee setipox DARLINGTON: CARABID BEETLES OF NEW GUINEA AMS) — Anterior-lateral pronotal setae absent; eyes larger but only moderately POLO MAI SING eerste ery Hae AN Sts vaya Ras vendita ay ote SUE OA al os Pierre cee nk a et 4 4, Apices of elytra more broadly and evenly rounded; elytral microsculpture deeply aumpressedsmisodiametrichyaan hmm ann nen iceee eee ee: altipox (4a) Darker, with dark areas of elytra more extensive and blackish or Gark=browin (os 2S) crea teal eee neo ea (altipox s. s.) (4b) Paler, with dark areas of elytra less extensive and paler brown; (see alsond Scrip tion) m(os29)) eee ene eee (subsp. pallipozx) — Apices of elytra more narrowly and abruptly rounded (lobed) about opposite 4th intervals; elytral microsculpture less deep, somewhat transverse. . .5 5. Slender, not fusiform; prothoracic width/length 1.19; 3rd elytral interval HE pumctatex (pipe li9) ere te. oe sae ae ev Ree eae Nie, a Pee le aE er tafapox _- Fusiform; prothoracic width/length 1.46; 38rd elytral interval 2-punctate (222 0) PN te asa hc ess calindn, Telia Spin Ue a ea ae scaphipox MAcULAGONUM POX Nn. sp. Description. With characters of genus as described above. Form (Fig. 7) somewhat Agonum-like, but more elongate-oval than usual in Agonum s. s.; more or less dark-brown, lateral margins of prothorax yellow-translucent, elytra mottled with small yellowish spots which sometimes form rows along intervals or sometimes anastomose, ap- pendages yellowish-brown; microsculpture nearly normal but faint on head and pronotum, more distinct and only moderately transverse on elytra. Head .74 & .72 width prothorax; eyes large and prominent, with posterior supraocular setae about between their posterior edges; front convex, with usual slight anterior impressions and often (not always) also with an extra impression each side near eye, behind and inside of anterior supraocular seta. Prothorax of moderate size, much narrowed in front and moderately so behind; width/length 1.85 & 1.40; base/apex about 1.60 (angles too rounded for accurate measurement) ; sides nearly evenly rounded for all or most of length, sometimes more or less straight but hardly sinuate toward base; posterior angles obtuse, blunted or narrowly rounded; lateral margins moderately wide, wider toward base, only slightly reflexed; basal foveae roundish, shallow, impunctate; disc normal; anterior and posterior marginal lines light, sometimes interrupted at middle. Elytra rather long; basal margin rather strongly but somewhat obtusely angulate at humeri; lateral margins narrow; subapical sinuations obsolete or nearly so; apices rather strongly rounded about opposite 3rd intervals, then oblique forward and obtusely angulate or subdenticulate at suture; striae rather lightly impressed, impunctate (sometimes appearing punctate in soft specimens); intervals flat or slightly convex, 3rd 2-punctate. Secondary sexual characters normal except last ventral abdominal 216 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY segment of co abruptly notched at middle of apex (bottom of notch would be right or slightly acute if not narrowly rounded), of Q entire. Measurements: length 8.2-8.9; width 3.1-3.3 mm. Types. Holotype Q (Leiden Mus.) and 1 & 3 2 9 paratypes (2 99 in M.C.Z. No. 28,661) from Sigi Camp, Snow Mts., Neth. N. G., 1,500 m. (about 4,875 ft.), Feb. 17 (holotype), 19, & 25, 1939 (Giowoneus): and 1 Q paratype from Mist Camp, Snow Mts., 1,800 m. (about 5,850 ft.), Jan. 9, 1939 (Toxopeus). Measured specimens. The 2 holotype and naka jee Notes. This species is sufficiently compared with others in the key, above. The single o and 3 of the 4 Q 2 paratypes are so soft that their elytra have crumpled or are at least not well formed. The- copulatory organs of the o7 are too unformed to draw. MacuLAGONUM PLAGIPOX N. sp. Description. With characters of genus as described above. A little more slender than poz; color (of single, soft specimen) brown; lateral margins of prothorax yellow-translucent; elytra with 2 common, trans- versely oval, pale plagiae involving the 4 inner intervals of each elytron, centering slightly before middle of elytral length and about 14 from apex; appendages yellow; microsculpture absent or very faint on head and pronotum, moderately transverse on elytra. Head .77 width pro- thorax; eyes large and prominent, with posterior supraocular setae about between their posterior edges; front broadly impressed each side from behind anterior supraocular seta to clypeal suture (but im- pressions may be due to warping of soft integument). Prothorax a little longer and less narrowed behind than in pow; width/length 1.23; base/apex about 1.63; sides rather weakly arcuate anteriorly, nearly straight and scarcely converging behind middle; posterior angles a little obtuse and rather narrowly rounded; lateral margins moderate, wider toward base, only slightly reflexed; anterior and posterior marginal lines entire. Elytra warped, long but apparently of nearly normal outline; basal margin obtusely subangulate at humer!; lateral margins wider than in pox; apices apparently formed as in pow; striae appear rather deeply impressed and somewhat punctate, but might be lighter and impunctate in fully hardened specimens; intervals abnormally warped in single specimen, 3rd 2-punctate. Legs: 4 outer segments of both hind tarsi missing, but structure presumably as in other species of genus. Secondary sexual characters: last ventral segment of o notched as in pox. Measurements: length 8.6; width probably about 3.0 mm. DARLINGTON: CARABID BEETLES OF NEW GUINEA Dalen Type. Holotype @ (British Mus.) from Cyclops Mts., Neth. N. G., 3,500 ft., Mar. 1936 (Cheesman); unique. Measured specimen. The type. Notes. Although this species is represented by a single soft and somewhat warped specimen, and although it is structurally fairly close to pox, I have no doubt it is distinct. It differs from pox not only in color-pattern but also in form especially of the prothorax, and in greater width of the elytral margins. I have not wanted to risk re- mounting the single soft specimen, which is glued to a small card, and so have been unable to examine the lower surface except from the side. Fortunately the form of the apex of the last ventral segment can be seen from above. Also, I have not risked dissecting out the genitalia, which would probably be too unformed for study in any case. MACULAGONUM SETIPOX N. sp. Description. With characters of genus as described above. Half- fusiform, more than usually tapering anteriorly but not posteriorly; brownish-piceous with slight aeneous lustre; lateral margins of pro- thorax somewhat translucent posteriorly only; elytra mottled with paler brown, many of the pale marks confined to single intervals, others coalescing to form irregular blotches; appendages yellowish; micro- sculpture faint on head and pronotum, deeply impressed and isodia- metric on elytra. Head .78 width prothorax; eyes rather small but abruptly prominent, with posterior supraocular setae far behind line of their posterior margins; vertex somewhat swollen; front broadly, slightly, transversely impressed between eyes, and with a long antericr impression extending from above each eye to base of clypeus. Pro- thorax broadest at extreme base; width/length 1.10; base/apex about 1.67; sides weakly arcuate anteriorly, straight and slightly diverging in posterior half; posterior angles right, well defined; base slightly lobed at middle, slightly oblique at sides to angles; lateral margins very narrow, each with a short seta before middle as well as one at basal angle; basal foveae rather shallow and poorly defined, they and adjacent areas of pronotum punctate; disc more wrinkled (transversely ) than usual, otherwise normal; anterior and posterior marginal lines entire. Elytra with basal margin obtusely subangulate at humeri; lateral margins narrow; subapical sinuations slight; apices beyond sinuations nearly evenly, subindependently rounded to obtuse, slightly blunted sutural angles; striae shallow, vaguely punctulate; intervals flat, 3rd 2-punctate. Secondary sexual characters: last ventral abdomi- nal segment of o emarginate at middle of apex, with bottom ot 218 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY emargination almost rectangular. Male copulatory organs: Fig. 52. Measurements: length 7.7; width 2.9 mm. Type. Holotype & (British Mus.) from Mt. Tafa, Papua, 8,500 ft., Feb. 1934 (Cheesman); unique. Measured specimen. The type. Notes. This is an exceptionally distinct species, unique within the genus in form, form of eyes, impressions of front, presence of both pairs of lateral pronotal setae, and punctation and extra wrinkling of pronotum. MACULAGONUM ALTIPOX N. sp. Description. With characters of genus as described above. Form intermediate, between normal Agonum-like and (anteriorly) subfusi- form; dark-brown, pronotum with anterior and lateral margins narrowly and base widely yellow, the latter finely speckled with dark; elytra mottled with many small yellow spots, the larger ones very irregular in outline, forming rows along intervals, and anastomosing very little; appendages pale, except antennal segments 2 to 4 infuseate basally; microsculpture light but normal on head and pronotum, deeply impressed and isodiametric on elytra. Head .76 width prothorax; eyes only slightly shorter but less prominent than in poz, with posterior supraocular setae slightly behind line of their posterior edges; neck only slightly impressed above; front convex, with only slight and irregular anterior impressions. Prothorax small, widest in basal half, slightly tapering anteriorly; width/length 1.28; base/apex about 1.64; sides slightly arcuate anteriorly, almost straight and parallel in almost basal half; base broadly lobed at middle, somewhat oblique at sides; posterior angles slightly obtuse, slightly blunted; lateral margins very narrow; immediate baso-lateral areas (just inside angles) rather broadly swollen rather than depressed, not punctate, but basal area impressed each side inside of swelling; disc normal, but median line and transverse impressions slighter than usual; anterior and posterior marginal lines light, vague or interrupted at middle. Elytra relatively ample; basal margin obtusely angulate at humeri; lateral margins very narrow; subapical sinuations almost obsolete; apices broadly and almost conjointly rounded to obtuse sutural angles; striae light, not distinctly punctate; intervals flat or nearly so, 3rd 1-punctate. Second- ary sexual characters: & unknown; 2 with last ventral abdominal segment broadly subtruncate, virtually entire. Measurements: length 7.0; width about 2.7 mm. Type. Holotype @ (M.C.Z. No. 28,662) from Mt. Wilhelm, Bismarck Range, N-E. N. G., over 10,000 ft. (on open grassy slope just above timber line), Oct. 1944 (Darlington). DARLINGTON: CARABID BEETLES OF NEW GUINEA 219 Measured specumen. The type. Notes. This species is sufficiently compared with others in the key to species of Maculagonum, above. MAcULAGONUM ALTIPOX PALLIPOX n. subsp. Description. Similar to typical altipox but paler, with pale spots of elytra more extensive and tending to coalesce in certain areas, es- pecially in a common sutural blotch about 14 from apex, and along outer margins; microsculpture as in typical altipor. Head .79 & .80 width prothorax, formed about as in altipox. Prothorax almost as in altipox but slightly narrowed behind; width/length 1.24 & 1.27; base/apex about 1.50 & 1.57; baso-lateral areas only slightly swollen. Elytra about as in altipox but with apices more independently, less conjointly rounded; 3rd interval similarly 1-punctate. Secondary sexual characters: normal; and last ventral abdominal segment of o’ moderately emarginate at apex, the bottom of the emargination obtuse; last ventral of 2 subtruncate. Measurements: length 6.9-7.2; width about 2.8 mm. Types. Holotype o (Leiden Mus.) and 1 Q paratype (M.C.Z. No. 28,663) both from Moss Forest Camp, Snow Mts., Neth. N. G., 2,800 m. (about 9,100 ft.), Oct. 9-Nov. 5, 1938 (Toxopeus). Measured specimens. The types. Notes. Sufficiently compared with typical altipox above. The two specimens of pallipox are not quite fully hardened, and this may affect the intensity of the color, but not the pattern. I have dissected out the genitalia of the o type, but they are not hard enough to show characters properly. MACULAGONUM TAFAPOX N. sp. Description. With characters of genus as described above. Small and slender, almost Europhilus-like but more convex, with prothorax less narrowed behind and elytra more oval than usual in Ewrophilus; brown; prothorax with anterior margin less and sides and posterior margin more distinctly margined with yellow; elytra primarily yellowish with dark brown mottling especially on disc, and with dark marks coalescing to form an irregular blotch between 3rd and 7th striae about 14 from apex on each elytron; appendages yellowish except antennal segments 3, 4, and less distinctly 5 infuscate basally; micro- sculpture normal except only slightly transverse on elytra. Head .77 width prothorax; eyes somewhat smaller and much less prominent than in pox, but with posterior supraocular setae scarcely behind line 220 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY of their posterior edges; otherwise about as in pox. Prothorax small, subquadrate except rather strongly narrowed near front; width/length 1.19; base/apex about 1.46; sides only moderately arcuate anteriorly, straight and slightly converging behind middle; posterior angles slightly obtuse, blunted or very narrowly rounded; lateral margins very narrow; basal foveae poorly defined and only slightly impressed, with surface slightly swollen just inside angles; disc normal; anterior and posterior marginal lines more or less entire but not well marked. Elytra elongate-oval; basal margin almost rectangular at humeri; lateral margins very narrow; subapical sinuations slight; apices rather strongly lobed about opposite 4th intervals, then oblique forward to obtuse sutural angles; striae well marked but only lightly impressed, not distinctly punctate; intervals flat or nearly so, 3rd 1-punctate. Secondary sexual characters: & unknown, 2 with last ventral abdomi- nal segment entire. Measurements: length 5.7; width 2.0 mm. Type. Holotype @ (British Mus.) from Mt. Tafa, Papua, 8,500 ft., Feb. 1934 (Cheesman); unique. Measured specimen. The type. Notes. This species is sufficiently differentiated from others in the key to species of Maculagonum. MAcULAGONUM SCAPHIPOX N. sp. Description. With characters of genus as described above. Fusiform, convex; brownish-piceous; prothorax with anterior margin faintly and lateral and posterior margins more distinctly (but not strikingly) yellowish; elytra irregularly mottled with dark and yellowish, the dark color almost solid on anterior half of dise and near apices, the yellow color predominating in a submarginal zone in about anterior 24 of each elytron and in a small, irregular, common sutural blotch about 14 from apex; legs inconspicuously bicolored, femora and tibiae darker at middle than at ends; antennae yellowish with bases of several segments (especially 2nd to 4th) infuscate; microsculpture normal, but only moderately transverse on elytra. Head .64 width prothorax; eyes of nearly normal length but much less prominent than in pox, with posterior supraocular setae about between their posterior edges; neck not impressed above; front more convex than usual, with slight anterior impressions. Prothorax wider than usual in genus, very strongly narrowed in front, scarcely so behind; width/length 1.46; base/apex about 1.82; sides moderately arcuate for most of length, almost straight just before base; basal angles slightly obtuse, slightly blunted; lateral margins very narrow; baso-lateral areas smoothly convex near angles, but basal area impressed each side almost as near to middle as to sides; DARLINGTON: CARABID BEETLES OF NEW GUINEA 221 dise normal; anterior and posterior marginal lines entire but not deeply impressed. Elytra with basal margin about rectangular at humeri; lateral margins very narrow; subapical sinuations slight; apices strongly lobed about opposite 4th intervals, then oblique forward to obtuse, indistinctly denticulate sutural angles; striae better impressed than usual in genus, not distinctly punctate; intervals slightly convex, Sth narrow and much more convex toward apex, 3rd 2-punctate. Secondary sexual characters of o& unknown; @2 with last ventral ab- dominal segment broadly subtruncate. Measurements: length 6.0; width 2.3 mm. Type. Holotype @ (British Mus.) from Orrori, Papua, 3,500 ft., July 1933 (Cheesman); unique. Measured specumen. The type. Notes. ‘This may be related to the preceding (tafapox) but differs in many characters, of which only a few of the more obvious are noted in the key to species of Maculagonum. POTAMAGONUM new genus Diagnosis. Based on one species, so generic and specific characters not separable, but genus characterized by large size (14-15.5 mm.); unusually long and slender appendages; wide, translucent prothoracic margin; conspicuously interrupted elytral striae; conspicuous accessory setae of 5th hind-tarsal segment; and wing-and-seta formula +w, amir ee tee te Description. See that of genotype, below. . Genotype. Potamagonum diaphanum n. sp. (below). Generic distribution. As yet known only from the Bismarck Range, N-E. New Guinea. Notes. For comparison with other genera, see key to genera, above. POTAMAGONUM DIAPHANUM Nn. Sp. Description. Form as figured (Fig. 8). Rather large, slender, with unusually long, slender appendages; castaneous, appendages not much paler, but tibiae paler than femora in some individuals; lateral margins of prothorax widely and conspicuously translucent; rows of pale or translucent spots along elytral striae in some individuals; shining; microsculpture of head and pronotum indistinct, of elytra light and moderately transverse. Head only moderately elongate (in tribe), .70 & .71 width prothorax; eyes rather large and prominent; both pairs supraocular setae present, posterior pair about between posterior edges of eyes; front normally convex, not wrinkled between eyes, impunctate, 222 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY with slight frontal impressions; neck-constriction rather deep but not sharply defined; mentum tooth triangular, somewhat blunted or rounded at apex. Prothorax rounded except apex emarginate and base slightly convexly subtruncate; width/length 1.27 & 1.23; base/apex not determined because of broad rounding of both anterior and pos- terior angles; basal foveae small, deep, not distinguishable from ends of lateral margins, impunctate; disc of pronotum normally convex, with light median line and rather deep transverse impressions, im- punctate; anterior and posterior marginal lines entire, deeplyimpressed. Elytra long, subparallel or slightly narrowed anteriorly, convex; basal margin entire, obtusely angulate at humeri; lateral margins very narrow; subapical sinuations moderate (variable), broad, leading onto acute angulations or short spines about opposite 4th intervals; apices then oblique forward and more or less emarginate to denticulate sutural angles; striation entire but striae conspicuously interrupted, reduced to series of short impressed lines; intervals approximately flat but very irregular, Sth and 9th not much modified toward apex, 3rd normally but inconspicuously 3-punctate. Inner wings fully developed. Lower surface impunctate; abdomen not pubescent; prosternal process simple; metepisterna long. Legs normally formed but slender; hind tibiae not sulcate along outer edges; hind tarsi slender, sulcate each side above; 4th hind-tarsal segment deeply emarginate and with short lobes, outer scarcely longer than inner; 5th hind-tarsal segment with a row of con- spicuous accessory setae on each side below; claws simple. Secondary sexual characters normal. Male copulatory organs as figured (Fig. 53). Measurements: length about 14-15.5; width about 3.5 mm. or slightly more (elytra too warped and spread for accurate measurements). Types. Holotype & (M.C.Z. No. 28,664) and 8 paratypes all from Chimbu Valley (some specimens labeled Mt. Wilhelm), Bismarck Range, N-E. N. G., 5,000-7,500 & 7,000-10,000 ft., Oct. 1944 (Darling- ton); all taken among wet stones and in other cover on the spray- drenched banks of the turbulent Chim River at various altitudes within the limits given. Measured specimens. The o& holotype and 1 Q paratype. Notes. This species has perhaps been derived from a Colpodes-like ancestor, but it seems distinct enough to stand as a separate genus endemic to the high mountains of New Guinea. GASTRAGONUM new genus Diagnosis. In most ways similar to Notagonwm but more convex, with eyes always so reduced that posterior supraocular setae are well behind line of their posterior edges, and often with wings reduced too; DARLINGTON: CARABID BEETLES OF NEW GUINEA 223 small (5.8-8.3 mm.), Gastrellarius- or broad Stenolophus- or even Trechus-like forms; brown or piceous, elytra sometimes slightly iri- descent, appendages somewhat but usually not strikingly paler; wing- and-seta formula =w, ++, ++,-+4++. Description. Form and color as described above; microsculpture variable, rarely absent. Head about as in Notagonum, except for re- duction of eyes; both pairs supraocular setae always present; antennae a little shorter than usual in Notagonum but normally formed; frontal impressions variable; mentum tooth somewhat variable, often shorter or shorter and wider than usual in Notagonum. Prothorax usually more convex, but otherwise as in Notagonum, with both pairs lateral pro- notal setae always present. Hlytra also more convex than usual but otherwise within common range of variation of Notagonwm; apices always simple, with sutural angles not denticulate; outer intervals never much modified toward apex; 3rd interval always normally d-punctate. Inner wings full, dimorphic, or vestigial. Lower surface as usual in Notagonum, not or only locally and slightly punctate; abdomen rarely pubescent, usually not; prosternal process simple; metepisterna variable, shortened in some species with reduced wings. Legs as in Notagonum; hind tarsi moderately slender, slightly or not distinctly suleate at sides above; 4th hind-tarsal segment different in different species; 5th hind-tarsal segment without obvious accessory setae; claws simple. Secondary sexual characters as in Notagonum (normal). Male copulatory organs simply agonine (Figs. 54-56). Genotype. Gastragonum terrestre n. sp. (below). Generic distribution. Mountains of New Guinea, probably mostly in unforested areas. Notes. This genus is presumably derived from Notagonum. Its obvious adaptations (reduction of eyes, atrophy of wings in some species, and associated structural changes) are those which occur most often among ground-living mesophile Carabidae in temperate areas. The only species of Gastragonum which I have myself collected in any numbers (terrestre) was common under stones, and under strawberry plants in a missionary garden, in open country, not in forest, and always away from water. I found no other agonine in this situation. The several species of Gastragonum, perhaps excepting laevisculptum, apparently form a natural mesophile group which may be more or less confined to open areas of the mountains of New Guinea. This would account for the small number of specimens secured by Cheesman and Toxopeus, and so for the unfortunate number of uniques described below. These collectors apparently worked chiefly in forest (which, as experienced collectors well know, is the richest environment in New Guinea) or along brooks, and took many of their Carabidae with light 224 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY traps, which would hardly catch at least the flightless species of Gastragonum. Key to the Species of Gastragonum 1. Whole upper surface polished, without reticulate microsculpture; elytral striae strongly, punctater(p:.1224 nia oe eer ee laevisculptum — Microsculpture present; elytral striae not punctate.................... 2 2. Prothorax more broadly rounded, with sides relatively widely margined and not sinuate before base; (relatively weak frontal impressions distinguish this from all following species except trechoides) (p. 225)... subrotundum — Prothorax less broadly rounded, often subcordate, with sides more narrowly margined and more or less sinuate before base.................-.--- 3 3. Sides of prothorax strongly sinuate before base; (frontal impressions rather deep; short-lmear but not punctiform)).,.(..55 4... 22800. e eee dee 4 — sides of prothorax;weakly, sinuatela. oa 6.-les > eee ee ae eee 5 4. Elytra not transversely impressed before middle; form a little broader; sides of prothorax more abruptly sinuate before base (p. 226)....... .terrestre -— Elytra transversely impressed before middle; form a little more slender; sides of prothorax more broadly sinuate before base (p. 227). .terrestroides 5. Elytra normal, subquadrate; frontal impressions very small, deep, puncti- forma’ (DY Q27Ts SPEAR een ees SC SO te Ria Oa tn ee frontepunctum — Elytra oval; frontal impressions shallow, poorly defined (p. 228). .trechoides GASTRAGONUM LAEVISCULPTUM n. sp. Description. With essential characters of genus as described above. Larger, broader, and less convex than usual (in genus); piceous-brown, appendages paler brown; upper surface polished, without detectable microsculpture (at 54) but finely sparsely punctulate. Head .70 width prothorax; eyes small but abruptly prominent; genae about long as eyes, very oblique, convex in profile; front irregular, with moderate anterior impressions. Prothorax rather large; width/length 1.41; base/apex 1.26; anterior angles rather prominent, moderately rounded; sides weakly arcuate, vaguely subangulate at anterior-lateral setae, broadly but not strongly sinuate before nearly right, only slightly blunted posterior angles; lateral margins wide (in genus), moderately reflexed toward base; basal foveae moderate, irregular but scarcely punctate; disc normal; anterior and posterior marginal lines more or less entire but not deeply impressed. Elytra moderately wide, sub- parallel at middle, relatively convex compared with prothorax, not impressed; basal margin rounded and more than usually elevated at humeri; lateral margins rather wide (in genus); striae moderately im- pressed, rather strongly punctate; intervals convex. Lower surface impunctate except abdomen with some fine sparse punctation and DARLINGTON: CARABID BEETLES OF NEW GUINEA 225 inconspicuous pubescence. Inner wings vestigial; metepisterna shortened, not much more than 1% longer than wide. Legs: 4th hind- tarsal segment rather shallowly emarginate. Male copulatory organs: Fig. 54. Measurements: length 8.3; width about 3.4 mm. Type. Holotype o (Leiden Mus.) from Letterbox Camp, Snow Mts., Neth. N. G., 3,600 m. (about 11,700 ft.), Sept. 1-12, 1938 (Toxopeus) ; unique. Measured specimen. 'The type. Notes. Although this interesting species has the technical characters of Gastragonum, I am not sure that it is really related to the other members of the genus. It may be independently derived, presumably from Notagonum. 'The characters which set this species apart from the others are general form, abruptly prominent ‘‘popped”’ eyes, absence of reticulate microsculpture, slight elevation of basal margin of elytra at humeri, punctation of elytral striae, pubescence of abdomen, and form of 4th hind-tarsal segment. I know of no species of any genus to which the present one is closely similar. GASTRAGONUM SUBROTUNDUM DN. sp. Description. With characters of genus as described above. Form of convex Stenolophus; piceous, more or less iridescent (varying in differ- ent lights), appendages brown, lateral margins of prothorax moderately translucent-brown; microsculpture light and isodiametric on head, very light, fine and transverse on pronotum and elytra. Head .62 & .64 width prothorax; eyes small, moderately prominent; genae nearly as long as eyes, oblique, straight or slightly convex in profile; front strongly convex, frontal impressions slight; mentum tooth a little smaller than usual, triangular. Prothorax rounded-transverse; width/ length 1.38 & 1.29; base/apex 1.30 & 1.33 (base measured at posterior- lateral setae) ; anterior angles only normally prominent, rather narrowly rounded; sides broadly arcuate through all or much of length, not sinuate posteriorly; posterior angles completely and broadly rounded out (o' from Bismarck Range) or merely very obtuse (2 from Snow Mts.); lateral margins moderate, wider basally, moderately reflexed; basal foveae rather small, scarcely punctate; disc normal; anterior and posterior marginal lines fine but entire. Elytra subquadrate, moder- ately elongate, rather strongly convex, slightly or scarcely impressed before middle; basal margin not or vaguely angulate at humeri; lateral margins rather narrow; striae moderately impressed, impunctate. Lower surface virtually impunctate; abdomen not pubescent. Inner wings fully developed in both specimens; metepisterna long. Legs: 4th hind-tarsal segment with rather short outer and still shorter inner 226 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY lobe. Male copulatory organs: Fig. 55. Measurements: of o& from Bismarck Range, 6.6 by about 2.5; of 2 from Snow Mts., 8.0 by about 3.0 mm. Types. Holotype &@ (M.C.Z. No. 28,665) from Chimbu Valley, Bismarck Range, N-E. N. G., 5,000-7,500 ft., Oct. 1944 (Darlington), from an unrecorded habitat but not in forest; and 1 9 paratype from Ibele (Iebele) Camp, Snow Mts., Neth. N. G., 2,250 m. (about 7,325 ft.), Nov. 1938 (Toxopeus). Measured specimens. ‘The types. Notes. This species is sufficiently distinguished from others in the key to species of Gastragonum, above. The Snow Mts. specimen is larger than the type, with relatively narrower prothorax with evident, though very obtuse, posterior angles. If these differences prove to hold in series, they should be of at least subspecific value. GASTRAGONUM TERRESTRE N. Sp. Description. With characters of genus as described above. Form (Fig. 9) of Gastrellarius (subgenus of Pterostichus); brownish-piceous, not distinctly iridescent, appendages yellowish-brown, lateral margins of prothorax moderately translucent-brown; microsculpture light and isodiametric on front (more distinct on occiput), very light and moder- ately transverse on dise of pronotum, more distinct (but still light) and moderately transverse on elytra. Head .65 & .64 width prothorax; eyes small but rather prominent; genae shorter than eyes, oblique; front strongly convex, with deep, short, linear frontal impressions ex- tending onto clypeus; mentum tooth shorter than usual in Notagonum. Prothorax quadrate-subcordate; width/length 1.27 & 1.26; base/apex also 1.27 & 1.26; sides arcuate for much of length, strongly sinuate near base; basal angles approximately right, not blunted; lateral margins average (in genus); basal foveae moderate, poorly defined, more or less extensively punctate especially toward disc; disc normal; anterior and posterior marginal lines entire. Elytra subquadrate, of normal outline, rather strongly convex, not impressed on dise; basal margin not or slightly subangulate at humeri; lateral margins rather narrow; striae rather deeply impressed, irregular but not distinctly punctate. Lower surface sometimes with a few punctures at sides of mesosternum; abdomen not pubescent. Inner wings dimorphic, ves- tigial in 17 specimens (including type), fully developed in 2 specimens with same data; metepisterna rather long but a little variable, longest in the fully winged specimens. Legs: 4th hind-tarsal segment lobed but outer lobe short and inner one even a little shorter. Male copu- DARLINGTON: CARABID BEETLES OF NEW GUINEA DPA latory organs as figured (Fig. 56). Measurements: length 6.2-6.9; width about 2.4—2.6 mm. Types. Holotype o& (M.C.Z. No. 28,666) and 18 paratypes all from Chimbu Valley, Bismarck Range, N-E. N. G., 5,000-7,500 ft., Oct. 1944 (Darlington). Measured specimens. ‘The o holotype and 1 9 paratype. Notes. This species is adequately compared with others in the key to species of Gastragonum. Its habitat is described in notes under the genus. GASTRAGONUM TERRESTROIDES Nn. sp. Description. Very close to terrestre and answering to the same de- scription (see above) except in details given below. Larger and a little more elongate than terrestre; color same, but elytra faintly iridescent in some lights, with elytral microsculpture finer. Head .70 & .64 width prothorax. Prothoraz: width/length 1.20 & 1.22 (narrower than in terrestre); base/apex 1.28 & 1.32; sides a little more broadly but still strongly sinuate before base. Elytra a little more elongate and with basal margin more distinctly angulate at humeri than in terrestre, and with disc distinctly impressed before middle in all (4) specimens. Inner wings fully developed in all specimens. Other characters as in terrestre (except male copulatory organs not compared). Measwrements: length about 7.2-7.6; width slightly under 3.0 mm. Types. Holotype o (Leiden Mus.) and 1 broken o paratype (M.C.Z. No. 28,667) from Ibele (Iebele) Camp, Snow Mts., Neth. N. G., 2,250 m. (about 7,325 ft.), Nov.—Dec. 1938 (Toxopeus); 1 @ paratype from Moss Forest Camp, Snow Mts., 2,800 m. (about 9,100 ft.), Oct. 9-Nov. 5, 1938 (Toxopeus); and 1 @ paratype from Mt. Misim, Morobe Dist., N-E. N. G. (Stevens, M.C.Z.). Measured specimens. The & holotype and @ paratype from the Snow Mts. Notes. This is so close to terrestre that, if it occurred only on the Snow Mts., I should probably consider it a subspecies, but its occur- rence also on Mt. Misim has led me to treat terrestroides as a full species at least for the time being. GASTRAGONUM FRONTEPUNCTUM 0. sp. Description. With characters of genus as described above. Similar to terrestre (of which see description, above) but differing in the follow- ing details. Similar to terrestre in form (except as noted below), color, and microsculpture. Head .65 width prothorax; eyes larger but less 228 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY prominent than in terrestre, with genae short and oblique; frontal im- pressions very small but deep, round, punctiform, just behind clypeal suture. Prothorax: width/length 1.28; base/apex 1.20 (base relatively narrower than in ferrestre); sides much less sinuate than in ferrestre; basal angles obtuse, somewhat blunted. FElytra, lower surface, ete. about as in terrestre. Inner wings vestigial in single specimen, but metepisterna not much shortened. Measurements: length 6.7; width about 2.4 mm. Type. Holotype @ (British Mus.) from Mt. Tafa, Papua, 8,500 ft., Feb. 1934 (Cheesman); unique. It is unfortunately somewhat crushed and broken, but the essential characters are not affected. Measured specimen. The type. Notes. The form of the frontal impressions is of course the out- standing character of this species. GASTRAGONUM TRECHOIDES 0. sp. Description. With characters of genus as described above. Very small; form of a convex Trechus; color and microsculpture as in terrestre (see above). Head .65 width prothorax; eyes relatively larger than in fterrestre but much less prominent; genae very short, oblique; frontal sulci slight. Prothorax rather small (in genus); width/length 1.24; base/apex 1.23; sides rather gently arcuate for most of length, slightly sinuate before base; basal angles obtuse, moderately rounded; lateral margins rather narrow, a little wider toward base, moderately reflexed; basal foveae moderate, not distinctly punctate; disc normal; anterior and posterior marginal lines entire. Elytra suboval, a little more narrowed in front than behind; basal margin obtusely but dis- tinctly angulate at humeri; striae moderately impressed, not punctate; other details normal for genus. Lower surface virtually impunctate; abdomen not pubescent. Inner wings vestigial; metepisterna a little shortened. Legs: 4th hind-tarsal segment as in terrestre. Measure- ments: length 5.8; width 2.3 mm. Type. Holotype 2 (Leiden Mus.) from Baliem C amp, Snow Mts., Neth. N. G., 1,700 m. (about 5,525 ft.), Nov. 16-27, 1938 (Toxopeus) ; unique. Measured specimen. The type. Notes. A distinct little species, sufficiently defined within the genus in the key to species of Gastragonum. DARLINGTON: CARABID BEETLES OF NEW GUINEA 229 IDIAGONUM new genus Diagnosis. Rather large (12.9-15.3 mm.), usually dull, black or brown, mountain-living forms, with wings vestigial, elytra with basal margin incomplete and with a partial extra (10th) interval, prosternal process setose, and wing-and-seta formula —w, ++, ++, ——~—. Description. Idiastes-like (Fig. 10); size and color as given above; microsculpture variable. Head large, more or less oval, with moderate or deep neck constriction; eyes very small but abruptly prominent, well separated from mouth below; both pairs supraocular setae present, anterior ones a little above anterior edges of eyes, posterior ones far behind and above eyes; antennae normal, with 3rd segment longer than 4th and 2 or more times long as 2nd; front very broadly, a little irregularly convex, slightly impressed each side anteriorly; mentum with a strong, triangular tooth, sometimes narrowly truncate at apex; ligula broad, bisetose; paraglossae slender, a little longer than ligula, free and bent inward toward apex; palpi slender, labial ones with 2nd segment bisetose. Prothorax more or less subcordate; anterior angles moderate or very prominent anteriorly; both pairs of lateral pronotal setae present, anterior ones at or a little before middle, posterior ones at basal angles. Elytra with basal margin absent inwardly, ending near bases of 4th striae, more or less angulate and prominent anteriorly at humeri; apices simple or nearly so; usual 9 intervals present and also a partial 10th one between 9th and margin posteriorly, outer intervals not much modified toward apex, 3rd without dorsal punctures; a slight longitudinal fold inside each elytron near outer edge, not reaching margin. Inner wings vestigial; metepisterna (without epimera a little longer than wide. Lower surface at most vaguely punctate; abdomen not pubescent; prosternal process with tip not margined but with conspicuous setae. Legs normally formed; hind tibiae not sulcate along outer edges; hind tarsi rather slender, lightly sulcate each side above; 4th hind-tarsal segment rather deeply emarginate but not or very briefly lobed; 5th hind-tarsal segment without obvious accessory setae; claws simple; sole of first 4 hind-tarsal segments rather densely setose each side below but with middle of sole narrowly bare. Secondary sexual characters normal except o front tarsi less dilated than usual. Male copulatory organs as figured (Fig. 57). Genotype. Idiagonum asperum n. sp. (below). Generic distribution. At present known only from the Bismarck and Snow Mt. Ranges of New Guinea. Notes. This new genus is superficially rather similar to Idiastes Andrewes of Mt. Kinabalu, Borneo, but it differs from Jdiastes in the following notable characters: all normal supraocular and lateral pro- notal setae present, basal margins of elytra incomplete, partial 10th 230 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY elytral interval present, and prosternal process with setae. No one o these characters is necessarily very important, but together they sug gest that Idiagonum is not directly related to Idiastes but has been independently derived perhaps from a Colpodes-like ancestor in the mountains of New Guinea. One of the alticoline Colpodes of New Guinea (acuticauda, described above) has some of the expected charac- teristics of such an ancestor. I should add that I have examined at the British Museum a dissected cotype of Idiastes alaticollis Andr. and that as a result I feel sure that Idiastes should go in the Agonini near Colpodes, not in the Pterostichini where Andrewes put it. The inner costa of the elytron does not reach the margin and is not really like that of a true pterostichine, and the male copulatory organs are agonine. The partial extra (10th) elytral interval occurs in all specimens of all species of Idiagonum. A smaller or less well defined rudimentary 10th interval occurs also in some Fortagonum (below), but this is probably an example of parallelism rather than an indication of relationship. Key to the Species of Idiagonum 1. Pronotum not transversely wrinkled (p. 280)...............-. inasperum — Pronotal disc densely and extensively transversely wrinkled............ 2 2. Base of pronotum subequal to or narrower than apex, not distinctly mar- pineds(p. DOW) ey Pee Me ae eA ac se asperum — Base of pronotum slightly wider than apex, margined................. 3 3. Eyes, though small and abruptly prominent, forming only very obtuse angles with genae in profile from above; pronotum with many shallow transverse wrinkles, very finely longitudinally rugulose near both anterior and posterior margins; posterior marginal bead below level of surface of base ofspronotumay (51232) ieee a cea ceecaeene lee oe ee muscorum — Eyes more abruptly prominent, forming conspicuous though still] somewhat obtuse angles with genae; pronotum with fewer but deeper transverse wrinkles, more coarsely longitudinally rugulose near anterior margin but not distinctly rugulose near posterior margin; posterior marginal bead not below level of surface of base of pronotum (p. 232)........ asperior IDIAGONUM INASPERUM n. sp. Description. With characters of genus as described above, but more shining than usual, without transverse wrinkling of pronotum. Piceous-black or brownish-piceous (immature); moderately shining, not iridescent; upper surface finely and sparsely punctate and with reticulate microsculpture light and isodiametric on front of head, fine and transverse on pronotum, coarser and variable on elytra (9) (nearly isodiametric on disc anteriorly, more transverse laterally and DARLINGTON: CARABID BEETLES OF NEW GUINEA Dill apically). Head .76 & .74 width prothorax; some transverse wrinkling behind and below eyes; mentum tooth truncate at apex. Prothorax rather large (in genus); width/length 1.28 & 1.25; base/apex 1.10 & 1.12; anterior angles moderately prominent; sides arcuate through much of length, not angulate at lateral setae, broadly sinuate before obtuse but nearly right and well formed basal angles; lateral margins rnoderately wide, reflexed; basal foveae poorly defined, shallow, im- punctate, separated from margins by slight swellings (or they could be described as broad and deep, reaching margins, with bottoms slightly swollen); dise moderately convex; anterior and posterior marginal lines entire. Elytra slightly narrowed toward base, moder- ately convex; subapical sinuations absent or nearly so; apices simple, almost conjointly rounded, but sutural angles independently narrowly rounded; striae rather deep, impunctate; intervals moderately convex. Measurements: length 12.9-13.4; width 4.6-4.8 mm. Types. Holotype 2 (Leiden Mus.) and1 @2 paratype (M.C.Z. No. 28,668) both from Moss Forest Camp, Snow Mts., Neth. N. G., 2,600-2,800 m. (about 8,450-9,100 ft.), Oct. 9-Nov. 5, 1938 (Toxopeus) Measured specumens. The types. Notes. This species is sufficiently distinguished from others in the key, above. The paratype has a few longitudinal wrinkles on head and pronotum which are, I think, due to warping of the surface. The specimen is slightly immature. IDIAGONUM ASPERUM N. sp. Description. With characters of genus as described above. Form as figured (Fig. 10). Rather dull black, legs and antennal bases piceous, outer segments of antennae brown; microsculpture of head fine but deeply impressed, isodiametric; that of pronotum fine, trans- verse; that of elytra apparently still finer and transverse but scarcely visible at 54. Head .75 & .70 width prothorax; front with fine sparse punctulation as well as microreticulation and a little irregular wrinkling; much transverse wrinkling behind and below eyes; mentum tooth bluntly pointed. Prothorax subcordate; width/length 1.16 & 1.21; base/apex .98 & .98 (& .92); anterior angles moderately prominent; sides irregularly arcuate or nearly straight (and converging) anteriorly, angulate at anterior-lateral setae, then strongly converging posteriorly and strongly sinuate well before basal angles; latter right, accurately defined; lateral margins rather wide, reflexed; basal foveae deep, not punctate; disc strongly transversely rugulose, impunctate; apical and basal marginal areas longitudinally rugulose; anterior marginal line entire, posterior one vague or absent. Elytra distinctly narrowed 232 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY toward base and rather strongly convex; subapical sinuations nearly absent; apices conjointly rounded except sutural angles slightly di- vergent and slightly produced (slightly variable); striae deep, not or slightly punctulate; intervals convex, with at most extremely fine, sparse, inconspicuous punctulation. Male copulatory organs as figured (Fig. 57). Measurements: length 13.8-15.3; width 4.7-5.2 mm. Types. Holotype @ (M.C.Z. No. 28,669) and 12 paratypes all from Mt. Wilhelm, Bismarck Range, N-E. N. G., 7,000-10,000 ft., Oct. 1944 (Darlington), taken under various cover on the ground in heavy forest. Measured specimens. The &@ holotype and 1 Q paratype (and in parentheses prothoracic base/apex ratio of another 9). Notes. Sufficiently compared with other species in the key above, and in the descriptions of the two following species. IDIAGONUM MUSCORUM DN. sp. Description. With characters of genus as described above. Color and sculpture of asperwm (of which see description, above) and form nearly similar, -but broader, with elytra less narrowed anteriorly. Head .70 & .70 width prothorax, like that of asperwm in all characters given. Prothorax: width/length 1.30 & 1.31; base/apex 1.08 & 1.10; anterior angles more prominent than in asperum; sides more evenly arcuate anteriorly, less angulate at anterior-lateral setae, and less strongly sinuate posteriorly, with posterior angles on the obtuse side of right; basal foveae and disc about as in asperum; posterior as well as anterior marginal line distinct, and posterior margin slightly below level of surface of base of pronotum. FElytra relatively broader (es- pecially anteriorly) than in asperum; otherwise about same except sutural angles narrowly rounded or simply angulate, usually less prominent than in asperum. Measurements: length 13.4-15.2; width 4.9-5.4 mm. Types. Holotype eon Mus.) and 5 paratypes (2 in M.C.Z. No. 28,670) all from Moss Forest Camp, Snow Mts., Neth. N. G., 2,600-2,800 m. (about 8,450-9,100 ft.), Oct. 9-Nov. 5, 1938 (ioconen) Measured specimens. The @ holotype and 1 9 paratype. Notes. This is sufficiently compared with asperum above, and with other species in the key. IDIAGONUM ASPERIOR N. sp. Description. With characters of genus as described above. Color of the two preceding species (asperum and muscorum); form intermediate DARLINGTON: CARABID BEETLES OF NEW GUINEA 233 between the two. Head .72 & .74 width prothorax; eyes much more abruptly prominent than in the other species; sides of front irregularly flattened or depressed; neck-constriction deeper; mentum tooth bluntly pointed; sculpture of head about as in asperum. Prothorax: width/ length 1.38 & 1.32; base/apex 1.09 & 1.07; anterior angles very promi- nent, making prothorax appear longer than measurements above suggest; sides angulate at anterior setae, rather broadly, moderately sinuate before basal angles; latter slightly obtuse or nearly right, very well defined; basal foveae deep and simple, as in asperum; dise with fewer but deeper transverse wrinkles than in asperwm or muscorum and with some fine sparse punctulation, coarsely longitudinally rugulose near apex, irregular but not distinctly rugulose near base; posterior as well as anterior marginal line distinct, posterior margin on same level as surface of base of pronotum. Elytra with apices simple, broadly conjointly rounded (as in other species) and with sutural angles narrowly rounded; striae and intervals about as in preceding species but intervals a little more plainly though still finely and sparsely punctulate. Measurements: length 13.1-13.8; width 4.9-5.3 mm. Types. Holotype o (Leiden Mus.) and 4 paratypes (2 in M.C.Z. No. 28,671) from Moss Forest Camp, Snow Mts., Neth. N. G., 2,600- 2,800 m. (about 8,450-9,100 ft.), Oct. 9-Nov. 5, 1938; and 1 paratype from Top Camp, Snow Mts., 2,100 m. (about 6,825 ft.), Jan. 29, 1939 (all collected by Toxopeus). Measured specimens. The & holotype and 1 @ paratype from Moss Forest Camp. Notes. It is surprising to find a second rugulose species of this genus with exactly the same data as muscorum. Possibly the two occur in different forest tracts, or are otherwise separated. It is conceivable that the two are forms of one dimorphic species, but this seems to me unlikely. Each is represented by several specimens and each is uni- formly characterized by several seemingly independent structural details. MONTAGONUM new genus Diagnosis. Based on one species, so generic and specific characters not separable, but genus characterized by convex Calathus-like form, atrophied wings, wing-and-seta formula —w, ++, —+, +++ (but single punctures of 3rd elytral interval sometimes missing) with pos- terior-lateral pronotal setae on flat surface of margins well before basal angles, absence of obvious accessory setae on 5th hind-tarsal segment, etc. Description. See that of single species, below. Genotype. Montagonum toxopeanum n. sp., below. 234 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Generic distribution. At present known only from one locality at a high altitude (at or slightly above timber line) on the Snow Mountains of Netherlands New Guinea. Notes. In atrophied wings and to some extent in appearance this new genus resembles certain Nebriagonum (below), but this is probably a result of similar adaptation to a similar habitat rather than an indi- cation of relationship. The pattern of standard setae is different from any Nebriagonum. The position of the posterior-lateral pronotal setae, on the flat surface of the margins, is somewhat like the position of these setae in certain Altagonum (latilimbus, paralimbus, scapha), although in the latter the setae are not so far forward. I suspect that Monta- gonum has been derived independently from an Altagonum by atrophy of wings and other appropriate changes. MoNTAGONUM TOXOPEANUM 0. sp. Description. Convex Calathus-like (Fig. 11); piceous-black, ap- pendages dark-reddish, lateral margins of prothorax and elytra faintly if at all translucent; upper surface impunctate, moderately shining, not iridescent; microsculpture normal, very distinct. Head rather small, .57 & .57 width prothorax; eyes reduced in size but more promi- nent than usual; genae about as long as eyes, oblique; both pairs of supraocular setae present, posterior ones well behind line of posterior edges of eyes; antennae rather short, normal in structure; neck not impressed above; front irregularly convex, with slight frontal im- pressions; mentum tooth triangular. Prothorax rather long, widest near middle, rather strongly narrowed in front, moderately so behind; width/length 1.13 & 1.11; base/apex 1.41 & 1.50; anterior angles moderately advanced, rounded-acute; sides weakly arcuate for much of length, more or less straight (and converging) or even faintly sinuate toward base; basal angles a little obtuse and rather narrowly rounded; lateral margins very narrow anteriorly, broader but very poorly defined posteriorly; anterior-lateral setae absent, posterior-lateral ones on flat margins just inside of thickened marginal beads about 1% of prothoracic length before apparent basal angles; basal foveae very shallow and poorly defined,sometimes with aslight swelling at middle, not punctate; dise convex, with light median line, and with transverse impressions scarcely indicated; anterior marginal line entire, posterior one faint or broadly interrupted at middle. Elytra with sides more rounded and disc more convex than usual in tribe; disc not impressed; anterior margin entire, about rectangular at humeri; lateral margins moderate; subapical sinuations absent, sides curving smoothly almost to suture, so apices almost conjointly rounded; sutural angles narrowly rounded, DARLINGTON: CARABID BEETLES OF NEW GUINEA oo slightly dehiscent but not produced, not denticulate; striae moderately impressed, not punctate; intervals slightly convex, 8th and 9th not much modified toward apex, 3rd normally 3-punctate except punctures somewhat variable in anterior-posterior spacing, and single punctures sometimes missing (left anterior puncture missing in type, right pos- terior one missing in 2nd measured specimen). Inner wings vestigial; metepisterna slightly shortened. Lower surface impunctate; abdomen not pubescent; prosternal process simple. Legs: 4th hind-tarsal seg- ment simply emarginate; 5th hind-tarsal segment with apparent ves- tigial but usually not obvious accessory setae (type has 1 obvious seta on outer-lower edge of right 5th hind-tarsal segment about middle of its length). Secondary sexual characters normal. Male copulatory organs as figured (Fig. 58). Measurements: length 8.7-9.3; width 3.2-3.5 mm. Types. Holotype o (Leiden Mus.) and 5 paratypes (2 in M.C.Z. No. 28,672) all from Letterbox Camp, Snow Mts., Neth. N. G., 3,600 m. (about 11,700 ft.), Sept. 1-12, 1938 (Toxopeus). Measured specimens. The o& holotype and 1 @ paratype. Notes. The possible relationships of this species are discussed under the genus. I take great pleasure in naming this distinct and interesting high- altitude form for Mr. L. J. Toxopeus, who obtained the types and so many other fine Carabidae on the Snow Mts. of New Guinea. I have used the name toxopeanum rather than toxopei because the latter has been employed by Andrewes for a species of Colpodes from Buru. NEBRIAGONUM new genus Diagnosis. Within the New Guinean agonine complex this genus is most simply characterized by atrophied wings plus ezther a very large head (.90 or more width prothorax) or presence of obvious accessory setae on the 5th hind-tarsal segment. Additional noteworthy generic characters are eyes small; head with distinct but variable neck- constriction; prothorax small or moderate in size and usually rather elongate; elytra more or less oval, with dorsal punctures irregular or absent. The wing-and-seta formula is confusingly variable: —w. (+)+, (—) (—), (—) (—) (—). Some of the species have rather the appearance of large-headed, convex species of Nebria with oval elytra. The size-range is from 7.7 to 14.3 mm. Description. Form variable; black or brownish with appendages not much paler; upper surface only moderately shining, not or (transvtzor) faintly metallic, not iridescent; upper surface including pronotal foveae virtually impunctate; microsculpture always distinct, about isodia- metric on head, more or less transverse (sometimes only slightly so) 236 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY on pronotum and elytra. Head moderate to very large; eyes rather small, variable in form; both pairs supraocular setae present except anterior pair usually absent in percephalum; posterior pair slightly or much behind line of posterior edges of eyes; neck-constriction always present, but variable, sometimes only slightly impressed above; an- terior frontal impressions moderate, irregular; front often impressed also between eyes, the impressions often being 4 in number and forming a rectangle narrowest anteriorly; antennae normal, varying in length; mentum tooth triangular with apex variable (pointed, blunted, or slightly emarginate). Prothoraz rather narrow but variable in form; lateral pronotal setae present only in cephalum, absent in other species; basal foveae moderate or small, not sharply defined; disc more or less strongly convex, with usual middle line and transverse impressions more or less lightly impressed; anterior marginal line entire, posterior one also usually entire but less impressed. Elytra more or less oval, usually strongly convex; basal margin entire, differently subangulate or angulate at humeri in different species; subapical sinuations absent or nearly so; apices usually simply and more or less independently rounded, rarely (arboreum) each with a very short spine; striae deeply to very lightly impressed in different species, not or (arboreum) faintly punctulate; Sth and 9th intervals not much modified toward apex; 3rd interval with or without (irregular) dorsal punctures. Lower surface impunctate or nearly so; abdomen not pubescent; prosternal process simple. Inner wings vestigial; metepisterna more or less shortened. Legs normally formed; hind tibiae not sulcate along outer edges; hind tarsi slender, not or only slightly suleate at sides above; 4th hind-tarsal segment simply emarginate in subcephalum, lobed in other species, with outer lobe longer than inner; 5th hind-tarsal seg- ment with or without obvious accessory setae in different species. Secondary sexual characters normal. Male copulatory organs as figured (Figs. 59 & 60). Genotype. Nebriagonum cephalum n. sp. (below). Generic distribution. At present known only from 5 species from the Bismarck Range and 1 from the Snow Mountains, New Guinea. Notes. Although there is extraordinary variation in some characters in this new genus, I am convinced that it is a natural one, except perhaps for N. subcephalum which is somewhat isolated in structure (see notes under its description) as well as geographically. The 5 species from the Bismarck Range form a remarkable and nearly con- tinuous series. N. cephalum has all normal supraocular and lateral pronotal setae except that one or both posterior-lateral pronotal setae are missing in a few individuals; it has some (irregular) dorsal punctures on the 38rd elytral intervals; and the 5th hind-tarsal segment lacks DARLINGTON: CARABID BEETLES OF NEW GUINEA Dai obvious accessory setae. N. percephalum is superficially very similar, differing only slightly in details of form, depth of elytral striae, etc., but has lost the anterior supraocular setae (except that the anterior one is present on one side in one individual), both pairs of pronotal setae, and all dorsal punctures of the 3rd elytral intervals; (the 5th hind-tarsal segment is without obvious accessory setae, as in cephalum). N. transitum too is not very different from cephalum superficially (the prothorax is more rounded but still with narrow margins, the elytra are less deeply striate, etc.), but it has lost both pairs of lateral pro- notal setae and most of the dorsal punctures of the 3rd elytral intervals (it has both pairs of supraoculars), and it has a few (usually 2 or 3 on each side) small but distinct accessory setae on the 5th hind-tarsal segment. N. transitior in turn is not very different from transitum, with elytral apices still simple, but it has wider prothoracic margins and better developed accessory setae on the 5th hind-tarsal segment, and it has completely lost all dorsal punctures of the 3rd elytral intervals. Finally, arborewm is not very different from transitior, but is larger, with short-spined elytral apices, and different in other details, including retention of at least the posterior puncture of the 3rd elytral interval. These 5 species form an interesting sequence in habits, too. N. cephalum and percephalum are hydrophiles which occur primarily by rapid mountain brooks in and above the highest forest on Mt. Wilhelm, and cephalwm occurs also in seepage areas up to about 14,000 ft., the highest altitude at which I found any Carabidae on the mountain. JN. transitum is a mesophile which is common on the ground under cover away from water, chiefly in open, grassy country above timber line. JN. transitior is another ground-living mesophile, but my 6 specimens of it were all found in forest. And N. arborewm is ap- parently arboreal; both my specimens of it were found above the ground, one in the thatch of an old shelter. I do not mean to imply that these 5 species form a simple linear series. Their relationships are probably complex. But I do feel sure that they represent one stock which has radiated on the Bismarck Range, or at least in a limited area in the mountains of New Guinea. These species are now isolated from each other at least partly by ecological factors, but it would be unwise to assume that they have evolved in their present positions as a result of ecological isolation. They are very distinct species and they may have had complex histories (cf. “Role of geographical isolation” in the introduction, above). Key to the Species of Nebriagonum 1. Fourth hind-tarsal segment emarginate, not lobed; (see also notes under Species) m(snows Mitss) i (os 2aS) aa ae neon i earn c orn subcephalum 238 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Fourth hind-tarsal segment lobed, outer lobe longer than inner (Bismarck FRAT ele ale ME TT a TOTS Nn Oe rr 2 2. Head very large, about .90 or more width prothorax; 5th hind-tarsal segment without obvious accessory setae....................2.-0-- 3 — Head relatively narrower; 5th hind-tarsal segment with obvious accessory 3. Both pairs (or sometimes only anterior pair) of lateral pronotal setae, both pairs of supraocular setae, and some punctures of 8rd elytral intervals DEESEN tC 5,239) as. lays moet leet sd faek teit tay a ena ap oa a cae i oe cephalum — Both pairs of lateral prothoracic setae, anterior supraoculars (usually), and all punctures of 3rd elytral intervals absent (p. 240)....... percephalum 4, Prothorax narrowly margined; length 8.4-10.3 mm. (p. 241)... .transitum — Prothorax more widely margined; size larger). «22-5. 205 see eee 5 5. Elytral apices simple; length 11.3-11.9 mm. (p. 241)........... transitior — Elytral apices each with a short spine; length 12.9-14.3 mm. (p. 242)..... arboreum NEBRIAGONUM SUBCEPHALUM nN. sp. Description. With main characters of genus as described above, but exceptional in several details. Form more normally agonine than in other Nebriagonum, with head narrower, prothorax less modified, and elytra less oval. Head .73 & .76 width prothorax; eyes small, only moderately prominent; genae about as long as eyes, slightly convex in profile; both pairs supraocular setae present; neck constriction vague, shallow;front onlyslightly impressed behind usual anterior impressions. Prothorax of moderate size, only moderately elongate; width/length 1.09 & 1.08; base/apex 1.18 & 1.21; anterior angles only slightly promi- nent; sides broadly, not strongly arcuate, then slightly or moderately sinuate before well formed, nearly right, only slightly blunted posterior angles; lateral margins rather narrow, without setae. Elytra with basal margin distinctly but obtusely angulate at humeri; apices simple, nearly as in cephalum (below); striae moderately impressed; 3rd inter- val impunctate. Legs: 4th hind-tarsal segment simply emarginate, not lobed; 5th hind-tarsal segment with obvious accessory setae. Measurements: length 9.8-10.0; width 3.6-3.8 mm. Types. Holotype o (Leiden Mus.) and 3 paratypes (1 in M.C.Z. No. 28,673) all from Lake Habbema, Snow Mts., Neth. N. G., 3,300 m. (about 10,725 ft.), Oct. 2, 19388 (Toxopeus). Measured specumens. The o& holotype and 1 9 paratype. Notes. This species is exceptional in Nebriagonum in that the head is only moderately large, with vague neck-constriction; the posterior angles of the prothorax accurately formed; and the 4th hind-tarsal segment only emarginate, not lobed. However, the species is not too different from some more typical Nebriagonum in form; the arrange- DARLINGTON: CARABID BEETLES OF NEW GUINEA 239 ment of fixed setae (both pairs of supraoculars present, both pairs of lateral pronotals absent, and all punctures of 3rd elytral intervals absent) is the same as in some Nebriagonwm; it has obvious accessory setae on the 5th hind-tarsal segment as do some Nebriagonwm; and it probably occupies a habitat like that of Nebriagonwm, occurring on the open grassy slopes or in the highest forest fringes above Lake Habbema in the Snow Mountains much as Nebriagonwm transitum occurs on the slopes above Lakes Aunde and Piunde on Mt. Wilhelm. I think that subcephalum probably is genetically a Nebriagonwm but that it is less closely related to any of the species of the Bismarck Range than the latter are to each other. NEBRIAGONUM CEPHALUM n. sp. Description. With characters of genus as described above. Form as figured (Fig. 12). Head appearing wider than prothorax but actually a little narrower, .96 & .90 as wide in measured specimens; eyes small, only moderately prominent; both pairs of supraocular setae present; genae as long or longer than eyes, convex in profile; neck-constriction rather deep; front usually with irregular impressions behind normal anterior ones. Prothorax small, appearing longer than wide but by measurement slightly wider; width/length 1.09 & 1.07, with greatest width about 14 behind apex; base/apex 1.03 & 1.02 if base measured across posterior setae, but base narrower if measured at apparent posterior angles; anterior angles hardly at all produced; sides irregu- larly, rather weakly arcuate, usually slightly sinuate before and sub- angulate at posterior-lateral setae, and then extended backward and slightly or strongly inward to narrowly rounded apparent posterior angles; lateral margins narrow, each with usual 2 setae in most cases, but posterior setae missing (not broken off) on one or both sides in several individuals. Elytra with basal margin distinctly but very obtusely angulate at humeri; apices irregularly broadly rounded to near sutural angles; latter narrowly rounded; striae moderately im- pressed; each 3rd interval with usually 3 or 4 (sometimes only 2) dorsal punctures irregularly placed as to both length and width of interval, and often very differently arranged on opposite elytra of one individual. Legs: 4th hind-tarsal segment lobed; 5th hind-tarsal seg- ment without obvious accessory setae. Male copulatory organs as figured (Fig. 59). Measurements: length 7.7-9.4; width 2.8-3.3 mm. Types. Holotype &@ (M.C.Z. No. 28,674) and 62 paratypes all from Mt. Wilhelm, Bismarck Range, N-E. N. G., Oct. 1944 (Darlington). The type and 21 paratypes are from above the forest line (above 10,000 ft.) and 5 additional paratypes are from still higher, about 240 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY 14,000 ft., not far below the rocky summit of the mountain, which I suppose to be 15,400 ft. high. The remaining 36 paratypes are from the forest zone between 7,000 & 10,000 ft. Actually all the specimens were taken along the Chim River or its highest tributaries, beside the running water, in and above the forest zone, or in seepage areas at still higher altitudes. Measured specimens. The o& holotype and 1 2 paratype with same data. Notes. Both the structure and habits of this species have been com- pared with those of other members of the genus in the generic dis- cussion, above. . NEBRIAGONUM PERCEPHALUM Nn. Sp. Description. With characters of genus as described above. Larger and a little more slender than the preceding (cephalum). Head ap- pearing wider than prothorax, but actually only .95 & .93 as wide; eyes small, more abruptly prominent than in cephalum; posterior supraocular setae present, anterior ones absent except present on right side in one individual; genae longer than eyes, strongly convex in profile; neck-constriction strongly marked; front conspicuously im- pressed between eyes, the impressed area divided into about 4 poorly defined, more or less longitudinal parts. Prothorax elongate, appearing much longer than wide but actually about as long; width/length .99 & 1.04; base/apex 1.03 & 1.08; anterior angles a little prominent an- teriorly; sides very broadly and slightly arcuate, more or less strongly sinuate before nearly right but slightly blunted or very narrowly rounded posterior angles; lateral margins narrow, without setae. Elytra with basal margin strongly (obtusely to almost rectangularly) angulate at humeri; apices simple, as in cephalum; striae deeply im- pressed at least at sides of elytra; 3rd intervals impunctate. Legs: 4th hind-tarsal segment lobed; 5th hind-tarsal segment without obvious accessory setae. Measurements: length 9.3-10.5; width 3.1-3.6 mm. Types. Holotype o&* (M.C.Z. No. 28,675) and 11 paratypes all from Mt. Wilhelm, Bismarck Range, N-E. N. G., Oct. 1944 (Darlington). The type and 10 paratypes are from the forest zone between 7,000 & 10,000 ft.; 1 paratype, from above the forest (above 10,000 ft.). Actually all were taken on the banks of the Chim River or its tribu- taries, in company with cephalum. Measured specimens. The o& holotype and 1 9 paratype. Notes. This species is sufficiently discussed and compared under the genus and in the key to species of Nebriagonum. DARLINGTON: CARABID BEETLES OF NEW GUINEA 241 NEBRIAGONUM TRANSITUM n. sp. Description. With characters of genus as described above. Slightly stouter than any of the preceding species, very convex, almost like a stout Broscus in appearance. Head .82 & .82 width prothorax; eyes small, rather abruptly prominent; genae longer than eyes, convex in profile; both pairs supraocular setae present; neck-constriction well marked, moderately impressed above; front rather lightly impressed (4 poorly defined impressions) between eyes. Prothorax of moderate size; width/length 1.12 & 1.13; base about wide as apex (angles too rounded to measure base exactly); anterior angles slightly advanced; sides more arcuate and much more converging posteriorly than in cephalum, straight or slightly sinuate before obtuse, rather narrowly rounded basal angles; lateral margins narrow, without setae. Elytra with basal margin vaguely or distinctly (but very obtusely) angulate at humer1; apices simple, about as in cephalum; striae lightly impressed; dorsal punctures of 3rd elytral intervals variable, rarely entirely absent, anterior puncture usually present on one or both elytra and posterior one often present on one or both elytra too, but middle puncture rarely if ever present. Legs: 4th hind-tarsal segment lobed; 5th hind-tarsal segment usually with small but distinct accessory setae (usually about 2 each side of segment, but in some cases they are broken off or possibly absent). Measurements: length 8.4-10.3; width 3.1-3.7 mm. Types. Holotype o& (M.C.Z. No. 28,676) and 48 paratypes all from Mt. Wilhelm, Bismarck Range, N-E. N. G., Oct. 1944 (Darlington); the type and 36 paratypes are from open grassy slopes above 10,000 ft.; the remaining 12 paratypes, from forest between 7,000 & 10,000 ft. (presumably from the upper forest fringes). Measured specimens. The o holotype and 1 2 paratype with the same data. Notes. This species is discussed under the genus and defined in the key to species of Nebriagonum. NEBRIAGONUM TRANSITIOR n. Sp. Description. With characters of genus as described above. Larger and more slender than the preceding (transitum), with a faint purple tinge on elytra not present in other species of genus. Head .84 & .83 width prothorax; eyes small, rather abruptly prominent; genae longer than eyes, convex in profile; both pairs supraocular setae present; neck-constriction moderately impressed above; front with 4 slight impressions between eyes. Prothorax long; width/length .98 & 1.01; base about wide as apex; anterior angles scarcely at all prominent; 242 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY sides rather weakly arcuate for much of length, rather strongly con- verging and straight or slightly sinuate posteriorly; posterior angles obtuse, rather narrowly rounded; lateral margins moderate (wider than in preceding forms) but scarcely wider posteriorly than anteriorly and not much reflexed, without setae. Elytra with basal margin moderately though somewhat obtusely angulate at humeri; apices simple, about as in cephalum; striae rather lightly impressed; 3rd intervals im- punctate in all specimens. Legs: 4th hind-tarsal segment lobed; 5th hind-tarsal segment with several obvious accessory setae each side. Male copulatory organs: Fig. 60. Measurements: length 11.3-11.9; width 3.7-4.0 mm. Types. Holotype &@ (M.C.Z. No. 28,677) and 5 paratypes all from Mt. Wilhelm, Bismarck Range, N-E. N. G., 7,000-10,000 ft. (in forest), Oct. 1944 (Darlington). Measured specimens. The & holotype and 1 @ paratype. Notes. This species too has been discussed and compared under the genus and in the key to species of Nebriagonum. NEBRIAGONUM ARBOREUM N. sp. Description. With characters of genus as described above. Still larger than preceding (transitior), with elytra less convex, more elongate, and more narrowed toward humeri. Head .79 & .78 width prothorax; eyes rather small, not abruptly prominent, almost con- tinuing lines of genae; latter long as or slightly longer than eyes, slightly convex in profile; both pairs supraocular setae present; neck- constriction distinct but not much impressed above; front scarcely impressed between eyes behind usual anterior impressions. Prothorax long; width/length .99 & 1.06; base appearing slightly narrower than apex, but posterior angles too rounded for accurate measurement of base; anterior angles slightly advanced; sides broadly arcuate anteri- orly, strongly converging from about anterior 14 toward base, slightly sinuate before broadly rounded posterior angles; lateral margins rather wide (wider than in ¢transitior) but scarcely wider posteriorly than anteriorly, moderately reflexed, without setae; base broadly, slightly arcuate or lobed, the lobe smoothly rounded into posterior angles. Elytra rather long and slender, much narrowed anteriorly, less convex than usual in genus; basal margin strongly but not quite rectangularly angulate at humeri; apices each with a very short, stout spine about opposite 3rd or 4th interval, base of spine running sinuously but smoothly into both lateral and sutural margins of elytron, without other angulation; striae very lightly impressed, irregular or faintly punctulate; 3rd interval of co’ type with only posterior puncture DARLINGTON: CARABID BEETLES OF NEW GUINEA 243 present, near top of declivity on each elytron, of 2 paratype with only posterior puncture on left elytron but middle and posterior ones on right elytron. Legs: 4th hind-tarsal segment lobed; 5th hind-tarsal segment with a row of conspicuous accessory setae each side. Measure- ments: length (o’ Q ) 12.9-14.3; width 4.2-4.8 mm. Types. Holotype o& (M.C.Z. No. 28,678) and 1 9 paratype both from Mt. Wilhelm, Bismarck Range, N-E. N. G., 7,000-10,000 ft. (forest), Oct. 1944 (Darlington). Measured specimens. The types. Notes. See again the generic discussion and the key to species of Nebriagonum. The copulatory organs of the o’ type are too unformed to draw. LAEVAGONUM new genus Diagnosis. Small (5.5-8.4 mm.), Europhilus- or Calathus- or cistelid- like; prothorax and elytra with outlines smoothly but more or less inde- pendently rounded and with dises smoothly convex; basolateral foveae of pronotum obsolete; wing-and-seta formula —w,+-+,(—)—,-—-—. Description. Form as indicated above; brown or piceous, with ap- pendages brownish or yellowish; surface moderately shining, not iri- descent, nearly impunctate above except elytral intervals often vaguely punctulate; microsculpture normal except only slightly transverse on elytra, usually less so than on pronotum. Head small, more or less elongate; eyes more or less reduced in size and only slightly prominent; both pairs supraocular setae present, posterior ones slightly or dis- tinctly behind line of posterior edges of eyes; antennae normal; neck slightly or not impressed above; front convex, with small anterior impressions; mentum tooth triangular with apex more or less blunted. Prothorax with sides smoothly rounded for whole length, or sometimes straighter (and of course more or less converging) toward base; lateral margins very narrow, usually without setae but anterior-lateral ones present in subcitum; anterior angles more or less distinct, obtuse or narrowly rounded, not produced beyond curve of broadly emarginate anterior edge of prothorax; posterior angles moderately or broadly rounded; disc with usual median line (sometimes very light) but trans- verse impressions slight or absent; anterior marginal line usually entire but often faint and sometimes interrupted at middle, posterior one faint or widely interrupted or obsolete. Elytra with basal margin entire or nearly so, rectangular or nearly so at humeri except only obtusely angulate in subcitum; lateral margins more or less narrow; sides forming nearly smooth curves from humeri to apices; subapical sinuations obsolete or nearly so; apices narrowly, more or less independently 244 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY rounded, rarely with sutural angles denticulate (citum only); striae rather lightly impressed, not distinctly punctate but often irregular or vaguely punctulate; intervals slightly convex, outer ones not much modified toward apex, 3rd impunctate. Inner wings vestigial. Lower surface impunctate or nearly so; abdomen not pubescent; prosternal process simple. Legs: hind tibiae not suleate along outer edges; 4th hind-tarsal segment emarginate, not lobed; 5th hind-tarsal segment without obvious accessory setae; claws simple. Secondary sexual characters normal. Male copulatory organs as figured (Figs. 61-63). Genotype. Laevagonum cistelum n. sp. (below). Generic distribution. Known only from high altitudes on the Bismarck Range, N-E. N. G. Notes. The presence of anterior-lateral pronotal setae in one species argues against a derivation of this genus from Altagonwm. It has perhaps been derived independently from Notagonwm or from a Gastragonum-like ancestor. The species of Laevagonum, which are perfectly distinct but obviously rather closely related to each other, constitute another good example (cf. Nebriagonum) of apparent speci- ation in a limited mountainous area. All the species were found in or just above one continuous piece of heavy mountain forest. I have no record of the habitat of swbcitwm, the most distinct of the four, except that it occurred in forest. Of the other three species, citwm, with the smallest eyes, was found under comparatively deep stones and logs in forest and may be incipiently subterranean; cistelum, under various cover on the ground in forest; and subeistelum, under cover on the ground in open grassy areas just above the upper edges of the forest. The fact that these species are all flightless is consistent with their apparent isolation in slightly separated habitats. However, it does not necessarily follow that they originated as a result of ecological isolation, as I have already noted in the introduction. All the species of this genus share many characters covered in the preceding generic description, so their individual descriptions can be brief. Key to the Species of Laevagonum 1. Anterior-lateral pronotal setae present; elytral margin very obtusely angu- late at humeri; (Hwrophilus-like) (p. 245).................... subcitum Anterior-lateral (as well as posterior-lateral) pronotal setae absent; elytral | mareinrectangular, or nearly, seat, humeni. |) 405. eee eee 2 2. Europhilus-like; slender, prothorax more elongate, width/length .98 & .99; sutural angles of elytra denticulate (p. 245)...................: citum Cistelid- or Calathus-like; prothorax shorter, width/length 1.15 to 1.39; suturallanclesor elytra not denticulatescn sass oe eee eee eee 3 DARLINGTON: CARABID BEETLES OF NEW GUINEA 245 3. Larger (7.0-8.4 mm.); browner; elytra relatively longer and less rounded (0, BAG) &: 2 6 baleen ieee ae SRE icine ola oi bu bas Sines eg marie rad cistelum — Smaller (5.5-6.4 mm.); blacker; elytra relatively shorter and more rounded (G05 HIG) oa racic & eager Aika me ON RRC Decne ruben ican pao Bt 9 np subcistelum LAEVAGONUM SUBCITUM N. sp. Description. With characters of genus as described above. Euro- philus-like. Head only moderately elongate, .69 & .65 width prothorax; eyes slightly longer than and somewhat more prominent than genae. Prothorax subquadrate but with moderately rounded sides; width/ length 1.07 & 1.15; base slightly or scarcely wider than apex (angles too rounded for exact measurement of base); anterior-lateral seta (or puncture marking its position) present slightly before middle of pro- thoracic length on both sides in both specimens. Elytra with basal margin only very obtusely angulate at humeri; lateral margins only moderately narrow; apices with sutural angles obtuse, poorly defined, not denticulate. Male copulatory organs: Fig. 61. Measurements: length 6.3-6.4; width about 2.2 mm. Types. Holotype & (M.C.Z. No. 28,679) and 1 @ paratype both from Mt. Wilhelm, Bismarck Range, N-E. N. G., 7,000—-10,000 ite (forest), Oct. 1944 (Darlington). Measured specimens. The types. Notes. Sufficiently compared with other species in key above. LAEVAGONUM CITUM 0. sp. Description. With characters of genus as described above. Euro- philus-like, elongate. Head relatively elongate, .63 & .62 width pro- thorax; eyes as long as or slightly longer than and scarcely or slightly more prominent than genae. Prothorax elongate, appearing longer than wide and by measurement very slightly so; width/length .98 & .99; base slightly wider than apex (angles too rounded for exact measurement of base). Hlytra with basal margin about rectangular at humeri; lateral margins rather narrow; sutural angles usually conspicu- ously denticulate, but only vaguely so in 1 paratype. Male copulatory organs: Fig. 62. Measurements: length 7.2-8.4; width 2.4-2.8 mm. , Types. Holotype & (M.C.Z. No. 28,680) and 3 (2 2) paratypes all from Mt. Wilhelm, Bismarck Range, N-E. N. G., 7,000-10,000 ft. (forest), Oct. 1944 (Darlington). Measured specimens. The o holotype and 1 9 paratype. Notes. This species is distinguished from others in the key above, and its habits are indicated in discussion under the genus. 246 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY LAEVAGONUM CISTELUM N. sp. Description. With characters of genus as described above. Rather slender Calathus- or (even more) cistelid-like (Fig. 13). Head small and not very elongate, .53 & .51 width prothorax; eyes longer than genae but scarcely or only slightly more prominent. Prothorax strongly narrowed in front, much less so behind; width/length 1.24 & 1.39; base/apex about 1.8 & 1.7 (angles too rounded for exact measurement of base); sides variably, weakly to rather strongly rounded; basal angles also rather variable, moderately to very broadly rounded. Elytra with basal margin about rectangular at humeri; lateral margins narrow; apices with sutural angles narrowly rounded or somewhat distinct, but not denticulate. Male copulatory organs as figured (Fig. 63). Measurements: length 7.0-8.4; width 2.4-2.9 mm. Types. Holotype &@ (M.C.Z. No. 28,681) and 9 paratypes all from Mt. Wilhelm, Bismarck Range, N-E. N. G., 7,000—10,000 ft. (forest), Oct. 1944 (Darlington). Measured specimens. The o& holotype and 1 9 paratype. Notes. This too is sufficiently compared with other species in the key above, and its habitat is indicated in notes under the genus. LAEVAGONUM SUBCISTELUM n. Sp. Description. With characters of genus as described above. Calathus- like. Head small, only moderately elongate, .60 & .62 width prothorax; eyes longer than and a little more prominent than genae. Prothorax a little longer and less narrowed in front than in cistelum; width/length 1.20 & 1.15; base/apex 1.4, more or less (angles too rounded for exact measurement of base); sides moderately rounded throughout or straighter toward base; basal angles rather broadly rounded. Elytra relatively shorter and more rounded than in cistelum; basal margin strongly but usually somewhat obtusely angulate at humeri; lateral margins narrow; apices with sutural angles narrowly rounded, or at most faint and obtuse, not denticulate. Measwrements: length 5.5-6.4; width 2.0-2.3 mm. Types. Holotype @ (M.C.Z. No. 28,682) and 7 paratypes all from Mt. Wilhelm, Bismarck Range, N-E. N. G., slightly above 10,000 ft., Oct. 1944 (Darlington), under cover in open grassy places a little above the forest line. Measured specimens. The & holotype and 1 9 paratype. Notes. Adequately compared in key above, and mentioned also in discussion under the genus. DARLINGTON: CARABID BEETLES OF NEW GUINEA 247 FORTAGONUM new genus Diagnosis. Moderate-sized (8.8-12.4 mm.), heavily built, broadly subparallel or fusiform, more or less strongly convex; inner wings vestigial; wing-and-seta formula —w, —(+), ——,(—) (—) (—) (note anterior supraoculars absent). Description. Form as indicated above; black or brownish-black, elytra sometimes more or less iridescent or purplish; lateral margins of prothorax and elytra not distinctly translucent; appendages piceous or reddish; upper surface moderately shining; microsculpture more or less normal, but differing in detail in different species. Head variable; mandibles sometimes (not always) strikingly long, slender, and nearly straight; eyes more or less reduced but very variable in prominence; anterior supraocular setae absent, posterior ones present except absent in bufo, slightly behind or almost between posterior edges of eyes; antennae normal; neck not impressed above; front variable; mentum tooth triangular, sometimes more or less blunted or subtruncate at tip. Prothorax with anterior angles broadly and strongly advanced, with apices right-acute except as blunted or narrowly rounded; lateral margins wide at least basally but sometimes not sharply differentiated from disc; lateral pronotal setae absent; basal foveae obsolete or nearly so, scarcely distinct from wide basal parts of margins; base more or less normal; anterior marginal line entire, posterior one faint or obso- lete. Elytra always broad at base and usually rather short, not im- pressed on disc; variable in many details but always more or less narrowly margined at sides, with apices not armed; striae moderately to rather deeply impressed, impunctate; a more or less distinct partial 10th interval often (not always) present posteriorly. Inner wings vestigial. Lower surface impunctate or partly punctate; abdomen not pubescent; prosternal process usually simple, but margined and tu- berculate at apex in bufo. Legs: hind tibiae not sulcate along extreme outer edges; 4th hind-tarsal segment emarginate or moderately lobed; 5th hind-tarsal segment without obvious accessory setae, but some- times (at least in some fortellum) with minute rudimentary ones (as in many other New Guinean Agonini). Secondary sexual characters normal. Male copulatory organs as figured (Figs. 64-66). Genotype. Fortagonum fortellum n. sp. (below). Generic distribution. High mountains of New Guinea. Notes. This is an apparently natural group of surprisingly diverse species, of which many more probably remain to be discovered at high altitudes on different mountain ranges. 248 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Key to the Species of Fortagonum 1. Prothoracic margins very wide anteriorly as well as posteriorly; elytral margin rounded at humeri; elytra purplish; (form very broadly sub- quadrate; eyes small but very abruptly prominent) (p. 248)..... lamum — Prothoracic margins very narrow anteriorly, very wide only posteriorly; elytral margins abruptly angulate at humeri; elytra not purplish; (form and eyes variable). 6) gcc Jos oye Ponce wen Sake Sab ger Bee 2 2. Eyes small and scarcely more prominent than genae; mandibles long, slender, only slightly arcuate: -). 225-6 1490-4 cio (ae eee eee 3 — Eyes small but rather prominent, much more so than genae; mandibles variable but never so produced as above..................-:.4-0++: 4 3. Rather broadly subparallel; basal margin of elytra entire (p. 249). . forceps — Oval-fusiform; basal margin of elytra joining bases of 3rd striae, obliterated imwardly AC QoO)/s seecics. eke ea ENS ee eee cychriceps 4, Posterior supraocular setae present; form moderately broad, not fusiform Gaie2OL) IS a ieisinc tysgescs cryntens Gnechbsusntc tated a hisas | ae. oe fortellum — Posterior (as well as anterior) supraocular setae absent; form very broadly rounded=fhusiform: (o.252)\ ace elciae eocle.s cose act ae aie eee eee eae bufo FORTAGONUM LIMUM N. sp. Description. With characters of genus as described above. Very broadly subquadrate; elytra inconspicuously purplish, not iridescent; upper surface (at 54X) finely and sparsely punctulate, with some additional coarser but rather superficial punctation on lateral margins of prothorax; microsculpture normal. Head relatively small but with neck thick and swollen; head about .47 width prothorax; mandibles not especially elongate; eyes small but excessively prominent; front conspicuously, irregularly, transversely impressed between eyes, and anterior frontal impressions extending backward into very deep channels above and behind eyes. Prothorax very wide; width/length about 1.78; base/apex about 1.33; sides strongly rounded for most of length, becoming nearly straight near base; posterior angles obtuse, slightly blunted; lateral margins very wide anteriorly as well as pos- teriorly, moderately reflexed; disc with light median line and vague transverse impressions. Elytra broad and relatively short, rather strongly convex; humeri broad but prominently rounded rather than pointed; basal margin entire, rounded rather than angulate at humeri; sides straight and probably subparallel to behind middle, then strongly arcuate to distinct but not strong subapical sinuations; apices rather narrowly independently rounded; intervals a little convex, not much modified toward apex; marginal (10th) interval narrow and poorly defined; 3rd interval impunctate on left elytron, 1-punctate (just behind middle) on right one. Lower surface impunctate or nearly so, DARLINGTON: CARABID BEETLES OF NEW GUINEA 249 but epipleurae roughened. Legs: 4th hind-tarsal segment with moder- ate outer and much shorter inner lobe. Measurements: length 11.0; width about 4.8 mm. Type. Holotype 2 (M.C.Z. No. 28,683) from Mt. Misim, Morobe Dist., N-E. N. G., altitude and date not given (Stevens). Measured specimen. The type. : Notes. ‘The single specimen of this species is partly crushed, with the prothorax split lengthwise, so that the measurements and pro- portions given are only approximate. The species has the generic characters of Fortagonum and is probably related to the other species here placed in that genus, but it differs from all the others in a number of striking characters: color of elytra, excessive prominence of eyes, form of front of head, form of prothorax (with margins very wide anteriorly as well as posteriorly), and form of humeri (rounded rather than pointed). FoRTAGONUM FORCEPS Nn. sp. Description. With characters of genus as described above. Rather broadly subparallel; upper surface impunctate except very finely and sparsely punctulate, rather shining, elytra somewhat iridescent in strong light only; microsculpture normal, with elytral meshes very fine and transverse. Head .57 & .56 width prothorax; mandibles unusually long, slender, and only slightly arcuate; eyes small and only slightly more prominent than genae; latter about as long as eyes, convex in profile; front moderately convex, more or less impressed transversely between eyes, with anterior impressions moderate and extending vaguely backward above eyes but not forming deep channels there. Prothorax: width/length 1.28 & 1.35; base/apex about 1.33 & 1.35; sides rather weakly arcuate for much of length, straighter posteriorly; posterior angles a little obtuse, narrowly rounded; lateral margins relatively narrow anteriorly, very wide posteriorly, flat, only slightly reflexed; disc with median line moderately impressed, transverse im- pressions almost obsolete. Elytra broad and rather short, only a little more than normally convex; basal margin entire, strongly advanced and acute at humeri; subapical sinuations faint or absent; apices rather narrowly, more or less conjointly rounded (a little more independently rounded in the 9 paratype); intervals moderately convex, not much modified toward apex; an extra (10th) interval present for much of elytral length, moderately wide, flat or slightly convex; 3rd interval impunctate. Male copulatory organs: Fig. 64. Lower surface im- punctate. Legs: 4th hind-tarsal segment emarginate, not lobed. Measurements: length 12.4; width 4.8-4.9 mm. 250 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Types. Holotype & (Leiden Mus.) and 1 9 paratype (M.C.Z. No. 28,684) both from Moss Forest Camp, Snow Mts., Neth. N. G., 2,600-2,800 m. (about 8,450-9,100 ft.), Oct. 9-Nov. 5, 1938 (Toxopeus) Measured specimens. 'The types. Notes. Sufficiently compared with other species in the key above. This and the following species (cychriceps) are unique among New Guinean Agonini in the form of their mandibles which may be adapted to feeding on small snails or some other special food. The present species, except that it is larger, superficially resembles F. fortellum (below) of the Bismarck Range, but fortellum has approximately normal mandibles and differs in other technical details. FoRTAGONUM CYCHRICEPS N. sp. Description. With characters of genus as described above. Oval- fusiform; upper surface virtually impunctate except lateral margins of prothorax vaguely punctulate; shining, head and pronotum opalescent or slightly iridescent, elytra more iridescent; microsculpture normal on head, very fine, transverse, and scarcely visible at 54 on pronotum and elytra. Head relatively long and narrow, .48 width prothorax; mandibles long, slender, only weakly arcuate; eyes small, scarcely more prominent than genae; latter about as long as eyes, only faintly convex in profile; front convex, obliquely flattened above eyes, with slight, almost indistinct frontal impressions. Prothorax widest very near base, strongly narrowed anteriorly; width/length 1.27; base/apex about 1.8; sides weakly arcuate for whole length; posterior angles would be right except narrowly rounded; lateral margins very narrow anteriorly, very wide basally, even less reflexed (more nearly in plane of disc) than in other species of genus; disc more convex than usual, with median line lightly impressed, transverse impressions almost obsolete. Elytra rela- tively longer than in other species of genus, rather gradually tapering posteriorly, more convex than usual; basal margin reaching and joining ends of 3rd striae but obsolete inwardly, strongly advanced and acute at humeri; sides almost evenly rather weakly arcuate from humeri to apices, without subapical sinuations; apices somewhat independently rounded; striae deeper and intervals more convex than usual; outer intervals not much modified toward apex; marginal (10th) interval present for most of elytral length, rather wide especially posteriorly, nearly flat; 3rd interval impunctate. Lower surface virtually im- punctate. Legs: 4th hind-tarsal segment emarginate, not lobed. Measurements: length 11.5; width 4.3 mm. j Type. Holotype 9 (Leiden Mus.) from Mist Camp, Snow Mts., Neth. N. G., 1,800 m. (about 5,850 ft.), Jan. 1939 (Toxopeus) ; unique. DARLINGTON: CARABID BEETLES OF NEW GUINEA 25 Measured specimen. The type. Notes. Adequately characterized and compared with other species in the key above. ForRTAGONUM FORTELLUM N. sp. Description. With characters of genus as described above. Not fusiform (Fig. 14); upper surface impunctate (except elytral intervals sparsely punctulate at 54), moderately shining, elytra faintly iri- descent in strong light only; microsculpture normal except so fine as to be scarcely visible on elytra. Head rather small and narrow, .54 & .55 width prothorax; mandibles not especially long; eyes rather small, moderately prominent; genae about as long as eyes, oblique; front almost evenly convex, with very slight anterior impressions. Prothorax widest behind middle, strongly narrowed in front, much less so behind; width/length 1.41 & 1.44; base/apex 1.47 & 1.42; sides broadly, almost - evenly arcuate for whole length, or sometimes vaguely straighter both anteriorly and posteriorly; posterior angles a little obtuse, narrowly rounded; lateral margins relatively narrow anteriorly, very wide pos- teriorly, flat, scarcely reflexed; disc with middle line light, transverse impressions virtually obsolete. Elytra broad and rather short but otherwise of nearly normal outline in tribe; sides slightly arcuate at middle, more strongly so apically; dise a little more than normally convex; basal margin entire, about rectangular at humeri; subapical sinuations absent; apices conjointlyrounded to obtuse, blunted, slightly dehiscent sutural angles; intervals flat or slightly convex, very variable toward apex, outer ones sometimes not much modified, or 8th and 9th and sometimes others deeply longitudinally impressed toward apex; variable short fragment of extra 10th interval present in some speci- mens absent in others, when present, convex and sharply defined, at outer apical curve of elytron; 3rd interval variably punctate (e.g. type has left 3rd interval normally 3-punctate except middle puncture is: farther forward than usual, right 3rd interval 3-punctate with all punctures at 2nd stria; second measured specimen has left 3rd interval 2-punctate with punctures at 2nd stria about 24 from base and 14 from apex, and right 3rd interval 1-punctate with puncture at 2nd stria about 14 from base). Lower surface with sides of sterna superficially punctate. Legs: 4th hind-tarsal segment emarginate, not lobed. Male copulatory organs as figured (Fig. 65). Measurements: length 8.8-10.8; width 3.7-4.4 mm. Types. Holotype o& (M.C.Z. No. 28,685) and 53 paratypes all from Mt. Wilhelm, Bismarck Range, N-E. N. G., 7,000-10,000 ft., Oct. 1944 (Darlington), taken in and under various cover on the ground in heavy forest. 252 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Measured specimens. The o holotype and 1 @Q paratype. Notes. This is, of course, a thoroughly characterized species, dis- tinguished from others in the key above. The variation in form of the outer elytral intervals (variably impressed or not impressed), in presence or absence of a variable fragment of a 10th interval, and in punctures of the 3rd interval is amazing, but the variation is erratic, with the different characters not correlated, and the whole series is from a limited area of continuous forest and has the look and in many ways the characters of a single population, which I have no doubt it is. FoRTAGONUM BUFO N. sp. Description. With characters of genus as described above. Very broadly oval, very convex; upper surface virtually impunctate, only moderately shining, not iridescent; microsculpture normal. Head .49 & .51 width prothorax; mandibles somewhat longer and straighter than — usual, but less so than in forceps and cychriceps; eyes small, rather prominent, but much less so than in limwm; genae short and oblique; front convex, with slight anterior impressions. Prothorax widest in basal half, strongly narrowed anteriorly, not or scarcely so posteriorly ; width/length 1.52 & 1.48; base/apex about 1.73 & 1.56; sides slightly sinuate behind anterior angles, then moderately arcuate, then straighter and subparallel toward base; basal angles nearly right, blunted; lateral margins very narrow anteriorly, very wide posteriorly, strongly flattened posteriorly but scarcely reflexed; disc strongly convex, with middle line moderately impressed, transverse impressions almost obso- lete. Elytra widest well behind humeri, very convex, with strongly rounded sides; basal margin entire, a little obtusely (almost rectangu- larly) angulate at humeri; subapical sinuations slight or virtually absent; apices rather abruptly rounded (almost subangulate) about opposite 2nd striae, with sutural angles obtuse; striae rather deep; intervals moderately convex, not much modified toward apex; no dis- tinct extra (10th) interval; 3rd interval impunctate. Lower surface with sides of mesosternum rather lightly punctate. Legs: 4th hind- tarsal segment with a moderate outer and slightly shorter inner lobe. Male copulatory organs: Fig. 66. Measurements: length 10.4-11.3; width 5.0-5.2 mm. Types. Holotype o' (Leiden Mus.) and 1 2 paratype (M.C.Z. No. 28,686) both from Mist Camp, Snow Mts., Neth. N. G.., 1 800 m. (about 5,850 ft.), Jan. 1939 (Toxopeus). Measured specimens. ‘The types. Notes. This species is unique among New Guinean Agonini in form as well as in absence of both pairs of supraocular setae. cS \ AN \ | ‘ 4 fey! f 4 Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. a oo PWNEFHODOMONAMRwWN Ee PLATE 1! . Tarsagonum latipes new genus and species. Notagonum externum new genus and species. . Plicagonum fulvum new genus and species. Lithagonum annulicorne dilutior new genus and subspecies. . Iridagonum quadripunctum new genus and species. . Altagonum caducum new genus and species. . Maculagonum pox new genus and species. . Potamagonum diaphanum new genus and species. . Gastragonum terrestre new genus and species. . Idiagonum asperum new genus and species. . Montagonum toxopeanum new genus and species. . Nebriagonum cephalum new genus and species. . Laevagonum cistelum new genus and species. . Fortagonum fortellum new genus and species. V5: Right middle tarsus (from above, with bristles omitted) of Tarsa- gonum latipes (Q paratype). Fig. 16. Right 4th hind-tarsal segment (from above, with bristles omitted) of Tarsagonum latipes, with ‘fa long outer lobe but almost no inner one” (@ paratype). Fig. 17. Same of Euplenes apicalis, “with 2 long, nearly equal lobes” (2 paratype from Dobodura). Fig. 18. Same of Notagonum externum, with ‘‘moderate outer and shorter inner lobe” (Q paratype from Dobodura). Fig. 19. Same of Notagonum subpunctum capitis, with apex ‘“‘simply emargi- nate, not lobed” (@ paratype). 1 Figures are to different scales; see descriptions for sizes of insects. Dar.ineton. Carasio BEetLes oF New Guinea. PLATE 1 BULL. MUS. COMP. ZOOL. PLATE 2 Fig. 20. Male copulatory organs (middle lobe from left, parameres from below, detached) of Arhytinus major (type). Fig. 21. Same of Tarsagonum latipes (type). Fig. 22. Same of Euplenes apicalis (type). Fig. 23. Same of Dicranoncus queenslandicus 8]. (Guadaleanar Is.). Fig. 24. Same of Lorostemma informalis (type). Fig. 25. Same of Notagonum angustellum (type). Fig. 26. Same of Notagonum reversior (type). Fig. 27. Same of Notagonum externum (paratype). Fig. 28. Same of Notagonum vaporum (type). Fig. 29. Same of Notagonum gibbum (paratype). Fig. 30. Same of Notagonum dentellum (type). Fig. 31. Same of Notagonum malkini (type). Fig. 32. Same of Notagonum subrufum (type). Fig. 33. Same of Notagonum spinulum (type). Fig. 34. Same of Colpodes violaceus Chd. (Dobodura). 3924 Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. PLATE 3 . Same of Colpodes helluo (type). . Same of Colpodes laetus Er. (Dobodura). . Same of Colpodes habilis 8). (Wasian). . Same of Colpodes antedens (type). . Same of Colpodes acuticauda (paratype). . Same of Colpodes sinuicauda (type). . Same of Plicagonum rugifrons (type). . Same of Plicagonum fuluum (paratype). . Same of Lithagonum annulicorne dilutior (type). . Same of Iridagonum quadripunctum (type). . Same of Iridagonum subfusum (type). . Same of Altagonwm vallicola (type). . Same of Altagonum magnox (type). . Same of Altagonum caducum (paratype, Mt. Misim). . Same of Altagonum sphodrum (type). . Same of Altagonum paralimbus (type). Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. 51. 52. 53. 54. 55. 56. of. 58. 59. 60. 61. 62. 63. 64. 65. 66. PLATE 4 Same of Altagonum nudicolle (type). Same of Maculagonum setipox (type). Same of Potamagonum diaphanum (type). Same of Gastragonum laevisculptum (type). Same of Gastragonum subrotundum (type). Same of Gastragonum terrestre (type). Same of Idiagonum asperum (paratype). Same of Montagonum toxopeanum (paratype). Same of Nebriagonum cephalum (type). Same of Nebriagonum transitior (type). Same of Laevagonum subcitum (type). Same of Laevagonum citum (type). Same of Laevagonum cistelum (paratype). Same of Fortagonum forceps (type). Same of Fortagonum fortellum (paratype). Same of Fortagonum bufo (type). as teh Ve a wie? y cf a 3 ane Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vou. 107, No. 4 REVISION OF THE SPECIES CURRENTLY REFERRED TO ALEPOCEPHALUS, HALISAURICEPS, BATHYTROCTES AND BAJACALIFORNIA WITH INTRODUCTION OF TWO NEW GENERA By A. E. Parr American Museum of Natural History CAMBRIDGE, MASS., U.S. A. PRINTED FOR THE MUSEUM September, 1952 PUBLICATIONS ISSUED BY OR IN CONNECTION WITH THE MUSEUM OF COMPARATIVE. ZOOLOGY AT HARVARD COLLEGE BULLETIN (octavo) 1863 — The current volume is Vol. 107. Breviora (octavo) 1952 — No. 6 is current. Menmorrs (quarto) 1864-1938 — Publication was terminated with Vol. 55. JOHNSONIA (quarto) 1941 — A publication of the Department of Mollusks. Vol. 2, no: 30 is current. OccASIONAL PAPERS OF THE DEPARTMENT OF MOLLUSKS (octavo) 1945 — Vol. 1, no. 16 is current. PROCEEDINGS OF THE NEw ENGLAND ZOOLOGICAL CLUB (octavo) 1899- 1948 — Published in connection with the Museum. Publication terminated with Vol. 24. These publications issued at irregular intervals in numbers which may be purchased separately. Prices and lists may be obtained on application to the Director of the Museum of Comparative Zoélogy, Cambridge 38, Massachusetts. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vou. 107, No. 4 REVISION OF THE SPECIES CURRENTLY REFERRED TO ALEPOCEPHALUS, HALISAURICEPS, BATHYTROCTES AND BAJACALIFORNIA WITH INTRODUCTION OF TWO NEW GENERA By A. E. Parr American Museum of Natural History CAMBRIDGE, MASS., U.S. A. PRINTED FOR THE MUSEUM September, 1952 a “ Aa uey No. 4.— Revision of the Species Currently Referred to Alepocephalus, Halisauriceps, Bathytroctes and Bajacalifornia With Introduction of Two New Genera By A. E. Parr The discovery of a second specimen of Bathytroctes nasutus in the collections made by the ‘‘Atlantis” off Cuba, and now deposited in the Museum of Comparative Zoology, has provided an opportunity to re-examine more closely the relationships within the genus Bathytroctes as a whole, with the result that it now seems possible to effect several desirable subdivisions of the very heterogeneous assembly of species heretofore included in Bathytroctes. The collections of the Museum of Comparative Zoology also contain the types of several species of both Alepocephalus and Bathytroctes, which have not been redescribed since they were first introduced into the literature by Garman in 1899, and are therefore in need of redefi- nition in order to bring their identification up to date in relation to the other species known today. This would seem to be most easily accomplished by the publication of new keys to the genera involved, without burdening this account with the detailed redescriptions of the species, which will form part of a monograph on the Alepocephalidae now being prepared by the writer. In regard to Alepocephalus asperifrons Garman 1899 (p. 291, pl. LIX, fig. 1), type No. 28472 M.C.Z., this has already (Parr, 1951, p. 8) been made the type of a new genus, Bruunichthys. The other species of Alepocephalus described by Garman may be identified by the use of the following key, which extends to the species of this genus the revised key to the genera of the Alepocephalidae given by Parr, 1951, pp. 4-10. The functions used to express the proportions in per cent of the length without caudal fin, identified by the letter L, are those intro- duced and explained by Parr, 1949. To find the indicated minimum size of the head of a specimen of A. umbriceps of 250 mm. L, according to the key, solve the function (389.5 — .015L) as follows: 39.5 — .015 x 250 = 39.5 — 3.75 = 35.75 The head should be more than 35.75 per cent of L. If the head of the specimen is less than the minimum indicated for A. umbriceps, e.g. 33.9 per cent of L, one may, for comparison with umbriceps, estimate a corresponding formula for the specimen in one of the following ways. Either indirectly, by subtracting 33.9 from the value (35.75) which the formula gave as a minimum for wmbriceps at 250 mm. L, and subsequently subtracting (adding if the head of the specimen had been 256 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY larger than in wmbriceps) this difference (1.85) from the constant (39.5) of the wmbriceps formula. This gives the corresponding formula for the head of the specimen as (37.65 — .015L) per cent of L. Or directly, by solving the function for wmbriceps for the measure- ments of the specimen, treating the constant of the function as the unknown, K, as follows: (K — .015 x 250) = 33.9 K = 33.9 + .015 x 250 = 33.9 + 3.75 = 37.65 which gives the head of the specimen as: (37.65 — .015L) per cent of L. Note that this does not establish the actual formula for the species to which the specimen belongs. It only establishes a possible basis for comparison with winbriceps in purely descriptive terms. If the species is closely related, and the formula for wmbriceps is well established empirically, either for the species (as it is not in this case), or as an approximate descriptive generalization for the genus (which applies here) it is likely that the new formula will also be approximately valid for the new species, but for one specimen it only offers a method of evaluating the difference from wmbriceps at the size of the specimen under consideration. Key to the genus Alepocephalus A. Anal fin long, with 28-32 rays, 8-11 rays more than in dorsal fin. Anal fin base 20-24 per cent of L, or 5-9 per cent of L longer than dorsal fin base. Heads very small, less than (35.5 — .025 L), with an average of about (34.5 — .03 L) per cent of L. Orbits small, less than (12 — .027 L), average about (11 — .027 L) per cent of L. Snout short. Fins advanced, snout to V 46-50, snout to D 62-66, snout to A 62.5-66.5 per cent of L. 61-63 scales in longitudinal series. Only 8-11 simple pyloric caeca... . SALE Eee SNe Ai Aap nnn Tartine Cent a Ian le sh aan ere p Lloydiella, new subgenus Genotype: Alepocephalus bicolor Alcock 1892 B. Anal fin moderate to short, with 16-25 rays, not in excess of 4 rays more than in dorsal fin. Anal fin base only 12-19 per cent of L, and less than 5 per cent of L longer than base of dorsal fin. Heads more than (34.4 — .025 L), generally from (35.5 — .025 L) to (44 — .015 L) per cent of L. Orbits not less than (9 — .008 L) per cent of L. Snout to V 46-61, snout to D 65-76, snout to A 66-74 per cent of L. I. A long and pointed snout formed by the upper jaws extends beyond the point of lower jaw by more than 3 per cent of L. Length of snout about (15 — .006 L) per cent of L, or more. Vertical fins very posterior, distance from snout to both dorsal and anal fin (according to figure) about (72.5 + .01 L) per cent of L. Only 7 “rudimentary”’ VA OLICHC ae Came aa eee Subgenus Halisauriceps (Fowler 1934) A. longiceps Lloyd 1909 PARR: SPECIES REVISION OF ALEPOCEPHALID GENERA 257 II. Snout short and blunt, its length less than (13 —.006 L) per cent of L, its tip not extending beyond the point of lower jaw by more than one per cent of L, measured horizontally. One of the distanees from snout to dorsal or to anal fin always less than (70 + .01 L) per cent of L, both distances usually less than this value and always less than (2:21 Ole le A pericentiro fallin ase yee. setae yemle ek So feo spr Les 8 DRO tan tcre he ReEY dato vat tear vars Subgenus Alepocephalus (Risso 1820) a. Only 40-60 scales in a longitudinal series. 1. Distance from snout to anal fin 5-8 per cent of L longer than snout to dorsal. Only 9-10 simple pyloric caeca. Anal fin with 17-20 rays, its base 12-13 per cent of L. Head moder- ate, about (88 — .015 L) per cent of L. Orbit small, (10.2 — .008 L) per cent of L. Mouth small, upper jaws only about (11 — .008 L) per cent of L, lower jaws less than (Gi 008 1) tperccentionla: sain ese cla ee spel opt rays NE ei Sil ho ccda ch ae bevneeaNE A. andersoni Fowler 1934 2. Distance from snout to anal fin less than 3 per cent of L longer than snout to dorsal. 15-23 pyloric caeca. x. Anal fin with 20-24 rays, its base 17-20 per cent of L. 19-23 pyloric caeca. Heads more than (36 — .018 L) but less than (88 — .15 L) per cent of L. Orbits about (12.4 — .008 L) per cent of L. Mouth moderate, upper jaws less than (12 — .002 L), lower jaws not more than about (18 — .008 L), but more than (15 — .008 L) OCP. GSM OF Is bods ocuevcuoou. A. rostratus Risso 1820 xx. Anal fin with 16-20 rays, its base 12-15 per cent of L. 15-17 pyloric caeca.! y. Head and eyes small, head about 27-28, corresponding to (33.4 — .018 L) or (85.5 — .025 L) per cent of L, eye only about 7—7.2, corresponding to about (9. 6 — .008 L) per cent of L, at 310 mm. L. Fo So Rey Oe nae eT ta A. owstoni Tanai 1908 -yy. Head large, more than (37 — .018 L), orbits more than (11 — .008 L) per cent of L. z. Eyes moderate, orbits less than (12.5 — .008 L) penicentohisy eae A. australis Barnard 1923 1. Head large, from about (39 — .015 L) to (41 — .015 L) per cent of L............. Sy snr iae A. australis australis (Barnard 1923) 2. Head moderate, less than (39 — .015 L) DETCEMU OR serait: Sek vets Mi aMAM cas See A. australis barnardi (Norman 1930) zz. Hyes large, orbits about (14 — .008 L) per cent of L. Head large, corresponding to about: (Gu = OS Ib) jose CSN OH Wiss goss gods doe BARS CET tice be ates cal clea A. macrops Lloyd 1909 1 Not known for A. owstoni. 258 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY b. About 64-70 scales in a longitudinal series. Snout to dorsal fin from 3 per cent of L shorter to 5 per cent of L longer than snout to anal] fin. 13-19 pyloric caeca. 1. Anal fin with 21-25 rays, its base 15-20 per cent of L. D, 20-23. Distance from snout to ventrals 46-51.5, snout to dorsal 65-69, and snout to anal fin 66-71.5 per cent of L. Head small, less than (89.5 — .015 L) per cent of L...... Sct eee dt nas ne AA A. bairdi Goode and Bean 1879 2. Anal fin with only 17-19 rays, its base only 12-15 per cent of L. D, 16-18. Distance from snout to ventrals 54-58, snout to dorsal 69.5—72.5, snout to anal fin 71-73 per cent of L. x. Head small, less than (39.5 — .015 L) per cent of L.... ee a he ae ecriektthe acct Aatye ae A. blanfordi Alcock 18921 xx. Head large, more than (39.5 — .015 L) per cent of L. y. Upper jaw reaches approximately to below the middle of the orbit, being nearly 5 per cent of L longer than the snout. Suboperculum with two distinct ridges ending posteriorly in two separate points. 18'pyloricicaeca.y ou oc aie oe eee ee MSc tne A. umbriceps Jordan and Thompson 1914 yy. Upper jaws ending distinctly in advance of the centers of the orbits, being less than 3 per cent of L longer than the snouts. Suboperculum ends in a single rounded point, and has only a single ridge. 14-15 pyloric caeca........ A. productus Gill 1883 ce. §0-105 scales in a longitudinal series. 1. Head large, more than (41.5 — .017 L) per cent of L. 21-24 Jong, simple pyloric caeca. Width of skull more than (14 — .0028 L) per cent of L. Upper jaws more than (16) — 2008) per.centiof L252... 5.2.5) eee Sobre oe sinc Ree DAE kde A. agassizi Goode and Bean 1882 2. Head moderate, less than (40 — .017 L) per cent of L. Only 12-16 long, simple pyloric caeca. Width of skull not more than (14 — .0080 L) per cent of L. Upper jaws less than (14 — .002 L) per cent of L.2 x. 15-16 scales between lateral line and origin of dorsal fin. Upper jaws not more than (12 — .002 L), not less than (11 — .002 L) per cent of L. Head less than (37.5 — .018 L), width of skull not more than (12.5 — .008 L) Ac ne is very uncertain whether A. blanfordi is actually distinct from A. productus. The type ¢ the former is no longer available, and only the statement that its head is one-third of the A the body permits a tentative distinction to be made, until new material may be i . © _ ? Note the changes in coefficients of slope between A. agassizi and the other species under ec, in'regard to widths of skull (.0028 compared with .008) and lengths of upper jaw (.008 compared with .002). These formulas can, however, only be taken as purely descriptive of the data now available which are quite insufficient for the establishment of definitive norms. PARR: SPECIES REVISION OF ALEPOCEPHALID GENERA 259 per cent of L. Suboperculum irregularly quadrangular, with a broad, sloping posterior margin, and without distinct main ridge........ A. tenebrosus Gilbert 1891 xx. Only 10-12 scales between lateral line and origin of dorsal fin. Upper jaws not less than (12 — .002 L) per cent of L. Suboperculum with distinct main ridge ending in a point at the posterior margin. y. Head more than (88.5 — .017 L), width of skull about (13.7 — .008 L) per cent of L. Suboperculum narrow, sickle-shaped, ending posteriorly in a single point. Center of anus about one-third as far from the origin of anal fin as from the bases of anterior ventral finrays......... A. fundulus Garman 1899 ‘ yy. Head less than (86 — .018 L), or possibly less than (37.5 — .025 L), width of skull about (12 — .008 L) per cent of L. Suboperculum wide, but with a short, blunt, point in the posterior outline at the end of the main ridge. Center of anus not more than one-fourth as far from the origin of anal fin as from the bases of anterior ventral finrays...... A. convexifrons Garman 1899 LLOYDIELLA new subgenus The long anal fin separates A. bicolor, in an easily defined manner, from all other species of Alepocephalus. Actually A. bicolor would seem to be at least equally, if not more, unique with reference to several other features which cannot be quite so simply and easily ex- pressed. In scatter diagrams showing proportions in reference to abso- lute length the measurements of A. bicolor form a group entirely apart from the rest in regard to the length of the head and of the lower jaw; almost entirely apart with regard to the orbits, interorbital width, and the lengths of the upper jaws; and largely apart in regard to the width of the skull and the distances from the snout to dorsal and anal fin. Characteristic of the groups of measurements for which a trend of change with size can be fairly clearly established, is the fact that the coefficients of slope for the A. bicolor are much greater than those that seem indicated for any other species of Alepocephalus. Thus the average size of the head in A. bicolor may be fairly expressed as (34.5 — .03 L) per cent of L, while no other species of which there is enough material available gives indications of a coefficient of change greater than —.018L. Similarly the orbits and lower jaws of A. bicolor indicate a coefficient of —.027L, while the measured orbits and lower jaws of all other species are in fair agreement with coefficients of — .OO8L, or less. 260 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY While the proportions dealt with in the preceding paragraph might be viewed as only partly independent and partly dependent upon the size of the head, this cannot be said of the relationships of the inter- orbital width to the diameter of the orbit and to the width of the skull. In both subspecies of A. bicolor the interorbital widths expressed in per cent of the diameter of the orbit, and also in per cent of the width of the skull, clearly indicate a coefficient of change of both percentages of about +.2L. In all other species of Alepocephalus of which there are a reasonable number of measurements, the changes with size of the same percentages are well fitted by a coefficient of only +.06L. This is specifically true of A. tenebrosus and A. convexifrons, two of the species that approach most closely to A. bicolor in regard to the size of both the orbits, and the length as well as the width of the head. With so much concurrent evidence to indicate that A. bicolor oecu- pies a quite separate position within the genus, at least a separate subgenus seems in order. Subgenus HALSAURICEPS (Fowler 1934) Genus Halisauriceps Fowler 1934, p. 247. The new genus introduced by Fowler, 1934, was based solely upon the figure and brief description of Alepocephalus longiceps published by Lloyd, 1909. Unfortunately the type has been destroyed by accident, and no further information can be obtained about this species until fresh material is obtained. The figure shows no indication of a shoulder organ, so one may reasonably assume that the species is a true alepocephalid. But nothing is known of the supramaxillaries, or of the manner in which the promi- nent snout is formed by the upper jaws, the extent of the possible modification of the premaxillaries and of their connection with each other, and other features that might be involved. Lloyd mentions “seven rudimentary caeca,’’ which would be exceptional as a normal condition in a species of Alepocephalus. The present writer has several times found “rudimentary”’ caeca as an obviously abnormal condition of individual specimens, e.g. associated with parasites in the abdominal cavity or with unmistakable malformations. In some species of other genera, with only one or two caeca, these, although normal, may be so small that they might be described as rudimentary. But, with as ‘many as seven caeca present, their rudimentary size or condition seems most likely to represent an individual abnormality. In certain in- stances of an obviously abnormal rudimentary condition of the caeca, there is also a strong indication that the number of discernible caeca PARR: SPECIES REVISION OF ALEPOCEPHALID GENERA 261 may likewise have been reduced. It is therefore quite possible that the normal number of caeca in A. longiceps may be higher than seven. It is thus evident that the actual status of Haltsawriceps is quite obscure and must remain obscure until new material becomes available. The writer would hazard a guess that such material would serve to sharpen the distinction between A. longiceps and other alepocephalids, but, in the absence of factual knowledge, it seems best to retain the species within the genus in which it was just introduced, and in which its presently known features will still fit without broadening of the generic definition. BATHYTROCTES AND RELATED GENERA In the key to the genera of the Alepocephalidae previously published by the writer (Parr 1951, pp. 4-10) all the genera and subgenera dealt with in the following were still retained within a single genus, Bathy- troctes, with a definition broad enough to cover the peculiarities of all of these forms. The following key may therefore be used as an ex- tension of the previous key, starting after the point (J, B, 2, b, xz, zz, yy, v) on page 5, at which Bathytroctes, sensu lato, is defined. In a scatter diagram showing the relative lengths of the heads plotted against absolute lengths of the bodies (L) for all specimens of all species here considered, those that have been previously referred to the genus or subgenus Bajacalifornia, plus Bathytroctes calcaratus, which also belongs here, form an exceptionally well defined and well separated group, with all species apparently fitted by a single formula: Length of head (35.4 — .04L)"2 per cent of L. The same coefficient of change also applies with unusual accuracy to the two known speci- mens of “B.” nasutus, as follows: head (37.25 — .04L)*°° per cent of L. Among the rest of the species we also find two well separated groups, one with small, and one with large heads. Among the former we have only the measurements of single specimens of each species, so it is not possible to speak with confidence of the coefficient of slope that may actually apply to this group. But, using the coefficient of slope indi- cated for the species with large heads, we find that the measured lengths of the heads in the small-headed group do not exceed (35 — 03L) per cent of L, while the measured heads in the large-headed group all exceed (37.5 — .03L) and may even exceed (45 — .03L) per cent of L. It may be probable that the forms with smaller heads have a larger coefficient of slope, but even so, the measured heads would be less than (86.5 — .035L) and also less than (37.5 — .04L) 262 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY per cent of L. On any assumption the small-headed species thus would seem to form a well separated group. This separation of the small-headed group from those with large heads is confirmed by the position of the anus (at A, versus removed from A), agrees with very evident differences in general habitus which are not so easily defined in words, and greatly facilitates the general taxonomy of the group as a whole. A separate genus Grima- troctes is therefore introduced for these forms. The separation of Grimatroctes from the large-headed species re- tained in Bathytroctes also makes it possible to re-establish the genus Bajacalifornia. In Bathytroctes, sensu stricto one may find various degrees of moderate development of a symphyseal knob, which, although it never compares with the symphyseal knobs in large speci- mens of Bajacalifornia, may even exceed the relative magnitude of the knob in smaller specimens such as those from which Bajacalifornia draket is known. Grimatroctes on the other hand remains sharply distinct from Bajacalifornia in regard to symphyseal knob and the profile of snout and lower jaw. Bajacalifornia can thus be sharply distinguished from Bathytroctes, sensu stricto, by its very small head, and from Grimatroctes by the features first used to define Bajacalifornia as a separate genus. The genus Rinoctes is so sharply distinct from the others by the shape and structure of its snout, that no confusion is possible. The fact that it agrees with Bajacalifornia in the formula for the length of the head, sharpens the distinction beyond any practical need by opposing the projecting snout of Rinoctes against the prominent lower jaw of Bajacalifornia. Within the genus Bathytroctes, sensu stricto, B. michaelsarsi repre- sents a species of strikingly different appearance from that of the others. This difference in general appearance is confirmed by entirely different gillraker, and pectoral fin ray counts, and a new subgenus, at least, would seem in order. Key to genera of the Bathytroctes group A. Head small, less than (87 — .035 L) per cent of L. 1. Premaxillaries meet dorsally in a long and very firm, almost rigid symphysis closely and strongly joined by integuments, so as to form a hard, beak-like, pointed snout, projecting beyond the tip of lower jaw by more than | (about 1.3) per cent of L. Anus only one-half to two-thirds as distant from the insertion of the anterior ventral fin rays as from the origin of anal fin. Upper and lower jaws of approximately equal length, or upper jaws slightly longer. 6-8 PARR: SPECIES REVISION OF ALEPOCEPHALID GENERA 263 pyloric caeca. 20-25 gillrakers in first arch. P.8........2...... BB EE eo ect ST SAEs CREAN Cathy RIE EINER ES Wen uN Rinoctes, new genus Genotype: Bathytroctes nasutus Koefoed 1927 2. Premaxillaries do not form a prominent, pointed and beak-like snout, and are not joined in a long, almost rigid symphysis. Anus at anal fin, its centre about 3-15 times as distant from the insertion of anterior ventral fin rays as from the origin of anal fin. a. Symphysis of lower jaw with a prominent ventral knob, projecting forward beyond the vertical from the point of the snout, and continuing the dorsal profile of the head downward and forward. Head very small, not over (86 — .040 L) per cent of L. Eyes small, less than (14 — .034 L) per cent of L. Snout long, not less than (50 + .4 L) per cent of orbit. Length of lower jaw exceeds length of upper jaw by more than (4 — .01 L) per cent of L. 11-21 pyloric caeca. P. 13-17. About 50-65 scales in a longitudinaliseniestnsmsere sc ce mre ae eee er ike chrono ay esata Genus Bajacalifornia Townsend and Nichols 1925 b. Symphysis of lower jaw not prominent beyond the snout, with only a slight point ventrally. Eyes larger, more than (14 — .034 L) per cent of L. Snout shorter, not over (75 + .1 L) per cent of orbit. Jaws subequal or with lower jaw less than (4 — .01 L) per cent of L longer than upper jaw. 9-13 pyloric caeca. 32-38 gillrakers in first arch. 55-80 scales in a longi- tudinal series. P. 11-12. Measured heads not over (85 — .030 It) SDELRCenthOl spe et ua ea teat ee Grimatroctes, new genus Genotype: Bathytroctes grimaldi Zugmayer 1911 B. Head large, more than (87 — .030 L) per cent of L. Anus removed from anal fin, its center from about one-third to twice as distant from the insertion of anterior ventral fin rays as from the origin of anal fin. Less than 60 (about 42-55) scales in a longitudinal series. No beak-like snout. Symphysis of lower jaw not greatly enlarged and prominent. 6-9 pyloric CRE Canaetore Uyak Gen es Genus Bathytroctes Giinther 1878 RINOCTES new genus Genotype Bathytroctes nasutus Koefoed 1927, p. 50, pl. III, fig. 10. Diagnosis. No shoulder organ. Two supramaxillaries. Premaxil- laries with continuous free edge. Interoperculum normal, elongate, oriented obliquely upwards and backwards, covered anteriorly by the preoperculum. Body completely scaly in the adults. No scales on head. Pectoral and caudal fins normal, without produced rays. Origin of dorsal fin well in advance of the origin of anal fin, which has less than 20 rays. Teeth in jaws in single series; dentition of maxillaries approximately equal to, or more extensive than that of the premaxil- laries, which meet dorsally in a long, firm symphysis so as to form a 264 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY hard, beak-like, pointed snout, projecting beyond the point of lower jaw by more than | per cent of L. Anus well in advance of the middle of the distance between the insertion of the anterior ventral fin rays and the origin of anal fin. 6-8 pyloric caeca. 20-25 gillrakers in first arch. Pectorals small, with only 8 rays. Head small, less than (37 — .035 L) per cent of L. The genus contains only one species. RINOCTES NASUTUS Koefoed 1927 This species was originally described from a single specimen taken in the eastern portion of the Sargasso Sea, southwest of the Azores (N. 35°59’, W. 33°1’). It has not been reported again in the literature. It was therefore of considerable interest to find a second specimen in the collections of the Museum of Comparative Zoology (No. 35587) taken by the “Atlantis” inside of the Antillean chain of islands, off the south coast of Cuba, N. 20°47’, W. 80°24’, at Station 2966, Febru- ary 26, 1938. The fact that both specimens were taken at. extremely great depths, the type in 2865 meters, the “‘Atlantis” specimen in 3886 meters depth, suggests that we are here dealing with one of the most abyssal forms of living teleosts, since the records of alepocephalids of this general character plainly indicate that they belong to the bottom fauna, as explicitly recognized by Koefoed (1927) with whom the writer entirely agrees. On the assumption that the species may belong to the very great depths only, it is also of interest to note that the two speci- mens were taken in entirely separate ocean basins. Although the effectiveness of this separation must remain uncertain until the life history of the species is known, it does make it desirable to give the counts and measurements of the ‘‘Atlantis’’ specimen in some detail, as follows. D.141%. A.12%. P.8. V.7. Br. 7. Gillrakers in first arch 4/1/15. Pyloric caeca 6 (2 + 4). Length without caudal fin 122 mm. Proportions in per cent of length without caudal fin: Head 32.4. Orbit, longitudinally 9.1. Orbit, vertically 5.8. Snout 10.6. Snout to top of gill slit 29.1. Snout to top of preopercle 23.6. Interorbital width 2.95. Sphenotie width of skull 11.5. Pterotie width of skull 12.3. Combined length of upper jaws 17.1. Width of upper jaws (Max. + Supramax.) 3.3. Length of lower jaw 17.2. Length of premaxillary 5.9. Snout to dorsal fin 62.5. Snout to anal fin 74.5. Snout to ventrals 58.4. Base of dorsal 13.6. Base of anal fin 8.2. Insertion of ventral fins to center of anus 5.6. Greatest depth of body 13.9. Least depth of caudal peduncle 6.8. Longest gillraker 3.3. Longest pyloric caecum 5.3. Stomach sipho-caecal, i.e. with a bluntly pointed end, extending PARR: SPECIES REVISION OF ALEPOCEPHALID GENERA 265 beyond the pyloric arm by less than the diameter of the pyloric arm at the base. It has, unfortunately, not been possible to obtain usable counts of the scales. Genus BAJACALIFORNIA Townsend and Nichols 1925 Key to the species A. Very slender, not over 10 scales in transverse count, about 50-55 in longitudinal series. Diameter of orbit more than (60 — .075 L) per cent of lower jaw. 21 pyloric caeca. Stomach siphonal. About 26 gillrakers in first arch, about 19 in lower limb. Western Atlantic.............. ie ee RL OR Hees clo LUd lend ta reona twats RTL eNO Beasts B. drakei Beebe 1929 1. Only about 11 short pyloric caeca. Stomach siphonal. About 32 gillrakers in first arch, about 24 on lower limb. Diameter of orbit more than (60 — .075 L) per cent of lower jaw. 50-55 scales in longitudinal series. Pacific (Gulf of California)................. Pgh dA AME SANS WANES Cats Aeseaen eS B. burraget Townsend and Nichols 1925 2. 16-18 long pyloric caeca. Stomach caecal, with pyloric arm inserted in middle third -of the combined length of caecum and ventricle. Diameter of orbit less than (55 — .075 L) per cent of lower jaw. About 60-62 scales in longitudinal series. 24-25 gillrakers in first arch, 18-19 on lower limb. Indo-Pacific. .B. calcaratus Weber 1913 Syn.: B. burraget Norman 1939 GRIMATROCTES new genus Genotype Bathytroctes grimaldi Zugmayer 1911a, p. 1; 1911hb, p. 6, pl. I, fig. 2. Diagnosis. No shoulder organ. Two supramaxillaries. Premaxil- laries normal, with continuous free edge. Interoperculum normal, elongate, oriented obliquely upwards and backwards, covered anteri- orly by the preoperculum. Body completely scaly in the adults. No scales on head. Pectoral and caudal fins normal, without produced rays. Origin of dorsal fin well in advance of the origin of anal fin, which has less than 20 rays. Teeth in jaws in single series; dentition of maxillaries approximately equal to, or more extensive than that of the premaxillaries, which are not joined dorsally in a firm symphysis and do not form a prominent, pointed snout. Anus at anal fin. Sym- physis of lower jaw not prominent beyond the short snout, and with only a slight point ventrally. Head small, less than (37 — .035 L) per cent of L. Eyes large. 32-38 gillrakers in first arch. 9-13 pyloric caeca. 11-12 rays in pectoral fins. Seales small, 55-80 in a longitudinal series. 266 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Key to the species of Grimatroctes A. About 70-78 scales in a longitudinal series, 18-22 in transverse count. Pectoral fin bases completely surrounded by normal squamation in front and above, without naked band from gill opening to axil. 1. Only 9 pyloric caeca. Orbit about (13.5 — .02 L) per cent of L. BEG. 5 Skips Pee PEP Ca ee G. grimaldi (Zugmayer 1911) 2. About 13 pyloric caeca. Orbit about (14.5 — .02 L) per cent of L. IES Eisen ghee tes ttt ae eee en Oe ane G. microlepis (Giinther 1878) B. About 55-61 scales in a longitudinal series, less than 18 in transverse count. 1. Caudal peduncle deep, its depth (at 160 mm. L) about 10-11 per cent of L, and about one-half of the depth of the body at the shoulder. Sphenotic width of skull about (14.5 — .015 L); diameter of orbit (13.5 — .03 L) per cent of L. Width of skull at anterior end of orbit about 4.7 per cent of L..........6.80.: G. danae, Parr 1951 2. Caudal peduncle slender, its depth (at 220 mm. L) only about 6-6.5 per cent of L, and about one-third of the depth at the shoulder. Sphenotic width of the skull only about (13.3 — .015 L); diameter of orbit about (15.8 — .03 L) per cent of L. Width of skull at anterior end of orbit only about 3.4 per cent of L............... G. zugmayeri (Fowler 1934 Genus BATHYTROCTES Giinther 1878 Key to the subgenera and species A. Pectorals with 16-18 rays. 28-32 gillrakers in first arch, 20-23 in lower limb. 45-55 scales in longitudinal series. ... . Nomoctes, new subgenus Genotype: Bathytroctes michaelsarst Koefoed 1927 B. Pectorals with only 10-12 rays. Only 18-22 gillrakers in first arch, 13-15 in lower limb. 42-48 scales in longitudinal series................--- dha ceet ar Sie SEP MRA Gee Ae A Ae Subgenus Bathytroctes (Giinther 1878) 1. Head moderate, less than (39.5 — .03 L), sphenotic width of skull less than (17 — .015 L) per cent of L. a. Anus somewhat nearer to the origin of anal fin than to the in- sertion of anterior ventral fin rays. Lower edge of premaxillaries highly arched in lateral view, upper part without lateral tooth- plates. Eyes large, diameter of orbit corresponds to (19 — .03 L) per cent of L. Articulation of lower jaw approximately below middle of orbit. Anterior supramaxillary extends forward well beyond posterior supramaxillary...........-23- 4-6 -—- PAD ey ctor Smt ar ee te ea Cs IN B. inspector Garman 1899 b. Distance from the insertion of anterior ventral fin rays to center of anus only about 14-24 of distance from anus to anal fin. Premaxillary with a series of horizontal, semi-elliptic platelike PARR: SPECIES REVISION OF ALEPOCEPHALID GENERA 267 teeth along its upper portion, lower edge not highly arched. Eyes moderate, diameter of orbit less than (16 — .03 L) per cent of L. Articulation of lower jaw below posterior one-fourth of orbit. Anterior supramaxillary small, does not extend forward beyond posterior supramaxillary....B. alvifrons Garman 1899 2. Head large, more than (40.5 — .03 L), sphenotic width of skull more a. than (17 — .015 L) per cent of L. Diameter of orbit more than (16.5 — .03 L) per cent of L. Articulation of lower jaw below the posterior one-third of orbit. Anterior supramaxillary extends forward beyond posterior supramaxillary by one-fourth of its length, or more. Head about (41 — .03 L), sphenotic width of skull less than (17.5 — .015 L) per cent of L. Interorbital width of skull less than (7.5 + .15 L) per cent of orbit. Posterior supramaxillary extends forward somewhat beyond the end of the anterior one-third of the anterior supramaxillary. Indo-Pacific........ Soiled teens a aoe en eR eee B. macrolepis Giinther 1887 b. Head more than (41.5 — .03 L) to (44 — .03 L) per cent of L. Sphenotic width of skull more than (17.5 — .015 L) per cent of L. Interorbital width of skull more than (7.5 + .15 L) to (25 + .15 L) per cent of orbit. Posterior supramaxillary barely reaches to, or slightly beyond the middle of anterior supra- maxillary. Premaxillaries with upper, exterior row of horizontal, semi-elliptic platelike teeth. Atlantic. ..B. koefoedi, Parr 1951. 268 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY BIBLIOGRAPHY Aucock, A. 1892. Natura! history notes from the... “Investigator.” Ann. Mag. Nat. Hist., ser. 6, vol. 10, no. 59, pp. 357-362. BARNARD, K. H. 1923. Diagnoses of new species of marine fishes from South African waters. Ann. So. Afr. Mus., Cape Town, vol. 13, pp. 489-444. BreEse, W. 1929. Deep sea fish of the Hudson Gorge. Zoologica, New York, vol. 12, pp. 1-19. Fow er, H. W. 1934. Descriptions of new fishes obtained 1907 to 1910 chiefly in the Philippine Islands and adjacent seas. Proc. Acad. Nat. Sci., Philadelphia, vol. 85, pp. 233-367. GARMAN, S. 1899. Reports on an exploration... by the... “Albatross,” during 1891. Mem. Mus. Comp. Zool., Harvard Coll., vol. 24, pp. 1-431. » GILBERT, C. H. 1891. Scientific results of explorations by the... ‘Albatross.’”’ Proc. U.S. Nat. Mus., vol. 14, pp. 545-546. (Canin, Al, 1883. Diagnosis of new genera and species of deep-sea fish-like verte- brates. Proc. U.S. Nat. Mus., vol. 6, p. 256. GoopgE, G. B., and T. H. BEan 1879. Description of Alepocephalus bairdii. Proc. U. S. Nat. Mus., vol. 2, p. 55. Gunruer, A. 1878. Preliminary notices of deep-sea fishes collected during the voyage of H.M.S. “Challenger.” Ann. Mag. Nat. Hist., ser. 5, vol. 2, pp. 248-251. 1887. Report on the deep-sea fishes collected by H.M.S. “Challenger” during the years 1873-76. Challenger Report, vol. 22, pt. 57, pp. 1-268. JORDAN, D.S., and W. F. THompson 1914. Record of fishes obtained in Japan in 1911. Mem. Carnegie Mus., Pittsburgh, vol. 6, pp. 205-313. KoEFoED, E. 1927. Fishes from the sea-bottom. Rept. Sci. Res. “Michael Sars’ N. Atlantic Exped. 1910, vol. 4, pt. 1, pp. 1-147. PARR: SPECIES REVISION OF ALEPOCEPHALID GENERA 269 Luoyp, R. E. 1909-10. A description of the deep-sea fish caught by R.I.M.S. ship “Investigator”? since the year 1900. Mem. Indian Mus., vol. 2, pp. 139=180. Norman, J. R. 1930. Oceanic fishes and flatfishes collected in 1925-1927. Discovery Report, vol. 2, pp. 261-370. 1939. Fishes. Sci. Rept. John Murray Exped., London, vol. 7, no. 1, pp. 1-116. Parr, A. E. 1949. An approximate formula for stating taxonomically significant proportions of fishes with reference to growth changes. Copeia, 1949, no. 1, pp. 47-55. 1951. Preliminary revision of the Alepocephalidae, with the introduction of a new family, Searsidae. Amer. Mus. Novitates, no. 1531, pp. 1-21. Risso, A. 1820. Memoire sur un nouveau genre de poisson. Mem. Ac. Sci. Nat. Torino, vol. 25, pp. 270-272. Tanaka, 8S. 1908. Notes on some Japanese fishes. Jour. Coll. Sci. Tokyo, vol. 23, art. 7, pp. 1-54. TownseEnD, C. H., and J. T. NicHois 1925. Deep sea fishes of the ‘‘Albatross’’? Lower California expedition. Bull. Amer. Mus. Nat. Hist., vol. 52, pp. 1-20. WEBER, M. 1913. Die Fische der Siboga Expedition. Siboga Expeditie. Leyden. pp. 1-710. ZUGMAYER, H. 191la. Diagnoses des poissons nouveaux provenant des campagnes du yacht ‘‘Princesse Alice.’’ Bull. Inst. Oceanogr. Monaco, no. 193, pp. 1-14. 1911b. Poissons provenant des campagnes. Res. Camp. Sci. Monaco. fase. 35, pp. 1-159. = nm AGp canaster **erythrophthalmus > rileyr Breeding Males.2) ARKANSAS: Washington Co., 4 (AMNBH). CONNECTICUT: Fairfield Co., 3 (AMNH); New Haven Co., 2 (AMNH). GEORGIA: Fannin Co., 1 (MMP), 2 (USNM); Haber- sham Co., 1 (USNM); Rabun Co., 4 (USNM), 1 (LSU); Townes Co., 1 For explanation of color nomenclature see discussion on p. 329. 2 Breeding material includes all those individuals utilized in arriving at the estimates of the peeactens of the population. For this reason some atypical specimens are included under this eading. 288 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY 1* (USNM); Union Co., 1* MMP), 2 (USNM); White Co., 1 (USNM). ILLINOIS: Cook Co., 4 (CNHM); Fulton Co., 1 (CNHM); Lake Co., 1 (AMNH); Marshall Co., 1 (CNHM); Pulaski Co., 4 (CNHM); Will Co., 7 (CNHM). INDIANA: Knox Co., 1 (AMNH); Lake Co., 1 (CNHM); Marion Co., 1 (UMMZ); Wells Co., 2 (CNHM). IOWA: Linn Co., 1 (CNHM); Marion Co., 1 (AMNH). KENTUCKY: Madi- son Co., 1 (USNM); Union Co., 1 (USNM); Harlan Co., 1 (USNM). MARYLAND: Montgomery Co., 1 (CM); Prince George Co., 1 (AMNH). MASSACHUSETTS: Barnstable Co., 1 (AMNH); Hampshire Co., 1 (AMNH); Middlesex Co., 4 (AMNH); Norfolk Co., 1 (AMNH), 1 (USNM). MICHIGAN: Berrien Co., 1 (UMMZ); Calhoun Co., 1 (UMMZ); Charlevoix Co., 3 (UMMZ); Cheboygan Co., 1 (MMP); Chippewa Co., 1 (UMMZ); Huron Co., 1 (UMMZ); Tonia Co., 1 (AMNH); Kalamazoo Co., 2 (UMMZ); Livingston Co., 3 (UMMZ); Menominee Co., 1 (UMMZ); Oscoda Co., 1 (UMMZ); Washtenaw Co., 4 (MMP), 3 (GMS), 3 (UMMZ); Wexford Co., 1 (AMNH). MINNESOTA: Hennepin Co., 1 (AMNH); Olmstead Co., 2 (AMNH). MISSOURI: Wayne Co., 1 (USNM). NEW HAMP- SHIRE: Hillsboro Co., 1 (AMNH). NEW JERSEY: Bergen Co., 4 (AMNH); Essex Co., 4 (AMNH); Mercer Co., 1 (AMNH); Middle- sex Co., 1 (AMNH); Morris Co., 2 (AMNH); County Unknown, 2 (AMNH). NEW YORK: Erie Co., 1 (AMNH); Jefferson Co., 1 (AMNH); Kings Co., 3 (AMNH); Nassau Co., 1 (MMP), 7 (AMNH); New York Co., 3 (AMNH); Orange Co., 1 (AMNH), 4 (USNM); Putnam Co., 1 (AMNH); Queens Co., 15 (AMNH); Rensselaer Co,. 1 (AMNH); Richmond Co., 1 (AMNH); Rockland Co., 3 (AMNH); Suffolk Co., 1 (CM), 1 (UMMZ), 52 (AMNH); Westchester Co., 6 (AMNH). NORTH CAROLINA: Buncombe Co., 5 (USNM); Burke Co., 1 (USNM); Cherokee Co., 1* (USNM); Greenville Co., 1 (USNM); Jackson Co., 1 (MCZ); Macon Co., 1* (USNM), 2 (USNM), 1* (DZUG), 6 (DZUG); Sampson Co., 1 (USNM); Tran- sylvania Co., 2 (USNM); Watauga Co., 2 (USNM); Yancey Co., 1* (USNM), 2 (USNM). NORTH DAKOTA: Towner Co., 1 (CNHM), 6 (UMMZ). OHIO: Portage Co., 1 (AMNH). PENNSYLVANIA: Beaver Co., 7 (CM); Bedford Co., 1 (CM); Blair Co., 1 (AMNH); Butler Co., 1 (CM); Cambria Co., 1 (AMNH), 2 (CM); Columbia Co., 1 (CM); Crawford Co., 3 (CM); Erie Co., 2 (AMNH); Fayette Co., 1 (LSU); Pike Co., 1 (AMNH); Washington Co., 1 (CM). RHODE ISLAND: Providence Co., 4 (AMNH). SOUTH CARO- LINA: Anderson Co., 1 (USNM); Cherokee Co., 1 (USNM); Green- ville Co., 2 (CHAM), 1* (USNM), 1 (USNM), 1* (AMNBH), 2 (AMNH); Pickens Co., 2* (USNM). TENNESSEE: Anderson Co., 1 (USNM); Campbell Co., 2 (UMMZ); Cocke Co., 1 (USNM); Cum- DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 289 berland Co., 1 (USNM); Johnson Co., 5 (USNM); Munroe Co., 1 (UMMZ); Sullivan Co., 1 (USNM); Van Buren Co., 1 (USNM). VIRGINIA: Accomac Co., 4* (USNM), 1 ({USNM); Arlington Co., 1 (AMNH); Augusta Co., 3 (USNM); Fairfax Co., 1* (AMNH), 1 (AMNH); Highlands Co., 2 (USNM); Norfolk Co., 1 (LSU), 1* (USNM); Northampton Co., 2 (USNM); Princess Anne Co., 1* (UMMZ); Rockbridge Co., 5 (USNM); Smythe Co., 1 (USNM); Spotsylvania Co., 1 (USNM); Wythe Co., 4 (MCZ). VERMONT: Wyndham Co., 1 (AMNH). WASHINGTON, D. C.: 2 (CM). WEST VIRGINIA: Calhoun Co., 1 (USNM); Barbour Co., 1 (USNM); Brooke Co., 1 (MMP); Nicholas Co., 1 (USNM); Pendle- ton Co., 1 (MMP); Pocohontas Co., 2 (USNM); Randolph Co., 1 (USNM); Tucker Co., 1 (UMMZ), 1 (USNM). WISCONSIN: Burnett Co., 1 (USNM); Dodge Co., 7 (CNHM); Juneau Co., 2 (AMNH); Rock Co., 1 (FSM); Walworth Co., 3(AMNH);?Selinctous? ie (CMB): Breeding females. CONNECTICUT: New Haven Co., 2 (AMNH). GEORGIA: Habersham Co., 1 (USNM); Rabun Co., 1 (USNM); Townes Co., 2 (USNM); White Co., 1* (USNM). KENTUCKY: Harlan Co., 1 (USNM); Union Co., 1 (USNM). MARYLAND: Montgomery Co., 5 (USNM); Prince George Co., 1 (AMNH). MASSACHUSETTS: Barnstable Co., 1 (AMNH); Middlesex Co., 1 (AMNH). MICHIGAN: Berrien Co., 1 (UMMZ); Charlevoix Co., 2 (UMMZ); Chippewa Co., 1 (UMMZ); Huron Co., 1 (UMMZ); Jackson Co., 2 (UMMZ); Kalamazoo Co., 1 (UMMZ); Lapeer Co., 1 (UMMZ); Leelanau Co., 1 (UMMZ); Washtenaw Co., 4 (MMP), 1 (GMS). NEW YORK: Kings Co., 1 (AMNH); Nassau Co., 2 (AMNH), 1 (USNM); Orange Co., 2 (AMNH); Queens Co., 2 (AMNH); Suffolk Co., 7 (AMNH), 1 (UMMZ). NORTH CARO- LINA: Buncombe Co., 1 (LSU), 1 (USNM); Macon Co., 1 (MCZ); Sampson Co., 1 (USNM); Wake Co., 1 (NCS); Watauga Co., 1 (USNM). NORTH DAKOTA: Towner Co., 3 (UMMZ). PENN- SYLVANIA: Beaver Co., 1 (CM); Cambria Co., 1 (CM); Crawford Co., 1 (CM); Westmorland Co., 2 (CM). SOUTH CAROLINA: Greenville Co., 2 (AMNH), 1 (CHAM), 1 (USNM). TENNESSEE: Cocke Co., 1 (USNM); Cumberland Co., 1* (USNM), 1 (USNM); Johnson Co., 1 (USNM). VIRGINIA: Accomac Co., 1 (USNM); Arlington Co., 1 (AMNH); Giles Co., 1 (AMNH), 2 (MCZ); High- lands Co., 2 (USNM); North Hampton Co., 1 (CM); Elliot Knob, 1 (USNM). WASHINGTON, D. C.:1(AMNB), 2 (CM), 1 (USNM). WEST VIRGINIA: Barbour Co., 1 (USNM); Ohio Co., 1 (MMP); Randolph Co., 1 (USNM); Tucker Co., 1 (AMNH), 1 (GMS); Zeld, 1 (USNM). WISCONSIN: Vilas Co., 1 (USNM); Walworth Co., 2 (AMNH). 290 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Wintering Males.1| ALABAMA: Autauga Co., 1 (USNM); Houston Co., 1* (USNM); Montgomery Co., 2 (USNM). ARKANSAS: Crawford Co., 1 (USNM). FLORIDA: Alachua Co., 1 (AMNH), 4 (FSM), 1 (JCD); Citrus Co., 1 (FSM); Duval Co., 2 (MCZ), 4 (USNM); Escambia Co., 2 (USNM); Leon Co., 1* (CNHM), 1 (CNHM); Levy Co., 1 (AMNH), 1 (MCZ), 1 (JCD); Nassau Co., 3 (AMNH); Okaloosa Co., 7 (CM); Okeechobee Co., 1 (USNM); Pinellas Co., 1 (MCZ); Polk Co., 1 (USNM); Putnam Co., 2 (USNM); 2 (CNHM); St. Johns Co., 2 (AMNH); Santa Rosa Co., 8 (CM); Wakulla Co., 2 (USNM), 1* (USNM), 2 (FSM), 1 (JCD). GEORGIA: Barrow Co!, 1 (USNM); Bibb Co., 1 (USNM); Camden Co... 1 (MCZ); Catoosa Co., 1 (USNM); Charlton Co., 3 (USNM); Chatham Co., 2 (MCZ); 1 (USNM); Cherokee Co., 1 (USNM); Clarke Co., 8 (USNM), 1* (USNM); Cobb Co., 1 (CNHM); De Kalb Co., 5 (USNM), 2* (USNM); Early Co., 1 (USNM); Fulton Co., 6 (UUSNM); Glynn Co., 1(MCZ); Hall Co., 1 (USNM); Heard Co., 1 (USNM); McIntosh Co., 1 (AMNH); Thomas Co., 1 (AMNH), 1 (USNM). KENTUCKY: Butler Co., 1 (USNM). LOUISIANA: Cameron Parish, 1 (LSU); East Baton Rouge Parish, 2 (CNHM), 4 (LSU), 1* (LSU), 1 (USNM); Ouachita Parish, 1 (LSU); Orleans Parish, 1* (LSU), 2 (USNM); West Feliciana Parish, 2 (LSU). MARYLAND: Worcester Co., 1 (USNM). MISSISSIPPI: Bolivar Co., 1 (CNHM); Harrison Co., 6 (USNM), 4 (LSU); Warren Co., 2 (LSU). NEW JERSEY: Morris Co., 1 (AMNH). NORTH CAROLINA: Bruns- wick Co., 1 (NCS); Buncombe Co., 3 (USNM), 1 (MCZ); Carteret Co., 1 (USNM); Robeson Co., 1 (CNHM); Transylvania Co., 2 (USNM); Yancey Co., 1 (USNM). OHIO: Pickaway Co., 1 (AMNBH). SOUTH CAROLINA: Beaufort Co., 2 (MCZ); Berkeley Co., 1 (CHAM); Charleston Co., 5 (CHAM), 1 (LSU), 1 (USNM); Horry Co., 1 (CHAM); Kershaw Co., 5 (USNM); Pickens Co., 1 (CHAM). TENNESSEE: Giles Co., 2 (USNM); Hamilton Co., 1 (USNM); Shelby Co., 3 (LSU). TEXAS: Lee Co., 1 (AMNH); Nueces Co., 1 (AMNH). WEST VIRGINIA: Brooke Co., 1 (GMS); Cabell Co., 1 (USNM). Wintering females. ALABAMA: Autauga Co., 1 (USNM); Jackson Co., 1* (USNM); Ardell, 2. (USNM); Orange Beach, 1 (USNM). FLORIDA: Alachua Co., 2 (FSM), 1 (JCD); Duval Co., 1 (MCZ); Levy Co., 1* (MCZ); Okaloosa Co., 5 (CM); Okechobee Co., 1 (USNM); Santa Rosa Co., 2 (CNHM), 1* (CNHM); Volusia Co., 1 (MCZ); Wakulla Co., 5 (USNM), 3 (FSM); Cow Creek, 1* (CNHM). GEORGIA: Clarke Co., 2 (MMP), 1 (USNM); Charlton Co., 1 1 Wintering and migrant specimens were identified as individuals on the basis of physical characteristics. DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 291 (USNM); Echols Co., 1 (USNM); Richmond Co., 1 (USNM); Black- beard Island, 1 (USNM). KENTUCKY: Hopkins Co., 1 (USNM). LOUISIANA: Baton Rouge Parish, 1 (CNHM), 2 (LSU); Orleans Parish, 1 (LSU), 1 (USNM); Ouachita Parish, 1 (LSU); St. John the Baptist Parish, 1 (LSU); Washington Parish, 1 (LSU); Chef Menteur, 1 (CNHM). MISSISSIPPI: Harrison Co., 4 (USNM), 1 (LSU); Jackson Co., 1 (CNHM); Warren Co., 1 (LSU); Lobdell, 1 (LSU). NORTH CAROLINA: Buncombe Co., 1 (USNM); Pasquotank Co., 1 (USNM); Rockingham Co., 1* (USNM); Wayne Co., 1 (USNM). SOUTH CAROLINA: Charleston Co., 1 (CHAM); Georgetown Co., 1 (USNM); Kershaw Co., 1 (USNM), 1* (USNM). TENNESSEE: Giles Co., 1 (USNM); Roane Co., 1 (USNM). TEXAS: Bee Co., 1 (AMNH); Cook Co., 1 (USNM); Galveston Co., 1 (AMNH). Hardin Co., 1 (AMNH). WEST VIRGINIA: Brooke Co., 1 (GMS). Migrant Males. ALABAMA: Jackson Co., 1 (USNM); Tuscaloosa Co., 1 (USNM). CONNECTICUT: Windham Co., 2 (AMNH). FLORIDA: Alachua Co., 1 (DBUF); Dixie Co., 1 (FSM); Escambia Co., 1* (LSU); Nassau Co., 1 (AMNH); Wakulla Co., 1 (FSM). GEORGIA: Bullock Co., 1 (DZUG); Camden Co., 3 (MCZ); Chat- ham Co., 1* (USNM), 1 (USNM); Clarke Co., 7 (USNM); Cobb Co., 5 (CNHM), 1 (MMP); De Kalb Co., 1 (USNM); Fannin Co., 1 (USNM); Hall Co., 1 (USNM); Liberty Co., 1 (USNM); Thomas Co., 1 (AMNH); Union Co., 1 (USNM). ILLINOIS: Wabash Co., 1 (AMNH). KENTUCKY: Bell Co., 3 (USNM); Carroll Co., 1 (USNM); Hopkins Co., 1 (USNM); Rockcastle Co., 1 (USNM); Rowan Co., 1 (USNM); Trigg Co., 1 (USNM). LOUISIANA: East Baton Rouge Parish, 2 (LSU), 1* (LSU). MARYLAND: Cecil Co., 1 (AMNH); Montgomery Co., 2 (AMNH), 2 (CM); Prince George Co., 1 (AMNH). NEBRASKA: Lancaster Co., 1 (AMNH); Bald Island, 1 (USNM). NEW YORK: New York Co., 1 (AMNH); Queens Co., 1 (GMS); Suffolk Co., 2 (AMNH). NORTH CARO- LINA: Buncombe Co., 1 (AMNH); Dare Co., 1 (USNM); Iredell Co., 1 (USNM); Pasquotank Co., 1 (USNM); Wake Co., 3 (NCS). PENNSYLVANIA: Erie Co., 1 (CM). SOUTH CAROLINA: Charleston Co., 4 (MCZ), 1 (CNHM); Cherokee Co., 1* (USNM); Georgetown Co., 1 (USNM); Richland Co., 1 (CNHM). TEN- NESSEE: Fayette Co., 2 (USNM); Lake Co., 1 (USNM); Lincoln Co., 1 (USNM); Obion Co., 1 (USNM); Roane Co., 2 (USNM); Stewart Co., 1 (USNM). VIRGINIA: Alexandria Co., 3 (USNM); Arlington Co., 1 (USNM); Essex Co., 1 (USNM); Fairfax Co., 2* (MCZ), 5 (USNM); Orange Co., 1 (USNM). WEST VIRGINIA: Cabell Co., 1 (USNM); Greenbriar Co., 1 (AMNH); Logan Co., 1 (USNM). 292 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Migrant Females. ALABAMA: Colbert Co., 2 (USNM); Walker Co., 1 (USNM); Sand Mt. 1 (USNM). FLORIDA: Alachua Co., 2 (FSM), 1 (JCD); Wakulla Co., 1 (FSM). GEORGIA: Chatham Co., 1 (MCZ); Clarke Co., 1 (USNM); Cobb Co., 1 (CNHM). ILLI- NOIS: Wabash Co., 1 (USNM). KENTUCKY: Carroll Co., 1 (USNM); Mead Co., 1 (USNM); Trigg Co., 1 (USNM). LOUISI- ANA: Orleans Parish, 1 (LSU). MICHIGAN: Jackson Co., 1 (GMS). MISSISSIPPI: Bolivar Co., 1 (USNM), 1* (USNM); Harrison Co., 2 (LSU), 4 (USNM). NEW JERSEY: Morris Co., 1 (AMNBH). NEW YORK: Orange Co., 1 (USNM); Queens Co., 1 (AMNH); Dunwoodie, 1 (AMNH). NORTH CAROLINA: Charleston Co., 2 (USNM). TENNESSEE: Carter Co., 1* (USNM); Roane Co., 1 (USNM); Shelby Co., 1 (USNM); Union Co., 1 (UMMZ). TEXAS: Chambers Co., 1 (AMNH). VIRGINIA: Fairfax Co., 2 (USNM); Surrey Co., 1 (USNM). WASHINGTON, D. C.:1 (USNM). WEST VIRGINIA: Cabell Co., 1 (USNM); Logan Co., 1 (USNM); Pendle- ton Co., 1 (USNM); Rorer, 1 (CM); Fourleen, 1 (USNM). PIPILO ERYTHROPHTHALMUS ALLENI Coues Pipilo alleni Coues (1871: 366, footnote), original description. Type locality: Dummitt’s Grove, Indian River, Florida. Pipilo erythrophthalmus var. alleni, Baird, Brewer, and Ridgway (1874: 112). Pipilo leucopis Maynard (1878: 113, pl. IV), original description. Type lo- cality: Dummitt’s Grove, Florida. Pipilo erythrophthalmus subsp. Pipilo alleni, Sharpe (1888: 746). Description Diagnosis. A small, medium-billed, pale-eyed race, showing very little white on the rectrices. Average dimensions of males. Wing, 80.47 + .29 (c, 2.62); tail, 91.43 + .42 (c, 3.74); exposed culmen, 15.05 + .07 (c, 0.61); width of lower mandible, 8.48 = .04 (c, 0.33); tarsus, 26.78 = .11 (¢, 0.95); middle toe without claw, 18.88 = .09 (c, 0.81); length of white on inner web of outermost rectrix, 19.71 + .50 (c, 4.44). (See Table 3). Average dimensions of females. Wing, 76.50 + .41 (o, 2.08); tail, 85.42 + .67 (c, 3.40); exposed culmen, 14.71 + .10 (c, 0.48); width of lower mandible, 8.47 = .06 (c, 0.28); tarsus, 25.89 + .17 (c, 0.89); middle toe without claw, 18.57 = .14 (c¢, 0.71); length of white on inner web of outermost rectrix, 15.58 + .69 (¢, 3.52). (See Table 4.) Average color of males. Back, Iron Gray; flanks ,12-F-10; breast, Iron Gray; top of head, Iron Gray. DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 293 Average color of females. 14-J-9; top of head, 15-E-12. Measurement Wing Tail Tail Spot Culmen Mandible Width Tarsus Middle Toe Measurement Wing Tail Tail Spot Culmen Mandible Width Tarsus Middle Toe Iris color. Straw. Breeding range. Florida (from Bay, Franklin, Levy, Alachua and Putnam Counties south through the peninsula); absent from the Florida Keys. (See Map 2.) Winter range. In Florida (from Santa Rosa, Franklin, Leon, Colum- bia and Duval Counties south through the peninsula) ; absent from the Florida Keys. (See Map 3.) Mean 80.47 91.43 IAEA 15.05 8.48 26.78 18.88 Mean 76.50 85.43 15.58 14.71 8.47 25.89 18.57 Back, 15-E-7; flanks, 13-D-11; breast, Table 3 P. e. alleni Males Standard| Standard Error of | Devi- Mean ation .29 2.62 42 3.79 50 4.44 .07 61 04 3383 sllil 95 10 81 Table 4 P. e. alleni Females Standard| Standard Error of | Devi- Mean ation 41 2.08 .67 3.40 .69 Bia 10 48 .06 .28 SU? .89 14 aif Mean + Standard Deviation 77.85-83.08 87.69-95.17 15.26-24.15 14.44-15.65 8.15-— 8.81 25.84-27.73 18.07-19.70 Mean + Standard Deviation 74.42-78.58 82.03-88.82 12.06-19.09 14.23-15.19 8.19— 8.74 25.00-26.77 17.85-19.28 Observed Range 73.0— 89.0 78.0-102.2 6.1— 27.5 13.9— 16.1 GG= Os 24.7— 29.1 17.0— 21.0 Observed Range 73.0-81.5 79.3-92.1 6.0—20.2 13.1-15.5 8:0=7 9:0 24.2-28.0 17.0—20.0 294 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Comments. P. e. alleni is the best defined of the four races of P. erythrophthalmus. The uniformity of iris color, small amount of white on the outermost rectrix and the short wing make this race easily recognized. Its habitat preference during the breeding season varies with local- ity. Along the coast the towhee is the commonest bird of the dune association of Quercus geminata, Q. myrtifolia, Pinus clausa, Serenoa repens and Ceratiola ericoides. This same scrub habitat, inland, is also heavily populated with P. e. allenz. In areas where scrub does not occur the birds are found in second-growth associations which approxi- mate the natural scrub habitat in physical aspect. Often large num- bers are found in close proximity to cities and towns, where cultivation has created similar conditions. During the winter alleni does not seem to be so specific in its habitat demands, and is found in a great variety of situations. Individuals have been taken in urban areas, pine flatwoods, mesophytic ham- mocks, tropical hammocks and mangrove bays. During this season flocks of fifteen or twenty birds are not uncommon and on some occa- sions these flocks may contain resident and migrant individuals. Specimens Examined Atypical specimens are indicated as follows: *alleni > rileyi Breeding Males. FLORIDA: Alachua Co., 1 (AMNH), 1 (BDUF), 2 (PB), 2 (JCD); Bay Co., 1* (UMMZ); Brevard Co., 3 (AMNH), 3 (CNHM), 1* (CNHM), 1 (USNM), 2 (MCZ), 1 (PB), 1 (JCD); Broward Co., 2 (PB), 1 (JCD); Charlotte Co., 2 (CM); Collier Co., 1 (USNM); Dade Co., 1 (JCD), 1 (USNM), 1 (UMMZ); Franklin Co., 1 (LSU), 2 (JCD), 1* (JCD); Highlands Co., 1 (USNM); Hillsborough Co., 4 (JCD); Lee Co., 3 (SCD); Levy Co., 3 (JCD); Martin Co., 1 (JCD); Palm Beach Co., 1 (CM), 1 (USNM); Pasco Co., 2 (USNM); Pinellas Co., 4 (MCZ); 1 (UMMZ), 2 (AMNH), 4 (USNM); Polk Co., 5 (JCD); Putnam Co., 4 (GN), 1 (USNM), 1* (JCD); Saint Lucie Co., 1 (AMNH); Sarasota Co., 1 (AMNH); Volusia Co., 6 (AMNH). Breeding Females. FLORIDA: Alachua Co., 1 (JCD); Benton, 1 (USNM); Brevard Co., 1 (AMNH); Broward Co., 4 (JCD); Char- lotte Co., 1 (CM); 1 (JCD); Hillsborough Co., 2 (JCD); Highlands Co., 1 (LSU); Levy Co., 2 (JCD); Nassau Co., 1 (USNM); Orange Co., 1 (CM); Pinellas Co., 1 (AMNH), 1 (MCZ); 4 (USNM); Polk Co., 1 (JCD); Putnam Co., 1 (GN); Volusia Co., 1 (AMNH), 1 (MCZ). DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 295 Wintering Males. FLORIDA: Alachua Co., 3 (AMNH), 1* (AMNH), 4 (JCD), 1(PB); Brevard Co., 1 (MMP), 1 (MCZ), 3 (CNHM), 1* (CM), 2 (USNM), 18 (AMNH); Charlotte Co., 1 (UMMZ), 9 (MCZ); Collier Co., 1 (MMP), 4 (USNM); Columbia o-,2) (LSU);) Hscambia, Co., 1* (LSU): Franklin Co.) 1) (USU): Hernando Co., 1 (JCD), 1 (PB); Highlands Co., 6 (USNM); Hills- borough Co., 1 (CM); Indian River Co., 1 (LSU); Lake Co., 1 (JCD); Lee Co., 2 (AMNH), 4 (CNHM); Leon Co., 1 (USNM); Levy Co., 2 (MCZ), 1* (MCZ), 1 (PB), 2 JCD), 6 (AMNH), 1* (AMNH); Nas- sau Co., 3 (CNHM), 1* (CNHM); Okaloosa Co., 1 (USNM); Okee- chobee Co., 3 (USNM); Orange Co., 3 (GMS); Osceola Co., 3 (USNM), 1* (USNM); Palm Beach Co., 5 (CNHM); Pinellas Co., 4 (AMNH), 4 (MCZ), 1* (MCZ), 2 (CNHM), 1 (FSM), 7 (UMMZ); Polk Co., 1 (USNM), 2 (UMMZ); Putnam Co., 3 (CNHM), 1* (CNHM); Santa Rosa Co., 2 (CNHM), 2* (CNHM); St. Lucie Co., 1 (USNM); Volusia Co., 2 (AMNH), 1 (MCZ). GEORGIA: Camden Co., 1* (AMNH), 1* (MCZ), 1 (USNM); Chatham Co., 1* (LSU); Liberty Co., 1* (DZUG), 1* (USNM). Wintering Females. FLORIDA: Brevard Co., 1 (AMNH), 3 (CNHM), 1* (CNHM); Collier Co., 2 (USNM); Dade Co., 1 (UMMZ); Desoto Co., 2 (USNM); Duval Co., 1* (MCZ); Franklin Co., 1 (MMP), 1 (LSU); Highlands Co., 1 (USNM); Indian River Co., 3 (MCZ); Lee Co., 1 (CNHM), 6 (MCZ); Nassau Co., 1 (CNHM); Orange Co., 1 (GMS); Osceola Ce., 1 (USNM); Okaloosa Co., 1 (USNM); Palm Beach Co., 1 (USNM), 2 (CNHM); Pinellas Co., 1 (AMNH), 1 (CNHM), 4 (UMMZ), 1* (UMMZ), 3 (MCZ); Putnam Co., 3 (CNHM); Santa Rosa Co., 2 (CNHM), 4* (CNHM); Volusia Co., 2 (MCZ), 1 (AMNH); Kissimmee Prairie, 1 (USNM), 1 (MCZ). GEORGIA: Camden Co., 1* (MCZ); Chatham Co., 1* (CHAM); Levy Co., 1* (MCZ). Migrant Females. FLORIDA: Brevard Co., 1 (CNHM); Lee Co., 1 (MCZ); Orange Co., 1 (MMP); Palm Beach Co., 2 (CNHM); Pinellas Co., 1 (MCZ), 1 (UMMZ). PIPILO ERYTHROPHTHALMUS CANASTER Howell Pipilo erythrophthalmus var. erythrophthalmus, Baird, Brewer, and Ridgway (1874: 108), part. Pipilo erythrophthalmus canaster Howell (1913: 202), original description. Type locality: Spring Hill, Alabama. Pipilo erythrophthalmus leptoleucus Oberholser (1938: 641), original description. Type locality: New Orleans, Louisiana. 296 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Description Diagnosis. A large, large-billed, pale race, with variable eye color, showing a medium amount of white on the rectrices. Average dimensions of males. Wing, 87.30 + .21 (o, 2.50); tail, 94.95 = .30 (c, 3.53); culmen, 15.44 = .05 (c, 0.60); width of lower mandible, 8.94 = .03 (¢, 0.37); tarsus, 20.19 = .08 (c, 1.00); middle toe without claw, 20.19 = .07 (c, 0.88); length of white on inner web of outermost rectrix, 31.79 = .40 (¢, 4.71); tail spot/tail ratio, 35.59 + .45 (¢, 5.17). (See Table 5.) Average dimensions of females. Wing, 83.45 = 42 (, 2.31); tail, 90.43 = .02 (c, 3.37); culmen, 15.20 = .10 (c, 0.55); width of lower mandible, 8.79 =.07 (c, 0.36); tarsus, 27.91 = .20 (c, 1.03); middle toe without claw, 19.66 = .16 (c, 0.85); length of white on inner web of outermost rectrix, 26.98 = .70 (c, 3.64); tail spot/tail ratio, 29.41 + .74 (c, 3.87). (See Table 6.) Average color of males. Back, Olivaceous Black (3); flanks, 13-D-11; breast, Olivaceous Black (3); top of head, Olivaceous Black (3). Table 5 P. e. canaster Males | Standard) Standard| Mean = Measurement N | Mean | Error of | Devi- Standard Observed Mean ation Deviation Range Wing 145 | 87.30 220 2.50 | 84.81-89.80 | 78.0— 93.0 Tail 136 | 94.95 .30 3.53 | 91.42-98.48 | 86.0-105.0 Tail Spot | 136 | 31.79 40 4.71 27.08-36.51 | 18.0— 43.6 Culmen | 142 | 15.44 05 50 | 14.85-16.04 | 14.5- 17.0 Mandible Width} 141 8.94 .03 Ry | 8.57— 9.30 | 8.2— 9.7 Tarsus 145 | 28.19 08 1.00 | 27.19-29.19 | 25.2— 31.0 Middle Toe 145 | 20.19 07 88 | 19.31-21.07 | 18.0— 22.5 Tail Spot/Tail 130 | 35.59 A5 5.17 | 27.42-37.76 | 19.0— 45.0 Average color of females. Back, 15-E-7; flanks, 13-K-9; breast, 15-C-12; top of head, 7-A-12. Tris color. Table 21.) Breeding range. Variable, red to pale orange, occasionally yellow. (See From eastern Louisiana (Iberia, Pointe Coupee, and West Carroll Parishes), and western Mississippi (Bolivar County), DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 297 Table 6 P. e. canaster Females Standard) Standard) Mean + Measurement N | Mean | Error of | Devi- Standard Observed Mean ation Deviation Range Wing 31 | 83.45 FAD 2.31 81.14-85.76 | 79.5-87.0 Tail 27 | 90.48 .62 3.37 | 87.06-93.80 | 83.5-96.0 Tail Spot 25 | 26.98 .70 3.64 | 23.35-30.62 | 18.8-35.6 Culmen DG || WES PX) 10 05 14.67-15.75 | 14.0-16.5 Mandible Width} 28 8.79 07 36 8.43-— 9.16 | 8.2- 9.4 Tarsus Dae NN PaCfeke LE .20 1.03 26.88-28.94 | 25.6-29.5 Middle Toe 27 | 19.66 16 85 18.82-20.51 | 17.6-21.0 Tail Spot/Tail 27 | 29.41 74 3.87 | 25.54-33.27 | 22.0-39.0 north to southern Tennessee (Shelby and Wayne Counties), northern Alabama (Colbert, Limestone, Shelby, and Calhoun Counties), and central Georgia (Chattooga, Floyd, Paulding, Cobb, Gwinnett, Jack- son, Clarke, and Oglethorpe Counties to Taliaferro, Hancock, Put- man, Jasper, Lamar, and Merriwether Counties), south to middle eastern Alabama (Russell and Montgomery Counties), thence south to the Gulf Coast, in western Florida (Okaloosa and Santa Rosa Coun- ties), southern Alabama, Mississippi, and southeastern Louisiana (St. Bernard, Assumption, and Iberia Parishes). Also in central-southern North Carolina (Richmond County). (See Map 2.) Winter range. From southeastern Louisiana, central Mississippi (Warren County), central Alabama (Jefferson County), and northern Georgia (Rabun County), to coastal South Carolina (Beaufort, Charleston, and Georgetown Counties) and eastern North Carolina (Pitt County), south to southern Georgia (Toombs, Lowndes, and Brooks Counties), to western Florida (Wakulla and Leon Counties), and thence westward along the Gulf to southeastern Louisiana. (See Map 3.) Comments. The habitat preference of this form as noted near the type locality is apparently slightly different from that of P. e. allent or P. e. rileyi. Exploration of what appeared to me to be suitable habitats, on the basis of my own experience in Florida, produced not a single towhee. Specimens taken near Mobile, Alabama, came from extremely dense hammock-like habitats. This form is not completely sedentary in its habits, but on the other hand it does not move as far south in winter as do erythrophthalmus and riley. 298 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Specimens Examined Atypical specimens are indicated as follows: * canaster > rileyi ** canaster > erythrophthalmus Breeding males. ALABAMA: Baldwin Co., 1 (UMMZ); Calhoun Co., 2 (USNM); Colbert Co., 3 (USNM); Limestone Co., 3 (USNM); Mobile Co., 4 (USNM), 2 (JCD), 1 (LSU); Russell Co., 1 (USNM); Shelby Co.,,\2 (USNM). FLORIDA: Okaloosa Co., 3* (CM), 1 (JCD); Santa Rosa Co., 1 (JCD), 1 (USNM). GEORGIA: Carroll Co., 1 (USNM); Chatooga Co., 2 (USNM); Clarke Co., 2 (LSU), 1 (MMP), 9 (USNM), 1** (USNM), 2* (USNM); Clayton Co., 2 (USNM); Cobb Co., 1 (USNM), 1* (USNM); DeKalb Co., 2 (LSU), 1* (MMP), 4 (USNM); Douglas Co., 1 (USNM); Floyd Co., 2 (USNM); Fulton Co., 2 (MMP), 2 (USNM), 1** (USNM); Gwinnett Co., 2 (USNM); Hancock Co., 1 (USNM); Haralson Co., 1 (USNM); Henry Co., 1 (USNM), 1* (USNM); Jackson Co., 3 (USNM); Jasper Co., 1 (USNM); Lamar Co., 1* (USNM); Meriweather Co., 1 (USNM); Morgan Co., 1 (USNM); Oglethorpe Co., 1 (USNM); Paulding Co., 2 (USNM); Putnam Co., 1 (USNM); Rockdale Co., 1 (USNM); Taliaferro Co., 2 (USNM); Walton Co., 1 (USNM.) LOUI- SIANA: Assumption Parish, 1 (LSU), East Baton Rouge Parish, 2 (LSU), 5 (USNM); Iberia Parish, 1 (MMP); Orleans Parish, 7 (USNM), 6 (LSU), 1 (MMP); Pointe Coupee Parish, 1 (GMS); Saint Bernard Parish, 2 (USNM); West Baton Rouge Parish, 1 (CHAM); West Carroll Parish, 1 (LSU); West Feliciana Parish, 1 (USNM). MISSISSIPPI: Adams Co., 2 (USNM); Bolivar Co., 1 (USNM), 1 (CNHM); Hancock Co., 1 (USNM); Harrison Co., 4 (LSU), 11 (USNM); Warren Co., 1 (USNM). TENNESSEE: Shelby Co., 3 (LSU); Wayne Co., 3 (USNM). Breeding Females. ALABAMA: Calhoun Co., 1 (USNM); Mobile Co., 1 (USNM), 1 (JCD); Russell Co., 1 (USNM). FLORIDA: Okaloosa Co., 1 (CM). GEORGIA: Clarke Co., 2 (MMP), 1 (USNM); Cobb Co., 1 (CNHM), 2 (USNM); DeKalb Co., 2 (USNM); Fayette Co., 1 (USNM); Fulton Co., 1 (MMP), 2 (USNM); Hancock Co., 1 (USNM); Oglethorpe Co., 1 (USNM); Taliaferro Co., 1 (USNM). LOUISIANA: East Baton Rouge Parish, 1 (USNM); Orleans Parish, 1 (USNM); Pointe Coupee Parish, 1 (GMS). MISSISSIPPI: Bolivar Co., 1 (CNHM); Harrison Co., 3 (USNM); Pearl River Co., 1 (AMNH). NORTH CAROLINA: Richmond Co., 1 (USNM). Wintering Males. ALABAMA: Baldwin Co., 1 (USNM); Jefferson Co., 1 (LSU); Lee Co., 1 (USNM); Mobile Co., 1 (LSU). FLORIDA: DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 299 Leon Co., 1 (CNHM); Okaloosa Co., 3 (CM); Santa Rosa Co., 2 (CNHM);Wakulla Co., 1 (USNM). GEORGIA: Barrow Co., 1 (USNM); Brooks Co., 1 (USNM); Clarke Co., 6 (USNM); DeKalb Co., 2 (USNM), 1 (DZUG); Fulton Co., 2 (USNM); Lowndes Co., 1 (USNM); Madison Co., 2 (USNM); Oconee Co., 2 (USNM); Walton Co., 1 (USNM). LOUISIANA: East Baton Rouge Parish, 3 (LSU), 1 (USNM); Orleans Parish, 3 (LSU), 2 (CNHM). MISSISSIPPI: Amite Co., 1 (USNM); Forrest Co., 1 (LSU); Harrison Co., 1 (LSU), 6 (USNM); Rankin Co., 1 (MMP); Warren Co., 1 (LSU). NORTH CAROLINA: Pitt Co., 1 (USNM).. SOUTH CAROLINA: Beau- fort Co., 1 (CHAM), 1 (MCZ). Wintering Females. ALABAMA: Houston Co., 4 (USNM); Orange Beach, 1 (USNM). FLORIDA: Escambia Co., 1 (USNM); Leon Co., 1* (CNHM); Okaloosa Co., 1 (CM); Santa Rosa Co., 1 (CNHM); 1* (CNHM); Wakulla Co., 1 (FSM), 1* (USNM). GEORGIA: Cobb Co., 1 (CNHM), 1* (CNHM); DeKalb Co., 1 (USNM); Rabun Co., 1 (USNM). LOUISIANA: East Baton Rouge Parish, 1 (CNHM), 1 (LSU), 1 (USNM); Orleans Parish, 1 (LSU), 2 (USNM); Chef Menteur, 1 (CNHM). MISSISSIPPI: Harrison Co., 4 (LSU), 1 (USNM); Rankin Co., 1 (MMP). SOUTH CAROLINA: Charleston Co., 1 (CHAM); Georgetown Co., 1* (USNM). Migrant Males. ALABAMA: Baldwin Co., 2* (USNM); Mobile Co., 1* (USNM); FLORIDA: Escambia Co., 1 (USNM); Okaloosa Co., 1 (CM). GEORGIA: Clarke Co., 2 (USNM); Cobb Co., 2 (CNHM); DeKalb Co., 1 (USNM); Harris Co., 1 (USNM); Jefferson Co., 1 (USNM); Toombes Co., 1 (USNM); Treutlen Co., 1 (USNM); Washington Co., 1 (USNM). LOUISIANA: East Baton Rouge Parish, 1 (LSU), 2 (USNM); Orleans Parish, 2 (LSU); West Feliciana anishy (GIVES). MULSSISSIPE DT: sHarrison) Corl (USN M25 (USNM), 1 (MMP); Pearl River Co., 1 (AMNH); Warren Co., 2 (CNHM). NORTH CAROLINA: Carteret Co., 1* (USNM). SOUTH CAROLINA: Georgetown Co., 1 (USNM). TENNESSEE: Roane Co., 1 (USNM). Migrant Females. GEORGIA: Clarke Co., 1 (MMP); Harris Co., 1 (USNM). LOUISIANA: East Baton Rouge Parish, 1 (USNM); Orleans Parish, 1 (LSU). MISSISSIPPI: Harrison Co., 2 (USNM). PIPILO ERYTHROPHTHALMUS RILEY] Koelz Pipilo leucopis Maynard (1878: 113, pl. IV), part. Pipilo erythrophthalmus var. alleni, Baird, Brewer, and Ridgway (1874: 112), part. Pipilo alleni Koelz (1939: 121), original description. Type locality: Brunswick, Georgia. 300 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Description Diagnosis. A medium-sized, large-billed race, with variable eye color, and showing less white on the rectrices than its northern rela- tives. Average dimensions of males. Wing, 85.53 = .26 (¢, 2.62); tail, 94.59 + .36 (c, 3.43); culmen, 15.64 + .62 (c, 0.62); width of lower mandi- ble, 8.85 = .03 (c, 0.31); tarsus, 28.25 + .10 (c, 0.99); middle toe without claw, 19.89 = .07 (¢, 0.68); length of white on inner web of outermost rectrix, 25.50 + .41 (c, 3.72). (See Table 7.) Average dimensions of females. Wing, 80.32 = .46 (c, 2.81); tail 89.70 = .67 (c, 3.67); exposed culmen, 15.34 = .09 (c, 0.53); width of lower mandible, 8.69 + .05 (c, 0.27); tarsus, 27.54 = .19 (oc, 1.15); middle toe without claw, 19.38 = .11 (c, 0.67); length of white on inner web of outermost rectrix, 22.85 + .61 (c, 2.93). (See Table 8.) Average color of males. Back, Olivaceous Black; flanks, 13-D-11; breast, Olivaceous Black; top of head, Olivaceous Black. Average color of females. Back, 15-C-11; flanks, 13-K-11; breast, 15-C-9; top of head, 8-H-11. Table 7 P. e. rileyr Males Standard) Standard) Mean = Measurement N | Mean | Error of | Devi- Standard Observed Mean ation Deviation Range Wing 100 | 85.54 .26 2.62 | 82.91-88.15 | 30.0— 92.0 Tail 91 | 94.59 36 3.43 | 91.16-93.03 | 87.0-103.0 Tail Spot 84 | 25.50 Al 3.01 21.78-29.22 | 17.0- 34.0 Culmen 98 | 15.64 .06 .62 | 15.01-16.26 | 14.1- 17.0 Mandible Width} 96 8.85 .03 so 8.54— 9.16 | 8.0- 9.5 Tarsus 100 | 28.25 .10 99 | 27.26-29.24 | 25.3- 30.6 Middle Toe 100 | 19.89 07 .68 | 19.21-20.57 | 17.0- 21.5 Tris color. Variable, straw to orange or red. (See Table 21). Breeding range. From western Florida (Walton County), south- eastern Alabama (Houston County), northeast through southeastern Georgia (Early, Dougherty, Crisp, Jones, Warren, McDuffie and Rich- mond Counties) to coastal South Carolina (Jasper, Beaufort, Charles- ton, and Georgetown Counties) and coastal North Carolina (Bruns- wick, New Hanover and Carteret Counties) south along the coast of DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 301 Table 8 P.e. rileyr Females Standard| Standard) Mean + Measurement N | Mean | Error of | Devi- Standard Observed Mean ation Deviation Range Wing 388 | 80.32 .46 2.81 77.51-83.13 | 72.5-85.0 Tail 30 | 89.70 .67 3.66 86.04-93.36 | 80.9-97.0 Tail Spot 23 | 22.84 61 2.93 19.92-25.78 | 19.9-30.3 Culmen 388 | 15.34 .09 53 14.81-15.87 | 14.5-16.5 Mandible Width| 37 8.69 -05 Pall 8.42— 8.96 | 8.2- 9.4 Tarsus 36 | 27.54 19 Es 26.39-28.69 | 25.4-29.6 Middle Toe 88 | 19.38 alli .67 18.71-20.06 | 17.5-20.5 Georgia (Chatham, Glynn and Camden Counties) thence west through southern Georgia (Charlton County) and northern Florida (Madison and Leon Counties) to the Gulf Coast (Wakulla County). (See Map 2.) Winter range. From western Florida (Escambia County), eastern Alabama (Lee County) north to north-central Georgia (Cobb, Walton and Clarke Counties) to coastal South Carolina (Beaufort, Jasper, Dorchester, Charleston and Georgetown Counties) and eastern North Carolina (Robeson, Carteret, Pitt and Hyde Counties) south to mid- peninsular Florida (Brevard, Charlotte and Pinellas Counties). (See Map 3.) Comments. This pale-eyed race is clearly distinguished from P. e. allent by its larger wing and greater amount of white on the rectrices. It can be distinguished from P. e. erythrophthalmus and P. e. canaster by the color of its irides and by the intermediate amount of white present on the rectrices. The color of the flanks in rileyi, especially in the males, is remarkably uniform in the material examined and is much darker than that found in P. e. allenz. My experience with this race at the type locality during the breeding season agrees with that of Walter Koelz, the original describer. The preferred habitat is in scrubby, moderately dense growth. Specimens were taken on St. Simons Island, and on the mainland a few miles to the north and south of Brunswick, Georgia, in much the same type of habitat that is preferred by P. e. alleni in Florida. Material from west Florida, in Walton County, was also taken in this same scrub asso- ciation. 302 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Specimens Examined Atypical specimens are indicated as follows: * rileyi > canaster ** rileyt > erythrophthalmus *EX riley > alleni Breeding Males. ALABAMA: Houston Co., 1 (USNM). FLORIDA: Leon Co., 4 (USNM); Madison Co., 1 (USNM); Walton Co.,6 (JCD), 1* (JCD), 1 (AMNH). SOUTH CAROLINA: Beaufort Co., 1 (MCZ), 1 (AMNH), 1 (USNM); Charleston Co., 1 (CHAM), 1** (CHAM), 3 (MCZ), 1* (MCZ), 2 (USNM); Georgetown Co., 1 (USNM), 1* (USNM); Jasper Co., 1 (USNM). NORTH CARO- LINA: Brunswick Co., 2** (NCS); Carteret Co., 1 (NCS); New Han- over Co., 1 (USNM), 1* (USNM). GEORGIA: Baker Co., 3 (DZUG), 2 (USNM); Ben Hill Co., 1 (USNM); Bibb Co., 1 (USNM); Bullock Co., 1 (USNM); Burke Co., 1 (USNM), 1* (USNM); Camden Co., 1 (DZUG), 1 (MCZ); Candler Co., 1 (USNM); Charlton Co., 1 (DZUG), 1 (USNM); Chatham Co., 1 (CHAM), 3 (USNM); Coffee Co., 1 (USNM); Colquitt Co., 1 (USNM); Cook Co., 1 (USNM); Crisp Co., 2 (USNM); Decatur Co., 1 (DZUG); Dodge Co., 1 (USNM); Dougherty Co., 1* (USNM), 1 (DZUG); Early Co., 1 (DZUG); Effingham Co., 1 (USNM); Glynn Co., 7 (JCD), 1 (PB), 1 (DZUG); Grady Co., 1*** (USNM); Irwin Co., 1 (USNM); Jones Co., 1 (LSU), 2 (USNM); Lowndes Co., 1 (USNM); Macduffie Co., 1 (MMP); Pierce Co., 1 (USNM); Pulaski Co., 1 (USNM); Richmond Co., 1 (MMP), 3 (USNM); Thomas Co., 2 (USNM), 1* (USNM); Turner Co., 1 (USNM); Ware Co., 2 (USNM); Warren Co., 1* (USNM). Breeding Females. FLORIDA: Franklin Co., 1 (USNM); Leon Co., 1* (USNM); Walton Co., 5 (JCD), 1* (USNM). GEORGIA: Bacon Co., 1 (USNM); Baker Co., 1 (DZUG), 1 (USNM); Ben Hill Co., 1 (USNM); Brantly Co., 1 (DZUG); Candler Co., 1 (USNM); Chatham Co., 1 (USNM); Colquitt Co., 1 (USNM); Cook Co., 1 (USNM); Crisp Co., 1 (USNM); Decatur Co., 1 (DZUG); Dodge Co., 1 (USNM) Dougherty Co., 2 (DZUG); Glynn Co., 3 (JCD), 1 (PB); Grady Co., 1 (USNM); McDuffie Co., 1 (USNM); Sumter Co., 1 (MMP); Ware Co., 1 (USNM). NORTH CAROLINA; Brunswick Co., 1 (USNM); Carteret Co., 1 (USNM). SOUTH CAROLINA: Charleston Co., 1 (MCZ), 1** (MCZ), 2 (USNM); Georgetown Co., 1 (USNM). Wintering Males. ALABAMA: Lee Co., 2 (UMMZ). FLORIDA: Alachua Co., 1 (AMNH); Baker Co., 1 (USNM); Brevard Co., 2 (CNHM); Charlotte Co., 1 (MCZ); Columbia Co., 1 (LSU); Duval Co., 1 (USNM); Escambia Co., 1 (LSU), 1 (USNM); Franklin Co., DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 303 2 (LSU), 1 (USNM); Hamilton Co., 1 (USNM); Levy Co., 4 (AMNH); Nassau Co., 2 (AMNH); Okaloosa Co., 9 (CM); Pinellas Co., 1 (AMNH); Putnam Co., 1 (CNHM); Santa Rosa Co., 12 (CNHM); Walton Co., 1 (USNM). GEORGIA: Berrien Co., 1 (USNM); Cal- houn Co., 1 (USNM); Camden Co., 1 (DZUG), 2 (MCZ), 5 (AMNH); Charlton Co., 3 (USNM); Chatham Co., 2 (USNM), 4 (CHAM); Clarke Co., 1 (USNM); Echols Co., 1 (DZUG), 1 (USNM); Glynn Co., 1 (LSU), 1 (MMP), 1 (UMMZ); Liberty Co., 1 (USNM); Mc- Intosh Co., 2 (AMNH); Tatnall Co., 1 (USNM); Walton Co., 1 (USNM). NORTH CAROLINA: Carteret Co., 3 (USNM); Pitt Co., 2 (USNM). SOUTH CAROLINA: Beaufort Co., 5 (MCZ), 2 (USNM), 1 (AMNH); Charleston Co., 6 (CHAM), 2 (USNM); Georgetown Co., 1 (USNM). Wintering Females. FLORIDA: Alachua Co., 1 (CNHM), 1 (USNM); Brevard Co., 1 (MCZ); Escambia Co., 1 (LSU), 1 (USNM); Levy Co., 1 (AMNH); Nassau Co., 1 (CNHM); Okaloosa Co., 7 (CM); Columbia Co., 1 (LSU); Pinellas Co., 1 (CNHM); Santa Rosa Co., 9 (CNHM); Volusia Co., 1 (MCZ); Wakulla Co., 1 (FSM), 1 (USNM); Walton Co., 1 (USNM). GEORGIA: Berrien Co., 1 (USNM); Brooks Co., 1 (USNM); Camden Co., 1 (MCZ); Chatham Co., 2 (CHAM); Clarke Co., 2 (USNM); Echols Co., 1 (DZUG); Glynn Co., 1 (LSU), 1 (MMP), 1 (UMMZ); Hinesville, 1 (USNM). NORTH CAROLINA: Carteret Co., 2 (USNM); Hyde Co., 1 (USNM); Robeson Co., 1 (CNHM). SOUTH CAROLINA: Charleston Co., 3 (CHAM); Dorchester Co., 1 (CHAM); Georgetown Co., 1 (USNM); Jasper Co., 2 (USNM). Migrant Males. FLORIDA: Levy Co., 1 (AMNH); Okaloosa Co., 1 (CM); Santa Rosa Co., 1 (USNM), 2 (CNHM). GEORGIA: Bryan Co., 1 (USNM); Camden Co., 1 (USNM); Chatham Co., 1 (MCZ); Cobb Co., 1 (CNHM); Cook Co., 1 (MMP); Decatur Co., 1 (DZUG); Evans Co., 1 (USNM); Liberty Co., 2 (USNM); Long Co., 1 (USNM); McIntosh Co., 1 (UMMZ); Tatnall Co., 1 (DZUG), 1 (USNM); Telfair Co., 1 (USNM); Thomas Co., 1 (AMNH); Washington Co., 1 (USNM). NORTH CAROLINA: Pitt Co., 1* (USNM). SOUTH CAROLINA: Charleston Co., 1 (CNHM), 3 (MCZ); Georgetown Co., 1 (USNM). Migrant Females. FLORIDA: Escambia Co., 2 (USNM); Franklin Co., 1 (JCD); Okaloosa Co., 1 (CM); Santa Rosa Co., 2 (CNHM), 1 (USNM). GEORGIA: Camden Co., 2 (MCZ), 1 (USNM); Cook Co., 1 (MMP); Fulton Co., 1 (USNM); Tatnall Co., 1 (USNM); Treutlen Co., 1 (USNM). SOUTH CAROLINA: Beaufort Co., 1 (MCZ), Charleston Co., 1 (MCZ), 2 (USNM); Georgetown Co., 1* (USNM); Greenville Co., 1* (USNM). 304 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY GEOGRAPHIC VARIATION Following preliminary examination, the material at hand was sepa- rated into six samples (Map 1). Mean, standard error of the mean, and standard deviation of the mean were calculated for each of the mensural characters investigated. These samples were treated sepa- rately for each sex, except for samples 1, 2 and 3, of which only the males were considered (Tables 9, 10 and 11). For each sex samples 1, 2 and 3 were later treated as a single sample, representing the northern population, P. e. erythrophthalmus. For greater clarity, these three northern samples are referred to in the following discussion as eastern, central and western, respectively. Sample 4 represents P. e. canaster, 5 represents P. e. riley, and 6 represents P. e. alleni. Means were regarded as being significantly different when two standard errors on either side of the means did not overlap in the samples under consideration. A character was assumed to be of diag- nostic value if it furnished more than 75 per cent correct separation of mixed samples. The degrees of separation furnished by the various characters are presented in Tables 12-19. In these tables, the distance in standard units to the point of intersection of the standard distributions is repre- sented by d/c. Per cent separation is the measure of area of the standard curves which lie to either side of the point of intersection of these curves. Division point is the theoretical point of maximum sepa- ration in millimeters for each of the characters indicated. X indicates that the difference is not statistically significant, and Z indicates that the degree of separation accomplished by the character is less than 50.0 per cent. Figures 1-15 present this same information in the form of Hubbs- Perlmutter diagrams. For a discussion of the use of this method of demonstrating differences between populations see Hubbs and Perl- mutter (1942). In each diagram a vertical line marks the mean; a rectangle to either side indicates one standard deviation; the black part of each rectangle indicates twice the standard error of the mean; a solid line shows the observed range of variation; a broken line shows the limits of three standard deviations to either side of the mean, beyond observed range. Wing Length Males (Figure 1). Among the four races here recognized, erythroph- thalmus has the longest wing, 87.57 = .15 mm. This form represents the northern part of the range of the species and extends from the eastern seaboard to the junction with Pipilo arcticus maculatus on the DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 305 west. It shows a gradual increase in wing length from east to west. The eastern sample shows a mean wing length of 86.63 = .16 mm., the central sample 87.32 + .28 mm., and the western sample 89.77 + .28mm. The difference in the means of the eastern and central samples is not statistically significant. ERYTHROPHTHALMUS N=145 | : CANASTER 4 ALLEN! 73 74 «#75 8 90 95 f°) 85 LENGTH IN MILLIMETERS Figure 1. Variation in wing length of males. See page 304 for explanation of figure. The western sample is significantly different from both the eastern and central samples. Ridgway (1901: 424) comments on variation within this race. He states that birds “from opposite sides of the Alleghenies differ but slightly in average measurements.” He gives as substantiating evidence, measurements of 9 adult males from “east of the Alleghenies” — wing 89.92, “8 adult males from Mississipp1 valley’ — wing 89.15. In this connection it is pertinent to note that the present study shows considerable difference in samples taken from the extreme eastern and western portions of the range of this race — enough to allow 75 per cent of the western birds to be distinguished from the eastern birds. This character, however, is of no diagnostic value in separating the central population from either the eastern or western samples. The wing of canaster is not significantly shorter than that of ery- throphthalmus. The mean length for the wing of canaster is 87.30 + 21 mm. Rileyi is significantly shorter in wing length than either canaster or erythrophthalmus. The mean length for the wing is 85.53 = .26 mm. The difference, however, is not of such magnitude as to allow its use as a diagnostic character. 306 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY The wing of allent is much shorter, averaging 80.47 = .29 mm. Wing length is of value diagnostically, furnishing the following per- centages of separation! from the various races: from rileyi, 83.5 per cent; from erythrophthalmus, 91 per cent; and from canaster, 91.5 per cent. The general trend in the variation of wing length is from northwest to southeast, with maximum length occurring in the northwest. The most abrupt change occurs at the junction of rileyz and allent. 5 ERYTHROPHTHALMUS a <----- 4 CANASTER 70 75 80 85 90 LENGTH IN MILLIMETERS Figure 2. Variation in wing length of females. See page 304 for explanation of figure. Females (Figure 2). The same pattern of variation observed in the males appears in the females. Erythrophthalmus and canaster have the longest wings, averaging 83.38 = .28 and 83.45 + .42 mm., re- spectively. The means are not significantly different. The wing in rileyt is shorter, 80.32 + .46 mm., and furnishes 74.8 per cent sepa- ration from canaster, and 65.7 per cent from erythrophthalmus. Alleni has the shortest wing, 76.50 = .47 mm. Separation of 91 per cent from erythrophthalmus, 94.2 per cent from canaster and 78.3 per cent from riley2 is obtained by using this character. 1In comparing the degree of separation of two samples the ‘‘per cent separation’’ was deter- mined by averaging the separate degrees of separation afforded. Thus, if 76 per cent of Form A was separable from 74 per cent of Form B, the ‘‘per cent separation’”’ was 75 per cent. DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 307 Measurement Wing Tail Tail Spot Culmen Mandible Width Tarsus Middle Toe Measurement Wing Tail Tail Spot Culmen Mandible Width Tarsus Middle Toe Table 9 P. e. erythrophthalmus (Sample 1) Males Standard| Standard N | Mean | Error of | Devi- Mean ation 211 | 86.63 .16 2.37 205 | 92.04 4. 3.83 200 | 36.10 ROD, 4.58 206 | 14.51 .04 61 201 8.61 .02 sol 208 | 27.42 .06 .87 205 | 19.58 .05 502 Table 10 P. e. erythrophthalmus (Sample 2) Males Standard| Standard N | Mean | Error of | Devi- 63 63 62 62 63 58 63 Mean ation 87.32 .28 2.21 92.70 43 3.42 36.85 61 4.79 14.56 .06 45 8.61 .03 .22 27.05 .09 .76 19.64 .08 .63 Mean = Standard Deviation 84.26-89.00 88.21—-95.87 31.52-40.68 13.90-15.13 8.30— 8.92 26.56-28.29 18.86-20.30 Mean = Standard Deviation $5.12-89.53 89.28-96.11 32.07—41.64 14.11-15.00 8.39— 8.83 26.29-27.82 19.01—20.27 Observed Range 80.0— 94.0 83.0-103.1 24.0— 55.0 Observed Range 82.1— 93.1 85.9-103.0 25.1— 50.2 13.6— 15.5 8.1-— 9.2 24.8— 29.1 18.1— 21.1 308 Measurement Wing Tail Tail Spot Culmen Mandible Width Tarsus Middle Toe Table 11 P. e. erythrophthalmus (Sample 3) Males Standard) Standard N | Mean | Error of | Devi- Mean ation 85 89.77 .28 2.61 §3 96.78 .36 3.05 83 38.10 42 3.82 84 14.60 .06 4583) 82 8.72 .03 .29 84 27.50 .09 Sl 83 19.80 .08 ne, Table 12 Mean = Standard Deviation 93.73-99.83 34.2841 .92 14.06-15.13 8.43— 9.01 26.69-28.30 19.08—20.51 87.16-92.38 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Observed Range 84.9- 96.0 89.1-104.0 29.1— 50.0 13.0— 15.9 Sel Wop 25.2— 19.5 17.4— 22.4 Pipilo erythrophthalmus erythrophthalmus versus P. e. alleni, P. e. canaster and P. e. rileyi Males Vs. allent Vs. canaster Vs. rileyt Per Per Per Measure- cent Di- cent Di- cent Di- ment d/o | Sepa- | vision| d/a | Sepa-| vision| d/a | Sepa- | vision ration | Point ration | Point ration | Point Wing 1.33 | 91.0 | 83.8 x x xX Z Z Z Tail Z Z Z Z Z Z Z Z Z Tail Spot 2.09 | 97.1 | 29.0 5 | (68.0) | 3424 E62 O42 5s a3 O87 Culmen Z Z Z 81 | 79.0 | 15.0 96 | 88.0 | 15.0 Mandible Z Z Z Z Z Z Z Z Z Tarsus Z Z Z Z Z Z Z Z Z Toe ie | OO) |) UO RL I Z Z Z Z Z Table 13 Pipilo erythrophthalmus erythrophthalmus versus P. e. canaster, P. e. alleni and P. e. rileyi Females Vs. canaster Vs. alleni Vs. rileyt Per Per er, Measure- cent cent | Di- cent | Di- ment d/a | Sepa-| vision| d/a | Sepa- | vision| d/a | Sepa-| vision ration | Point ration | Point ration | Point Wing se ¢ D, 140 92.0 | 79.6 | .53 | 60.0 | 81.9 Tail x x Z Z Z x x xX Tail Spot “80° CRONO2SES A P21AS OSto 8240036 9210 o7e3 Culmen 65 74.5 Z Z Z 79. GIO) || ILS ,IL Mandible ag x x x x xX x x Tarsus Coe Ao Wezel eZ Z Z s Z Z Z Toe ge Z x x xX xX ».4 x Table 14 Pipilo erythrophthalmus canaster versus P. e. alleni, P. e. erythrophthalmus and P. e. rileyi Males Measure- ment Wing Tail Tail Spot Culmen Mandible Tarsus Toe Vs. alleni Vs. erythrophthalmus Vs. rileyi Per Per Per cent cent Di- cent Di- Sepa- | vision | d/o | Sepa-| vision| d/o | Sepa- | vision ration | Point ration | Point ration | Point 91.0 x xX xX Z Z Z 7, Z Z Z Bxe x x 91.0 | 25.6 Oy | uO | seb) wil | BRR | BOD Z sell | OL fo LO) OK xX x 75.0 Z Z Z x x x A | 2raD Z Z Z xX xe xX 78.0 Z Z Z Z Z Z SSS a Table 15 Pipilo erythrophthalmus canaster versus P. e. alleni, Females Measure- ment Wing Tail Tail Spot Culmen Mandible Tarsus P.e. rileyi and P. e. erythrophthalmus Vs. allent Vs. rileyt Vs. erythrophthalmus Per Per Per cent | Di- cent | Di- cent | Do- d/o | Sepa-| vision| d/o | Sepa-| vision| d/o | Sepa- | vision ration | Point ration | Point ration | Point Fie 94.0 | 79.9 “57. 75.0 “32.1 ea x x ie CHD | SY |) 2 x x = ca Rise" 158 94.0 | 21.2 “67” 75.0 | 24.6 | .76 | 78.0 | 28.8 ee Z Z x xX xX “60. 72.0 | 14.9 kn Z Z DE xe x eS xX aon 1.05 85.5 | 26.8 | X x Xx ca x oe 73 CUD USE | 2S x x naa Z Z Toe Table 16 Pipilo erythrophthalmus rileyt versus P. e. allena, P. e. erythrophthalmus and P. e. canaster Males Vs. allena Vs. erythrophthalmus Vs. canaster Per Per Per Measure- cent D- cent Di cent Di- ment d/o | Sepa-| vision| d/o | Sepa-| vision | d/o | Sepa-| vision ration | Point ration | Point ration | Point Wing 97 83.5 | 83.0 RB Z Z eee Z Z Tail oon Z Z ay Z Z et x x Tail Spot 73° GEO |e 23:0 Fear 92.0 | 30.72 ae 78.0 | 28.2 Culmen ae Z Z 86 80.5 | 15.0 ee x x Mandible 60 77.5 8.7 ea Z Z MSc | se a Tarsus ie) Wao | 27-5 a Z Z xX De | a Toe on 76.0 | 19.5 a Z Z Z Z Z Females Measure- ment Wing Tail Tail Spot Culmen Mandible Tarsus Toe Table 17 Pipilo erythrophthalmus rileyi versus P. e. alleni, P. e. canaster and P. e. erythrophthalmus Vs. alleni Vs. canaster Per Per | cent | Di- cent | Di- d/o | Sepa-| vision| d/o | Sepa- | vision ration | Point ration | Point au 79.5 |. 78.0 65 74.5 | 82.1 62 73.0 | 87.4 | xX x x 134 91.0 | 19.5 | .59 | 72.0 | 24.6 co 74.0 | 15.0 x x fra ceils ocala 93 80.0 | 26.6 | X oem x ae 67.5 19.0 at ce) x Table 18 Vs. erythrophthalmus Per cent Di- d/o | Sepa- | vision ration | Point A 7A |) Bil x xX xX LE 53m OAL OM ones An ide Ou| oul xX aXG x Z Z Z x xX x Pipilo erythrophthalmus alleni versus P. e. riley, P. e. erythrophthalmus and P. e. canaster Males Measure- ment Wing Tail Tail Spot Culmen Mandible Tarsus Toe Vs. riley Per cent Di- d/a | Sepa- | vision ration | Point Of | Bh |) SBMO) Z Z Z S| 7S || PRO Z Z Z BN) 7ALKo) 8.8 76 | 78.0 | 27.5 .64,/| 74.0 | 19.5 Vs. erythrophthalmus Per cent Di- d/a | Sepa- | vision ration | Point 1.31 | 91.0 | 83.8 Z Z Z 12M OMA || 2852 Z Z Z Z Z Z Z Z Z .46 | 67.5 | 19.3 Vs. canaster Per cent Di- d/a | Sepa- | vision ration | Point 1.35 | 92.0 | 85.0 Z Z Z 1.31 | 91.0 | 26.6 Z Z Z 64 | 74.0 8.7 |) CAO BB 18 | 78.5 | 19.5 Bl BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Table 19 Pipilo erythrophthalmus alleni versus P. e. rileyt, P. e. canaster and P. e. erythrophthalmus Females Vs. rileyt Vs. canaster Vs. erythrophthalmus Per Per Per Measure- cent | Di- cent | Di- cent | Di- ment d/o | Sepa-| vision| d/o | Sepa- | vision | d/o | Sepa- | vision ration | Point ration | Point ration | Point Wing AA 700) 78:0) 1.62) | 94.5 1) 79:9) | deol | 93" aan laa 9eo Tail EDO Oro, | Otek) | eMA WN GeOn |e Sis On || taeZ Z Z Tail Spot 1.26 | 87.5 | 19.5 | 1.60} 94.5 | 21.2 | 2.39 | 99.2 | 24.0 Culmen SOL 7a 00 1 LOO 4 OF Gta yf 429s eZ Z Z Mandible ean Z Z wae Z Z ca x x Tarsus a 80.0 | 26.6 1.06. 85.5 | 26.8 ae Z Z Toe 60. 77.5 | 19.0 .68 | 75.5 | 19.1 xX x x Tail Length Males (Figure 3). Canaster and riley: have the longest tails. Canaster averages 94.95 = .30 mm. and rileyz 94.59 = .86 mm. They are not significantly different in this respect. Both canaster and riley. differ TAIL mo CO 1 Ne 351 E— 4 N=136 CANASTER — "1 RILEY! ALLEN} “4 77 78 79 80 100 105 6 so LENGTH IN MILLIMETERS Figure 3. Variation in tail length of males. See page 304 for explanation of figure. DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 313 significantly from erythrophthalmus and alleni. Erythrophthalmus has a mean tail length of 92.91 = .20 and alleni 91.43 + .42 mm. The trend of geographic variation in wing length within P. e. erythrophthalmus is repeated in tail length. The three samples from east to west average respectively, 92.04 = .27, 92.70 + .44 and 96.78 + .36 mm. The western sample is again significantly different from the other two. Ridgway’s measurements show the reverse in so far as general trend — nine eastern males 94.74, eight western males 93.73 mm. Tail length, while of no value as a diagnostic character, shows an interesting pattern of geographic variation: maximum length in the far northwest, next largest in the coastal plains and piedmont areas, smaller again in the north-central and eastern areas, and smallest in peninsular Florida. Females (Figure 4). Canaster and rileyi have the longest tails, the former averaging 90.43 = .62 mm., and the latter 89.70 = .67 mm. The difference is not statistically significant. The tail in erythroph- thalmus is not significantly shorter than these, averaging 88.22 + .41 mm. Alleni has a much shorter tail, 85.43 = .67 mm., and this charac- ERYTHROPHTHALMUS : eens tremens erer ——| N=27 pone - oe owe. N=26 ee 7S 60 6S 90 35 100 LENGTH IN MILLIMETERS Figure 4. Variation in tail length of females. See page 304 for explanation of figure. ter serves to separate the Florida race from rileyi and canaster, with 74.3 per cent and 77 per cent accuracy, respectively. The mean tail length in alleni, though significantly different from that of erythroph- thalmus, furnishes less than 50 per cent separation from this form. Tail Spot Males (Figure 5). Erythrophthalmus shows the greatest linear extent of white on the outermost rectrix, 36.70 + .24 mm. The western 314 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY sample of this population averages 38.10 = .15 mm., a significant difference from the eastern (36.10 = .32) but not from the central sample (36.85 + .61 mm.). Ridgway, in commenting on variation in this character in P. e. erythrophthalmus, found the reverse of the results presented here. His nine eastern males and eight western males averaged 40.64 and 38.35 mm., respectively. TES OT C ERYTHROPHTHALMUS N2136 = =< CANASTER -4 Gosesees, ALLEN! oe ornny 6 10 15 20 25 30 35 40 45 . 30 35 LENGTH {N MILLIMETERS Figure 5. Variation in length of white spot on outermost rectrix of males. See page 304 for explanation of figure. The tail spot of canaster averages 31.79 + .40 mm., a significant difference from that of erythrophthalmus. Howell (1913) used this as a diagnostic character in separating canaster from erythrophthalmus. However, I find this to be unreliable under statistical scrutiny, since only 69.5 per cent of the erythrophthalmus sample is separable from that of canaster. Rileyi has the next smallest amount of white on the tail, averaging 25.50 + .41 mm. This race shows a significant difference from ery- throphthalmus, canaster and allent. In my samples rileyi shows 77.3 per cent separation from canaster, 93.3 per cent from erythrophthalmus, and 76.3 per cent from alleni. This character is of diagnostic value in separating rileyi from canaster, erythrophthalmus and allenv. P. e. allent has the smallest amount of white on the tail, the spot on the outermost rectrix averaging 19.71 + .50 mm. The difference is statistically significant and allows the use of this measurement for diagnostic purposes. The tail spot of alleni affords 97.1 per cent sepa- ration from erythrophthalmus, 76.3 per cent from riley, and 90.8 per cent from canaster. The variation in this character shows a northwest-southeast trend, the amount of white diminishing as populations are examined from North Dakota south and east to peninsular Florida. Of interest is the fact that among the Florida material studied, those having the smallest DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 315 amount of white (6-8 mm.) came from coastal localities within the peninsula. VAIS Oh {0} = = ERYTHROPHTHALMUS. CCS SOCECEE 4 §678 910 35 40 45 20 25 30 LENGTH IN MILLIMETERS Figure 6. Variation in length of white spot on outermost rectrix of females. See page 304 for explanation of figure. Females (Figure 6). P. e. erythrophthalmus shows the greatest amount of white on the rectrices. The average extent is 33.18 + .46. This character is of diagnostic value in separating erythrophthalmus from canaster (78.5 per cent), from rileyi (93.0 per cent) and from allen (94.3 per cent). Canaster in turn is separable from rileyi (73.5 per cent) and from alleni (94.3 per cent). Rileyi-allent separation on the basis of this character is 89.3 per cent correct. The females show the same pattern of geographic variation as do the males. Culmen Length Males (Figure 7). The length of culmen is quite different in its geographical variation from those characters previously discussed. Howell (1913: 202) first called attention to the longer bill found in the race he designated as canaster. In the material examined in this study it was found that the culmen of canaster averages 15.44 = .05 mm. Erythrophthalmus, to the north, averages 14.49 + .03 mm. On the basis of this character 75.3 per cent separation of erythrophthalmus from canaster obtains.! In the three northern samples (erythrophthalmus) no significant difference in culmen length was found. The eastern segment of the 1 From a practical standpoint it is unfortunate that the optimum point of division falls at 14.96 mm. The culmen measurement is at best a difficult one to make on a short-, heavy- billed bird such as the towhee. In addition the fact that this point lies near an even millimeter, possibly causing bias in measurement, is not helpful. This does not negate the difference, how- ever, and the culmen length is certainly an indicator of the racial distinctness of this form. 316 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY CULMEN ef N= 352 ERYTHROPHTHALMUS Fe--- ons ell LENGTH IN MILLIMETERS Figure 7. Variation in culmen length of males. See page 304 for explanation of figure. population averaged 14.51 + .04, the central segment 14.56 + .06 and the western segment 14.60 + .06 mm. Ridgway found a greater difference in the material he had at hand, nine eastern males averaging 14.22 and eight western males averaging 13.97 mm. The culmen of rileyi is significantly longer than that of canaster, averaging 15.65 = .06 mm. The difference is not sufficient to separate canaster from rileyi. However, 84.3 per cent of riley2 and erythroph- thalmus are separable. Culmen length in allenz is significantly less than that of canaster and rileyi,t he average length being 15.05 = .07 mm. Although serving to indicate a difference in the population, it is not of diagnostic value. Maximum culmen length is obtained in canaster and rileyi. To the north and south it diminishes, and the northwest-southeast trend is not as pronounced as in other characters. Females (Figure 8). P.e. canaster and rileyi females have the longest culmens, averaging 15.20 + .10 mm. and 15.34 + .09 mm., re- spectively. Canaster is separable with 73.5 per cent accuracy from DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 317 CULMEN a ERYTHROPHTHALMUS wane n-----4 CANASTER ---4 RILEY! 13 14 15 16 17 18 LENGTH IN MILLIMETERS Figure 8. Variation in culmen length of females. See page 304 for expla- nation of figure. erythrophthalmus, which averages 14.47 = .07 mm. Correct identifi- cation of erythrophthalmus and rileyi obtains 78.0 per cent of the time. Allent, averaging 14.71 = .10 mm., is significantly smaller than ery- throphthalmus, canaster and rileyi. The degree of separation of alleni from the other three races is less, however, than 75.0 per cent. The pattern of geographic variation in the females is the same as in the males. Width of Lower Mandible Males (Figure 9). Lower mandible width shows much the same geographic variation as does culmen length. Canaster has the widest bill, with a mean of 8.94 = .03 mm. It is significantly different from erythrophthalmus in this measurement, the average width in the latter being 8.68 + .02 mm. Within the northern race the western birds appear to have a slightly larger bill, with a mean of 8.72 = .03 mm. This is near the borderline of significance in relation to the eastern birds, with a mean of 8.61 = .02 mm. The material examined from the northcentral area has a mean mandible width of 8.61 + .03 mm. 318 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY It shows no significant difference from the adjacent segments to the east and west. LOWER MANDIBLE N= 346 ERYTHROPHTHALMUS CANASTER RILEY ALLEN] 7 8 9 10 Ut 12 WIDTH IN MILLIMETERS Figure 9. Variation in width of lower mandible of males. See page 304 for explanation of figure. The bill of riley? has a mean width of 8.85 = .03 mm., which is not significantly different from that of canaster. It is, however, signifi- cantly different from erythrophthalmus. Rileyi is separated from allent 74.3 per cent of the time. Alleni has a smaller bill, averaging 8.48 = .04 mm. A significant difference is shown in relation to canaster and rileyi. Alleni is separable from canaster 74.5 per cent of the time. Of interest is the fact that erythrophthalmus and alleni show no dif- ference in this character. Once again a northwest-southeast gradient is present, with maximum size in canaster and rileyt. Females (Figure 10). The width of the lower mandible is of no diagnostic value in this sex. Canaster and rileyi have the greatest average width, 8.69 + .05 and 8.79 + .07 mm., respectively. They are not significantly different in this character. Erythrophthalmus averages slightly smaller (8.60 = .04 mm) but is not significantly different from canaster and rileyi. Alleni is significantly smaller than DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 319 LOWER MANDIBLE O N=90 ERYTHROPHTHALMUS [22 Fos|| N=28 CANASTER ; b---- ----4 N=37 RILEY! F-- ----4 N=22 ALLENI b--- ---4 7 8 9 10 Vw WIOTH IN MILLIMETERS Figure 10. Variation in width of lower mandible of females. See page 304 for explanation of figure. TARSUS OF ERYTHROPHTHALMUS i] N=145 : un CANASTER 24 25 26 27 28 29 30 31 LENGTH IN MILLIMETERS Figure 11. Variation in length of tarsus of males. See page 304 for expla- nation of figure. 320 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY rileyt and canaster, averaging 8.47 = .06 mm. Alleni and erythroph- thalmus are not significantly different from each other. Tarsus Length Males (Figure 11). The length of tarsus shows much the same geographic trend of variation as do the characters previously dis- cussed. Rileyi and canaster have the longest tarsi, 28.25 + .10 and 28.19 = .08 mm., respectively. The difference is not statistically significant in this case. Erythrophthalmus has a shorter tarsus, averaging 27.37. .05 mm. Although the tarsus is significantly different from that of canaster, rileyi and allenz, the character is of no diagnostic value. Alleni has the shortest tarsus, average length being 26.78 + .11 mm. Tarsus length is of diagnostic value in separating alleni from canaster and rileyi. In canaster-alleni separation, 76.8 per cent cor- rectness of identification is achieved, and in rileyi-alleni 77.3 per cent. Within the population here considered as erythrophthalmus I find no significant difference in the three samples studied. The eastern sample averages 27.42 = .06, the central sample 27.05 = .01, and the western sample 27.50 + .09 mm. Ridgway obtained the same results in his earlier study, the nine eastern and eight western males each averaging 28.45 mm. TAmoUS ERYTHROPHTHALMUS -----4 N=36 [eR es 24 25 26 27 28 29 30 LENGTH IN MILLIMETERS Figure 12. Variation in length of tarsus of females. See page 304 for explanation of figure. DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 321 Females (Figure 12). Alleni has the shortest tarsus, averaging 25.89 + 18 mm. It is significantly different from erythrophthalmus, which averages 26.60 = .10 mm. Allenz is 80.0 per cent separable from rileyr, whose tarsus averages 27.54 + .19 mm. There is 85.5 per cent separation of allenz from canaster, whose tarsus averages 27.91 + .20. Canaster and rileyi are not significantly different from each other in this character. Erythrophthalmus and canaster are 75.5 per cent separa- ble. Less than 50 per cent separation obtains in rileyi-erythrophthalmus samples. The pattern of geographic variation is the same as that found in the males. Middle Toe Without Claw Males (Figure 13). Canaster, with a mean middle toe length of 20.19 = .07 mm., has the longest toe, and is significantly different from erythrophthalmus, rileyi and alleni in this measurement. Never- theless, only in the case of alleni and canaster is this character of diag- nostic value, furnishing 78.3 per cent correct identification. ERYTHROPHTHALMUS N=145 CANASTER +----- -4 RILEY! ————— a =8I ALLEN! a 17 18 19 20 2i 22 23 LENGTH IN MILLIMETERS Figure 13. Variation in length of middle toe without claw of males. See page 304 for explanation of figure. Rileyi has the next smallest middle toe, averaging 19.89 = .07 mm. As pointed out above, it is significantly different from canaster in this measurement. In addition it is significantly different from alleni to the south and from erythrophthalmus to the north. oae BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY The toe of erythrophthalmus averages 19.64 = 04 mm. ‘This is sig- nificantly shorter than that found in canaster and riley, and signifi- cantly longer than that found in adlent. No significant difference was found in the samples examined from the eastern, central and western portions of the range of erythrophthalmus. Ridgway found that there was little difference in his nine eastern and eight western males. The average middle toe measurement was 19.81 and 19.56 mm., respec- tively. Allent has a much shorter toe than any of the other races, averaging IS.88 e& 09 mm. On the basis of this measurement allen is separable from canaster in 78.8 per cent of the material examined. Tt is also separable in 75,0 per cent of the cases from riley’. Correct separation from erythrophthalmus is possible in less than 70 per cent of a mixed sample of the two populations, —... ERYTHROPHTHALMUS N®27 _ CANASTER RILEY! =e ALLENI 7 18 19 20 2t 22 LENGTH IN MILLIMETERS igure I. Variation in length of middle toe without claw of females, See page S04 for oxplanation of figure, A general northwest-southeast clinal trend is observable in this character as in the others already discussed — maximum size occurring once again in caraster and riley’, with diminishing size to the north- west and southeast. Females (Figure 14). The middle toe in the female is not as variable DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 323 as in the male, although the same pattern of geographic variation ap- pears. Canaster and rileyi, not differing significantly from each other, have the longest middle toes, averaging 19.66 = .16 and 19.38 + .11 mm., respectively. Canaster is significantly different from erythroph- thalmus, which averages 19.06 = .OS mm., and from alleni, which averages 18.57 + .14 mm. Accuracy of 76.3 per cent is attained in separation of canaster from alleni. Rileyr is significantly different from alleni, but not from erythrophthalmus in this measurement. Tail Spot: Tail Males (Figure 15, upper). The ratio of extent of white on the outer- most rectrix to length of tail proves to be of value in separating ery- throphthalmus and canaster. In these two races the tail length and tail spot vary reciprocally, and for this reason the differences are magnified by calculation of this simple ratio. In erythrophthalmus, the tail spot is 39.36 + .27 per cent of the tail length, and in canaster the percentage is 32.59 + .45. This furnishes 74.5 per cent separation of these two forms. TAILSPOT LA AIL N=340 , ERYTHROPHTHALMUS - i 2 CANASTER enn eeennn ---4 20 25 0 35 40 45 50 55 60 PERCENT (TAILSPOT + TAIL) Figure 15. Variation in length of white spot on outermost rectrix: males (upper) and females (lower). See page 304 for explanation of figure. Females (Figure 15, lower). In the females P. ¢. canaster has a mean of 29.41 + .74 per cent, and erythrophthalmus a mean of 37.47 + .51 per cent. The disproportionate increase in length of tail spot to length 324 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY of tail makes for an even greater divergence in this sex, with 82.3 per cent separation obtaining. Number of Rectrices Showing White on the Inner Web Males and Females (Table 20). The greater extent of white on the inner web of the outermost rectrix in the northern races has been com- mented on earlier. In addition to extent of white it is also true that more individual rectrices show white in the northern population. The Table 20 Variation of White on Rectrices Number of pairs involved, expressed in per cent of total sample Pairs of Rectrices with White Subspecies 1 2 3 4 5 erythrophthalmus (N = 347) 0.6 43.8 525i 2.9 g canaster (N = 148) ene Zeal 75.5 22.4 = rileyt (N = 97) Were ae 7.6 87.0 5.4 ey allenit (N = 79) ae 32.9 59.5 7.6 a erythrophthalmus (N = 89) Heil 69.7 29.2 iS canaster (N = 25) aie 12.0 80.0 8.0 ae riley (N = 24) i. 20.0 79.2 a 3 allent (N = 24) 4.2 75.0 20.8 BaLea Ps males in P. e. erythrophthalmus furnish the upper extreme, occasion- ally showing white on as many as five pairs of feathers. One female of alleni showed white on only a single pair. In general males tend to show white on more feathers than do females. The character is of little diagnostic value in relation to the populations but may prove helpful in individual identifications. Table 20 presents the data for this character. DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 325 Iris Color Males and Females. Variation in iris color presents some difficulty in analysis for the reasons outlined in the introduction of the present paper. P. e. erythrophthalmus and P. e. alleni may be largely disre- garded, since they uniformly have “red” and ‘‘white’’ irides, respec- tively. Canaster and rileyz, on the other hand, show mixtures in varying degrees of these two basic colors, and it appears that a knowledge of variation in iris color is quite important in gaining a clearer under- standing of the racial relationships within the species. Basically it is readily apparent that the intermediate eye colors are found in the areas of geographic intermediacy between the northern and southern races (Map 3). The map does not attempt to weigh the different iris colors according to relative percentages but simply furnishes a record of the limits of distribution of the three colors plotted. Arbitrary classification of the multiplicity of color notations made by various collectors was quite necessary. An attempt was made to be as conservative as possible in the interpretation of the color notation made on the labels. Weighing of the several colors was ac- complished by use of the Chi-square test (Table 21). The degree of association of iris color with geographic locality is very high, so high that it may safely be assumed that the probability of such distribution occurring by chance alone is nil. Iris color in canaster may be “‘orange”’ or “‘yellow” but it is usually red (84.25 per cent). In rileyi the color may be “red” or “orange” but it is usually yellow (79 per cent). Material from the type localities is unfortunately slightly confusing. A canaster topotype (JCD 269) taken near Mobile, Alabama, had yellowish irides. Specimens (PB 15771 and JCD 263) taken by Dr. Pierce Brodkorb and myself near Brunswick, Georgia, were close to salmon (10-G-3) and buff (10-I-5), respectively. There seems to be no reason to call upon any genetic principle more complicated than multiple factors or multiple allels for explanation of this distribution. No clues are available at present to indicate the number of genes involved, but the almost perfect blending of pigments seems to follow the pattern usually associated with these types of inheritance. The splotchy distribution of color found in some indi- viduals, and reports of pie-shaped segments of different colors in the irides furnish further support for this hypothesis. The iris color in juvenals has apparently been overlooked by most investigators dealing with the pale-eyed birds of the southeast. Maynard (1881: 114) states that Florida nestlings have light brown irides. The young birds of the north have iris color that is usually described as ‘‘muddy brown’, “brownish”, “‘sepia-brown’’, ‘dark The Alabama, Mississippi, Louisiana, west Florida and central Table 21 Variation in Iris Color Georgia records represent P. e. canaster. The coastal North Carolina, coastal South Carolina and southeastern Georgia records represent P. e. rileyt. [=4} (eel [oad (enh? @) (S) le < frel Sh fo PG) CO -<| Fel Ite = ©) Ole MALES Ala., Miss., La., west Fla., central Georgia Coastal IN. C:, S. C., southeast Georgia RED 119 IRIS COLOR ORANGE 19 x? = 158.54 (Surely significant) MALES Ala., Miss., La., west Fla., central Georgia Coastal N. C., S. C., southeast Georgia BOTH SEXES Ala., Miss., La., west Fla., central Georgia Coastal N. C., S. C., southeast Georgia IRIS COLOR RED ORANGE 20 3 0 7 x? = 41.52 (Surely significant) IRIS COLOR RED ORANGE 139 22 2 19 x2 = 190.67 (Surely significant) YELLOW 59 YELLOW 20 YELLOW us DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 327 brown’, “brown’’, or some other color designation which appears to fall in the brown category. Sutton (1935: pl. III) pictures a 15-day old male bird from Michigan as having dark brown irides. Fully- fledged juvenals (JCD 180) taken in Florida, and in Georgia, 5 miles north of Brunswick, Glynn County, (JCD 260 and 264) have the iris color near “Pale Neutral Gray” in Ridgway’s system and 15-A-3 in Maerz and Paul. Plumage Characteristics of Females Efforts to establish color classes for the various parts of the plumage were not successful. The range and variability of the colors is such that clear-cut differences are not observed. Four areas of plumage were scrutinized: back, flanks, breast and crown. After it was decided that it was not possible to use the techniques of color classes for statistical analyses of the plumage characters, the material was identified on the basis of mensural characters and notations made as to range of color in the various races. The possibility of color change due to age of skins was considered, and old skins (1930 and earlier) were compared with fresh material (1940 and later). I was not able to detect any change due to museum age. Back color. P. e. erythrophthalmus as a whole is more reddish brown than P. e. alleni and P. e. rileyi. Some skins (CM 2290) are quite reddish (15-A-12); others (USNM 363240) though reddish are very pale, (15-H-9). USNM 23598 is much darker (15-H-9) than the average. USNM 56535 is dark but not quite as reddish (8-H-12). The average color of this race appears to be best represented by UMMZ 96950, (15-J-8). Birds in fresh plumage are slightly darker than those showing wear, and a skin from Jackson County, Michigan (GMS, Oct. 11, 1949), is typical of these fresh-plumaged birds (8-E-11). In P. e. rileyi the older skins show more red pigment than the fresh material. In this case, however, the older material is from the northern and inland localities where riley: meets canaster and erythrophthalmus, both of which are on the average more reddish than rileyi. MCZ 208060 is typical (15-C-11). In this case it seems quite likely that the difference is due to geographic rather than temporal factors. The coastal localities provide material that is very similar to alleni in general color (15-E-7). Some individuals are much darker (JCD 258 and PB 15772). The canaster sample shows no detectable differences due to age of skins. USNM 38247 (1946) matches CNHM 166708 (1916) in back color (15-E-11). The average color of the sample is typified by USNM Q9 328 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY 340494 (15-E-9). The material examined appears to be more olive and reddish than the allenz sample and slightly grayer than ery- throphthalmus. Alleni is conspicuously grayer than erythrophthalmus and canaster but not grayer than rileyi. Some individuals, AMNH 368079 (15-J-11), approach canaster in redness, but the general color of the sample is typified by USNM 261709 (15-E-7). There is no detectable change in color due to age of skin. The color of the back does not show any sharp breaks which are correlated with geographic distribution. The northern race shows much more red pigment than the peninsular Florida population, but in the intervening populations the colors are apparently the result of varying admixtures of these extremes. Flanks. P.e. erythrophthalmus has darker flanks on the average than have the other races. Average color is represented by USNM 339625, 257922 and CM 2762 (13-H-12). Some skins are much darker (14-E-12) than the average, as in USNM 268666, 235598, 306445 and UMMZ 66998. The lightest flanks (13-J-10) in the sample at hand are from a skin taken at Wheeling, West Virginia, on May 2, 1936, now in the M. M. Peet collection. P. e. rileyi has very uniform coloration of the flanks and is quite like erythrophthalmus in color. The darkest specimen (MCZ 10355) is 14-I-11, the lightest (USNM 298673) is 13-K-11, and the average color, as represented by USNM 382391, is 14-C-12. P. e. canaster is more variable in flank color. The darkest specimen, USNM 382361, is 14-E-12, and the lightest is 13-K-9. The average color is lighter than erythrophthalmus and slightly darker than riley. It is represented by 13-K-9, as found in USNM 378909. P. e. allent has much lighter flanks than the other races. USNM 261711, with 13-D-11 flanks, is considered as being typical. The darkest color present in my material is 13-G-10 (USNM 133091). The lightest, 12-I-10, is found in LSU 8148. Breast. P. e. erythrophthalmus is quite variable in the coloration of this area of plumage. Skins from the far northwest, Grafton, North Dakota, are very dark, 15-C-12, as seen in UMMZ 66998 and 56535. The lightest breast in this sample, 14-H-9, is found in USNM 348791. Some are quite reddish, 15-A-12, USNM 306445. The average color appears to be 14-K-9, as found in AMNH 367858. P. e. rileyi has more gray pigment (or less red) than erythroph- thalmus. The grayest, 15-A-6, is JCD 259, while the most reddish, 15-C-11, is found in MCZ 208060. The average color is best repre- sented in MCZ 212455, as 15-C-9. DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 329 P. e. alleni shows considerable variation in breast color. The average, 14-J-9, found in USNM 300219, is paler and less red than rileyz and canaster. JCD 176 is as reddish as the average canaster color, 15-C-12. USNM 261711 is the darkest in this sample, 15-C-6, and LSU 8148 is the lightest. Crown. P. e. erythrophthalmus has the darkest crown of the four races. USNM 338022 and UMMZ 56535 are very dark, 8-E-12. UMMZ 74532 is the lightest, 14-E-8. Some skins show considerable redness, as in CM 2290, which is 15-A-12. GMS 8917 appears to be representative of the average color, 8-L-12. Canaster is paler than erythrophthalmus; the average color, found in USNM 258980, is 7-A-12. One skin, USNM 379723, is very dark, 8-H-12. The lightest color observed, 15-C-12, occurs in a specimen taken in Fulton Co., Georgia, May 2, 1928 (M. M. Peet collection). P. e. riley is slightly paler on the top of the head than is canaster. An average skin is represented by USNM 298673, 8-H-11. MCZ 208060 (see breast) is the most reddish, 7-A-12. PB 15772 is the darkest, 8-J-12. AMNH 55406 is very light, 14-L-10. Alleni is the palest of the four races. The average color, 15-E-12, is present in USNM 261708. JCD 176 (see breast) is quite reddish, 14-C-12. The darkest specimen, AMNH 368317, is 8-H-12. Summary. There appears to be a general intensification of red and black pigments in the northern areas in all the feathers showing color. There is considerable difference between erythrophthalmus and allent, but in so far as I am able to determine, the blending of colors is so gradual and variability is so excessive that coloration is of taxonomic value in this species only when large series are available for comparison. Plumage Characteristics of Males Back, Crown and Breast. The color of these areas of plumage varies little in the four races. The average color of the back in erythroph- thalmus is Sooty Black!. Canaster averages slightly paler, Olivaceous Black (38). Olivaceous Black appears to represent the average color in riley. Allent is conspicuously paler, Iron Gray. In all of the races there is considerable individual variation, and Sooty Black individuals occur in all four. As a result of wear and bleaching, the darker races occasionally show very pale back colors. The color of the crown and breast presents the same pattern of geographic variation as does the back. To my eye the color of these areas is the same as that of the back. 1Colors approaching black are not adequately treated in Maerz and Paul, and Ridgway (1912) was used for these colors. 330 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Flanks. P. e. erythrophthalmus has the richest color in the flank feathers. The darkest specimen examined was CM 2831, which was 6-A-12. CM 7576 was the lightest of the northern birds, 12-D-11. The average color was matched by 5-D-12 in AMNH 367998. Canaster is slightly paler than erythrophthalmus, darker than alleni, and paler than riley. The darkest specimen, USNM 378920, was 13-K-11. The average color present in this race is 13-D-11, as found in USNM 338901. Some birds are quite light, as light as average allent specimens. The lightest, 12-F-10, was found in USNM 240167. Rileyi shows little variation in flank color. In a series of eight birds from the type locality seven are remarkably uniform. One (JCD 256) is as dark as the darkest specimen of P. e. erythrophthalmus, 6-A-12. The average color is near 13-C-12, as in JCD 253. The palest specimen, JCD 255, is 12-H-8. The palest of the races, alleni, has an average flank color which is near 12-F-10, as present in JCD 178. In the only cotype of this race available to me, MCZ 10721, the color is 12-G-10. JCD 249 is the palest specimen examined and is near 11-H-8. The darkest flanks were found in JCD 178, 13-A-12. Summary. In this sex the same general intensification of color occurs in the northern races. Alleni is conspicuously paler in all plumage showing color. Rileyi and canaster are intermediate in color. The sample representing canaster shows much greater variation than does that of rileyi. NON-GEOGRAPHIC VARIATION The most unusual individual variant was discovered in a bird of the year taken by W. H. Osgood. The bird was collected in Maryland, 10 miles north of Washington, D. C., on August 1, 1897, and is now No. 367895 in the collection of the American Museum of Natural History. The specimen is labeled as a male, and the greater part of its plumage is clearly of this sex. A large patch of female plumage, however, is present on the upper back, extending forward to the neck and around the right side. No flight feathers appear to be involved. Heterosexual plumage changes have been recorded on many occasions in domestic birds, and Brodkorb (1935) reports what may have been a similar situation in Falco sparverius. His explanation of that case as gynandromorphism does not furnish an adequate solution for the pattern present in this case. I am inclined to believe that a non- bilateral pattern is the result of somatic change rather than hormone interplay. In this case there is no information available concerning the gonads, but it does not seem likely that such a spotted distribution DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 331 would be produced in a sexually dimorphic character by lack of balance in the male and female hormones. If it is assumed that the towhee has the same type of sex determi- nation as poultry, then another proposal can be made. If this specimen developed from a fertilized ovum carrying a pair of X chromosomes, as such it was destined under normal circumstances to become a male. Distortion of the normal mitotic processes on many occasions has been known to result in the deletion of parts of or whole chromosomes. The number of daughter cells showing the effects of the deletion is de- pendent upon the stage of development of the individual at the time the aberration occurs. Development of female secondary sex charac- ters from the heteroploid cells, produced in such fashion, is due to the absence of the X chromosome rather than the presence of the Y. Such a chain of events appears to have occurred in this individual. White-tipped feathers at the bend of the wing, involving the tips of the secondary coverts, and varying degrees of white streaking in the scapulars have long caused speculation as to the affinities of P. ery- throphthalmus and P. maculatus (Baird, Brewer and Ridgway 1874, Allen 1878, Coues 1878, and more recently by Sibley 1950.) The appearance of white at the above mentioned points has been looked upon as evidence of an exchange of germ plasm between the eastern and western species, or as an indicator of common origin of the two forms. There seems to be some evidence that such an exchange does at least have the opportunity to occur in that the ranges of the two forms closely approach each other. In the breeding material examined in the course of the present study varying degrees of white tipping of the secondary coverts was found in 136 male and 29 female specimens, taken over the entire range of the eastern forms, even in Florida and New York. USNM 302208, a male taken near Athens, Georgia, on February 24, 1930, shows this spotting and streaking of the coverts and scapulars to a marked degree. The iris color, as recorded by T. D. Burleigh, was ‘‘dirty yellow.”” This would seem to rule out the possibility that the bird was a stray from the west. On the basis of mensural characters the Georgia bird is identified as P. e. canaster. The degree of marking is equal to that found in GMS 10172, a male taken at Lincoln, Nebraska, on May 13, 1946. George M. Sutton, the collector, comments on the label ‘‘white on back re- duced — P. m. arcticus approaching P. e. erythrophthalmus.” The recommendation made by Sibley (op. cit.) that erythrophthalmus and maculatus be considered conspecific would seem to be a wise action. Although I did not have sufficient material at hand to critically analyze the geographic distribution of this variation I am inclined to think 332 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY that it is indicative of common origin rather than interbreeding in recent time. Albinism is quite rare in the species. Only one specimen, an adult female P. e. erythrophthalmus from North Carolina, AMNH 104434, shows any appreciable degree of this plumage change. The iris color was recorded as “pearl gray.” The pattern of white is bilaterally symmetrical and extends from the front to the rectrices. The neck is white, and the contour feathers of the body are generously sprinkled with white. Several of the primaries and secondaries are involved. Twenty-one other specimens were found showing from one to perhaps several hundred randomly placed white feathers. In both sexes there is a variation which involves the tips of the crown feathers. In seven specimens from scattered localities these feathers were conspicuously tipped with rufous. This variation may be still another indication of the affinity that exists between the eastern and western segments of the populations. The pattern presented by this variant is similar to that found by Sibley in the hybrid population of Cerro Viejo, Jalisco. (See his plate 12, MVZ numbers 115243 and 115215). MIGRATORY BEHAVIOR It has long been known that P. e. erythrophthalmus moves well southward during the winter. The practicality of distinguishing red- eyed from pale-eyed individuals in the field has had two effects: (1) to confuse the literature with numerous sight records of this race when the birds may have been P. e. canaster and (2) to allow the retention of Howell’s original concept of the sedentary nature of canaster. It is felt that on the basis of the present study, valid criteria for the identi- fication of the four races have been determined. On the basis of these criteria, wintering material has been examined and identified with interesting results. Admittedly some of the individual identifications are not correct. However, it is assumed that sufficient numbers of specimens were at hand to minimize the danger of false conclusions with reference to the general winter behavior of the four races. Map 4 shows the result, in summary form, of this phase of the investigation. P. e. erythrophthalmus withdraws practically all of its numbers from the breeding grounds. Some few individuals for unknown reasons do remain in the northern United States during the winter, but the majority leave, spreading south and west as far as Nueces and Lee Counties, Texas, and south and east in Florida as far as Hillsborough, Okeechobee and Volusia Counties. DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 3833 P. e. canaster, regarded by Howell as remaining on the breeding grounds during the winter, certainly does this, in so far as a portion of the population is concerned. On the other hand, winter specimens, taken in Florida as far south as Wakulla County on the northern Gulf Coast, indicate some post-breeding movement of individuals. In the material examined there were no specimens of canaster from the east coast of Florida. There are numerous winter-taken individuals of canaster from south-central Georgia, in the area occupied by rileyi during the breeding season. There appears to be a slight retreating of the population along the northern extent of its breeding range. Winter specimens, with one exception, from southern Tennessee, northern Alabama, Mississippi and Louisiana are typical of ery- throphthalmus. P. e. rileyi spreads north, south and west, during the winter. The wandering to the north and west is not marked, but to the south rileyz extends its population to mid-peninsular Florida. Winter specimens of this race are available from as far south as Charlotte County on the west coast and Brevard County on the east coast. It is apparent, however, that many of the birds do remain within the breeding range outlined for this form. It is interesting to note that the series of cotypes established by Coues’ action in describing the Florida race contains some four individuals (C. J. Maynard 2559, 2669, 2592, 2513) whose wing measurements as given indicate that they may have been wintering birds from the north (rileyi). This series was taken during February and March. Only one specimen of this series (MCZ 10721) was available to me, and it is typical of allent. My measurement of 77.0 mm. for length of wing as compared with 2.92 inches [74.2 mm.] as given by Allen (1871) probably indicates that these earlier wing length figures represent “chord”? measures.! P. e. alleni apparently is largely sedentary. Howell (1932: 448) did not have records available for breeding birds in the extreme sovthern part of peninsular Florida but did indicate that alleni spread southward during the winter to this area. Summer specimens (JCD 185, UMMZ 114394 from southern Dade County, Florida) furnish evidence that this form now extends into the extreme southern tip of Florida as a breeding bird as well as in winter. JCD 185 was taken while in com- pany with females and juvenals. There is no evidence of any extensive post-breeding northward wandering in this race. 1 Two discrepancies in the table cf measurements given by Allen are worthy of note. A wing length of 3.90 inches (99.1 mm.) for MCZ 10726 must be due to error in measuring or in publication. This length far exceeds any seen by me in any of the four races. There is duplica- tion of MCZ 10729 in having this number assigned to two of Maynard’s skins, 2668 and 2669. 334 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY VARIATION IN RELATION TO ENVIRONMENT Ecological Rules The data available have been examined from the standpoint of several pertinent ecological rules suggested as being applicable to warm-blooded vertebrates. Mayr (1942: 88-92) has summarized these rules and his wording of them is followed below. Bergmann’s Rule. ‘The smaller-sized races of a species are found in the warmer parts of the range, the larger-sized races in the cooler districts.” Insufficient data concerning body weight were available, and hence this measure of size could not be examined adequately. Other measures, wing, tail, culmen, width of lower mandible, tarsus and middle toe, which may possibly be indicators of total body sizel, present a puzzling picture. Alleni is smaller than any of the other races in respect to these characters. In keeping with Bergmann’s Rule the adjacent races to the north, canaster and rileyi, are larger. Hry- throphthalmus from still further north however, is smaller than the intermediate races, though not as small as alleni. It is of course diffi- cult to estimate the selective effect of lowered winter temperatures on migratory forms. Erythrophthalmus moves well down into those areas where a large portion of canaster remains as a resident form during the winter, and for the most part does not experience lower temperatures than the resident southern individuals. A segment of the population of rileyi accomplishes this same movement in relation to alleni and as a population is probably not subjected to much lower temperature than the peninsular Florida race. In this case, then, it appears that if the breeding ranges are con- sidered, allenz follows the rule in relation to canaster, rileyi and ery- throphthalmus. Canaster and rileyi, however, do not seem to follow the rule in relation to erythrophthalmus. Allen’s Rule. ‘‘Protruding body parts, such as tails, ears, bills, extremities, and so forth, are relatively shorter in the cooler parts of the range of the species than in the warmer parts.” Lack of infor- mation concerning total body size made valuation of this rule difficult. Examination of culmen length, mandible width, tarsus and _ toe, however, shows that the most southern race is the smallest of the four. Canaster and rileyi are larger than alleni in these measurements but erythrophthalmus is smaller than the former two. In this case ery- throphthalmus may conform to the rule in relation to canaster and rileyt. Alleni, however, apparently does not conform in relation to any of the other forms. 1 Miller (1941: 354), in his detailed investigation of variation in Junco, concluded that wing and tail length in this genus were largely independent of body size. DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 335 Gloger’s Rule. ‘The melanins increase in the warm and humid parts of the range. Reddish or yellowish brown phaeomelanins prevail in arid climates where the blackish eumelanins are reduced. The phaeo- melanins are reduced in cold climate, and in extreme cases also the eumelanin (polar white).”” In examination of this rule it was felt that the breeding season was the critical period. For this reason July values for this environmental factor were examined (Weather Bureau: 1897). The breeding range of alleni has the highest relative humidity. The average for two stations is 83.0. Rzleyi inhabits the next most humid area, two stations averaging 81.0. The range of canaster is slightly less humid, averaging 80.3 for six stations. The more northern and inland range of erythrophthalmus is considerably less humid; 12 stations average 70.2. In both the males and females the northern form, inhabiting an area of lowest relative humidity, shows an increase in the reddish pigments present. Canaster is grayer than erythrophthalmus. Rileyz shows more red on the plumage than does canaster, but not as much as erythrophthalmus. The plumage of alleni is not darker than these, but rather more gray, as if due to bleaching. Thus, riley? and allent appear to conform to this rule as a unit. Canaster and ery- throphthalmus, when viewed together, also appear to conform to Gloger’s rule. When the whole species is considered, however, a lack of conformity is evident. Rensch’s Clutch Rule. ‘The races of a species which live in the cooler parts of the range of that species lay more eggs per clutch than the races in the warmer parts of the range.’’ Todd (1940) and Roberts (1932) report for western Pennsylvania and Minnesota, respectively, average egg clutches of four to five. Howell (1932) records the average size of clutch in Florida as three. Sprunt and Chamberlin (1949) indi- cate clutches of two to five for canaster in South Carolina. With the scanty data available general agreement with this rule appears to be the case. Rensch’s Wing Rule. ‘“The wings of races that live in a cold climate or in the high mountains are relatively longer than those of the races that live in the lowlands or in a warm climate.” Allenz is the only race which shows conformity with this rule. In its relationship to riley, canaster and erythrophthalmus, the wing of alleni shows considerable shortening. Erythrophthalmus, however, has a mean wing length which is not longer than that found in canaster and rileyz, but shorter. Mayr’s Rule (1942: 92). “Races in the cooler climates are more strongly migratory than the more southerly one.” Pipilo erythroph- thalmus does not conform to this rule as indicated in the discussion of migratory behavior in this report. 336 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Biotic Areas Dice’s (1943) Biotic Provinces have been scrutinized in this con- nection and little conformity of racial distribution with those provinces is evident except in erythrophthalmus. This race is distributed in an approximation of Dice’s Canadian, IIlinoian and Carolinian provinces. The other three races, however, are mainly contained in his single Austroriparian province. Conformity of racial distribution of towhees to the Life Zones of Merriam et al. (1910) presents much the same picture. P. e. ery- throphthalmus breeds in the eastern part of the Upper Austral and Transition zones. The differentiation of canaster, rileyi and alleni within the Lower Austral Zones does not conform with the proposals of either Dice or Merriam with regard to environmental sameness for this area. Temperature Zones Visher (1944) presented a series of seventeen maps concerned with freezing temperatures in the United States. His Figure 17, based on the duration and severity of freezing temperatures, divides the United States into six zones. These zones approximate fairly closely those of Merriam, except in the extreme southeast. Visher’s data indicate that southern Florida, northern Florida and the Gulf Coast, and the Piedmont and Coastal Plains areas of the southeast should be con- sidered as different from one another in respect to freezing temper- atures. Calhoun (1947) concluded that there was correlation of size in Passer domesticus with the thermal lines drawn by Visher. Visher’s zones 4, 5, and 6 cover the range of erythrophthalmus. The southern limit of zone 4 closely approximates the southern limit of this race during the breeding season. Zone 3 contains canaster in its western, and rileyi in its eastern portion. The southern limit of this zone conforms favorably with the southern limit of these two races. P. e. allent ranges over Visher’s zones 1 and 2. There is approximate con- formity of racial distribution to these zones. It appears that the various ecological rule, biotic province, life zone and temperature zone correlations are of most importance in dealing with non-migratory species. Migrating populations for the most part avoid the selective effect of lowered winter temperatures by simply moving away from them. The few stragglers that remain in the far north must be of little importance when considered with respect to the whole population. DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 337 Historical Factors In view of the fact that present environmental factors do not appear to furnish an entirely adequate explanation for the geographic vari- ation within the species, some other factors must be considered. As mentioned earlier great difficulty is experienced in arriving at a con- clusion as to whether the (1) southeastern races represent an extension of the range of the northern population (or vice versa) or whether 2) canaster and riley: are representative of extensions of formerly isolated populations which are now meeting and intergrading in the middle ground of the southeast. The possibility of stabilization of genotype as a result of a narrowing front as the continental population pushed south into Florida must be considered. The narrowing would not become pronounced until the species moved down onto the Coastal Plain. It furnishes no expla- nation for the differentiation of canaster and erythrophthalmus, whose line of junction is not appreciably narrowed. The junctions between canaster and rileyi, and riley2 and alleni are successively narrower. It is possible to visualize, particularly in the case of alleni, stabilization taking place as a result of this radical narrowing. P. e. alleni, however, is not the most perplexing of the races here considered. In many respects, as already noted, this form shows the effects of environmental selection in agreement with several of the ecological rules surveyed. In connection with the second proposal, a review of some of the geological events which occurred in the Cenozoic is to the point. Cooke (1945: 3) concludes that the Floridian Plateau has always been a part of the continental mass as distinguished from the deep sea. There appears to be little question that throughout Tertiary time peninsulas or islands of varying shapes and sizes existed in the present area which constitutes Florida. There certainly have been periods during pre-Oligocene epochs when all of this area was under water, but the evidence available seems to indicate the possibility that several island areas may have been above water ever since that time. White (1942: 29-47, figs. 5-9) discusses the history of this area based on evidences from geology and paleogeography. Much of his infer- mation comes from a vertebrate fossil deposit in Gilchrist County, Florida, near Bell. He concludes that following the withdrawal of the Eocene seas, there was a series of crustal movements which resulted in the formation of the Central Florida Dome. At the same time there was a down-warping of the strata across the northern end of the plateau to form the Okefenokee Trough. During the Lower Oligocene the sea invaded the Okefenokee Trough only far enough to form a large bay at either end, with a pedunculate land mass extending southeastward from the mainland of North America. 338 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY During Middle Oligocene (Marianna time), with further crustal movements occurring, there was a deepening of the Okefenokee Trough, and the Gulf communicated with the Atlantic across north Florida and south Georgia through straits 50 or 60 miles wide. At the close of this period of submergence there was a general withdrawal of the seas, and Florida was again connected with the mainland. The late Upper Oligocene saw the reduction of this land mass to a small island, located in what is now the northwestern part of the peninsula, roughly 150 miles from the nearest mainland. The verte- brate fossil material at Bell, Florida, contradicts Schuchert’s (1935: 231) earlier concept of the general submergence of the Florida pen- insula by the advancing Lower Miocene seas. White (op. cit.: 42) states : “During the period of time represented by the fluvial deposit in Gilchrist Co. [Lower Miocene], Florida was a limestone island cut off from the mainland by a shallow sea fifty or sixty miles wide. In Tampa time Florida was an elliptical island roughly 220 miles north-south by 100 east-west... . “Tf the structure of Florida. .......... during the Lower Miocene was at all similar to that of today the highest part of the island would have had an elevation of about 200 feet. This is not enough seriously to affect the climate. There is no reason to suppose that the climate was very different then than now.” Throughout the Miocene there was further withdrawal of the seas. It appears that from this time forward there has certainly been some land, in the form of large islands or a group of keys, present in central Florida. The emerging land mass, indicated as having appeared during the late Miocene, in Pliocene time became a peninsula forecasting the shape of the state today. The area south of the present latitude of Lake Okeechobee was covered by a shallow sea (Campbell, 1940: 104). During the Pleistocene, peninsular conditions were permanently established, following the several oscillations of sea level attendant to the fluctuation of the ice caps. The Pleistocene history of the extreme southeastern United States has been reviewed by Cooke (1939). He states that there were at least six fluctuations of sea level during this period — from 270 feet above, to 230 or 300 feet below the present level. The highest rise of water, attained at the beginning of the Pleistocene, produced the Brandywine Terrace. At this time all of Florida was submerged, except for a group or groups of islands located in what is now mid-peninsular Florida, at about the latitude of Tampa. A scattered chain of small islands extended northward into middle southeastern Georgia. Much of the present range of P. e. rileyi was DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 339 under water at this time. Following this high level of water, there was a fall to 230 or 300 feet below the present sea level; a rise of 100 feet above and an intermittent fall to 60 feet below; an intermittent rise to 25 feet above and a drop to an undetermined low; and a last rise to the present level. Cooke further states that there has been no crustal movement during this period, as evidenced by the unbroken beach terraces in this area created by the oscillations of sea level. The present Florida race, allen, may be a relict form. Carr (1940: 6) in his study of the relict herpetofauna of this area places the Florida relicts in two classes: ‘‘(1) those derived in situ, from living or extinct or subsequently modified ancestral stocks, either by isolation on a Pliocene island or islands (or as I believe less likely, on Pleistocene Islands), or else by ecesic isolation; and (2) those which represent the remnant of a once widespread pre-Pleistocene stock.” That birds inhabited these island and peninsular land masses there is no doubt. Wetmore (1943) has examined avian material from Pierce, Polk County, Florida, and from the Bell locality. The material from Bell has been assigned to Lower Miocene by White and contains “a peculiar shorebird of a hitherto unknown type, a dove, and a wood warbler.’ The shorebird has been designated by Wetmore as the type of an extinct family. The other two species have not been identified. The Pierce fossils have been attributed to Middle Pliocene deposits by White and consist of Gavia palaeodytes Wetmore, Diomedea anglica Lydekker and Phalacrocorax auritus (Lesson). These forms, all marine or aquatic in habits, do not necessarily indicate the existence of nearby land masses, except as the needs for breeding grounds arose. Of ad- ditional interest in connection with the present discussion of variation is Wetmore’s comment concerning his assignment of the cormorant material of Pliocene age to the modern species. He says, “‘unquestion- ably they appear to belong to this the modern species... . ”’ Certainly the possibility exists that either during Pliocene or post- Pliocene time a segment of a continental population of birds might have become isolated in this area. The barrier causing isolation does not have to be visualized as consisting of simply the straits or narrow necks of land that have existed between Florida and the continent at various times during this period. Even with the connection re- established for varying periods of time, ecesic barriers may have pre- vented re-uniting of the previously isolated segments. If these events have occurred then one might expect to find some evidence of their occurrence reflected in the present populations. In the review of the migratory behavior of the four races, it was pointed out that the movements of riley? and erythrophthalmus are much more pronounced than those of canaster and alleni. If rileyi is viewed as the 340 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY northward extension of alleni, it may be thought of as returning to its point of origin during the winter. The same may be said of erythroph- thalmus as it extended from the range of canaster northward, following the retreating glaciers of the Pleistocene. The blending of certain characters, such as iris color and tail spot are certainly not in disa- greement with such a proposal. The intensification (enlargement) of others such as bill size, wing, tarsus and toe may be due to the hybrid nature of the intermediate populations. In conformity with Allen’s Rule, if erythrophthalmus is viewed as the extension of canaster northward, it is found that the extremities in this race do show reductions in size. There is no significant difference in wing size in these forms, in contradiction to Rensch’s Wing Rule. Mayr’s Rule on migration, however, confirms this view, asdoes Gloger’s Rule on color. Riley, when compared with alleni, shows some increase in bill, tarsus and toe, in contradiction of Allen’s Rule. The increase in wing length is in agreement with Rensch’s Wing Rule, and the migratory behavior of these two races conforms to Mayr’s Rule. In the same manner, as pointed out earlier, there is evidence of conformity with Gloger’s Rule. Habitat preference, particularly in the peninsular Florida race, is of interest. P. e. alleni shows definite association with the various scrub- type plant communities found near the coasts and in the central lake region. These plant communities are, according to Laessle (1942: 96), representative of the earliest stage of a xerosere, in the area he studied in Putnam County, Florida. Throughout the state they occur on old dune areas where the soil is almost pure white (St. Lucie) or yellow (Lakewood) sand. In this pioneer association, the only. tree of any size is the sand pine (P. clausa), with a dense undergrowth of dwarfed trees and shrubs. This usually includes several oaks (Quercus geminata, Q. myrtifolia, Q. Chapmanni) saw palmetto (Serenoa repens), and in some localities rosemary (Ceratiola ericoides). From April to September the towhee is found in abundance in such habitats. The oft-mentioned tendency, noted throughout the range of the species, to move into cut-over, second-growth areas may be attributed to the physical simi- larity of such habitats to those found in pioneer plant associations. It is reasonable to suppose that such favorable habitats were abundant on the transitery island and peninsular land masses that existed during the latter portions of the Cenozoic. Under such con- ditions then many opportunities were presented for immigration of a segment of the mainland population. Several other birds have distributions that lend support to the theory of insular isolation. The Florida Jay, Aphelocoma coerulescens, is usually held to be specifically distinct from the far western members DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 9441 of its genus. Geographically far removed from its relatives, this species is best viewed as a relict form. The habitat preference of this species limits its distribution in Florida to typical scrub associations, along the coast and inland. The Florida Grasshopper Sparrow, Ammodramus savannarum floridanus, occurs only in the central part of the state. It has allied races in some of the islands of the West Indies, and also has an insular-type distribution in Central America. The discontinuous distribution of the southern races contrasts with the continuous distri- bution of the forms in the northern United States. Several other birds, including the Pine-woods and Bachmann’s Sparrows (Aimophila aestivalis aestivalis and A. a. bachmanni), Eastern and Florida Cardinals (Richmondena cardinalis cardinalis and R. c. floridana) and the Northern and Southern Crested Flycatchers (Myiarchus crinitus crinitus and M. c. boreus) show lines of junction in west Florida. It is pertinent to note that P. e. alleni, canaster and rileyt meet in this same area. In other groups of animals, endemism in the Florida peninsula is well recorded. Hobbs (1942: 12) lists 17 species of freshwater cray- fishes (Cambarinae) which he considers as endemic forms. Six of these he believes are relicts. Carr (1940: 6) states that 11 amphibians and reptiles of Florida may be either relict or isolated species. Byers (1930: 289) concludes that the initial Florida Odonata fauna was a Nearctic one, isolated by a sea barrier. According to Berner (1950: 24), there is no necessity to hypothesize that Pleistocene islands existed in Florida so far as the Ephemeroptera are concerned. He adds, however, that there is no evidence to the contrary in the distribution of these forms today. Professor H. K. Wallace, of the University of Florida, tells me that there is considerable evidence in the wolf spiders (Geolycosa) of such insular isolation. Professor H. B. Sherman, of the same institution, has called to my attention the interesting disjunctive distribution of the brown bat, Eptesicus fuscus, which has an endemic form in south Florida (Sherman 1936: 107). The geological history of the peninsula of Florida furnishes abundant evidence of the possibility of isolation occurring during the Pleistocene. Other animals, both vertebrate and invertebrate, apparently were trapped by fluctuating sea levels. Some of these forms, moreover, have never since been able to rejoin the continental stocks from which they were derived, and remain as relict forms within the present peninsula. Adams (1902) concludes that the southeastern states represent a center of dispersal from which many forms expanded their ranges, after the advance of the Pleistocene ice-caps. Variation in the eastern races of P. erythrophthalmus supports this conclusion in many respects. 342 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY It is unfortunate that Adams did not consider peninsular Florida in his study of this problem. If his view is correct, and if a segment of this population was isolated on the islands, it appears that the present trends in geographic variation have a rational basis. I do not feel, on the basis of the evidence at hand, that it is possible to come to any definite conclusion concerning this problem, as it is reflected by this species. It does appear, however, that the greater weight of evidence is on the side of isolation and subsequent merging of populations. Further it seems reasonable to suppose that the pale-eyed Florida stock was derived from living or subsequently modified forms. It does not appear to be the remnant of a once wide- spread pre-Pleistocene stock. SUMMARY A review of the historical status of thespecies Pipilo er, yihvo piahalnas from 1731 to the present is given. A statistical study of geographical variation in size of body parts and color of irides is presented. A subjective analysis of variation in plumage color is presented. On the basis of these studies recognition of four geographic races within the species is possible. For each of these races there is given: . An analytical key to identification. . A synopsis of names applied in the past. . A description of the characters by which it may be recognized. . A discussion of its habits. . A statement of the breeding range and migratory behavior. . A list of specimens examined. A detailed analysis of geographical variation in the several characters examined is presented. Non-geographic variation is described. Possible reasons for geographic variation in the species are explored and suggestions are presented in explanation of the patterns observed. Oorwhd rt ACKNOWLEDGMENTS Dr Pierce Brodkorb was helpful on numerous occasions during the preparation of the original manuscript from which this paper has been condensed. Drs. H. B. Sherman, C. Francis Byers, R. A. Ed- wards and B. B. Leavitt read and criticised the manuscript and to them Iam most grateful.Dr. Arnold B. Grobman was generous in help- ing with the statistical methods utilized. Dr. John D. Kibly checked many of the statistic] computations for accuracy. Mr. J. L. Peters was cooperative in danswering certain questions as to the fate of C.J. DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 343 Maynard’s skins that were in the Museum of Comparative Zoology. Dr. Alexander Wetmore kindly furnished detailed locality data on certain specimens. Mr. James C. Greenway, Jr. kindly furnished the transcript of Vieillot’s description of Pipilo ater. Dr. A. F. Carr, Jr. checked several literature references for me at the American Museum of Natural History. Many of my colleagues in the Department of Biology have been pleasant companions in the field. Dr. Frank N. Young, in particular, was of material assistance in helping with the preparation of study skins. To all of these people I express my deep appreciation for the services they have so kindly rendered. Miss Esther Coogle, Staff Artist of the Department of Biology, recorded the iris color of much of the locally taken material. She also prepared the original charts which are reproduced here. To Mr. George K. Reid I express my gratitude for the final preparation of the plates. Mr. Leonard Giovanolli helped in tracking down the county locality for many of the obscure place names given on the museum labels. My wife has lent considerable moral support and in addition accomplished the tiresome chore of listing the specimens examined. 344 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY LITERATURE CITED Apams, C. C. 1902. Southeastern United States as a center of geographical distribution of flora and fauna. Biol. Bull., 3(8): 115-131. ALLEN, J. A. 1871. On the mammals and winter birds of East Florida, with an exami- nation of certain assumed specific characters in birds, and a sketch of the bird faunae of North America. Bull. Mus. Comp. Zool., 2: 161-450, 5 pls. 1872. Notes of an ornithological reconnaissance of portions of Kansas, Colorado, Wyoming and Utah. Bull. Mus. Comp. Zool., 3(6): 113-183. 1878. Addenda to Coues (1878: 41). Notes on spotted birds, at Ft. Leavenworth. Bull. Nutt. Orn. Club, 3(1): 42. Amapon, D. 1949. The seventy-five per cent rule for subspecies. Condor, 51(6): 250-258. AMERICAN ORNITHOLOGISTS’ UNION COMMITTEE 1886. The code of nomenclature and check-list of North American birds adopted by the American Ornithologists’ Union (New York, Am. Ornith. Union). i-vili + 1-392. 1895. Check-list of North American birds. Second Ed. (New York, Am. Ornith. Union). i-xi + 1-372. 1910. Check-list of North American birds. Third Ed. (New York, Am. Ornith. Union). 1-431, 2 maps. 1916. Changes in the A.O.U. check-list of North American birds pro- posed since the publication of the sixteenth supplement. Auk, 33: 425-430. 1923. Highteenth supplement to the American Ornithologists’ Union check-list of North American birds. Auk, 40: 513-525. 1931. Check-list of North American Birds. Fourth Ed. (Lancaster, Pa., Am. Ornith. Union). i-xix + 1-526. Bairp, 8. F., T. M. Brewer, and R. Ripgway 1874. A history of North American birds. Land birds, 2: i-vi + 1-590 + 6, 30 pls., 3: i-vi + 1-560 + 28, 8 pls. BERNER, L. 1950. The mayflies of Florida. Univ. Fla. Bio. Sci. Ser. 4(4): 1-267, 88 figs., 24 pls., 19 maps. BRoDKoRgB, P. 1935. A sparrow hawk gynandromorph. Auk, 52: 183-184. Byers, C. F. 1930. A contribution to the knowledge of Florida Odonata. Univ. Fla. Bio. Sci. Ser., 1(1): 1-827, 19 figs., 11 pls. DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 345 Bureau, T. D. 1937. Bird life on the North Carolina coast. Auk, 54: 452-460. 1944. The bird life of the Gulf Coast region of Mississippi. Occ. Pap. Mus. Zool. La. State Univ., 20: 329-490, 1 map, 3 figs. CaLHoun, J. B. 1947. The role of temperature and natural selection in relation to the variations in the size of the English Sparrow in the United States. Am. Nat., 81: 203-228, 5 figs. CAMPBELL, R. B. 1940. Outline of the geological history of peninsular Florida. Proc. Fla. Acad. Sci., 4: 87-105. Carr, A. F., Jr. 1940. A contribution to the herpetology of Florida. Univ. Fla. Bio. Sci. Ser., 3(1): 1-118. CatesBy, M. 1731. The natural history of Carolina, Florida and the Bahama Islands. London: at the expense of the author, Vol. I: i-xii + 1-100, 100 pls. 1754. The natural history of Carolina, Florida and the Bahama Islands. Second Ed. London, Vol. I: 2 + i-viii + 1-120, 120 pls. 1771. Histoire naturelle de la Caroline, de la Florida et des Isles de Bahama. Third Ed. London: Benjamin White, Vol. 1:2 + i-vii + i-xlivy, 1-100, 100 pls., 1 map. Cazier, M. A., and A. L. Bacon 1949. Introduction to quantitative systematics. Bull. Am. Mus. Nat. Hist., 93(5): 349-388, 12 figs., 11 tabs. CookgE, C. W. 1939. The scenery of Florida as viewed by a geologist. Bull. Fla. Geol. Surv., 17: 1-118, 58 figs. 1945. Geology of Florida. Fla. Geol. Surv., Bull. 29: i-ix + 1-339, 1 map, 47 figs. Cougs, E. 1871. [Untitled footnote.] Am. Nat., 5: 366. 1878. Pipilo erythrophthalmus with spotted scapulars. Bull. Nutt. Orn. Club, 3(1): 41-42. ID ero, Ibe ete 1943. The biotic provinces of North America. Ann Arbor, University of Michigan Press, i-vii + 1-73, 1 map. Dwicut, J., JR. 1900. The sequence of plumages and moults of the passerine birds of New York. Ann. N. Y. Acad. Sci., 13(1): 73-360, 7 pls. 346 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY GMELIN, J. F. 1788. Caroli A. Linné, Systema naturae per regna tria naturae, secundum classes, ordines, genera, species, cum characteribus, differentiis, synonymis, locis. Ed. 13 (Lipsiae, impensis Georg. Emanuel. Baer), 1: part 2, 501-1032. Hosss, H. H. 1942. The crayfishes of Florida. Univ. Fla. Bio. Sci. Ser. 3(2): 1-179, 24 pls. Howett, A. H. 1913. Descriptions of two new birds from Alabama. Proc. Biol. Soc. Wash., 26: 202. 1928. Birds of Alabama. 2nd ed. Birmingham: Dept. of Game and Fisheries of Alabama. 1-384, 7 pls., 31 figs. 1932. Florida bird life. Florida Dept. of Game and Fresh Water Fish in cooperation with Bureau of Biological Survey, U.S.D.A. i-xxiv + 1-579, 58 pls., 72 figs. Husss, C. L., and A. PERLMUTTER 1942. Biometric comparison of several samples, with particular refreence to racial investigations. Am. Nat. 76: 582-589. KoeEtz, W. 1939. Three new subspecies of birds. Proc. Biol. Soc. Wash., 52: 121-122. Lagssip, A. M. 1942. The plant communities of the Welaka Area. Univ. Fla. Bio. Sci. Ser., 4(1): 1-148, 22 figs., 14 pls. Linnaegus, C. (Caroli Linnaei) 1758. Systema naturae per regna tria naturae, secundum classes, ordines, genera, species, cum characteribus, differentiis, synonymis, locis. Ed. 10 (Holmiae, impensis direct. Laurentii salvii) 1: 1-823. Maerz, A., and M. R. Pau 1930. A dictionary of color. New York: McGraw-Hill. i-vi + 1-207, 56 pls. Maynapp, C. J. 1878. The birds of Florida, with the water and game birds of eastern North America. C. J. Maynard and Co., Newtonville, Massa- chusetts. Part 5: 113-136. 1881. The birds of eastern North America; with original descriptions of all the species which occur east of the Mississippi River, between the Arctic Circle and the Gulf of Mexico, with ful] notes on their habits, ete. C. J. Maynard and Co., Newtonville, Massachusetts. Rev. ed., i-iv + 1-532, 32 pls. Mayr, E. 1942. Systematics and the origin of species. New York: Columbia Bio- logical Series, Columbia University Press. i-xiv + 1-334. DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 347 Merriam, C. H., V. Bartey, E. W. NEuson, and E. A. PREBLE 1910. Fourth provisional zone map of North America. Washington, D. C.: U.S. Bio. Surv. Miter, A. H. 1941. Speciation in the avian genus Junco. Univ. Cal. Pub. Zool., 44: 173-434, 33 figs. 1949. Some concepts of hybridization and intergradation in wild popu- lations of birds. Auk, 66: 338-342. Muprpue_y, E. E. 1937. Observations on the bird life of the middle Savannah Valley, 1890-1937. Contr. Charleston Mus. 9, i-vii + 1-61, 1 map. OBERHOLSER, H. C. 1938. The bird life of Louisiana. State of La. Dept. of Conservation, New Orleans, Bull. 28, i-xii + 1-834, 45 pls. Pearson, T. G., C. 8S. Brimuey, and H. H. Brimiey 1942. Birds of North Carolina. Raleigh, N. C.: North Carolina De- partment of Agriculture, i-xxxii + 1-416, 37 pls., 141 figs. Pupp, P. B. 1910. Annotated list of birds observed (with B. S. Bowdish’s bird photographing in the Carolinas). Auk, 27: 312-322. Ranp, A. L., and M. A. TRayLor 1950. The amount of overlap allowable for subspecies. Auk, 67: 169-183. Ripeway, R. 1901. The birds of North and Middle America. Bull. U. S. N. M. 60, Pt. I, i-xxx, 1-715, 19 pls. 1912. Color standards and color nomenclature. Washington, D. C., published by the author. i-iii + 1-43, 53 pls. Roserts, T. S. 1932. Birds of Minnesota. Minneapolis: Univ. of Minnesota Press. Vol. II, i-xv + 1-821, 90 pls., 606 figs. ScHUCHERT, C. 1935. Historical geology of the Antillean-Caribbean region. New York: John Wiley and Sons, Inc., i-xxvi + 1-811, 16 pls., 107 figs. SHARPE, R. B. 1888. Catalogue of the birds of the British Museum. 12 (Fringillidae). London, British Museum. i-iv + 1-871; 16 col. pls. SHERMAN, H. B. 1936. A list of the recent land mammals of Florida. Proc. Fla. Acad. Sci., 1: 102-128. SIBLEY, C. G. 1950. Species formation in the Red-eyed Towhees of Mexico. Univ. Cal. Pubs. Zool., 50(2): 109-194, 6 pls., 18 figs. 348 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Simpson, G. G., and A. Roz 1939. Quantitative zoology. New York: McGraw-Hill. i-xvii + 1-414, 52 figs. 1942. A standard frequency distribution method. Am. Mus. Novit. No. 1190, 1-19, 5 figs., 6 tabs. Sprunt, A., and EK. B. CHAMBERLIN 1949. Birds of South Carolina. Cont. Charleston Mus. 11, i-xx + 1-585, 34 pls., 48 photos. Sutton, G. M. 1935. The juvenal plumage and post-juvenal molt in several species of Michigan sparrows. Cranbrook Inst. Sci. Bull., 3, 1-36, 8 pls. Topp, W. E. C. 1940. Birds of western Pennsylvania. Pittsburgh: Univ. of Pittsburgh Press. i-xv + 1-710, 23 pls., 1 map. VIEILLOT, L. J. P. 1816. Analyse d’une nouvelle ornithologie, elementaire. Paris. 1-70. 1819. Nouveau dictionnaire d’histoire naturelle. Paris. 34 (8): 1-578, 8 pls. 1824. La galerie des oiseaux (etc.). Paris. 300 col. pls., 33 pls. VISHER, S. S. 1944. Freezing temperatures in the United States. Ecology, 25: 113-117, 17 figs. WEATHER BUREAU, REPORT OF THE CHIEF 1897. Report of the Chief of the Weather Bureau, U. S. D. A., House Document 42, No. 166: 1-481, 61 pls. WeErTMokg, A. 1937a. Observations on the birds of West Virginia. Proc. U.S. Nat. Mus. 84: 401-441. 1937b. Notes on the birds of North Carolina. Proc. U. 8. Nat. Mus. 90: 483-530. 1943. Fossil birds from the Tertiary deposits of Florida. Proc. New England Zool. Club, 22: 59-68, 2 pls. Wuite, T. E. 1942. The Lower Miocene mammal fauna of Florida. Bull. Mus. Comp. Zool., 92(1): 1-49, 14 pls. Witson, A., W. JARDINE, and C. L. J. BoNAPARTE 1832. American ornithology. London and Edinburgh. (2) i-vii + 1-390, 33 pls. WorrTHINGTON, W. W., and W. E. C. Topp 1926. The birds of the Choctawhatchee Bay region of Florida. Wilson Bull., 38 (old series) (4): 204-229. DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 349 woon le” cores 8 LEGEND e@ LOCALITIES LIMITS OF SAMPLE AREAS Map 1. Localities and limits of sample areas from which breeding specimens were examined. Samples 1, 2 and 3 represent P. e. erythrophthalmus; 4, repre- sents P. e. canaster; 5, represents P. e. rileyi and 6, represents P. e. allent. 390 ) == i MN / cies of Pipilo erythrophthalmus. DICKINSON: GEOGRAPHIC VARIATION IN RED-EYED TOWHEE 351 Map 3. Ey an ® LE Hr | athens Be ay Ke, il ! BBLS peaDAe ae ae a se Ls S Ce, Ser é agreed Va ORANGE YELLOW hs mi Geographic variation in iris color in the southeastern United States. 352 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY olleni Oct.-April rileyi Nov.-Mar conoster Nov.-Mar. erythrophthalmus Nov.-Moar, WINTER RANGE —-—7— conoster —1— x— erythrophthalmus —e——alleni Map 4. Winter ranges. Note that erythrophthalmus and rileyi move farther south in peninsular Florida than does canaster. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vo. 107, No. 6 CYATHASPID FISHES FROM THE VERNON SHALE OF NEW YORK By R. H. FLower anp R. WayYLAND-SMITH With Wiear PLares CAMBRIDGE, MASS., U. S. A. 12 Jey JE IN) 201, ID) Te CO) 1k AP TSC Is, CWS) 1m WW) IY October, 1952 PUBLICATIONS ISSUED BY OR IN CONNECTION WITH THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE BULLETIN (octavo) 1863 — The current volume is Vol. 107. BREviIoRA (octavo) 1952 — No. 9 is current. Mewmorrs (quarto) 1864-1938 — Publication was terminated with Vol. 55. JOHNSONIA (quarto) 1941 — A publication of the Department of Mollusks. Vol. 2, no. 30 is current. OccASIONAL PAPERS OF THE DEPARTMENT OF MOLLUSKS (octavo) 1945 — Vol. 1, no. 16 is current. PROCEEDINGS OF THE NEW ENGLAND ZOOLOGICAL CLUB (octavo) 1899- 1948 — Published in connection with the Museum. Publication terminated with Vol. 24. These publications issued at irregular intervals in numbers which may be purchased separately. Prices and lists may be obtained on application to the Director of the Museum of Comparative Zoédlogy, Cambridge 38. Massachusetts. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE VoL. 107, No. 6 CYATHASPID FISHES FROM THE VERNON SHALE OF NEW YORK By R. H. FLower anp R. WaAyYLAND-SMITH Wits Erecut PLaTEs CAMBRIDGE, MASS., U.S. A. PRINTED FOR THE MUSEUM October, 1952 mata iD oN any “as Ni * ‘ ‘ fi i De ‘ , . ¥ on La i un i . b Vy ‘ i ey cies ’ cain No. 6. — Cyathaspid Fishes from the Vernon Shale of New York By R. H. Frower and R. WayLanp-Suirs Jyathaspid fishes are rare elements in Silurian faunas. Only three occurrences have been previously noted in North America, that of Diplaspis acadica Matthew from the Upper Silurian of New Brunswick, of Palacasys americana and P. bitruncata described by Claypole (1885) from the Upper Silurian of Pennsylvania, and of Cyathaspis wardelli (Ruedemann) and C. van ingeni Bryant from the Shawangunk forma- tion of southeastern New York. The present occurrence, in the Vernon shale in central New York, although it has yielded only eight identifiable specimens, compares very favorably in preservation and completeness of material with previous American finds. It is regrettable that the nature of the exposures makes it extremely improbable that more material can be obtained, particularly since the presence of several species in the asso- ciation makes the matching of dorsal and ventral plates an extremely inferential matter. ACKNOWLEDGMENTS The writers are indebted to Dr. A. S. Romer and Dr. J. C. Bradley for advice concerning problems of nomenclature, and to the following for aid in a search for the original material of Claypole’s Palacasms: Dr. Bobb Schaffer, Dr. David Dunkle, Dr. J. O. Fuller and Dr. W. C. Kraatz. OCCURRENCE The Vernon shale, of the Upper Silurian of New York, overlies the Pittsford shale, and underlies the Camillus shales, which are in turn succeeded by the Bertie waterlime. These formations constitute the Salina group, which is overlaid by higher Silurian limestones, the Waterlime group, consisting of the Cobleskill limestone, the Rondout and Manlius limestones. The Vernon shale, characterized by the presence of red and green beds, has yielded a few sparse faunas, but has come to be considered so proverbially barren that the suggestion has been made that it is, at least in a large part, composed of wind blown loess. 356 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Two previous occurrences of fossils have been reported. The first was a recurrent Pittsford fauna, sparsely distributed in shales of the lower part of the Vernon near Pittsford, New York, from which Ruedemann (1920) reported a fauna of sixteen species. Later Eaton (1924) reported a fauna, from which he described several new species, from Elbridge township, fourteen and one-half miles northwest of Syracuse, New York. The cyathaspid fish described in the following pages came from a layer of grey shale 1 ft. 9 in. thick, in the middle of the Vernon shale. This layer has been found exposed only in two small streams, tribu- taries of Oneida Creek, about two miles southeast of Kenwood, New York. The complete section, together with the invertebrate faunas, will be described upon another occasion. It will suffice here to mention that this thin layer has yielded several hundred specimens of inverte- brate fossils, including abundant pelecypods (Pterinea, Modiolopsis and Nuculites), abundant but extremely fragmentary eurypterids of the genera Hughmilleria and Pierygotus, ostracods, several poorly pre- served nautiloid cephalopods, including one brevicone and several orthoconic genera, brachiopods representing the genera Lingula and Camarotoechia, an annelid jaw, and a single large siphonophore (Flower and Wayland-Smith, 1947). The invertebrate elements are very closely allied to those of the underlying Pittsford shale and also to those of the younger Bertie waterlime. The type specimens described in this paper are deposited in the collections of the Museum of Comparative Zoology at Harvard. CLASSIFICATION OF CYATHASPID FISHES The history of the study of the Cyathaspida has been summarized by Kiaer and Heintz (1935). In an earlier posthumous paper edited by Heintz, Kiaer (1932) presented a rather elaborate classification of the group. A generation ago, all dorsal plates of these rare and usually fragmentary fossils, were assigned to the genus Cyathaspis. After considerable controversy it came to be recognized that a second genus, Scaphaspis, was nothing more than the ventral plate of the same animal. It is a far cry from this to the classification of Kiaer, who recognizes the suborder Cyathaspida as divided into two tribes, the Poraspidei and Cyathaspidei, and further divided into seven families and fourteen genera. Unfortunately, problems of recognition and of the legal availability of some of the names used in Kiaer’s work, are raised by the fact that some of the species which are new are only listed, while others also new, are listed and illustrated, but not de- scribed. This unfortunately applies to the type species of several of FLOWER AND WAYLAND-SMITH: CYATHASPID FISHES 357 the genera erected in this work and to the genus Eoarchegonaspis Kiaer and Heintz 1932. Under the present International Rules of Zoological Nomenclature, no species or genus described after 1930 shall be considered valid for the purposes of nomenclature and priority unless it is accompanied by either a description showing how it differs from other species or genera, or by a bibliographic reference to such a description, and in the case of a genus, is accompanied by a clear and unequivocal designation of a genotype. Later work has resulted in the description of few additional genera and species, mainly those of Bryant and Woodward, but has produced marked and confusing changes in nomenclature. Many of the generic names used by Kiaer (1932) were found to be junior homonyms. New names were proposed by Strand (1934), by Whitley (1940), and by White and Moy-Thomas (1940-1941). Whitley, and White and Moy-Thomas unfortunately proposed new names for the same generic group at about the same time. The names of White and Moy-Thomas are more widely known, but those of Whitley clearly have priority and must therefore be used. Recent trends in classification have been to employ as families the groups which Kiaer recognized as tribes, and to ignore the finer family divisions as employed by Kiaer. This more conservative treatment is a great improvement, for the much finer family groups which Kiaer used involve both serious problems of nomenclature, as well as serious morphological problems in their recognition. Further, critical analysis of the genera indicates that even the use of two families may imply a more distinct division within the Cyathaspida than actually exists- Moy-Thomas (1939) employed the families Palaeaspidae and Cyat- haspidae. The use of Palaeaspis as the type of a family, is unfortunate as the genus is inadequately known. Romer (1945) used the families Poraspidae and Cyathaspidae, as employed in the present outline, again treating them as two of the five families of the Heterostraci. The relationships of the Poraspidae and Cyathaspidae, as noted above, are much closer than a casual reading of Kiaer’s paper would lead one to believe and, in fact, serious difficulties attend any attempt to draw a clear line between them. But there is no close connection between the cyathaspids on the one hand, and either the pteraspids or the drepanaspids on the other. It seems that it would be better to either recognize the two families as members of the suborder Cyatha- spida, or else to suppress the family Poraspidae, and recognize only a single family, the Cyathaspidae. In an attempt to determine the generic affinities of our material, it has been necessary to analyse the previous work rather closely. A short account of the results is presented here, partly to clarify the 358 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY rather intricate nomenclatorial problems, and partly to summarize the present additional morphological information. Particularly relevant to our present problem has been a comparison with the Silurian cyathaspids previously described from America. Of these, the genus Diplaspis Matthew proves to be still so poorly known that close com- parison is not possible, and its exact position in the modern classifi- cation of cyathaspids is extremely dubious. Palaeaspis Claypole is still not adequately known, in spite of the work of Bryant (1926). In the same work, Bryant described additional cyathaspids from the Shawangunk formation of southeastern New York as Cyathaspis wardelli (Ruedemann) and C. van ingeni Bryant. Kiaer has reunited’ the two species, we think mistakenly, and erected for them the new genus Eoarchegonaspis, for which no description is vouchsafed. As shown below, under discussion of the genus, the original suite of speci- mens of C. wardelli consists of inadequate fragments of several cyatha- spid plates, representing an indeterminate number of species and certainly including representatives of three different genera. What Bryant described as C. wardelli cannot be recognized among Ruede- mann’s suite of illustrated types. However, in that materia! there is at least one dorsal shield of Cyathaspis van ingeni Bryant. In any case, the genus Eoarchegonaspis is a nomen nudum, being without a de- scription. None of the fish in the Vernon shale assemblage is conspe- cific with the Shawangunk materials. CYATHASPIDA Kiaer The Cyathaspida are Heterostraci characterized by single dorsal and ventral plates, and a pair of smaller branchial plates. Additional hard parts consist only of large scales which covered the posterior part of the body, and small lateral plates which are rarely found and are poorly known. Kiaer divided the Cyathaspida into two “‘tribes’’, and the tribes into a number of families. This is unfortunate, as tribes are properly a category beneath families and subfamilies in rank. Romer (1945) in his summary of classification of fossil vertebrates, does not employ the families of Kiaer, which are too finely drawn, but instead employs the “tribes”? Poraspidei and Cyathaspidei, as families Poraspidae and Cyathaspidae. These families are grouped with others in the Hetero- straci without use of the term Cyathaspida to differentiate them from other members of the order. We are employing the term Cyathaspida here because these forms constitute an easily recognizable group, but appalling difficulties were encountered in trying to differentiate be- tween the Poraspidae and Cyathaspidae. Indeed, our own material FLOWER AND WAYLAND-SMITH: CYATHASPID FISHES 359 and the present descriptions and illustrations, lead to the conclusion that these families intergrade to such an extent that their differenti- ation may have very little merit. The Poraspidae are defined as having the dorsal plate undivided, and a lateral line system which is well developed, and consists usually of continuous series of canals within the dorsal and ventral plates. The Cyathaspidae are defined as having the dorsal plate differenti- ated into a central disc, a rostrum, and two lateral plates, but the lateral line system is incomplete or wanting. Actually, neither the differentiation of the dorsal plate into distinct areas nor the development of the lateral line system serve to divide the Cyathaspida into two distinct families. Impressions of the interior of the dorsal plate of some poraspids show branchial regions fairly well set off from a central disc. In the genera placed in the Cyatha- spidae, the distinction is reported as complete in Traquairaspis, as poor in Tolypelepis, clear externally in Cyathaspis, and variable in Archegonaspis. It is apparently clear in the invalid genus Eoarche- gonaspis Kiaer and Diplaspis Matthew. This may be more apparent than real, for these genera are to date represented in the literature only by extremely diagrammatic outline drawings. In the extant illustrations of the Cyathaspida there is no evidence of a clear sepa- ration of the dorsal plate into four discrete plates; rather the differ- entiation is produced mainly by the surface pattern. Indeed, one of our most vexing problems in the description of our new material has been the generic position of a specimen which resembled in some respects Anglaspis of the Poraspidae and in others Archegonaspis of the Cyathaspidae. So close are these two genera, that the decision was finally made on the basis of the texture of the surface pattern rather than on criteria mentioned in previous descriptions of the families and genera. On the other hand, differentiation of the two anterior lateral plates of the dorsal plate, in Vernonaspis, is clear on the interior but not on the exterior. Separation of rostral and lateral areas is not indicated in the suite of syntypes of Cyathaspis wardelli Ruedemann, the type species of Eoarchegonaspis. It is necessary to conclude that the differentiation of the dorsal plate is not a good basis for the recognition of two families in the Cyathaspida. The lateral line system is no better as a family criterion. It is a feature difficult to demonstrate, frequently impossible to detect unless there is abundant and well preserved material. It can be seen best in specimens which have been exfoliated, or where the canals are so large that they have been accentuated as depressions on specimens which have been subjected to the slight crushing which accompanies com- paction of the sediments. It must be remembered that the Poraspidae 360 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY from which the lateral lines have been described, are known from the only source of abundant and well preserved cyathaspids, the Spitz- bergen material. Genera assigned to the Cyathaspidae, for which in general no lateral lines are reported at all, are known from associations yielding fewer individuals, and apparently, individuals in a much poorer state of preservation. There are certainly grounds for regarding with suspicion the assumption that the lateral line system is wanting in the majority of the genera of the Cyathaspidae simply because it has not been reported. Its value as a family character is therefore doubtful at the best. It is true that in some poraspids — those genera grouped currently under the Poraspidae — the lateral line system forms a complete and sometimes an elaborate series of canals. There are also others in which the canals in the inner layers of the plates form an incomplete and disconnected system of tubes. One such specimen (Pl. 3, fig. 6; Pl. 7; Pl. 8, fig. 3; text fig. 2), consists of a ventral plate. The interior is smooth, but shows a faint pattern of grooves. Upon removing the inner surface by etching (PI. 7), a disconnected series of tubes of the lateral line system was found. It should be noted that this form, assigned tentatively to Archegonaspis of the Cyathaspidae, has the discontinuous canals which are said to characterize that family. However, in Vernonaspis (Pl. 1; Pl. 2, fig. 8), there is apparent a series of grooves similar in aspect to a lateral line system. Closer study shows that they are so askew in relation to the symmetry of the organism, that they are more probably wrinkles. Further, if this should prove to be a lateral line system, it would be one comparable with that of the Poraspidae, but this genus, with its distinct rostrum and two anterior lateral plates, is in other respects closer to the Cyathaspidae. It is necessary to conclude that the boundary between the Pora- spidae and Cyathaspidae is, in the present state of our knowledge, too tenuous to be recognized. The problems of morphology, taxonomic recognition of genera, and their validity from a purely legalistic and nomenclatorial viewpoint, present such intricate problems, that they may be best summarized by a discussion of the individual generic groups. These are arranged, following the system of Kiaer (1932), with the addition of a few subsequently described genera. PORASPIS Poraspis Kiaer 1932 (= Holaspis Lankester 1873, not Gray 1863) is based upon Holaspis sericea Lankester. Dorsal plate entire, surface with essentially longitudinal markings of dentin ridges. Lateral line FLOWER AND WAYLAND-SMITH: CYATHASPID FISHES 361 system represented externally by rows of pores, and well developed. Eye notches clear, branchial and postbranchial regions separated by a faint lateral constriction; posterior margin extended into a blunt median process. Ventral plate truncate in front, similar to dorsal plate in longitudinal markings. Lateral line system less well developed, with two main lateral branches. The genus is best known from the beautiful series of species from Spitzbergen described and admirably illustrated by Kiaer and Heintz (1935). Eleven species are listed by Kiaer (1932) including those later described by Kiaer and Heintz (1935). Internal molds of the dorsal plate of Poraspis are quite similar to those of other genera, but Poraspis shows the following distinctive features: the first two pairs of branchial impressions are well anterior to the semicircular canals and prominent, instead of faint; lateral areas are distinct and bear lateral branchial impressions. In general, the semicircular canals are more prominent than the pineal body. HoOMALASPIDELLA Homalaspidella Strand 1934 (= Homalaspis Kiaer 1932, not Rein- hardt 1860; = Homaspis Kiaer and Heintz 1935, not Foerster 1868 or Skuse 1888) contains only the type species, H. nitida (Kiaer). It is differentiated from Poraspis by the polished surface and the very narrow grooves separating the dentin ridges. The lateral line system of the venter is more advanced, showing anterior median lateral commissures as well as the marginal lateral commissures. These two genera are closely allied, but do not seem distinct enough from the following forms, particularly Anglaspis, to justify their sepa- ration into a family by themselves as was done by Kiaer (1932). AMERICASPIS Whitley (1940) proposed the name A mericaspis to replace Palaeaspis Claypole 1885, not Gray. Claypole proposed this generic name for two species from the Upper Silurian of Pennsylvania, P. americana and P. bitruncata, the former being the genotype. He later stated (1892) that he considered P. bitruncata to represent a ventral plate of P. americana. His descriptions are accompanied by drawings, giving for P. americana the shape of the plate and something of its surface pattern but only the outline of P. bitruncata. Bryant (1926) restudied the species on the basis of material in the Princeton University collections, but without having access to the type specimens. He re-illustrated the species with exceptionally poor photo- graphs, presented revised descriptions, and dismissed with little com- ' 302 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY ment Claypole’s suggestion that P. bitruncata was the ventral plate of P. americana. It is unfortunately quite evident that what Bryant called P. bitruncata does not agree at all closely with Claypole’s original description or outline drawing of that species, but was almost certainly the dorsal plate of quite another form. Indeed, Claypole’s original drawing of P. bitruncata quite evidently does have the form of a ventral and not a dorsal plate, and is quite similar in shape to the specimen here described and illustrated as the ventral plate of Vernonaspis leonardi (PI. 2, figs. 6-7). Unfortunately, an effort to locate the Claypole types proved fruit- less. This material was deposited in the collections of Buchtell College, Akron, Ohio, where Dr. Claypole taught from 1884 to 1897. The building which housed these collections was destroyed by fire in 1899, and there is little doubt that the types were destroyed at that time. Dr. Walter C. Kraatz assures me that the material is not in the collections of the University of Akron, which houses the extant records of Buchtell College; the collections were apparently completely destroyed. Kiaer (1932) treats this genus under the name Palaeaspis, and apparently bases his conclusions upon Bryant’s redescriptions of the two species. He considers that P. americana should be united with his genus Dinaspis, and P. bitruncata with the genus Poraspis, accepting both as dorsal shields. He then calmly proceeds to erect for it the family Palaeaspidae. Actually, so little is known of the two species which constitute the genus Americaspis (= Palaeaspis Claypole) that restudy is necessary if the genus is to be considered in relation to the present classification of cyathaspid fishes. In view of the destruction of the type material, careful restudy of the problem is required to determine whether it will be possible to re-establish these species and the genus, on the basis of neotype material. Otherwise it will be necessary to admit that P. americana and P. bitruncata cannot be recognized with certainty. Therefore the genus Americaspis (= Palaeaspis) is regarded as valid from the nomenclatorial point of view, but one which is so inadequately known that no species can be referred to it. Even its type species cannot be recognized. Judging from the association of the Vernon shale, and also those in the Shawangunk and Longwood formations of southeastern New York, it will be remarkable if only one genus is present in the association of the Upper Silurian shales of Pennsylvania, which yielded the original material of Palaeaspis, as well as all subse- quent specimens attributed to the genus and its two species. FLOWER AND WAYLAND-SMITH: CYATHASPID FISHES 363. DINASPIDELLA Strand Dinaspidella Strand 1934 was proposed to replace Dinaspis Kiaer 1932, not Leonardi 1911. The generic group may be a valid one, and it is therefore discussed here. However, both of the species which Kiaer places in it are nomina nuda, since neither has been described. One is only listed; the other, the type species of the genus, is listed and illustrated but not described. In this and in the following genus the ventral plate is five-sided, broadly truncate in front, widest near the posterior margin, which is produced behind into a broad blunt point. In Dinaspidella two pairs of longitudinal commissures mark the lateral line system of the dorsal plate. IRREGULARASPIS Zych Irregularaspis Zych was described only in Polish, which did not help in its general recognition in the least. White and Moy-Thomas (1940) state that it is the same genus as Dictyaspis Kiaer 1932. Dictyaspis is another genus which Kiaer described without any valid species. He refers to it three species, again illustrating but not describing the genotype. The genus is characterized by a ventral plate similar in form to that of the preceding genus. On both the dorsal and ventral plates the lateral line system is enlarged into a complex reticular network, quite unlike that of any other cyathaspid. It should be noted that the family Dinaspidae of Kiaer 1932, erected for Dinasps and Dictyaspis, is not legally available since Dinaspis rests upon a species which is a nomen nudum. In any case, since Dinaspis is a junior homonym, a new family name would have to be proposed if there is any point in doing so. While the striking similarity in the shape of the ventral plates suggests a close relationship between these genera, it does not seem that a group embracing them should have family rank. ANGLASPIS Jaekel This genus is defined for the first time by Kiaer, who also erects a family, Anglaspidae, for it. This does not seem necessary. Anglaspis, based upon Cyathaspis mcecullought Woodward, has been more ex- tensively restudied and re-illustrated by Wills (1935). Kiaer considers as one of the significant features of this genus the fact that the dorsal plate is relatively flat, the ventral plate very strongly arched. The lateral line system is very similar to that of Poraspis, with two pairs of longitudinal commissures in the dorsal plate, two series of short 364 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY transverse pairs of lines on the venter, and a pair of longitudinal commissures at the extreme sides. Dentin ridges are well defined, prominent, in the main, forming a longitudinal pattern. An anterior triangular area is a conspicuous feature, embracing the rostral area and terminating in a point just behind the pineal body. The margins ‘of this area correspond to a part of the lateral commissures of the lateral line system. In a part of this area a faint transverse line differentiates the rostrum on the exterior, and laterally the lateral line system is again responsible for differentiation of two lateral areas from a central disc. In this respect, the genus Anglasps approaches the Cyathaspidae to such an extent that it might as well be considered a member of that family (= tribe Cyathaspidei as used by Kiaer) instead of the Poraspidae (= Poraspidei of Kiaer). The shape of the dorsal shield is somewhat varied, but the eye notches are developed, the sides are more convex than in Poraspis, and the branchial and postbranchial regions more poorly differentiated. The posterior margin is pointed at an angle sharp enough that the lateral angles at its sides are obtuse and sometimes obscure. Impressions of the interior of the dorsal plate resemble those of Poraspis, but the pineal body is more prominent, the mesocephalon more obscure; the anterior branchial grooves are more obscure, the anterior portion more transverse, and the nasal pits more pronounced. The ventral plate is quite similartothat of Poraspis; all species show much finer texture in the surface markings than does Poraspis. It should be added that the branchial plates are well known as are the posterior body scales, on the basis of which Kiaer (1932) has reconstructed the entire animal. In addition to the genotype, Kiaer lists four new species and one variety, which have not yet been described. CTENASPIS Kiaer 1930 This genus is set apart by its striking form and ornament. Anterior part of dorsal shield transverse, sides rounded, widest near posterior margin, which is pointed behind, and joins the convex sides without a definite angle; main part of shield pustulose, sides with definite spines at the margins. Ventral shield more sharply transverse in front, posterior margin somewhat more distinct from sides. Lateral line system well developed, somewhat more advanced than in Poraspis. Two species from Spitzbergen, C. dentata and C. cancellata are all that are known. The former is the genotype. Kiaer (1930) erected a family for the reception of this genus alone. FLOWER AND WAYLAND-SMITH: CYATHASPID FISHES 365 CYATHASPIS Lankester The genus Cyathaspis, which formerly embraced all cyathaspid fish, is now considered to be confined to the type species, C. banksi Huxley and Salter. Dorsal shield a broad oval, longer than wide, narrowly rounded in front, and narrowing also behind; without a clear boundary between the lateral and posterior margins. Main part of plate with dentin ridges forming a concentric pattern, in which a prominent feature is a depression separating the rostral and lateral areas from the central disc; the separation of the rostrum from the lateral plates is more obscure. A low narrow raised ridge occurs on the posterior part of the dorsal shield. The ventral shield, originally described as Scaphaspis truncatus Lankester, was the center of some controversy before its identity with C. banksi was generally admitted (see Kiaer and Heintz, 1935, pp. 32-33); and as stated by Woodward (1891, p. 172), it is not at all certain that all of the ventral shields placed under this name belong to the same species. Kiaer (1932, pl. 8) has presented beautiful illustrations of dorsal and ventral shields. Differ- ences in the character of the ornament are shown here that remind one of Woodward’s statement, but may be due to different conditions of preservation, for the dorsal shield appears to have the internal features impressed on the exterior. Nothing is known of the lateral line system. ARCHEGONASPIS Jackel This genus, proposed by Jaekel (1927) without any very clear diagnosis, is based upon Cyathaspis integer (Kunth, 1872). It is better known from Kiaer’s (1932) definition. Dorsal plate with rostral and lateral areas well differentiated on the basis of the surface pattern; dentin ridges on rostrum transverse, arranged in two coalesced whorls, one centered near each side. As in Anglaspis, the rostrum forms part of a triangular area of the surface pattern which terminates in a point just behind the pineal body. Though nothing is stated concerning this area, its boundaries are in all probability a part of the lateral line system, which is not otherwise evident from extant descriptions or illustrations. Additional pores are present on the dorsal surface, sug- gesting a more extensive lateral line system. Ventral plate sharply truncate in front, broadest near posterior margin which is convex, apparently not greatly produced, and with only a short production of the center into a point, or else transverse. The posterior margin of the dorsal plate is faintly convex, but nearly transverse, and not pointed. Kiaer (1932) lists four species, all valid, except that he fails 366 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY to state on what basis he erects a species A. lindstrg@mi for the specimen which Lindstrgm (1895) assigned to Cyathaspis? schmidti. This is, however, probably the best known of the species, from the illustrations of Lindstrgm and Kiaer. This species is from the Upper Silurian of Skaane, Sweden. The other four are from erratic blocks in northern Germany. The interior of the dorsal shield presents no features by which it can be separated with certainty from Anglaspis, other than the more transverse condition of the posterior margin. Indeed, the genera are quite similar. Kunth’s (1872) illustrations of the type species indicate that, as in Anglaspis, the ventral shield is much more strongly arched than the dorsal shield; Kiaer states that the body scales are imperfectly known but also show a similarity with Anglaspis as far as can be told. These genera, then, appear to be separated mainly on matters of degree; in Anglaspis the ribbing of the surface is coarser; the posterior margin is more produced, the lateral line system is better developed, or perhaps only better observed, and the rostral and lateral areas are less distinctly set off from the central shield. One could wish for clearer differences. EOARCHEGONASPIS Kiaer 1932 Original description: “This new genus is established for the two forms known from America and described under the names of Cyathaspis wardelli Bryant and Cyathaspis van ingeni Bryant. “The present writer regards these two forms as representatives of the angusta and lata forms, and therefore proposes to retain only the name LHoarchegonaspis wardelli Bryant. After the descriptions of Bryant it must be regarded as being closely related to the Cyathaspis and Archegonaspis, but it shows some features which make it necessary to establish a new genus for it. All the material of this form is known from the beds of Red shale in the part of the Yerguard Quarzit, Orange County, N. Y. These beds probably belong to the Medina formation.” This genus, happily, need not be recognized under the International Rules of Zoological Nomenclature. Kiaer states that in his opinion the genus is different, but does not say what the features are which impelled him to erect this new genus. There are several minor errors in this description. The type species is to be attributed not to Bryant, but to Ruedemann, who described it as a species of Anatifopsts Barrande, a genus supposedly belonging to the cirripedes. FLOWER AND WAYLAND-SMITH: CYATHASPID FISHES 367 The material came from a formation proposed by Van Ingen, the Guymard quartzite, of Orange County, New York. Today this formation name is not generally used, and these beds are included in the Shawangunk conglomerate which grades upward into the Long- wood shale. These beds have been considered by many to be of Medina age, Lower Silurian, and therefore older than the other American Silurian cyathaspids. The supposedly greater age may have been at least a strong influencing factor in impelling Kiaer to erect a new genus for this material. The greater age of these beds is not, however, proved. The Medina age determination rests mainly upon the presence in the Shawangunk formation of the trail Arthrophycus, which is, to be sure, the abundant and conspicuous feature of the Medina sandstone in western New York, but is hardly an adequate or a reliable criterion for age determination. Overlying the Shawangunk and the almost equally barren Longwood shales, are fossiliferous beds of upper Cayugan age, the equivalent of the Cobleskill limestone, and it is therefore possible for these clastic beds to be as young as the Salina, lower Cayugan. It has been suggested that the Shawangunk may be Clinton rather than Medina in age, but there is no real evidence to oppose the view of Clarke (1907) or Hartnagel (1907) that the entire sequence may be no older than the base of the Salina group. If so, the Shawangunk fish are not materially older than those known from the Upper Silurian of Pennsylvania. In the hope of clearing up the difficulties surrounding the recogni- tion of Eoarchegonaspis and its two species, or two forms according to Kiaer, the original types of Anatifopsis wardelli Ruedemann and supplementary material from among which they were selected, were examined. It indicates, in brief, that these forms are cyathaspid fish, but too fragmentary for definite specific recognition. There are obviously several species and even several genera in the suite of type specimens. Further, among Ruedemann’s type material there are some dorsal shields which are similar to what Bryant called Cyathaspis van ingeni, but not a single specimen which can be identified with what Bryant called Cyathaspis wardelli. The type specimens are described briefly as follows: 1. New York State Museum, no. 9612. A strongly convex plate, sides subparallel, the whole rather strongly curved from one side to the other, surface with rather coarse longitudinal lineation, one end bluntly pointed, the other obviously incomplete. This is probably an incomplete ventral shield, but it might be also a part of a rather large branchial plate. Its generic position cannot be determined in relation to the present rather exacting classification. 368 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY 2. N. Y. State Museum no. 9613. A small elongate plate, slightly curved and longitudinally striate. One end is drawn to a blunt oblique point; the other is broken. This is certainly a branchial plate. 3. N. Y. State Museum no. 9614. An incomplete plate, rounded at the complete end, and asymmetrical; ridges are mainly longitudinal, but concentric around the rounded border. This is obviously one end of a branchial plate. 4. N. Y. State Museum no. 9615. This is the anterior end of a plate, showing the ridges centering about two anterior lateral areas, the front transverse, truncate, but rather obscure. This is, by its shape and surface markings, a ventral plate similar to those known in Anglaspis and Archegonaspis. The test is remarkably thick on this specimen in comparison to its size. 5. N. Y. State Museum no. 9616 is the impression of the interior of a ventral plate, truncate in front, strongly produced behind, and suggestive of the (invalid) genus Dinaspidella and also to some extent, Irregularaspis. Unlike most of the other ventral plates, it is extremely flat. 6. N. Y. State Museum no. 9617 (PI. 3, fig. 7). is a crushed dorsal plate, preserving the anterior end. The front is bluntly pointed; there are good eye notches. The extreme anterior part of the head has a remarkably thick shield, and shows a smooth surface. There is no clear separation of rostral and lateral plates. Ridges, which appear behind the anterior margin, are extremely fine and faint, and quite closely spaced. This appears to be quite similar to the Cyathaspis van ingent of Bryant (1926). 7. N. Y. State Museum no. 9618. (PI. 3, fig. 1.) A slender curved plate, with a definite ridge on one side, fractured transversely at several points, with longitudinal markings. This is quite plainly the narrow posterior end of a branchial plate. 8. N. Y. State Museum no. 9619 is a broader triangular plate, com- parable in form, but not in surface marking to the unidentified tri- angular plate figured and described here from the Vernon shale. 9. N. Y. State Museum no. 9620. A plate, strongly curved from one side to the other, incompletely exposed, sides subparallel, very thin, with fine linear markings. 10. N. Y. State Museum no. 9621. A strongly convex plate in- completely exposed, of somewhat oval outline, suggesting the strongly convex ventral shields of Archegonaspis and Anglaspis. No. 9615 may be a smaller conspecific individual. 11. N. Y. State Museum no. 9622. A small piece of a ventral or dorsal plate, weathered, showing the dentin and cancellous layers FLOWER AND WAYLAND-SMITH: CYATHASPID FISHES 369 with typical cyathaspid structure. It is too poorly preserved for any attempt at generic or specific determination. 12. N. Y. State Museum no. 9623 is the fragmentary impression of the outer surface of a ventral or dorsal plate, showing in addition to the linear markings faint pustules, smaller than those of Tolypelepis and Traquairaspis, and not closely comparable to those of Vernonaspis. A good drawerful of material has failed to yield any specimens from the original locality that are materially better, but does serve to indi- cate more strongly that several genera are present. There are, among such specimens, some with broad flat dentin ridges separated by narrow grooves, suggestive, some of Poraspis and Homalaspidella, others of Anglaspis. None of this material contains anything at all similar to the form which Bryant called Cyathaspis wardelli. One such specimen was found (Pl. 2, fig. 4,), but it is from another collection, from the red Longwood shales which lie above the Shawangunk conglomerate. This is the anterior end of a broadly rounded plate, with very faint obscure eye notches. The surface bears dentin ridges which are irregularly arranged, tending to form small whorls over the anterior portion, and more linear, but still irregular and anastomosing, farther back. The pineal body is indistinct, there is no clear separation of the rostrum from the central disc, but there is some indication of a distinc- tion of the lateral plates on the basis of surface irregularities, but this may be false because the whole surface pattern is so irregular, and quite probably variable from one specimen to another. The surface is folded into small wrinkles, a condition which was plainly not original. This form poses quite another problem. It is quite similar to what Bryant figured as Palaeaspis americana (1926, Pl. 1, fig. 1) as well as to his Cyathaspis wardelli (his Pl. 4, fig. 1,) but quite unlike a some- what narrower specimen with more prominent orbital notches, figured by him on Plate 2, figure 6, as Cyathaspis wardelli. Again, it appears similar in shape to the internal impressions of the dorsal plate which Bryant figured as C. wardelli on his Plate 2, figure 6, and Plate 3, figure 1. It is essential that a lectotype be designated for Archegonaspis wardelli. If a recognizable dorsal plate is chosen from among the suite of type specimens it will have to be a specimen which is con- specific with Cyathaspis van ingeni of Bryant. In any case, at least one new name, and possibly two, will be needed for the C. wardelli of Bryant, 1926. A more radical step will at least reduce these name changes by one, and eliminate from all possible revival the undefined genus Hoarchegonaspis. This can be done by designating instead of a recognizable plate an unrecognizable fragment as the lectotype of 370 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Anatifopsis wardelli Ruedemann. I therefore designate as the lecto- holotype of this species New York State Museum no. 9613, which I have illustrated on Plate 2, figure 5. This will at least leave C. van ingent Bryant as based upon much better type specimens than would be the case otherwise, and will require new names only for the material which Bryant called Cyathasms wardelli. It is extremely dubious whether these forms are distinct generically from what has been called Palaeaspis, but as noted in the discussion of that genus, restudy of the original materials is impossible, and Palaeaspis and its two species may be completely unrecognizable. Any revision will at least require the re-establishing of the type species on neotype material. I propose no new name, specific or generic, for the form which was the C. wardella of Bryant, being of the opinion that such proposals should rest upon better preserved and a larger suite of materials than are now available. TOLYPELEPIS Pander Toly pelepis Pander has priority over Tolypaspis Schmidt; the former name was employed first for an isolated scale; Tolypaspis was pro- posed for an essentially complete dorsal shield. Kiaer employs Schmidt’s generic name and erects the family Tolypaspidae. Dorsal shield oval in outline, very much as in Cyathaspis, posterior margin not adequately known. The separation of the rostral and lateral areas is reminiscent of that of Archegonaspis. The dorsal surface is given a distinctive appearance by elongate wart-like tubercles, or scale-like ridges. Kiaer states that these structures are marked by a broad median ridge and finer lateral ridges. Actually, from his photograph, each one of these protuberances is broad enough that its elevated portion consists of a central dentin ridge and two lateral ridges, which usually continue beyond these elevated areas, though the sur- face is quite irregular in this respect. The protuberances usually bear pores, frequently two or more pores to one protuberance. On this basis, something of the lateral line system can be reconstructed. One pair of commissures passes from the margin of the head in front of the eyes obliquely toward a point behind the pineal body, but does not join. The pattern of the remaining pores is less clear, but from what can be seen, there is a lateral line system here not strikingly different from that of Poraspis. The elongate wart-like ridges are quite evi- dently not all equipped with pores, and are not the direct result of the development of a lateral line system. Rather they are the effect pro- duced by the breaking up of such longitudinal ridges, as are present in Cyathaspis, into individual elongate units. The only species known is the genotype, 7. wndulata Pander, from the upper Ludlow of Oesel, and possibly present also in the Ludlow of England. FLOWER AND WAYLAND-SMITH: CYATHASPID FISHES 371 DIPLASPIS Matthew This genus is known only from the genotype, Diplasys acadica Matthew, 1886. The description is accompanied by an outline drawing showing three plates in addition to a central disc. Although Matthew presents a rather detailed description, the exceedingly diagrammatic illustrations leave some doubt as to what the characters of this genus really are, particularly in view of the fact that diagrams of related forms have been found to exaggerate the distinctness of rostral and lateral plates. Kiaer (1932, p. 25) makes a new family for this form, on which, we gather, he is no better informed than we are: “It differs from all other previously known forms, and must therefore be re- garded as a representative of a new family. As the author could not study the original specimen, he refers to Matthew’s original descrip- tion.” TRAQUAIRASPIS Kiaer Kiaer based this genus upon Cyathaspis campbell Traquair, a species which had been described but not previously illustrated. Kiaer figures a ventral plate and two scales. Kiaer (1932, pp. 25-26) described the genus as follows: ‘Middle size Cyathaspider with a complete clefting of the dorsal shield into the different parts (rostral, lateral, and the central disc). The anterior part of the lateral plates probably divided as a separate, small supra-orbital plate. “The central disc roundish in the posterior part without any median keel or spine. Pineal area indistinctly limited. Traces of the pineal organ cannot be seen. The body scales small, probably in more rows than in Poraspis. “The sculpture of the plates with fine, distinct Psammosteus-like ridges, which usually are sharply divided into short portions. On the central disc of the dorsal shield the ribs are more irregularly arranged than on the ventral shield. The latter has a well-marked median keel, and the ribs form an elliptical septum with very fine ribs between the gross ones. “On the lateral and branchial plates a clear median keel with a strong ridge is developed. On both sides of the latter, fine ridges, regularly longitudinally arranged, are placed. “This genus is represented only by one specimen, Traquairasyis cambelli [sic] Trag. The specimen was found by Professor Cambell [sic] in the Downtonian series in Stonehaven area, Scotland.” In the last paragraphs obviously, the word “specimen”’ is a misprint for “species.”’ Traquair died shortly after publishing a short descrip_ One BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY tion of this species without illustrations. Kiaer (1932) has presented beautiful illustrations of a ventral plate and two scales, but the dorsal plate, which supplies the most crucial generic characters remains unillustrated. The longitudinal ridges on the surface are broken, and suggest the longitudinal ridges of a Cyathaspis more clearly than do the ridges in Tolypelepis. Pores have not been observed. CORVASPIS Woodward 1934 The genus Corvaspis is based upon a single species, Corvaspis kingt Woodward 1934, which is known only from a ventral plate. The plate is strongly concave in front, more deeply so than in any other known cyathaspid. The sides of the ventral shield are extremely tuberculate. Dentin ridges are linear except on the sides where they are broken up into tubercles, but are crossed by an irregular network of fine grooves, which is faintly reminiscent of the condition of the lateral line system portrayed by Kiaer (1932) for Irregularaspis, under the name of Dictyaspis. However, that these ridges are not the lateral linesystem, is shown clearly by the fact that they obviously have nothin g to do with the pores. A similar tuberculated border is unknown in other eyathaspids, although it is approached, though not closely, in the anterior margin of the ventral plate of Archegonaspis lindstrémi Kiaer, which was illustrated by Lindstrém as Cyathaspis schmidt. The emarginate anterior border of Archegonaspis drummondi ap- proaches the form of Corvaspis, but does not attain it. This species lacks the complex network of Corvaspis. It should be noted that without the dorsal plate, it is manifestly impossible to tell whether Corvaspis should be assigned to the family Poraspidae or to the Cyathaspidae. Two Devonian genera have been assigned to the Cyathaspida, which will only be mentioned. Cyrtaspidicthys Whitley (1940) ante- dates by a month Eucyrtaspis White and Moy-Thomas. The name replaces Cyrtaspis Bryant 1932, not Fischer 1853. Likewise, Allo- eryptaspis Whitley precedes Bryantasps White and Moy-Thomas, proposed to replace Cryptaspis Bryant 1934. DESCRIPTION OF THE VERNON SHALE CYATHASPIDS As noted in the introduction, the cyathaspids of the Vernon shale consist of only eight good specimens. Considerable vexation has attended attempts to match dorsal and ventral plates, as well as plates showing the outer and inner surfaces. The species are, as a consequence, based upon dorsal shields showing the external features, FLOWER AND WAYLAND-SMITH: CYATHASPID FISHES 373 and the reference of ventral shields or dorsal interiors has been necessarily somewhat inferential. As a consequence, it has seemed wisest to discuss each of these plates separately, leaving as tentative the specific identification of the ventral shields as well as one beautiful dorsal interior. The material may be summarized as follows: 1. A fine dorsal shield, the holotype of Vernonaspis allenae. 2. A much smaller dorsal plate, incomplete, but showing a fine im- pression of the exterior of the crucial anterior end. This is the holo- type of Vernonaspris leonard.. 3. A fine large dorsal plate, exposing only the inner surface and the impression of the interior. This is assigned tentatively to V. leonardt. 4. An essentially complete internal impression of a ventral shield, together with the impression of the exterior of the posterior portion. On the basis of form and surface this is assigned tentatively to the same species. 5. A fine internal impression of a dorsal plate, described as the holotype of Archegonaspis drummondt. 6. An impression of the exterior of a nearly complete ventral shield, clearly an Archegonaspis, tentatively assigned to the same species as the above. 7. A small triangular plate, possibly a branchial plate, but of uncer- tain affinities which is briefly described and illustrated, but not named. 8. A ventral plate exposing the interior, and etched to show the lateral line system. This is described as Archegonaspis sp. VERNONASPIS n. gen. Genotype: Vernonaspis allenae n. sp. This genus is known only from the dorsal shield, which is slender, the anterior end obscurely pointed, widened behind the eyes, slightly constricted before reaching the branchial region, branchial region gently expanded, sides subparallel in postbranchial region, prominent posterior lateral angles, the posterior margin obtusely pointed in the center, straight on either side; marginal band of posterior margin poorly developed. The surface shows an obscure division of the dorsal shield into a central disc, a rostral plate, and two small anterior lateral plates. The lateral plates proper are not distinguished from the central disc throughout most of their length; it is questionable as to whether the two anterior lateral plates are derived from the anterior end of the lateral plates, or from the rostrum. The surface of the rostrum is complexly whorled, centered about two points in the anterior lateral margin, pustulose in the center, transverse behind. Of the two lateral plates, the anterior one is smooth internally; the 374 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY posterior one bears coarse polygonal markings. On the exterior the ridges are slightly whorled in front, but become linear behind, blending with the pattern of the lateral plates and central disc. On the central disc the pineal body, itself not prominent or sharply elevated, is the center of radiating curved ridges which finally swing back joining the pattern of longitudinal ridges which occupy the greater part of the dorsal surface. They are modified at irregular intervals by pustules arranged in irregular groups. Some of these pustules bear minute pores, an expression of the lateral line system. Discussion. This remarkable form is unique in the differentiation of the rostrum, in the two anterior lateral plates of small size, and the absence of any clear differentiation of the lateral plates from the central disc. In this respect, the genus is somewhat comparable to the extremely inadequately known genus Diplaspis, judging from the out- line drawing of the one and only known specimen. Our form is very different in shape, and indeed, Diplaspis acadica is so inadequately known that close comparison is not possible. The pustules of the surface suggest a comparison with Tolypelepis and Traquairasyis, but the nodes of the surface are quite different in appearance in those genera. In Tolypelepis the nodes are larger and more wart-like; in Traquairaspis they are obviously isolated elements of broken up coarse longitudinal ridges such as are seen in a perfect state in the genus Cyathaspis. Traquairaspis according to the description of the genus, is said to have the rostral and lateral plates completely sepa- rated, and the anterior end of the lateral plate is possibly separated as a distinct plate. In this respect, Traquazraspis is closer to Vernonaspis than to any other described genus, but in Vernonaspis there is not one anterior lateral plate, but two, and the main parts of the lateral plates are not set off from the main part of the central disc. Strangely, the form most similar superficially to V. allenae is not a eyathaspid, but a pteraspid, the form described as Pteraspis leriche Zych (see Pauca, 1941, Pl. 2, fig. 2,) and P. lericht var. plana Brontzen (see Pauca, 1941, Pl. 2, fig. 3.) This form is distinct in a number of features, having apparently distinct pineal and ocular plates and a posterior groove for the reception of a dorsal spine on the posterior part of the shield. Fine surface details are not well enough shown to permit a close comparison, but there is no indication of a pattern of linear ridges and pustules. P. lericht is apparently a pteraspid and its resemblance to this cyathaspid is superficial, but it is not at all similar in aspect to Pteraspis, sensu stricto, as illustrated by White (1935). FLOWER AND WAYLAND-SMITH: CYATHASPID FISHES 375 VERNONASPIS ALLENAE N. sp. Plate 1; Plate 2, figures 3, 8. This species is known from a single dorsal plate, 31 mm. long, ocular width 11 mm., branchial width 16 mm., with the sides sub- parallel in the postbranchial region, no definite postbranchial expan- sion, 13 mm. wide at the posterior lateral angles. Pineal index 7.5 mm., rostrum 5 mm. long, rather bluntly pointed in front, the anterior part with the dentin ridges arranged in whorls about two anterior lateral centers, the central part pustulose, ridges broken and rather irregular, faintly radial in arrangement, transverse along posterior border. In the type, a slightly oblique break behind the rostrum is more con- spicuous than the posterior boundary of the rostrum itself, but is clearly adventitous. On the sides of the rostrum are two small anterior lateral plates, which are roughly rounded. The inner surface of the more anterior of these plates is smooth; the posterior one shows a surface of small polygonal facets, not closely similar to the usual pattern formed by the cancellous layer, but of a considerably coarser pattern. The external impression of this part of the type was available, and a portion of the test of this lateral region was broken away, and a rubber impression was then taken (PI. 2, fig. 3,) which shows in part the surface pattern over these regions. Though whorled in front, the surface over the greater part of the area occupied by these two plates consists of longitudinal ridges which are very similar to the arrangement of the ridges over the lateral plates and the central disc. The lateral plates are not at all differentiated from the central disc on the basis of the surface features, and their margins are indicated only by broad shallow poorly defined grooves. The greater part of the central disc bears fine longitudinal ridges, modified at intervals by irregular groups of raised rounded pustules. In some of these pustules minute pores are seen, which evidently connect with the lateral line system. The pineal body is in itself poorly defined on the exterior and is not well elevated; in comparison to our other forms it is extremely small and inconspicuous. It is, however, a center from which the ridges of the surface radiate. Those ridges which extend obliquely forward or directly laterad are curved, eventually turning toward the posterior margin, and joining the general pattern of longitudinal ridges. There is no arrangement in the anterior-lateral portion of these ridges into whorls, as in Vernon- aspis leonardi, and neither is there a strong V-shaped marking, point- ing forward and terminating at the pineal body, as in Anglaspis and Archegonaspis, which, when present, represents a portion of the lateral line system. 376 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY The surface of the type shows several faint linear depressions, best shown in Plate 2, figure 8, which simulate the lateral line system, particularly in showing two prominent longitudinal grooves. These grooves are probably adventitious, as they are not centered at all. Indeed the one on the left passes forward from near the median posterior angle directly to the pineal body, which is a behavior shown by no normal part of the lateral line system in other cyathaspids where, instead, neither of the main commissures intersect the pineal body, but pass forward on either side of it. There is indication of the thin posterior margin, which retains the dentin layer without the cancellous layer beneath, but it is not as well defined in this species as in our other forms. Discussion. This specimen is remarkable for the very thin condition of the dorsal plate, which was evidently poorly calcified and quite flexible, for it has been flattened somewhat, but bent in the process without developing any of the cracks which are shown by other of our specimens. Although the test was broken in several places, it proved extremely thin, and no definite trace of the usual layers could be found. The cancellous layer was evidently very thin in comparison to that of Vernonaspis cf. leonardi. However, a similar condition was found in the small holotype of V. leonardi, where no cancellous layer could be detected, indicating that this layer was probably not well developed in small and probably immature individuals. V. allenae is distinguished from V. leonardi by the quite different shape, particularly of the anterior end, the constriction of the lateral margins in front of the branchial region and behind the two anterior lateral plates, and by the much larger pustules of the surface. From other species, none of which it resembles closely, it can be distinguished by the characters of the genus. VERNONASPIS LEONARDI N. sp. Plate 2, figures 6-7; Plate 3, figure 5; Plate 5, figure 2. This species is represented by a small incomplete dorsal plate. There is an impression of the interior to which fragments of the plate adhere, the anterior end incomplete, the posterior end missing. The outline is gently rounded at the sides. The anterior end is shown clearly by an external impression, from which a rubber cast was made, which serves as the basis of our illustration (Pl. 3, fig. 5). This shows the anterior end to be slightly truncated, the sides very faintly notched for the reception of the eyes. The pineal body is large and strongly elevated. The rostrum is distinctly set off from the central disc, though the division between the central disc and the lateral plates is again obscure. At the anterior end, prominent depressions separate the anterior end of the lateral plate from the remainder, though both parts FLOWER AND WAYLAND-SMITH: CYATHASPID FISHES Odd bear essentially linear ridges. In front of this, and seemingly attached to the rostrum, is a similar small plate the surface of which shows the ridges developed into whorls, suggestive of the lateral plate figured by Kiaer and Heintz (1935, Pl. 33, fig. 1,) for Homalaspidella nitida, but with finer ornamentation. The pustules of the ridges on the middle part of the plate are tiny, but well elevated. The markings on the posterior part are somewhat irregular, but essentially linear. A ventral plate tentatively assigned to the same species, on the basis of form, is represented by an internal impression (PI. 2, figs. 6-7) and an impression of the exterior from which a rubber mold was taken (JEL, Hay, ates 2) The internal impression, 26 mm. long, 16 mm. wide, is quite strongly convex and, in flattening, the margin on the right has been bent under the remainder of the specimen. The anterior end which is imperfect, appears to be strongly truncate in front, and scarcely emarginate if at all. The posterior outline is obtusely angled in the middle, the oblique sides straight and meeting the lateral margins at a prominent angle. Lateral margins curved, greatest width attained shortly behind the middle of the plate. The impression of the interior is relatively smooth but (PI. 2, fig. 7,) shows a series of very faint ridges, which, however, it has proved impossible to distinguish with certainty from the cracks which are the result of slight crushing of the plate in flattening. In part, at least, these represent a portion of the lateral line system as in Archegonaspis sp., but no clear pattern can be made out (Fig. 1). Fig. 1. Ventral shield attributed to Vernonaspis leonardi, showing ob- servable pattern of fine ridges, and cracks of the surface, which cannot be properly differentiated. Part of this doubtless represents the lateral line system, which is here very imperfectly preserved. 378 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY The surface of the interior in natural color, shows the reticular pattern of the cancellous layer of the shell (PI. 2, fig. 6.) A small piece of the plate, adhering just orad of the center, fails to show any of the surface features or structure. The impression of the exterior of the same plate shows only the posterior portion (Pl. 5, fig. 2). The ridges are fine, closely spaced, linear, with irregularly scattered pustules, very tiny in relation to those of lV’. allenae, but suggesting affinities with that form, and com- parable to the much smaller pustules of the dorsal plate which is the holotype of Vernonaspis leonardi. The internal mold shows a distinct and rather prominent posterior margin to the ventral plate, where, as usual, the cancellous layer is absent and only the dentin layer present. This ventral plate bears a striking resemblance in outline to Clay- pole’s original drawing of Palaeaspis bitruncata, and indicates that this is a true ventral plate as stated by Claypole (1892) and not a distinct species as stated by Bryant (1926). Strict comparison is impossible. Claypole’s types have been destroyed, and the original illustration is only an outline drawing. Palaeaspis bitruncata cannot, under these circumstances, be recognized with certainty. The reference of this plate to V. leonardi rests upon similarity of form and the similarity of the tiny pustules to those of the dorsal plate and holotype. Cf. Vernonaspis leonard Plate 3, figures 2-4; Plate 4; Plate 5, figure 1. Ironically, the largest and finest dorsal plate in the entire collection of Vernon cyathaspids is known from an impression of the interior, and the plate itself, the dorsal surface of which is embedded in the matrix. Under these circumstances, nothing can be ascertained con- cerning the dorsal surface pattern, which supplies the generic criteria under our present rather finely divided classification. As a consequence, the generic and specific affinities of this form must remain uncertain. The plate has a maximum length of 32 mm., pineal index 6 mm., ocular width 10 mm., branchial width 18 mm., reduced to 17 mm. at the posterior lateral angles. The posterior margin is very thin, there being a well defined marginal band, and the shape of the posterior margin itself is somewhat obscure. There is clearly a blunt median point, as in V. allenae, the sides straight or very nearly so on either side of it. The lateral margins of the plate are slightly convex, with no clear differentiation of branchial and postbranchial regions, the anterior end rounded, the ocular impressions faint and very obscure. There is no clear differentiation between the marginal branchial region and the remainder of the plate, which in most cyathaspids is FLOWER AND WAYLAND-SMITH: CYATHASPID FISHES 379 well marked by shallow grooves on the interior of thed orsal plate, even when no clear separation exists on the basis of the exterior alone. The impression of the interior shows no clear indications of the nasal sacs. The pineal body is clearly defined and well elevated, behind which is the mesencephalon, represented by two faint vertical ridges, essentially in contact with the semicircular canals, and not distinct from them as in Archegonaspis drummondt. The tips of the anterior branches of the semicircular canals are bifid. The medullary ridge is long and well defined in contrast to 4. drummondi, slightly depressed in the center of the most elevated portion. The branchail impressions are faint, poorly outlined, but the seven pairs can be distinguished. No marginal branchial impressions can be seen. The interior of the dorsal plate (Pl. 4) shows with remarkable clarity the structure of the cancellous layer, in color. The broken edge, shown in further enlargement in Plate 5, figure 1, shows the usual cyathaspid structure, the thickest and most conspicuous element being the cancellous layer. The surface markings are indicated only in a few small areas where the plate has broken away from the shale, showing a small part of the impression of the exterior. The linear markings in the portion shown are not diagnostic, and agree closely with those of A. drwmmondi as known from the ventral plate, and also V.leonardi. The interior of the dorsal plate shows no traces of either a lateral line system, or the differentiation of anterior lateral plates. Shght flattening has caused the plate to be intersected by a series of cracks, widening toward the margin (PI. 3, figs. 2-3). These have no significance, and fail to parallel any course of a probable lateral line system. Discussion. 'The taxonomic position of this dorsal plate is uncertain without the dorsal surface. The transverse condition of the posterior margin is unlike that of most of the “Poraspidei,” and is more closely approximated in Archegonaspis and in Vernonaspis than in any other of the known genera. The character of the interior of the dorsal shield is quite different in several particulars from that of Arche- gonaspis drummondi, particularly in the obscure distinction of the lateral areas, the lack of lateral expansion over the branchial area, and the straight lines which make up the posterior margin of the shield; in Archegonaspis these lines are sinuate. A striking feature of this specimen is the thickness of the cancellous layer. Either this is thickened through old age, which is possible, for none of our other specimens approximates this one in size, or else our form is strikingly distinct from any of the other dorsal plates in the present collection. In shape, this plate agrees strongly with the much smaller holotype of Vernonaspis leonardi, and it is tentatively assigned to that species 380 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY mainly because of this strong resemblance. Two features oppose this determination. The first is the thickness of the cancellous layer which, as noted above, need not necessarily be a determining factor. The second is the absence of any discernible anterior lateral plates. Again, if old age has produced a thickening of the cancellous layer, it may well be accompanied by a secondary fusion of these plates. As has been noted, the present species shows significant differences in proportions with the previously described genera. None of the Poraspidae is closely similar, and of the Cyathaspidae the only genus known to contain species at all similar in proportions is Archegonaspis, and the resemblance even here is not close. The original material of Palaeaspis americana was considerably different in outline, the greatest width of the plate being farther forward, and the sides ap- proaching each other behind the middle. However, as pointed out, it is dubious whether this species or genus can be recognized in view of the loss of the type material. ARCHEGONASPIS DRUMMONDI 0. sp. Plate 2, figures 1-2; This species is based upon a beautifully preserved specimen, show- ing only the impression of the interior of a dorsal shield. This shield, 22 mm. in length, pineal index 5.5 mm., has an ocular width of 8 mm., increasing to a branchial-width of 13.5 mm. There is no widening toward the posterior end to a definite postbranchial area. The branch- ial regions are depressed at the edges, and the elevated portion of the plate between them has margins which are essentially straight from the ocular region, diverging gently to the posterior lateral angles, which are 12 mm. across. The angles are rounded, in contrast to their sharp condition in Vernonaspis, the posterior margin sinuate, broadly convex in the center instead of coming to a blunt point. A clear narrow posterior border 1 mm. wide is sharply defined; the internal mold shows that here, as in other species, there is only the dentin layer, the cancellous layer being absent, as shown by faint longitudinal striations on the internal mold. Pineal body large, round, well elevated; nasal sacs at anterior end relatively obscure, not prominently elevated on internal mold as in Anglaspis. Three short linear ridges behind the pineal body represent the mesencephalon; they lie between the anterior branches of the semicircular canals which are well elevated and clearly defined. The seven pairs of branchial impressions are broad, shallow, and about of equal clarity from the first to the last. Marginal branchial impres- sions are not at all evident. The surface over the lateral areas is faintly FLOWER AND WAYLAND-SMITH: CYATHASPID FISHES 381 pustulose, due in part to the presence of fragments of the plate, which as in our other material, was evidently more strongly calcified near the lateral margins. The medullary ridge is forked in front, and lies some distance behind the semicircular canals, between the posterior pair of branchial impressions. It is short, not penetrating into the post- branchial region. Discussion. Reference of this form to any genus is somewhat hazardous, since the essential generic characters are drawn largely from the features of the exterior rather than the interior of the dorsal plate. However, the extant illustrations are adequate to show that our form is not closely comparable with any of the previously de- scribed genera except Archegonaspis. Poraspis differs in the somewhat more produced condition of the anterior end, the greater prominence of the nasal sacs, the more anterior position of the pineal body in relation to the branchial impressions, the lateral expansion of a postbranchial area, the obscurity of the posterior margin, and the presence of well defined marginal branchial impressions. The known forms of Homalaspidella, though more like the present species in the prominent posterior margin (Kiaer and Heintz, 1935, Pl. 30, fig. 2; see also text fig. 56), resemble Poraspis more closely in other features. Much closer to our form in proportions are the interiors of Anglaspis and Arche- gonaspis, both in general shape and the arrangement of the impres- sions of the branchial apparatus and brain. Archegonaspis is favored for several reasons. In Anglaspis the front of the head tends to be somewhat produced, and the nasal sacs are more prominent, the orbital notches more pronounced, and the posterior margin more strongly produced. Also, a ventral plate, tentatively assigned to this species on the basis of proportions, shows the relatively fine closely spaced striae of Archegonaspis instead of the much coarser and deeper markings of an Anglaspis. Our material does not show what is con- sidered the crucial difference between the genera, namely whether the dorsal shield is entire as in Anglapis, or whether there is a separation of rostral and lateral plates on the basis of the surface marking, as in Archegonaspis. As noted in the discussion of these genera, this differ- ence is relatively slight, and actually it is here that there is the closest link between the families Poraspidae and Cyathaspidae. Archegonaspis cf. drummondi Plate 6 This ventral plate is known only from an impression of the exterior. Our present illustration is taken from an artificial rubber cast. The plate is narrow in front, broadest in the posterior third, the posterior 382 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY margin bluntly pointed in the middle, the outline essentially straight on either side, and meeting the lateral margin at a well defined angle. The anterior end, obscure, was evidently considerably emarginate. Dentin ridges are longitudinal over the greater part of the length of the shield, but are centered about two whorls on the anterior lateral portions. The ridges are fine, closely spaced. No structures are seen which can be interpreted with certainty as pores of the lateral line system. The plate has a length of 21 mm., a width which increases from 6 mm. at the anterior end to a maximum of 11 mm. Discussion. This plate in form and surface pattern is typical of Archegonaspis. Reference on the basis of such scant material - of course uncertain, but the general proportions suggest that this imay have well been the ventral plate of Archegonasms drummondi. It is markedly different in proportions from Archegonaspis sp., described below, and differs considerably in texture as well as in outline from the more parallel-sided plate which is tentatively assigned to Vernonaspis leonardi. ARCHEGONASPIS ? sp. Plate 3, figure 6; Plate 7; Plate 8, figure 3; text figure 2. Differing from all other forms in its considerable breadth, is a beautifully preserved ventral plate, remarkable for the sharp and slightly eccentric emargination of the anterior end, and the clarity with which the elements of the lateral line system are preserved. This form, we have been unable to identify with any of the other species described above, which is particularly unfortunate in view of the evidence it supplies as to the structure of the lateral line system. The reference to Archegonaspis is necessarily tentative, and is based upon the resemblance of its surface pattern to that of A. drummondi and also to A. lindstrém:. The plate is 24 mm. long, with a maximum width of 14 mm., with the dorsal surface embedded in matrix. The inner surface was essentially smooth, showing cracks due to flattening and a series of faint linear impressions. These are better seen on the impression of the inner surface (PI. 8, fig. 3, Text fig. 2), than on the inner surface itself. The plate was subjected to gentle etching, which exposed the cancellous layer, the canals of the lateral line system and, in places, the surface pattern of the dentin layer of the exterior, as shown on Plate 7. In shape, this plate is considerably broader in proportion to its length than other ventral plates encountered in the Vernon shale, showing a strong emargination of the anterior end, best shown in the specimen in its natural color (PI. 3, fig. 6,) and which is clearly strongly FLOWER AND WAYLAND-SMITH: CYATHASPID FISHES 383 askew. The ventral shield was originally quite strongly curved. Flattening has tended to bend one side more than the other, but is apparently not entirely adequate in itself to account for the eccentric position of the anterior notch. The posterior border meets the lateral edges at sharp and definite angles; it is bluntly pointed in the middle, straight on either side. The posterior thin border of the plate is not as sharply set off from the remainder of the plate as in our other ventral shields, but, as in the other specimens, apparently represents an area in which the cancellous layer is lacking. The surface bears longitudinal dentin ridges as in most species of Archegonaspis. These ridges are fine, closely spaced, and probably appear shallower on the exterior than they do from the etched interior shown on Plate 7, for they are not nearly as sharply defined on the impression of the exterior which is shown in the lower left part of the plate as oriented in our figure. Etching has revealed in addition to the normal features of the dentin and cancellous layers, a series of rather irregularly spaced short transverse canals which represent the lateral line system. Though more irregular and less complete, as series, than those shown for Poraspis (Kiaer and Heintz, 1935, fig. 3b, p. 45,) they are clearly analogous to the median portions of the ventral transverse commissures, and the lateral portions of the ventral transverse commissures, though only a very few individual canals of this last series can be detected, and it is not easy to distinguish them from the cracks in the ventral plate which are the result of slight flattening. Fig. 2. Outline of Archegonaspis? sp., showing reconstruction of canal system from the grooves of the inner surface. Visible portions are indicated in solid lines; indistinct portions by broken lines, frankly inferred connections by dotted lines, where no trace of the canal system could be seen. 384 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Though the matter does not seem to have been discussed in print, it stands to reason that the lateral line system, as a series of neural and sensory organs, must have had a direct connection with the brain and the central nervous system. If so, the incomplete series of canals which are embedded in the ventral plate must have been connected by nerve cords which lay within it, in the tissue of the animal. On the impression of the interior, such a system of canals is suggested by a series of faint ridges, representing, of course, grooves of the inner surface of the plate. They are obscure, low, poorly defined, and some- what less prominent than the cracks of the ventral plate, in which we are not interested. Their appearance is shown in Plate 8, figure 3. Under favorable lighting, aided by turning the specimen, they can be reconstructed into a much more complete series, illustrated in text figure 2. This shows at least a reasonable series of canals which align with the portions which are embedded in the ventral plate. One possible objection to this explanation is that, prior to etching, no openings were found on the inner surface of the plate which would indicate a connection between the internal canal system, and that embedded in the ventral plate itself. Discussion. This plate, the only one of our specimens showing clear evidence of the lateral line system, does not seem to match any of the species known from dorsal plates. This is particularly unfortunate because the lateral line system is so imperfectly known in the Cyathaspidae, to which apparently all of our forms belong. It is, however, evident that this form indicates that the differences in the lateral line systems of the Poraspidae and Cyathaspidae are not as marked as has been supposed, Indeed, it must be remembered that the Poraspidae, in which the lateral line system is fairly well known, have been studied on the basis of abundant and well preserved ma- terials, which has not been true of any of the genera assigned to the Gyathaspidae. This fact, together with our present specimen, suggest that the differences are perhaps much more apparent than real and will decrease when more of the Cyathaspidae have been studied from better and more abundant materials. Unidentified plate Plate 8, figures 1-2. This plate is roughly quadrangular, with one angle obtuse and rounded, or may be considered triangular, with one side convex and strongly curved in the middle. The upper side, as oriented in our figure, is perfectly straight; the under side strongly curved, the posterior side straight, forming an acute angle with the upper edge. | | | FLOWER AND WAYLAND-SMITH: CYATHASPID FISHES 385 Most of the plate has been exfoliated, showing traces of the reticular structure of cancellous layer. Small portions indicate that the surface bore linear markings parallel to the long axis of the plate. As such, the most logical interpretation would be that of a branchial plate, but this form does not agree at all closely with any known branchial plate figured for cyathaspids, the best known being those of Poraspis (Kiaer and Heintz, 1935) and Anglaspis (Wills, 1935). The form is too blunt and too obtusely pointed to suggest any known spine or scale. Short of a specimen showing a plate similar to this in relation to dorsal and ventral plates, an extremely remote contingency, it will not be possible to refer the present specimen to any genus or species with certainty. Therefore all that is possible is to call attention to this plate and to illustrate it. 386 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY BIBLIOGRAPHY Bryant, W. L. 1926. On the structure of Palaeaspis and on the occurrence in the United States of fossil fish belonging to the family Pteraspidae. Proc. Amer. Phil. Soc., 65: 256 - 271, pls. 1 - 4, 3 figs. CuaRKE, J. M. 1907. The Eurypterus shales of the Shawangunk mountains in eastern New York. Bull. New York State Mus.; 107: 295-326, pls. 1-8, 2 figs. CLAYPoLg, E: W. 1885. On the recent discovery of pteraspidian fish in the upper Silurian rocks of North America. Quart. Jour. Geol. Soe. London, 41: 48-64, 8 figs. 1892. On the structure of the American pteraspidian Palaeaspis Clay- pole, with remarks on the family. Quart. Jour. Geol. Soc. London, 48: 542-561, 8 figs. Eaton, H. E. 1924. A Vernon shale (Silurian) fauna in central New York. Bull. New York State Mus., 253: 111-116, 1 pl. Frower, R. H. and R. WayYLaNnbD-SMITH 1947. New fauna from the Vernon shale of New York. Bull. Geol. Soc. Amer., 58: 1180 (abstract). HARTNAGEL, C. A. 1907. Upper Siluric and Lower Devonic formations of the Skunnemunk mountain region. Bull. New York State Mus, 107: 39-54. Heintz, A. 1933. Neuer Fund von Archegonaspis in einem obersilurischen Geschiebe. Zeitschr. f. Geschiebeforschung, Berlin, 9: 123-181, 5 figs. 1938. Uber die iiltesten bekannten Wirbeltiere. Die Naturwissen- schaften, 26, heft 4: 49-58, 4 figs. JABKEL, O. 1927. Die Atemorgane der Wirbeltiere in phylogenetischer Beleuchtung. Zool Anz., 70: 273-303. KIAER, J. 1930. Ctenaspis, a new genus of cyathaspidian fishes; a preliminary report. Skr. om Svalbard og Ishavet, 33: 1—7, 4 figs. 1932. The Downtonian and Devonian vertebrates of Spitsbergen. IV. Suborder Cyathaspida. (Posthumous; edited by A. Heintz). Skr. om Svalbard og Ishavet, 52: 1-26, 11 pls., 12 figs. Krakr, J., and A. Heintz 1935. The Downtonian and Devonian vertebrates of Spitsbergen. V. Suborder Cyathaspida. Part I, Tribe Poraspidei, Family Poraspidae. Skr. om Svalbard og Ishavet, 40: 1-139, 40 pls., 57 figs. Kunrtu, A. 1872. Ueber Pteraspis. Zeitschr. Deutsch. Geol. Ges., 24: 1-8, Taf. 1. FLOWER AND WAYLAND-SMITH: CYATHASPID FISHES 387 LINDSTROEM, G. 1895. On remains of a Cyathaspis from the Silurian strata of Gotland. K. Svenska Vet.-Akad, Handlingar, 21, Afd. 4, No. 3: 1-15, 2 pls. MarrHEw, G. F. 1886. A preliminary notice of a new genus of Silurian fishes. Bull. Nat. Hist. Soc. New Brunswick, 6: 1-5, 1 fig. 1886. Discovery of a pteraspidian fish in the Silurian rocks of New Brunswick. Canadian Records of Science, 2, no. 4; 242-252, 1 fig. Moy-Tuomas, J. A. 1939. Palaeozoic fishes. 148 pp., 31 figs. London. Pauca, M. 1941. Poissons paléozoiques de la rive bucovinienne du Dniester. C. R. Inst. Geol. Roumanie, 26: 14-31, 2 pls., 4 figs. Romer, A. 8. 1945. Vertebrate paleontology. 687 pp. Chicago. RUEDEMANN, R. 1916. Spathiocaris and the Discinocarina. Bull. New York State Mus., 189: 98-112, pl. 32. 1920. A recurrent Pittsford (Salina) fauna. Bull. New York State Mus., 219, 220: 205-222. STRAND, E. 1934. Zoologische und palaiontologische Ergebnisse von den Svalbard- und Eismeer- Untersuchungen Norwegens. Folia Zool. Hydro- biol., 5: 326-830. VASCAUTANU, T. 1931. Les formations siluriennes de la rive roumaine du Dniester. Ann. Inst. Geol. Roumanie, 15: 585-661. Wuits, HE. I. 1935. The ostracoderm Pteraspis Kner and the. relationships of the various agnathous vertebrates. Phil. Trans. Royal Soc. London, (B), 225: 381-457, pls. 25-27, 97 figs. Waite, E. I., and J. A. Moy-THomas 1940-1941. Notes on the nomenclature of fossil fishes. I. Homonyms A-C. Ann. Mag. Nat. Hist., 5 (1940): 502-507. IT. Homonyms D-L. [bid., 6 (1940): 98-103. III. Homonyms M-Z. [bid.,7 (1941): 395-400. WHITLEY, G. P. 1940. The Nomenclator Zoologicus and some new fish names. Australian Naturalist, May, 1940: 241-245. Wits, L. J. 1935. Rare and new ostracoderm fishes from the Downtonian of Shrop- shire. Trans. Royal Soc. Edinb., 58: 427-447, 7 pls., 4 figs. Woopwarpb, A. 5S. 1889-1901. Catalogue of fossil fishes in The British Museum, 4 vols. London. 1934. Note on a new cyathaspidian fish from the Upper Downtonian rocks of Corvedale. Quart. Jour. Geol. Soc. London., 90: 566—567, 1 pl. PLATE 1 Vernonaspis allenae, n. sp. Dorsal plate, x 5%, holotype. Note surface detail and exfoliation of plates on anterior lateral region, where are shown outlines of a small anterior lateral plate with a smooth interior, and behind it a second lateral plate with an internal polygonal pattern. The apparent sharp line separating the rostrum from the central disc is clearly a crack and not an original feature. BULL. MUS. COMP. ZOOL. Flower & WayYLann-SmitH. CyatHaspio FisHes. PLATE 1 PLATE 2 1-2. Archegonaspis drummondi n. sp. Holotype, an internal impression of a dorsal plate, x2. (1) whitened, (2) natural color. 3. Vernonaspis allenae n. sp. Rubber impression from external mold of holotype, x7, showing a portion of the surface pattern over the lateral plates, which bear longitudinal lines, making these plates inconspicuous externally. Same specimen as Plate 1. 4. Cyathaspis “wardellv” of Bryant. Longwood shale, Shin Hollow, near Port Jervis, New York. x2. New York State Museum. 5. Cyathaspis wardellt Lectoholotype, a probable branchial plate, x2. New York State Museum, no. 9613. 6-7. Vernonaspis leonardi n. sp. Ventral plate tentatively assigned to this species. (6) Internal mold, natural color, showing traces of cancellous layer; (7) whitened, showing form more clearly, also faint cracks and possible traces of canals. See also text figure 1. 8. Vernonaspis alienae n. sp. Holotype, x2, showing more clearly than the enlargement on Plate 1, grooves of a possible lateral line system complicated by wrinkles of no organic sig- nificance. BULL. MUS. COMP. ZOOL. Flower & WayLANb-SmitH. CyatHaspio FisHes. Prate 2 Tie his Ls Jeg PLATE 3 1. Cyathaspis wardelli (Ruedemann) Syntype, a fractured branchial plate. New York State Museum no. 9618. 2-4. Cf. Vernonaspis leonardi (2) Interior of dorsal plate. (3) Impression of same, showing more clearly pineal body, branchial impressions, semicircular canals, and impression of brain. (4) Rubber mold from interior shown in figure 2, showi ae more clearly the posterior margin of the plate. 5. Vernonaspis leonardi Rubber impression from external mold of anterior end of dorsal plate, showing nature of ornament of dentin ridges. x7. 6. Archegonaspis ? sp. Interior of ventral plate, natural color, after etching, exposing canals of lateral line system, shown more clearly in Plate 8, figure 3, and some surface features of the dentin ridges. See also text figure 2. 7. Cyathaspis wardelli (Ruedemann) An incomplete dorsal plate. Syntype. New York State Museum no. 9617. Xe Piate 3 CyaTHAsPiD FisHEes Flower & WayYLANo-=SMITH . MUS. COMP. ZOOL. BULL PLATE 4 Cf. Vernonaspis leonardi Interior of dorsal plate, x4, showing in natural color, the pattern of the cancellous layer, and the posterior margin with dentin layer only. See also Plate 3, and Plate 5, figure 1. Flower & WayLanp-SmitH. CyaTHAspio FisHes. PLATE 4 BULL. MUS. COMP. ZOOL. RG PLATE 5 1. Cf. Vernonaspis leonardi Broken edge of dorsal plate shown in Plate 4, showing traces of layers. x16. 2. Vernonaspis leonardi n. sp. Rubber impression from external mold of ventral plate, from same specimen as that shown as an internal mold on Plate 2, figures 6 and 7. Flower & WaYLAND-SmiITH. CYATHSAPID FisHes. PLATE 5 . COMP. ZOOL. MUS BULL. Vie PLATE 6 Archegonaspis cf. drummondi Rubber impression from external mold of ventral plate tentatively assigned go this species, showing anterior end with pattern characteristic of Arche- tonaspis. BULL. MUS. COMP. ZOOL. Flower & WayYLAND-SmITH. CyaTHaspid FisHse. PLate 6 PLATE Archegonaspis ? sp. Interior of ventral plate, whitened, etched, showing cancellous layer and the interrupted canals of the lateral line system, also in places, penetrating to the dentin layer. In the lower left quadrant the plate is removed completely, showing the external mold of the pattern of the dentin layer; note that the grooves do not appear to be as deep in this portion. See also Plate 3, figure 6, and Plate 8, figure 3, and text figure 2. Swi MUS: COMP. ZOOL: FLoweR & WayLAND-SmitH. CyaTHasPip FisHes. PLaTE7 PLATE 8 1-2. Branchial plate of unknown taxonomic affinities. x2. (1) whitened; (2) natural color. 3. Archegonaspis ? sp. Internal impression of ventral plate. Same specimen as Plate 7, and also Plate 3, figure 6 and text figure 2. BULL. MUS. COMP. ZOOL. Flower & WayYLAND-SmitH. CyatHaspip Fish s. PLaTe8 Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Voz. 107, No. 7 NOTES ON SOME PETRELS OF THE NORTH PACIFIC By Outver L. Austin, JR. CAMBRIDGE, MASS., U. 8. A. PRINTED FOR THE MUSEUM November, 1952 PUBLICATIONS ISSUED BY OR IN CONNECTION WITH THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE BULLETIN (octavo) 1863 — The current volume is Vol. 107. Breviora (octavo) 1952 — No. 9 is current. Memorrs (quarto) 1864-1938 — Publication was terminated with Vol. 55. JOHNSONTA (quarto) 1941 — A publication of the Department of Mollusks. Vol. 2, no. 30 is current. OccASIONAL PAPERS OF THE DEPARTMENT OF MOLLUSKS (octavo) 1945 — Vol. 1, no. 16 is current. PROCEEDINGS OF THE NEw ENGLAND ZOOLOGICAL CLUB (octavo) 1899- 1948 — Published in connection with the Museum. Publication terminated with Vol. 24. These publications issued at irregular intervals in numbers which may be purchased separately. Prices and lists may be obtained on application to the Director of the Museum of Comparative Zodlogy, Cambridge 38. Massachusetts. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vou. 107, No. 7 NOTES ON SOME PETRELS OF THE NORTH PACIFIC By Otutver L. Austin, JR. CAMBRIDGE, MASS., U.S. A. JER ICN ALI IO IMO) gy AP IEl IH Ml Osis] OLA! November, 1952 No. 7.—Notes on Some Petrels of the North Pacific. By Otiver L. Austin, JR. In the course of preparing a comprehensive work on the birds of Japan! it has been necessary to review the extra-limital forms of several genera of Tubinares. As the results of this study are of sig- nificance in other areas, they are presented here instead of being held for inclusion in the work on Japan now in progress. The Tubinares, particularly the smaller species, are a difficult group systematically because of their relatively slight morphological varia- tion. Many of the populations occupy restricted but widely separated breeding grounds, to which it may be assumed they show a high degree of individual site tenacity. The group being an ancient one, some of the present breeding colonies may have been occupied con- tinuously since late Tertiary time, possibly longer. Yet despite wide geographical separation, the birds have all existed under such similar ecological conditions that no structural modifications have been en- couraged, and only minor fortuitous changes have been perpetuated genetically. The recognition of geographical populations by their slight mor- phological differences can be of great importance, particularly in the petrels which wander so far over the trackless seas in the non-breeding season. Other than by extensive banding on the breeding grounds, the year-round movements and distribution of the various breeding populations can be learned only by the ability to assign to their proper nesting grounds all the specimens taken away from them. Systematic study of the geographical races of petrels has been hampered by the paucity of comparable specimen material, which must be taken on the breeding grounds and be of similar age, both of the individual when collected and of the specimens themselves. The problem is aggravated by a bewildering synonymy, a plethora of names given to minor variations shown by small, inadequate series, and in some cases on geographical grounds alone. The group has been attacked and mutilated by some of the most liberal and radical of splitters, and studied as well by some of the soundest and most conservative systematists. No one relishes being considered either an unscrupulous ‘‘splitter”’ 1 Financed in part by a J. S. Guggenheim Memorial Foundation Fellowship. 392 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY or an unconscionable ‘“‘lumper’, but at the moment my tendencies seem, in common with the general trend, to lean toward the latter and the “‘conservative right”. A subspecific name designating a geo- graphical population is of no practical use unless at least three- quarters or more of the individuals of that population can be identified correctly by their morphological characteristics alone. The recognition of variations in color is still largely a matter of individual ability, integrity, and judgment because we have yet to find a satisfactory method of measuring color accurately.’ For mensurable characters such as size, the most practical standard is the statistical device (see Simpson and Roe, 1939) whereby two populations are considered separable only when the means of a given character in each population are as far or farther apart then the sum of their standard deviations. This-allows the separation of at least 84 per cent of each population from 84 per cent of the other. Among the petrels it is rare indeed to find the means of any character separated by two standard deviations, allowing a 97 per cent separation. Many of the named petrel races have been described, despite almost complete overlap, on claimed “average’’ differences. Statistically these “average’’ differences are of no significance unless it can be demonstrated that two standard errors on either side of the respective means do not overlap, and even then it is not possible to identify individual specimens with any degree of confidence. It may eventually be desirable to name these “average” populations, but for present purposes I can see no practical value in a name unless it can be applied with assurance, and without reference to the collecting data on the specimen’s label. In this study I have adhered as closely as possible to the ‘84 per cent from 84 per cent”’ convention. Genus PTERODROMA — the Gadfly Petrels The only members of this genus in the northwest Pacific belong to the difficult group separated by Mathews (1934b, 169) into the genus Cookilaria, but which Murphy (1929 and 1936) lumps with Pterodroma and divides into two species and ten subspecies. A more natural divi- sion of the ten recognized races both morphologically and geo- graphically is into three species. I propose revising Murphy’s key to the genus (1929, 2) as follows: A.1. Crown and nape close to a “neutral gray’’ and concolor with the back: AUSTIN: SOME PETRELS OF THE NORTH PACIFIC 393 Pterodroma cookti cookit (New Zealand) Pterodroma cookii axillaris (Chatham Islands) Pterodroma cookw nigripennis (Kermadec Islands) Pterodroma cookii orientalis (breeding grounds unknown) Pterodroma cookii defilippiana (Masatiera Id., Juan Fernandez) A.2. Crown and nape mainly sooty black, much darker than the back, with which it forms a sharp contrast: B.1. Inner web of outer primaries with a wedge-shaped white patch extending at least half the distance of the feathers from the base: 1. Pterodroma leucoptera leucoptera (east Australia and nearby islands) 2. Pterodroma leucoptera masafuerae (Masafuera Id., Juan Fernandez) 3. Pterodroma leucoptera longirostris (breeding grounds unknown) B.2. Primaries wholly dark, with no white on inner webs: 1. Pterodroma brevipes brevipes (New Hebrides and Fiji Ids.) . 2. Pterodroma brevipes hypoleuca (Hawaiian and Bonin Ids.)! Pterodroma cookit does not occur in the northwest Pacific. Ptero- droma leucoptera is represented there only by the rare and little-known P.L.longirostris which, though described originally from Mutsu Bay in northern Honshu, Japan, will probably be found eventually to nest in southern seas and to occur in the north Pacific only as a migrant. This form is still known only from ten specimens: Stejneger’s type and cotype in the Yamashina Museum, Tokyo; two undated Owston skins from the type locality in the American Museum of Natural History gS CON 1Jt is extremely unlikely that Krusenstern Island, the type locality of hypoleuca (and also of Salvin’s Puffinus cuneatus) is Ailuk in the Marshall group as stated by Fisher (1946, 588). According to Baker (1951, 65, 70), who nevertheless accepts Fisher’s hypothesis, the only Micronesian records for Pterodroma brevipes hypoleuca and Puffinus pacificus cuneatus are the type specimens of each from the mysterious ‘‘Krusenstern’’. It is far more likely, as James C. Greenway, Jr. first pointed out to me in 1948, that when Captain H. J. Snow collected these specimens he was engaged on a feather raid at Laysan where both species are common, and used the mythical and unlocatable Krusenstern on the labels of the skins he sent back to England to hide the scene of his activities from the authorities and from his competitors. At Mr. Greenway’s suggestion I searched in Japan for records of the old Yokohama firm of Owston and Snow, which might supply a clue to the island’s identity, but no former associate of the firm is alive today, and any of its records that might have existed prior to the war were de- stroyed by the fire raids of 1945. The question can never be answered with certainty, but Mr. Greenway’s logical hypothesis was accepted by Murphy (1951, 18) for the type locality of cuneatus. The type locality of Salvin’s Cstrelata hypoleuca is hereby designated as Laysan Island in the Hawaiian group. 394 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY in New York; and the six specimens in the Chicago Museum of Natural History taken at sea by the Crane Expedition on 17 August 1929 about 600 miles east of the type locality (cf. Murphy 1930, 14-15). Pterodroma brevipes breeds in the Hawaiian, Bonin, Fiji, and New Hebrides islands. No color differences have been ascribed to any of these populations, nor have I been able to discern any in the series I have examined. As suspected by most authorities from the few speci- mens previously available, an adequate series of brevipes from Japan and the Bonin Islands proves indistinguishable in measurements from the Hawaiian population, P.b.hypoleuca (table 1, fig. 1). However, the southern population of the Fijis and New Hebrides is separable from the populations of Hawaii and the Bonins on the basis of its smaller dimensions (table 1, fig. 1), particularly in the tail, to a lesser sf} (Gl 22 43 24 25 26 27 28 29 Prettaafecttasrtnenfienpees; wiNnG T ANA | ee ee S| SS He | +o ects as {28> oe | ey Fea ae ee = =a gr, [ows bewes wees bow bere Peete Population No Fig. 1. Measurements of Pterodroma brevipes populations from Table 1 graphed in the form of Hubbs-Perlmutter diagrams. In each diagram the solid horizontal line represents the observed range of variation, the central vertical line represents the mean, the rectangle indicates one standard devi- ation to either side of the mean, and the black part of each rectangle indicates twice the standard error of the mean. (For a discussion of the use of this method to demonstrate differences and similarities between populations see Hubbs and Perlmutter, 1942.) extent in the wing and culmen. The type of brevipes in the U. S. National Museum unfortunately has the longest tail in the series of southern birds, 109 mm., but the cotype’s tail (also in the U. S. Nat. Mus.) measures only 94 mm. The exact collecting locality of these two birds is doubtful (cf. Murphy, 1929, 15), but the name is valid both for the species and the southern race. Genus OCEANODROMA — Storm Petrels The Storm Petrels have a strong claim to the dubious honor of AUSTIN: SOME PETRELS OF THE NORTH PACIFIC 395 being one of the most difficult systematically of avian groups. Their morphological differences are in many cases so slight that many forms of specific rank cannot be identified with certainty in the field, and some of them only with difficulty in the hand. Representative popu- lations of some species, though oceans apart geographically, can be told apart if at all only by average measurements which allow fewer than half the individuals to be identified with assurance. The characters used are size and color. Size variation is very slight, and color is frequently even less definitive and reliable. The various shades of black employed by many students of the group are deceptive and misleading. They vary within populations with the time of year or state of molt of the individual, and particularly with the age of the specimen. The lead “bloom” characteristic of several forms fades and foxes with age, both of the individual in life and of the museum speci- men. Nevertheless color is a valid character, and in some cases more definitive than measurements, even though the differences can some- times be seen only with a series of fresh specimens in the hand and in _ good light. As I have been concerned with only three aggregations of this complex society of small black sea birds, I do not feel qualified to judge the larger systematic relationships of the group as a whole. Several of its elements have been separated as distinct genera, but under the present tendency to use the genus as a collective rather than a distinctive category, I believe it best to regard them all as congeneric. Nor does the establishment of subgenera seem warranted, though several of them combine into well-defined ‘‘species groups’’. OcEANODROMA CASTRO — Madeiran Fork-tailed Petrel This is one of the most wide-spread but least variable of the small black petrels. A small, white-rumped species, it is distinct from the other similarly colored and dimensioned species only in the shallower forking of its tail and in having the white rump feathers broadly tipped with black. Neither of these characters can be discerned without having the bird in the hand. It breeds in both hemispheres in widely- separated island colonies. Each of the known breeding populations has been named on minor color and size differences, and the synonymy is a long one. I can find no valid color differences in the material I have examined, none that are not ascribable to age-foxing or seasonal wear and fading. Each population exhibits minor size differences, but figure 2 and table 2 396 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY show the futility of trying to assign any but a few extremes among the available specimens to their proper populations on measurements. Some of the series I have measured are admittedly small (though much larger than those used by the describers of all but the proposed eA ie GOR 6ST Ove ou O) WING 4O I45 ISO 155 160 165 Population |No Atlantic Hawaii | 1 ae | Galapagos) |8 — {i Japan "se Iso. 15S 160 16S Fig. 2. Diagrammatic presentation of measurements of Oceanodroma castro populations from Table 2: Atlantic races), but they are adequate by the tenets of modern sta- tistical analysis. Not only is the overlapping almost complete, but there is no significant difference in their means, and larger series will undoubtedly show even slighter average differences. None of the proposed subspecies is tenable, and systematically the species is indivisible. The MELANIA-MARKHAMI-TRISTRAMI-MATSUDAIRAE complex Limited to the Pacific are four very similar forms of puzzling and ambiguous systematic relationship. They are the largest of the small, fork-tailed storm petrels, remarkably similar in size, and all are black-rumped. Two of them, markhami and melania, occur only in the eastern Pacific; the other two, tristrami! and matsudairae* only in 1When they described Cymochorea owstoni, Mathews and Iredale (1915, 581) discarded tristrami as of doubtful identity, in which they have been followed by Hartert (1915, 1415) and the Ornithological Society of Japan Special Committee (Hand-List 1942, 135). Although the type of tristrami was lost, the descriptions of it by Salvin in his ‘‘Key to the Species’’ of Oceano- droma (1896, 347) where the name tristramz is first mentioned, and by Ridgway in Salvin (idem, 354-355) are clearly identifiable in the light of the more adequate material now available. Three of the characters given are diagnostic, the tarsus length, the ‘‘plumbeous’”’ color of the head and mantle, and particularly the light edgings of the tertials and wing-coverts. Coupled with the Sendai Bay type-locality (the bird is fairly common in those and adjacent waters, and none of the three similar forms with which it might be confused has ever been taken in north-. eastern Honshu), they leave no doubt whatever as to which form the name must be applied. 1 have examined the type of owstoni which, as suspected by Peters (1931, 74, footnote), is unquestionably a synonym of tristramz. ie 2 Mis-spelled matsudariae in the origenal description (Kuroda 1922, 311),but named in honor of Marquis Matsudaira, the first serios student of the patrels in Japan. The original spelling is here- corrected. AUSTIN: SOME PETRELS OF THE NORTH PACIFIC 397 the central and western. Hitherto they have been regarded tenta- tively as two subspecies respectively of two species, tristrami as a race of markhami, and matsudairae of melania. I consider them equally of specific rank, and key them as follows: A.1. Tarsus 30 mm. or more: O. melania (breeds on islands off southern and Baja California.) A.2. Tarsus less than 30 mm.: . B.1. Primary shafts white at base, no lead bloom on head and shoulders: O. matsudairae (breeds in the Bonin Islands.) B.2. Primary shafts dark at base, a lead bloom on head and shoulders: C.1. Tarsus less than 25 mm., wing bar poorly marked: O. markhami (breeding ground unknown, perhaps in the southern Peruvian or northern Chilean Andes [Murphy, in lit.|. The one egg known was taken from a bird shot off central Peru.) C.2. Tarsus 25 mm. or more, the upper secondaries and tertials lighter at the tips, forming a distinct wing bar: O. tristrami (breeds in the Bonin, southern Izu [Tori- shima], and Hawaiian [Laysan] islands.) Table 3 and figure 3 show the average and inclusive measurements of the specimens examined. Although all four species show average differences, their dimensions overlap so that they are of diagnostic Species wing tail tail-fork tarsus ~ . [16 170 180 190 200 BO 90 100 110 120 20 30 40 20° 25 930-35 markhami WY melania matsudairae tristrami " Laysan Ids " Bonin, Izu Ids Agr Fig. 3. Average and inclusive measurements of the Black Fork-tailed Petrels from Table 3. 398 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY value only in melania and markhami. O. melania has a longer tarsus with practically no overlap, and a shorter bill, tail, and tail fork than any of the other three forms, though the latter three characters have too much overlap to be of much value. O. markhamz is distinet in having the shortest tarsus of the four. Color differences are of more use in separating the four forms. The heads and shoulders of markhami and tristrami in fresh plumage show a pronounced lead-gray bloom which is lacking in the uniformly brownish-black melania and matsudairae. The bloom fades with time, however, so that old specimens of tristramz can be told from matsudairae only by the color of the primary shafts which in the latter are uniquely white at the base when viewed dorsally, less so ventrally. O. tristrama is distinctive in the light margins of the upper secondary and tertial coverts, which form a pronounced light wing patch. The differences and similarities of the four forms may be tabulated as follows: Primary Heads and Tarsus shafts shoulders Wing patch melania 30-34.5 mm. dark brownish not pronounced matsudairae 25—2830 light brownish not pronounced markhami 22-25 dark lead gray not pronounced tristrami 26-30.5 dark lead gray pronounced Only in the presence or absence of the lead bloom on the heads and shoulders do the four species divide themselves into two equal groups with one representative of each on each side of the Pacific. Each form otherwise has its own distinctive characteristics not shared by the other three: melania its long tarsus, markhami its short one, matsudairae its light primary shafts, tristrami its light wing patch. O. tristrami and matsudairae are unquestionably specifically distinct, because they breed sympatrically in the same colonies in the Bonin Islands. Although the breeding place of markhami is unknown, neither it nor melania so far as known is sympatric with the other or with either of the two western forms. Hence either or both might be conspecific with either tristramz or matsudairae. We have only external morphological characters to go by, and these are neither indicative nor salient. Actually there is no way of determining the relationships of these allopatric forms other than arbitrarily, and until better evi- dence is available of their possible subspecific affinities, all four are best regarded taxonomically as of full specific rank. They have all doubtless evolved from a common ancestor, and form a fairly tight- AUSTIN: SOME PETRELS OF THE NORTH PACIFIC 399 knit “species group” in the genus Oceanodroma. On the strength of its long tarsus, Mathews (1934a, 119) proposed placing melania in the monotypic genus Loomelania, in which he has been followed by Murphy (1936, 744). I cannot see that this is warranted, or accomplishes any nomenclatural improvement, for the short tarsus of markhami and the white primary shafts of matsudairae make them just as worthy (or unworthy) of generic rank. Figure 3 also shows by an adequate series of measurements that the two widely separated populations of tristramz, one on Laysan Island and the other on the Bonins and southern Izus, are inseparable. OcEANODROMA LEUCORHOA — Leach’s Fork-tailed Storm Petrel This complex of small, fork-tailed petrels breeds in the northern hemispheres of both the Atlantic and the Pacific. Most authorities agree that the populations of the Atlantic and of the northwest Pacific are inseparable, and recognize the northeastern Pacific population, WING Au. # caren ee “Oo 45 [50 65 160 (65 tho) aS) K=Xo) | (JS) 8X0) 130 «14 Sieh One a7; oe I f Subspecies leucorhoa NORTH ATLANTIC ALEUTIANS & |4 WESTERN ALASKA COMMANDERS & 85 NORTHERN JAPAN TOTAL 94 beali sitka TO CEN- | 30 TRAL CALIFORNIA as Se a oe HEH AH His conoweoos ios |27| + EF} a oe mt HE on TOTAL Syi/ chaomani {| SAN BENITO IDS. (Ze) monorhis JAPAN (HONSHU) & KOREA socorroensis |25 GUADALUPE 1D. A awake ai 140 145 150 155 (60 (65 ke) 7) SK) SO) 13 14 15 16 17 Fig.4 Diagrammatic representation of measurements of Oceanodroma leucorhoa populations from Table 4. 400 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY O. 1. beali, as smaller, and forming a cline in the variability of the white rump patch as it progresses southward on the west coast of North America, reaching its climax in the black-rumped population of San Benito Island, 0. 1. chapmani. As yet no harmony has been attained in the number of intermediate forms to be recognized in the eastern Pacific. In the western Pacific, breeding on islands off Honshu, Kyushu, and Korea, is another black-rumped population, monorhis, which has been considered specifically distinct, but which is distinguishable from the San Benito population only by slight measurements. No cline either in color or size has yet been demonstrated between monorhis and the white-rumped leucorhoa population breeding on islands off northern Honshu and Hokkaido, though the latter shows some variability in the amount of white on the rump. In view of the cline between the white- and black-rumped populations in the eastern Pacific, and the fact that the Asiatic black- and white-rumped populations are allo- patric, monorhis is best regarded as conspecific with lewcorhoa. Table 4 and figure 4 show how little difference exists in the measure- ments of the known populations of lewcorhoa that have been studied and named. On dimensions alone these are all very weak races. Despite significant differences between the means of some characters, fewer than half of the individuals of many of the recognized subspecies are distinguishable. In the eastern Pacific, however, color can be used with size to advantage. As Loomis (1918) and van Rossem (1942) have pointed out, the white rump patch becomes darker from north to south, and the lead bloom of the upper parts fades to a browner black. The latter character varies with age of the specimen, the lead grays fading to brown in time just as they do in the previous group. This cline runs southward evenly from Sitka, Alaska, to the Farallon Islands off San Francisco, and breaks at Los Coronados Islands, the population of which is intermediate in color between the all-white- rumped birds to the north and the all-black-rumped birds to the south. None of these populations can be differentiated on size alone, and no sharp lines can be drawn between them in color. As none of the intermediate forms in the cline running from southern Alaska to Los Coronados Islands can be recognized, I synonymize both beldingi and willetti with bealt. The Guadalupe Island population, on the other hand, is quite variable in color, but distinctive in its small size. It is intermediate in its rump patch. Only one of the 23 Guadalupe specimens I have AUSTIN: SOME PETRELS OF THE NORTH PACIFIC 401 seen is entirely black-rumped, and the amount of white present in the remainder varies from four or five white feathers (as in the type of socorroensis which, as van Rossem pointed out (1942) is indeed the Guadalupe bird) to an intermediate condition with half the rump white, about as in the population of the central California coast. From the material I have examined I believe the O. leucorhoa complex should stand as follows: QO. l. leucorhoa: Breeds on islands in the north Atlantic, and in the north Pacific from northern Japan (Hokkaido) through the Kuriles, Commanders, and Aleutians to islands off the west coast of Alaska. It is the largest of all in all measurements, and whitest in the rump patch. 0. 1. beali: Breeds from southern Alaska (Sitka region) southward on islands off the west coast of North America to Los Coronados Islands off northern Baja California. Although their means are significantly smaller, on measurements only about half the population is dis- tinguishable from leucorhoa. The rump patch varies from almost identical to that of the nominate form in the north, to absence in 50 per cent of Los Coronados population. O. 1. chapmani: Breeds on the San Benito Islands, off central Baja California. Indistinguishable from beali in size, but always with a black rump. 0. l. monorhis: Breeds on islands off southern Japan (Honshu and Kyushu) and Korea. Black rumped, and distinguishable from chapmani by its shorter bill. Its wing averages longer and its tail shorter than those of both chapmani and socorrocnsis. O. 1. socorroensis: Breeds on Guadalupe Island off Baja California. Distinguishable from all other races except monorhis by its smaller, slenderer bill. Rump patch variable, almost always with some white, but never as much as in typical beali. ACKNOWLEDGEMENTS I gratefully express my obligations to my friend and mentor, the late James Lee Peters of the Museum of Comparative Zoology. I will always be indebted to him for his encouragement when I was a student, and for the friendship of the ensuing 26 years which gave me the privilege of recourse at all times to his sound critical judgment, his council, and advice. My thanks are also due to Dr. Herbert Friedmann for his courtesy in giving me access to the U.S.N.M. collections, and to Dr. R. C. 402 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Murphy for his generosity not only in making available to me the excellent A.M.N.H. series of his own special group, but also for going over them with me, reading a draft of this paper, and making pertinent suggestions which I have been pleased to follow. Acknowledgements are made particularly to Nagahisa Kuroda, my collaborator in Japan, who measured for me the extensive series of petrels in the Japanese collections, and finally to Dr. J. C. Dickinson, Jr., for his assistance with the statistical methods employed. BIBLIOGRAPHY Baker, Roun H. 1951. The avifauna of Micronesia, its origin, evolution, and distribution. Univ. Kansas Pub., Mus. Nat. Hist., 3 (1): 1-359. Fisuer, Harvey I. 1946. The type localities of Puffinus pacificus cuneatus Salvin and Pterodroma leucoptera hypoleuca (Salvin). Auk, 63: 587-588. Gopman, F. DuCangE 1907-1910. A monograph of the petrels. London, 376 pp., plates. Harrtert, Ernst 1912-1921. Die Végel der paliarktischen Fauna. Berlin, vol. II, pp. i-xxiv +1764. Husss, C. L., and A. PERLMUTTER 1942. Biometric comparison of several samples with particular reference to racial investigations. Amer. Nat. 76: 582-592. Kuropa, N. 1922. Remarks on the Japanese petrels of the genus Oceanodroma. Ibis, 1922, 309-314. Loomis, L. M. 1918. ° 9 v0j02 "M awe °o B05 + kuons140 te) SSNOVWUAS FF oo} ov 0 S31IN 4O 31V9S os Onbiopuouog 8104 @hO9U0H Y31S3H9ON ‘TSI Ao 4g @ O1vI45ne GRIFFIN: AIRPLANE OBSERVATIONS OF HOMING PIGEONS 415 at greater distances during 1947, most of the longer flights were observed from the air, so that at the time of each successive release something would be known of the birds’ previous experience and familiar territory. The general topography of the area in which all of these flights took place is shown in Figure 1; and the details of individual flights are. described below, and illustrated in the maps of Figures 2 to 6. EXPERIMENT I 1. General plan and prelinunary training. This experiment employed four white pigeons four to five months old (Nos. 6, 8, 9, and 21); they were first observed during a flight over familiar ground and were then followed from the air as they sought to find their way home from unknown territory. They were given pre- liminary training by numerous releases in an area extending ten miles northwest, seven miles west, two miles southwest, three miles south, two miles southeast, three miles east, and 21 miles northeast. After one of the later training releases in familiar territory northeast of home, they were followed by airplane. They were next carried 24 miles WNW and released in quite unfamiliar territory where the topography resembled that of a familiar area bordering the south end of Cayuga Lake. Most of the preliminary training flights were made in the company of other pigeons; but each bird was also released alone on several occasions, so that all had had experience in returning from familiar territory without any possibility of guidance by others of the flock. Nos. 6, 8, and 9 had each completed a total of 39 training flights while No. 21 had made 25. The more distant of these release points em- ployed in the preliminary training are shown in Figure 2 by circles with an enclosed cross. The releases were timed so that the birds never had more than two to three hours of daylight in excess of time required for a direct flight home. If they were back at the loft by nightfall or early the next morning, it was very unlikely that they had wandered to points far removed from the area included in Figure 2. Two birds, Nos. 6 and 9, never remained away from the loft over- night during this training period, but the other two did have occa- sional lapses. Number 8 remained out overnight after one release at 7 miles, and after another at 10 miles; after a second release at 7 miles it remained out for two days. Number 21 stayed out overnight 416 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY following two releases at 7 and 11 miles, and it remained away from the loft for 36 hours after a release at 8 miles. It was thus possible that these two birds had wandered extensively, and had a large area of familiar territory, although the experience of pigeon racers suggests that such lapses at a relatively early stage in a pigeon’s training are caused by preoccupation with rest or feeding rather than extensive wandering. Furthermore, when the group was later released in unfamiliar territory, the bird that exhibited by far the best homing performance was No. 9, which had never remained out overnight, rather than No. 8 or No. 21. 2. Airplane observation of a flight over familrar territory. After the preliminary training outlined above, the four white pigeons were released 11:45 a.M., October 21, 1946, at the Cortland airport and followed during their entire flight to the home loft. As can be seen from Figures 1 and 2, the Cortland airport is 17 miles north- east of the loft. The birds had been released there once before on October 16, and in addition they had been released October 20 at Homer, four miles farther from the loft, so that they had flown over Cortland area on two previous occasions. After a 10 minute period of circling in the immediate vicinity of the release point (the usual behavior which preceded homing flights by these pigeons), the four birds were joined by a larger flock of local pigeons. Together with these birds they circled over an area south- west of the airport, but finally separated from the local birds and began, about 15 minutes after release, to move along the route shown in Figure 2. During the first five miles they circled continuously, but the trend of their circling was always westward, i.e., slightly to the north of the true course home. When about five miles from the release point their manner of flight changed rather abruptly to straight flight along the course shown. Thereafter their progress was steady with very few sudden turns until they reached their home loft 70 minutes after release. Disregarding preliminary circling, the actual flight path was about 21 miles in length, while the airline distance from release point to home was 17 miles. 3. Airplane observations in unfamiliar territory with confusing to- pography. The Cortland flight showed that the birds, while neither as fast nor as sure of themselves as highly trained racing pigeons, nevertheless flew an essentially straight course when released in familiar territory. GRIFFIN: AIRPLANE OBSERVATIONS OF HOMING PIGEONS A417 Fig. 2. Route flown by four pigeons over familiar territory in Experiment I. The actual flight path dur- ing the first three miles contained even more loops and turns than are indi- cated. Previous release points are shown on the map by circles with an en- closed cross; note that while the birds flew down the same valley as that traversed during the transportation to earlier release points they did not fly directly past these points. CITIES of TOWNS 418 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY The next step was to repeat the observation from roughly the same distance into unknown territory. The geography of the Finger Lakes area permitted the addition of a geographical factor which offered some chance of revealing how much the birds depended upon to- pography. The home loft lay six miles southeast of the southern end of Cayuga Lake, in one of the two well-defined valleys leading south from this end of the lake. Such topographic features, in addition to the presence of the small city of Ithaca at the end of the lake, should be useful for visual orientation of birds released near Ithaca, as these birds had often been. The Finger Lakes are all similar in geological origin and general shape (see Fig. 1); Cayuga and Seneca Lakes have especially similar topography at their southern ends. In both cases there are (1) flat alluvial plains extending one to three miles immedi- ately south of the end of the lake, (2) steep hillsides on both sides of the lake along the southern half of its length, and (3) hills directly south, beyond the alluvial plain, which are divided by two valleys, each containing a small creek. Furthermore, there are cities on both alluvial plains, although these are quite different in appearance. It thus seemed possible that pigeons familiar with the south end of Cayuga Lake might be misled by this similarity when released at the south end of Seneca Lake. The first release point in unfamiliar territory was therefore placed near Burdett at point A of Figure 3, 23.6 miles WNW from the home loft. This point was analogous to a release point on the eastern shore of Cayuga Lake where the four birds had been released several times during their preliminary training (see Fig. 2). Ii the pigeons relied upon topography for their orientation, they might be expected to fly south from release point A into the valley leading southeast from Seneca Lake, whereas a true course towards their home would take them ESE. The difference between these two hypothetical courses was about 85 degrees. The actual release at point A was made at 11:05 a.m., October 27, 1946. In the interval since the flight from Cortland the pigeons had been given two short exercise flights by releases five miles west and 12 miles NNW from the loft. On the 27th they circled for about 15 minutes within a mile of release point A, and then turned abruptly south. The flight south was continued for about 214 miles, and at this distance from release point A the birds turned to the east as shown in Figure 2. By this time they were in a region where the topography was clearly different from that of their familiar territory near Ithaca. Fig. 3. Route flown over unfamiliar territory by the same four pigeons (as in Fig. 2). Point A was 23.6 miles WNW from the home loft, at a point near the shore of Seneca Lake that was analogous to a previous release point near Cayuga Lake. Note that on two occasions the birds flew south when departing from the vicinity of the release point; this is the direction they would be expected to take if they were misled by the topographical similarity between the two lakes. 420 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY After their turn to the east the pigeons returned to Burdett, circled over the village, turned east through a valley and continued along the course shown in Figure 3, which led them back to the vicinity of the release point. Again the birds flew south from Burdett, but this time flew three miles before turning east and then north once more. At one hour and forty minutes low fuel supply forced us to discontinue the observation at a time when the birds were heading northeast. Throughout the period of observation the four birds stayed together, and they flew 35 miles altogether, aside from local circling. Yet at the end of the observation they were only seven miles from the release point, heading 77° north of the correct course towards the home loft. Only one of the four, No. 9, returned to the loft at all, and it required four hours and forty-five minutes for the flight. This should be con- trasted with one hour and ten minutes required for the return of the entire group from Cortland over familiar territory. Pigeon No. 9 was again released at the south end of Seneca Lake (at point B of Fig. 3) at 11:08 a.m., November 9th. The airplane observation of No. 9 on this date was unfortunately terminated after a few minutes, when the pigeon was lost to view against a cluster of white houses in the city of Watkins Glen. But it was back at its loft when I first returned to look for it 11% hours later. The results of Experiment I show clearly that these four young white pigeons were unable to determine the correct direction of their home when released at 23.6 miles in what was evidently unfamiliar territory. The nearest of their previous release points to Seneca Lake was only seven miles from home, so that they had been suddenly carried across 16 miles of hilly territory over which they had no occasion to fly during the preliminary training. The general pattern of their flight suggested that they lacked any clear orientation and confcrmed to expectations of theories based upon recognition of visual Jandmarks. As in most homing flights of pigeons there was a clear separation between the initial period of circling or undirected short flights within a mile or so of the release point, on the one hand, and the rather sudden beginning of a definitely ‘‘cross-country”’ flight which continued in approximately the same direction for a matter of miles. In this case the initial period lasted 15 minutes, and the shift to cross-country flight was abrupt and recognizable at the time it occurred. As noted above, both cross-country flights away from the release point were directed south, just as one would expect if the birds were guided by GRIFFIN: AIRPLANE OBSERVATIONS OF HOMING PIGEONS 421 the topographical similarity between the southern ends of the Cayuga and Seneca Lake basins. It is not surprising that this topographical delusion failed to persist all the way to a point analogous to the location of the home loft southeast of Cayuga Lake. When the birds had flown south for two to three miles the different appearances of the two lake basins should have become quite obvious; for this flight was largely over open fields, whereas the analogous flight path south from a corresponding point on the shore of Cayuga Lake would have taken the birds over a residential section of Ithaca and over the Cornell campus. Another not uncommon behavior pattern of inexperienced or dis- oriented pigeons is to fly for several miles in a closed circuit that returns to the release point after 20 to 30 minutes. This type of flight has been described to me by pigeon racers and, as shown in Figure 3, it occurred after this release of pigeons at point A near Seneca Lake. EXPERIMENT II 1. General plan, and previous experience of the birds. In the spring of 1947 a flock of 13 experienced pigeons was available, all approximately one year old; seven were white (Nos. 9, 34, 35, 36, 45, 48, and 49), and six were dark colored (Nos. 4, 12, 14, 15, 16, and 17). The previous experience of No. 9 has been described under Experiment I; after the second release near Seneca Lake it had made only short exercise flights and was never out overnight. Nos. 34-49 had been purchased in the fall of 1946 as young birds never flown from their natal lofts; but all of the dark colored birds had been raised at my loft in 1946 and had acquired considerable flying experi- ence. In the course of 26 preliminary training flights from March 23 to May 22, 1947, this flock was brought into good flying condition by releases at several points within an area extending three to five miles east, south and southwest from the loft, 21 miles west, 13 miles northwest, and 5 miles north. In addition Nos. 4, 12, 14, and 15 had made three longer flights during 1946 from release points to the northeast of the loft, all of which are shown in Figure 1. On September 22, 1946, these birds were released at DeRuyter, N. Y., 42 miles northeast, at 3:45 p.M.; all were back at the loft by 6:30 p.m. On September 26 they were released at West Eaton, N. Y., 52 miles northeast, at 4:15 p.m.; three were again back by nightfall and the fourth, No. 12, was home by 7:30 a.m. the following morning. The 422 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY longest of these flights was made October 4, when three of the birds, Nos. 4, 14, and 15 were released at 1:00 P.M. near Oriskany Falls, 65 miles northeast of their home loft. Nos. 14 and 15 reached the home loft at 4:26 p.m., but No. 4 did not return until sometime between 5 p.m. October 5 and 9:00 a.m. October 6. Thus all flights except the return of No. 4 from Oriskany Falls occurred with sufficient speed to assure that no extensive wandering from the direct course home could have taken place. In Table 1, | have summarized the previous experience of the 13 pigeons used for Experiment IJ, together with the most serious cases where their return was slow enough to allow for possible wandering into the area of later releases. In each case the instance of slow homing listed was the worst one in the bird’s record, and none of its other lapses could have significantly increased the probability that the later release points of Experiment II had been visited. In Table 1, the homing times refer to the time when the bird was actually seen at the loft; in some cases a day or two elapsed between complete checks, so that some of these pigeons probably returned considerably sooner than indicated in the table. The general plan of Experiment II was to release the flock at progressively greater distances northwest from Ithaca, following the birds from the air after each release in unfamiliar territory. Because of occasional slow returns before the final releases at 55 miles or more, it is not possible to establish with complete certainty that any indi- vidual release was made in what was unfamiliar territory for every bird in the flock. But the probability that all of the more distant release points had been visited previously appears extremely small, and the results as a whole cannot reasonably be explained as orien- tation based exclusively upon familiar landmarks and exploration. The routes flown by various birds are presented in Figures 4, 5, and 6, and the pertinent observations on individual behavior are contained in a condensed summary of each observation. It was unfortunately not possible to provide a continuous watch for returning birds at the loft; but unless stated otherwise, all of the birds returned soon enough after release to establish the fact that they had flown home without extensive detours. 2. Description of individual flights. Covert, 19 miles northwest; birds released 12:35 p.m., May 24, sky almost fully overcast, warm, hazy, visibility five miles, calm. The flock circled within half a mile for 10 minutes, then in larger circles tending more to the south UMNUTXEUL otf} SOsBd 1oy}O [[B UL fauOY YYSIY YYSreays ApreoU MOYsS uy ‘quettiodxe oy} Surmp souewsojsed Burvioy ‘UMO] JOU SBA PATG oY) 4BY} 07 BoIPUT ‘(_) ‘sosoyjuareg 480] 480] Wd OFT ‘18999 eyaRyeg "SI, sInqsueur ‘say $0-02 SO] |-Miy, vou} 4 go7 480] ¥2-02 WY ZO:IT ‘ZI 990 S[[@,q oAOoMO FY PET 480] IT “90 * * * * * * ‘Wd QT:f 7 QT:% ‘2 900 «ensepueueD J048qa MQ x * * * * a * * “Wd §@:@ ‘Z 390 BAoUaL) * * r ” * * * * “WV 09:6 ‘28 349g PJAO ssep ‘SIU 66 “SIY 6G * Bo * * x Wd G€:¢ ‘€% “deg weye[iayuy tI =|{ * * 480] * naa (Ge) x * * Wd GHG “Gg AlNG ensepueue/) * * x * * SABD fF * ; ': ‘Wd [0:¢ ‘EZ AINE eAauary Wap 12 4[ne SABP 2 | ostadexo fs Shep % 4SO| < 5 * i pueyqory * BAep Z% WY C@:TT ‘GI Aine oqyaceg “Ay UO 480] * * 480] * * gf euns ik * * * * * * Wd €@:7 ‘ET oUNE pIeTPIAL YUAN * qystus9A0 my PYSUIOAO |4YZIUIOAO |4qRWIaAO ee Re 44 stUI0A0_ + 4qS1UsaA0 ee 4qs1u19A0 ‘Wd 12:8 ‘Of ounp JoWIOD 8.43009 * x * * * * * * * * * * * ‘Wd 0¢:% ‘J ounfe uosyepJoyuy * * * (>) ( )) * * ry * * * a * Wd G2:8 ‘Tg AB 910400 CD) ‘SIY $6 x SIY $3 SACD Z * * * SARD Z * “ “SIY $2 * Wd 10:6 ‘92 ABI UoHe]JoVUY ( ) eal * Nes Lin | eee = a fo Seek ih Sh *” a * ca * * *” ‘Wd ara § ‘BZ ARN J1VAO() “MSM “wy | “MS “MSM | ‘UD GT 98 |. Jomoy | stemr0Y (9F61 (OF6T 9P6T | Jomoy | (LF6T) (9F61) “WTSI | JVskep |uTsye | wTZye |yqatusaao | pooze | poowe | 4yaIg ‘Vqag | Ul UL {] | poose “AN ‘AN #4S1p SULUIOY SNotAadd ysamojg sAUp p | [BtoAos | “SIy gp SI 8h sAvale | sVmye 4S) ysqy) | ge gyda | sAvmye jug 3e | ‘ureg ye skep ¢ | skepg -19A0 “SI SP “SI If “MNM UN “HN “HN “UN “UN @) (=) (C_-) (ee) @) (Ge) “WG TH CG ) “ur ¢9 “ul G9 “Weg “wd ¢9 OPOT Ul 4qs1p 4sesuOT req Ieq req yooyo 1eq ayn eqn OIG SOUL ENN oqrq an ayy UT U Pan pea eniq antq an{q ang eng BC) (079) 6F 8F cP 98 GE ve 6 LT 9 GT I ai 7 yoquinad woes “UOATS ST OUT) SUTUIOY 0} uinjor ATIve ApUOTONS B soyvorput ‘, “YsTIO}sE Urey} pue “TT yuowedxy ut pasn suoasid jo A1oysty snotaetg ‘T aTaV AD4 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY and east than north and west. Four or five times they returned to the release point. At 20 to 25 minutes they came within sight of Cayuga Lake and suddenly turned towards its shore (see Fig. 4). On reaching the lake they flew south along the shore until at four miles north of Ithaca they crossed the lake and continued south along the east shore and thence directly home to reach the loft 55 minutes after release. Interlaken, 22 miles northwest; birds released 3:07 p.m., May 26; clear, visibility unlimited, polar continental air mass, wind southwest, moderate, sharp updrafts, cumulus clouds forming by the end of observation. The birds circled within 0.3 mile for 13 minutes, then flew east or ESE to reach Cayuga Lake (which was visible from the release point) at about the same point as on May 24. They then flew south along the west shore of the lake as shown in Figure 3 and were lost to view over Ithaca. Numbers 12, 36, and 48 were out overnight, and Nos. 16 and 35 did not return until two days after release. Covert, 19 miles northwest; birds released 3:25 p.m., May 31. They were not followed by air, but were observed from the ground to fly off to the west, return ten minutes later to the release point, and then head south. Interlaken, 22 miles northwest; birds released 2:30 p.m., June 1; solid overcast, visibility eight to ten miles, gentle SSE wind, light rain just before and after flight. The birds circled 15 minutes, mostly to the south, then headed east. Two miles from release point the flock split; one group continued southeast and another, which I followed because it contained three white birds, con- tinued SSE along the main highway, thus passing close to all of the previous release points (see Fig. 4). In fact, they seemed to seek out these spots, flying within 200 to 300 yards of each previous release point and circling over the one at Trumansburg, although at other times they deviated a quarter to a half mile from the highway. Scott’s Corner, 25 miles northwest; birds released 3:27 p.m., June 10; clear, warm, slight haze, visibility eight miles, a few cumulus clouds forming, light south wind. The birds circled near the release point for 17 minutes, flying more often to the south and east than in other directions. They then flew ESE with some veering, reached Cayuga Lake near Shelldrake Point, and turned south along the Jake shore after first flying 200 yards out over the water (see Fig. 4). Fifty-eight minutes after release the flock under observation contained five white and three dark birds, and these eight landed on the shore of the lake near Jacksonville so that they could not be followed farther from the air. Numbers 4, 14, 16, 34, 35, 36, and 48 remained out overnight; that is, three dark and four white birds. This is only one bird less than the number seen to land, so that it seems likely that other activities than wide wandering caused this slow homing. Up to June 10 the birds were given every opportunity to learn the topography and landmarks. They were carried to the release points in a cage attached to the roof of an automobile and not only was each release point within sight of the previous one from an altitude of 100 to 200 feet, but all were within two GRIFFIN: AIRPLANE OBSERVATIONS OF HOMING PIGEONS 425 Cayuga Lake N 6 Scott s i> © Release points Corners QW) _ Cities and towns CN =~. interlaken Sy \ \ NAS XN Covert AWN \ ») Xr re) 5 miles Fig. 4. Routes flown by pigeons during the early stages of Experiment II : all these release points were within a mile or two of familiar territory. Note that after the release at Interlaken on June 1 the birds that were being followed flew directly over several of the previous release points. 426 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY miles of Cayuga Lake. The birds returned in all observed cases by following the shore line of the lake, or in one case by flying along the highway leading past all of the previous release points. After June 10, however, the birds were carried in closed boxes inside the automobile or airplane and were taken to release points separated by greater distances. Table 1 shows that the losses increased sharply when these changes were made. Clearly the releases after June 10 provided a more strenuous test of the pigeons’ homing ability. Willard, 32 miles northwest; birds released 1:23 p.m., June 13, partly over- cast, moderate south wind, visibility about 15 miles, moderately bumpy air with high stratus and stratocumulus clouds. This release provided an oppor- tunity to test the influence of topography, for the birds had previously been released close to Cayuga Lake (Fig. 4) and on reaching it had only to turn right and follow the shoreline in order to reach Ithaca. At Willard, however, they were released about 11% miles from Seneca Lake as shown in Figure 5. If they relied upon simple topographical orientation one would expect them to circle, notice Seneca Lake, fly to it, turn right along its shore, and head north — almost directly away from home. The birds circled for only about five minutes, then flew southwest and reached the shore of Seneca Lake at ten minutes. So far they had behaved roughly as would be expected on the basis of the topography. On reaching the lake, however, the flock turned south, flew inland about one-half mile to a railroad track along which they continued southward. About one mile south of the village of Willard the flock split, four white birds continuing south, and others turning back to the north. There were at least two white birds in the latter group, and almost certainly they were Nos. 34 and 45 which failed to find their way home. One of these, No. 34, was picked up at Newark, N. Y., on June 16, and later returned tome. Newark lies 28 miles NNW from Willard (see Fig. 5) and it seems quite possible that. these two white birds broke away from the flock because they were indeed misled by the topographic resemblance to previous release points near Cayuga Lake. The group which continued south, containing four white birds, flew a surprisingly straight course towards home, as shown in Figure 5, despite the fact that this course led them over many miles of territory never previously visited when under observation. But some of these pigeons might have visited the Willard area during one of the longer periods away from home after an earlier release (for example, No. 35 had been out for two days after the release at Scotts’ Corner, and Nos. 12, 16, 21, and 45 may have been out as long as 48 hours after a release 21 miles WSW of the loft). After reaching the city of Ithaca the group made a detour across the valley extending southwest from the end of the lake, instead of flying southeast to- their home loft. This detour inside of very familiar territory is difficult to- understand; it should be compared with another detour within familiar terri- tory that occurred on July 25 after the first release at Canandaigua. West Fayette, 38 miles northwest; birds released 11:25 a.m., June 15; high. GRIFFIN: AIRPLANE OBSERVATIONS OF HOMING PIGEONS 427 Honeoye Falls o Canandaigua / y! | Loke o Release points oO Villcges Keuka Lake} Cities and towns wen Railroads 10 20 30 miles Geneva Nework Coyuga Lake \ \ \\. Infertaken iA / SN A> © 13/6 Seneca Lake Home: Fig. 5. Routes flown in Experiment II after releases at 32 to 55 miles from home, in what was probably unfamiliar territory. Note the generally correct direction taken in most instances. The observed flight path of No. 34 after its release at Willard is indicated by the small arrow branching away from the path taken by the main flock; this bird was recovered at Newark, and may have been misled by the lake topography. Note also the case of an apparent “overshoot” across familiar territory after the release at Canandaigua on July 25. nm Botavia & 4 “=~ A spy Honsoye Folls Dorlen 1 ___ Route 20 oy ~\ \ “ Ni Qos \-—B povittion ‘ “77 *S-41/10 | Webster Crossing o Releose points D Villeges Cities ond towns —=— Prominent highwoy 19 10 20 30 miles Conandolguo , (4,2 Cayuga Loke Genevo = 7/10 & Infericken Taughonnock Greek Keuko Loke Ithace © Home Seneca Leke Fig. 6. Routes flown after the second releases at Geneva and Canandaigua, and after the final releases at Honeoye Falls and Batavia in what was almost certainly unfamiliar territory. Note the nearly correct initial flight direction in all cases. 428 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY solid overcast with lower cumulus clouds at 2500 feet, scattered showers, hazy, visibility 4 to 5 miles, moderate SSE wind. These were the most adverse weather conditions, and also the poorest visibility, of any release involving airplane observation. The birds circled for 20 minutes, then started southeast as shown in Figure 5, but after 114 minutes turned east and flew to within three miles of Cayuga Lake (which, however, was not visible). There they turned south, and then back to the west, next northwest, and finally returned to the release point 35 minutes after they had been set free. They next flew ENE and east and were finally lost to view 55 minutes after release. Pigeon No. 34 was released separately from the other birds in order to see whether its wanderings between Willard and Newark had had any effect. It was not followed from the air, but it headed west from the release point and never returned. The main group also did poorly. The first birds returned after six hours, and No. 14, which had previously appeared to be one of the best fliers, was found six days after release at Richland, N. Y. — 68 miles northeast of the release point. Numbers 4 and 35 were out for two days and No. 49 was not back until a week after release. Here was striking confirmation of the pigeon racers’ experience that weather involving very poor visibility has an adverse effect on pigeons’ homing performance. Geneva, 43 miles northwest; birds released 3:01 p.m., July 23; almost solid overcast at 5000 to 6000 feet and half coverage of cumulus clouds at about 3000 feet, strong updrafts, wind generally west. The birds circled near the release point for 17 minutes, with two false starts to the south along the shore of Seneca Lake; after each of these southward flights they returned to the release point. They next flew south for two miles, turned sharply east across the lake, and turned south again along its eastern shore as shown in Figure 5. This portion of the flight crossed terrain over which they had flown on June 18, but now they continued south along the lake shore about two miles farther than on the 13th. There the flock split, and we followed two white birds which flew straight east along a road which led them into Interlaken. Next they flew southeast to Cayuga Lake and continued home along its west shore just as they had done on so many previous flights. Number 16 remained out for 4 days after this release. Canandaigua, 55 miles northwest; birds released 2:42 p.m., July 25; four- tenths coverage with cumulus clouds, visibility eight to ten miles, modified polar continental air mass, warm, light westerly winds with surprisingly smooth air. This release represented the longest step yet taken into presumably unfamiliar territory. The birds circled the release point for 23 minutes, in- cluding a flight one mile to the south; but they returned frequently to the release point. They then flew west over the city of Canadaigua, returned once again to the release point, and next headed ESE along the route shown in Figure 5. When the main flock made a turn to the north one white bird separated from the rest by turning farther and heading northwest. The main group flew along a railroad for some distance as shown in Figure 5, and after GRIFFIN: AIRPLANE OBSERVATIONS OF HOMING PIGEONS 429 reaching Seneca Lake followed roughly the same route as that flown from Geneva to the west shore of Cayuga Lake on July 23. Here, however, they surprised us by continuing directly east across Cayuga Lake and holding to the same direction for 15 miles more. As shown in Figure 5, they eventually turned south and reached territory which must have been familiar from flights made in 1946. It is even possible that the whole area ten to fifteen miles east of Cayuga Lake had been visited by one of the birds either in 1946 when returning from releases up to 65 miles northeast, or during the long period required to return from the West Fayette releases on July 15. But why these birds flew so far east of Cayuga Lake is difficult to explain, for they were already at the shoreline along which they had flown home on previous oc- casions. The whole performance is reminiscent of pigeon racers’ description of an “‘overshoot’’. This is reputed to be a flight in which pigeons continue flying in the same direction, passing their own loft and landing at some other loft in the same city which lies farther from the release point than their own home. Number 35 was lost on this flight, and it may well have been the single white bird that broke away to the northwest shortly after the flock left the vicinity of Canandaigua. After this 12-mile step into presumably unfamiliar territory, the birds took essentially the correct direction, except for short flights near the release point, just as they had done at Willard. It is far less likely that they could have found familiar landmarks near Canandaigua, however, for no exploratory flights west of Geneva had been observed. Nevertheless, there had been opportunity for wandering into the Canandaigua area after the release at West Fayette, when Nos. 4 and 35 were out for two days and No. 49 may have been out for as long as seven days, and also after the release at Geneva, when No. 16 may have been out as long as four days. During August it was not possible to make further observations of these pigeons. They were given two short practice flights from Interlaken (3:35 P.m., September 23) and Ovid (9:50 a.m., September 27); on the first flight, No. 9 stayed out overnight and Nos. 16 and 49 remained away from the loft for four days. Evidently their homing ability, or more likely their physical stamina for long flights, had been impaired by two months with no activity except exercise flights in the immediate vicinity of the loft. Geneva, 43 miles northwest; birds released 2:53 p.m., October 2; high broken altocumulus clouds and occasional sunshine, visibility 10 to 12 miles with patches of haze, wind northwest 10-12 mph at the outset but dropping to five mph at the end of the observations. The birds circled for twenty minutes at the release point, making during that period two flights of one half mile to the south and back to the release point. They next flew north over the city of Geneva, returned once again to the release point, and finally headed south along the route shown in Figure 6. During the ensuing southward flight along the lake shore they twice flew out over the water for 200 to 300 yards and then returned to the land. It was tempting to infer that two dark birds which 430 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY in each case flew farther from land than the rest were attempting to lead the flock across the lake. The flock did turn sharply to the east and crossed Seneca Lake only when it arrived at the point where the birds had reached the lake on their flight from Canandaigua on July 25. The remainder of this flight followed routes flown on previous occasions. Canandaigua, 55 miles northwest; birds released October 7; clear, hot, visi- bility 15 miles with slight haze in patches, gentle southwest winds. The flock was divided into two groups; the first group to be released (Nos. 4, 16, 36, and 49) were carried from Geneva to Canandaigua at the window of the airplane in an attempt to “show” them the terrain. After being set free at 2:15 p.m. these four birds interrupted their initial circling near the release point by a sojourn on the roof of a barn and silo in the company of local pigeons; but at 30 minutes after release the four birds circled several times over the city of Canandaigua and the north end of Canandaigua Lake. At this time the two white birds separated from the others and were followed south along the route shown in Figure 6. After another return to the release point at 68 minutes after release these two white birds spent another 20 minutes over the city of Canandaigua before being lost. Neither returned to the loft; but one, No. 36, was reported from Webster, N. Y., 25 miles southwest from Canandaigua, between October 11 and 13. Evidently the previous flight from Canandaigua three months before, together with the view of the landscape from the airplane window, had not sufficed to permit successful orientation by visual landmarks. Numbers 4 and 16 found their way home, but No. 16 remained out overnight. ; The remaining four birds flew SSE immediately after release at 4:15 p.m. with no circling at all, but were unfortunately lost to view six miles SSE from the release point when they apparently landed in a patch of woods (see Fig. 6). This relatively straight flight departed almost immediately from the area over which these birds had flown after their previous release at Canandaigua on July 25. Three birds of this group homed promptly, but No. 17 remained out overnight. Honeoye Falls, 72 miles WNW;; birds released 11:02 a.m., October 12; clear and warm, visibility eight to ten miles with slight haze, wind SSW eight to ten mph. The birds circled near the release point for 13 minutes before flying the route shown in Figure 6. Their first heading was south for four miles, but they then turned east and circled back to the release point 23 minutes after release. On their second departure to the southeast they flew considerably farther. The lone white bird, No. 9, separated from the others at the point indicated on the map and could not be followed farther. The returns from this release were very poor; three birds with excellent previous records were never heard from again (Nos. 12, 15, and 17), and the two which did return (Nos. 4 and 9) arrived somewhere between 20 and 24.5 hours after release. The re- maining bird, No. 16, was reported found dead near Trumansburg on October 25 (apparently shot). It was thus close to the direct line between the release GRIFFIN: AIRPLANE OBSERVATIONS OF HOMING PIGEONS 431 point and home (see Fig. 4). Despite the heavy losses, it was clear that the flock followed from the air -did set out in nearly the correct direction while still within two miles of the Honeoye Falls release point. The only slow returns on previous flights had occurred after releases at Interlaken, West Fayette, and Geneva, 29 to 50 miles east of Honeoye Falls. It is surely difficult to believe that even one bird had wandered by chance to the immediate vicinity of this release point at Honeoye Falls, so that the flock must have been able to select the approxi- mate direction of home without benefit of visual landmarks. Batavia, 100 miles WNW; birds released 1:40 p.m., October 23; clear and warm, haze in patches caused visibility to vary from four to nine miles, wind WNW, light. The birds circled for five minutes, landed in a tree near the release point for five minutes more, and then flew rather directly south along the road leading into the city of Batavia. They circled over the city and returned to the release point. They then separated, so that only the white bird, No. 9, could be followed as it circled back to the south and continued southeast along the course shown in Figure 6. The sharp turn to the east occurred as it reached a prominent four-lane highway, along which it flew for about three miles. At Pavillion it landed on a housetop and remained for the period from 78 to 117 minutes after the time of release. During this interval the visibility increased considerably as a patch of haze over Pavillion moved to the east. After leaving Pavillion along the route shown on the map, No. 9 headed in a generally westerly direction and at 168 minutes after release it again came to the same highway near Darien. Once more it turned left and flew along this highway until lost to view among a large flock of local pigeons which it joined approximately three hours after its release at Batavia. The flight of No. 9 was the longest traced from the air, covering 54 miles during the 214 hours of actual cross-country flight. Despite this bird’s success in all previous flights it seemed lost and disoriented towards the end of the period of observation. Yet it can be seen from Figure 6 that even in this flight the first hour constituted progress in a generally correct direction. Both birds in this final experiment had returned from Honeoye Falls in 20 to 24.5 hours, of which only between 9 and 14 hours were daylight, so that the chance of their having acquired a familiarity with the vicinity of Batavia is very remote indeed. Once again it is necessary to conclude that these pigeons could select the correct direction without reliance upon visual landmarks. DISCUSSION An experienced pigeon racer would immediately recognize from these case histories that my pigeons, especially the white birds, did not display the same physical stamina for long flights as his successful racers. The losses became serious at shorter distances than in the training of the best strains of racing pigeons, and the several occasions 432 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY when the birds landed showed that they were not racing for home. It should be recalled, however, that my birds were intentionally de- prived of the company of more experienced pigeons on any of their training flights. In most pigeon races, birds of widely differing experi- ence, and from several lofts, are released together. Some of the present observations indicated a tendency for the birds to follow linear landmarks such as railroads, prominent highways, and the shores of the Finger Lakes. In the releases at Willard and Burdett topographic factors seem to have led certain birds astray. Most striking, however, was the tendency for the birds to fly rather straight courses for several miles in a generally correct direction. On several occasions, as noted above, this occurred in territory where the birds had never previously been observed to fly. The ‘‘overshoot”’ after the first Canandaigua release is especially interesting in view of pigeon racers’ reports of similar occurrences. To explain the essentially correct direction taken in several cases one might assume that the birds had actually visited the release area after some previous release, during the excess time above that required for a direct flight home. The releases at Honeoye Falls and Batavia are good examples; for the flock took the correct initial direction within two miles of the release point. Individual birds had remained out for one to four days after releases at Interlaken, West Fayette, or Geneva. But it seems improbable that they had indulged in enough wandering to have carried them over the area immediately surrounding Honeoye Falls or Batavia. Moreover the birds often deserted the correct direction after about an hour of flying; if they were indeed familiar with the countryside one would not expect this to occur. It should be borne in mind that all of the more distant release points in Experiment II lay nearly on a line extending approximately north- west from the home loft. The birds flew in a generally southeasterly direction after release in what was almost certainly unfamiliar terri- tory. This could be considered either as an ability to fly towards home, or alternately as an ability to fly southeast, learned in the course of several previous flights in that direction. The obvious critical experi- ment of a release at 75 to 100 miles in some totally different direction — was planned, but could not be carried out because of the rapid dwindling of the flock. The possibility that the sun is used as a means of orientation by birds should not be overlooked; such a notion has become more at- tractive in recent years because of the remarkable discoveries of von GRIFFIN: AIRPLANE OBSERVATIONS OF HOMING PIGEONS 433 Frisch (1950) concerning the precise orientation of insects with refer- ence to the sun. The results of the release at West Fayette coincide with racers’ experience that heavy losses are likely to occur in the presence of a heavy overcast and poor visibility. It should also be recalled that in pigeon races of 100 miles or more it is customary to make all releases in the early morning, and furthermore to select re- lease points at various distances so that they all lie in roughly the same direction from the home lofts. Such arrangements might pos- sibly have been adopted empirically because they facilitated an initial orientation relative to the rising sun. In Experiment II the releases were made in the middle of day, with one or two exceptions. It had been my original plan to vary the hour of release in order to test the possibility that the sun’s position was an important factor, but the heavy losses at distances beyond 50 miles made this impossible. The relatively slight variation in time of day at which pigeons were released did not appear to be correlated with the direction of their flight except in one case. In the second release at Canandaigua on October 7, 1947, the flock of eight birds was divided into two groups of four, and the second group to be released had been held in their shipping box for somewhat longer than was usual in these experiments. The behavior of this group after release was striking in that they began a cross-country type of flight almost immediately after their release without the ten to twenty minutes of circling in the immediate vicinity of the release point that occurred in almost every other case. The hour of this release, 4:15 P.M., was distinctly later than that of any previous release at more than 20 miles (for example, the five preceding releases had been at 2:53 p.M., October 2; 9:50 a.M., Sep- tember 27; 3:35 p.m., July 25; 3:01 p.m., July 23; and 11:25 a.m., July 15). The approximate heading of the pigeons during this straight flight observed from the air between 4:15 and 4:30 p.m., October 7, was 160°; this may be compared with a heading of 135° flown by the flock followed after a previous release at Canandaigua on July 25, after the birds had begun an essentially straight flight over what was probably unfamiliar territory. This shift of about 25° to the south corresponds to the sun’s movement across the sky in one hour and forty minutes, a rough agreement with the time interval separating this release from the customary release time between 2 and 3 P.M. This single correlation can, however, be considered merely as a sug- gestive indication. Subsequent to these observations four additional studies of pigeon 434 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY homing have been reported. Hitchcock (1950, 1952) has followed flocks of U.S. Signal Corps racing pigeons or their immediate progeny over both familiar and unfamiliar territory. The greatest distance of transportation from the home loft was 190 miles. Hitchcock observed many deviations from the direct course home while following these pigeons from small airplanes — even within familiar territory. The control of previous flights was apparently less complete than in the experiments reported above, so that it is more difficult to set precise limits to the birds’ familiar territory. But in three significant cases birds trained to return to their home loft in New Jersey from release points up to 100 miles to the west were carried 100 to 190 miles north and followed after release in what was certainly unfamiliar territory. In all three cases many birds headed to the east or southeast for the first 25 to 50 miles, demonstrating a tendency to take the direction of the previous training flights even though this direction led them away from home. On the other hand, in these and other flights observed by Hitchcock some birds were clearly able to head towards home even from unfa- miliar territory. In the clearest case birds were first trained by releases up to 35 miles to the northwest, 17 miles to the east, 28 miles to the southwest and 87 miles to the southeast. They were then released at Massena, N. Y., 106 miles northwest of the home loft at Middlebury, Vermont. These birds were followed over a route reminiscent of the flight paths described above at Honeoye Falls and Batavia. The general trend of the flight path clearly lay between east and south, so that these pigeons were taking the correct general direction. There was, however, one complicating geographical factor in the Massena experiment. Virtually all the previous flights of these pigeons had been in hilly or mountainous regions of Vermont or the Adirondack area of New York. Massena lies in the broad, flat valley of the St. Lawrence River, and at the time the birds were observed the Adi- rondack mountains were apparently visible. Hence the pigeons might possibly have been using these mountains as an ecological cue (in the meaning discussed earlier, Griffin 1944); but in view of other evidence discussed below it seems far more likely that these pigeons were in fact able to select roughtly the correct direction in unfamiliar territory. Yeagley (1947 and 1951) has reported extensive experiments with homing pigeons designed to test a theory of bird navigation based upon an assumed sensitivity to terrestrial magnetism and to the Coriolis force. This theory has been adequately and severely criticised GRIFFIN: AIRPLANE OBSERVATIONS OF HOMING PIGEONS 439 from several points of view (Thorpe 1949); but some of Yeagley’s recent work has involved airplane observations of pigeons, both in familiar and unfamiliar territory. Again the pigeons sometimes showed wide deviations from the direct course, even within familiar territory. Fourteen birds were followed for significant distances after individual releases in totally unfamiliar territory (several hundred miles from home); thirteen of these were released at the same point and one at another point several miles to the north. Nine of the first 13 birds flew in a northwesterly direction, while the remaining four scattered in other directions. The single bird set free at a different release point also flew to the northwest. It is difficult to accept Yeagley’s expla- nation of these flight paths for reasons made clear by Thorpe and other critics; and there does not seem to have been any clear relation between the northwesterly heading and previous training flights. The selection of the same general direction of flight by so large a proportion of these birds complements the evidence discussed above for some orienting factor that is related to direction, rather than to position of the home loft. A third recent contribution by Kramer and his associates (Kramer and St. Paul, 1950; Kramer and Seilkopf, 1950; and Dinnendahl and Kramer, 1950) offers further support for this point of view. Pigeons released in unfamiliar territory, at some distance from the line along which previous training releases had been made, tended to start their flights in the direction of the earlier training flights; and they required longer, on the average, to reach their home than when released along the line connecting the previous release points with the home loft. All this evidence, together with Skinner’s recent demonstration (1950) of the striking ability of pigeons to retain visual memories of landmarks, suggests that orientation is based upon visual factors, but not solely upon topographic memory of familiar territory. Matthews (1951) has recently presented a substantial contribution to the problem of pigeon navigation. Using birds raised especially for his experiments by experienced and successful racers he judged their homing performance both by the customary standards of speed and percentages of returns and also by the initial direction of flight ob- served through binoculars from the release point. In his earlier experi- ments with birds of three to four months trained up to distances of 18 miles, and trained predominantly in one direction, Matthews found that releases in unfamiliar territory tended clearly to result in flights that approximated the training direction. In this the results were 436 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY similar to those reported by Kramer ef al. and by Hitchcock. But in Matthews’ later experiments with more experienced birds (training flights of 25 to 127 miles) the initial flight directions after release in unfamiliar territory centered around the true direction of home rather than approximating the training direction or scattering at random. Matthews showed, by releases at different times of the day, that no simple insect-like flight at a fixed angle to the sun could account for even those cases where birds in unfamiliar territory did hold to the direction of their training flights. In other experiments he obtained entirely negative results from a repetition of Yeagley’s experiments in which magnets attached to homing pigeons were reported to worsen their homing performance. Matthews found that on the average the homing performance of his birds was better on clear or partly cloudy days than under a solid overcast, and he is inclined towards some explanation of pigeon navigation based upon the position of the sun. Taking all of these results into consideration we may conclude that pigeons vary so greatly in homing ability that it is advisable to dis- tinguish three distinct levels of navigational ability. What we may conveniently call type I homing is reliance upon ‘“‘contact flying” within familiar territory. When released in totally unfamiliar territory birds which exhibit only type I homing show a quantitative homing performance that can easily be accounted for by theories of exploration or even by a completely random search for familiar landmarks. They may be misled by topographic features resembling portions of their familiar territory. Clear examples would be the pigeons of Experiment I. Pigeon racers generally believe that young and inexperienced pigeons home in this way, and, among wild birds, the gannets followed by airplane gave every evidence of employing type I homing. A second level of homing, conveniently designated type II, is the ability to maintain flight in roughly a fixed direction, usually that adhered to in previous training flights. Examples would be the birds from Fort Monmouth, New Jersey, followed by Hitchcock after being released in central New York State; these birds had been trained to fly east from training releases in Pennsylvania, and they were followed by air as they flew easterly courses over unfamiliar territory. The earlier experiments of Matthews and the behavior of pigeons described above after releases at Canandaigua, Honeoye Falls, and Batavia appear to be type II. Since most pigeon races involve a series of release points all lying in roughly the same direction from home, flying in a fixed direction may be of predominant importance in pigeon GRIFFIN: AIRPLANE OBSERVATIONS OF HOMING PIGEONS 437 racing. Furthermore the results obtained with wild birds by Schiz (1934, 1949), by Riippell (1944) and by Rowan (1946) suggest that cross-country flights in roughly a fixed direction may be of importance in the natural migrations of many wild birds. Finally there is a third level of homing ability, conveniently desig- nated as type III, which permits birds to fly approximately straight towards home from unknown territory regardless of the direction in which home lies. Such an ability has long been tacitly assigned to homing pigeons and to wild birds; but aside from unsatisfying, indirect evidence such as that obtained with petrels (Griffin, 1940), the recent work of Matthews with pigeons affords the first rigorous evidence for the existence of this third and most remarkable type of homing. In addition, at least one of the pigeon flights traced from the air by Hitchcock appears to fall into type III. Most recently of all, Kramer and von St. Paul (1952) have reported two additional cases of type II homing in pigeons. In differentiating between these three types of homing one must not forget that the same flock of pigeons may exhibit two, or even all three, of these types of orientation at different times during a single homing flight. Thus Hitchcock’s birds which had been trained to fly east and were then released in unfamiliar territory did follow an easterly course that continued for an hour or two. But then they shifted to a different, and usually a more correct, direction, perhaps because the flight in the direction of previous training had not brought any familiar landmarks into view. These birds appear to have used type II homing at first and then shifted to something approximating type III. Pigeon No. 9 released at Batavia followed a route, described above, which suggests type II homing during the first hour, followed, after failure to reach familiar territory, by type I. In conclusion, it must be admitted that the central problem of pigeon navigation remains unsolved; the sensory basis of orientation in both type II and type III homing is still a matter for speculation. There is, however, good reason to hope that the real progress achieved within the past few years will continue during the reasonably near future, and that the definite demonstration of these three categories of homing in pigeons may lead to their further analysis, and eventually to experiments that will clarify the sensory basis of types II and III. Indeed significant progress in this direction has been reported very recently by Kramer and his associates, who have apparently succeeded in arranging an experimental situation in which birds demonstrate 438 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY their choice of a definite compass direction while confined within a small cage. The results achieved to date point towards a sort of celestial navigation based upon the sun and sky brightness patterns; but a discussion of this subject is outside the scope of the present paper, especially since the current status of the larger subject of orientation in wild birds as well as pigeons has recently been reviewed elsewhere (Griffin, 1952). GRIFFIN: AIRPLANE OBSERVATIONS OF HOMING PIGEONS 459 BIBLIOGRAPHY CLAPAREDE, E. 1903. La faculté d’orientation lointaine. Arch. de Psychol., Geneva, 2: 133-180. EXNeER, S. 1905. Uber das Orientierungsvermégen der Brieftauben. Sitzungsber. Kaiser. Akad. Wiss. Wien, Math.-Naturwiss. KI. [11, 114: 763-790. DINNENDARL, L., and G. KRAMER 1950. Heimkehrleistungen italienischer und deutscher Reisetauben., Die Vogelwarte, 15: 237-242. Friscu, K. v. 1950. Bees, their vision, chemical senses, and language. Cornell Univ. Press, Ithaca, N. Y. GrirrFin, D. R. 1940. Homing experiments with Leach’s petrels. Auk, 67: 61-74. 1943. Homing experiments with herring gulls and common terns. Bird Banding, 14: 7-33. 1944. The sensory basis of bird navigation. Quart. Rev. Biol., 19: 15-31. 1952. Bird navigation. Biol. Rev., 27 (4): in press. GrirFin, D. R., and R. J. Hock 1949. Airplane observations of homing birds. Ecology, 30: 176-195. GuUNDLACH, R. H. 1932. A field study of homing in pigeons. Journ. Comp. Psychol., 13: 397-402. Hernrortu, O., and K. Hernroru 1941. Das Heimfinde-vermogen der Brieftauben. Journ. f. Ornithol., 89: 213-256. Hircncock, H. B. 1950. Aerial observations of homing pigeons. Anat. Rec. 108, (3): 571-572. 1952. Airplane observations of homing pigeons. Proc. Amer. Philos. Soc., 96: 270-289. Hopags, C. F. 1894. The method of homing pigeons. Popular Science Monthly, 44, (April): 758-775. Kramer, G. 1952. Experiments on bird orientation. Ibis, 94: 265-285. Kramer, G., and U. v. St. Pau 1950. Ein wesentlicher Bestandteil der Orientierung der Reisetaube: Die Richtungsdressur. Zeitschr. f. Tierpsychol., 7: 620-631. 1952. Heimkehrleistungen von Brieftauben ohne Richtungsdressur. Zool. Anz. Suppl. Bd., 16: 172-178. 440 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Kramer, G., and H. SEILKOPF 1950. Heimkehrleistungen von Reisetauben in Abhangigkeit vom Wetter, insbesondere vom Wind. Die Vogelwarte, 15: 242-247. Martruews, G. V. T. 1951. The experimental investigation of navigation in homing pigeons. Journ. Exp. Biol., 28: 508-536. Prart, C. S., and R: 8. DARE 1945. The homing instinct in pigeons. Science, 101: 439-440. Rapaup, E. 1928. How animals find their way about. (Chap. V.), Harcourt, Brace and Co., N. Y. RrivibreE, B. B. 1929. ‘The homing faculty” in pigeons. Proc. Sixth Int. Ornithol. Congr. Copenhagen, 1926: 535-555. Rowan, W. 1946. Experiments in bird migration. Proc. and Trans. Roy. Soc. Canada, (3) 40, sect. 5: 123-135. RUprewy, W. 1944. Versuche itiber Heimfinden ziehender Nebelkrahen nach Ver- frachtung. Journ. f. Ornithol., 92: 106-132. Scutiz, E. 1934. Vom Storch-versuch 1933 der Vogelwarte Rossitten. Der Vogelzug 5: 21-25. 1949. Die Spat-Auflassung ostpreussischer Jungstérche in Westdeutsch- land durch die Vogelwarte Rossitten 1933. Die Vogelwarte 15, (2): 63-78. SKINNER, B. F. 1950. Are theories of learning necessary? Psychol. Rev., 57: 193-216. Tuorpe, W. H. 1949. Recent biological evidence for the methods of bird orientation. Proc. Linnean. Soc. London, 160: 85-94. Warson, J. B., and K. 8. LasHLey 1915. Homing and related activities of birds. Papers Dept. Marine Biol. Carnegie Instit., 7: 1-104. (Carn. Inst. Publ. No. 211) YEAGLEY, H. L. 1947. A preliminary study of a physical basis of bird navigation. Journ. Appl. Physics, 18: 1035-1063. 1951. A preliminary study of a physical basis of bird navigation, II. Journ. Appl. Physics, 22: 746-760. Manuscript received for publication September 3, 1952. Bulletin of the Museum of Comparative Zoology AXAW Vel AN 18% WAN IR ID) (GOI IL Is Gee Vou. 107, No. 9 THE APSEUDID CHELIFERA OF THE EASTERN TROPICAL AND NORTH TEMPERATE PACIFIC OCEAN By Ropert JAMES MENZIES CAMBRIDGE, MASS., U.S. A. PRINTED FOR THE MUSEUM February, 1953 PUBLICATIONS ISSUED BY OR IN CONNECTION WITH THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE BULLETIN (octavo) 1863 — The current volume is Vol. 107. BREVIORA (octavo) 1952 — No. 9 is current. Memorrs (quarto) 1864-1938 — Publication was terminated with Vol. 55. JOHNSONIA (quarto) 1941 — A publication of the Department of Mollusks. Vol. 2, no. 31 is current. OccasIONAL PAPERS OF THE DEPARTMENT OF Mo.Luusks (octavo) 1945 — Vol. 1, no. 17 is current. PROCEEDINGS OF THE NEW ENGLAND ZOOLOGICAL CLUB (octavo) 1899- 1948 — Published in connection with the Museum. Publication terminated with Vol. 24. These publications issued at irregular intervals in numbers which may be purchased separately. Prices and lists may be obtained on application to the Director of the Museum of Comparative Zoélogy, Cambridge 38, Massachusetts. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vot. 107, No. 9 THE APSEUDID CHELIFERA OF THE EASTERN TROPICAL AND NORTH TEMPERATE PACIFIC OCEAN By Ropert JAMES MENZIES CAMBRIDGE, MASS., U.S. A. PRINTED FOR THE MUSEUM February, 1953 No. 9 — The Apseudid Chelifera of the Eastern Tropical and North Temperate Pacific Ocean By Rosert JAMES MENZIES Research Fellow, Allan Hancock Foundation The University of Southern California INTRODUCTION The crustacean order Tanaidacea (auct. Chelifera) has for many years been considered to be constituted by the families Apseudidae and Tanaidae. The recent subdivision of the Tanaidae by Lang (1949) into the Tanaidae (sensu stricto), and the Paratanaidae suggests that a further division of the Apseudidae may also be in order, since it too is composed of a heterogeneous group of genera. Currently, however, there is no good evidence that a splitting of the Apseudidae can be made successfully, because the characteristics of too many of its genera and species are imperfectly known. The only monographic account treating American Tanaidacea is Richardson’s (1905) ““A Monograph on the Isopods of North Amer- ica’’, in which the Tanaidacea are considered a suborder of the Isopoda. This lumping of the tanaids and isopods into one order represents an outdated classification; however, because one finds that classification in use today it seems desirable that a few of the major differences be- tween the tanaids and isopods be pointed out. In the Tanaidacea a carapace is present. This consists of a fusion of the first peraeonal somite with the cephalon. Contained within the carapace in a branchial chamber are the “‘cephalic’’ gills which consist of delicate, foliaceous appendages attached to the maxillipeds. The peraeon consists of only six free somites. The first pair of peraeopods, which are attached below the carapace, are invariably chelate. The eyes are usually located on eyelobes which are separated from the cara- pace. The rami of the uropods are multiarticulate. In complete con- trast, the Isopoda have no carapace and no cephalic gills, although in some species the first peraeonal somite does fuse with the cephalon. Respiration in the Isopoda is carried out by means of foliaceous ap- pendages called pleopods which are attached in pairs to the somites 1 Contribution No. 104 from the Allan Hancock Foundation, University of Southern Califor- nia, Los Angeles, Calif. 444 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY of the abdomen or pleon. The first pair of peraeopods of isopods is never chelate although in a few genera, those of the Anthuridae par- ticularly, subchelate peraeopods occur. The eyes of the isopods are invariably fused with the head, and separated eyelobes are not known to occur. The rami of the uropods are usually flattened; each ramus consists of a single article. The first species to be described from the area under consideration were Apseudes meridionalis Richardson (1912a) and A pseudes tropicalis Richardson (1912b). Both were collected from below 400 fathoms, the former, off the Galapagos Is., and the latter from off Cape San Lorenzo, Ecuador. These species have not been recorded since. Until the dis- covery of Dalapseudes (Boone 1923), a probable synonym of Para- pseudes, at Laguna Beach, California, not a single record existed of an apseudid from the Pacific shores of North America. The discovery of Synapseudes intumescens Menzies (1949) from Dillon Beach, Marin County, California, brought the number of previously known species from the area under consideration to four. : The writer has examined numerous collections of Tanaidacea from localities north of California. None contained specimens of apseudids and to date these animals are not known from Alaska to the southern border of Oregon. In localities in California and points south apseu- dids do not seem to be rare. About the only factors which might ac- count for their obscurity up to the present time are their small size and a paucity of investigators interested in the group. Seven genera are characterized in this paper. Two are described as new. Five of the genera are new to the fauna of the region. Seven- teen species are considered in this paper, of which thirteen are de- scribed as new to science. ACKNOWLEDGMENTS The writer expresses his appreciation to the Director of the Allan Hancock Foundation, Captain Allan Hancock, for his liberal support of this work. Special thanks are due Dr. John S. Garth, in whose laboratory this study was made, for generously yielding valuable re- search time to discuss with the writer the several taxonomic problems which became apparent during the investigation and for his assistance in the preparation of this manuscript. Mr. Al VanAuker, staff artist, provided the excellent copies of A. meridionalis and A. galapagensis (Fig. 1) which were taken from Richardson’s (1912a, 1912b) original MENZIES: EASTERN PACIFIC APSEUDID CHELIFERA 445 figures. The assistance of the Museum of Comparative Zoology in the publication of this paper is particularly appreciated. Key to the Families of the Tanaidacea A. First antenna without an accessory flagellum. B. Marsupium formed of one pair of oostegites which proceed from the proximal] inner margin of the fifth pair of peraeopods..... Tanaidae* B!. Marsupium formed by four pairs of oostegites which proceed from the proximal inner margin of the second to fifth pairs of peraeopods. seston cute sence uy SHEE aE ety SAREE SNS SEE Tare ele rte Paratanaidae* A!. First antenna with an accessory flagellum.................. A pseudidae Family APSEUDIDAE As can be seen from the key, the Apseudidae may be told from the Tanaidae and Paratanaidae due to their having an accessory flagellum on the first antenna. In addition, they usually have a scale attached to the second antenna and often have a triarticulate epipod attached to the first (gnathopod) and second pairs of peraeopods. The antennal scale and epipods are absent from the Tanaidae and Paratanaidae. Key to the Genera of Apseudidae Known from the Eastern Tropical and North Temperate Pacific Ocean A. Second antenna without a scale. B. Pleon with three somites including telson....... Synapseudes (p.461) B'. Pleon with six somites including telson....... Pagurapseudes (p.470) A‘. Second antenna with a scale. B. Mandibular palp with less than three articles. . Kalliapseudes (p.471) B'. Mandibular palp triarticulate. C. First somite of pleon much narrower than other somites of pleon Be apeat ret Be cd, aa>be, dd about = $C. Nephropores, when recognizable, on or near cd lines. Spermathecal pores, on 9/10-10/11, on c¢ lines. Female pores, on Xiv, on setal arc, slightly lateral to 6. Male pores, in a transversely slit-like depression of a marked tumescence which extends nearly to b and ¢ and to setal ares of xiv and xvi, obliterating 14/15 and 15/16. Clitellum, on 3$xxix-xxxv (10), xxx-xxxv (49), xxx-3xxxvi (2), XXX-xXxxvi (3), in earliest stages recognizable only on xxxi-xxxlv or xxxv. Ventral margins, as indicated by disappearance of clitellar opacity, about at level of lateral margins of tubercula, in mid be. Tubercula pubertates, on xxxili-xxxiv (except on variant specimens considered below), with bilobed median margin, incision at site of GATES: EARTHWORMS OF ARNOLD ARBORETUM 519 33/34. In earliest stages each tuberculum is represented by two swollen areas, of shortly elliptical outline, not quite in contact at 33/34. Genital tumescences, include ab: on ix (85), xi (84), xii (97), xxix (28), xxx (71), xxxi (70), xxxii (104), xxxv (89), xxxvi (31), xxxvii (13). In 14 specimens the region of ab of one side of x appeared to be slightly swollen, and in another worm there seemed to be a slight tumescence on each side of x. Internal anatomy. Calciferous sacs, paired, large, in x (10). Calci- ferous glands, in xi-xii (10). Intestinal origin, in xv (10). Typhlosole, begins gradually in region of xx, at maximum development nearly filling intestinal lumen, gradually decreasing in size posteriorly. An anterior portion looks like a pile of coins with a string on the ventral face but posteriorly the lateral lamellae gradually disappear. The typhlosole ends abruptly in region of cili-cix: cili, worm with 118 segments; civ, with 124 segments; cix, with 125 segments. Last hearts in xi (10). Often there appears to be a pair of hearts in xii, as large as or larger than those of xi, but the vessels pass from the dorsal trunk anteroventrally, through 11/12 and then anteriorly median to the hearts of xi. Subneural trunk present. Seminal vesicles, in ix—xii (10), those of ix—x smaller and often with brownish masses therein. Spermathecae project forward freely into ix and x and, when filled with sperm, may touch 8/9 and 9/10 or even be bent upwards. Distribution. Previously reported from: Hamburg, (and Holstein?) Germany; Ziirich, Switzerland; and three towns in Belgium. Original home still unknown. Life history. Through early May at least, the clitellum of these worms seemed not to be fully developed, the epidermal thickening often less pronounced than might be expected, and rudiments of inter- segmental furrows still faintly visible. Nevertheless 11 of the early May specimens had spermatophores externally, and other worms had spermatozoa in the spermathecae, showing that copulation had taken place. Amputation and regeneration. Posterior amputees, 6 (+1?). Pos- terior regenerates, 1— of 11 segments, at 104/105. Probably a few segments only had been lost in each case. Abnormality and homoeosis. Metameric abnormality: involving xi-xvi and probably result of some parietal damage; v—vi split on right side and in addition a spiral abnormality posteriorly; in xii-xviil; 520 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY spiral abnormality involving xxxili-xxxv. Several other specimens had metameric abnormality in a postclitellar portion of the body. Parasites. Several of the May specimens had parasitic masses in coelomic cavities of various segments in a postclitellar portion of the body. As yet there had been no extensive accumulations of such masses in the last few segments. Autotomy following accumulation of such masses seems to be an adequate explanation for the cases of posterior amputation observed. Remarks. Yn the arboretum sites at which A. limicola was obtained, no castings were found. Presumably all intestinal ejecta are deposited underground. . Variant individuals Peat bog, March—April—May, 4 aclitellate and 15 clitellate specimens. Segments, (of four normal worms): 123 (2), 124 (1), 129 (1). Clitellum, on 4xxix-xxxv (2), 43xxix—3xxxvi (3), xxx-}xxxvi (4), XXX-xxxvl (5), xxxi-xxxv (1, but clitellar glandularity only very slightly developed). Tubercula pubertates, on clitellate worms, ex- tending also across xxxv, as well as xxxiii-xxxiv; on left side only (4), on right side only (2), on both sides (9). On each of the six asym- metrical specimens that portion of the tuberculum on xxxv is slightly smaller than the other two parts, not reaching as far mesially as on xxxlll and xxxiv. A similar condition prevails on xxxv of two of the symmetrical clitellate specimens. On other clitellate individuals the tubercula have a trilobed median margin, incised at 33/34 and 34/35. On four aclitellate worms, with intersegmental furrows as yet unin- terrupted, tubercula are each clearly composed of three equisized circular portions. Genital tumescences, of 15 normal specimens: on 1x) @5); x13) pex(15)), xxix (3), sexx (112) ocx i) xox xoxox (9 + 12), xxxvi (8 + 1?), xxxvii (5). One worm appeared to have a tumescence on one side of x. The typhlosole ends in civ, in worm with 123 segments; cv, in worm of 124 segments; cix, in worm of 123 segments; exii, in worm of 129 segments. Spermatophores were present on one of the May worms. Posterior amputees, 2. Posterior regenerates, 1 — of 12 segments at 101/102. Abnormality. Two specimens have metameric abnormalities in a postclitellar region of the body. GATES: EARTHWORMS OF ARNOLD ARBORETUM 521 Remarks. The asymmetrical specimens with a tuberculum of one side limited to xxxili-xxxiv as in the worms previously characterized, provide a transitional condition between the symmetrical worms with tripartite tubercula and those with bipartite tubercula. In these cir- cumstances naming of worms with one or both tubercula tripartite seems unnecessary. ALLOLOBOPHORA LONGA Ude 1885 Railroad bog, early May, 7 clitellate specimens. Late May, 1 juvenile, 1 aclitellate and 6 clitellate specimens. Peter’s Hill, late May, 22 juvenile, 4 aclitellate and 3 clitellate specimens. (16 other juveniles probably belong here.) Newton, garden, April, 7 aclitellate and 1 clitellate specimens. Early May, 21 clitellate specimens. Late May, 1 clitellate specimen. Segments, 183, 186, 193, 203. The posteriormost segments have a tendency to flatten dorsoventrally somewhat as in L. terrestris L., though not so marked as in that species. Pigmentation, brown. Two specimens from Peter’s Hill had a deep blue appearance anteriorly and several others had a less marked bluish appearance of that part of the body. Clitellum, on 3$xxvii-xxxv (2), xxvili-xxxv (8), xxix- xxxv (1). Tubercula pubertates, on xxxii-xxxiv, except on one worm on which the right tuberculum extended across all of xxxv, the left reaching slightly onto xxxv. Genital tumescences, include ab, (of 48 worms): on viii (2), ix (48), x (48), xi (48), xxviii (4), xxix (2), xxx (1), XXxl (45), xxxili (47), xxxiv (47). Male tumescences obliterate 14/15 and 15/16. Distribution. A. longa has been reported once each from Maine and Connecticut but without further specification as to locality, and from Greencastle, Indiana. Life history. On a number of worms, with no indications whatever of development of clitellum or tubercula pubertates, rudiments of genital tumescences are obvious, on ix—xi only, or also posteriorly. Most of the specimens characterized as clitellate actually have little or no tumescence of clitellar epidermis. Clitellar segments are however clearly indicated by a much darker brown or more yellowish coloration than that of other segments. On eight specimens only is the clitellum well developed. 22) BULLETIN: MUSEUM OF COMFARATIVE ZOOLOGY Amputation and regeneration. One specimen had lost its first four segments, a ventral part of v, and a little of the anterior margin of vi ventrally. Missing parts of v-vi had been somewhat abnormally re- placed and a new head of three segments regenerated. Another worm has a head regenerate at 3/4 but metamerism is abnormal (furrows quite irregular, setae and nephropores unrecognizable). Tail regenerates are as follows: at 100/101, of 100 segments; 101/102, 98 segments; 117/118, 84 segments; 132/133, 70 segments; 133/134, 70 segments (juvenile from Peter’s Hill); 186/137, 59 segments; 153/154, 29 segments; 154/155, 47 segments; 170/171, 14 segments. Head regenerates have a uniform, light pinkish appearance, with delicate tissues, and presumably are of fairly recent age. Tail regener- ates, on the contrary, all appear to be older though known recent tail regenerates of this species are not available for comparison. In France, in A. longa, head regeneration is supposed to be possible all year round with tail regeneration possible only during a summer diapause. Presumably then the two heads had been regenerated during the winter and the tails during the previous summer. Amputations of the garden specimens can be attributed to cultiva- tion. Some other agency presumably was involved in case of arboretum amputations for both sites from which they were obtained showed no indications of recent disturbance. The posterior amputations appear to have been too extensive for autotomy resulting from parasitic activity. Moles were not noticed. As earthworms of this species are supposed to cast on the surface of the ground, tails presumably could have been lost to birds while casting. Remarks. Earthworms ‘‘throw up plenty of castings in the United States”, according to Darwin (1881, p. 121). No authority was cited for that statement. We have to assume that the information on which it was based was obtained from correspondents who, quite probably, lived in a northeastern portion of the country. In that part of the States, only earthworms of the family Lumbricidae would be involved. Of the lumbricids of that region worms of only one species, A. longa, are known to deposit intestinal ejecta on the surface of the ground as “castings”. As already noted above, that species has been reported but three times from the entire country. (Worms of one other lum- bricid species, A. nocturna, are known to cast on the surface of the ground but that species has been recorded hitherto only from two localities in England.) GATES: EARTHWORMS OF ARNOLD ARBORETUM 523 During the months of March-April-May, as well as in an early por- tion of June, little casting was observed in the Boston region. Such casts as were found were small in size, usually associated with rather small juveniles, and altogether seemed to be of insignificant volume especially in comparison with that to be seen in many parts of the oriental tropics after beginning of the rainy season. In the Boston region, A. /onga must have a fairly extended, ex- ternally imposed period of winter rest in addition to its internaily imposed or obligatory summer diapause. Or, in colonizing a region with a different climate have ancient habits been modified? Genus DENDROBAENA Eisen 1874 DENDROBAENA OCTAEDRA (Savigny) 1826 Under litter, June 4, 1 clitellate specimen. Clitellum, on xxix—xxxiil. Tubercula pubertates on xxxi-xxxili. This specimen has no seminal vesicles. DENDROBAENA SUBRUBICUNDA (Eisen) 1874 Manure heap, April, 1 clitellate specimen. Railroad bog, May, | clitellate specimen. Under litter, June 4, 3 clitellate specimens. Clitellum, on $xxiv—}xxxii. Tubercula pubertates, on xxvili- xxx. Genital tumescences, on xvi (5), xxiv (1), xxvii (3), xxxi (8). Genus HISENIA Malm 1877 EISENIA FOETIDA (Savigny) 1826 Manure heap, April, 29 clitellate specimens. Under litter, June 4, 1 aclitellate and 15 clitellate specimens. (A number of other specimens died before reaching the laboratory and were not counted.) All worms from the litter are slender and brevicaudate. The bands without the red pigment are inconspicuous in the manure worms, searcely visible to unaided eye in worms from the litter. Genital tumescences, in front of xv, are too indistinct to be recognized with certainty. Behind xv, on worms from the litter, tumescences were recognized as follows: on xxi (8), xxii (8), xxiii (12), xxiv (4), xxv or 324 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY XXvi to xxxii or xxxili on each specimen. Posterior amputees, 2. Regenerates, none. The left tuberculum of one worm is on xxvii-xxix but the right is on XXVili-xxx (no abnormality of intersegmental furrows in clitellar region recognizable). One worm from the litter has a split metamere in the clitellar region. Every other worm from the litter has one or more metameric anomalies in the postclitellar region of the body. Remarks. In the manure heap worms were present in great profu- sion, the 29 being from a single double handful of the manure, and apparently in good condition. In the litter, one specimen was found that had been dead for several hours at least (the only dead worm en- countered in arboretum collecting) and other specimens from the same source seemed to be in poor condition. EISENIA ROSEA (Savigny) 1826 Natural woods, May, 2 clitellate specimens. Peat-bog, March-April, 2 clitellate specimens. May, 3 clitellate specimens. Railroad bog, May, 1| clitellate specimen. Newton, garden, late May, 1 clitellate specimen. Segments, 115, 118 (peat bog). Clitellum, on }xxv-xxxii (8), xxv—-3xxxli (1), xxv—-xxxii (1). Tubercula pubertates, on xxix—-xxx (1), on xxix-3xxxi (3, + 1 with tuberculum of one side extending clear across Xxxl), xxix-xxxi (5). In latter case, each tuberculum is clearly marked off into three parts. Genital tumescences; including ab, of ix (2), xii (4), xxv (1), xxvi-xxxii (10), xxxiii (1); including cd, ix (1), x (1), xii (4), xiii (1). Tumescences of xxvi-xxxii are united or continuous with clitellar opacity except where tubercula intervene. A more or less definite furrow between tuberculum and tumescences may be continued all round the tuberculum. Genus KISENIELLA Michaelsen 1900 EISENIELLA TETRAEDRA (Savigny) 1826 Ponds, April, 1 posterior fragment. Peat bog, March-April, 1 clitellate specimen. May, 3 clitellate specimens. Spermatophores were present externally on the May specimens. Clitellate worms are of the forma usually referred to as typica. GATES: EARTHWORMS OF ARNOLD ARBORETUM 525 Genus LUMBRICUS L. 1758 LuUMBRICUS CASTANEUS (Savigny) 1826 Natural woods, May, 3 clitellate specimens. Peter’s Hill, late May, under litter, 1 clitellate specimen. Railroad bog, late May, under litter, 14 clitellate specimens. Newton, garden, April, under leaf litter, 1 clitellate specimen. Segments, 64 (amputee), 72 (amputee), 81 (Peter’s Hill), 85, 88, 89, 90 (2), 91 (2), 92 (2), 98, 94 (2), 95. Clitellum, on xxviii—xxxiii (14), $XXVill-3xXxxill (2). Opacity extends about to the a line, occasionally a setae of xxix—xxxli included. Tubercula pubertates, on xxix—xxxil, with a narrowed extension of translucent portion onto anterior half of xxviii (16). Genital tumescences, include ab, on x (16), xii (1), xxiv (1). Tumescences are better developed or at least much more obvious than in L. rubellus. Homoeotics, 2. In one the right female pore is on xiii and the right male pore is on xiv (bog). In the other (natural woods), all organs are one segment in front of the normal location. There are several metam- eric abnormalities in the postclitellar region. Presumably develop- ment had taken place in unfavorable conditions (metameric abnor- malities resulting) and some time after hatching a posterior portion as well as an anterior portion had been amputated. The lost anterior part had then been replaced by a regenerate, hypomeric by one seg- ment, which had become indistinguishable from the substrate. As the last segment (Ixiv) had undergone little change, posterior amputa- tion may have been more recent. LUMBRICUS RUBELLUS Hoffmeister 1843 Natural woods, May, 3 aclitellate and 12 clitellate specimens. (18 juveniles may be of this species.) Peat bog, March-April, 5 clitellate specimens. May, 2 juvenile, 2 aclitellate and 11 clitellate specimens. (3 juveniles probably belong here.) Railroad bog, early May, 19 clitellate specimens. Late May, 4 clitellate specimens. Ponds, April. (10 very small to small juveniles may be of this species.) Under litter, June 4, 6 clitellate specimens. (Other specimens apparently of this species died before reaching the laboratory.) Newton, garden, April, 10 clitellate specimens. 526 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Segments, 58, 69, 78, 81, 84 (2), 93, 95, 98, 99, 101 (52 x 4 mm.), 106, 107, 108, 109 (2), 110, 111 (8, including one juvenile), 112 (3), 113 (2), 114, 115 (2), 116 (4), 117 (3), 119, 120 (2). In the peat bog series number of segments varied from 69 to 116, and in the railroad bog series from 93 to 120. A juvenile, 35 x 3 mm., and without indica- tion of amputation, had 69 segments. The largest specimen, ca. 75 x 5.5 mm., had 116 segments. The anal segment in most specimens seems rather large, and is pigmented. Clitellum, on xxvii-xxxi (2), xxvii-}xxxil (10), xxvii-xxxii (21), $xXvii-4xxxil (5). Tubercula pubertates not distinct from but ap- parently united with the clitellum. In a number of specimens the ventral margin of clitellar opacity is slightly median to the a lines, with the ab setae of xxvili—xxxi included in the clitellum. In other worms the opacity does not quite reach the b lines. The tubercula of these specimens are recognizable only as translucent areas within the clitellar opacity. These translucent areas may be limited to xxvili—xxxi (3), or a narrowed extension may cross well into xxvii (35). On at least a dozen specimens vestiges of intersegmental furrows are recognizable across the translucent areas. Genital tumescences, include ab: on vil (3); 1x (2), x (7), xi (17), xi (1), xin 1), x1 and xi @)) xxv) exe (35), xxvii (2), xxix (1), xxxii (1). Anterior to xv tumescences may be entirely lacking, or present only on one side of one segment, and often are only slightly developed and not easily recognizable. Worms in copula were found in the leaf pile (natural woods) late in the morning of May fifth. Many of the clitellate specimens of the same lot had spermatophores externally. Amputees, 2 only recognized as such. One posterior regeneration of a two-segment tail at 51/52 (Newton, garden). Homoeotics, 2. In one, the left male pore is on xvi. Setae are lack- ing on left sides of xv—xvi but intersegmental furrows, septa and other organs are normal. On the other worm the right male pore is on xiv lateral to the female pore. Remarks. A number of the specimens are obviously brevicaudate, with clitellum about at middle of the body. Even in those worms in which the clitellum seems well in front of the middle of the body, the number of segments is well below the species maximum of 150. ~I GATES: EARTHWORMS OF ARNOLD ARBORETUM 52 LUMBRICUS TERRESTRIS L, Peat bog, March-April, 8 clitellate specimens. (8 juveniles may belong here.) May, | aclitellate and 1 clitellate specimens. Natural woods, April, 10 clitellate specimens. May, 4 juvenile, 1 aclitellate and 10 clitellate specimens. Peter’s Hill, early May, 2 juvenile and 4 clitellate specimens. (5 small juveniles probably belong here.) Late May, 5 juvenile and 3 clitellate specimens. Under litter, June 4, 1 clitellate specimen. Newton, garden, late May, 5 clitellate specimens. Length, 125-210 mm. Diameter, 7-12 mm. Segments, on juveniles of about adult size but without indication of sexual maturity: 149, 152, 155, 157. A somewhat smaller juvenile has 145 segments and one that is 120 mm. long has 108 segments, the last three of which apparently lack setae. Clitellate specimens from the bog (and Peter’s Hill) are obviously brevicaudate; one, measuring 130 x 7.5 mm., has 98 seg- ments. On none of the brevicaudate forms was evidence of posterior amputation or subsequent reorganization recognized. Nephropores of a side, in Peter’s Hill specimens, occasionally are in one location and in a nearly straight rank for several successive seg- ments, aS Many as nine and eleven noted. Clitellum, on xxxii-3xxxvii (1), xxxii-xxxvil (25), xxxii-}xxxviii (3), xxxli-xxxvili (1). Tubercula pubertates of aclitellate worms are markedly protuberant and obviously of quadripartite origin — inter- segmental furrows still continued deeply across the tumescence. In clitellate specimens the tubercula are usually recognizable only as areas of sharply delimited translucence. In some cases a peripheral rim portion of opacity is more or less distinctly marked off. Tubercula reach across all or almost all of xxxili-xxxvi (11), or reach slightly onto xxxvii (1), or across all of xxxvii (1, worm with clitellum reaching to 38/39). In other specimens, the tubercula, even including any specially opaque rim portion, fall quite definitely short of reaching 32/33 and 36/37. Genital tumescences, include ab: on xxv (9), xxvi (22), xxvii (2), xxvili (1), xxix (1), xxx (6), xxxi-xxxvili (82). Tumes- cences may also be present on some of segments vii-xii but if so are feebly developed and not certainly recognizable. Amputees, none recognized. Regeneration: posterior, none; anterior, one. In this case (Newton, garden), i-iv and probably part of v on right side had been lost. The regenerate is a functional head, pinkish — without pigment, tissues still delicate. The prostomium is only slightly 528 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY epilobous. Metamerism is quite abnormal, no normal intersegmental furrows at all, setae and nephropores lacking. Abnormality and homoeosis, 3. In one worm the prostomium is practically prolobous. Pigmentation is norma! and tissues are of usual firmness. If the unusual condition is due to regeneration, the regen- erate is of an age as to be no longer recognizable as such. In another worm the left male pore is on xiv, lateral to the female pore. The left calciferous gland of xii is lacking (intersegmental furrows, septa and other organs normal). The third worm has eight spermathecal pores, about on the c and d lines of 9/10 and 10/11. Each of the eight sperma- thecae is markedly iridescent indicating presence of spermatozoa re- ceived in copulation. Each spermathecal duct is separated from that of its near neighbor by a sizable strand of longitudinal musculature. Genus OCTOLASIUM Orley 1886 OcTOLASIUM CYANEUM (Savigny) 1826 Peter’s Hill, May, 2 clitellate specimens. Railroad bog, early May, | clitellate specimen. Newton, garden, late May, 4 aclitellate and 2 clitellate specimens. Segments, 151, 155. The last few segments, in living worms, were characterized by a brilliant yellowish appearance. During preserva- tion most of the yellow masses responsible for that appearance were discharged from the coelomic cavities through the dorsal pores. Genital tumescences, include ab: on xviii (1), xx (1). Distribution. O. cyaneum has been reported once before from this country, from Fairfield, Iowa. Ocrotastum LacteuM (Orley) 1881 Peat bog, March-April, 2 clitellate specimens. May, | clitellate specimen. Peter’s Hill, May, 3 clitellate specimens. Newton, garden, late May, 3 clitellate specimens. One of the worms from the Newton garden has a brownish coloration similar to that of A. longa. Genital tumescences, include ab: on xxi (4), xxii (6). Worms in copula were found in the Newton garden late in the afternoon, in late May. GATES: EARTHWORMS OF ARNOLD ARBORETUM 529 Each of the Peter’s Hill worms has parasitic cysts in considerable numbers in the coelomic cavities of the postclitellar portion of the body. The single amputee, which had lost its tail at 96/97, may have auto- tomized the last segments to get rid of an accumulation of such para- sitic masses. Distribution. Reported once before in New England, from an unspecified locality in Connecticut. Family GLOSSOSCOLECIDAE Genus SPARGANOPHILUS Benham 1892 SPARGANOPHILUS EISENI Smith 1895 Ponds, April-May, over a hundred specimens. Amputation and regeneration. In a series of 61 specimens: very recent amputees, posteriorly, 28; with small, unsegmented posterior regenerate about the size of one segment, 15; with a metamerically segmented tail regenerate, 9. Although these worms appear to be unpigmented the substrate is sharply distinguished from the regenerate by an almost blackish ap- pearance. One worm had been amputated posteriorly twice. The second regenerate is unsegmented but of quite different color from the first. One worm had lost a posterior portion of a regenerate very re- cently, possibly in collection. In fact all of the very recent amputees may have been broken in course of collection. Parasites. A number of the worms had masses of parasitic bodies in the coelomic cavities of the postclitellar portion of the body. Remarks. The longest of the worms with a normal posterior end has a very obvious and large growth zone in front of the anus. Family MEGASCOLECIDAE Genus PHERETIMA Kinberg 1866 PHERETIMA sp. Under litter, June 4, 1 small juvenile. Intestinal caeca, manicate. Remarks. The only species of Pheretima hitherto found in Massa- chusetts (in greenhouses, near Northampton), P. hawayana (Rosa) 530 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY 1891, has simple intestinal caeca. Two species with manicate caeca have been recorded from North America: P. schmardae (Horst) 1883, from the Bermudas and Barbados; and P. agrestis (Goto and Hatai) 1899, from Baltimore. DISCUSSION Eighteen species are represented among the thousand odd earth- worms collected from only six sites within the confines of the Arnold Arboretum in the city of Boston. Of these species, A. arnoldi, chloro- tica, limicola, longa, molita, D. octaedra, O. cyaneum and lacteum are recorded from Massachusetts for the first time, A. limicola for the first time outside of Europe, A. arnold: and molita for the first time anywhere. Most of the species that might have been anticipated were obtained, except the following: Bimastos parvus (Eisen) 1874, B. tenuis (Eisen) 1874, Eisenia lénnbergi (Michaelsen) 1894, each of which was already known from Massachusetts; D. rubida (Savigny) 1826, already known from Maine and New Hampshire, Eisenia hortensis (Michaelsen) 1890, California and Ohio, Lwmbricus festivus Savigny 1826, several Canadian provinces, and Dendrobaena mammalis (Savigny) 1826, intercepted on plant materials imported into this country (Gates, MS). A. limicola, of course, as well as arnoldi and molita, were not anticipated. Nor was any species of Pheretima expected, as species of that genus have hitherto been known, in the colder parts of this country, mainly from greenhouses, as in the case of P. hawayana (Rosa) 1891 at Hadley, Mass. The important exception is P. hwpeiensis (Michaelsen) 1895, from golf courses of western Connecticut to Washington, D. C. Among the Lumbricidae of the arboretum are species such as A. caliginosa and FE. foctida which seem to have been recognized almost everywhere that Europeans have settled. In several of such areas, that have been searched, about thirteen other lumbricids have been found. Occurrence of any of those fifteen species in the arboretum, ac- cordingly, is of no particular significance; all may have been present in the Boston region before the arboretum was established late in the last century. Nor is importance attached to failure to find species such as I. hortensis or B. parvus in view of the limited number of sites stud- ied and the small volume of material that was dug up or searched through. The presence in the arboretum, and there alone, of A. limicola ap- GATES: EARTHWORMS OF ARNOLD ARBORETUM ddl parently is to be explained as a result of the introduction of exotic, live plants in soil. Other such introductions presumably have been responsible for presence of Sp. eisent and the species of Pheretima. The original sources of the 22 Massachusetts species must be looked for in three widely separated parts of the world. The most important, of course, is Europe which has provided all species of the genera Allo- lobophora (arnoldi and molita for the present considered to be of extra- American origin), Dendrobaena, Evseniella, Octolasiwm, Lumbricus, and two species of Eisenia. Another source — southwest of New England and New York — has contributed E. lonnbergi, presumably one species of Bimastos,! and the glossoscolecid Sparganophilus eisent. The last source, of the species of the megascolecid Pheretima, must be somewhere in that large area comprising Burma, Malaya, Thailand, Indo-China, China, Korea, and the islands of Japan, Malaysia, etc., but excluding Australia and New Zealand. (The two last, with India, Africa and South America, have apparently contributed nothing to the arboretum fauna.) Some of the arboretum species appeared to be restricted as to habitat. Thus, Sp. eisent was found only in bottom mud, A. limicola only in bogs, L. castaneus only under litter or in piles of decaying leaves, and E. tetraedra only in saturated soil. E. foetida, sold for ten years throughout the country as a garden, orchard or farm cultivator of the soil, was found only in manure or under thick leaf litter. Such few specimens of D. octaedra, subrubicunda and A. chlorotica as were ob- tained were also found only in manure and under or near litter. Other species seemed to have no such restrictions, and were found in saturated bog soils, much drier garden soil, and, in certain cases, away from soil, in leaf piles. A number of such species were usually found in close neighborhood with each other as well as with those of more re- stricted habitat. In five of the sites the number of species varied from six to ten (see Table V). Almost certainly extension of collecting only a few feet in one direction or another would have yielded one to three more species in case of the railroad bog, Peter’s Hill and the natural woods. As the table shows, nine species were obtained at three or more of those five sites, and two, A. arnoldi and caliginosa, at all of them. The one factor obviously common to all of those places was presence of a considerable amount of organic material (plant). 1 The other “‘species’’, B. tenuis, is at present, rather tentatively considered an athecal mutant of D. rubida and/or subrubicunda. Several specimens recently studied apparently can be so regarded (Gates, MS). 532 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY TABLE V SITES Natural Garden Peat Railroad -Peter’s woods in Species bog bog Hill leaf pile Newton A. arnold — +D + + +D caliginosa + + + + + chlorotica + limicola +D +. longa + a ar D. subrubicunda + E. rosea + + + tetraedra + L. castaneus - — + + rubellus + + + + terrestris + + +D + O. cyaneum + + ae lacteum + + + Number of species 8 10 of 6 9 D, dominant In each of the sites the earthworm population probably could be characterized as dense, somewhat less so perhaps in the Newton garden than in the other places. A. limicola clearly was the dominant species in an apparently undis- turbed portion of the peat bog next to the brook. The same earth from which all worms presumably had been removed in previous collecting was again dug over early in May. During the intervening period 50 (+) worms, of three species only, had moved into the disturbed soil. Of those, 44 were limicola, the rest belonging to the calzginosa complex. In the railroad bog, where there had been considerable filling, A. limicola was much less important, ranking fourth in the following series of descending order: A. arnoldi, caliginosa, L. rubellus, A. limicola, longa. A. arnoldi again appeared to be dominant in the Newton garden followed by longa, caliginosa, rubellus and L. terrestris in descending numerical importance. L. terrestris probably was domi- nant in the leaf pile of the natural woods, not only in number (only a few of the specimens that were seen having been brought to the laboratory and counted) but also in volume — all of the adults un- usually large and many of them 2 to 3 mm. thicker than the maximum recorded by Cernosvitov and Evans (1947). Next in numerical order GATES: EARTHWORMS OF ARNOLD ARBORETUM Sia) was the calzginosa complex and then rubellus. No specimens of E. foetida were found in that leaf pile or nearby. Yet, sometime later, some fifty or so feet away, under rather thick natural litter, foetida was found and in about the same number as rubellus. In the leaf litter site of the natural woods, presence of one dead worm and of one or more metameric anomalies in every specimen examined, seemed to indicate some unfavorable factor in the environ- ment, in spite of the numbers present. In other sites, worms appeared healthy and provided no indication of unusual incidence of homoeosis or abnormality in spite of low pH. However, the percentage of more or less obviously brevicaudate individuals in the leaf pile and bogs seemed unusually high. In copulation, participating individuals always have been clitellate, in each species of earthworm that has been studied — LE. foetida, L. terrestris, Eutyphoeus waltont Michaelsen 1907, and Pheretima com- munissima (Goto and Hatai) 1899. A clitellum was also present on each individual of the copulating pairs of L. rubellus and O. lactewm mentioned on preceding pages. It is therefore of interest that some arboretum specimens of A. _arnoldi and caliginosa seemed to have copulated before development of the clitellum, or, in case of A. limicola, when the clitellum was only very slightly developed. Proof of copulation was provided by presence of “‘spermatophores” externally and of spermatozoa in the sperma- thecae. Iridescence on male funnels indicated, in at least one case, that the individual’s own sperm had been mature. After beginning of clitellar regression, in some species, presence of mature spermatozoa may still be indicated for some time by iridescence on male funnels and in spermathecae. No indications of clitellar regression were recognized in any of the arboretum specimens under consideration. SUMMARY Of 18 species collected in the arboretum, eight — Allolobophora arnoldi, chlorotica, limicola, longa, molita, Dendrobaena octaedra, Octo- lasium cyaneum and lactewm— are recorded from Massachusetts for the first time, cyanewm for the first time in New England, limzcola for the first time outside of Europe, arnoldi and molita for the first time anywhere. A species of Pheretima was found outside of greenhouses fad 534 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY for the first time in Massachusetts. Presence of A. limicola, Spargan- ophilus eiseni and Pheretima sp. is considered to be the result of importation of live plants from three widely separated parts of the world. A. limicola, restricted to bogs, is dominant in an undisturbed peat bog, arnoldi in a bog with fill as well as in a Newton garden, L. terrestris both by number and volume in a leaf pile. REFERENCES Crrnosvirov, L. and A. C. Evans 1947. Lumbricidae. No. 6 in Synopses of the British fauna. London, Linnean Society. Darwiy, C. 1881. The formation of vegetable mould through the action of earth- worms. London, J. Murray. Evans, A. C. 1946. On a new species of earthworm of the genus Allolobophora. Ann- Mag. Nat. Hist. (11), 18: 98-101. 1948. On some earthworms from Iowa, including a description of a new species. Ann. Mag. Nat. Hist. (11), 14: 514-516. Gates, G. E. 1952. New species of earthworms from the Arnold Arboretum, Boston. Breviora, 9: 1-3. GavRILov, K. 1937. Regenwiirmer aus Siidfrankreich. Zool. Anz., 118: 145-154. Manuscript received for publication September 16, 1952. Loy iy a An eat! : Anan rmsd Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vou. 107, No. 11 A NEW FOSSIL TORTOISE FROM THE THOMAS FARM MIOCENE OF FLORIDA By Ernest WILLIAMS CAMBRIDGE, MASS., U. S. A. PRINTED FOR THE MUSEUM February, 1953 PUBLICATIONS ISSUED BY OR IN CONNECTION WITH THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE BULLETIN (octavo) 1863 — The current volume is Vol. 107. BREVIORA (octavo) 1952 — No. 9 is current. Memorrs (quarto) 1864-1938 — Publication was terminated with Vol. 55. JOHNSONIA (quarto) 1941 — A publication of the Department of Mollusks. Vol. 2, no. 30 is current. OccASIONAL PAPERS OF THE DEPARTMENT OF MOLLUSKS (octavo) 1945 — Vol. 1, no. 16 is current. PROCEEDINGS OF THE NEw ENGLAND ZOOLOGICAL CLUB (octavo) 1899- 1948 — Published in connection with the Museum. Publication terminated with Vol. 24. These publications issued at irregular intervals in numbers which may be purchased separately. Prices and lists may be obtained on application to the Director of the Museum of Comparative Zoélogy, Cambridge 38, Massachusetts. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE VO, UO; ING. 1 A NEW FOSSIL TORTOISE FROM THE THOMAS FARM MIOCENE OF FLORIDA By Ernest WILLIAMS CAMBRIDGE, MASS., U.S. A. Je JE TEINY 4013) 1D) IPO) 1 AP IBLITT WE UPSD Wy Wl February, 1953 No. 11-—A New Fossil Tortoise From the Thomas Farm Miocene of Florida By Ernest WILLIAMS Several plastra, a nearly complete carapace and additional frag- ments of a tortoise from the Miocene of Thomas Farm, Gilchrist County, Florida, pertain to an undescribed form apparently ancestral to the giant tortoise of the Florida Pleistocene, Testudo sellardst. The new species may appropriately be named in honor of Dr. Theodore White whose work on other components of the Thomas Farm fauna is so well known: TESTUDO TEDWHITEI, new species Type. M.C.Z. No. 2020, a complete plastron. Type locality. Thomas Farm, Gilchrist Co., Flerida. Horizon. Arikareean Miocene. Diagnosis. With the characters of Testudo sellardsi Fay (as described by Loomis 1927) but much smaller, so far as known not exceeding 400 mm. in plastral length. Pectoral scute 1/6 to 1/7 the abdominal in length of median sulcus; gulars more triangular, anals smaller; nuchal scute well developed, reaching anterior margin; costo-vertebral sulci less deeply incised; free margins less reverted. Referred material. M.C.Z. No. 2021, “plastron 2”’. M.C.Z. No. 2022, “plastron 3”’. M.C.Z. No. 2023, “plastron 4’’. M.C.Z. No. 2024, “plastron 5”’. M.C.Z. No. 2025, a carapace lacking some of the left peripherals, O38 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY neurals 4 and 5, part of pleural 3 and all of pleural 5 of the left side. M.C.Z. No. 2026, a miscellaneous lot containing three anterior cara- pace margins. a posterior lobe of a plastron, a small femur, a humerus, a scapula and acromion, and neurals, peripherals and other carapace fragments not further identified. Character Analysis 1. A feature of considerable interest in the new species is the size of the pectoral scute, especially its medial anteroposterior dimension. In 1950 I used a difference in this dimension to assist in distinguishing Miocene [Hesperotestudo! from Miocene Gopherus and Miocene Chelo- novdis, ascribing to the first a pectoral scute with ‘‘median suleus 1/15 to 1/30 of the abdominal median sulcus”, to the second a pectoral scute with ‘‘a median sulcus usually about 1/5, never less than 1/10 of the abdominal median sulcus’’, and to the last a pectoral scute with the “median sulcus 1/5 the abdominal median sulcus”. Testudo ted- white’ has the following measured values for pectoral and abdcminal median sulci: Specimen pectoral abdominal type 20 mm. 116 mm. plastron 2 20 118 ey 3 17 114 fe 4 18 129 This range of values gives ratios of approximately 1/6 to 1/7. According to the 1950 criteria this should rank Testudo tedwhite either with Gopherus or with Chelonoidis. But in major characters the new Florida form is not a Gopherus, and the presence of a nuchal scute excludes it from Chelonoidis. If we look now at Oligocene Hesperotestudo we find that 7. brontops had the pectoral median sulcus between 1/4 and 1/5 that of the ab- dominal, while 7. amphithorax had the same suleus between 1/3 and 1/4 that of the abdominal. Testudo tedwhitei is, therefore, intermediate in this character be- tween the Oligocene species of Testudo and those Miocene species of Testudo which previously have been well enough known to be assigned definitely to the subgenus Hesperotestudo. Very fortunately we are not dealing here with a single specimen in which such an aberrant or 1A subgenus of Testudo (type: Testudo osborniana, Miocene of Colorado) that includes all the North American members of the genus (Williams 1950). All living North American land tortoises belong to the genus Gopherus. WILLIAMS: A NEW FOSSIL TORTOISE 539 apparently aberrant character might well be an individual variation. The pectoral-abdominal ratio is clearly shown in four specimens and readily inferred in a fifth. In this regard 7’. tedwhitei seems clearly to manifest as a species character a condition more primitive than that found in 7. osborniana and its closer relatives. 7. sellardsi as de- seribed by Loomis (1927) is nearer in this feature to 7. osborniana than to 7. tedwhitei.' 2. In all five plastra of 7. tedwhitei the pectoral-humeral sulcus is separated from the entoplastron by a distance equivalent to about 1/2 to 1/3 the median length of the pectoral scute itself. In 7. osborniana and its relatives the entoplastron tends to be very close to or in con- tact with the pectoral-humeral boundary in spite of the extreme nar- rowing of the pectoral scute. In 7. brontops and T. amphithorax of the Oligocene the entoplastron is also in contact with the pectoral- humeral boundary, but that is less surprising here since the pectoral scute is quite large. The difference which 7. tedwhitci manifests in this respect from the other Miocene and the Oligocene species is not great and may probably be bridged by individual variation when more specimens are known. It does, however, seem to indicate a trend in 1. tedwhiter different from that in 7’. osborniana, ete. With regard to this character, 7’. sellardsi and T. tedwhitei are quite in agreement. 3. The entoplastron cannot be accurately measured in 7. tedwhitcei except in the type and in plastron 5. In the type, the width of the en- toplastron is 71 mm., its length 67 mm.; the same values in plastron 5 are 67 mm. and 59 mm. respectively. The entoplastral length is thus about 9/10 its width. In 7. osborniana and in the closely similar forms of the Miocene and Pliocene, the length was about 3/4 the width, as it was also in the Oligocene 7’. brontops. T'. amphithorax, on the other hand, departs in the other direction from the condition of T. tedwhitei, having the entoplastron slightly longer than wide (104 mm. wide, 108 mm. long, Hay 1908). According to Loornis’ figure the Jength of the entoplastron in 7’. sellardsi must have been about 9/10 its width, as in 7’. tedwhiter. 4. Four of five plastra? of 7. tedwhitei show the gular region scarcely distinct from the general contour of the anterior lobe. The same region is somewhat more developed as a projecting gular prominence in 7’. 1Loomis gives no measurements, but from his figure the pectoral-abdominal ratio must be between 1/12 and 1/15. 2In plastron 2 this area is missing. D540 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY osborniana and its group and much more developed in 7. arenivaga of the Lower Miocene and in 7. brontops of the Oligocene. Oligocene T. amphithorax and Eocene T. wintensis, on the other hand, have the gular prominence just as little distinct from the contour of the anterior lobe as in T.. tedwhiter. This is equally true of T. sellardsi. The striking similarity of the four plastra of 7’. tedwhitei in this as in other regards, permits much greater confidence in the use of these rather minor characters in this group than would otherwise be at all possible. The nearly identical differentiation of the gular region (at least in ventral view) in these four specimens is in strong contrast with the situation in Gopherus and in Chelonoidis in which individual and sexual differences in this region can be very great. 5. The dorsal aspect of the gular region, in the four specimens of T. tedwhitec in which this region is known, presents some interesting differences in the anteroposterior length of the swollen area and the degree of excavation of its caudal margin. The length of this ‘“‘epi- plastral lip” and the degree of excavation are directly correlated and, in three of the plastra, (the type and plastra 3 and 4) there is a con- sistent increase-in both characters with size. In plastron 5, however, though it is larger than the type, both the dorsal length of the lip and the amount of posterior excavation are less, so that this region is less differentiated than in any of the other plastra. The dorsoventral thickness of the lip again varies directly as the dorsal length. 6. The inguinal scute in 7. tedwhitei was apparently large and reached the femoral. It is satisfactorily discernible only on the left side of the type, but the partial sulci present in other specimens seem consistent with this description. A large inguinal scute is apparently characteristic of Hesperotestudo. A small scute is figured by Hay (1908) for T. amphithorax, but I have been unable to verify this on the type material at the American Museum. A small inguinal, not reaching the femoral, seems to be very characteristic of Stylemys nebrascensis, according to the many specimens I have examined. In Recent forms this character seems to be sufficiently constant to help in discriminating species groups; it may assist also with fossil forms. 7. The xiphiplastral notch is distinct and angular but not deep in the four plastra or partial plastra of 7. tedwhitei in which it is pre- served. In contrast, it is less distinct, wide and very shallow in 7’. osborniana and its close relatives. In both 7’. brontops and T. amphi- thorax the notch is very like that in T. tedwhite?. Loomis (1927) WILLIAMS: A NEW FOSSIL TORTOISE 541 specially commented on the distinctness of the notch in 7. sellards:; as he figures it, it is, indeed, very similar to that in 7. tedwhiter. 8. With regard to the characters of the carapace we are not as fortunate as with those of the plastron, since we have only one cara- pace which is even approximately complete. From this, however, we may frame a rough estimate of the size and shape of the shell. The length of the carapace may be estimated as about 370 mm., the width as approximately 300 mm.; it was therefore about 4/5 as wide as long and distinctly parallel-sided, not globular. The sides were quite vertical, and the height may have been in the neighborhood of 150 mm. In its parallel-sided contour T. tedwhiter was like T. sellards: and rather unlike 7. osborniana and its relatives which tend to a more nearly hemispherical carapace shape. 9. The anterolateral corners of the carapace flare above the limbs in 7. tedwhitei, but between these flared corners, the anterior margin is essentially straight. This condition is rather characteristic of the whole assemblage which I called Hesperotestudo in my 1950 paper. The general impression is one of an indentation of the anterior margin, but this is not due to a real notch at the nuchal region as in some other turtles but solely to the considerable flare of the anterolateral margins. 10. The nuchal scute is preserved in three specimens in 7’, tedwhitev. In the more complete carapace it is about twice as long as wide, but in the other two instances it is significantly broader, though still longer than wide. This is another feature in which it is in general agreement with the members of the broad group Hesperotestudo. T. sellardsi, on the other hand, was described by Loomis (1927) on the basis of the Amherst specimen as having a very small nuchal which did not reach the anterior margin. Material referred to 7. sellardsi at the M.C.Z. shows this scute to be of more normal Hesperotestudo-like character, longer than broad and reaching the anterior shell margin. 11. The vertebrals and neurals, so far as known, are essentially as in the other members of Hesperotestudo in which they are known. Vertebral 1 is very broad but does not reach, by a considerable interval, the second marginal of either side. Vertebral 4 1s not completely known but was evidently longer than wide. Neurals 2 and 4 are octagonal (seen in two specimens). The first neural is elongate, oval in the usual fashion; neurals 3 and 5 are quadrilateral; neurals 6, 7 and 8 are hexagonal, short-sided in front. p42 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY 12. The suprapygals are not distinctive; the pygal is rather narrow and bowed outward as in Loomis’ figure of Testudo sellardsi. 13. The pleurals in 7. tedwhitei are quite primitive in not displaying any trace of the alternate widening and narrowing of their distal ends, a feature characteristic of most advanced tortoises and seen well- developed in 7. osborniana and T. orthopygia, but scarcely developed in 7. brontops and not present in 7. amphithorax, T. impensa and T.. sellardsi. 14. The costo-vertebral sulei in 7. tedwhite: are distinct but not deep, by no means so incised as sometimes in 7’. sellardsi. The latter, however, is a giant form, and this feature of deeply incised sulci may be suspected to be correlated with its great size. The other Hespero- testudo which are sufficiently known, none of quite comparable size, never show this deeply incised condition of the scute boundaries. The anterior and posterior margins are also more reverted in 7’. sellardsi than in the other forms. 15. A humerus and a small femur, both of normal testudine char- acter, are referred to 7. tedwhitci.. The humerus has a slender rounded shaft without notable compression in any plane and with only a roughened area of attachment for the latissimus dorsi, not a pit. The femur, on the other hand, like that of 7. osborniana, has the shaft compressed in the plane of the head and hence the shaft is rather quadrate in section. Phyletic Relationships The diagnosis above of 7. tedwhitei has explicitly compared the new species with 7’. sellards? as described from a specimen at Amherst, by Loomis (1927). I have made this special comparison because I want to evade, in the present discussion, the issue of the identity of the Testuco sellardsi of Loomis, described from a ccmplete shell from Melbourne, Florida, with Testudo sellardsi Hay, the type of which is a xiphiples- tron from Vero, Florida. Loomis’ equating of his form with that of Hay may or may not be correct, but that point is not germane in the present instance.! If now we compare Testudo scllardsi, as so defined, with 7. os- borniana, the type species of Hesperotestudo, certain differences are 1T feel it necessary to deprecate the description by anyone, even by one so experienced as Hay, of a new species of tortoise from material as poor as the type of T. sellardsi. Especially is it unfortunate that Hay described two species of giant tortoise from Vero from different parts of the shell! WILLIAMS: A NEW FOSSIL TORTOISE 543 evident. These differences, however, are not major ones. It would seem initially probable that they are merely specific and not group differences, since they are outweighed by the resemblances held in common between 7’. sellardsi and T. osborniana, and which separate them very clearly indeed from the species of Gopherus. It therefore seemed logical, in 1950, to infer the descent of the Florida giant tortoise from 7’. osborniana or T. osborniana-like ances- tors. But the discovery of 7. tedwhitet changes this picture radically. It is obviously, in view of its many special resemblances, a much more suitable Miocene ancestor of the Pleistocene giant than 7’. osborniana or its close relatives. So with the entrance of 7. tedwhitei into the picture, we see not one but two phyletic lines within Testudo in North America, separate at least since the Lower Miocene. It is now desirable to set down as a first approximation the features which seem to distinguish these two phyletic lines. Thus, the 7. tedwhitei lineage appears to differ from the 7’. osborniana series by having: (1) the carapace parallel-sided rather than rounded; (2) the xiphiplastral notch distinct and angular rather than indis- tinct and rounded; (3) the pectoral scute appreciably posterior to the entoplastron rather than in contact with it; (4) the entoplastron about as wide as long rather than noticeably wider than long; (5) the gular region less differentiated; (6) the pectoral scute less narrowed. None of these differences can be conceived of as an absolute differ- ence. We are dealing not with key differences but with assemblages of characters that, in my Judgment, empirically set apart two groups of species. These two groups of species are indicated as closely related because of the common possession of the following characters: 1. a nuchal scute longer than wide; 2. a pectoral scute tending (with time) to become more and more narrowed ; 3. an elongate fourth vertebral scute; 4. a differentiated neural sequence early acquired (in the time series) ; 5. markedly flared anterolateral and posterolateral margins; 544 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY 6. a more convex shell than that of compared forms. In the 7. osborniana series in which alone skulls associated with shells are known, the characters listed above are associated with dis- tinctive features of the premaxillary alveolar surface (a pit without a median ridge) and of the external surface of the dentary (fine vertical ribbing). These skull characters are assumed to hold for the 7. ted- whitei series also. In thus assuming two closely related parallel series we have two potential sources of confusion. (1) In later members of the two series, parallel variation may make closer the resemblance of forms which were distinct over a long period. Especially is this probable if, as in the pres- ent case, there appear to be similar trends with time (as in the narrow- ing of the pectoral scute) but trends pursued at different rates in the two series. Reversal of evolutionary trend is also a possibility. (2) The earlier members of the two series, as they approach in time their common ancestor, should be progressively less and less distinct one from the other. Some instances of the first possibility of confusion may be expected to turn up as knowledge of Tertiary and Quaternary tortoises in- creases. The other point we may consider at this time in terms of the known Oligocene and Eocene tortoises. In the Oligocene two very distinct species of T'estudo have been described: 7. brontops and T. amphithorax. On the character of the gular region, 7’. brontops with this region highly differentiated belongs with the 7’. osborniana series, and 7. amphithorax with the same region not at all distinct, belongs with the 7. tedwhitei series. The width- length ratio of the entoplastron arranges the two species in the same way. The pectoral scutes in both are very much wider than in Miocene forms, but in 7. brontops the ratio of pectoral length to abdominal length is somewhat less than 1/4, in T. amphithorax somewhat more than 1/4, possibly an indication that the trend to greater narrowing of the pectoral was present in 7’. brontops to a somewhat greater degree than in 7’. amphithorax. This again would be consistent with a posi- tion of 7’. brontops in the T.. osborniana line, T. amphithorax in the T. tedwhite? line. On the other hand, in certain other characters on which it is possible to distinguish the Miocene forms, no distinction is possible in the case of the Oligocene forms. In both species the pec- toral-humeral sulcus is in contact with the lower margin of the ento- plastron; in both the xiphiplastral notch is distinct and angular. T. WILLIAMS: A NEW FOSSIL TORTOISE 545 brontops had the carapace parallel-sided though possessing a wide shell; the shell of 7. amphithorax, though not completely known, was prob- ably narrower but also parallel-sided. In the Eocene, only 7. wintensis Gilmore has previously been re- ferred to the genus 7’estudo. In that form the gular region is not more differentiated than in 7. amphithorax or T. tedwhitei; it should therefore belong to that series. The entoplastron is about as wide as long in the unique specimen of the species; this also might count it as a member of the 7. tedwhitei series. On the other hand, the carapace has a rounded rather than a parallel-sided contour; this might place it in the 7. osborniana series. In still other respects it is much more primitive than any of the forms previously cited. The pectoral scute is about half the abdominal scute in length. The supracaudal scute is divided as in emydines or Hadrianus. If 7. wintensis is placed as a member of the 7. tedwhitei series on the basis of the absence of gular differentiation, it will be to species referred to ‘“Hadrianus” that we will have to look for the antecedent to the T. brontops — T. osborniana series. All so-called Hadrianus have a gular prominence well differentiated; this is true even of the Wasatch species. All are otherwise very primitive and differ from the forms ealled ‘‘Testudo” only in their primitiveness. Their supposed generic separation is no ground for doubting their ancestral relation to the later forms called ‘“Testudo.” Rather it is preferable to regard the generic distinction as invalid or at best of subgeneric value. ‘‘Hadri- anus” corsoni is in most respects primitive enough to have given rise to any of the later Testudo of the Western Hemisphere. In one respect only is there a difficulty and that not an important one. All the “species” called Hadrianus have rather parallel-sided shells combined with a differentiated gular region. 7. wintensis has an undifferentiated gular region combined with a rather rounded shell. This is quite the reverse of the character combination we found to be present in Miocene and later species. In the Oligocene we have seen that both the assignable species have parallel-sided shells. Presumably this means only that the shell contour character, 7f valid at all, was not firmly fastened on either series until the Miocene. This mixture and merging of the characters of the two series in the Eocene is indeed what we should expect. The series which are quite distinct in the Miocene, appear more closely approximated in the Oligocene and merge in the Eocene. 546 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY With the clearer view which the concept of two lineages within Hesperotestudo gives us, let us look now at certain previously problem- atical Miocene species. These forms poorly understood till now fall into place. T. ducatelli Collins and Lynn (Hemingfordian Miocene of Maryland) is a form with the gular region of the plastron not distinct from the contour of the anterior lobe, the entoplastron very slightly wider than long, pectoral scute about 1/5 the abdominal in median length and touching the entoplastron on one side of the unique specimen, inguinal scute large, reaching femoral, xiphiplastral notch well-marked but not deep. The carapace except for the octagonal second neural and a few peripherals and partial pleurals is unknown. I placed this (1950, pp. 27-28), with some misgiving, as possibly a Gopherus, but with the new information now available I place it with much more confidence as a relative of 7. tedwhiter. T. farri Hay (Barstovian Miocene of Montana) is known from most of a crushed shell. The nuchal scute is longer than wide. The antero- lateral corners of the carapace are not preserved. The second and fourth neurals are octagonal. The pleurals are alternately narrowed and widened distally. The gular region is not distinct from the contour of the anterior lobe. The entoplastron is just as wide as long. The pectoral scute, its anterior margin very close to but not touching the entoplastron, is about 1/7 the abdominal in median length. The xiphiplastral notch is distinct but not deep. This species, even more certainly than 7. ducatelli, is a relative of T. tedwhiter. With the addition of these forms the record of the twin lineages within Hesperotestudo becomes much more nearly complete, and the formerly obscured evolutionary picture is in part at least resolved into clearly defined elements. A few difficulties, however, remain: (1) The discovery of the two lineages within Hesperotestudo forbids us to assign to either one of them the form T'estudo gilberti Hay, known only from a skull. I am quite uncertain of the value of the differences described by Hay between the skulls of 7. gilberti, T. impensa, T. osborniana, and T. orthopygia, and I do not think this problem will be amenable to solution until we have still more skulls associated with shells. (2) There is a problem also in the allocation of Testudo crassiscutata Leidy of the Florida (Peace Creek) Pleistocene. In this form, unfortu- nately incompletely known, the anterior plastral lobe is much as in the WILLIAMS: A NEW FOSSIL TORTOISE 547 1’. tedwhiter series, but the xiphiplastral notch is as shallow as in the T. osborniana series. This species may, perhaps, be provisionally placed in the 7. tedwhitei series, but it cannot be pretended that this rather arbitrary placement implies that the situation is understood. What is the relationship of this giant form to 7’. sellardsi? What are Pliocene 1’. lowisekressmani Wark and 7. hay? Sellards and Pleistocene T. luciae Hay? Were there both giant Testudo and giant Gopherus in the Florida Pleistocene? The scattered remains are tantalizing and the problems of nomenclature frustrating. Furthermore, in addition to the puzzling Florida giants there are the giant thick shelled forms of the Ashley River beds of South Carolina, in regard to which Leidy’s name Hupachemys obtusus based on a single peripheral must be con- sidered. This latter may possibly represent a late member of the series of which 7’. tedwhitei is now the best known example. (3) An unfortunate nomenclatorial tangle exists. I have traced two lines of what I previously called Hesperotestudo back to the base of the Oligocene. If, as seems entirely probable, these two lines find their common ancestor in the earlier Eocene in some species of what has been called ‘“‘Hadrianus’’, then according to modern concepts it is impossible to retain these two lineages in Testudo, if Hadrianus is accepted as a full genus. However, as I have indicated and as Gilmore had already suggested in 1915, Hadrianus is very imperfectly defined as against the genus 7’estudo broadly conceived. Its characters may be matched elsewhere in that genus, and it is impossible to retain the name at all unless as a subgenus. If then Hadrianus is regarded as a sub- genus, that will obviate the major difficulty, and that solution is pro- posed here. This will not, however, solve all our problems. If the apparently plausible hypothesis of the separate descent of our two lineages from forms called Hadrianus is true, either Hadrianus, as the oldest avail- able name, must be used as the inclusive subgenus name of all North American Testudo (it then becomes undefinable), or alternatively a new subgenus name is required for the line of which 7. tedwhite7 is a central member. In the latter case, probably but not certainly, Leidy’s name Hupachemys is available. Fortunately, failure to solve this problem at this time involves no serious difficulty since all the forms concerned belong quite certainly to the genus Testudo. DAS BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Figure 1 presents diagrammatically the relationships as I now see them of the determinable members of North American Testudo. Faunal Associations of the New Species In addition to the excellent tortoise material which forms the hypodigm of 7. tedwhitei there are a few additional turtle fragments which can be assigned with fair certainty to the pond turtle genus Pseudemys but which cannot be more precisely placed. The frag- ments in question (Plate 4) are one complete (unillustrated) and two partial nuchal plates, a single complete pleural, a left xiphiplastron and a left epiplastron. The delicacy of the surface sculpturing plus the greater resemblance in some minor details to individual members of the Pseudemys floridana group probably imply that the fragments belong to some extinct member of that group, but in the absence of more material and particularly more comparative material it is quite useless to further discuss these remains at this time. Conspicuously absent from the preserved testudinate fauna of the Miocene of Thomas Farm are trionychids and chelydrids which are very important in the fauna of Florida today and which certainly were abundant also in the Pleistocene faunas (as considerable material at the M.C.Z. proves). This comparative rarity of aquatic types (the absence, indeed, so far as the present record shows, of two very important aquatic turtle families) probably points as does the mammalian record (the pre- dominance of horses, etc., Romer, 1948) to an ecology for the Miocene of Florida very different from that now characteristic of the area. “Tt was then, as now, a low country — but a low plain, relatively dry and grass-covered — a prairie in the western rather than the floridian sense of that term’? (Romer, 1948, p. 10). The amphibian fauna re- cently described by Tihen (1951) would appear to point in the same direction — most abundant a species of Bufo, less frequent a Rana. Acknowledgment: The photographs are to be credited to Hazel and Peter Vaughn. WILLIAMS: A NEW FOSSIL TORTOISE 549 REFERENCES Couns, R. L. and W. G. Lynn 1936. Fossil turtles from Maryland. Proc. Amer. Phil. Soc., vol. 76, pp. 151-173. GILMoRE, C. W. 1915. The fossil turtles of the Uinta formation. Mem. Carnegie Mus., vol. 7, pp. 101-161. Hay, O. P. 1908. Fossil turtles of North America. Carnegie Institution of Wash- ington. Publication No. 75, 568 pp. 1916. Description of some Floridian fossil vertebrates belonging mostly to the Pleistocene. Rept. Florida Geol. Sury., vol. 8, pp. 41-76. Loomis, F. B. 1927. A giant tortoise from Florida. Amer. Jour. Sci., ser. 5, vol. 13, pp. 435-439. Rome_r, A. 8. 1948. The fossil mammals of Thomas Farm, Gilchrist County, Florida. Quart. Jour. Florida Acad. Sci., vol. 10, pp. 1-11. TinEN, J. A. 1951. Anuran remains from the Miocene of Florida with the description of a new species of Bufo. Copeia, 1951, pp. 230-235. Wiuurams, E. E. 1950. Testudo cubensis and the evolution of Western Hemisphere tor- toises. Bull. Amer. Mus. Nat. Hist., vol. 95, pp. 7-36. 500 BULLETID > MUSEUM OF COMPARATIVE ZOOLOGY Plate 1. Testudo tedwhiter new species, ventral view of type plastron. do WILLIAMS: A NEW FOSSIL TORTO Nite JON Plate 2. Testudo tedwhitei new species, dorsal view of type plastron. BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Plate 3. . eins z 5 Z Testudo tedwhitei new species, referred carapace. x about 14. Plate 4 Pseudemys sp., various fragments from the Thomas Farm Miocene. x i, Snjnosniow | UDIUDIDSDM 1U0SIJOD (SNUDIIPDH) | ubduebpug snplwny 1 sdojuosq |) DDDAIuaID | $dadijsnbuo Th DSUddWI! | DUDIUJOGSO (OpNjsajOs9dSaH) | DIBAdoujyso 1 UDIBIDYUY layIYMpa} TL ]]a}DONP || upips0jybulway THIDS ISpad||aS “1 DJDJNOSISSDIO | é DSnjqo (Skwayondna) | 2 9us009 aUusd05NO QUdIOIW QUdd0!}q 9U9D04SI9}q Diagram of presumed phvletic relationships of North American Jone, il Testudo species. ww i iy wi eae cainethe ¥