pend adtahi eta As oh week Olle dts y, of fo Btist vot abethutean a Sone Met a2 Ae eas nt the Nee 60 Bast toe ; . : ate 3 ty “ v - ‘ he hater ‘ ; eis - ; wae Hb sthe eee peered tes 5 r : wma Bann An @ - m : . a fore at a ash yas no me eiate. oreeeny at TET ne - she fakin tied cape : HARVARD UNIVERSITY e Library of the Museum of Comparative Zoology { i Min RORY OF hyd: eaten Donnie nA «nite dilecslei gale ES Aiea an tt a i anparsia 7 se ai eer ai i P Jovan, py Ai ila mek) ie i | 9): bara mh Ch ee hie ) LOCI. Aba ) PL UOR TE Gnd nalobieeMiaatk Ciiaias ‘ i Chg ace ay Op ia \s He at Sa a co auc ievt-a) 08 i MARDI HAL, POI exer Le tie ey Sips 4) i CET FAT CREM ome RSP OTS £) Le see bAr RDN ' eect if 4 ay, 2. eae hePp it. ke ae ee mai tan. +: rete uF a f nat LASRAP FAN Das) ih Tie a od si PLARL BA 4 pak VR OO Saag | Ba hy ve Dea ee cueiaeess 169-08: Me eerie tT eR te Tce eee Le a a a ieee Bo oe ae Anais i ane Sa Pere Te ve eS hier: a ioe ‘pak Tema ay ae array i ee oe maa eins wi: pat B. Keine | ; my een nic Pis: oA ban SUR ME ‘unaeie a inhebtey = (oie aaa HS hh ay bear mas a ee wo Cre PRG CAT & i ne i" < A: ‘ can i wee a i cae 4 Occasional Papers of the Museum of Natural History _ University of Kansas 71-95 (1978-81) va. KARYOLOGY AND EVOLUTION OF THE PLAINS POCKET GOPHER, GEOMYS BURSARIUS. E. Blake Hart 72. A RE-ASSESSMENT OF THE TELMATOBIINE LEPTODACTYLID FROGS OF PATAGONIA. John D. Lynch 73. MAMMALS OF THE WOLF RANCH LOCAL FAUNA, PLIOCENE OF THE SAN PEDRO VALLEY, ARIZONA. Jessica A. Harrison 74. REPRODUCTION IN MARGINAL POPULATIONS OF THE HISPID COTTON RAT (SIGMODON HISPIDUS) IN NORTHEASTERN KANSAS. L.R. McClenaghan Jr. & M.S. Gaines eS. AN EARLY MIOCENE (ARIKAREEAN) FAUNA FROM NORTH- CENTRAL FLORIDA (THE SB-1A LOCAL FAUNA). David Frailey 76. A NEW SPECIES OF LIOLAEMUS (SAURIA: IGUANIDAE) FROM THE ANDEAN MOUNTAINS OF THE SOUTHERN MENDOZA VOLCANIC REGION OF ARGENTINA. Jose M. Cei a7 BRACHYPOTHERIUM FROM THE TERTIARY OF NORTH AMERICA. Daniel Yatkola and Lloyd G. Tanner 78. SYSTEMATIC STUDIES OF DARTERS OF THE SUBGENUS CATONOTUS (PERCIDAE), WITH THE DESCRIPTION OF A NEW SPECIES FROM CANEY FORK, TENNESSEE. Marvin E. Braasch and Lawrence M. Page 79. THE FOSSIL FAUNA FROM LOST AND FOUND QUARRIES (HEMPHILLIAN: LATEST MIOCENE), WALLACE COUNTY, KANSAS. Debra K. Bennett 80. GLACIATION AND SPECIES RICHNESS OF BIRDS ON AUSTRAL SOUTH AMERICAN ISLANDS. Philip S. Humphrey and Jaime E. Pefaur 3674-087 81. NOTES ON AN IMPORTANT NINETEENTH CENTRY COLLECTION Ose tO ays NORTH AMERICAN BIRDS MADE BY N.S. S e ° . Jenkinson and R.M. Mengel ’ ? c a a . f ft r " ” : yroseil Is \ 49 peewiterfg to Broqed fenokeaae “i we Ms ms ee ae gett? iL SToOVuGs fey z : . my a £ a Sb ) Pe» | - TJ cry f i Ta] i é On of fe ATAQWAT ) 2008F PAPA a WO. ya >} ZIAMBAR SAAGSIAA .Va f. OMA - HT 30 I TTS , wyseeey i arialt r GQ GAUL a2 7 CHP Sa -MBATSASNT AOR 41 (eG Ttcin 20uc wic) TA TIOD : > a ' nin ef s) r a x | Sora oS ‘ SStt i fie “Get, 7 o' nets BITS 93 ofan 7 r CAA, « “HTHOU, MONT AMUAG CRAMARAN TS a gNGI0TM YIAA Doe t Al-de 7S Ra eu a ARInaS » CARUAT TAD : : Stes) Caves ce * ; ; pide & f / (TACTAAUOL sATAUAE) -2URAAIOLI 0 CAT sate wet 7 A asoaitay. ARART UO? ant 10: SMTATHUOM MASCMA OAT Mog be ; | | ’ ram = 60 1M seot AGETMSOSA 70 AOL ODADAS |G AQUITRSMA WT AO WO YHAETAST. diet MOR Mates oa +i hee, ‘yeninsT oo told br slo sey foding ae aes a 4" SAATHAG FQ dentine. TEAM qe 2 a | ‘A FO MOL LTT LAQRSCL ‘SHT H¥Iv ARE 4. ee AdeeaeMaT ANON YAdAd HOF | a seh -gani. st annerwsd: pre. dowma rt | a Siva - ‘ eaEaIA cdo GHA. seat. HOH en Cy ready a PEALOS 994 tA Pret gn 4: T ‘as ae 81. 82. 83. 84, 85. 86. 87. 88. 89. 90. Ake. - 92. 93. NOTES ON AN IMPORTANT NINETEENTH CENTURY COLLECTION OF CENTRAL AND NORTH AMERICAN BIRDS MADE BY N.S» GOSS. M.A. Jenkinson and R.M. Mengel REPLICATION OF HABITAT PROFILES FOR BIRDS. Richard F. Johnston THE RELATIONSHIPS OF THE AMPHIBERINGIAN MARMOTS (MAMMALIA: SCIURIDAE). R.S. Hoffmann et al. A NEW MARSUPIAL FROG (HYLIDAE: GASTROTHECA) FROM THE ANDES OF ECUADOR. W.E. Duellman & R.A. Pyles ADDITIONS TO THE KNOWLEDGE OF HIPPOCAMELUS, CTENOMYS AND MYOCASTOR FROM THE MIDDLE PLEISTOCENE OF THE TARIJA BASIN, BOLIVIA. David Frailey et al. REPRODUCTIVE BIOLOGY OF LIZRDS ON THE AMERICAN SAMOAN ISLANDS. Terry D. Schwaner A REVIEW OF THE PHYLLOMEDUSA BUCKLEYI GROUP (ANURA: HYLIDAE). David C. Cannatella THREE NEW SPECIES OF CENTROLENID FROGS FROM THE PACIFIC VERSANT OF ECUADOR AND COLOMBIA. William E. Duellman REDESCRIPTION OF ETHEOSTOMA AUSTRALE AND A KEY FOR THE IDENTIFICATION OF MEXICAN ETHEOSTOMA (PERCIDAE). Lawrence M. Page THE GENERA AND SUBGENERA OF DARTERS (PERCIDAE, ETHEOSTOMATINI). Lawrence M. Page A NEW HYLID FROG OF THE GENUS PLECTROHYLA FROM A CLOUD FOREST IN HONDURAS. James R. McCranie and Larry David Wilson A NEW GYMNARTHRID MICROSAUR FROM THE LOWER PERMIAN OF KANSAS WITH A REVIEW OF THE TUDITANOMORPH MICROSAURS (AMPHIBIA). H.-P. Schultze LIFE HISTORY OF THE SLENDER MADTOM, NOTURUS EXILIS, IN SOUTHERN ILLINOIS (PISCES: ICTALURIDAE). Richard L. Mayden and Brooks M. Burr SrrSaJJod YAUIMAD. i TMATAMTA TAATAMM 44 MO.2atOM A024 Jos , P . ‘oor want legit ae fa QUAM & | WADTRUMA HTEACE GHA . IAATHAD FO fousion «Meh huts oenttael .AuM 2200 «Belk Sista HOF au.iI4OR4 TATIGAH FO “OLTAOL sopemiol «7% bredoif STOMAAM MAIO“IAaoLHIMA SHI 40 @FIHeMOITAIZA Gee ‘4 yw 4 7 werno ef us ahem ne to - oGeh - e% GAG fs Jide cAl. AMMAN > HORT CAVSHTOATEAD os oe Oe SP) I I Len ION} De hon I =) HAoocowmMmnaAnnnAaea a) Ao SBX9T, “UOSRIY SBXOT, “[[AISVD sexo, ‘Apousy Sexo], “O1oND sexo y, “orvnyeuy Sexo], “A}ID Sexo], SexaT, “BS 252][0D SPXOT, “BUUO'T ¥}[V ‘eT “elOWUs|) ‘eT ‘asnoy Joy ‘ePTO “UeULION ‘eIYO ‘yesumnoa J, "RTIO ‘sso[A\og "By POed "YLV “OT[LASTYSIAA , uey ‘Aouadye AA "G°N “YSION ‘qeN ‘UoIpeyD "qeN “te0ueds ‘uey ‘AVID espoqd ‘uey ‘Ad[sury ‘uey ‘Teloqry ‘eDO Wea ‘XO 'N ‘uepas *XOW 'N ‘UOVRID “xO “N ‘peysrury 0L-89=NA “GL-0L=UG dnois sisuaxaz (2 GL=NA ‘OL=UG dnois 1wajonqyyo (9g CL=NA ‘FL=US dnois sdasiaaiq (GC OL=NA ‘GL=UZ 86-98=Na ‘GL="UZ dnois suagsajny (fF OL=NA ‘ZL=US dnois smiujsnpur (¢ SL=NA “GL=UG 12 OCCASIONAL PAPERS, MUSEUM OF NATURAL HISTORY The initial derivation in the sequence from the breviceps karyo- type is postulated to be a hypothetical intermediate stock of 2n=72, FN=72 with one pair of biarmed autosomes produced by centric fusion and retention of the biarmed X chromosome. A single speci- men with this karyotype was secured from Norman, Oklahoma, between adjacent populations of dutcheri and major; the identical karyotype also has been reported in G. personatus streckeri (Davis, et al., 1971). The Oklahoma specimen documents the existence of a potential chromosomal link between the typical breviceps karyo- type (from 4 mi. E., Norman, Cleveland, Co., Oklahoma) and the more northerly major (Norman). If interbreeding occurs, intra- populational polymorphisms may persist within the population. Alternatively, the 2n—72 form may be prone to rise in geographically peripheral populations of breviceps. In northern bursarius and majusculus samples, several different chromosomal polymorphisms were found (see above); however, in no other part of breviceps range were any encountered. From the intermediate stock described, typical karyotypes of the majusculus and attwateri types can be derived in single chromo- somal events from the ancestral form. The formation of attwateri (2n—=70, FN=72, metacentric X) can be assumed to have resulted from an autosomal centric fusion in the intermediate form; this event probably occurred as early as formation of the major karyotype, and it is possible that attwateri already has attained full specific status because its 2n—=70 contrasts with the 2n—74 of contiguous populations of breviceps to the east (Fig. 1). This suggests that gene exchange is limited, perhaps nonexistent between gophers possessing these two chromosomal forms. Geomys personatus megapotomus to the south of attwateri (Fig. 1), possesses a grossly similar karyotype. There are several possible sequences of chromosomal change that could have resulted in G. b. lutescens, the distinctively variable subspecies occurring in the variable environment of western Ne- braska, western Kansas, and eastern Colorado (Fig. 1). This area of mixed and short grass prairie may be the most marginal, xeric habitat occupied by any population of G. bursarius. The most likely derivation karyotypically is one identical to extant G. per- sonatus, from attwateri, or from the postulated “intermediate” type. The latter is supported by the present northern distribution combined with similarity of skulls (Davis, 1940) between personatus and lutescens; moreover, G. arenarius has a chromosomal system similar to lutescens. Berry (1969) reported the 2n of G. arenarius as 70, FN as 100 or 102; whereas Davis et al. (1971) reported it as 2n—70, FN==102. Both chromosomal forms were found to possess a biarmed X and highly variable FN’s, and are probably more closely related, chromosomally, then either is to gophers of any other sub- THE PLAINS POCKET GOPHER 13 group. An intensive study has been underway for some time to determine the chromosomal characteristics of individuals along the zone of potential contact between G. b. lutescens and other sub- species (Hart, upublished data). PLEISTOCENE HISTORY Accurate designation of the time when these postulated chromo- somal arrangements could have occurred is difficult at best. Hib- bard e¢ al., (1965) suggested that probably all Recent mammalian species originated during the Quaternary; Blair (1954) argued that most speciation in grassland taxa likely occurred during repeated north-south migrations of the Wisconsin glacial period. Russell (1968) summarized the fossil history of Geomys bursarius; it first appeared in the [linoian of Kansas and Oklahoma. Specimens of Geomys, mostly referrable to living species, are common from the Sangamon Interglacial of Kansas, Nebraska, and Texas, and Geomys bursarius is recorded from the Wisconsin of Kansas, Texas, I]linois, Wisconsin, and Nebraska. Lundelius (1967) reported G. bursarius fossils from numerous localities in Texas, representing late Pleisto- cene assemblages. It appears from the present distribution of karyotypic forms, some involving disjunctions of hundreds of miles, that gophers hav- ing a variety of chromosome types have successively occupied large areas of the midwest, only to be displaced. It also appears that climatic changes accompanying Pleistocene and post-Pleistocene events decimated large portions of established populations; later, optimal conditions provided ample opportunity for invasion of op- portunistic, chromosomally variable populations from southern and other refugia, with invaders encroaching on and competing with earlier inhabitants. Significantly, the Gulf Plains were open throughout the Pleistocene and undoubtedly acted as an important dispersal corridor (Auffenberg and Milsted, 1965). Pocket gophers are fossorial and are consequently limited in their dispersal ability by soil and physiography. Although the Great Plains does contain a variety of soil types and physiographic features, there is little doubt that the relative paucity of significant barriers has resulted in fewer readily detectable gopher pheno- types than occurred in areas such as the Rockies, where physiogra- phy has been more efficient in isolation of small populations of the montane pocket gopher (Thomomys), with subsequent dramatic evolutionary events occurring independently of near neighbors (Durrant, 1952). In Geomys, the spatial isolation required for intensive inbreed- ing by small remnant or marginal populations, and resulting in irreversible genetic independence (Lewis, 1966), may be inferred from the recent past history of the Great Plains. Physiography has 14 OCCASIONAL PAPERS, MUSEUM OF NATURAL HISTORY had some influence as an isolating mechanism in the Great Plains (large rivers, natural geological formations of indurate soils and rocks) but, overall, such influence appears to have been minimal. Instead of insurmountable topographic barriers, I suggest that climatic events accompanying glacial advance and recession were mainly responsible for decimation, often termed catastrophic, of extensive populations occupying originally. optimal habitats. CHROMOSOMAL EVOLUTION Although the evidence at hand does not warrant. unequivocal conclusions regarding the dynamics of population dispersal of plains pocket gophers and formation of chromosomal morphs, the following ideas, based upon the postulated sequences of chromo- somal rearrangements operating within glacial and_ postglacial climatic changes, are presented. The hypothetical ancestral breviceps-like form is thought to have originally given rise, during the middle to late Pleistocene, to major-like gophers which radiated northward and occupied much, if not all, of the midwest. It is assumed that glacial advance or meltwater rivers bifurcated populations of major, thus isolating illinoensis in the east. During the next epoch of optimal habitat availability, major persisted in western parts of its range, but was replaced by another form derived from major, the bursarius karyo- type, in the east. Ancestors of the bursarius group became isolated from northern peripheral populations of the major group, ultimately forming and spreading the bursarius karyotype into northern (partially unoccupied) localities, even approaching glacial margins. Peculiarities of their biology may have enabled pocket gophers to survive glacial epochs in fairly close proximity to the actual glacial front (Russell, 1968). Perhaps natural selection impinging upon periglacial major populations at the northern periphery resulted in the bursarius karyotype. Peripheral populations or individuals of bursarius are assumed to have crossed the Mississippi River some- what later and today retain similar chromosomal phenotypes (G. D. wisconsinensis). Perhaps emigrants from the latter population crossed the Wisconsin River during the intrusion of the “prairie peninsula” and spread east of the Mississippi. Thus independent parallel chromosomal events may account for the similarity between illinoensis and major. G. b. lutescens populations, derived from southern breviceps or related forms, extended their range to the north, via Colorado and eastern Wyoming. The present distribution of industrius also ap- pears of rather recent rearrangement via an extension of northern boundaries from more southerly extant breviceps-derived an- cestors; Hoffmann and Jones (1970) described similar northward migrations of southern mammals during or following the late THE PLAINS POCKET GOPHER 15 Pleistocene. Also, it is conceivable that while lutescens expanded to the north, it also gave rise to G. arenarius, which possesses a uniquely similar chromosomal makeup, and which is now disjunct. Perhaps the non-Robertsonian polymorphic conditions found only in G. arenarius and G. b. lutescens indicate parallel chromosomal responses to the relative severe environments in which both taxa are presently found. Neotoma micropus, likewise inhabiting xeric areas of northern Mexico, Texas, and New Mexico has been re- ported to have a unique system of polymorphisms (Baker and Mascarello, 1970). Approximately at the time of the initial independence of bur- sarius from major in the north, attwateri may have formed from southern peripheral populations of breviceps. Exact systematic evolutionary relationships among G. b. attwateri, G. b. ammophilus, and G. personatus are presently obscure. However, affinities among them may well be closer than that of the attwateri group to brevi- ceps, with which the latter is currently allied taxonomically. So few samples of G. b. texensis and G. b. llanensis were ana- lyzed that the most tentative of generalizations concerning origin can be proposed here. The similarity of this 2n—70-72 karyotype to those reported by Berry (1969) and Baker and Genoways (1975) in western Texas indicate dendritic expansion along river systems. The ancestral stock was likely similar to the totally acrocentric G. b. major (2n=72). Elsewhere in the genus, Geomys pinetis (2n—42, FN—78—Hart, 1971; see also Williams and Genoways, 1975) and Geomys tropicalis (2n—38, FN=74—Davis et al., 1971) possess significantly lower diploid numbers (see Fig. 1). G. pinetis is reported to have been isolated since Sangamon times (Russell, 1968) and obviously has undergone periodic geographic restrictions, to which fluctuating sea-levels along the eastern coasts during the Pleistocene probably contributed substantially (Auffenburg and Milstead, 1965). CONCLUSIONS When considering the karyotypic variation represented by the chromosomal groups within Geomys bursarius, as well as the intra- populational variation occasionally encountered, it appears that a taxonomic system based on few characters is unlikely to be valid. In the present study, animals were found that varied markedly in terms of external morphology, even though possessing similar chromosomes. On the other hand, morphologically similar animals often exhibited great karyotypic differences, even though some were collected only a few miles apart. Reduced diploid numbers in the relatively restricted populations of G. b. attwateri and G. b. ammophilus (2n—=70) suggests that accumulation of two pairs of biarmed autosomes may represent an 16 OCCASIONAL PAPERS, MUSEUM OF NATURAL HISTORY initial Robertsonian fusion process which, if continued over a period of time, will lower 2n, while maintaining FN. Closely re- lated, parapatric Geomys personatus exhibits diploid numbers ranging from 72 to 68 (Davis et al., 1971); significantly, the forms possessing diploid numbers lower than 72 have increased numbers of biarmed autosomes. On the other hand, in populations of the breviceps group, one may account for the persistence of the higher diploid number (2n=74) on the basis of the extensive distribution of the group and resultant gene flow throughout most of this area as well as long-term stability of the southern plains environment. Davis (1940) contended that at least G. b. breviceps in Louisiana and G. b. terricolus in Texas are presently isolated due to unfavor- able soils. The eventual fate of these recently isolated populations may well follow the trend toward decreasing diploid number as in the attwateri group and ultimately in G. tropicalis and G. pinetis. Systematic implications of the present study suggest that there are several chromosomal types within Geomys bursarius and that reproductive compatibility between them may be seriously ques- tioned, most notably “Southern” vs. “Northern” lutescens, major, breviceps, and attwateri. Extensive work is in order, especially between the attwateri-ammophilus group, and breviceps to the east and Geomys personatus to the south. ACKNOWLEDGEMENTS I express my sincere gratitude to A. Dean Stock and J. Keever Greer for all assistance rendered throughout the duration of this study. Troy Best, Larry Brown, Ervin “Hoplanc Richard Savage, and Dwight Spencer aided me in obtaining specimens. I thank James L. Patton for critically evaluating this manuscript, J. Knox Jones for editorial advice, and especially Robert $. Hoffmann for constructive advice, editorial suggestions, and encouragement. I express my deepest gratitude to my wife, LuAnn, who has sup- ported and assisted in innumerable ways. Technicians were Renata Waters and Priscilla Zink. Thomas Swearingen assisted with the figures. Phyllis Stacy assisted with manuscript preparation. This research was supported in part by the University of Okla- homa Graduate College, University of Oklahoma Department of Zoology, Stovall Museum of Science and History, University of Oklahoma Biological Station, and a Grant-In-Aid from the Society of the Sigma Xi. LITERATURE CITED \urrenbenc, W., and W. W. Minsreap. 1965. Reptiles in the Quaternary of North America. Pp. 557-568, in the Quaternary of the United States (Hi. iE. Wright and D, G. Frey, Eds.), Princeton Univ. Press. Baken, RK. J., and H. H. Genoways. 1975. A new subspecies of Geomys THE PLAINS POCKET GOPHER 17 bursarius (Mammalia: Geomyidae) from Texas and New Mexico. Occasional Papers, The Museum, Texas Tech Univ., No. 29:1-18. Baker, R. J., and J. T. MascareLito. 1969. Karyotypic analyses of the genus Neotoma (Cricetidae, Rodentia). Cytogenetics, 8:187-198. Baker, R. J., S. L. WiixiaMs and J. C. Patron. 1973. Chromosomal variation in the pocket gopher, Geomys bursarius major. J. Mammal., 54:765-769. Berry, D. L. 1969. Karyotypes and chromosomal evolution in west Texas pocket gophers (Rodentia, Geomyidae). Unpub. M.S. Thesis, Texas Tech Univ., Lubbock. Berry, D. L., and Rosertr J. BAxer. 1972. Chromosomes of pocket gophers of the genus Pappogeomys, subgenus Cratogeomys. J. Mammal., 53: 303-309. Best, T. L. 1973. Ecological separation of three genera of pocket gophers (Geomyidae). Ecology, 54:1311-1319. Biair, W. F. 1954. Mammals of the mesquite plains biotic district in Texas and Oklahoma, and speciation in the central grasslands. Texas J. Sci., 6:235-264. CocxruM, E. L. 1952. Mammals of Kansas. Univ. Kansas Publ., Mus. Nat. Elisteie l=308: Cross, J. C. 1931. A comparative survey of the chromosomes of rodents. J. Morphol., 52:373-376. Davis, B. L., S. L. Wimu1ams, and G. Lopez. 1971. Chromosomal studies of Geomys. J. Mammal., 52:617-620. Davis, W. B. 1940. Distribution and variation of pocket gophers (genus Geomys) in the southwestern United States. Bull. Texas Agric. Exp. Sta., 590:4-38. DurranT, S. D. 1952. Mammals of Utah. Univ. Kansas Publ., Mus. Nat. Hist., 6:1-549. Hat, E. R., and K. R. Ketson. 1959. The mammals of North America. The Ronald Press Co., New York, 1:xxx-+-1-546++79. Hart, E. B. 1971. Karyology and evolution of the plains pocket gopher, Geomys bursarius. Ph.D. thesis, Univ. Oklahoma, Norman, 110 pp. Hart, E. B. 1973. A simple and effective gopher live trap. Amer. Midland Nat., 89:200-202. Hart, E. B. 1975. Chromosomal evolution in Geomys bursarius plains pocket gopher. Mamm. Chromo. Newsletter 16(1):12-16. Hisparp, C. W., D. E. Ray, C. E. Savace, D. W. TAytor, and J. E. Gumpay. 1965. Quaternary mammals of North America. Pp. 509-525, in The Quaternary of the United States. (H. E. Wright, Jr. and D. G. Frey, Eds.) Princeton Univ. Press, Princeton. HorrMann, R. S., and J. K. Jones, Jr. 1970. Influences of late-glacial and post-glacial events on the distribution of Recent mammals on the northern Great Plains. Pp. 353-394, in Pleistocene and Recent environ- ments of the central Great Plains (W. Dort and J. K. Jones, Jr., eds.), University Press of Kansas, Lawrence. Jones, J. K., Jr. 1964. Mammals of Nebraska. Univ. Kansas Publ., Mus. Nat. Hist., 16:1-356. Kennerty, T. E. 1958. Comparison between the ranges of two allopatric species of pocket gophers. Texas J. Sci., 10:133-147. Kim, Y. J. 1972. Studies of biochemical genetics and karyotypes in pocket gophers. Unpub. Ph.D. Disser., U. Texas, Austin. (Original not seen; in Selander, et al., 1974). Lewis, H. 1966. Speciation in flowering plants. Science, 152:167-172. Lunvbeuius, E. L., Jr. 1967. Late-Pleistocene and Holocene faunal history of central Texas. Pp. 287-319, in Pleistocene extinctions: the search for 18 OCCASIONAL PAPERS, MUSEUM OF NATURAL HISTORY a cause (P. S. Martin and H. E. Wright, Jr., eds.), Yale Univ. Press, New Haven, Connecticut. Martrnuey, R. 1960. Quel est le nombre maximum de chromosome sonnu chez un mammifere? La formula chromosomique de Geomys_ bursarius breviceps Swenk. Mammalia, 24:165-170. McLaucuun, C. A. 1958. A new race of the pocket gopher Geomys bursarius from Missouri. Contr. in Sci., Los Angeles Co. Mus., No. 19. 4 pp. Merriam, C. H. 1895. Monographie revision of the pocket gophers family Geomyidae (exclusive of the species of Thomomys). N. Amer. Fauna, 8:1-258. Patron, J. L. 1967. Chromosome studies of certain pocket mice, genus Perognathus (Rodentia: Heteromyidae). J. Mammal., 48:27-37. Patron, J. L. 1970. Karyotypic variation following an elevational gradient in the pocket gopher, Thomomys bottae grahamensis Goldman. Chromo- soma, 27:41-50. Patron, J. L. and R. E. Dincman. 1970. Chromosome studies of pocket goph- ers, genus Thomomys. II. Variation in T. bottae in the American south- west. Cytogenetics, 9:139-151. RusseLx, R. J. 1968. Evolution and classification of the pocket gophers of the subfamily Geomyidae. Univ. Kansas Publ., Mus. Nat. Hist., 16:473-579. SELANDER, R. K., D. K. KaurMan, R. J. BAKER, and S. L. WituiaMs. 1974. Genic and chromosomal differentiation in pocket gophers of the Geomys bursarius group. Evolution, 28:557-564. TuHAeE eR, C. S., Jn. 1968. Karyotypes of sixteen populations of the Thomomys talpoides complex of pocket gophers ( Rodentia-Geomyidae). Chromo- soma, 25:172-183. WeEntTWworTH, G., and D. A. Surron. 1969. Chromosome variation in the pocket gopher, Thomomys townsendii. Southwestern Nat., 14:157-161. Wixuiams, S. L., and H. H. GeNoways. 1975. Karyotype of Geomys pinetis (Mammalia: Geomyidae), with a discussion of the chromosomal rela- tionships within the genus. Experientia, 31:1141-1142. APPENDIX: SPECIMENS EXAMINED Voucher specimens are deposited in the Collection of Recent Mammals of the Stovall Museum of Science and History at the University of Oklahoma. Numbers with no prefix are museum catalogue numbers; specimens with a prefix of EBH (E. B. Hart), TLB (T. L. Best) and ADS (A. D. Stock) are also available in the museum. Numbers with an asterisk (*) refer to speci- mens with fewer than eight (four to seven) chromosomal counts per animal due to suboptimal preparations: Geomys bursarius ammophilus 2N=70. TEXAS: DeWitt Co.: 1 mi. SE Cuero, 14, EBH571°; 1 9, TLB3799°. Geomys bursarius attwateri 2N=70. TEXAS: Karnes Co.: 8 mi. NE Kenedy, 24 4, TLB3806, TLB 3807; 1 9, TLB 3809. Geomys bursarius brazensis 2N=74. TEXAS: Brazos Co.: 0.5 mi. W College Station, 19, EBH567; 4 mi. W College Station, 1¢, 19, EBH569, EBH570. Geomys bursarius breviceps 2N=74. LOUISIANA: Morehouse Parish: 2 mi. W Mer Rouge, 1 4, 7698; 2 mi. NW Mer Rouge, 4 9 9, 7707, 7708, 7710, 7711. Geomys bursarius bursarius 2N=72. IOWA: Clayton Co.: 4 mi. E Monona, 1 %, 7535°; Monona, 1 4, 10082, 6 99%, 10085°, 10086, 10087, 10088, 10089, 10091. Dubuque Co.: 4 mi. NE Bankston, 1 9, 7543; 4 mi. NE Garbin, 1 4, 7536, 1 9, 7535. 2N=70. SOUTH DAKOTA: Bon Homme Co.: i mi. E Scotland, 19, EBH 307. THE PLAINS POCKET GOPHER 19 Geomys bursarius dutcheri 2N=74. ARKANSAS: Pulaski Co.: 3 mi. E Wrightsville, 4. 3 4, 7696; TLB3748; EBH562, EBH563, 2 2°, 7694, TLB3749. LOUISIANA: Grant Parish: 1.6 mi. N Pollock, 12, 3763. OKLAHOMA: Cleveland Co.: near Norman, 5 ¢ 4, EBH657, EBH662, EBH663, EBH664, EBH665; 4 2 2°, EBH658, EBH661, EBH666, EBH669. Pottawatomie Co.: 1 mi. W Tecumseh, 1 2, EBH635. Seminole Co.: near Bowlegs, 1 2, EBH206. Geomys bursarius illinoensis 2N=72. ILLINOIS: DeWitt Co.: 1 mi. S Clinton, 1 ¢, 7552. Madison Co.: 4.5 mi. S Collinsville (O'Fallon Rd), 1 ¢, 7724, 4 2 2, 7723, 10065, 10066, 10067. St. Clair Co.: East St. Louis, 1 ¢, 7718. Geomys bursarius industrius 2N=72. KANSAS: Edwards Co.: Kinsley, 2 6 4, 7475*, 10116; 5 mi. W Kinsley, 12, 7474. Ford Co.: 4 mi. S Dodge City, 1 $, 7477, 1 2, T7479. Geomys bursarius jugossicularis 2N=72. KANSAS: Seward Co.: 10.5 mi. S Liberal, 32 2, 7480, 7485, 7486. NEW MEXICO: Union Co.: 2.4 mi. W, 3.2 mi. N Amistad, 1 ¢, EBH 544; 13.5 mi. S Clayton, 1 ¢, EBH541; 10.1 mi. S Clayton, 1 2, EBH545; 3 mi. W, 1.6 mi. N Sedan, 12, EBH542. OKLA- HOMA: Cimmaron Co.: Feli, 1 2, EBH540. Geomys bursarius llanensis 2N=71. TEXAS: Llano Co.: 6 mi. E Castell, 1 g, EBH575. Geomys bursarius ludemani 2N=74. TEXAS: Chamber Co.: 7 mi. S Anahuac, 1 6, TLB3786, 4 2 2°, TLB3771, TLB3772, TLB3775, TLB3785. Geomys bursarius lutescens 2N=72. KANSAS: Trego Co.: 20 mi. S Wakeeney, 1 2, 7473. NEBRASKA: Antelope Co.: 4 mi. S Neligh, 1 ¢, 10100*; Boyd Co.: 4.5 mi SE Spencer, 1 ¢, 7531. Dawes Co.: 11 mi. S Chad- ron, 1 2, 7495; 20 mi S Chadron, 1 ¢, 7507; 2 mi. SW Chadron, 1 9°, 7504. Geomys bursarius major 2N=72. OKLAHOMA: Cleveland Co.: 3 mi. N Lexington, 2 2 2, 9001, EBH513; Norman, 10 ¢ ¢, 9005, 9006, 9008, 9331, EBH515, EBH536, EBH552, EBH649, EBH652, EBH670; 19 2 2, 9002, 9003, 9004, 9007, 9009, 9093, 10060, 10062, EBH202, EBH503, EBH516, EBH554*, EBH644, EBH645, EBH646, EBH647, EBH648, EBH659, EBH660. Grant Co.: 14 mi. W, 1 mi N Hawley, 1 4, 9115, 2 2 2, 9108, 9116. Kay Co:: Ponca City, 1 ¢, EBH512. Stephens Co.: 0.25 mi. E Claude, 1 ¢, 9096, 1 2, 9095; 0.25 mi E 0.75 mi. S Claude, 2 2 2, 9097, 9320. Tillman Co.: Tipton, 1 g, ADS130. NEW MEXICO: Roosevelt Co.: 9 mi. S, 1 mi. W Tolar, 1 92, EBH543. Geomys bursarius majusculus 2N=70. KANSAS: Chase Co.: 13 mi. W, 1.3 mi. N Emporia, 1 2, EBH547; Douglas Co.: 3 mi. N Lawrence, 3 2 9, EBH683*, EBH 682*, EBH684*. 2N=72. NEBRASKA: Antelope Co.: 1 mi. W Tilden, 1 ¢, EBH628. Geomys bursarius missouriensis 2N=72. MISSOURI: Franklin Co.: 2.5 mi. E Sullivan, 1 2 , 7727. Geomys bursarius pratincola 2N=74. LOUISIANA: Rapides Parish: 2.3 mi. N, 8.8 mi. W Glenmora, 2 6 6, TLB3762, TLB3767, 6 2 2, EBH566, TLB3763, TLB3764, TLB3765, TLB3768, TLB3769. Geomys bursarius sagittalis 2N=74. TEXAS: Galveston Co.: Alta Loma, 1 ¢, TLB3794°, 29 9, TLB3795, TLB3796. Geomys bursarius terricolus 2N—74. TEXAS: Galveston Co.: 1 mi. N Texas City, 2 ¢ ¢, TLB3791, TLB3793, 1 9, TLB3792. Geomys bursarius texensis 2N=70-72. TEXAS: Mason Co.: Mason, 1 ¢, TLB38l2; 2 2 2, TLB3811°, TLB3813°". Geomys bursarius wisconsinensis 2N=72. WISCONSIN: Richland Co.: 0.5 mi. E Gotham, 1 ¢, 7541, 5 2 9, 7540, 7544, 7547, 7548, 7549; Gotham, 2 $ 6, 10073, EBH605, 2 2° 2, 10072, EBH601. UNIVERSITY OF KANSAS PUBLICATIONS MUSEUM OF NATURAL HISTORY The University of Kansas Publications, Museum of Natural History, beginning with volume 1 in 1946, was discontinued with volume 20 in 1971. Shorter research papers formerly pub- lished in the above series are now published as Occasional Papers, Museum of Natural History. The Miscellaneous Publica- tions, Museum of Natural History, began with number 1 in 1946. Longer research papers are published in that series. Monographs of the Museum of Natural History were initiated in 1970. All manuscripts are subject to critical review by intra- and extra- mural specialists; final acceptance is at the discretion of the publications committee. f Institutional libraries interested in exchanging publications may obtain the Occasional Papers and Miscellaneous Publica- tions by addressing the Exchange Librarian, University of Kan- sas Library, Lawrence, Kansas, 66045. Individuals may pur- chase separate numbers of all series. Prices may be obtained upon request addressed to Publications Secretary, Museum of Natural History, University of Kansas, Lawrence, Kansas 66045. Editor: Ep Wi Lry PRINTED BY UNIVERSITY OF KANSAS PRINTING SERVICE LAWRENCE, KANSAS ON! mus, COMP, ZOOL. 2/22 LIBRARY of the HARVARD MUSEUM OF NATURAL HISTORY The University of Kansas Lawrence, Kansas NUMBER 72, PAGES 1-57 SEPTEMBER 20, 1978 A RE-ASSESSMENT OF THE TELMATOBIINE LEPTODACTYLID FROGS OF PATAGONIA By Joun D. Lyncx’* As presently constituted, the leptodactylid frog genus Telma- tobius contains some 30 species distributed on the Andean cordil- leras from Ecuador (0° 15’ N) to southern Argentina (ca. 43° S). The species from Bolivia, Ecuador, Peri and from northern Argen- tina and Chile are Andean frogs whereas those from southern Argentina include Andean and extra—andean species. Although known for only the past decade, the generic relationships of these southern species are the subject of considerable uncertainty. My characterization of the genus (Lynch, 1971) was incomplete in that only three species were available for study; even so, considerable heterogeneity was evident. The unusual species of that limited sample was T. patagonicus, the first of the Patagonian species described. Six Patagonian species are now known from Neuquén and Rio Negro Provinces, Argentina, and a seventh was recently described from southern Chile (ca 49° S). The relationships of these species with other telmatobiine leptodactylids are obscure. The seven spe- cies are variously assigned to three genera. Initially, my study was to be one of the seven “extra—andean” species but it soon became evident that their relationships were not so close to the Andean Telmatobius (and allied genera) as might be surmised from perusal of the literature. The study was further enlarged by the fortuitous availability of several rare and/or recently described Argentine and Chilean telmatobiines. The availability of these frogs prompted a re-assessment of the relationships and proposed classification of Alsodes, Eupsophus, Insuetophrynus, and Telmatobufo, as well as a 1 Associate Professor of Zoology, School of Life Sciences, The University of Nebraska, Lincoln, Neb. 68588; and Associate in Herpetology, Museum of Natural History, The University of Kansas, Lawrence 66045. 2 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY re-evaluation of the putative tribes Alsodini, Batrachylini, and Telmatobiini of the leptodactylid subfamily Telmatobiinae. The extra—andean assemblage of Telmatobius consists of seven species: grandisonae Lynch, nitoi Barrio, patagonicus (Gallardo); praebasalticus Cei and Roig, reverberii Cei, solitarius Cei, and somuncurensis Cei. This assemblage of frogs is distributed south of 39° in patagonian South America (Fig. 1). In an attempt to re-evaluate the relationships and classification of the telmatobiine leptodactylid frogs of Patagénia, I have first reviewed the morphology and taxonomy of the seven Patagonian Telmatobius. Secondly, I have expanded our knowledge of the osteology of the several Patagonian genera heretofore poorly known. Following this review, I present an analysis of the taxonomic char- acteristics used for the lower telmatobiines. This analysis forms the basis for an analysis of the relationships of the various genera. In the concluding discussion and summary, I provide a biogeographic rationale and propose a classification that reflects my current under- standing of the relationships of the genera of this leptodactylid stem group. - me 2) i ied SS. «ip er I ! zi L amet os. Fic. l—(A) Distribution of Atelognathus: A. grandisonae (square), A, nitoi (triangle), A. patagonicus (all circles), A. praebasalticus (open circle), A. reverberii (hexagon), and A. solitarius (inverted triangle). (B) Distribu- tion of Telmatobius (sensu stricto). TELMATOBIINE LEPTODACTYLID FROGS OF : PATAGONIA 3 THE PATAGONIAN Telmatobius Gallardo (1962, 1970) placed patagonicus in Batrachophrynus, an assignment contested by Cei (1970a), Cei and Roig (1968), and Lynch (1971). The other six species were described as Telmatobius species (Barrio, 1973, Cei, 1969a, 1970a, Cei and Roig, 1968, and Lynch, 1976), but Gallardo (1970) assigned praebasalticus and reverberii to Alsodes. Lynch (1971, 1972) questioned the propriety of separating Alsodes and Eupsophus and placed praebasalticus and reverberii in Telmatobius. Much of the confusion and debate surrounding the generic placement of these species stems from the limited morphological data. In an attempt to resolve the issues I studied examples of all seven species. The following descriptions include primarily osteo- logical data; brief synopses of soft anatomies are appended. Taxo- nomic conclusions are deferred to a section following the account of morphology. Telmatobius grandisonae is imperfectly known; the following account does not pertain to T. grandisonae. Morphology Cranial skeleton.—In all species the frontoparietals are paired and of limited extent, exposing a large frontoparietal fontanelle. The fontanelle is least extensive in nifoi and somuncurensis. The frontoparietals rest on the posterior edge of the sphenethmoid and do not contact the nasal bones. Posteriorly, the frontoparietals meet in nitoi and nearly do in somuncurensis. In all six species the frontoparietals extend laterally and posteriorly onto the otoccipitals, but in none of the species are the frontoparietals fused with the otoccipitals. The frontoparietals are not ornamented, nor is there any trace of supraorbital flange or crest formation. The nasal bones are comparatively large elements in_ nitoi, patagonicus, praebasalticus, reverberii, and solitarius, but are smaller in somuncurensis (Fig. 2). In all of these frogs the nasals are nar- rowly separated or in tenuous contact. Laterally, the nasals are narrowly separated from the pars facialis of the maxilla except in somuncurensis (in contact). The maxillary arch is incomplete in all six species. The deleted element is the quadratojugal. The alary processes of the premaxillae are directed dorsally or somewhat anterodorsally in all species ex- cept somuncurensis in which they are directed posterodorsally (Fig. 2). The pars dentalis bears pedicellate teeth in all species. Gallardo (1962) originally described patagonicus as a Batrachophrynus. This species differs from the other Patagonian species in having weakly ankylosed teeth on the premaxiallae, maxillae, and prevomers but is never completely edentate. In seven juvenile patagonicus, the 4 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Fic. 2.—Dorsal views of skulls of Patagonian “Telmatobius.” (A) soli- tarius (IBA-UNC uncatalogued 5A), (B) nitoi (CHINM 6875), (C) somun- curensis (IBA—UNC 2135/8), and (D) patagonicus (IBA—-UNC uncatalogued 2A). Line equals 5 mm. number of premaxillary teeth varies from 3-6 (X = 4.50, 2 SE = 0.50, N = 14) per premaxilla. The number of teeth per maxilla in these 7 individuals varies from 0-7 (x= 2.93, 2SE = 1.14,N = 14). The prevomers have 0-1 teeth per odontophore (x = 0.43, 2 SE = 0.27, N = 14). In 11 adults, the number of teeth per maxilla ranges from 3-16 (x = 10.14, 2 SE = 1.55, N = 22); the number of teeth per premaxilla ranges from 4-10 (x= 7.09, 2 SE = 0.64, N = 22); and the number of teeth per prevomerine odontophore varies from 0-2 (x = 1.27, 2 SE = 0.27, N = 22). The palatine process is prominent and easily distinguished from the pars palatina, which is narrow in each of the species but broadens slightly near the point of articulation with the maxilla. The premaxillae are relatively nar- row (as in most advanced frogs)—the maxilla length/premaxilla width ranges from 3.5 (reverberii, solitarius, somuncurensis) to 5.0 (aquatic patagonicus). TELMATOBIINE LEPTODACTYLID FROGS OF PATAGONIA 5 The maxillae bear teeth in all but some young specimens of patagonicus. The pars facialis bears a prominent preorbital process but below and posterior to the orbit the pars facialis is insignificant. The pars palatina is narrow through the length of the maxilla al- though it narrows more markedly posterior to the pterygoid-maxilla articulation. No pterygoid process is developed. The parasphenoid is triradiate in all six species and the cultri- form process extends anteriorly to just posterior to the level of the palatines in nitoi, patagonicus, solitarius, and somuncurensis but extends more anterad (just anterior to level of the palatines) in praebasalticus and reverberii. The parasphenoid alae are oriented at right angles to the cultriform process (or feebly deflected pos- teriorly ). No odontoids or ridges appear on the ventral surface of the bone. Fic. 3.—Ventral views of skulls of Patagonian “Telmatobius.” (A) soli- tarius (IBA—-UNC uncatalogued 5A), (B) nitoi (CHINM 6875), (C) somun- curensis (IBA-UNC 2135/8), and (D) patagonicus (IBA-UNC uncatalogued 1A). Line equals 5 mm. 6 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY The prevomers are entire in each of the Patagonian Telmatobius but are relatively small bones. The dentigerous processes are usually present (except in small individuals) and bear teeth. The processes are situated between the choanae. The prevomers do not contact any of the other dermal bones of the skull. The palatines are uni- formly present but are small, not contacting the maxillae or the sphenethmoid except in somuncurensis (Fig. 3). In all of the spe- cies there is an anterior flange on the -palatines—these are best developed in nitoi and solitarius. The triradiate pterygoids are relatively small. The anterior ramus articulates with the maxilla in an articulation extending nearly to the planum antorbitale (patagonicus, somuncurensis) or is only briefly in contact with the maxilla (nitoi, solitarius). The median ramus is reduced in all of these frogs and does not bear a bony articulation with the otic capsule in nitoi, reverberii, and soli- tarius; the median ramus is larger and does contact the otic capsule in patagonicus, praebasalticus, and somuncurensis. The zygomatic rami of the squamosals are present and moder- ately long. The otic rami are small and directed medially toward the cristae paroticae but do not articulate or overlap the cristae paroticae. The septomaxillae are uniformly present and proportionately large in each of these species; the elements are least large in somuncurensis. The sphenethmoid is weakly ossified in five species but is extensive in nitoi (extending anteriorly to near the anterior edge of the nasals and posteriorly to enclose the optic foramen). In the other species, the ossification extends to beneath the posterior edge of the nasals (patagonicus, solitarius) or fails to reach the nasals (other three species). The otoccipital is weakly ossified in all species except nitoi and the prootic and exoccipital are clearly defined. The epiotic .emi- nences are relatively low and the crista parotica is best described as short and stocky. An operculum is evident in each of the species but the plectrum (columella) is present only in somuncurensis. (The cavum tympanicum and tympanic annulus are likewise absent in nitoi, patagonicus, praebasalticus, reverberii, and solitarius; these features are present in somuncurensis). The occipital condyles are not stalked and are not confluent. The condyles are slightly sepa- rated in all species except in nitoi (Fig. 2-3). Postcranial axial skeleton—All six species are uniformly pro- coelous and have 8 presacral vertebrae; the cervical and second vertebra are not fused. The cotylar facets of the atlas are broadly to narrowly separated. Few specimens exhibit the clearly type II atlantal condition previously reported for Telmatobius (Lynch, 1971). Only solitarius and somuncurensis exhibit the type II (closely TELMATOBIINE LEPTODACTYLID FROGS OF PATAGONIA 7 juxtaposed cotyles) atlantal condition. The other four species have more widely separated cotyles approaching a type I condition. I am reluctant to term the atlases of these four species as type I. The neural arches are not imbricate (Fig. 4), although the exposure of the nerve canal varies concordantly with general levels of ossification (least non-imbricate in nitoi and solitarius—most non-imbricate in patagonicus and reverberii). The neural arches lack crests. Verte- brae II through IX (sacrum) bear transverse processes and lack ribs (even in the newly metamorphosed individuals). With the exception of reverberii (see below), the transverse processes are prominent on vertebrae II-VIII. Those of vertebrae IJ-IV are per- ceptibly broader than those of vertebrae V—VIII. The transverse processes are deflected anteriorly on vertebrae II, III, VII, VIII, and sometimes VI, and are deflected posteriorly on vertebrae IV, V, and sometimes VI. Those of VI are oriented anteriad in nitoi and patagonicus but posteriorly in praebasalticus, solitarius, and so- muncurensis. The widths of the transverse processes (measured from the tip of the process to the base of the prezygapophysis along the anterior edge of the process) are in order of decreasing length—IIJ, II and IV, V, VI, VII, VIII. In patagonicus, the transverse processes of vertebrae II-IV are broader than or as broad as the sacral dia- pophyses (98-118%). The processes of vertebrae V—VIII are much narrower than the sacral diapophyses (49-68%). The transverse processes of vertebra III are broader than the sacral diapophyses in praebasalticus and somuncurensis but not in nitoi or solitarius (Fig. 5). In these last two species the processes on [I-IV are Fic. 4.—Anterior views of atlases of Patagonian “Telmatobius.” (A) nitoi (CHINM 6875), (B) patagonicus (IBA-UNC uncatalogued 1A), (C)_soli- tarius (IBA—UNC uncatalogued 5A), (4) somuncurensis (IBA-UNC 2135/8), and (E) Alsodes gargola (CHINM 7082). Line equals 5 mm. 8 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY QS } 3 ae c 9 | > Fic. 5.—Dorsal views of vertebral columns of Patagonian “Telmatobius.’ (A) somuncurensis (IBA-UNC 2135/8), (B) nitoi (CHINM 6875), (C) patagonicus (IBA—UNC uncatalogued 1E), and (D) solitarius (IBA-UNC uncatalogued 5A). Lines equal 5 mm (scale between A and B applies to both). broader than the processes on V—VIII. The same situation is the case in praebasalticus. In somuncurensis, the processes of vertebrae II, II, and IV are subequal in width and not strikingly broader than the processes of vertebra VI. The sacral diapophyses are deflected posteriorly in each species. The diapophyses are most dilated (ratio of length to width) in nitoi (0.96) and patagonicus (0.84), but are much less dilated in praebasalticus (0.64), solitarius (0.71), and somuncurensis (0.54). In no case could the diapophyses be described as “expanded or dilated” (and thus equivalent to the character state seen in many bufonids), but they are more dilated than is the case in most telmatobiine frogs. . In terms of vertebral morphology, reverberii is the most distinc- tive of the Patagonian Telmatobius. The transverse processes of vertebrae V-VIII are extremely short. The deflection of the proc- esses of the posterior vertebrae cannot be satisfactorily described but the deflection of the anterior vertebrae (II-IV) conforms with that seen in the other Patagonian species. The transverse processes of vertebrae II-IV are broader than the sacral diapophyses (103-147%). The sacral diapophyses are not dilated but are weakly deflected posteriorly. The centra are proportionately very broad (compare Fig. 6 A and D) when compared with those of adults of the other five Patagonian Telmatobius. The skeletons of reverberii available to me include a number of juveniles; the largest specimens are not mature (26.5-27.5 mm SVL contrasting with adult size range of 35-38 mm given by Cei, 1969a). Comparing the skeletons of the TELMATOBIINE LEPTODACTYLID FROGS OF PATAGONIA 9 largest reverberii available to me with those of immature patagoni- cus (Fig. 6 C) prompts the conclusion that the broad centra of my reverberii are reflections of immaturity. The failure of the centra of the anteriormost vertebrae to ankylose with the neural arches is also a reflection of immaturity. The lack of sacral diapophyseal dilation may also be due to age but the brevity of the transverse processes of the posterior vertebrae in reverberii appears not to be due to age. The centra appear to be stegochordal but my prepara- tions do not permit a definitive statement. The ilial shaft is elongate bearing little or no dorsal crest (Fig. 7). The dorsal prominence is small, conical, and directed dorso- laterally (a protuberance is not clearly evident). The dorsal acetab- ular expansion is small with little dorsal vector. The ventral acetab- ular expansion is moderate-sized. The angle of ventral acetabular expansion is about 90°. The acetabulum is large and weakly emargi- nate dorsally. The preacetabular zone is broad. The ischium is small and lacks prominent processes. The pubis is small and car- tilaginous. The hindlimb skeleton is not noteworthy. The femur and tibio- fibula lack processes or hooks. The tibulare and fibulare are not fused. Two tarsalia are present. The metatarsals and phalanges are not modified, and the primitive phalangeal formula (2-2-3-4-3) The sacrococcygeal articulation is bicondylar in all six species. The coccyx does not bear transverse processes or prezygapophyses. is unmodified. The terminal phalanges are knobbed. If a prehallux — — CK Fagg os Fic. 6.—Vertebral columns of Patagonian “Telmatobius.” (A and B) dor- sal and ventral views, reverberii (IBA—UNC uncatalogued 4A); ventral views of that of a (C) juvenile patagonicus (IBA—UNC uncatalogued 2C) and (D) an adult patagonicus (IBA-UNC uncatalogued 1A). Scale for A and B be- tween them; lines equal 5 mm. 10 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Fic. 7.—(A) prepollex, thumb, and second metacarpal of left hand of nitoi (male, CHINM 6875), (B) palmar view of hand of patagonicus (male, IBA—UNC uncatalogued 3A), (C) prepollex, thumb, and second metacarpal of left hand of somuncurensis (male, IBA-UNC 2135/8), and (D) pelvis of patagonicus (IBA—UNC uncatalogued 1E). Line equals 5 mm. is present, the element is cartilaginous and does not pick up Alizarin stain. Pectoral girdle and appendages.—The pectoral girdles of all six species are arciferal. Procoracoid cartilages are present extending laterally along % to % the width of the clavicles. The procoracoids are fused anteromedially in patagonicus but are free in somuncuren- sis. The epicoracoidal cartilages are free for their entire lengths (except at the precoracoid or procoracoid bridge) and end pos- teriorly as prominent epicoracoidal horns lateral to the sternum. The omosternum is uniformly present. It is narrow and elongate in all species except somuncurensis in which it broadens anteriorly (Fig. 8). I found no omosternal ossifications. The clavicles are strongly arched (tips extending anterior to a line between the anterior edges of the scapulae) and in contact medially. The posterior border (on the medial %) bears a thin flange which increases the length of the bone (this is simply ossification of 4 TELMATOBIINE LEPTODACTYLID FROGS OF PATAGONIA 11 the procoracoidal cartilages). The clavicle does not overlay the pars acromialis of the scapula. Clavicle width/scapula width is 0.95-1.2 in the six species. The scapula is proximally bicapitate. The coracoids are narrowly dilated at their distal and proximal ends, narrower than the clavicles, and somewhat thicker elements than the clavicles. The sternum is longer than broad in all six species. In all but somuncurensis, the sternum is plate-like, weakly expanded postero- laterally, weakly pointed posteriorly, and lacks internal ossification (Fig. 8). In somuncurensis, the sternum is deeply incised posteriorly (for about one-half its length) and somewhat expanded postero- laterally. Anterior to the incision are two rectangular ossifications. The skeleton of the forelimb is not noteworthy. The humerus bears a normal-sized crista ventralis (deltoidea), a narrow crista lateralis, and a slightly larger crista medialis. On the humeri of nitoi, the spina tuberculum medialis is visible (Fig. 9). The fore- limbs of the males of these species are not massive when compared to those of females, and that is reflected in the modest development of humeral crests. The phalangeal formulae for the manus is the primitive condition (2-2-3-3), and the terminal phalanges are knobbed. The prepollex consists of three elements and is small (less than half the length of the first metacarpal). The inner meta- carpal of males of nitoi and somuncurensis bears a short flange at the midpoint of the medial margin of the bone (Fig. 7). This condi- tion is not seen in patagonicus, praebasalticus, or solitarius. I do not have adult males of reverberii. Secondary sex characteristics—Males of each of the six species have nuptial pads on the thumb. In some species the excrescences Fic. 8—vVentral views of pectoral girdles (excluding scapulae, supra— scapulae, and cleithra) of (A) nitoi (CHINM 6875), (B) patagonicus (IBA- UNC uncatalogued 3A), and (C) somuncurensis (IBA-UNC 2135/7). Line equals 5 mm. Cartilage is stippled. 12 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Fic. 9.—Right humeri of male Patagonian “Telmatobius”; medial, anterior, lateral, and posterior views, respectively, of Alsodes gargola (A, A’, A”, A’; CHINM 7082), Atelognathus nitoi (B, B’, B”, B’”’; CHINM 6875), and Somuncuria somuncurensis (C, C’, C”, C’’: IBA-UNC 2135/8). Line equals 5 mm. extend onto the second and even third fingers. In every case the nuptial spines are small, close—-set, and numerous, forming a rough- ened pad. None of the six species shows any trace of nuptial arma- ment on the chest, and all species appear to lack spicules on the TELMATOBIINE LEPTODACTYLID FROGS OF PATAGONIA 13 skin of the dorsum and hind legs. As mentioned above, the fore- arms of the male are not greatly enlarged. On the basis of data provided by Cei (1969a, 1970a, 1972), Barrio (1973), and Cei and Roig (1968), there seems to be no appreciable size difference between males and females. Vocal sacs and slits are absent in all species. Other traits —The tympanic annulus is visible only in somuncu- rensis; this structure as well as the cavum tympanicum and plectrum is lost in the other five species. The pupil is a horizontal slit in each of the six species. The meniscus noted by Cei (1969a, 1972) in somuncurensis is not found in the other five species. The thumb is slightly longer than the second finger. The skin of the dorsum is smooth or weakly granular in the six species and not provided with obvious parotid, lumbar, or inguinal glands. This skin of the venter is smooth. The peculiar cloaca of somuncurensis (see Cei, 1969a, for illustrations) is unique. Tarsal folds occur in five species but are absent in somuncurensis. The folds are weakly developed and nar- row in praebasalticus, but prominent in nitoi, patagonicus, reverb- erii, and solitarius. Two metatarsal tubercles, the inner larger than the outer and neither spade-like, occur in each of the six species. Considerable variation is exhibited in toe webbing. The toes of nitoi and patagonicus are fully webbed although the litoral popula- tions of patagonicus have one phalanx of the fourth toe free of webbing. Two to two and one-half phalanges of the fourth toe are free of webbing in praebasalticus and reverberii (these species may be described as having the foot “half-webbed”). Three phalanges are free (one-third webbed) in somuncurensis and three and one- half (one-fourth webbed) in solitarius. The posterior edge of the tongue is free (non-adherent to floor of mouth) and not notched in any of the species. Prevomerine odontophores are usually visible (if not visible, palpable) and are situated between the moderate-sized choanae. -Karotype data are available for five of the species. Four have diploid counts of 26 chromosomes (Barrio, 1973, Barrio and Rinaldi de Chieri, 1970, Cei, 1969b )—nitoi, patagonicus, praebasalticus, and reverberii. Barrio (1973) reported a secondary constriction on chromosome 6. The karotypes of nitoi, patagonicus, and praebasalti- cus consist of five pairs of large chromosomes and eight pairs of small chromosomes. Cei (1969b:269) mentioned that somuncuren- sis has a 2N of 22. Taxonomic Conclusions The morphology of the Patagonian “Telmatobius” provides sup- port for the following taxonomic conclusions: 14 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 1) Gallardo’s (1962, 1970) assignment of patagonicus to Ba- trachophrynus cannot be supported because patagonicus has teeth on the maxillary arch and prevomer ( Batrachophrynus is edentate ), and patagonicus has a narrow pars palatina of the premaxilla and lacks a quadratojugal. In patagonicus the transverse processes of the posterior presacral vertebrae are shortened and the sacral dia- pophyses are weakly dilated. The tongue is present and its posterior edge is free in patagonicus, unlike the condition in the two species, of Batrachophrynus. The similarity of patagonicus to Batrachophry- nus is best viewed as convergence through adaptation to similar habitats. 2) Gallardo’s (1970) assignment of praebasalticus and reverberii to Alsodes is rejected on two bases. Firstly, neither species exhibits the greatly enlarged forearm (and humeral spines or flanges) pur- portedly characteristic of males of Alsodes; neither species has nup- tial armature on the chest of the male, and neither species exhibits the expanded and posteriorly notched sternum characteristic of Alsodes (Gallardo, 1970). Secondly, these species share a number of character states of skeletal features with nitoi, patagonicus,-and solitarius (viz., lack of quadratojugal, very large frontoparietal fon- tanelles, and relatively large juxtaposed nasals) in contrast to the character-states for the same characteristics in the other species usually assigned to Alsodes (gargola, montanus, monticola, and nodosus ). 3) The six species are not members of Telmatobius. Unlike Telmatobius, the Patagonian species lack quadratojugal bones. Loss of the quadratojugal is rare among leptodactyloid frogs—only seven other leptodactyloid genera exhibit the loss (the Cycloranine Nota- den, the Elosiine Crossodactylus, the Leptodactylines Pleurodema and Pseudopaludicola, and three Telmatobiines, Batrachyla, Hy- lorina, and Insuetophrynus ). 4) One of the Patagonian species, somuncurensis, exhibits a low character-state congruence with the other five. Unlike those species, somuncurensis has moderate-sized, medially separated, ellipsoid nasal bones, contact between the maxilla and nasal, posterodorsally directed alary processes of the premaxillae, fully developed ears, a large, spatulate omosternum, a deeply notched (bifurcate)_ster- num bearing a pair of endochondral ossifications, 11 pairs of chromo- somes, and lacks a tarsal fold. One other trait of somuncurensis, unusual among Neotropical leptodactyloids, is inguinal amplexus (Cei 1972; 447, fig. 5). The architecture of the skull, vertebral column, and pectoral girdle of somuncurensis are similar to the conditions seen in the leptodaetyline genus Pleurodema. The chromosome number for somuncurensis (2N = 22) agrees with the modal count for the TELMATOBIINE LEPTODACTYLID FROGS OF PATAGONIA | 15 Leptodactylinae rather than that for the lower telmatobiines (2N = 26). Although the sternal condition in somuncurensis approaches the condition seen in several Plewrodema, no sternal style is devel- oped. The type II cervical cotylar arrangement in somuncurensis is unlike the type I arrangement found in Pleurodema and the other leptodactylines. The occurrence of inguinal amplexus is unusual among Neotropical leptodactyloids but is duplicated in Pleurodema bufonina and the bufonids of the genus Osornophryne. The low character—state congruence between somuncurensis and the other Patagonian Telmatobius coupled with a number of simi- larities between the seven species and between somuncurensis and the leptodactylines of the genus Pleurodema suggest that somuncu- rensis should be placed in its own genus and viewed as annectant between the Patagonian Telmatobius and Pleurodema. 5) The remaining Patagonian species cannot be placed in any currently recognized genus of the Leptodactylidae. The architecture of the pectoral girdle, skull, and vertebral column, as well as the simple digits (no dermal scutes, no discs or pads), and presence of aquatic larvae in the life cycle support assignment to the Telmato- biinae. Within the Telmatobiinae, the Grypiscini and Eleutherodac- tylini exhibit more advanced and/or specialized features than seen in the Patagonian species (Lynch, 1971). The Odontophrynini have a ceratophryine-type ilium and markedly differ in the arrangement of skull bones and in pectoral architecture (Lynch, 1971). The Patagonian species do exhibit resemblances to those telmatobiines placed in the tribes Alsodini, Batrachylini, and Telmatobiini. These three putative tribes include the following genera (not all previously recognized by all authors): Alsodes, Batrachophrynus, Batrachyla, Caudiverbera, Eupsophus, Hylorina, Insuetophrynus, Telmatobufo, Telmatobius, and Thoropa. Barrio and Rinaldi de Chieri (1971) suggested that Limnomedusa (placed in the Leptodactylinae by Lynch, 1971) also belongs with this assemblage. Atelognathus new genus Type species ——Batrachophrynus patagonicus Gallardo, 1962. Diagnosis.—A genus of Telmatobiine Leptodactylids unique in having large, frontoparietal fontanelles, short palatine bones (not contacting the maxilla or calcified sphenethmoid), large nasal bones in median contact, lacking quadratojugals, plectra (columellae), tympanic annuli, and cavi tympani. The pectoral girdle is arciferal with an elongate, non-spatulate omosternum, and non-bifurcated sternum lacking endochondral ossifications. The cervical cotylar arrangement is type II or intermediate between types I (concave ) and II (convex). The eight procoelous presacral vertebrae are independent and lack a vertebral shield. The transverse processes of presacral vertebrae V-VIII are short. The sacral diapophyses are 16 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY weakly dilated. The ilium is of the leptodactyline type. Terminal phalanges knobbed. Tadpoles with % tooth rows, labial papillae interrupted anteriorly, pond type. Etymology.—Greek, atelés and gnathos, meaning incomplete jaw; in reference to the lack of a quadratojugal in the maxillary arch. Content.—Six species: A. grandisonae (Lynch), A. nitoi (Bar- rio), A. patagonicus (Gallardo), A. praebasalticus (Cei and Roig), A. reverberii (Cei), and A. solitarius (Cei). Remarks.—Atelognathus is readily associated with the ten genera listed above. In the lack of quadratojugals, Atelognathus resembles Batrachyla, Hylorina, and Insuetophrynus. The three genera differ from Atelognathus in having small, medially separated nasal bones, long palatine bones (resting against calcified sphenethmoid and maxilla), and in having complete ears. Batrachyla differs further in having T-shaped terminal phalanges, the first finger shorter than the second, and the larvae have a complete series of papillae about the mouth. Hylorina has greatly elongated digits, vertical pupils, and the larvae have 2/2 tooth rows. Insuetophrynus has large spines on the thumb and chest of the reproductively active male and has a very narrow frontoparietal fontanelle. Atelognathus grandisonae is the most southern species of the genus. It differs from the other five species in having a dilated ma- nubrial portion of the omosternum, narrowly separated nasal bones, and complete coverage of the frontoparietal fontanelle. More ma- terial is required to verify the generic assignment; perhaps the most critical datum is the breadth of the palatine bones. Somuncuria new genus Type-species.—Telmatobius somuncurensis Cei, 1969. Diagnosis —A genus of telmatobiine leptodactylids unique in having a large frontoparietal fontanelle, long palatine bones, mod- erate—sized nasal bones, plectra, cavi tympani, and tympanic annuli, and in lacking quadratojugal bones. The pectoral girdle is arciferal with an elongate, spatulate omosternum, and the sternum is bifur- cated and bears two endochondral ossifications. The cervical cotylar arrangement is type II. The eight procoelous presacral vertebrae are independent and lack a vertebral shield. The transverse proc- esses of presacral vertebrae V-VIII are short and the sacral diapoph- yses very slightly dilated. The ilium is of the leptodactyline type. Terminal phalanges are knobbed. Tadpoles with % tooth rows, labial papillae interrupted anteriorly, pond type. Amplexus inguinal. Etymology.—The generic name is taken from the name of the isolated Patagonian plateau on which the frog lives. Content.—Somuncuria somuncurensis (Cei); monotypic. Remarks.—Somuncuria somuncurensis is intermediate between TELMATOBIINE LEPTODACTYLID FROGS OF PATAGONIA 17 the Patagonian genus Atelognathus (Telmatobiinae) and the lepto- dactyline genus Pleurodema and is conceivably a survivor of a tran- sitional stage in the early evolution of the Leptodactylinae from the Telmatobiinae. Somuncuria resembles Pleurodema in the arrange- ment of skull bones, its karotype, and in having a sternum approach- ing the stylar condition seen in leptodactylines. In habitus, S. somuncurensis is quite similar to some Atelogna- thus and Telmatobius and rather unlike any Pleurodema or other leptodactyline. It differs from the other lower telmatobiines having lost the quadratojugals in the same ways they differ from Atelogna- thus (see above) and differs from Atelognathus in having broad palatines, a protruding cloaca (Cei, 1969a), and inguinal amplexus. THE Lower TELMATOBIINES Lynch (1971) recognized 25 genera in the leptodactylid sub- family Telmatobiinae and divided the subfamily into five tribes (Al- sodini, Eleutherodactylini, Grypiscini, Odontophrynini, and Telma- tobiini). Subsequently (Lynch, 1973), the Alsodini was partitioned into the Alsodini and Batrachylini. The tribes Alsodini, Batrachylini, and Telmatobiini are collectively termed the lower telmatobiines. Twelve genera make up the lower telmatobiines: Alsodes, Atelog- nathus, Batrachophrynus, Batrachyla, Caudiverbera, Eupsophus, Hylorina, Insuetophrynus, Somuncuria, Telmatobius, Telmatobufo, and Thoropa. Most are rare and accordingly morphological and biological data are scarce. A number of them are illustrated in Fig. 10. Heyer (1975) recognized “the telmatobines,” an informal group- ing including Batrachophrynus, Batrachyla, Caudiverbera, Eupso- phus, Hylorina, Insuetophrynus, Telmatobius, and Telmatobufo. He placed Thoropa in an ecological assemblage containing elosiines and grypiscines as well. In the past five years I have been able to study skeletons of Insuetophrynus and Telmatobufo as well as those of Alsodes gar- gola, A. montanus, Atelognathus, Batrachophrynus brachydactylus, Somuncuria, and several species of Telmatobius. The data derived from study of these critical taxa require a re-interpretation of the relationships of the lower telmatobiines and allow some conclusions not previously possible. Morphological data for these lower telma- tobiines follow: Alsodes gargola Gallardo and A. montanus (Lataste) Skull_—tThe frontoparietals are paired, not fused to the prootics, and moderately extensive exposing a frontoparietal fontanelle (Fig. 11). The fontanelle is narrower than that seen in most species of Atelognathus but more extensive than that illustrated in Telma- 18 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Fic. 10.—Representative tematobiine frogs: From left to right and top to bottom: Atelognathus patagonicus, KU 160427, @, 44.5 mm; Batrachyla taeniata, KU 161457, 2, 30.0 mm; Caudiverbera caudiverbera, KU 161411, 2, 106.5 mm; Eupsophus vertebralis, KU 162236, 6, 49.4 mm; Hylorina sylvatica, KU 161391, 2, 61.7 mm; Insuetophrynus acarpicus, KU 161413, ', 46.9 mm; Telmatobius peruvianus, KU 162058, 6, 53.5 mm; Telmato- bufo venustus, KU 161438, ¢, 75.6 mm. Photographs by William E. Duellman. TELMATOBIINE LEPTODACTYLID FROGS OF. PATAGONIA 19 tobius hauthali or T. marmoratus by Lynch (1971). Anteriorly, the frontoparietals broadly rest on the posterolateral margins of the sphenethmoid. Posteriorly, the elements rest on the otoccipitals. The sphenethmoid is moderate-sized and the nasals rest on the anterolateral margins of the calcified portions of the element. The nasals are broadly separated from one another and from the fronto- parietals; they bear long maxillary processes which nearly reach the maxillae. The nasals are relatively short and teardrop—shaped. The maxillary arch is complete but less massive than that of Telmatobufo venustus. The premaxillae are relatively narrow and bear alary processes of moderate length directed dorsally. The 11-14 premaxillary teeth are moderately long, pedicellate, and well- ankylosed to the jaw. The pars palatina is narrow except where the moderate—length palatine processes are found. The pars facialis of the maxilla is relatively deep and restricted to the snout. The maxilla bears about 30 well—-ankylosed teeth in a row extending to a point just posterior to the maxillopterygoid junction. The pars palatina is narrow and narrows abruptly at the posterior end of the maxil- Fic. 11.—Skulls of (A) Alsodes gargola (CHINM 7082) and (B) Insue- tophrynus acarpicus (CHINM 6903). Line equals 5 mm. 20 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY lopterygoid junction. No obvious pterygoid process is developed. The quadratojugal is of moderate-size and broadly articulates with the maxilla. The septomaxillae are small. The otoccipital is well-ossified with low epiotic eminences. The cristae paroticae are short and stocky in A. gargola and slightly broader and narrower in A. montanus. Both have opercula and lack plectra. The occipital condyles are not stalked and are narrowly separated. The zygomatic ramus of the squamosal is of moderate , length. The otic ramus is short. In A. gargola the otic ramus is directed medially and lacks an otic plate whereas in A. montanus it is not deflected medially but bears a small otic plate. In neither species does the otic ramus or plate contact the crista parotica. The parasphenoid is cruciform. The cultriform process is long and gradually narrows anteriorly in A. gargola. The anterior end of the cultriform process extends to between the palatines in A. gargola but not as far anteriad in A. montanus. The parasphenoid alae are oriented at right angles to the cultriform process in A. gargola but weakly deflected posteriorly in A: montanus. Both have a posterior ramus of the parasphenoid; the ramus is round -pos- teriorly (not indented). The pterygoids are relatively large. The anterior ramus rests on the palatal shelf of the maxilla but does not reach the palatines. The median ramus is of moderate length and abuts against the anterior face of the prootic overlapping the parasphenoid alae. The ptery- goid does not bear a ventral flange. The skull anterior to the palatines is not shortened. The palatines are broad and arched, extending from the maxillae to the spheneth- moid. The anterior flange is better developed in A. montanus but perceptible in A. gargola. The small prevomers lie well anteriad to the palatines, are broadly separated, entire, and bear rows of 6-7 well-ankylosed teeth on prominent odontophores lying between the choanae. Postcranial skeleton.—The vertebral column contains eight inde- pendent, procoelous, presacral vertebrae. The cervical cotyles are moderately separated. The neural arches are non-imbricate and lack crests. Vertebrae II-IX bear transverse processes and lack ribs. The transverse processes of II-IV are larger (thicker and wider) than those of V—VIII. In gargola, the processes of II, VI-VIII are deflected anteriorly and IV-V are deflected posteriorly. In mon- tanus, the processes of II, VII-VIII are deflected anteriorly and II-VI deflected posteriorly. In order of decreasing widths, the transverse processes of gargola are III, IX, II, IV, V, VI-VIII; those of montanus are III, IV and IX, II, V, VI, VII, VIII. The sacral diapophyses are oriented at right angles to the sagittal plane or very weakly deflected posteriorly. The sacral diapophyses are weakly TELMATOBIINE LEPTODACTYLID FROGS OF PATAGONIA 21 dilated (L/W 0.82-0.90 in gargola; 0.77-0.82 in montanus). The sacrococcygeal articulation is bicondylar. The coccyx lacks lateral flanges and is not inflated anteriorly. In A. montanus there is a pair of thin processes arising on the arch (Fig. 12); these are not verte- bral like structures. The ilium bears a very slight ridge—like dorsal crest; the dorsal prominence is the same as that in Insuetophrynus. The clavicles are massive and curved. The omosternum is short and bears a manubrial dilation. The sternum broadens posteriorly, is weakly inflated posteriorly, and is feebly notched. The sternum bears an irregular-shaped endochondral ossification. The humerus in males bears enlarged cristae medialis, lateralis, and ventralis (Bigy9): Batrachophrynus brachydactylus Peters The skeleton of a single specimen of this rare aquatic leptodac- tylid is available. The skull exhibits few departures from the de- scription and illustration of B. macrostomus given by Lynch (1971). The nasals are shorter than in macrostomus but in that species are broadly in contact medially and extend laterally to form sutures with the pars facialis of the maxilla and anterior ramus of the ptery- goid. In B. brachydactylus, the frontoparietals are separated for most of their length exposing a long, narrow fontanelle; the fronto- parietals are in sutural contact posterior to the fontanelle. The cultriform process of the parasphenoid is shorter in B. brachydacty- lus, extending anteriorly to the level of the palatines. Unlike B. Fic. 12.—Coccyges of telmatobiines. (A) Alsodes montanus (IBA-UNC 1646/2), (B) Telmatobufo venustus (KU 159811) and (C) dorsal, (D) lateral, and (E) ventral views of Batrachophrynus brachydactylus (SDSNH 46894). Lines equal 5 mm. 22 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY macrostomus, B. brachydactylus has completely lost the prevomers Both species are completely edentulous and both have broad pars palatina of the premaxilla and maxilla. The median ramus of the pterygoid articulates with the parasphenoid ala and rests on the prootic. A plectrum as well as operculum is present in B. brachy- dactylus. Postcranial skeleton—As in B. macrostomus, there are eight procoelous persacral vertebrae. The cervical and second vertebrae are not fused. The cervical cotyles are narrowly separated. The neural arches are imbricate on anterior vertebrae but non-imbricate on posterior vertebrae; the arches lack crests. Vertebrae IJ-IX bear transverse processes without ribs. The transverse processes of verte- bra II are narrower than those of IIJ-IX (subequal in width). The transverse processes of vertebrae III-VI and IX are deflected pos- teriorly, of IJ and VIII are deflected anteriorly, and of VII are perpendicular to the sagittal plane. The sacral diapophyses are not dilated (L/W 0.58-0.60). The sacrococcygeal articulation is bicon- dylar. The coccyx bears flanges along the anterior one-third of its centrum (Fig. 12) similar to those seen in Bufo blombergi. The anterior end of the coccygeal neural arch is conspicuously inflated and posterolateral to the inflated arch are large spinal nerve fora- mina. More posteriorly one finds the smaller foramina normally seen on the coccyx. No prezygapophyses or transverse processes are borne on the coccyx (in B. macrostomus short transverse processes are found). The pectoral girdle is short, broad, and arciferal. The omoster- num is long and narrow and lacks a manubrial portion. The clavicles are massive and strongly curved; their massiveness is in part due to ossification of the procoracoidal cartilages posterior to the clavicles. The coracoids are less broad and less massive than the clavicles. The sternum is as long as broad ( greatest breadth), gradually broad- ening posteriorly, is notched posteriorly, and bears a broad endo- chondral ossification. The humerus of males does not bear enlarged cristae medialis and lateralis. Insuetophrynus acarpicus Barrio Skull.—The frontoparietals are paired, not fused with the proot- ics, and extensive. A narrow frontoparietal fontanelle is exposed (Fig. 11). The frontoparietals rest on the posterior edge of the sphenethmoid and are broadly separated from the nasals. The nasal bones are broad and short, lie anterolateral to the calcified sphenethmoid, are broadly separated, and do not contact the maxil- lae. The maxillary arch is incomplete; the quadratojugal is absent. The alary processes of the premaxillae are short and directed antero- dorsally. The palatine process is short and broad. The pars palatina TELMATOBIINE LEPTODACTYLID FROGS OF PATAGONIA 23 is narrow except where it broadens into the palatine process. The premaxillae are relatively narrow. The maxillae bear 30-35 well- ankylosed teeth to a point posterior to the maxillopterygoid articula- tion. The pars palatina gradually narrows posterior to the maxil- lopterygoid articulation. No pterygoid process is developed. The otoccipital is weakly ossified. The epiotic eminences are low and the crista parotica stocky. The plectrum and operculum are present. The occipital condyles are not stalked and separated me- dially. The zygomatic ramus of the squamosal is moderately long. The otic ramus is short and directed medially toward the crista parotica but does not overlap the crista parotica. The parasphenoid is cruciform; the cultriform process is long and extends onto the sphenethmoid but does not reach the level of the palatines. The parasphenoid alae are feebly deflected pos- teriorly. The parasphenoid bears a short posterior ramus that is weakly emarginate. The pterygoids are of moderate size. The anterior ramus rests on the maxilla and nearly meets the palatines. The median ramus is relatively long, articulates against the anterior side of the prootic. The pterygoid does not bear a ventral flange. The portion of the skull anterior to the palatines is short com- pared to most frogs. The palatines are nearly straight, thin, and broad, resting on the maxilla and sphenethmoid. The palatines lack an anterior flange. The prevomers are moderately large, entire, well- separated, and bear prominent odontophores lying slightly posteriad to the choanae. The odontophores bear 6-7 well-ankylosed teeth. Postcranial skeleton—The vertebral column consists of eight procoelous presacral vertebrae; the cervical and second vertebrae are not fused. The cotylar facets are moderately separated (type Il). The neural arches are not imbricate (comparable to the state seen in A. patagonicus) and lack crests. Vertebrae II-IX bear trans- verse processes and lack ribs. The transverse processes are broadest on vertebrae II-V and narrow on vertebrae VI-VIII; the transverse processes are deflected anteriorly on vertebrae I, VII-VIII, slightly posteriorly on III, and not deflected on IV-VI. The width of the transverse processes are, in order of decreasing widths, IX, III, IV, II, V, and VI-VIII. The sacral diapophyses are deflected posteriorly and are not dilated (L/W 0.48-0.57). The sacrococcygeal articulation is bi- condylar. The anterior end of the coccyx is not inflated and does not bear transverse processes or any other vertebra-like form. The ilium lacks a crest. The dorsal prominence is elongate and about twice as large as that in Atelognathus. The pectoral girdle is functionally firmisternal but has free epi- coracoid horns. The omosternum is relatively short and bears a 24 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY dilated manubrial portion. The sternum is small and not dilated posteriorly. The humerus bears a normal-sized crista ventralis and insignificant cristae lateralis and medialis. Telmatobius My earlier account of Telmatobius (Lynch, 1971) included data for T. marmoratus and T. schreiteri (or T. hauthali schreiteri). In cranial morphology, T. barrioi, T. culeus, and T. niger agree in all respects with T. marmoratus and T. hauthali. The five species differ in vertebral morphology in that the transverse processes of the pre- sacral vertebrae are quite broad in T. culeus, less so in T. marmora- tus, and even less so in T. barrioi, T. niger, and T. hauthali (Fig. 13). The intrageneric variation seen in this characteristic renders the characteristic of little or no value in inferring relationships. The sacral diapophyses are very slightly dilated (L/W values are 0.56-0.64 for culeus, 0.56-0.59 for marmoratus, 0.56-0.65 for niger, and 0.65 for hauthali). Comparing the pectoral girdles of some additional species to my illustration (Lynch, 1971: 60, fig. 34B), T. barrioi and T. marmora- tus have even less inflation of the omosternum (in each it narrows anteriorly) whereas T. brevipes, T. culeus, and T. niger have a slightly greater inflation than that illustrated, and T. sanborni has considerably greater inflation of an elongate omosternum. All seven species have the same sternal form except that the posterior notch is generally less pronounced. The humerus bears enlarged crests in T. barrioi, T. marmoratus, and T. hauthali, but not in T. brevipes, T. niger, T. sanborni, or T. vellardi. Telmatobufo venustus (Philippi) Skull—The frontoparietals are paired, not fused with the proot- ics, and extensive, nearly meeting along the midline anteriorly and sutured posteriorly. Anteriorly, the frontoparietals rest on the sphe- nethmoid; posteriorly they broadly rest on. the otoccipitals. The occipital artery passes through a roofed canal on the posterolateral edge of the frontoparietals (Fig. 14). The calcified sphenethmoid extends anterior to the nasals which lie anterolateral to the spheneth- moid, are broadly separated, and comparatively short. The nasals have an elongate maxillary process that does not reach the maxilla. The nasals are not in contact with the frontoparietals. The maxillary arch is complete. The premaxillae are relatively narrow and bear short, posterodorsally directed alary processes. The pars palatina is narrow except medially where it expands slightly forming short palatine processes. The premaxillary teeth (10-12) are long, fang-like, pedicellate, and well-ankylosed to the ~ TELMATOBIINE LEPTODACTYLID FROGS OF PATAGONIA 25 ap ae i A WC AS QOD Oy Fic. 13.—Vertebral columns of telmatobiine frogs. (A) Telmatobius culeus, KU 135684; (B) T. hauthali, KU 72879; (C) Alsodes gargola, CHINM 7082; (D) Batrachophrynus brachydactylus, SDSNH 46894; (E) Caudiver- bera caudiverbera, AMNH 51510; and (F) Telmatobufo venustus, KU 159811. Lines equal 5 mm. jaw. The pars facialis of the maxilla is moderately deep and re- stricted to the nasal region. The maxilla bears 34-36 long, pedicel- late, well-ankylosed teeth in a row extending to the level of the pterygoid process. The pars palatina is relatively narrow anterior and broadens into a pterygoid process posteriorly. The quadrato- jugal is of moderate size and broadly articulates with the maxilla. The septomaxillae are moderate-sized bones. 26 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY The otoccipital is well ossified with low epiotic eminences. The cristae paroticae are short and stocky. An operculum (but no plectrum ) is present. The occipital condyles are not stalked and are in median contact. The zygomatic ramus of the squamosal is long and curved. The otic ramus is shorter, turned medially, and bears a broad otic plate which overlays the crista parotica. The parasphenoid is triradiate. The cultriform process is long and extends anteriorly between the prevomerine odontophores. The , parasphenoid alae are oriented at right angles to the cultriform process. Posteromedially, the parasphenoid is emarginate or notched. The pterygoids are relatively small but massive. The anterior ramus articulates with the pterygoid process of the maxilla as well as overlays the pars palatina of the maxilla and nearly contacts the palatine. The median ramus is short but firmly articulates with and overlaps the parasphenoid ala as well as resting against the anterior face of the prootic. The pterygoid does not bear a ventral flange. The skull anterior to the palatines is not especially short as in Insuetophrynus. The palatines are broad, extending from the max- illa onto the edges of the sphenethmoid. The palatines are nearly straight, thin, and bear an anterior flange. The prevomers are rela- tively small, entire, broadly separated, and bear prominent odonto- phores between the choanae. The odontophores bear 7-8 long, fang-like teeth in a transverse row. Fic. 14,—Skull of Telmatobufo venustus (KU 159811). Line equals 5 mm. TELMATOBIINE LEPTODACTYLID FROGS OF PATAGONIA 27 Vertebral column.—The cervical and second vertebrae are fused; the column consists of eight procoelous, presacral vertebrae (only seven elements). The cervical cotyles are narrowly separated. The neural arches are nearly completely imbricate and bear low, ridge— like crests. Vertebrae IJ-IX bear transverse processes without ribs. The transverse processes are broadest on vertebrae IJ-IV, narrowest on V-VIII; all are narrower than those of the sacrum. The sacral diapophyses are broadly dilated (L/W = 1.13-1.17). The sacro- coccygeal articulation is bicondylar. The anterior end of the coccyx bears some vertebral form. Short transverse processes are present as are two pairs of spinal nerve foramina; the anterior end of the coccyx is slightly inflated (Fig. 13). The posterior edge of the neural arch of the sacrum bears a shelf covering part of the gap between the sacrum and coccygeal neural arch. Taxonomic note——As Donoso-Barros (1972) pointed out, Phi- lippi’s Bufo venustus is conspecific with Schmidt’s Telmatobufo bul- locki; the correct combination is Telmatobufo venustus, not Aruncus venustus as suggested by Donoso-Barros. Donoso—Barros also sug- gested that Aruncus valdivianus Philippi was an earlier name for Telmatobufo australis Formas but his argument is untenable for the reasons given by Lynch (1971) in refutation of Gallardo’s (1965) parallel proposal. ANALYSIS OF CHARACTERISTICS Certain characteristics (while often cited) are of little value in demonstrating relationships. These are characteristics having one or more unique character-states. As used here, unique states are those found in only one taxon (“singleton characteristics” of Le- Quesne, 1975). Such unique states are quite distinct from character states termed unique by Wilson (1965), Inger (1967), or LeQuesne (1972, 1975). In the group under discussion unique traits include the following: - Caudiverbera: Casquing of the skull; contact between fronto- parietals and nasals; broad parietal-squamosal contact. Telmatobufo: Fusion of first two vertebrae; sucker-like, ventral mouth of tadpole. Insuetophrynus: Functionally firmisternal pectoral girdle; very small sternum. Thoropa: Elongate, stream-adapted tadpoles. Batrachophrynus: Greatly reduced prevomers; edentulous pre- maxillae, maxillae, and prevomers; coccygeal flanges. Hylorina: Long digits. Atelognathus: Narrow palatine bones. Each of the traits listed above demonstrates the distinctiveness 28 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY of the genus they characterize but does not contribute data to an exercise devoted to examining similarities. Nineteen other characteristics are judged to be useful in assess- ing relationships. Each characteristic is a discontinuous variable and is partitioned into character-states. The character-states are then identified as primitive and derived implying evolutionary direction. The method of establishing evolutionary direction is that used by Tihen (1965) and Lynch (1973, 1975a, 1975b). . For the purposes of analysis, the following putative genera were recognized: (species studied follow in parentheses). Alsodes (A. gargola, A. montanus, A. monticola, A. nodosus, A. vanzolinii), Atelognathus (A. nitoi, A. patagonicus, A. praebasalticus, A. rever- berii, A. solitarius), Batrachophrynus (B. brachydactylus, B. macro- stomus), Batrachyla (B. leptopus, B. taeniata), Caudiverbera (C. caudiverbera), Eupsophus (E. roseus, E. vertebralis), Hylorina (H. sylvatica), Insuetophrynus (I. acarpicus), Somuncuria (S. somun- curensis), Telmatobius (T. barrioi, T. brevipes, T. culeus, T. hau- thali, T. marmoratus, T. niger), Telmatobufo (T. venustus), and Thoropa (T. lutzi, T. miliaris, T. petropolitana). - Of the twenty-three characteristics discussed below, four are rejected as inadequate in assessing relationships. The characteristics are numbered in the same order as they appear in the data matrix (Table 1) and on the cladograms (Figs. 16-17). 1. Maxillary arch—Two character-states are recognized here— a primitive condition in which the maxillary arch is complete (quad- ratojugal present) and a derived state in which the quadratojugal is lost. The derived state is present in Atelognathus, Batrachyla, Hylorina, Insuetophrynus, and Somuncuria. 2. Nasal bones.—The nasal bones are large and in broad median contact in Atelognathus, Batrachophrynus, and Caudiverbera (Fig. 2). This condition is taken as primitive following Tihen’s (1965) reasoning. In Somuncuria (Fig. 2), the nasals are small and tenu- ously in median contact, whereas the bones are small and widely separated in the other genera (Figs. 11, 14). The character-states seen in Somuncuria and the other genera are combined into a single derived state. 3. Exposure of frontoparietal fontanelle——In two taxa, Caudi- verbera and Telmatobufo, the frontoparietals are complete covering the frontoparietal fontanelles. In all other taxa, the frontoparietals exhibit appreciable separation medially exposing some expanse termed a fontanelle. The exposure is narrow in Insuetophrynus and I have termed the character-state as non-exposed. In all others there is at least a broader exposure anteriorly than posteriorly. In part, exposure of a fontanelle is a function of size in that younger, and smaller, individuals have a greater exposure of the fontanelle TELMATOBIINE LEPTODACTYLID FROGS OF PATAGONIA = 29 TABLE 1.—DatTA Matrix FoR LOWER TELMATOBIINE LEPTODACTYLIDS AND Pleurodema. OTU Code? Characteristics? ONS Ae On oO LO MIN 2hS [4 SG develo lo IAUES © WP est Ae oe ee OP IO =O Oi ORO ae OMOMO ne Lali aL PAGTR EL [ip crcett oe Lt a IP @) 2b WS OS Ie Oy Ort kk Oy a IBINGUS Peace rhe an ete OO te OS OS) 0 OM ib O Mab OM ab aw YATE A Ge ee eh ad re OM VOOM les ES ae OO kat 10 CAUD ptt eet eas eh OM Ot Mal oO Ol OO Ut uo OU O Ww 1212S ae ee ee OPO) Oe ONO OS SI ORO ROMO TB INGLAO) Ve ee ee Lott © O Ol O Oar OW oO) © O&O io IN SW ee iol © Osh OS ¢ Lak Owl lh @ il @ ab ih il NE TIN GIN eos ees eee Oa 1.0 i OO -© 2 © Oa Wt OSL © OAL © PEO Series Saad weet Pars O a Mail Mal Qa OM WO il dO @ i ut @ AT) BU Of Dis oe ee Joy ee OPI OMe ORO ROO ROO ROR ORORORO STPE TORR ae ee le OUR OM OPO “OMS Pie Sle Sob al ay) S@ NW resect ine. Le TL O On dh Oak b sks OO) athe dk LWi@s Creal ibs@) [PALA AU somite ve Gumi Oy AE a OF O00 00 el OO ei ar 0 * See legend for Fig. 16 (page 40). * See text, p. 28 ff. than do older and larger examples. As a general corollary adult males have a more extensive fontanelle than do adult females (inas- much as the latter are larger than the former). Hence, comparisons using specimens of different sizes and sexes may yield spurious results. Only the species of Atelognathus can be described as having extensive frontoparietal fontanelles. The fontanelles illustrated by Lynch (1971) for Batrachyla, Hylorina, and Thoropa, while exten- sive, are of a much lesser magnitude. Partly because of the sexual and size-related variation in this characteristic, I have partitioned it into only two character-states. The primitive character-state is the lack of an exposed fontanelle and is present in only three of the genera (Caudiverbera, Insuetophrynus, and Telmatobufo). 4, Canal for occipital artery —In most frogs, the occipital artery passes dorsally over the posterolateral corner of the frontoparietal. In a number of bufonid and leptodactylid frogs the occipital artery lies in a groove or in an enclosed canal. In Caudiverbera and Tel- matobufo, the occipital artery is enclosed in a bony canal. None of the other species has any suggestion of such a feature. The presence of the enclosed canal is considered to be derived. 5. Otic element and otic ramus of squamosal.—Caudiverbera exhibits the most marked otic element with the medial extension of the otic ramus of the squamosal to articulate with the parietals. No other genus of telmatobiines approaches this condition. The otic ramus of the squamosal has a small otic plate and curves me- dially in Hylorina and Telmatobufo. Batrachyla has a narrow otic 30 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY plate but the otic ramus is not deflected medially. In Batracho- phrynus the otic ramus is nearly obsolete but a small otic plate is developed. In the other six genera no otic plate is in evidence although the otic ramus of the squamosal is curved medially in all but Thoropa. Partitioning this characteristic is made more difficult by the lack of comparability between Caudiverbera and the other genera. The development of a temporal arcade in Caudiverbera has obliterated any medial curvature of the otic ramus of the squamosal; further- more, the length of the otic ramus in Caudiverbera may be partially distorted by the casquing of the skull. The otic ramus is not de- flected medially in Batrachyla or Thoropa; both also have otic rami that are relatively long. A solution might be to consider the characteristic as two char- acteristics—the development of an otic plate as one trait and medial deflection or curvature of the otic ramus as a second trait. Such a partitioning is difficult for two reasons: the casquing in Caudiver- bera and the absence of an otic ramus in Batrachophrynus. The latter difficulty could be resolved by considering the otic plate-in Batrachophrynus to be a homologue of the medially deflected otic ramus such as seen in Alsodes, Eupsophus, or Telmatobius. Although not entirely satisfactory, I have treated this character- complex as a single characteristic and have divided it into three character-states. The primitive character-state is that in which the otic ramus is medially deflected and an otic plate is developed; in spite of the objections given above, I have characterized Caudiver- bera, as well as Hylorina and Telmatobufo, as exhibiting this char- acter-state. The other two character-states are considered inde- pendently derived; one in which the otic ramus is long and straight ( Batrachyla and Thoropa) and a second in which the otic ramus is short and curved medially (other six genera). . 6. Pars palatina of premaxilla—Medially, a palatine process is developed on the pars palatina of the premaxilla. In Batrachophry- nus and Caudiverbera, the pars palatina is long; the element is slightly shorter in Eupsophus and Telmatobufo and is markedly shorter in all other genera. A short pars palatina is considered to be primitive and a lengthened element, derived. I have not considered the difference in length between that of Batrachophrynus and Tel- matobufo sufficient to warrant recognition of three character-states. 7. Pterygoid process of the maxilla.—Although the pars palatina of the maxilla narrows perceptibly posterior to the end of the maxil- lopterygoid junction in all species examined, only some species have an increase in breadth of the pars palatina anterior to the end of the maxillopterygoid junction. A narrow pterygoid process is de- veloped in Batrachyla, Eupsophus, and Thoropa (see figures in TELMATOBIINE LEPTODACTYLID FROGS OF PATAGONIA © 31 Lynch, 1971). In Caudiverbera and Telmatobufo a much wider process is developed, and a broad ventral suture develops between the process and the pterygoid. In the other six genera the maxilla tapers gradually or rather abruptly but no process is developed. The strengthened maxillopterygoid junction in Caudiverbera and Telmatobufo is considered to be derived, as is the absence of a pterygoid process. 8. Maxillopterygoid overlap.—The anterior ramus of the ptery- goid rests on the dorsal surface of the palatal shelf of the maxilla. In Atelognathus and Batrachyla the anterior end of the pterygoid extends anteriorly to about mid-orbit; in Eupsophus and Hylorina it extends considerably more anteriad and perceptibly less than in Alsodes, Batrachophrynus, Caudiverbera, Insuetophrynus, Telmato- bius, Telmatobufo, and Thoropa. I have recognized only two char- acter-states; a derived condition for Atelognathus and Batrachyla, and a primitive condition in the other nine genera. 9. Articulation of median ramus of pterygoid—The median ramus of the pterygoid articulates against the anteroventral surface of the prootic and may or may not contact the anterolateral edge of the parasphenoid ala. The character-state in which the pterygoid and parasphenoid are in contact (Batrachophrynus, Caudiverbera, Eupsophus, Hylorina, and Telmatobufo) is coded primitive; that seen in the other six genera is derived. 10. Length of cultriform process of parasphenoid.—The anterior ramus of the parasphenoid (cultiform process) is very long in Caudiverbera and Telmatobufo; in each the tip separates both the palatines and prevomers. The element is nearly as long in Ba- trachophrynus. In Alsodes, Atelognathus, and Telmatobius, the cultriform process is noticeably shorter but much longer than that in Batrachyla, Eupsophus, Hylorina, Insuetophrynus, and Thoropa. The greatest morphological hiatus is between the second and third groups, but I have separated the characteristic into three character- states. The first and third are treated as equally and independently derived from the second. 11. Separation of occipital condyles and intercotylar separation. —The occipital condyles are not stalked and narrowly separated in all but two genera, Batrachyla and Thoropa. In both, the occipital condyles, while not as markedly stalked as seen in eleutherodacty- line telmatobiines, are feebly stalked and very widely separated. In the other nine genera, the occipital condyles are more closely juxtaposed. The more lateral position of the occipital condyles in Batrachyla and Thoropa is reflected in their widely separated cervi- cal cotyles (type I of Lynch, 1971, concave atlas of Gallardo, 1961, 1965). In Caudiverbera and Telmatobufo the occipital condyles are nearly confluent and the cervical cotyles are narrowly separated 32 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY (type II of Lynch, 1971; convex atlas of Gallardo, 1961, 1965). In Alsodes, Atelognathus, Batrachophrynus, Eupsophus, Hylorina, In- suetophrynus, and Telmatobius the condyles (and cervical cotyles) are somewhat more separated and may approach the condition de- scribed as type I (see Figs. 4 and 13). Barrio (1970) characterized the atlas of Insuwetophrynus as concave (type I) but his illustrations and my specimens require a characterization as type II. The primi- tive condition is type II, the derived condition, type I. 12. Neural arches.—The neural arches are completely imbricate in Caudiverbera and Telmatobufo (Fig. 13); the arches are imbri- cate on the anterior half of the vertebral column and non-imbricate on the posterior half in Batrachophrynus. In the other eight genera, the neural arches are uniformly non-imbricate (Figs. 5, 6, 13). Imbricate arches are coded as primitive (following Tihen, 1965, but not Trueb, 1973) and non-imbricate arches as derived. 13. Coccyx.—Four species exhibit some vertebral form on the anterior end of the coccyx (Alsodes montanus, Batrachophrynus brachydactylus, B. macrostomus, and Telmatobufo venustus) but only Batrachophrynus and Telmatobufo have inflated, neural-arch like anterior ends of the coccyges. The short transverse processes seen in Alsodes montanus are similar to those of Discoglossus, whereas the modifications of the anterior end of the coccyx in Batrachophrynus and Telmatobufo more closely resemble the condi- tion seen in Mertensophryne, in which vertebral reduction is achieved via deletion through the sacrum. If the condition seen in Batrachophrynus and Telmatobufo reflects a recent vertebral dele- tion, we would have an enigmatic situation inasmuch as both have eight presacral vertebrae, the primitive number found in all non- leiopelmatid frogs, because one must argue that the ancestor(s) of Batrachophrynus and Telmatobufo had nine presacrals. I have coded the presence of marked vertebral-like modifications of the anterior end of the sacrum (inflated neural arch, large spinal nerve foramina) as primitive and the absence of such modifications as derived. 14. Omosternum.—All eleven genera have cartilaginous omo- sterna and lack endochondral ossifications within the element. In comparing the omosterna, some obvious variations occur. The omosterna of Caudiverbera and Telmatobufo are short, broad struc- tures (Fig. 15); in Alsodes the manubrial portion of the omosternum is more clearly defined (Fig. 15); in Batrachyla, Eupsophus, Hy- lorina, and Insuetophrynus, the omosterna are slightly longer and the manubrial portion equally well-defined or slightly more clearly defined. The omosterna are much more elongate in Atelognathus, Batrachophrynus, Telmatobius, and Thoropa (Figs. 8 and 15). No manubrium is developed in Atelognathus, Batrachophrynus, and TELMATOBIINE LEPTODACTYLID FROGS OF PATAGONIA = 33 Fic. 15.—Ventral views of pectoral girdles of telmatobiine frogs. (A) Eupsophus roseus, KU 160566; (B) Batrachyla leptopus, KU 125371; (C) B. taeniata, KU 124242; (D) Insuetophrynus acarpicus, CHINM 6907; (E) Hylorina sylvatica, KU 154546; (F) Telmatobius brevipes, KU 131691; (G) Thoropa miliaris, KU 92853; (H) Batrachophrynus brachydactylus, SDSNH 46894; (1) Alsodes gargola, CHINM 7082; (J) A. montanus, IBA-UNC 1646/4; and (K) A. vanzolinii, KU 162206. 34 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY some Telmatobius (barrioi and marmoratus). An ill-defined manu- brium is developed in the other Telmatobius studied (brevipes, culeus, niger, sanborni, and schneiteri) as well as in T. halli and T. peruvianus (Cei, 1962) and in the three species of Thoropa. I consider the elongate omosternum to be a derived state, because such a form would allow a more complex site of muscle attachment. 15. Sternum.—aAll taxa in this complex have sterna, but the sternum is markedly reduced in size in Insuetophrynus (Barrio, 1970). Endochondral ossifications were noted in Alsodes, Batracho- phrynus, Caudiverbera, Hylorina, Telmatobius, and Thoropa mi- liaris. These ossifications are irregular in shape and in occurrence (most frequent in large individuals) and are completely enclosed by cartilage. Such ossifications represent an advanced character-— state over completely cartilaginous sterna and may be ancestral to the character-state(s) where the ossifications form styles and are not enclosed by cartilage. In my previous study of Neotropical leptodactylids (Lynch, 1971) I treated the sternum as exhibiting only two character-states (viz., a primitive state in which the element was cartilaginous with or without calcifications; and a de- rived state in which the element was more complex having a bony style as well as a cartilaginous portion, the xiphisternum auctorum). Heyer (1975) argued that my treatment of the sternum was an oversimplification and used both shape and internal ossifications to partition the character into five character-states. The most common sternal shape (state A) among the 27 species for which I have data is one in which the sternum broadens pos- teriorly (Fig. 151). No obvious identations allow the recognition of separate metasternum and xiphisternum. The posterior edge is generally weakly indented. This character-state is evident in Al- sodes (gargola, montanus), Batrachophrynus, Batrachyla (lepto- pus), Caudiverbera, Telmatobius (7 sp. examined), and Telmato- bufo. A second character-state (state B) is one in which the sides of the sternum gradually narrow (over the anterior % to % of the sternum ) and then are markedly inflated (Fig. 15C and E). In this character-state one can easily imagine metasternal and xiphisternal portions of the sternum. ‘Laxa exhibiting this character-state in- clude Alsodes (nodosus), Batrachyla (taeniata), Eupsophus (ro- seus, vertebralis), and Hylorina. In Atelognathus (5 sp.) and Insuetophrynus the margins of the sternum are parallel and there is no posterior inflation of the element (Fig. 8, 15D). I have grouped these two taxa as state C even though the sternum of Insuetophrynus is very much shorter than that of Atelognathus. The sternum of Thoropa is style-like in shape as noted by Heyer (1975); it gradually narrows, lacks the inflated xiphisternal portion, and is deeply indented. The sternum of Thoropa (Fig. 15G) resembles TELMATOBIINE LEPTODACTYLID FROGS OF PATAGONIA = 35 those grouped as state B in that it gradually narrows; it resembles those of state C in that no xiphisternal portion is developed. Ac- cordingly, I consider the sternal apparatus of Thoropa to represent a fourth state (D). Two genera have species exhibiting two character-states. Two species of Alsodes exhibit state A and another state B. Cei’s (1962) figures suggest that A. monticola exhibits state B. One species of Batrachyla exhibits state A and another state B. Inasmuch as the posterior edge of the sternum in state A and state B are quite broad, these may be functionally identical. If the characteristic is to be used, each OTU must exhibit a single state; accordingly, state A and state B are combined as the primitive character-state; states C and D are considered independently derived. 16. Terminal phalanges.—The terminal phalanges are knobbed in 9 of the eleven genera; the knobbed condition is coded as primi- tive. The terminal phalanges are T-shaped in Batrachyla and Thoropa; the T-shaped condition is coded as derived. 17. Pupil shape——tIn the Patagonian telmatobiines, the pupil is either a vertical slit (primitive character-state: Caudiverbera, Hylorina, and Telmatobufo ) or a horizontal slit (derived character- state: other eight genera). 18. Outer metatarsal tubercle —Caudiverbera and Telmatobufo differ from the other nine genera in lacking an outer metatarsal tu- bercle. The presence of an outer metatarsal tubercle is considered derived. 19. Nuptial armature——Males of all species for which adults are known have spiny excrescences on the thumb. The excrescences are fine in most species and more coarse in others (Schmidt, 1954; Cei, 1962). Small horny spinules develop on the chest, limbs, and sometimes on the skin of the flanks and back in several Telmatobius (Vellard, 1951). The most impressive variation is seen in adult males of Alsodes and Insuetophrynus where patches of large spines occur on the chest (see illustrations by Barrio, 1970; Cei, 1962; Cei and Roig, 1965; and Gallardo, 1970). A few species of Telmato- bius have similar nuptial excrescences on the chest, but the spines are small and do not form discrete patches. The presence of dis- crete pectoral patches is coded as derived; their absence as primi- tive. The remaining characteristics have been widely used in the study of leptodactylids but must be rejected for purposes of infer- ring relationships because either variation is continuous (21, 22) or there is intra~-OTU variation in the states (20, 23). State C of trait 15 occurs in a single OTU and does not yield information; trait 15 is used in the analysis but the evolutionary step required to convert A to C is non-informational. 36 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 20. Development of ear—Although Grandison (1961) provided a clear cautioning relative to the appropriateness of this character- istic, it has continued to be cited in generic diagnoses as a major characteristic (Gallardo, 1970; Barrio, 1970). In many of the Pata- gonian telmatobiines the ear is reduced by covering the tympanum and annulus with skin, reduction of the annulus and tympanum, loss of those elements and the cavum tympanicum, and loss of those elements as well as the plectrum. The last and most extreme condi- tion is exhibited by Alsodes gargola, A. montanus, and A monticola, whereas A. nodosus has a complete, albeit concealed, ear (Grandi- son, 1961). All Atelognathus have lost the tympanum, tympanic annulus, cavum tympanicum, and plectrum, as have Telmatobufo and several species of Telmatobius. The ear is concealed in many other Telmatobius and visible in T. cinereus. The tympanum is visi- ble externally in Batrachyla, Caudiverbera, Eupsophus, Hylorina, and Thoropa, concealed in Insuetophrynus, and greatly reduced and concealed in Batrachophrynus. The intrageneric variability of this characteristic in Alsodes and in Telmatobius is the most conclusive reason for rejecting it as a basis for generic distinctions as well as for demonstrating relation- ships. Within this complex of frogs the limitations of the character- istic are to discussing the distinctions of and relationships of species within a genus. 21. Transverse processes of vertebrae V-VIII.—Within the com- plex of frogs under discussion there is considerable variation in the breadth and orientation of the transverse processes of vertebrae V-VIII (posterior presacral vertebrae of Lynch, 1971, 1973); the greatest difference is between Atelognathus reverberii (Fig. 6) and Batrachophrynus brachydactylus (Fig. 13). Most species exhibit an intermediate condition of slightly narrowed transverse processes. Broad transverse processes (nearly as broad as the sacral dia- pophyses—90%) occur in Batrachophrynus. Batrachyla, Telmato- bius culeus (Fig. 13), T. marmoratus, and Thoropa have only slightly narrower transverse processes (ca. 80% of sacral diapophyses width). The transverse processes gradually narrow in Caudiver- bera; only those of VIII are appreciably narrower than the sacral diapophyses (Fig. 13). The transverse processes are narrower (58-65%) in Alsodes, Eupsophus, and Telmatobius (niger and schreiteri); these species have approximately the same degree of narrowness of the transverse processes as seen in most Atelognathus (nitoi, patagonicus, praeba- salticus, and solitarius). The transverse processes are relatively nar- row (46-52%) in Insuetophrynus and Telmatobufo. The transverse processes are obsolete in Atelognathus reverberi. Intrageneric variation in Atelognathus and Telmatobius suggests TELMATOBIINE LEPTODACTYLID FROGS OF PATAGONIA — 37 that this characteristic is of little value. Telmatobius culeus is a markedly aquatic species and approaches the vertebral form of Batrachophrynus, an aquatic genus; narrower transverse processes are found in the less aquatic species of Telmatobius. However, broad processes should not be viewed as an aquatic adaptation in view of the narrow processes in the markedly aquatic species Atelognathus patagonicus and Xenopus as well as the broad proc- esses in the many relatively terrestrial Eleutherodactylus and Lep- todactylus. Atelognathus solitarius has narrow transverse processes (comparable to A. patagonicus) and is evidently relatively terres- trial (Cei, 1970a). 22. Dilation of sacral diapophyses.—Few Neotropical leptodac- tylids have dilated sacral diapophyses such as are seen in some Australian leptodactyloids and most bufonids. The Patagonian lepto- dactylids are often characterized as having weakly dilated dia- pophyses, a character-state suggesting intermediacy between di- lated and rounded (characterizing most Neotropical taxa). Without quantification, the three verbalized character-states seem discrete. In order to quantify the trait I have used two ratios to describe the degree of dilation. The first is the length (greatest anteroposterior measurement—near lateral edge of diapophysis) divided by the breadth (a point on a line connecting the lateral base of the pre- zygapophysis and the lateral edge of the sacral condyle to the tip of the diapophysis, at midpoint) as a measure of sacral dilation. For most species I have measured only a single specimen (right and left). Although some taxa are very readily distinguished (Batracho- phrynus and Insuetophrynus—0.48-0.60 compared to Caudiverbera and Telmatobufo—1.00-1.17), most taxa exhibit intermediate val- ues. Within Atelognathus values range from 0.64-0.96 suggesting that the L/D ratio is a continuous variable. The L/D ratio is useful only if the widths of the sacral diapophyses are proportionately equal. As an alternative, I measured the greatest length (L above) of the diapophysis as well as the least length to produce a GL/LL ratio. This ratio is independent of diapophysis width and may more closely approximate what many of us have attempted to describe us- ing such epithets as “narrow,” “rounded,” “somewhat dilated,” etc. I have data for 21 species. The smallest GL/LL ratio is seen in Batrachophrynus macrostomus (1.15), the largest in Telmatobufo venustus (2.54). These two species exemplify the character-states described as “rounded or narrow” and “dilated or expanded.” On the basis of these 21 taxa, the GL/LL ratio is a continuous variable. The extremes are readily identified but there seems to be no gap in the values correlating with any two or three character-states. The GL/LL values for five species of Atelognathus nearly span the spec- 38 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY trum of variation (A. praebasalticus, 1.55; A. reverberii, 1.62; A. nitoi, 2:32; A. solitarius, 2.42; and A. patagonicus, 2.49). While computing these values I noted a general inverse corre- lation between the widths of the transverse processes of vertebrae VI-VIII and the dilation of the sacral diapophyses. The correlation is not significant (r = -0.29, p > 0.05) for process widths (VI- VIII) and L/D ratio of the sacral diapophyses nor for process widths and the GL/LL ratios (r = -0.18, p > 0.05). . 23. Humerus of male—Gallardo (1970) emphasized the en- larged forelimbs of the male as a generic feature of Alsodes. Barrio (1970) characterized Insuetophrynus as having enlarged forelimbs and in part used the trait as evidence of a relationship between the two genera. In order to render the trait less subjective, I have redefined it as the development of large cristae lateralis, medialis, and ventralis. Although all species have a moderate to large crista ventralis, most do not have large forelimbs. The massive forelimbs correlate with the hypertrophy of the crista lateralis and crista medialis and are seen in Alsodes gargola, A. montanus, A. monticola, and A. nodosus (and probably A. illotus, if adult males were avail- able); the condition also occurs in several Telmatobius (s.s.), viz., marmoratus, schreiteri. The humerus of male Inswetophrynus re- sembles that seen in most frogs where the crista medialis and c. lateralis are insignificant ridges if detectable at all. Intrageneric variation in character-states precludes use of this trait in attempting to show relationships. ANALYSIS OF RELATIONSHIPS I analysed the relationships using the CLADON programs de- scribed by Bartcher (1966). In using this method, I agree with the assumptions set forth by Camin and Sokal (1965). Certainly of paramount importance is the assumption that shared primitive character-states are evidence of relationship, whereas shared de- rived character-states do not necessarily provide evidence of rela- tionship. I concur with Eaton’s (1970) reasoning that derived character-states may be independently acquired whereas primitive character-states cannot be—sharing primitive character-states is a result of inheritance. Frogs are a morphologically uniform group (Inger, 1967; Trueb, 1973). I am of the opinion that in such groups independent acquisi- tion of derived character-states is probably the rule. Bock’s (1963, 1969) work suggests that this is generally true in birds, another morphologically uniform group whose key adaptations have im- posed constraints on radiation. This is not to deny all value to derived character-states in inferring phylogeny but to restrict the value of such character-states. Independent acquisition of a de- TELMATOBIINE LEPTODACTYLID FROGS OF PATAGONIA — 39 rived state is more likely if two taxa share a single derived state than if two taxa share several derived states (a cluster of derived states could be used to infer relationship). Independent acquisi- tion is considered more likely for those characteristics showing low compatibility with other characteristics. The most parsimonious cladogram for the eleven telmatobiine genera (excluding Somuncuria) requires 46 evolutionary steps (Fig. 16). Three suprageneric groups may be identified. One consists of Caudiverbera and Telmatobufo; the second of Batrachyla, Eup- sophus, Hylorina, and Thoropa; and the third of Alsodes, Atelo- gnathus, Batrachophrynus, Insuetophrynus, and Telmatobius. The relationship of Alsodes, Atelognathus, Insuetophrynus, and Telma- tobius can be expressed with equal parsimony (subcladogram A’) by requiring extra steps in characters 14 (omosternum shape) and 19 (nuptial armature). The distribution of genera on the cladogram does not lend credence to the suprageneric arrangements I advo- cated previously (Lynch, 1971, 1973). No rational argument could be made to include Batrachophrynus, Caudiverbera, Telmatobius, and Telmatobufo in suprageneric group without also including AI- sodes, Atelognathus, and Insuetophrynus. I (Lynch, 1971) placed Batrachyla, Eupsophus (partim), Hylorina, and Thoropa in a tribe (Alsodini) but later (Lynch, 1973) advocated separating Batrachyla and Thoropa as a tribe. Barrio (1971) and Barrio and Rinaldi de Chieri (1971) sug- gested that Limnomedusa belongs to this complex of genera because Limnomedusa differs from all other Leptodactylines in having 26 chromosomes. Bogart (1973a) pointed out that Adenomera also has 26 chromosomes weakening Barrio’s argument. If Limnomedusa is included with the Patagonian telmatobiines in a cladistic analysis, an additional state for character 15 may be employed; the state is shared with Thoropa. The cladogram generated for these 12 gen- era requires 50 steps (Fig. 16). Limnomedusa is situated as a sister group to Hylorina-Eupsophus-Thoropa-Batrachyla. Four dif- ferent, equally parsimonious, trees may be generated (see subclado- grams A’ and B’). Only two evolutionary steps (i.e., derived char- acter-states) link Limnomedusa with Batrachyla and Thoropa; I viewed the linkage as spurious rather than suggestive of relationship. When Pleurodema and Somuncuria are also included in the anal- ysis (Fig. 17), the cladogram undergoes considerable configura- tional change. Batrachophrynus is linked with Caudiverbera and Telmatobufo rather than with Alsodes, Atelognathus, Insuetophry- nus, and Telmatobius. In the subcladogram I prefer, Pleurodema and Somuncuria replaced Batrachophrynus. Limnomedusa remains associated with Eupsophus and Hylorina. The failure of Limnome- 40 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY a ie) R 10) U ” ” 20 ee ee a. oo os ao Sa RS eg 2 oe a Re a SS oe = eZ Oo f-y& 2 < & Lt = xc uw fF a Lin ee ee 3 i2 19 8 5 1 14 1 1 4 14 +19 8 3 15 ae 3 17 6 A 10 2 14 9 3 2 12 10 13 13 12 5 13 4 17 6 7 18 10 =) » Oo a w Ne) >" SON ED SS Se So Wh alee opt Oe OL < co mi ie SPS ee ny ee U - a) a Hi f . ed r ae ~ a > te er iy 3 int cen ‘ (faa \ lie pea, Sua i Dele il hai 5 Created Ritenutomit hor dca . - 7 =f: 1, % iz 7 rsh) f ae Rarytisn A DhY . d bea Mae ; borin “ii i y ! i ORF. Mf . i i ' J + hale riled 2 eur Z V's 10 rl seas T “SNttter o Later Ae MAY JUNE JULY AUG SEPT OCT NOV DEC JAN FEB MARCH APRIL 1972 I973 Fic. 1.—Changes in relative abundance of Sigmodon hispidus populations estimated by live-trap catches at six areas in Douglas, Jefferson and Leaven- worth counties, Kansas. 4 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY study of a cotton rat population on the Nelson Environmental Study Area. Patterns of breeding activity—The proportions of males and females in breeding condition were calculated for each month to determine temporal changes in population breeding activity. Males were judged to be fertile if the epididymal tubules were convoluted rather than looped (Jameson, 1950). Reproductive activity in females was determined by monthly pregnancy rates based on the proportion of females weighing more than 50 g that were pregnant. One 45g female in the June sample was pregnant, but this would appear to be an exceptional case based on the weights at first reproduction observed in this study and reported by other investi- gators (Meyer and Meyer, 1944; Odum, 1955; Goertz, 1965a). The percentages of males and females breeding in each month are shown in Fig. 2. The breeding season for Sigmodon hispidus is seven months in duration and coincides with the normal frost-free period in northeastern Kansas (April to November). Significant differences in breeding activity between months were observed in males (x? = 79,4; dt = 11; P < 005) and in females (x77:= 46.8; df.— 1Je;Buee 005). Temporal changes in breeding activity in males and females appear to be synchronous, although males maintain breeding con- dition later into the fall and become reproductive earlier in the spring than do females. Female breeding activity changes closely paralleled changes in density shown in Fig. 1. Peaks in breeding activity in May and September were reflected in the peak densities observed in June- July and November, and the July decline in reproduction occurred just prior to the decline in numbers in August. There was no breeding activity during the winter months, as evidenced by the lack of pregnant or lactating females during this period. The pattern of breeding activity in males (Fig. 2) is supported by changes observed in average testicular weight. The relative size of the testes was calculated by dividing the testes weight by the body weight. Average testes weights (mg) per 10g of body weight for each monthly sample are presented in Fig. 3. The largest average testicular sizes occurred in June, August, and September. These months were also months of peak reproductive activity as determined by the proportions of fertile males present (see Fig. 2). Testicular regression began in the early fall (October). Smallest testes size was observed in December and the average testes size at that time was only 1/30th the size of testes observed in Sentem- ber. Testes weight increased in the small January and February samples. By April, average testes weight was over 700mg/10g as the breeding season began. Changes in sex ratio—One mechanism by which reproductive output can be increased is through the skewing of the sex ratio of REPRODUCTION OF THE HISPID COTTON RAT 5 PERCENT MALES BREEDING MAY JUNE JULY AUG SEPT OCT NOV DEC JAN FEB MAR APRIL IQ IS) PERCENT FEMALES PREGNANT MAY JUNE JULY AUG SEPT OCT NOV DEC JAN FEB MAR APRIL IN 1973 Fic. 2._Monthly proportions of males and females in breeding condition observed in Sigmodon hispidus populations in northeastern Kansas. a population in favor of females. In this study, males were more abundant than females, with 221 males (57.77%) and 163 females (42.2%) being captured over the 12 month study period. This dif- ference is highly significant (x?:-— 8.76; P < .005). Sex ratio was found to fluctuate from month to month. Ratios favoring males were observed in the winter and fall. Although only one of these’ ratios was statistically different from 1:1 (June), the trend favoring males in the winter was consistent. However, as the breeding season 6 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY progressed, sex ratios became more equitable with equal numbers of males and females being observed. Changes in uterine litter size —Pregnant females were separated into primiparous and multiparous groups to analyze changes in uterine litter size. Uterine embryo counts were made for each parity group within each monthly sample containing pregnant females. The mean numbers of embryos for each parity class in each month are given in Table 1. One-way analysis of variance (Sokal and Rohlf, 1969) indicates that the average number of embryos per pregnant female did not vary significantly from month to month for primiparous females (F, = 1.49; df = 6, 20; P < .10) or for multiparous females (F, = 1.41; df = 6, 36; P < .10). However, 200 100 oe 30 80 70 - fr AVERAGE TESTES WEIGHT (mG) PER!O Grams Body WEIGHT SL O ODANWwoOO MAY JUNE JULY AUG SEPT OCT NOV DEC JAN FEB MAR APRIL 972 I973 Fic. 3.—Average testicular weights for monthly samples of male Sigmodon hispidus from northeastern Kansas. REPRODUCTION OF THE HISPID COTTON RAT rh the average uterine litter size for all multiparous females was significantly larger than that for all primiparous females (ts; = 3.93; df — 68; P< .001). Significant correlations were also observed between litter size and body length (r = .26, df — 68; P < .05) and between litter size and body weight (r = 0.37; df = 68; P < .01). The average uterine litter size for all females was 9.04 + 0.31. Prenatal mortality —From autopsy of female reproductive tracts, it is possible to quantify prenatal mortality of two types. First, differences in the number of corpora lutea and the number of implantation sites provides a measure of pre-implantation loss. Partial post-implantation mortality (excluding whole-litter loss) can be determined by counting the number of embryos being resorbed. Table 1 provides monthly estimates of the percentage of ova lost via pre-implantation mortality and the percentage of embryos lost through resorption. Low levels of both pre- and post-implantation mortality were observed. Variation between months was non- significant for pre-implantation mortality (x? = 2.44; df — 6; P < .o) and for post-implantation mortality (x? = 5.26; df = 6; P < .10). No month showed a loss of ova in excess of 2% and the average loss over all months was 1.21%. An average loss due to resorption of 4.10% was slightly higher than that for pre-implantation morality but in no month was post-implantation loss greater than 6%. There is no discernible pattern of prenatal mortality with respect to either season or population density. Median weight at sexual maturity—Following the method of Leslie et al. (1945), a median weight at sexual maturity was cal- culated. Although some inaccuracy is inherent in using weight as a measure of age in S. hispidus (Chipman, 1965), we feel that weight should allow sufficient accuracy to obtain evidence for any seasonal trends in age at sexual maturity. Animals captured were separated according to sex and then grouped by 20g weight intervals into four-month periods. Female weights were adjusted for all pregnant females by subtracting the weight of the uterus and embryos. The percentage of sexually mature individuals was calculated for each weight class within a time interval. These percentages were then converted to probit values and plotted against the logarithm of average weight of individuals in each weight class. A straight line is fitted and an estimated weight at which 50% of the sample was sexually mature and a standard error of this estimate were calculated. The interval for March 1973 to June 1973 was not included in the analysis because the small sample size of this interval resulted in an indeterminant regression. The median weights at sexual maturity for males and females and their standard errors are given in Fig. 4. Despite some rather large standard errors resulting from small sample sizes in some intervals, there is a trend towards an increasing 8 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY TABLE 1.—Average number of embryos for multiparous and _ primiparous Sigmodon females and estimates of pre- and post-implantation mortality during the breeding season. Sample sizes are in parenthesis. May June July Aug Sept Oct April Total 1972 1972 1972 1972 1972 1972 1973 Mean embryos per multi- 8.67 11.17 9.14 8.50 9.80 10.17 7.00 9.91 parous females (3). (12). (1) ()> Oe) 1012) ates Mean embryos per primi- 6.50 8.29 7.67 8.80 5.00 7.40: 4.00 7.67 parous females (2) (7d, CB) vines oid Th) (5) (dp atm) % Ova lost 2.43: .1.49. 1.74. 1.61,..0.00. . 0.60 O.0OiaRat (41) (202) (115) (62) (57) (167) (18) (662) % Embryos resorbed 2.50 4.48 9.65 0:00 ..5.26 + :5.92. 5:foeeeee (40) (201) (113) (61) (57) (169) (18) (659) weight at sexual maturity as the breeding season progressed. Throughout the breeding season, both sexes showed this pattern and both males and females became sexually mature at approx- imately the same weight. Average body weight.—The reproductive potential of a popula- tion is largely a function of the age structure of the population. Since this age structure is rarely, if ever, static, it is necessary to MARCH - JUNE JULY- OCT. NOV. 1972—FEB.1973 1972 1972 100 100 oe 100 80 80 80 N=54 N=86 N=47 by 60 60 60 ° 9° ° 40 To 40 fo 40 20 20 20 © 30 60 90 120 150 © 30 60 90 120 150 © 30 60 90 120 I50 BODY WT. (GMS) BODY WT. (GMS) BODY WT. (GMS) 100- 100 OF 100 80 80 80 N=94 N=50 60 60 60 To %o % 40 40 40 20 | 20 20 © 30 60 90 120 150 180 © 30 60 90 120 150 © 30 60 90 120150 Booby WT. (GMs) BODY WT. (GMS) Boby WT. (GMS) Fic. 4.—Median weights at sexual maturity in males and females in Sig- modon hispidus. N is the number in each time interval from which the estimate was obtained. REPRODUCTION OF THE HISPID COTTON RAT 9 consider changes in age structure and how these changes might affect reproductive activity. For this study, body weight distribu- tions were determined for each four-month interval, again deleting the March 1973-June 1973 interval because of sample size. Distribu- tions for males and females are given in Fig. 5. In contrast to the pattern exhibited by median weight at sexual maturity, average body weight for both males and females showed a marked decrease as the breeding season progressed. Thus, the age structure of the cotton rat population shifted towards greater proportions of juvenile and sub-adult animals during the breeding season. DIscussion Cotton rat populations from a number of areas in the United States have been shown to undergo annual cycles in abundance which appear to be associated with local climatic conditions (Odum, 1955; Haines, 1961, 1963, 1971; Goertz, 1964; Fleharty et al., 1972; Joule and Cameron, 1974, 1975) and habitat quality (Goertz, 1964). This annual cycle is characterized by peak densities in the fall, a winter decline, and increasing numbers in the spring and summer. This general pattern was observed in this study for S. hispidus in northeastern Kansas. Abundance was lowest in the early spring, and it increased through the summer until peak abundance was reached in the fall. A dramatic reduction in abundance was coincident with the onset of winter and was prob- ably initiated by the cessation of breeding activity. This decline undoubtedly was accelerated by severe winter weather. “Cold weather starvation” (Howard, 1951) may have been a major source of winter mortality since cotton rats are not known to hoard food (Dewsbury, 1970) and extreme cold temperatures restrict foraging activity (Goertz, 1964). Although hard data from this study are lacking, predation may also have contributed significantly to this decline in abundance ( Wiegert, 1972). Reproductive activity was also observed to fluctuate seasonally, with the breeding season restricted to the seven months of the year that are normally frost-free (April to November). Changes in breeding activity were closely associated with abundance changes. Abundance peaks were preceded by periods of intense breeding activity and periods of low breeding activity were followed by reductions in abundance. Changes in breeding activity (Fig. 2) may also be indicative of shifts in population age structure (Fig. 5). For example, the decrease in breeding activity observed in July may reflect the shift from a population composed largely of older, sexually active individuals (spring population) to one in which a large proportion of the individuals in the population are young, sexually immature animals born in the spring and early summer. In contrast, Haines (1961) observed ovulation the year round in 10 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 140 FEMALES 100 - 80 60 40 MEDIAN WEIGHT AT SEXUAL Maturity (GMs) 20 MAR-JUNE JULY- OCT. NOV.- FEB. MAR-JUNE JULY-OCT. NOV-FEB. Fic. 5.—Body weight distributions for males and females in a population of Sigmodon hispidus from northeastern Kansas. cotton rat populations near Austin, Texas, and maximum testes size in February. Goertz (1965a) has suggested that, although the incidence of pregnancy and breeding activity declines during the winter, Sigmodon populations in Oklahoma are also potentially irruptive throughout the year. Some winter reproduction also takes place in cotton rat populations from Florida (Bigler and Jenkins, 1975) and the Gulf Coast of Texas (G. Cameron, pers. comm.), although the magnitude of breeding activity is reduced. Our data indicate that the cessation of breeding activity in cotton rats in northeast Kansas was complete. Pregnant and lactating females were not present in the population from November 1972 to March 1973. During this time, males showed a pronounced testicular regression and no fertile males were captured from December 1972 to March 1973. It appears that under normal winter conditions, cotton rats in northeast Kansas are not capable of reproduction during the winter. Whether or not reproduction occurs during un- seasonably mild winters in Kansas is an open question (Fleharty et al., 1972). Having examined temporal changes in abundance and associated changes in breeding activity, we next consider possible REPRODUCTION OF THE HISPID COTTON RAT 11 mechanisms by which reproductive output could be increased in these marginal cotton rat populations. If a species is polygamous, a sex ratio favoring females would increase a population's capacity for growth. However, on a month- to-month basis, only June 1972 showed a sex ratio significantly dif- ferent from 1:1, and in this month, males were in excess. Biased sex ratio data due to trap type has been reported for several species of small mammals (Keller, 1968; Myers and Krebs, 1972) and S. hispidus has shown differential trap response by weight class (Joule and Cameron, 1974). However, we feel that our three-day trap- ping regime with removal of animals as they were captured provides a reasonably accurate sample of both sexes and weight classes. The excess of males seen in this study is an observation supported by other studies (Goertz, 1965b; Fleharty e¢ al., 1972). One interesting aspect of sex ratio in this study was the statistically nonsignificant but consistent trend towards excess males during the winter and early spring months, ending with the significant difference in June. As the breeding season progressed, sex ratios more closely approx- imated 1:1. Although accurate survivorship data are not available for our study, the slight excess of males seen through the winter may reflect differential survival with regard to sex. Haines (1972) reported a reduction in longevity for sub-adult females in two separate populations of cotton rats from Texas. However, Goertz (1964) observed no differences in male and female survivorship during a winter in Oklahoma. In any event, equal sex ratios were observed during the major portion of the breeding season and there was no evidence of a shift in sex ratio to favor females and thus increase the population’s capacity for growth. From the uterine litter sizes observed in this study, female cotton rats in northeast Kansas are the most fecund yet reported for the species as a whole (Meyer and Meyer, 1944; Odum, 1955; Haines, 1961; Goertz, 1965a; Bowdre, 1971). Our litter size data support Kilgore’s (1970) evidence for latitudinal cline for litter size in S. hispidus. Clinal variation in litter size with both latitude and altitude has been reported in a variety of mammalian species (Lord, 1960; Dunmire, 1960; Keith et al., 1966). It would appear that litter sizes in Kansas cotton rats have been increased by natural selection to maximize reproductive success (Spencer and Steinhoff, 1968). However, this conclusion should be qualified since increased repro- ductive effort per se does not necessarily maximize reproductive success or individual fitness. Randolph et al. (1977) have reported decreased growth rates in neonatal cotton rats from litters that are larger than average and have pointed out that relative fitness of the young may be subject to maternal energetic constraints during pregnancy and lactation. The extent to which energy allocation to reproduction has been altered to accommodate increased litter sizes 12 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY in populations of S. hispidus from northeast Kansas is unknown. Some sort of “trade off’ between litter size and energy allocation may be important in these populations. As previously suggested (Randolph et al., 1977), insights into the evolution of litter size in cotton rats will come only from comparative studies that include considerations of energetics and the effects of the various biotic and abiotic ecological factors (i.e. spatial and temporal variation in resources, predation, etc. ). Cotton rats from northeastern Kansas displayed a pronounced differential reproductive output based on an individual’s parity or previous reproductive experience. Significantly larger litter sizes were observed in multiparous females when compared to primi- parous females. Previous studies dealing with cotton rat repro- duction (Haines, 1961; Goertz, 1965a) have not considered this relationship and comparative evidence is absent. However, no sig- nificant differences in uterine litter size between primiparous and multiparous females were found in 68 pregnant females captured in July 1973 and August 1974 from the states of Veracruz and Tamaulipas, Mexico, nor was litter size correlated with either body weight or body length (McClenaghan, unpubl.). In temperate latitudes, environmental factors acting as sources of stress may have a more pronounced effect on younger individuals reproducing for the first time and small litter sizes in these individuals may be. the result of natural selection operating to enhance the survival of mother and young (Lack, 1954). Goertz (1965a) and Kilgore (1970) both reported seasonal variability in litter size in cotton rats from Oklahoma and Kansas, respectively. In both these studies smaller litters were produced in the cold weather months and larger litters in warm weather months. Such a pattern was not observed in this study. It is possible that the cotton rat population studied here has become adapted to the environment to the extent that litter size is not affected by climatic conditions. It may also be that the results of Goertz (1965a) and Kilgore (1970) were produced by the parity-dependent reproduc- tive effect described in this paper. Cold weather populations, particularly in late winter and early spring, are likely to be com- posed largely of primiparous animals born the previous fall, whereas warm weather populations would contain a large proportion of multiparous females. Thus, a cotton rat population’s potential for reproductive output would be a function of the age structure of the population and the relative proportions of primiparous and multi- parous females. Prenatal mortality was generally low in the cotton rat populations monitored in this study. Pre-implantation mortality averaged only 1.21% of all ova produced and, of those implanted, an average of 4.10% were resorbed. In comparison, Keller and Krebs (1970) REPRODUCTION OF THE HISPID COTTON RAT 13 reported pre- and post-implantation losses in Microtus pennsylvan- icus from Indiana of 9.0% and 6.4%, respectively. Rose (1974) found these values to be 9.46% and 1.55% in M. ochrogaster in northeast Kansas. Kansas cotton rats exhibited no seasonal differences in either pre- or post-implantation loss, suggesting that neither climatic nor density factors play an important role in regulating reproductive output through prenatal mortality. The average age at sexual maturity has been demonstrated to be one of the most important parameters determining the rate at which a population increases in density (Cole, 1954). Lewontin (1965) has shown theoretically that by reducing the average age at sexual maturity by 10%, a population’s potential for increase is enhanced by 100%. Mayr (1963) suggested that there will be selec- tion for an early age at sexual maturity following a rapid decline in numbers in microtine rodent populations or in other species of organisms which are subject to pronounced decreases in density. Cotton rat abundance in northeast Kansas declined rapidly over the winter, after reaching a peak in the fall. Spring cotton rat populations are characterized by low densities following the winter decline and it was at this time that the lowest median weight at sexual maturity was observed. This finding should not be inter- preted as selection for an early age at first reproduction in spring cotton rat populations. In fact, these populations are composed entirely of individuals that have survived the winter and are, thus, at least four months old. Rather, this illustrates the fact that cotton rats frequently lose weight during the winter even though they are becoming chronologically older (N.A. Slade, pers. comm.). This observation underscores the need for caution when using weight as a criterion of age in studies of S. hispidus. As the breeding season progressed, the median weight at sexual maturity increased while the average population body weight decreased (Figs. 4 and 5). Temporal shifts in body weight distributions like those observed here have been reported for S. hispidus in other studies (Fleharty et al., 1972; Joule and Cameron, 1975). Keller and Krebs (1970) have shown that median weights at sexual maturity in populations of Microtus pennsylvanicus and M. ochrogaster in southern Indiana are lowest following a decline in population density and greatest when the populations are at peak densities. This pattern can be observed in Fig. 4. At least part of the large increase in median weight at sexual maturity seen in the November 1972-February 1973 interval may reflect the decrease in breeding activity in S. hispidus associated with the onset of winter. Occurring concomittantly with the cessation of breeding activity in late fall is a noticeable change in the quality of resources utilized by S. hispidus. The fall senes- cence of plant species used as food by cotton rats may act to inhibit reproduction (Negus and Berger, 1977). On the basis of the data 14 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY from this study, it is difficult to determine whether the changes in median weight at sexual maturity observed reflected a density- related response or a response to changes in resource quality. It is possible that both these factors are operating in concert on these populations of cotton rats. In summary, several aspects of reproduction in Sigmodon hisp- idus from northeastern Kansas distinguish these populations from others in different parts of the species distribution. In response to a shortened breeding season and a lack of winter breeding, reproduc- tive output in terms of uterine litter size has been increased to levels not previously reported for the species. In addition, individual reproductive output is dependent to a large extent on previous reproductive experience, with primiparous females having signif- icantly smaller litters than multiparous females. Both these phe- nomena could be a result of natural selection acting to maximize reproductive success. Reproductive effort has exceeded winter losses on the average, although severe winters probably result in the extinction of some local populations. The end result of these reproductive adjustments is the perpetuation of Sigmodon hispidus in temperate regions only recently colonized by the species. ACKNOWLEDGEMENTS We wish to acknowledge Robert K. Rose and Kevin Whittaker for their assistance in the field and laboratory. We also are grateful to James L. Hamrick III, Robert S$. Hoffmann, Norman A. Slade and Max R. Terman for their comments and criticisms concerning our manuscript. Financial support for this study was provided by grants from the National Science Foundation (GB29135), the University of Kansas Museum of Natural History Watkins Fund, the University of Kansas General Research Fund (3345-5038) and the University of Kansas Biomedical Sciences Support Fund (4076-5706). A Sum- mer Dissertation Fellowship from the Graduate School of the University of Kansas to L.R.M. facilitated the writing of this paper. Computer facilities were generously provided by the University of Kansas Computation Center. LITERATURE CITED Bicier, W. J., Jenxrns, J. H. 1975. Population characteristics of Peromyscus gossypinus and Sigmodon hispidus in tropical hammocks of South Florida. J. Mamm. 56:633-644, Bowpre, L. P. 1971. Litter size in Sigmodon hispidus. Southwest. Natur. 16; 126-128. CuipMan, R. K. 1965. Age determination of the cotton rat (Sigmodon hispidus). Tulane Stud. Zool. 12:19-38. Cockxrum, E. L. 1948. Distribution of the cotton rat in Kansas. Trans. Kansas Acad. Sci. 51:306-312. REPRODUCTION OF THE HISPID COTTON RAT 15 Cotz, L. C. 1954. The population consequences of life history phenomena. Quart. Rev. Biol. 29: 103-137. Dewssury, D. A. 1970. Food hoarding in rice rats and cotton rats. Psychol. Rep. 26:174. DunmireE, W. W. 1960. An altitudinal survey of reproduction in Peromyscus. Ecology 41:174-182. FLEHARTY, E. D., Mares, M. A., Cuoate, J. R. 1972. Population fluctuations in Sigmodon hispidus: factors influencing a “crash”. Bull. So. California Acad. Sci. 71:132-138. Genoways, H. H., Scuuitrer, D. A. 1967. Northward dispersal of the hispid cotton rat in Nebraska and Missouri. Trans. Kansas Acad. Sci. 69:356-357. Grier, H. T. 1967. The Kansas small mammal census: terminal report. Trans. Kansas Acad. Sci. 70:505-518. GoErtTz, J. W. 1964. The influence of habitat quality upon density of cotton rat populations. Ecol. Monogr. 34:359-381. GorrtTz, J. W. 1965a. Reproductive variation in cotton rats. Amer. Midl. Natur. 74:329-340. GoERTz, J. W. 1965b. Sex, age, and weight variation in cotton rats. J. Mamm. 46:471-477. Hanes, H. 1961. Seasonal changes in the reproductive organs of the cotton rat, Sigmodon hispidus. Texas J. Sci. 13:219-230. Haines, H. 1963. Geographical extent and duration of the cotton rat Sigmodon hispidus, 1958-60 fluctuation in Texas. Ecology 44:771-772. Hares, H. 1971. Characteristics of a cotton rat (Sigmodon hispidus) population cycle. Texas J. Sci. 23:3-27. Hisparp, C. W. 1960. An interpretation of Pliocene and Pleistocene climates in North America. 62nd Annual Report, Michigan Acad. Sci., Arts, Ltrs., 1959-1960, pp. 5-30. HorFrMann, R. S., Jones, J. K. 1970. Influence of late-glacial and post-glacial events on the distribution of Recent Mammals on the northern Great Plains, pp. 355-394. In Pleistocene and Recent environments of the Central Great Plains (W. Dort, Jr. and J. K. Jones, Jr., eds.). Spec. Publ. 3, Dept. Geol., Univ. Kansas, viii + 433 pp. Hooper, E. T. 1949. Faunal relationships of recent North American rodents. Univ. Michigan, Misc. Publ. Mus. Zool. 72:1-28. Jameson, E. W., Jn. 1950. Determining fecundity in male small mammals. J. Mamm. 31:433-436. JouLE, J., Cameron, G. N. 1974. Field estimation of demographic parameters: influence of Sigmodon hispidus population structure. J. Mamm. 55:309-318. Jouz, J., Cameron, G. N. 1975. Species removal studies. I. Dispersal strate- gies of sympatric Sigmodon hispidus and Reithrodontomys fulvescens populations. J. Mamm. 56:378-396. Kerrn, L. B., Roncstap, O. J., MEsLtow, E. C. 1966. Regional differences in reproductive traits of the snowshoe hare. Can. J. Zool. 44:953-961. KELLER, B. L., Kress, C. J. 1970. Microtus population biology III. Reproduc- tive changes in fluctuating populations of M. ochrogaster and M. penn- sylvanicus, in southern Indiana. Ecol. Monog. 40:263-294. Kitcore, D. 1970. The effect of northern dispersal on growth rate of young, size of young at birth, and litter size in Sigmodon hispidus. Amer. Midl. Natur. 84:510-520. Kress, C. J. 1966. Demographic changes in fluctuating populations of Microtus californicus. Ecol. Monog. 36:329-373. Lack, D. 1954. The natural regulation on animal numbers. Oxford University Press, Oxford. 343 pp. 16 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Leste, P. H., Perry, J. S., Watson, J. S. 1945. The determination of the median body-weight at which female rats reach maturity. Proc. Zool. Soc. London 115:473-488. Lewontin, R. C. 1965. Selection for colonizing ability, p. 77-94. In The genetics of colonizing species, Academic Press, New York. Lorp, J., Jr. 1960. Litter size and latitude in North American mammals. Amer. Midl. Natur. 64:488-499. Mayr, E. 1963. Animal species and evolution. Belknap Press, Cambridge, Mass. 797 pp. Necus, N. C., Bercer, P. J. 1977. Experimental triggering of reproduction in a natural population of Microtus montanus. Science 196:1230-1231. Opum, E. P. 1955. An eleven year history of a Sigmodon population. J. Mamm. 36:368-378. RANDOLPH, P. A., RANDOLPH, J. C., Matrincty, K., Foster, M. M. 1977. Energy costs of reproduction in the cotton rat, Sigmodon hispidus. Ecology 58:31-45. Rivker, C. G. 1942. An extension of the range of the Texas cotton rat in Kansas. J. Mamm. 23:439. Rose, R. K. 1974. Reproductive, genetic, and behavioral changes in popula- tions of the prairie vole (Microtus ochrogaster) in eastern Kansas. Ph.D. thesis, Dept. Syst. Ecol., Univ. Kansas. SoxaL, R. R., Rontr, F. J. 1969. Biometry. W. H. Freeman and Co., San F rancisco, xiii + 766 pp. SPENCER, A. W., STEINHOFF, J. W. 1968. An explanation of geographic variation in litter size. J. Mamm. 49:281-286. Wiecert, R. G. 1972. Avian versus mammalian predation on a population of cotton rats. J. Wildl. Manage. 36: 1322-1327. U nJ } MUS. GOM®, ZOOL. 2122 LISRARY OCCASIONAL PAPERS OCT 23 1978 of the UNIVERSITY MUSEUM OF NATURAL HISTORY The University of Kansas Lawrence, Kansas NUMBER 75, PAGES 1-20 OCTOBER 3, 1978 AN EARLY MIOCENE (ARIKAREEAN) FAUNA FROM NORTHCENTRAL FLORIDA (THE SB-1A LOCAL FAUNA) By Davip FRAILEY The history of terrestrial vertebrates on the Florida peninsula after its presumed emergence in the Oligocene (White, 1942) is recorded in beach strand-line deposits and scattered karst fillings in the widespread marine limestones (see Olsen, 1965, 1968, for a review of Tertiary localities). The fragmentation of specimens inherent in beach deposits and the rarity of known bone-bearing sinkholes, either through their absence or through lack of discovery, has created a situation in which a few small localities, often with poorly preserved specimens, achieve great importance in any recon- struction of mid-Tertiary faunal diversity in Florida. For example, the Late Oligocene-Early Miocene North American Land Mammal Age, the Arikareean, is presently represented in Florida by only three local faunas: Brooksville (Patton, 1967a), Franklin Phosphate Pit No. 2 (Simpson, 1930), and SB-1A, the subject of this paper. Of these three, SB-1A has yielded the most diverse fauna and the greater number of taxonomically useful specimens. Each of the three known Arikareean local faunas of Florida presents a slightly different aspect of faunal diversity in Florida in that each contains elements not found in the other three. SB-1A is the most unusual in this respect, and the most indicative of our present state of knowledge on this subject, in that none of its faunal members (with the possible exception of the higher taxon Anchi- theriinae) are found in any of the other two local faunas; all are new additions to the Arikareean record of Florida. *Museum of Natural History and Department of Systematics and Ecology, University of Kansas, Lawrence, KS 66045; and, Research Associate, Timber- lane Research Organization, Route 2, Box 212B, Lake Wales, FL 33853. 2 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY The SB-1A Local Fauna was first discovered by Mr. Wesley Hunt during a routine survey of quarries in North Florida in July, 1972, by the Timberlane Research Organization. A small amount of material was collected at this time but its significance was over- looked. Shortly thereafter, the site was rediscovered and brought to the attention of personnel in the Florida State Museum who re- ferred to it as the Live Oak Locality. Both groups made sporadic superficial collections until December, 1974, when Mr. John S. Waldrop received word that the area was being planned for use as a dump for compacted garbage. During December, 1974, and March, 1975, with intermittent trips in between, Mr. Waldrop and field crews from the Timberlane Research Organization removed, washed, and sorted the material that forms the basis of this study. Recent work has not added to the taxonomic diversity of the fauna nor added to the representation of the rarer elements and justifies a systematic description of the fauna at this time. The SB-1A Local Fauna occurs in an unstratified conglomeratic sequence above the Suwannee Limestone (Oligocene). The locality was exposed by limestone mining operations of Florida Rock Prod- ucts, Inc., Shands and Baker Division, from which the S and B of the faunal name is derived. The site is located about one mile north of the small town of Live Oak, Suwannee County, Florida. Tooth nomenclature follows that of Szalay (1969) and Patton and Taylor (1971). All measurements are in millimeters. “( )” indicates an approximate measurement. The following institutional abbreviations are used: AMNH, American Museum of Natural History; CM, Carnegie Museum; F:AM, Frick American Mammals, American ether of Natural History, TRO, Timberlane Research Organization, Rt. 2, Box 212B, Lake Wales, FL 33853; UF, Uni- versity of Floridaj UOMNH, University of OrgePns Museum of Natural History. ACKNOWLEDGEMENTS I am indebted to Mr. John S. Waldrop, Coordinator of Research, Timberlane Research Organization, for permission to study and report on this material. Individuals who have aided in the collection of this material include: Fred Barrett, Fred Helsel, Wesley Hunt, John Iskra, Greg Lail, Steve Martin, Emroy Powell, Aubie Rhoden, and Dale Wyatt. Washing and sorting of some of the matrix was done by John Beaudua. I would also like to thank colleagues of mine who aided me with various identifications and discussions during the course of this project: Dr. R. H. Tedford and Mssrs, B. Taylor and H. Galiano of the American Museum of Natural History; Dr. R. M. Hunt of the Nebraska State Museum; Dr. L. D. Martin of the University of Kansas Museum of Natural History, and Mr. Waldrop. Figures 1 A-B, H-I; and 4 were drawn by Mr. Ray EARLY MIOCENE FAUNA — NORTHCENTRAL FLORIDA 3 Gooris. Ms. Deb Bennett assisted with stippling and layout. A special note of gratitude is extended to the officials of Florida Rock Products, Inc., for permission to collect on their properties. GEOLOGY The following information was provided by Mr. John S. Waldrop. Site Stratigraphy—The general stratigraphic section in the pit area is 20 to 30 feet of exposed Suwannee Limestone unconformably overlain by a conglomeratic sequence that is also 20 to 30 feet thick and capped by a reddish clay deposit or residuum that ranges from 0 to 7 feet in thickness, excluding sinkhole fillings. The conglomer- atic sequence begins with a basal bed of reworked Suwannee Lime- stone boulders grading into an unstratified, poorly sorted, yellowish- white limestone conglomerate. The contact between the underlying Suwannee Limestone and the bounder bed is an undulating erosional surface. Three periods of activity, here referred to as Zones, are shown at the fossil site itself, as the main bone quarry rests on a second boulder bed consisting of boulders 3 to 4 feet in diameter and is overlain by a third boulder bed. The three conglomeratic zones are alternating, irregularly developed, unstratified beds with cobble-sized particles dominant in one area and _ pebble-sized particles dominant in another. There does not seem to be a pattern to the dispersion of these particles in the sequence, as they appear to alternate randomly within the plane of the bed. The thickness of the zones of the sequence is highly variable. Zone 1, the lowest bed, was the most nearly uniform bed, ranging in thickness from 8 to 10 feet. Zone 2, the middle bed, was observed to vary between 2 and 6 feet; and Zone 3, the upper bed, varied between 3 and 15 feet in the pit area. At the fossil site, Zone 1 is 8 feet thick, Zone 2 is 4 to 5 feet thick, and Zone 3 is 15 feet thick. The vertebrate fossils occur in the finer matrix of the unstratified, poorly sorted, yellowish-white limestone conglomerate in Zone 2. This finer matrix consists of a white to yellowish-white argillaceous calcarenite with blebs and stringers of almost pure green clay with white clay surrounding many of the larger particles. During quarry operations in March, 1975, a small lens of clay was encountered in the lower part of the quarry. This lens was about 2 feet thick and extended laterally for about 4 feet. The conglomeratic particles are reworked Suwannee Limestone, some of which are virtually un- weathered (although strongly recrystallized as is the Suwannee Limestone in this area) and some of which are weathered to a limestone saprolite. The white clay is probably derived from this saprolite as it generally occurs within crevices in and surrounds these saprolitic particles for 1 to 6 inches. The conglomeratic fraction ranges from pebble-sized to boulder-sized particles with the pebble and cobble sized particles being most numerous. The 4 OCCASIONAL. PAPERS MUSEUM OF NATURAL HISTORY more strongly recrystallized particles are well rounded but some are angular or subangular and nearly all are roughly spheroidal in shape, although many show concavities due to erosion. The bone is, overall, very well permineralized and preserved, although often badly broken by expansion and contraction of the clay inside and around the bone. A few bones are badly weathered and crumbly, however, and some were not collectable. Generally, the bone in the white clay either shows the least permineralization or greatest post-depositional weathering. Numerous well perminer- alized angular bone and tooth fragments occur throughout the site. There was little evidence of association of postcranial elements in the site and no cranial-postcranial association. In several instances associated groups of camel teeth were found in their natural sequence but without any trace of the jawbone. In other instances, the jawbones had been completely shattered but were in association with the teeth. This shattering appears to have occurred due to the expansion and contraction of the clay within the cavities in the jaws. This phenomenon would explain the extreme rarity of complete mandibles and the absence of the resistant skulls. - The main fossil quarry is located in Zone 2 although some bone fragments, possibly due to stratigraphic leakage, were found in Zone 1, and a few bone fragments, possibly reworked, occur in Zone 3. Ne‘ther Zone 1 nor Zone 3 were found to be rich enough to warrant quarry operations. The apparent absence or extreme scar- city of bone from the remainder of the conglomeratic sequence has not been satisfactorily explained. It should be pointed out, however, that due to the vertical and unstable nature of much of the pit wall, continuous direct observation could not be made in the interest of safety. Field observation resulted in no apparent reason for the concentration of bone in this one small area. Slumping and heavy vegetative growth have also hampered ‘research into this problem and the question is as yet unresolved. ° Although no well defined bedding planes were observed by close examination during excavation of the main bone quarry, the overall appearance of the conglomeratic sequence at a distance of one-half mile suggests that the three boulder beds are essentially horizontally bedded. Even though the zones thicken and thin and, as a result, the contacts between the basal boulder beds and the upper part of the underlying conglomeratic zones are wavy and undulating, no sharply dipping beds or angular disconformities were observed. Sinkhole development was not apparent within any of the zones. The horizontality of the beds was much more apparent at a distance, however, than close up, and no appearance of bedding was observed within any of the zones. The upper surface of the conglomeratic sequence is very irreg- ular and undulating, but it has not been determined whether this EARLY MIOCENE FAUNA — NORTHCENTRAL FLORIDA 5 was a naturally undulating erosional surface or was influenced by karst activity or perhaps by a combination of the two. The above mentioned horizontality of the beds would sugest that the dominant effect was one of erosion. The conglomeratic sequence is unconformably overlain by an orangeish-red and gray mottled clay or sandy clay representing what is thought to be a res'duum and soil profile, but is possibly a sedi- mentary deposit. This clay rests upon either the pebble or cobble conglomeratic material of Zone 3, depending upon the size particle that is dominant in that particular area. The unconformable nature of the contact between this upper clay and the upper part of the conglomeratic sequence, without the development of a weathering profile, might be one reason for considering the clay a sedimentary deposit or redeposited residuum. The absence of a weathering profile is puzzling as the geologic history of the area indicates a long period of subaerial exposure, a period of at least post-Hemphil- lian (Early Pliocene). Where the section along the pit wall has not been obscured or disturbed by mining activity, vegetative growth, or slumping, sinkholes are seen to be developed down into and through the conglomeratic sequence. These sinkholes are filled with a red or gray sandy clay or clay alone. The sinkholes near the main quarry that were studied in detail are filled with a blocky, orangeish-red and gray mottled clay or sandy clay. The overall ap- pearance is a reddish color but with much gray locally. As this clay seems to be the same as the surface veneer, the formation of, or at least collapse of, these sinkholes must have been subsequent to the time of deposition of the conglomeratic sequence. The depth of the observed sinkholes ranges from 1 to more than 50 feet. Mode of Formation and Significance of Deposit—The site would seem to be one of proximal deposition with little transport as evi- denced by the associated, but loose, teeth. Reworked Suwannee Limestone invertebrates are common in the site. These consist of casts of mollusks, echinoids and echinoid spines, and tubes or burrows of invertebrates. Younger invertebrates are totally absent. This is surprising due to the abundance of early to middle Miocene mollusks in deposits along the Suwannee River about 15 miles north- east of the site (Brooks, 1966). A similar yellowish clay about 10 miles north of the site contains numerous silicified specimens of Ostrea normalis associated with and stratigraphically below a mid- dle Miocene land vertebrate fauna. Except for some as yet uni- dentified frog and snake remains, the vertebrate fauna from SB-1A is strictly terrestrial. There is a definite absence of freshwater fishes, alligators, and freshwater turtles as well as the marine sharks and rays that are common in many other Miocene vertebrate faunas of Florida. The erosion and pitting on some bone and many of the teeth suggests that the bone lay in the open for a period of time before being covered. 6 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY The sediments, especially the cobbles and boulders, suggest considerable current energy, if transported by water. The poorly sorted sediments indicate rapid transport and deposition. The total absence of any marine or freshwater animals of similar age in association with the fauna suggests a means of transport other than water, although it is conceivable that the aqueous environment could have been present without these forms becoming entombed within the sediments. In a paper on the geological history of the Suwannee River, Brooks (1966) makes several statements that might suggest an answer to this seeming paradox. In speaking of the structures of an area approximately 15 miles to the northeast of SB-1A, Brooks discusses the stratigraphic displacement in the narrow belt between the karst plain and the high flatland physiographic entities. He also suggests the possibility that this displacement, which is commonly 60 or more feet, is due to faulting. The SB-1A Locality lies very near the boundary between the karst plain, which the site is located upon, and the high flatlands just to the north and northeast of the site. Displacement of 60 or more feet in the vicinity of SB-1A with-its attendant mass wasting and colluvial transport, could provide the conditions necessary for the rapid transport and deposition of poorly sorted sediments that contain particles ranging from clay to boulder size. This would also explain the extreme varibility in the dispersion of particle sizes and the unstratified nature of the sediments. The appearance in the fauna of only terrestrial vertebrates is also ex- plained as is the absence of marine or freshwater forms. If younger marine sediments were deposited in this area they have since been eroded, except perhaps for the upper clay veneer. It should be born in mind that definite proof for this hypothesis is lacking at present although it best explains this otherwise puzzling catastrophe of nature. Younger sediments occur at similar elevations to the north and northeast. A similar widespread catastrophe of a similar age is also present in south-central Florida, again in the same geologic setting. If this hypothesis can be proven, then it indicates that uplift and erosion of the Ocala Arch continued into the late Early Miocene and is perhaps the best direct evidence for uplift and erosion of the Ocala Arch. SYSTEMATIC ACCOUNTS OrpeR CARNIVORA Bowdich, 1821 Famity AMPHICYONIDAE Trouessart, 1885 Genus Mammacyon Loomis, 1936 Mammacyon cf. obtusidens Loomis, 1936 Figs. 1A-G Discussion.—F amilial rank follows Hunt (1972). The M* (TRO 390) of a large carnivore in the SB-1A Local Fauna is immediately EARLY MIOCENE FAUNA — NORTHCENTRAL FLORIDA 74 referable to either Mammacyon or Temnocyon on the characteristic prominence of three cusps, the paracone, metacone, and the cen- trally located protocone, and the anteroposterior expansion of the tooth at the protocone. The latter characteristic gives the tooth a rough “figure 8” outline in occlusal view. The very large size of H KISS Fic. 1—A-G. Mammacyon cf. obtusidens. A. M', TRO 390, labial view; B. M', TRO 390, occlusal view; C. Lower canine, TRO 388, labial view; D. Medial phalanx, TRO 387, lateral view; E. Medial phalanx, TRO 387, dorsal view; F. Proximal phalanx, TRO 386, lateral view; G. Proximal phalanx, TRO 386, dorsal view. H-I. Paroligobunis frazieri n. sp., UF 23928, holotype. H. Occlusal view of dentition; I. Labial view of ramus. 8 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY the protocone and the wide expansion of the tooth at the protocone and a heavy lingual cingulum are more like Mammacyon obtusidens and the referral is made on that basis. The size of the M* (17.7 x 24.4, length x width) is close to that of Mammacyon obtusidens although the relative expansion of the protocone area (expressed as a ratio of labial length/labial width across the protocone, Hunt, 1971) is greater than in M. obtusidens (1.1 vs. 1.32-1.36). This greater expansion of the tooth in this area may be no more than an extreme of individual variation or may later prove to be a species distinction. The uncertainty on this point prevents certain referral of this specimen to M. obtusidens.? A right lower canine (TRO 388, Fig. 1C), presumably of Mammacyon cf. obtusidens because of its size, has a long, flat wear facet from its tip to the base of the crown. The advanced wear of this tooth contrasts with the slight wear on the M! and indicates the presence of more than one individual in the deposit. A proximal and a medial phalanx from SB-1A (TRO 386; 387; Figs. 1D-G) are attributed to Mammacyon cf. obtusidens. In size and proportions they are similar to those of Daphoenodon superbus as described by Peterson (1910). Mammacyon obtusidens occurs in Arikareean beds (Monroe Creek and possibly Harrison formations) of South Dakota and Nebraska with a single specimen recorded from the Upper John Day beds ( Hunt, 1971). This referral indicates a Monroe Creek or possibly Harrison age equivalence for the SB-1A Local Fauna. SUBFAMILY CANINAE Gill, 1872 ?Mesocyon Scott, 1890 Discussion.—A medium-sized canid is represented in the fauna by a single, badly damaged P! (TRO 391) which is missing the protocone. This tooth is questionably referred to Mesocyon because it lacks a large parastyle, as do the P's of Mesocyon, and because Mesocyon is the only canid of this size which occurs in the Arikare- ean, the age indicated by the other faunal members. Tomarctus and Cynodesmus are of this size range but have a large parastyle on P* and occur in Hemingford and later times. Phlaocyon sp. Matthew, 1899 Figs. 2A-B Discussion.—The P', TRO 392, from SB-1A is smaller than the P' on the holotype of P. leucosteus (AMNH 8768) and has a rela- tively smaller hypocone. A small, accessory cusp is present on the 2 Dr, R. M. Hunt is currently studying the taxonomy and relationships of Temnocyon and Mammacyon. He has generously allowed me to draw upon his work in progress but a more complete discussion of these genera must await his paper. EARLY MIOCENE FAUNA — NORTHCENTRAL FLORIDA £ Fic. 2.—A-B. Phlaocyon sp., right P*, TRO 392. A. Occlusal view; B. Lingual view. C-D. Protosciurus sp., right ramus with lower incisor and P,, TRO 401. C. Occlusal view of Ps; D. Labial view of ramus. anterior cingulum of the SB-1A P* which is absent on that of Phlaocyon leucosteus. Phlaocyon marslandensis (McGrew, 1941) is even larger than P. leucosteus but shares the features of P* with P. leucosteus which contrast with the P! from SB-1A. The Pt of Phlaocyon found at SB-1A could be within the limits of variation for P. leucosteus. On the other hand, the species of Phlaocyon from SB-1A may be new but certain identification is not possible with the material at hand. 10 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Measurements of the P! from SB-1A and that of P. leucosteus are compared in Table 1. TABLE 1.—Comparative measurements (mm) of P4 of Phlaocyon sp. from SB-1A and of Phlaocyon leucosteus (AMNH 8768, holotype ). Character P. leucosteus Phlaocyon sp. greatest length 10.8 ay greatest width en 6.8 ( protocone across paracone ) width (3 : 6.3 (in occlusal position ) Famity MUSTELIDAE Swainson, 1835 Genus Paroligobunis Peterson, 1906 Paroligobunis frazieri, new species Fig. 1H-I Etymology.—Named for Michael Frazier, whose skill in prep- aration saved this specimen and in honor of his work in Florida paleontology. Holotype.—UF 23928, right ramus with C,, Ps-M,; P; and M2 alveoli. Diagnosis —Ramus more slender than in P. simplicidens or P. petersoni. Premolars less crowded than in either species. Paroligo- bunis frazieri is larger than P. petersoni but slightly smaller than P. simplicidens. Description and Comparisons—The shortened M, with a strong labial convexity is characteristic of Oligobunis and Paroligobunis and serves to separate these genera from other mustelids of this size range. The ramus from SB-1A is further referrable to Paroligobunis (using CM 1553, holotype of P. simplicidens) and not Oligobunis (using AMNH 6903, holotype of O. crassivultes) on dental features. Paroligobunis, and the SB-1A ramus, has a smaller M, than Oligo- bunis; a larger P, (P, of Oligobunis is reduced to a simple, insignifi- cant peg); no diastemata between the premolars (small diastemata are present between P,-P» and Ps-P; of Oligobunis); and in not having an entoconid on M, as does Oligobunis. Paroligobunis simplicidens (Peterson, 1910) and P. petersoni (Loomis, 1932) are the only species of this genus currently recog- nized as valid. Paroligobunis frazieri differs from P. simplicidens in being slightly smaller (see Table 2); in having a more slender ramus (in depth and width); the premolars of P. frazieri are less massive than those of P. simplicidens; P. and P, are less crowded (Pz. in P. simplicidens has been crowded in the tooth series to the point that it is rotated, labially, about 30°); Ps and P; of Paroligobunis frazieri EARLY MIOCENE FAUNA — NORTHCENTRAL FLORIDA 11 are not broadly widened, posteriorly, as are those of P. simplicidens; and Ps and P; of P. frazieri have a more distinct posterior cingulum than do these premolars in P. simplicidens. The apparent less crowd- ing of the premolars in P. frazieri may be an artifact of the recon- structed mandible but it appears to be natural. The rotation of P, in P. simplicidens places this tooth parallel to the sagittal section of the animal. In P. frazieri, P» is in line with the tooth row. Paroligobunis petersoni, according to Loomis (1932), is much smaller than either P. simplicidens and hence also of P. frazieri. Paroligobunis petersoni also has a massive mandible in agreement with P. simplicidens and in contrast to the light ramus of P. frazieri. Discussion.—Paroligobunis simplicidens occurs in the Harrison Formation (late Arikareean) of Nebraska (Peterson, 1910). Paroligobunis petersoni was discovered in beds of “Upper Harri- son” age (=Hemingfordian, Schultz and Falkenbach, 1968) near Van Tassel, Wyoming. These two species are probably more closely related to each other than either is to P. frazieri as a massive mandible, present in both species, is very likely a derived character in this genus. Oligobunis and Paroligobunis are undoubtedly closely related. Oligobunis is known from the Thomas Farm Local Fauna (Late Hemingfordian) of Florida (Oligobunis floridanus). The relation- ship between O. floridanus and P. frazieri is not readily apparent but is probably no closer than for other species in each genus. TABLE 2.—Comparative measurements (mm) of Paroligobunis frazieri, new species, and P. simplicidens. P. frazieri P. simplicidens Character (UF 23928 ) (CM 1553, holotype ) Length, P.-M2 49.0 54.2 ~C, (length) 9.5 11.2 (width) 753 8.2 P. (length) 7.4 8.8 ( width ) 4.8 sare P; (length) 8.1 9.7 ( width) 5.6 6.9 P, (length) 10.8 11.6 (width ) 6.1 7.0 M; (length) 15.7 (16.4) ( width ) Fal (7.6) Depth of mandible 21.8 25.4 (below M;) Width of mandible 10.2 13.0 (at M:) 12 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY OrpeR CARNIVORA Gen. et sp. indet. Fig. 3A-B Discussion—An isolated metatarsal II (TRO 402) of an uni- dentified carnivore is illustrated in Fig. 3A-B. The features of the articular surfaces are obscured by an apparent osteopathology which renders taxonomic referral inconclusive. The size of the metatarsal is between that expected for Phlaocyon or Mesocyon, the two carnivores in the fauna which are nearest in size, and the size one would expect for an animal like Leptocyon or Nothocyon. The carnivore which possesed this metatarsal was digitigrade and the metatarsals were not closely appressed. The distal articu- lation is rounded with only a posterior keel and has much the appearance of a felid distal articulation. Fic. 3.—A-B. Carnivora gen. et sp. indet., right metatarsal II, pathologic, TRO 402. A. Dorsal view; B. Lateral view. C-D. Protosciurus sp., left tibia, TRO 400. C. Anterior view; D. Lateral view. EARLY MIOCENE FAUNA — NORTHCENTRAL FLORIDA 13 OrpvEeR PERISSODACTYLA Owen, 1848 FamiLy EQUIDAE Gray, 1821 SuspFAMILY ANCHITHERIINAE Osborn, 1910 Gen. et sp. indet. Discussion.—An anchitherine horse about the size of Parahippus leonensis is represented in the SB-1A Local Fauna by a single lateral metapodial (TRO 393). The scarcity of horse material from this deposit, and the evident rarity of horses in the vicinity during deposition, is an intriguing problem for which there is no satis- factory answer. In this respect SB-1A stands in sharp contrast to the Hemingfordian faunas in Florida, especially Thomas Farm, in which horses are abundantly represented. OrpdER ARTIODACTYLA Owen, 1848 SUBORDER TYLOPODA Illiger, 1811 Famity CAMELIDAE Gray, 1821 Genus Nothokemas White, 1947 Nothokemas waldropi, new species Fig. 4 Etymology.—Named for John S. Waldrop, who collected this material, in respect for his diligent stratigraphic paleontology. Holotype.—UF 23927, left ramus with C,, Ps-Ms. Diagnosis.—About 1/2 the size of N. floridanus and N. hidal- gensis, the only previously recognized species of Nothokemas. Description and Comparisons—Lower Dentition: The lower incisors are spatulate, thin, and form a closely set, fan-shaped group in which I, and I; each slightly overlaps the incisor in front. The lower canine tooth is long and strongly recurved. The first lower premolar is not present in this genus and its absence has been used as part of a generic description by Patton (1969). As can be seen on a radiograph of the holotype of N. waldropi, on file at the Timberlane Research Organization, the P; is truly absent and not merely suppressed. A long diastema separates C, from Ps. The dorsal border of the diastema drops quickly between C, and Py» apparently as a result of the missing P). P. is a simple, linear tooth. The central cusp is prominent. P» and P; are relatively thinner than these teeth in N. floridanus and N. hidalgensis. P; is elongate and slender. A lingual stylid is present which arises just posteriorly to the protoconid and which may extend to the rear margin of P,; or less than half that distance depending upon the individual. P, is typically camelid in appearance with its widest part at the hypoconid giving the characteristic wedge-shape to the tooth. 14 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY — j oS 3 C §) Fic. 4.—Nothokemas waldropi n. sp. A. Right P'-M* with partial maxilla, TRO 389, labial view; B. TRO 389, occlusal view; C. UF 23927, holotype, left ramus with C,, Pe-Ms, occlusal view of dentition; D. UF 23927, holotype, labial view. The lower molars have the discontinuous and overlapping crests which are typical of oxydactyline camels in which the posterior crest unites first with the anterior crescent, and not the anterior crest, during wear. Intercolumnar styles are generally absent, but slight styles are present on molars of some individuals and are presumed to be only another example of individual variation in this species. Internal ribs are faintly discernible on the crests. A small extension of the posterior crest (entoconid) which is present on the M; of two individuals (TRO 357, 360) is suggestive of the “double-enamel loop” of Floridatragulus (Patton, 1969). The posterior terminus of the posterior crest of M; in N. waldropi is apparently also individually variable. ; Upper dentition: I’? are unknown; I* is large, curved, and labially-lingually flattened. Small blades are present on the anterior and posterior margins. The diastema between I? and I? was small, approximately 2 mm in length. C! is curved, caniniform, and larger than I* or P’. P! is a simple, short-bladed tooth as is also seen in Oxydactylus. P! is double-rooted but the roots are fully fused. P* is an elongate, tricuspid tooth which has a simple linear enamel pattern. The protocone is located anterior to the center of the tooth and is the most prominent cusp. Other than the P? of Floridatragulus, the P? of Nothokemas waldropi is the longest, rela- tive to P*, among all the genera examined. The length of P? approximates that of P*, P*, like P*, appears to be elongated antero-posteriorly. This effect is produced in part by the small size of the internal cingulum. EARLY MIOCENE FAUNA — NORTHCENTRAL FLORIDA 15 This cingulum does, however, have two distinct cuspules which are located above the medial root near the protocone. The overall appearance is like the P* of Oxydactylus longipes. P# is simple, camelid-like in shape and pattern and undiagnostic except by its small size. Only a very slight cingulum is present on the posterior part of the crescent on a few individuals while totally absent in others. The upper molars are characteristically oxydactyline in their low crown height, prominent ribs and styles (particularly the meso- style), and the non-linear overlap of the anterior crest (paracone) by the posterior crest (metacone). The external ribs are only slightly less prominent than the mesostyle (as in all oxydactyline genera except Miolabis). Intercolumnar styles are almost totally lacking although occasionally found on M?. These styles are more frequently distributed among other genera of camels with the oxydactyline molar pattern although subject to some degree of individual variation in Miolabis and Oxydactylus. The anterior crescent may bifurcate posteriorly, again determined by individual variation, reminiscent of the bifurcation seen in xiphodonts and in some individuals of Poebrotherium. Dental measurements of Nothokemas waldropi are given in Table 3. TABLE 3.—Measurements (mm) of the teeth of Nothokemas waldropi UF 23927 Character (holotype ) TRO 360 TRO 389 Depth of ramus beneath M; 18.5 14.5 — Depth of ramus beneath P, 18.3 13's — Po-Mz 59.9 56.3 — P3-Px 16.9 15.4 — P:-Mz 52.9 49.8 — M:-Me 36.2 ; 34.4 — - Ps (Lx W) Soo Tak (2.8 — Py 8.6x 4.5 8.3x 4.1 — Mi 9.7x 6.7 95x 6.4 — Me 11.0x 7.9 10.6 x 7.0 — Ms UGPsx T2383 14.6x 6.9 — C'-P' diastema — — 13.4 P'-P? diastema — — 11.9 P?-M° — — 58.8 E(LxW) — — 4.7x 2.2 [ee — — Sl x 35 12k = — 8.8x 4.9 iP — Usa TAS Udales fas) M* — 10.8 x 9.7 10.9 x 10.2 M? — 11.6 x 12.0 Tease ETL M® —_ Mas a, 12.6x 11.9 16 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Discussion —The genus Nothokemas was first recognized by White (1947) who placed it in a new family, the Nothokemadidae, incertae sedis, within the Hypertraguloidea. Patton (1969) rein- terpreted the hypertraguloid features of Nothokemas as _plesio- morphic for artiodactyls and transferred Nothokemas to the Family Camelidae, Subfamily Aepycamelinae Webb, 1965. The Aepycame- linae is probably a paraphyletic group and at present contains such genera as Oxydactylus, Aepycamelus, Paratylopus, and Miolabis (Webb, 1965; Simpson, 1945). Nothokemas is undoubtedly similar to Oxydactylus and could logically be placed within the Aepycamelinae, but, because of the questionably validity of that subfamily, I have chosen. instead to refer to Nothokemas as an “Oxydactylus-like camelid” in reference to the characteristic molar pattern shared by these two genera. Other Oxydactylus-like camelids are Miolabis, Gentilicamelus, and Floridatragulus, but decidedly not Paratylopus. Paratylopus is a slightly advanced poebrothere. Aepycamelus is probably closely related to Oxydactylus but has modified the Oxydactylus molar pat- tern through hypsodonty, a derived condition possibly indicative of relationship with Procamelus and later camels. The inclusion of Floridatragulus in this group appears at first to be inappropriate in view of the striking features of this genus which have led to its placement in the monotypic Subfamily Floridatragulinae Maglio, 1966. Many of the characters of Floridatragulus, however, are prim- itive ruminant characters as recognized by Patton (1969) and are of no use in phylogenetic grouping. The most unusual features are the greatly elongated muzzle and the double enamel loop or divided hypoconid on M;. The elongated muzzle is certainly autapomorphic but the double enamel loop may not be. The resemblance of the small posterior extension of the posterior crest as seen on some individuals of Nothokemas waldropi to the lingual half of the double enamel loop of Floridatragulus suggests an hypothesis for the origin of this characteristic feature of Floridatragulus. A further extension of the posterior crest (the entoconid) of Ms; onto the talonid would produce a typical double enamel loop. A similar degree of variation in the posterior limit of the posterior crest of Mz; is seen in some individuals of Miolabis (F:AM 68985, 68988) from Trinity River, Texas (Barstovian), and supports the contention that the double enamel loop of Floridatragulus is not as exceptional as White (1940, 1947) and Patton (1967b, 1969) believed. These enamel loops on the talonid of Ms, would then not be homologous with those of hypertragulids in which both crests of the talonid of M, probably arose from two small crests which run anteriorly from the single cusp on the talonid, the hypoconulid, as can be seen in Archaeomeryx and other early artiodactyls. Nothokemas was previously known from the Hemingfordian of EARLY MIOCENE FAUNA — NORTHCENTRAL FLORIDA 17 Florida and Texas (Patton, 1969). The occurrence of Nothokemas in the SB-1A fauna is the earliest occurrence of this characteristic Gulf Coast genus. OrDER RODENTIA Bowdich, 1821 SUBORDER SCIUROMORPHA Brandt, 1855 Famity SCIURIDAE Gray, 1821 TrisE SCIURINI Burmeister, 1854 Genus Protosciurus Black, 1963 Protosciurus sp. Figs. 2C-D, 3C-D Discussion.—A partial right ramus with the incisor and P,; (TRO 401) and a tibia (TRO 400) of a squirrel are the only indication of small mammals in the SB-1A fauna. The P, of the squirrel from SB-1A, unlike the Pys of most genera of squirrels, does not have a laterally compressed trigonid or rad- ically enlarged hypoconid and retains a basically unmodified tree squirrel tooth pattern as is seen in Protosciurus. The following description of the P, from SB-1A, using the terminology of Black (1963), is also applicable, except where noted, to the Pys of P. mengi and PProtosciurus jeffersoni (referral of Black, 1965). The P, of Protosciurus condoni has been figured by Black (1963) but it is damaged and little can be said about it. The Pys of the other species of Protosciurus are unknown. The trigonid of P, is composed of a clearly defined protoconid and metaconid which are well separated yet connected by a metal- ophid anterior to which is a trigonid basin and a small, knob-shaped anterior cingulum, the Pparaconid. The protoconid and metaconid are not well separated on the Py, of ?P. jeffersoni and an elongate anteroconid is present on the protoconid which may be equivalent to the knob on the anterior cingulum of P. mengi and the Proto- sciurus from SB-1A. The metaconid on the P, from SB-1A is strongly cone-shaped (as are the protoconid and hypoconid), whereas in P. mengi the metaconid is less clearly separable from the anterior rim of Py. The protoconid is placed slightly posteriorly to the metaconid. The protoconid and metaconid are not appressed* and are approximately of the same size. The hypoconid is slightly larger and creates a postero-labial swelling on the otherwise square to rectangular occlusal outline. In all other genera of squirrels, the * The P, on the holotype of P. condoni (UOMNH F-5171) is badly damaged but the protoconid and metaconid appeared to be closely appressed to Black (1963). On a referred specimen in the American Museum of Natural History, Protosciurus cf. condoni (F:AM 99254), the metaconid and protoconid of P, appear to be relatively closely placed due to the much larger size of the metaconid relative to the protoconid than is seen on the P,’s of either P. mengi or the squirrel from SB-1A. 18 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY relatively great size of the hypoconid, combined with the laterally compressed trigonid, produces a more triangularly shaped Py. The rim of the talonid is continuous between the hypoconid and a distinct entoconid. A mesoconid is present on a faint ectolophid. The mesostylid is represented by a slight rise on the posterior of the metaconid. The ramus also has features of Protosciurus as given by Black (1963) in his diagnosis of the genus: The massateric fossa ends anteriorly beneath M;; the diastema is short; the diastemal depres- sion is deep; and the mental foramen is positioned well below the depression of the diastema. Without including PP. jeffersoni, four species of Protosciurus have been described (Black, 1963). Listed in decreasing order of size, these are: Protosciurus condoni, P. tecuyensis, P. mengi, and P. rachelae. The P, of Protosciurus sp. from SB-1A is slightly smaller than that of P. mengi and differs in having a more distinct meta- conid. The P, from SB-1A measures 2.1 x 1.7 mm versus 2.5 x 2.2-2.5 for P. mengi (Black, 1963). The ramus from SB-1A has a small muscle scar anterior to the massateric fossa as does P. tecuyensis but unlike P. condoni (un- known in P. mengi and P. rachelae). The massateric fossa on the SB-1A ramus is deeper than on the species of Protosciurus for which this fossa is known (all except P. rachelae). A complete tibia (TRO 390) from the SB-1A locality is referred to Protosciurus sp. on the basis of its size and sciuromorph features. The significance of this specimen lies in its greater resemblance to tibiae of arboreal squirrels (Sciwrus, Tamiasciurus) than to those of ground squirrels (Tamias, Citellus, Spermophilus). Specimen TRO 390 is slender, relatively straight, the tibial crest terminates in the proximal one-fourth of the total length, and the distal articular grooves (especially the medial) are not well defined as in tree squirrels and in contrast to the sturdy, curved tibiae of ground squirrels in which the tibial crest reaches approximately one-third the length and the articular grooves are more sharply defined. These similarities support the conclusion of Black (1963), based on dental characters, that Protosciurus was a tree squirrel. Protosciurus occurs from the Chadronian (?Protosciurus jeffer- soni; Black, 1965) or Orellan to the Pre-Harrison Arikareean (Black, 1963). This occurrence of Protosciurus at SB-1A would appear to be one of its latest occurrences. AGE OF THE FAUNA The SB-1A Local Fauna contains genera which are found in Arikareean and Hemingfordian deposits outside of Florida and could conceivably be referred to either age. Phlaocyon (Matthew, 1901; Galbreath, 1953) and Nothokemas (Patton, 1969) are the most EARLY MIOCENE FAUNA — NORTHCENTRAL FLORIDA 19 apparent Hemingfordian forms. Paroligobunis is both an Arikareean (P. simplicidens) and Hemingfordian (P. petersoni) genus. Paroligo- bunis frazieri, however, is more like P. simplicidens in its features. Mammacyon is primarily an Arikareean genus; Mammacyon obtu- sidens occurs in the middle Arikareean (Monroe Creek Formation). Protosciurus has never been found later than the Middle Arikareean (Black, 1963). As a choice must be made on the reliability of the indicators, I think it more probable that the widespread and better known forms, Mammacyon and Protosciurus, are better age indica- tors and that the presence of Phlaocyon and Nothokemas in the SB-1A Local Fauna represent Arikareean occurrences of these genera. SUMMARY A small Arikareean fauna, the third known from Florida, is described. The fauna consists of Mammacyon cf. obtusidens, Phla- ocyon sp., PMesocyon, Paroligobunis frazieri n. sp., an indeterminate carnivore, an indeterminate anchitherine horse, Nothokemas wal- dropi n. sp., and Protosciurus sp. All the included taxa, with the exception of the higher taxa which represent indeterminate species, are new additions to the faunal record of Florida. Mammacyon, Protosciurus, and Paroligobunis are present in Arikareean faunas of the classic Great Plains sequence and allies the SB-1A Local Fauna with this North American Land Mammal Age. Mammacyon and Protosciurus may be conspecific with known forms but the small amount of material referable to these genera prevents a full taxo- nomic treatment at the species level. The presence of Phlaocyon in the SB-1A fauna is evidently an early occurrence of this more typ- ically Hemingfordian genus, and in some respects this specimen is more primitive than the later species. The new species of Nothoke- mas from SB-1A extends the temporal range of this characteristic Gulf Coast Hemingfordian camel back into the Arikareean. The SB-1A fauna as a whole is similar to the better-known Arikareean faunas of western North America with the added dimension of regional differentiation as exhibited in new species of Paroligobunis and the endemic Gulf Coast genus Nothokemas. Further distinc- tions among the rarer taxa, which are now conservatively attributed to individual or populational variation, may become apparent as the fossil record of Florida becomes better known. The fauna is from unsorted terrestrial outwash sediments over- lying the Suwannee Limestone (Middle Oligocene) which may have been associated with faulting caused by the uplift of the Ocala Arch in northcentral Florida. LITERATURE CITED Biack, C. C. 1963. A review of the North American Tertiary Sciuridae. Bull. Mus. Comp. Zool., 130 (3): 111-248. 20 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Brack, C. C. 1965. Fossil rodents from Montana. Part 2. Rodents from the Early Oligocene Pipestone Springs Local Fauna. Ann. Carnegie Mus., 38 (2): 1-48. Brooks, H. K. 1966. Geological history of the Suwanee River. In: Geology of the Miocene and Pliocene series in the north Florida-south Georgia area, N. K. Olsen (ed.), Atlantic Coastal Plain Geol. Assoc. and South- eastern Geol. Soc. Guidebook, Seventh Annual Field Conf., pp. 37-45. GALBREATH, E. C. 1953. A contribution to the Tertiary geology and paleon- tology of northeastern Colorado. Vertebrata; Univ. Kansas Publ., Art. 4: 1-120. Hunt, R. M. 1971. North American Amphicyonids (Mammalia; Carnivora). Unpubl. Ph.D. Diss., Columbia Univ., New York. Hunt, R. M. 1972. Miocene amphicyonids (Mammalia, Carnivora) from the Agate Springs Quarries, Sioux County, Nebraska. Amer. Mus. Nat. Hist. Novit, 2506: 39 pp. Loomis, F. B. 1932. The small carnivores of the Miocene. Am. Jour. Sci., 24: 316-329. MatrHew, W. D. 1901. Fossil mammals of the Tertiary of northeastern Colorado. Mem. Amer. Mus. Nat. Hist., Vol. I, Part VII: 353-447, 4 pl. McGrew, P. O. 1941. A new procyonid from the Miocene of Nebraska. Publ. Field Mus. Nat. Hist., Geol. Ser., 8: 33-36. OtseNn, S. J. 1965. Vertebrate fossil localities in Florida. Florida Geol. Surv., Spec. Publ. 12: 28 pp. OtseEn, S. J. 1968. Miocene vertebrates and north Florida shorelines. Paleo- geogr., Paleoclimatol., Paleoecol., 5: 127-134. Patron, T. H. 1967a. Oligocene and Miocene vertebrates from central Florida. In: Miocene-Pliocene Problems of Peninsular Florida, H. K. Brooks, E. C. Pirkle, R. C. Fountain (eds.), Guidebook, Southeastern Geol. Soc. Thirteenth Field Trip. pp. 3-10. Patron, T. H. 1967b. Revision of the selenodont artiodactyls from Thomas Farm. Quart. Jour. Florida Acad. Sci., 29 (3): 179-190. Patron, T. H. 1969. Miocene and Pliocene artiodactyls, Texas Gulf Coastal Plain. Bull. Florida State Mus., 14 (3 )s> 115-226. Patron, T. H., and B. E. Taytor. 1971. The Synthetoceratinae (Mammalia, Tylopoda, Protoceratidae). Bull. Amer. Mus. Nat., 145 (2): 123-218. PETERSON, O. A. 1910. Description of new carnivores from the Miocene of western Nebraska. Mem. Carnegie Mus., 4 (5): 205-278. * Scuuttz, C. B., and C. H. FaLKENBACH. 1968. The phylogeny of the oreo- donts, Parts 1 and 2. Bull. Amer. Mus. Nat. Hist., 139: 1-498. SIMPSON, G. G. 1930. Tertiary land mammals from Florida. Bull. Amer. Mus. Nat. Hist., 59 (3): 149-211. Smmpson, G. G. 1945. The principles of classification and a classification of mammals. Bull. Amer. Mus. Nat. Hist., 85: 350 pp. SZALAY, F . S. 1969. Mixodectidae, Microsyopidae, and the insectivore-primate transition. Bull. Amer. Mus. Nat. Hist., 140 (4): 195-330. 40 pl Wess, S. D. 1965. The osteology of Camel Man el Wom ee gy amelops. Bull. Los Angeles County Wuire, T. E. 1940 New Miocene vertel i England Zool. Club, 18: 31-38. uicmemner ci fame Wurre, T. E. 1942, Th ioc ‘lori Pel eee e lower Miocene fauna of Florida. Bull. Mus. Comp. Wuire, T. E. 1947. Additions to th Mi : i Mus: Come. 258i Bete aa iocene fauna of North Florida. Bull. ON!) MUS. Comp. $122 LIBRARY JAN 8 4979 HI OCCASIONAL PAPERS UNIVER Ey of the MUSEUM OF NATURAL HISTORY The University of Kansas Lawrence, Kansas NUMBER 76, PAGES 1-6 DECEMBER 29, 1978 A NEW SPECIES OF LIOLAEMUS (SAURIA: IGUANIDAE,) FROM THE ANDEAN MOUNTAINS OF THE SOUTHERN MENDOZA VOLCANIC REGION OF ARGENTINA By Josz M. Cer In the course of field research in South America, William E. Duellman and his associates obtained specimens of an undescribed iguanid lizard in the Paso El Choique area in southern Provincia de Mendoza, Argentina. A careful comparison of these lizards with other Argentine iguanids revealed that the new species is a member of the widespread, Andean-Patagonian genus Liolaemus; however, it is easily distinguished by a significant combination of several morphological characters from its congeners. In recognition of the discoverer of this new iguanid, it herewith is named after Dr. William E. Duellman, who has advanced so appreciably field and taxonomic studies on the herpetofauna of South America. Liolaemus duellmani, new species Figure 1 Holotype.——The University of Kansas Museum of Natural His- tory (KU) 161126, an adult male from Paso El Choique, 50 km SSW El Manzano, 2260 m, Provincia de Mendoza, Argentina (lati- tude 36°27’ S; longitude 60°50’ W), collected on 12 December 1974 by William E. Duellman and John E. Simmons. Paratopotypes—KU 161127-161128, adult and subadult male specimens collected with the holotype; Instituto Biologia Animal, 1Instituto Biologia Animal, Universidad Nacional de Cuyo, Mendoza, Argentina. 2 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Universidad Nacional de Cuyo (IBA-UNC) 139, an adult female, collected on 22 November 1961 by José M. Cei and Virgilio G. Roig. Diagnosis——Liolaemus duellmani can be distinguished from other Liolaemus by the following combination of characters: 1) slender body; 2) short legs; 3) subimbricate, almost juxtaposed polygonal dorsal scales; 4) small, juxtaposed, conical lateral scales; 5) wide, smooth ventral scales; 6) dorsally keeled caudal scales; 7) presence of antehumeral and neck folds; and 8) strongly clawed _ digits. Description—Head length about one-fifth body length (Fig. 1 and Table 1). Well-developed antehumeral fold present; neck folds distinct. Snout bluntly pointed.. Rostral scale large and wide, twice as wide as high; nasal not in contact with rostral, separated from it by two enlarged scales. Nostril equal in size to nasal scale, located posterolaterally at a point midway between eye and tip of snout. Ear opening elliptical, edged by small, regular-shaped scales becoming larger posteriorly. Fic. 1—Dorsal (upper) and ventral (lower) views of holotype of Liolaemus duellmani, an adult male (KU 161126); total length, 170 mm. Temporal scales smooth, irregularly shaped and convex. Scales of frontal, parietal, and occipital regions large and slightly convex. Interparietal scale small, nearly triangular, and bordered laterally by a pair of large, irregularly shaped parietal scales meeting pos- terior to interparietal. Largest supraoculars smaller than scales of interorbital region. Subocular scale enlarged, undivided, separated from supralabials by single row of slightly rugose, pointed scales. Supralabials 10; infralabials 7. Mental trapezoidal. Eyelids slightly fringed. Pterygoid teeth present. Scales of lateral neck granulated, A NEW SPECIES OF LIOLAEMUS (SAURIA: IGUANIDAE) 3 rounded, and smaller than subimbricate dorsal neck scales. Scales across gular region between tympana 48-50. Vertebral scale row absent. Dorsal scales heterogeneous, small, polygonal but almost rounded, faintly keeled or-smooth, faintly subimbricate, and almost juxtaposed. Granular scales present between large scales. Scales on sides smaller, rather conical, and juxtaposed. Ventral scales larger than dorsal scales, wide and smooth, and decreasing in size in gular region. Upper caudal scales large, square, and distinctly keeled. Ventral caudal scales subtriangular and smooth. Scales around middle of body 86-90. Scales of dorsal surface of forelimb large, imbricate, and slightly keeled; ventral forelimb scales granular and smooth. Dorsal thigh scales large, imbricate, and smooth; ventral thigh scales small and nearly granular. Dorsal tibia scales heterogeneous, subimbricate, and juxtaposed; ventral tibia scales large, imbricate, and smooth. Four indistinct secretory pores present anterior to vent. Subdigital lamellae of fourth finger 18, faintly keeled. Subdigital lamellae of fourth toe 20-21, tricarinate. Claws of all digits sharp, black, and 34 mm long. Hind limb rather short and stout; when adpressed, fourth toe not reaching axilla. When forelimb adpressed, fourth finger reach- ing middle of body. Coloration: Dorsum pale gray becoming reddish tan laterally, with dark brown transverse markings enclosing bluish white flecks. Dorsum of tail pinkish tan with dark brown bars. Throat pale gray mottled with black. In smaller individual, belly colored like throat; in others, belly black. Ventral surfaces of forearm and hand, femoral region, and foot bright yellow. Iris reddish brown. Tongue pink. Lining of throat gray. Coloration of preserved female like males except darker. Distribution.—This species is known only from the type locality. Remarks.—Liolaemus duellmani is found in a xeric, montane TABLE 1.—Measurements (mm) of Liolaemus duellmani. Holotype Paratopotypes CHARACTER SS KU KU KU IBA-UNC 161126, f 161127, ¢0 161128, ¢ 139, @ Total Length 170.0 —" —" — Snout-vent Length 80.0 83.0 60.0 82.0 Head Length 18.5 20.0 20.0 19.2 Head Width 14.6 15.5 12.0 14.6 Forelimb Length 24.5 25.0 21.0 28.0 Hind limb Length 40.5 42..0° 32.5° 42..0° Axilla-groin Length 42.0 42.0 31.0 42..0 * Tail short, regenerated. > Tail broken. ¢ When adpressed, hind limb reaches axilla. 4 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY habitat characterized by basaltic outcrops, rocky slopes and sandy soil. Vegetation consists of clumps of plants—bunch grass, cacti, Ephedra, and spiny legumes. The female contained two mature eggs, each 21 mm in diameter. The size and number of eggs suggest that this species may be oviparous. DIscussION Liolaemus duellmani cannot be regarded as a member of any presently recognized group of Andean-Patagonian Liolaemus. It differs from the psammophilous and burrowing forms of the fitzingeri complex by the lack of femoral patches, reduced number of preanal pores, and the number and shape of dorsal scales. Furthermore, the dorsal scales are distinctly imbricate and keeled in the fitzingeri species-group. Likewise, L. duellmani is distinct from the other species of Patagonian lizards belong to the kingi- archeforus, megallanicus, and bibroni-gracilis species groups. These groups are characterized by imbricate, keeled, and mucronate dorsal scales, and obviously different color patterns (Cei 1973, 1975a, 1975b). Liolaemus duellmani cannot be allocated to any presently recognized species-group of Chilean Liolaemus (Donoso Barros 1966; Peters and Donoso Barros 1970), nor referred to L. fitzgeraldi (Boulenger 1899), L. robertmertensi (Hellmich 1964), or L. dorbigny (Koslowsky 1898). It differs from the latter three spe- cies in characteristics of its dorsal and ventral scales, and color- ation. The species also can be distinguished from L. elongatus, L. austromendocinus, and L. buergeri, all of which live in the same region, on the bases of their imbricate, mucronate and keeled dorsal scales, larger tails, and distinctively different dorsal color- ations (Cei 1974). Although Liolaemus kriegi and L. ceii (Miller and Hellmich 1939; Donoso Barros 1971) are similar to L. duellmani in having heterogeneous lepidosis and granular scales between the larger scales of the dorsum, both species have many more scales at mid- body (97-115), a dark pileus, extremely fat base of the tail, slightly keeled dorsal scales, nasal scale in contact with rostral, and the fourth toe with more subdigital lamellae (28-30). Liolaemus ruibali, which inhabits the Uspallata mountains in the north of Mendoza Province (Donoso Barros 1961), seems the most closely related to L. duellmani. Its scales are heterogeneous, subimbricate, and almost smooth. The nasal is not in contact with the rostral, and it shares the same number of supralabials, infra- labials and subdigital lamellae with L. duellmani. Liolaemus ruibali differs from L. duellmani by its smaller size, shorter tail, and fewer scales at midbody (63-80). The lateral scales are less heterogeneous, and the color pattern is distinct. A NEW SPECIES OF LIOLAEMUS (SAURIA: IGUANIDAE) 5 There seem to be some convergent morphological features be- tween L. duellmani and northern Andean Liolaemus groups such as the signifer and multiformis groups, as well as with lizards of the genus Ctenoblepharis.2. These groups share with L. duellmani granular scales between the large scales, the heterogeneous lepi- dosis, and the nearly juxtaposed dorsal scales. ACKNOWLEDGMENTS I am grateful to The University of Kansas Museum of Natural History and to Dr. William E. Duellman, Curator, Division of Herpetology, for hospitality extended to me during my recent visit there. Dr. Duellman kindly allowed me to examine specimens and provided field notes on the ecology and coloration of this species. RESUMEN Se describe una nueva especie argentina de iguanido tropidurino, Liolaemus duellmani, de las regiones rocosas volcanicas de la Cordillera Andina, en el Sur de la Provincia de Mendoza, Paso El Choique, a 2000-2400 m de altura. La nueva forma, que se difer- encia netamente de todas las especies conocidas del género, es larga hasta 170 mm, y se caracteriza por la lepidosis heteronota, las escamas dorsales casi lisas, imbricadas y casi yuxtapuestas, las escamas ventrales mucho mas grandes que las dorsales, un numero relativamente alto de escamas alrededor del cuerpo (85-90) y una coloracién de bandas transversales oscuras, escotadas posterior- mente, sobre un fondo grisaceo rojizo, jaspeadas de puntos blancos azulados. La forma que mas se le aproxima es Liolaemus ruibali de los Andes de Uspallata, Mendoza, la que todavia difiere de L. duellmani por ser mucho mas pequena, por tener cola mas corta y un numero inferior de escamas alrededor del cuerpo, escamas laterales mas uniformes y un patron de coloracién muy diferente. Se ponen en evidencia algunas convergencias morfoldgicas entre la nueva forma y las especies del género Ctenoblepharis. LITERATURE CITED Bouencer, G. A. 1899. The highest Andes. In Fitzgerald. Rep. III. Methuen Co., London pp. 354-356. Camp, C. L. 1923. Classification of the lizards. Bull. Am. Mus. Nat. Hist. 48 :289-481. Cer, J. M. 1973. Herpetologia Patagénica. VII. Notas ecolégicas y mor- folégicas sobre Liolaemus bibroni y L. boulengeri. Physis 32(85): 459-469. Crt, J. M. 1974. Revision of the Patagonian iguanids of the Liolaemus elongatus complex. J. Herp. 8(3):219-229. *In spite of Donoso Barros’ statement (1971) to the contrary, the presence or absence of pterygoid teeth is not a significant character state to differentiate the genera Liolaemus and Ctenoblepharis (Camp 1923). 6 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Cer, J. M. 1975a. Liolaemus melanops Burmeister and the subspecific status of the Liolaemus fitzingeri group (Sauria-Iguanidae). J. Herp. 9(2): 217-222. Cer, J. M. 1975b. Southern Patagonian iguanid lizards of the Liolaemus kingi group. Herpetologica 31(1):109-116. Donoso Barros, R. 1961. Three new lizards of the genus Liolaemus from the highest Andes of Chile and Argentina. Copeia 1961(4):387-391. Donoso Barros, R. 1966. Reptiles de Chile. Edic. Univ. Chile, Santiago exlvi + 458 pp. ; Donoso Barros, R. 1971. A new Liolaemus from Neuquen (Argentina). Herpetologica 27(1):49-51. Hetimicu, W. 1964. Uber eine neue Liolaemus-Art aus den Bergen von Catamarca, Argentinien (Reptilia, Iguanidae). Senck. Biol. 45(3-5): 505-507. Kostowsky, J. 1898. Enumeracién sistematica y distribucién geografica de los Reptiles Argentinos. Rev. Mus. La Plata 8:161-200. Mutter, L., and Hettmicu, W. 1939. Liolaemus-Arten aus dem westlichen Argentinien. III. Uber Liolaemus kriegi, eine neue Liolaemus-Art aus der Gegend des Lago Nahuel Huapi. Zool. Anz 127:44-47. Peters, J. A., and Donoso Barros, R. 1970. Catalogue of the Neotropical Squamata. Part IJ. Lizards and amphisbaenians. Bull. U.S. Natl. Mus. 297 :1-293. &% ab Py, Maat ro UNIVERSITY OF KANSAS PUBLICATIONS MUSEUM OF NATURAL HISTORY The University of Kansas Publications, Museum of Natural History, beginning with volume 1 in 1946, was discontinued with volume 20 in 1971. Shorter research papers formerly pub- lished in the above series are now published as Occasional Papers, Museum of Natural History. The Miscellaneous Publica- tions, Museum of Natural History, began with number 1 in 1946. Longer research papers are published in that series. Monographs of the Museum of Natural History were initiated in 1970. All manuscripts are subject to critical review by intra- and extra- mural specialists; final acceptance is at the discretion of the publications committee. Institutional libraries interested in exchanging publications may obtain the Occasional Papers and Miscellaneous Publica- tions by addressing the Exchange Librarian, University of Kan- sas Library, Lawrence, Kansas, 66045. Individuals may pur- chase separate numbers of all series. Prices may be obtained upon request addressed to Publications Secretary, Museum of Natural History, University of Kansas, Lawrence, Kansas 66045. Editor: Ep Witry PRINTED BY UNIVERSITY OF KANSAS PRINTING SERVICE LAWRENCE, KANSAS UN! MUS, COMP. ZOO, 2122 CIBRARY OCCASIONAL PAPERS JAN 8 4975 : of the HARVARD MUSEUM OF NATURAL HISTORY. * The University of Kansas Lawrence, Kansas NUMBER 77, Pages 1-11 January 11, 1979 BRACHYPOTHERIUM FROM THE TERTIARY OF NORTH AMERICA By DANIEL YATKOLA’? AND LLOyp G. TANNER? Preface—The manuscript for this publication was near final form before the tragic automobile accident which took Dan Yat- kola’s life on March 12, 1976. The manuscript remains essentially as it was prior to his death, because we had already arrived at the conclusion that this rhinocerotid is the first record of Brachypother- ium on the North American continent. Present plans are that the stratigraphic and biostratigraphic evidence regarding the Marsland Formation and the Hemingfordian fauna which was accumulated by Dan Yatkola will be published. _ A “Teleoceras” skull (U.K. 9857) was reported by Galbreath (1953:108) from the Tertiary sediments of Martin Canyon in northeastern Colorado. It was found in University of Kansas col- lecting locality “Quarry A.” Dr. Larry D. Martin, University of Kan- sas, loaned the skull and limb bones to Lloyd Tanner for study in 1974. Recently, additional research was accomplished by Daniel Yatkola, as a part of his Ph.D. dissertation. This included a com- prehensive study of the stratigraphy and biostratigraphy of the Marsland Formation of northwestern Nebraska and adjacent areas. The research prompted a renewed interest in the Martin Canyon local fauna, including additional comparison of this unusual rhi- roceros skull, and the limb bones which were found in the quarry near the skull, with other rhinoceros material in the University of Nebraska Study Collection. We both had the opportunity to com- 1 Deceased, Doctor of Philosophy Candidate, Department of Geology, Uni- versity of Nebraska-Lincoln, Lincoln, Nebraska 68588. 2 Curator of Vertebrate Paleontology and Coordinator of Systematic Col- lections, University of Nebraska State Museum, University of Nebraska-Lincoln, Lincoln, Nebraska 68588. 2 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY pare the skull with other rhinoceros remains in the study collection of the American Museum of Natural History, and Frick Laboratory, New York. Yatkola also visited the Carnegie Museum in Pittsburgh and the Field Museum in Chicago to study fossil vertebrates from sediments connected with this study. We found no species com- parable to the Martin Canyon “Teleoceras” in these collections. We conclude that the specimen belongs to the Old World genus Brachypotherium as discussed below. Wood (1941:91) explained the “assignments” of the genera, ~ Brachypotherium and Teleoceras as follows: The genus Brachypotherium is represented by succes- sively more specialized species from the earliest. Miocene into the Pliocene of Eurasia and north Africa. A direct de- rivative, Teleoceras, appears in the upper Miocene of North America .. . so similar is a primitive Teleoceras to a mod- erately specialized Brachypotherium that assignment be- tween the two genera is based simply on locality, so that essentially indistinguishable specimens from Japan and Ne- braska have been assigned without objection to Brachypo- therium and to Teleoceras ( Mesoceras). z Below, we describe the first New World Brachypotherium and discuss the genus. In our description, the following abbreviations are used for Institutions: A.M.N.H., American Museum of Natural History; K.U., University of Kansas Natural History Museum; U.N.S.M., University of Nebraska State Museum. Brachypotherium Roger Type species—Rhinoceros brachypus Lartet, 1837. Included species—The type species, B. aurelianensis (Nouel, 1866), B. snowi (Fourtau, 1918), B. heinselini Hooijer (1963), B. goldfussi (Kaup, 1854), B. stehlini Roman and Viret (1934), B. lewisi Hooijer and Patterson, 1972; and B. americanus n. sp. (this : paper). Known range.—Lower Burdigalian through Pontian of Europe; lower Miocene through upper Pliocene of Africa; Upper Oligocene to Middle Miocene of Kazakhstan; Miocene of Japan; and early mid-Miocene of North America. Diagnosis.—The following features distinguish Brachypotherium from Teleoceras: foramen ovale and foramen lacerum medius sepa- rate, post tympanic process light and just in contact with post glenoid process, M2/ anteroposterior length not much greater than M3/ length, P1/ large, basilar mound on sphenoid small, greatest breadth across frontal lies above anterior portion of orbit, lacrimal expanded anteriorly, infraorbital foramen located outside of narial notch, teeth relatively low crowned, 2nd metapodials more elongate than in Teleoceras. BRACHYPOTHERIUM FROM NORTH AMERICA 3 Discussion.—Rhinoceroses with broad skulls, strong, chisel- shaped I1/, and short plump limbs are grouped together in the Tribe Teleoceratini Hay, 1902 (Heissig, 1973). The Afroeurasian teleoceratine rhinos are generally referred to the genus Brachy- potherium, while the North American teleoceratine rhinos are referred to Teleoceras. The specimen described below from north- eastern Colorado represents the first documented occurrence of Brachypotherium in North America. The University of Kansas specimen herein identified as Brachypotherium cannot be referred to any North American genus. The primitive Oligocene genera (Subhyracodon, Caenopus, Trigonias) as well as Menoceras and Diceratherium (see Tanner, 1969, p. 398) are easily distinguished from this skull. The brachypothere skull described below differs from the genoholotype skull of Aphelops (A.M.N.H. 8292: Cope and Matthew, 1915; Plates CXXV-CXXVII, CXXXV) in having a dis- tinct rugosity at the tips of the nasals, suggesting a nasal horn on the anterior portion of the nasal. The outline of the nasal is V- shaped in cross-section, whereas it is essentially flat in Aphelops. The narial notch is retracted to a point perpendicular to P3/ in Brachypotherium, while in Aphelops it is retracted to the anterior portion of P4/. The occiput is vertical, broad at top as well as at base and wider than greatest breadth across frontals. In Aphelops the occiput is tilted slightly posteriorly, considerably narrower at top than at base and narrower than greatest breadth across frontals. The occipital crest of Brachypotherium is anteroposteriorly more massive when compared to Aphelops. These characteristics sepa- rate the Colorado brachypothere from all valid species of Aphelops (see Tanner, 1967:11). The genoholotype skull of Peraceras (A.M.N.H. 1880: Cope and Matthew, 1915; Plate CXLIV a-b) differs from the Colorado brach- ypothere in having an anteriorly directed occiput, occipital crest thin, broader skull (especially across frontals), flattened, broader nasals, narial notch retracted to anterior of P4/, post tympanic process not in contact with post glenoid process, I1/ alveolus very small, and premaxilla reduced. In a forth-coming publication (Tanner) evidence will be presented that the nasals of Peraceras probably have paired lateral horns, and therefore differ significantly from those of Brachypotherium. Among North American rhinocerotids, only specimens referred to Teleoceras closely approach the morphology of K.U. 9857. This skull was compared with skulls of all Teleoceras species in the University of Nebraska State Museum collection. The following similarities were observed: skulls short and broad, occiput vertical, broad at top as well as at base, narial notch relatively short (above P3/), nasals have rugosities for terminal horn, I1/ alveoli large, and nasals U-shaped in cross-section. This combination of char- By OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY acters separates the Tribe Teleoceratini Hay (1902) from the Tribe Aceratherini Dollo (1885; cf. Heissig, 1973). The dentition, espe- cially upper molars two and three, of Teleoceras are very distinct from other rhinoceroses of the Pliocene. The known geologic span for Teleoceras is Valentinian through Kimballian, Tanner (1975: 23). It is unfortunate that the remaining teeth of K.U. 9857 are so badly worn that it makes it difficult to-discern the tooth pattern. _ However, if a comparison is made of K.U. 9857 with upper molars — two and three of U.N.S.M. 62097, a right maxillary (P3/-M3/) of a Teleoceras from Clarendonian age deposits, from U.N.S.M. Coll. Loc. Bw-101 (Quinn Rhinoceros Quarry of Skinner and Quinn), Brown County, Nebraska, there is a distinct difference. The com- bined greatest diameter of M2/-M3/ (along the buccal side) for K.U. 9857 is 90 mm, this same dimension for U.N.S.M. 62097 is 125, or about 40 percent larger. Other differences of Brachypotherium americanus new species from Teleoceras are included in the generic diagnosis in this paper. The following species of Teleoceras are valid: T. major; T. fossiger; T. hicksi; T. schultzi, Tanner (1975: 23). Comparisons were made with known illustrations of Brachy- potherium skulls, and also the cast of the holotype skull of B. aurelianensis in the American Museum of Natural History Collec- tion. The differences between K.U. 9857 and Teleoceras also ap- parently separate the Afroeurasian Brachypotherium species from Teleoceras. Based on this evidence we assign specimen K.U. 9857 from northeastern Colorado to Brachypotherium; and considering the small size of the skull and lack of a frontal rugosity, we propose a new species, Brachypotherium americanus. Brachypotherium americanus new species Fics. 1, 2, 3 Holotype.—K.U. 9857, a nearly complete skull. ; : Type Locality—Northeast one-quarter, Section 27, Township 11 North, Range 53 West, University of Kansas Museum of Natural History, Quarry A, Martin Canyon, Logan County, Colorado. Stratigraphic Occurrence.—Silty sands in the Pawnee Creek Formation, approximately 20’ above indurated, lithic conglomerate (see Galbreath, 1953; 26, Section XV). The evidence for the strati- graphic occurrence and biostratigraphic correlations of the Martin Canyon local fauna and deposits containing the vertebrates will be presented in Yatkola’s dissertation which will be published by the University of Nebraska Conservation and Survey Division in the near future, The vertebrate assemblage from Quarry “A” sediments is older than the Marsland Quarry assemblage U.N.S.M. Coll. Loc. Bx-22, Box Butte County, Nebraska, located approximately 25’ below the BRACHYPOTHERIUM FROM NORTH AMERICA 5 top of the Marsland Formation; probably older than the Bridgeport Quarries, U.N.S.M. Coll. Loc. Mo-113-114, Morrill County, Ne- braska; and younger than Runningwater Quarry U.N.S.M. Coll. Loc. Bx-58, located approximately 20’ above the base of the Mars- land Formation. The faunas collected from the Marsland Forma- tion were considered in the establishment of the Hemingfordian, North American Land Mammal “Age” (Wood et al., 1941). Quarry “A” Local Fauna of Martin Canyon is best considered to be earliest Hemingfordian, which as defined by the Wood Committee, repre- sents middle Miocene “time” in North America. Diagnosis.—Smaller than all described species of Brachypother- ium except B. aurelianensis var. gailiti (Borissiak, 1927). Only slightly smaller than B. aurelianensis (Nouel, 1866). Brachypo- therium americanus lacks the well-defined frontal rugosity which is a definite character of B. aurelianensis, it also lacks a distinct nasal cleft, and has a more excavated occiput. Description—tThe type skull (K.U. 9857), is only slightly dis- torted. Tooth wear indicates the remains of a very aged individual. The premaxilla is pushed dorsally reducing the depth of the nasal incision and its anterior end is missing. All the cheek teeth are missing except the M2/-M3/ on each side. The right zygomatic arch is pushed in slightly and the left side of the face was acci- dentally cut away. The specimen was discovered while making a soil core. The skull is short, broad and low. Fic. 1.—Dorsal view of the holotype of Brachypotherium americanus (K.U. 9857). Length of skull: 440 mm. The free portion of the nasal is rather long (144 mm from the narial notch to the tip of the nasals), U-shaped in cross-section, 6 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY inclined at an angle of about 10 degrees above the plane of the frontal, gradually tapers anteriorly and has a distinct rugose area at the anterior end. The posterior margin of the narial notch is perpendicular to the center of the P3/ alveolus. The infraorbital foramen is located outside and about 14 mm behind the posterior margin of the narial notch, placing it well out on the face. The lacrimal is large, ovoid and expanded well onto the face (53 mm from the front of the orbit to the anterior margin), which amounts to about 70 percent of the distance between the narial ~ notch and front of the orbit. The lacrimal foramen is large (7 mm in diameter) and located internal to a distinct tuberosity on the anterior margin of the orbit. The frontal is broadly concave and widest at a point directly above the anterior margin of the orbit. The frontal nasal suture is not evident. There is no apparent rugosity but arterial nutritive Fic. 2.—Posterior view of the holotype of Brachypotherium americanus (K.U. 9857). Breadth of skull: 285 mm. depressions radiate from a convex surface on the frontal as in B. aurelianensis. The frontal crests are low and extend forward at a 35 degree angle from the plane of bilateral symmetry. The occiput is vertical to the condyles and distinctly elevated above the plane of the frontal—the angle between the plane of BRACHYPOTHERIUM FROM NORTH AMERICA i the frontal and the anterior slope of the saggital crest is approxi- mately 145 degrees. The sagittal crest is broad, measuring 23 mm at a point 50.mm from the posterior margin of the occiput, dis- tinctly elevated above the parietal and sharply defined. In dorsal view the occipital crest is relatively broad, amounting to 63 percent of the maximum skull width. In posterior view (Fig. 2) the top of the occiput is almost as broad as the base, tapering only slightly dorsally, and is more broad than high. The external occipital protuberance forms a low crest with deep emarginations on either side. The lateral occipital area is also deeply pocketed. The occipital condyles are well developed, widely placed, amounting to 58 percent of the basal width of the occiput. The dorsal and ventral margins of the foramen magnum exhibit distinct U-shaped notches; the ventral notch is deeper. The basicranial length, measured from the center of the fora- men magnum to a point midway between the posterior end of the maxilla, makes up about 35 percent of the skull length. The post- glenoid process is very massive, measuring at the base, 36 mm anteroposteriorly and 31 mm transversely, and extends further ven- trally (57 mm below the basicranial plane) than the paraoccipital process. The anterior margin is oriented nearly perpendicular to the basicranial plane. It is anteroposteriorly elongate with the trans- verse width tapering gradually towards its tip. The paraoccipital process is much lighter than the post-glenoid and tapers to a less, anteroposteriorly elongate tip. The process is “emis. - Fic. 3.—Lateral view of the holotype of Brachypotherium americanus. Length of skull: 440 mm. roughly triate in form, being divided by distinct ridges extending the length of the process. The internal sulcus is the deepest. The hypoglossal foramina are small and located just internal to 8 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY the base of the paraoccipital processes, measuring 45 mm across. The foramina ovale are located just internal to the articular groove of the temporal bone and are separated by a large bony plate from the foramen lacerum medius. The bulla tympanica is not preserved. The mastoid process is much expanded into a broad light plate- like process in thickness that almost comes into contact with the post-glenoid process and extends ventral to the glenoid surface. The laterally expanded post-glenoid process and the mastoid en-_ close a long, horizontally oriented passageway leading towards the | external auditory meatus. The basal part of the occipital is not much expanded above the plane of the basicranium. There is a narrow, median, antero- posteriorly oriented crest located on the body of the sphenoid and a low basal mound. In Teleoceras there is a distinct arched basal mound on the body of the sphenoid. The body of the sphenoid is flat and broad in Brachypotherium but expanded in Teleoceras. The maxillary teeth extend nearly to the posterior margin of the maxilla, which projects well back into the orbitotemporal fossa. The maxillary tuberosity is not as well developed as in Teleoceras. The widely separated pterygoid processes flare externally and are roughened posteriorly. The palate is approximately the same width between M2/ (64 mm) as between P1/ (71 mm). The broad palatal notch ex- tends forward to the anterior end of the M3/ and fills most of the breadth of the palatal area. The zygomatic arch does not rise very high or steeply. The widest part of the arch is 67 mm at a point located just posterior to the M3/. The zygomatic process of the temporal is low, rec- tangular-shaped, and not much expanded below the ventral margin of the zygomatic arch. A distinct fossa is anterior to the zygomatic process of the temporal on the ventral surface of the arch. The malar extends posteriorly to a point 54 mm anterior of the front of the zygomatic process of the temporal and well in front of the fossa anterior of this process. In Teleoceras the M3/ extends pos- teriorly to the anterior end of this depression. The malar narrows anteriorly and extends slightly into the face, and forms the ventral margin of the orbit. The malar is very narrow (29 mm) below the center of the orbit. The facial crest is low and not well defined anteriorly. The premaxilla is somewhat distorted, but is broad and tapers only slightly anteriorly. The I1/ alveolus is large, implying a large incisor, The only teeth preserved are the last two molars in each side. These teeth are badly worn. The only observable feature is the presence of a low, weak internal cingulum around the internal BRACHYPOTHERIUM FROM NORTH AMERICA 9 margin of the tooth. The ectoloph length of the M2/ is not much greater than the ectoloph length of M3/. Discussion.—Skulls are known for only four of the described species of Brachypotherium: B. aurelianensis (Nouel, 1866), B. goldfussi (Kaup, probably a junior synonym of B. brachypus), and B. lewisi (Hooijer and Patterson). The remaining species are based on dental features. Brachypotherium americanus is smaller than all previously de- scribed species except B. aurelianensis var. gailiti (Table 1). B. lewisi is much larger and differs in the following features: shorter nasals, narial notch retracted to above anterior of P4/ and a broader skull. B. americanus differs from B. brachypus in its smaller size. TasLE 1.—Comparative measurements (mm) of Brachypotherium americanus new species and B. aurelianensis var. gailiti (after Borissiak 1927 ) B. americanus B. aurelianensis K.U. 9857 var. gailiti Occipital condyles to tip of premaxillary ..... (450 )* Occipital condyles to tips of nasals ......... 492, Midpoint occipital crest to tips of nasals .... 440 (369 ) Anterior margin of P’ to occipital condyles .. 390 Narial notch to occipital crest -............. 335 Palatal notch to foramen magnum .......... 210 Palatal notch to palatal foramina ........... Ze Nanialinoteh tostips of masals) i... 3 .0.0e0%.. 150 Zygomatic breadth (maximum) ............ 285 Width across palate between M’ ............ 64 Occipital height, base condyles at midpoint LOMCLES a eee oe chee ae mA eee wee 165 Occipital width (maximum) ...... base 195 midpoint .... 180 Condylar width (outer margins occ. condyles) 100 Post M> occipital condyles 20.0.3... ..+. 5: 187 itenetiae Wien (maximum) ss .0224 aaeeee sce 43.3 34.7 Wradtinwvien(maximiam) “el. fe l.'dtias veces (53.6) 41.2 ikenoth Mien ectoloph’)) >. -.2 ech aes ces 46.6 36.5 Macchi Ni: emmaximitm) “fe tse hs lemcre ee aces 45.6 36.9 Menotiae NI -mudiline, cas .e. sed ve sao else's 42.4 31.5 IE val Ketey0) RCL eee ae i ie Na rea ASP a ii2 iL trans. 14.5 Eee COlISHe edt). etetcare ere card sae ered AaP rs IAG trans .... a PST Post. margin narial notch to ant. end orbit .. 75 54 Width nasals at post. end narial notch ...... 94 Width supraoccipital processes ............ 156 Ant. border orbit—external aud. meatus ..... 228 Width post-glenoid process (external) ...... 124 Diem CLES OUALACASE: thers © 1a) aie yeu wi wiene joie) cue 3 144 Distance between foramina ovale ........... 59.4 * Numbers in parentheses are approximate diameters. 10 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Brachypotherium americanus is closest in morphology and size to B. aurelianensis. The holotype skull of B. americanus is 15 per- cent smaller than the holotype B. aurelianensis. This size difference seems significant, since both skulls probably represent males. The skull referred to B. goldfussi, which was originally referred to B. brachypus (Osborn, 1904a; Roger, 1900), is approximately 15 percent larger than B. aurelianensis. B. aurelianensis has a frontal rugosity, which is absent in B. americanus and all other species of Brachypotherium. It is present on Teleoceras mediocornutus, Os-~ born 1904b. The nasal cleft is not distinct in B. americanus, as in B. aurelianensis. This may be an old age feature, although very aged Teleoceras individuals retain a distinct nasal cleft. The occi- put of B. americanus is distinctly more excavated than B. aurelian- ensis, but this may be an artifact of preservation. Brachypotherium americanus is most closely related to B. aurelianensis. We con- sider B. americanus to be derived over B. aurelianensis, based on the lack of a frontal rugosity and an emarginate occiput. The presence of Brachypotherium in North America is of great significance for intercontinental correlation. Specimens of B. aurelianensis have been recovered from Neuville—aux—Bois in France. Mayet (1908) and most European vertebrate paleontol- ogists consider the age of the vertebrate assemblage from Neuville —aux—Bois, and its principal -correlatives, Chilleurs—Aux—Bois and Wintershof West, to be early Burdigalian. The Quarry “A” assemblage cannot be much younger than faunas from these Euro- pean localities. Wilson (1960, 1968) has already pointed out the similarity between the early Burdigalian rodents and insectivores and those of Quarry “A.” The limb bones from Quarry “A,” which are reported to have been associated with the skull K.U. 9857, are not being assigned to a species at this time. It is necessary that a further study of these post axial elements be made before a conclusive assignment_can be made. A preliminary study indicates that the dimensions are near those of Menoceras falkenbachi Tanner, from the Bridgeport Quarries U.N.S.M. Collecting Locality Mo 113-114, Morrill County, Nebraska. The elements are not the short heavy limb bones typical of the Teleoceras from the Pliocene deposits. ACKNOWLEDGEMENTS I appreciate the support and counsel from many of Dan Yat- kola’s friends in the most difficult task of preparing this publication. I am especially grateful to R. George Corner, Larry D. Martin, and Michael Voorhies for the many ways they helped, and to Rebecca Monke and Gail Littrell for their assistance in the typing of the manuscript. BRACHYPOTHERIUM FROM NORTH AMERICA 11 LITERATURE CITED Boriss1aAk, A. A. 1927. Brachypotherium aurelianense Nouel, var. nov. gailiti, from the Miocene deposits of the Turgai region. Bull. Acad. Sci. St. Peters- burg, 6( 21) :273-286. Corr, E. D., and W. D. Marruew. 1915. Hitherto unpublished plates of Tertiary Mammalia and Permian Vertebrata. Mono. Amer. Mus. Nat. Hist. (ser. 2): Pls. 125-144b. GaLBREATH, E. C. 1953. A contribution to the Tertiary geology and paleon- tology of northeastern Colorado. Univ. Kansas Paleont. Contrib., Verte- brata (art. 4):1-120, 26 figs., 2 pls. Hetssic, K. 1973. Die Unterfamilien und Tribus der rezenten and fossilen Rhinocerotidae (Mammalia). Saugetierkundliche Mitteilungen, BLV-Ver- lagsgesellschaft, Miinchen, 21(1):25-30. Hooryer, D. A., and B. Parrerson. 1972. Rhinoceroses from the Pliocene of northwestern Kenya. Bull. Mus. Comp. Zool., 144(1):1-26, 11 figs., 9 tables. Mayet, L. 1908. Etude des Mammiféres miocénes des Sables de l’Orléanais et des Faluns de la Touraine. Ann. Univ. Lyon, new series 1(24):316 pp., 12 pls. Ossorn, H. F. 1898. The extinct rhinoceroses. Mem. Amer. Mus. Nat. Hist., 1(3):75-164, 8 pls. Ossorn, H. F. 1904a. New Miocene rhinoceroses with revisions of known species. Mem. Amer. Mus. Nat. Hist., 20(27):307-326, 21 figs. Ossorn, H. F. 1904b. Phylogeny of the rhinoceroses of Europe. Bull. Amer. Mus. Nat. Hist., 13:229-269. Rocer, O. 1900. Ueber Rhinoceros goldfussi Kaup und die anderen gleich- zeitigen Rhinocerosarten. Ber. naturwiss. Ver. Schwaben Neuberg, 34: e522 ols! Stmmpson, G. G. 1945. The principles of classification and a classification of mammals. Bull. Amer. Mus. Nat. Hist., 85:ix-xvi, 1-350. Tanner, L. G. 1967. A new species of rhinoceros, Aphelops kimballensis from the latest Pliocene of Nebraska. Bull. Univ. Nebraska State Museum 6(1):1-16, 1 fig., 5 pls. Tanner, L. G. 1969. A new rhinoceros from the Nebraska ae: Bull. Univ. Nebraska State Mus., 8(6):395-412, 10 figs., 2 tables. Tanner, L. G. 1975. Cenozeie mamunals frome the Central Great Plas. Stratigraphic occurrences of Teleoceras with a new Kimballian species from Nebraska. Bull. Univ. Nebraska State Mus., 10(1), Part 2:23-33, Frontis- piece, 6 figs., 4 tables. Witson, R. W. 1960. Early Miocene rodents and insectivores from north- eastern Colorado. Univ. Kansas Paleont. Contrib., 24(7):1-92, 131 figs. Wison, R. W. 1968. Insectivores, rodents and intercontinental correlation of the Miocene. XXIII Int. Geol. Cong., 10:19-25, 1 table. Woop, H. E., II. 1941. Trends in rhinoceros evolution. Trans. New York Acad. Sci., II(3):83-99. Woop, H. E., et al. 1941. Nomenclature and correlation of the North Amer- ican continental Tertiary. Bull. Geol. Soc. Amer., 52:1-48, 1 pl. UNIVERSITY OF KANSAS PUBLICATIONS MUSEUM ‘OF NATURAL HISTORY The University of |Kansas Publications, Museum of Natural History, beginning with volume 1 in 1946, was discontinued with volume 20 in 1971. Shorter research papers formerly pub- lished in the above series are now published as Occasional Papers, Museum of Natural History. The Miscellaneous Publica- tions, Museum of Natural History, began with number | in 1946. Longer research papers are published in that series. Monographs of the Museum of Natural History were initiated in 1970. All manuscripts are subject to critical review by intra- and extra- mural specialists; final acceptance is at the discretion of the publications committee. Institutional libraries interested in exchanging publications may obtain the Occasional Papers and Miscellaneous Publica- tions by addressing the Exchange Librarian, University of Kan- sas Library, Lawrence, Kansas, 66045. Individuals may pur- chase separate numbers of all series. Prices may be obtained upon request addressed to Publications Secretary, Museum of Natural History, University of Kansas, Lawrence, Kansas 66045. - Editor: Ep WiLEy PRINTED BY UNIVERSITY OF KANSAS PRINTING SERVICE LAWRENCE, KANSAS } UN MUS. COMP. ZOOL. 2| 2.2 LIBRARY MAY 4 6 1979 HARVARD of the UNIVERSITY MUSEUM OF NATURAL HISTORY The University of Kansas Lawrence, Kansas OCCASIONAL PAPERS NUMBER 78, PAGES 1-10 APRIL 27, 1979 SYSTEMATIC STUDIES OF DARTERS OF THE SUBGENUS CATONOTUS (PERCIDAE), WITH THE DESCRIPTION OF A NEW SPECIES FROM CANEY FORK, TENNESSEE By Marvin E.. BRAASCH! AND LAWRENCE M. PAGE? In the middle 1960’s Warren U. Brigham, then at Tennessee Technological University, collected a previously unknown darter while conducting a census of the fishes of Putnam County, Tennessee. This darter, an endemic of Caney Fork and nearby tributaries of the Cumberland River, is herein described and com- pared with its closest relative, Etheostoma squamiceps Jordan. Etheostoma squamiceps is a polytypic form under examination by several investigators. Discussion of E. squamiceps herein is limited to a comparison with the new species. METHODS Characters——Characters examined were numbers of lateral scales, pored lateral line scales, unpored lateral scales, transverse scales above and below the lateral line, infraorbital canal pores, supraorbital canal pores, supratemporal canal pores, preoperculo- mandibular canal pores, lateral canal pores, coronal pores, dorsal fin spines and rays, branched caudal fin rays, anal fin spines and rays, pelvic fin spines and rays, and pectoral fin rays; pigmentation patterns; squamation of head, nape and breast; head length/stand- ard length (HL/SL), body depth/SL (BD/SL), caudal peduncle depth/SL (CPD/SL), pectoral fin length/SL (P1L/SL), pelvic fin length/SL (P2L/SL), preorbital length/HL (PreOL/HL), eye * Assistant Professor, Belleville Area College, Belleville, Illinois 62221 * Associate Taxonomist, Illinois Natural History Survey, Urbana, Illinois 61801 2 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY diameter/HL (ED/HL), first dorsal fin height/first dorsal fin base length (DH/DIL). Counts and measurements were made as described by Hubbs and Lagler (1964) and Page and Braasch (1976) with the follow- ing exceptions. The number of transverse scales was counted from the anal fin insertion anteriodorsally to the base of the first dorsal fin and expressed as two variables—scales above lateral line and scales below lateral line. The lateral line row was arbitrarily in- cluded in the ventral count and all scales no matter how small were counted. First dorsal fin height was the length of the first dorsal spine. Cheeks, opercles, nape and breast were classified as scaled if scales were exposed with free margins, embedded if scales were covered with skin (no exposed margins), and unscaled if no scales were present. Analysis—Counts and measurements were compared among samples determined by sex (dimorphic), by age (ontogenetic), and by stream system (geographic). A multivariate test (Stepwise Discriminant Analysis BMDO7M, KU Comp. Center) and one-way analysis of variance were used to determine significances of differ- ences in means of samples. Etheostoma olivaceum, new species Dirty Darter Holotype.—Illinois Natural History Survey 75734, an adult male 50.5 mm SL (Fig. 1), collected in Rock Springs Branch at Rock Springs Church, 2 km N Buffalo Valley (36° 10’ N, 85° 47’ W), Putnam County, Tennessee, on 14 June 1975 by B. M. Burr and L. M. Page. Fic. 1,—Etheostoma olivaceum holotype (INHS 75734), male, 50.5 mm SL, Rock Springs Branch, Putnam Co., Tennessee, 14 June 1975. SYSTEMATIC STUDIES OF DARTERS 3 Paratypes.—A total of 87 specimens deposited as follows: 18— Illinois Natural History Survey (INHS 75735, 10 specimens, 36-58 mm SL, same collection data as holotype; INHS 75576, 8 speci- mens, 35-64 mm SL, same locality as holotype, 11 Dec. 1976); 6—U. S. National Museum (USNM 216895, 34-46 mm SL, same collection data as holotype); 6—University of Michigan Museum of Zoology (UMMZ 200208, 35-48 mm SL, same collection data as holotype); 6—University of Tennessee (UT 91.1332, 37-45 mm SL, same collection data as holotype); 6—Tulane University (TU 101391, 37-50 mm SL, same collection data as INHS 75576); 6—University of Alabama (UAIC 5341.01, 39-48 mm SL, same collection data as INHS 75576); 6—Northeast Louisiana University (NLU 35160, 41-51 mm SL, same collection data as INHS 75576); 33—The University of Kansas (KU 11442, 34-59 mm SL, Dry Fork, 3 km W Gordonsville, Smith Co., TN, 7 April 1966). Etymology.—The name olivaceum refers to the olive color of nonbreeding individuals. The common name, dirty darter, was suggested by Dr. David A. Etnier, University of Tennessee, and refers to the drab color and indistinct mottling on the sides. Diagnosis.—The subgenus Catonotus of Etheostoma was diag- nosed by Kuehne and Small (1971). The presence of an uninter- rupted infraorbital canal in E. olivaceum and E. neopterum Howell and Dingerkus (1978) requires that the subgeneric description of the infraorbital canal be modified to “infraorbital canal interrupted Or uninterrupted.” In addition, the cheek, breast, nape and prepectoral area may be scaled or unscaled. Etheostoma olivaceum is distinguished from all other members of the subgenus by the following combination of characteristics: an uninterrupted infraorbital canal with eight infraorbital pores, 10 preoperculomandibular pores, 13 or more pored lateral line scales, scales present on nape and prepectoral area, scales absent on cheek and opercle, distinct black vertical bands on caudal fin, branchiostegal membranes slightly connected, no bar pattern on cheek, no red or blue pigments, no dark suborbital bar, usually nine dorsal spines, 12-13 dorsal rays, and seven or eight anal rays. Comparisons.—Etheostoma olivaceum is separated from all species of Catonotus except E. neopterum by the uninterrupted infraorbital canal and from E. neopterum by the absence of scales on the opercle. E. olivaceum is further separated from the barcheek species of Catonotus (virgatum, obeyense, barbouri, smithi, striatulum) by the lack of the bar pattern on the cheek, lack of red and blue pigments, and modally 7-8 anal rays; from E. flabel- lare, E. kennicotti, and the barcheeks by the presence of scales on the nape and prepectoral area; and from its closest relative, E. squamiceps, by the absence of scales on cheeks and opercles, ab- sence of a dark suborbital bar, modally seven scales above the 4 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY lateral line (Table 1), modally 12 pectoral rays (Table 2), and modally 12 dorsal rays (Table 3). Description—Etheostoma olivaceum is a moderately large species of Catonotus, largest specimen 67 mm SL; infraorbital canal uninterrupted, normally with 8 pores (9 in 8%); preoperculo- mandibular canal with 10 pores (9 in 8°); supraorbital canal with 4 pores (5 in 24°/); supratemporal canal interrupted medially, pores 2—2; premaxillary frenum moderate to broad; branchiostegal rays six; branchiostegal membranes with slight to moderate fusion; preopercle entire. Head unscaled; body including nape scaled, breast unscaled or with embedded scales, prepectoral area scaled (often em- bedded;) 44-58 lateral scales; 13-46 pored lateral line scales; 10-37 unpored lateral scales; 5-8 (usually 6-7) transverse scales above lateral line; 7-12 (usually 8-10) scales below lateral line. First dorsal fin with 8-10 (usually 9) spines, with small terminal knobs in both sexes; second dorsal fin with 11-14 (usually 12-13) rays; 2 anal spines; 6-9 (usually 7-8) anal rays; 14-16 branched caudal rays; 11-13 (usually 12) pectoral rays; 1 pelvic spine; 5 pelvic rays. Body proportions: HL/SL, X=0.29; BD/SL, 0.21; CPD/SL, 0.13; PIL/SL, 0.23; P2L/SL, 0.19; PreOL/HL, 0.25; ED/HL, 0.27; DH/DIL, 0.26. Color of freshly preserved specimens: Basic body color of females and nonbreeding males (Fig. 2) varies from yellowish tan to dark olive-brown. Pale individuals have dark markings on sides varying from an indistinct mottling to 8-10 fairly distinct vertical bars; bars do not connect dorsally or ventrally. Dark spots are sometimes present at base of caudal fin but are never darker than other markings. Dark individuals show no pattern and are uni- formly olive-brown laterally. k Dorsum usually lighter than sides, with 0-10 dark blotches. Venter much lighter than sides, ranging from olive or gray to white. Large dark humeral spot present. Anterior. lateral line pores un- pigmented and conspicuous on darker individuals. Head with dark pre-and postorbital bars; no suborbital bar (teardrop); head above bars dark, cheeks and throat light. First dorsal fin colorless except for a gold or brown subdistal band, and scattered melanophores concentrated on membranes along subdistal band and along basal part of fin; basal part of fin thickened; light colored knobs present on tips of spines. Second dorsal fin with four to seven dark bands; rays often produced beyond membranes making distal edge of fin appear ragged; basal part of fin thickened. Caudal fin with melanophores clustered along rays forming 6-9 dark bands; distal margin of fin clear. Anal and pelvic fins clear. Pectoral fins with 6-9 indistinct bands. SYSTEMATIC STUDIES OF DARTERS P " : yio,y Aouen oF L9°0 ECL OOT If ce 09 v ulnaovayo DUL0;S02Y4Iq € 19 LS 4 seit AUG) Bi 87 69°0 VGI 6S I VG CE G puepiequny) s[ppruz SP 19°0 OGL 99 G LE GG G puepequimny IoMo] sdaowupnbs pu0jsoay7q SSeS AOD as x N ial eI ral II oseurerq] skey ‘ON x satoedg a a ee a ee ‘unaovayo “iq pue sdao;wonbs vwoysoayzq ur skex [esIop jo sjun0Qg—g¢ wav L GE 6¢°0 TGL OOT GI v8 v yo AoueD uUNnaaDvaAyO DUL0jsSOaYyIT cI L6 eT ‘STV.LOL OF €v'0 S01 6S I LY II puepoquiny e[ppruz Tv 970 CIT 99 VI os G Ppueprequmy) 19M sdaowuonbs DULOISOBYIT SE AD as x N €I ra II Or oseureld skey ‘ON y satoeds ee ee ee ee aa ae ee ‘ungovayo “q pue sdaowoenbs vwuojsoayzq ut skex [e10}0ed Jo sjun0Q—zZ AIaVI, 66 C0 €6 T L c&é Gh OL & YG C90 = 2:9. (OF aco eee O0T yoy AoueD uinaavayjo Duo s0aYyIT Or GL 8&€ @ O Of 66 OT *S'TV.LO.L GL 90 98 G Vom LCn nc 06 aaah ead 1.0000 INR Set 0.5315 1.0000 eee ee ee — 0.7634 — 0.6583 1.0000 i = ~ 0.6408 _~ 0.9759 0.8067 1.0000 6 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY plete glaciation has had no effect on present species richness in the avifaunas of Fuego-Patagonian islands then predicted (Y;,BIRDS) and observed (Y;BIRDS) numbers of species of birds for each of the three northern islands should be the same. If the predicted number of species of birds for each of the northern islands is sub- stantially below the observed number, then the null hypothesis is rejected and one of three sets of circumstances pertains: - 1) the southern islands (Grande, Hoste, Wollaston, Los Estados) are below equilibrium and the northern islands (La Mocha, Chiloé, Las Guaitecas) are at equilibrium. 2) the southern islands are at equilibrium and the northern islands are above equilibrium. 3) the southern islands are below equilibrium and the northern islands are above equilibrium. The results of the test suggest that La Mocha has reached equilibrium for number of species of birds and that Chiloé and Las Guaitecas are above equilibrium. This outcome is not unex- pected in that La Mocha is very small in area (48 km?) and com- paratively distant (35 km) from land. Moreover, since the southern island regression model predicted the equilibrium species richness for the avifauna of La Mocha, it is reasonable to suggest that species richness of the avifaunas of each of the southern islands is also at equilibrium. In addition, the results of the test are consistent with the notion that extinction rates are inversely related to island area. Estimated rates of net loss in years per species for Chiloé and Las Guaitecas are inversely proportional to island area. Since La Mocha very likely was an island 13,000 years ago, the species richness of its avifauna may have reached equilibrium soon after post-glacial rise in sea- level at a net insular extinction rate that cannot be determined. Preliminary conclusions of our study are: 7 1. Numbers of species of birds are at present above equilibrium on Chiloé and Las Guaitecas. These islands were part of the continent during the Llanquihue Glaciation and since they were only partly glaciated, they had the continental species richness of birds when they became islands during glacial shrinkage between TasLe 5.—Prepicrep (Y;BIRDS) AND opservep (Y;BIRDS) NUMBERS OF SPECIES OF BIRDS ON THREE SOUTHERN SOUTH AMERICAN ISLANDS. THE 95% CONFIDENCE LIMITS FOR Y; ARE: LA Mocua +12.068, Curmo& +12.010, AND Las GuaIrEcAS +12.011. Distance Differ- Island from Island Y.BIRDSw~—Ya = Yi:BIRDS Y\BIRDS ence areakm?® land km La Mocha Li7 92 = 85 84 -] 48 35 Chiloé 160 64 = 96 119 +23 8138 8 Las Guaitecas 150 8&8 = 67 81 +14 234 30 BIRDS ON AUSTRAL SOUTH AMERICAN ISLANDS it 13,000 and 12,000 BP. They have been “relaxing” toward equi- librium ever since. 2. La Mocha probably was not part of the continent during the Llanquihue Glaciation; in any case, the island was not glaciated. The species richness of its avifauna is now at equilibrium. 3. Isla Grande de Tierra del Fuego, Los Estados, Wollaston, and Hoste all were completely glaciated during the Llanquihue Glaciation and their avifaunas extinguished. The species richness of the avifaunas of each of these islands now is at equilibrium. 4. Latitude, in concert with island area and distance to closest land, has an important influence on numbers of species of birds on austral South American islands. We have developed a preliminary model that estimates equi- librium values for species richness of birds of forested southern South American islands. Using these estimates we have calculated net insular extinction rates of partly glaciated or unglaciated “land bridge” islands that appear to be “relaxing” toward species-area equilibria, the sizes of which are directly related to area and in- versely related to both latitude and distance to land. Large south- ern South American land bridge islands that were partially or not at all glaciated about 13,000 years BP have present species rich- nesses for birds that are greater than equilibrium. Net insular ex- tinction rates of high latitude land bridge islands may be an order of magnitude slower than those of land bridge islands near the equator (for data on estimated past continental and present species richnesses of low latitude land bridge islands see Terborgh, 1974:717 and Diamond, 1973:760-761). Fuegian islands were completely glaciated during the late Pleistocene and lost their biotas. Follow- ing glacial shrinkage between 12,000 and 18,000 BP birds reap- peared on these islands and numbers of species reached equilibrium an unknown number of years before present. Our analysis and interpretation of data bearing on the difference in present species richness of birds between austral South American islands whose biotas were eliminated by glaciation and those whose biotas were not are tentative and speculative. Our sample of seven islands was small and the avifaunal data not as detailed as we would have liked. Neither of these deficiencies can be remedied without additional fieldwork. Acknowledgments.—Unpublished data concerning the birds of Isla de los Estados were obtained from the letters and field notes of Philip Angle, David Bridge, Jeffrey Boswall, H. Meade Cabot, Jr., Rae Natalie P. Goodall, Oscar Kuhnemann, and Robin J. Prytherch. We are grateful to William M. Adams for sending us a list of birds he saw during recent field work on Isla Hoste. We had many helpful discussions with Robert S. Hoffmann, James Koeppl, and Norman Slade at various stages in the develop- 8 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY ment of our study and are grateful for their enthusiastic interest. Dr. Slade calculated confidence intervals and assisted generously with statistical consultation. Drs. Beryl Simpson and Frangois Vuilleumier kindly read the manuscript and sent us many useful comments and suggestions. The map of southern South America was drawn by Debra Bennett. LITERATURE CITED BArz, A. M., and Scm.ato, G. J. Late Cenozoic Environmental Changes in Temperate South America. In press. Buttock, D. S. 1935. Las aves de la isla de la Mocha. Rev. Chilena Hist. Nat., 39:232-253. CasTELLANos, A. 1935. Observaciones de algunas aves de Tierra del Fuego e isla de los Estados. Hornero, 6:22-37. CasTELLANOS, A. 1937. Observaciones de algunas aves de Tierra del Fuego e isla de los Estados. Hornero, 6:382-394. CHAPMAN, F. M. 1933. Autobiography of a bird-lover. Appleton-Century, New York. : DraMonp, J. M. 1973. Distributional ecology of New Guinea birds. Science, 179:759-769. Hopkins, D. M. 1973. Sea level history in Beringia during the past 250,000 years. Quaternary Research, 3:520-540. Housse, R. P. R. 1924. Apuntes sobre las aves de la isla La Mocha. Rev. Chilena Hist. Nat., 28:47-54. Housse, R. P. R. 1925. Adicién a los “Apuntes sobre las aves de la isla La Mocha.” Rev. Chilena Hist. Nat., 29:225-227. Humpurey, P. S., Brince, D., REyNoxps, P. W., and Pererson, R. T. 1970. Birds of Isla Grande (Tierra del Fuego). Smithsonian Institution, Washington. Jounson, A. W. 1965-67. The Birds of Chile and adjacent regions of Argen- tina, Bolivia, and Peru. Vols. I and II. Platt Establicimientos Graficos, S. A. MacArtuur, R. H., and Witson, E. D. 1967. The Theory of Island Bio- geography. Monographs in Population Biology, I. Princeton. Mercer, J. H. 1976. Glacial history of southernmost South America. Qua- ternary Research, 6:125-166. MEYER DE SCHAUENSEE, R. M. 1966. The Species of Birds of South America and their Distribution. Livingston Publishing Co. Oxtroc, C. C. 1948. Observaciones sobre la avifauna de Tierra del Fuego y Chile. Acta Zool. Lilloana, 5:437-531. Oxtroc, C. C. 1950. Notas sobre mamiferos y aves del Archipiélago de Cabo de Hornos. Acta Zool. Lilloana, 9:505-532. Oxroc, C. C. 1958. Observaciones sobre la avifauna antartica y de alta mar desde el Rio de la Plata hasta los 60° de latitude sur. Acta Zool. Lilloana, 15:19-33. Oxnoc, C. C. 1963. Lista y distribucién de las aves argentinas. Opera Lilloana IX. Instituto Miguel Lillo, Tucuman, Argentina. Oustratet, E. 1891. Oiseaux. In: Mission Scientifique du Cap Horn 1882- 1883. Tome VI, Zoologie. Paris. Péraun, J. E., and Humpurey, P. S. Models to predict species richness of birds on austral South American islands. MS. Péraurn, J. E., and Humpurey, P. S. Preliminary study of avifaunal resem- blance for selected austral South American islands. MS. BIRDS ON AUSTRAL SOUTH AMERICAN ISLANDS 9 Peraur, J. E., and YANEz, J. Consideraciones biogeograficas sobre isla La Mocha, Chile. Bol. Mus. Hist. Nat. Valparaiso. In press. ReyNo tps, P. W. 1935. Notes on the Birds of Cape Horn. Ibis, 5:65-101. Simpson, B. P. Quaternary biogeography of the high montane regions of South America. In press. TERBORGH, J. 1974. Preservation of Natural Diversity: The Problem of Extinc- tion Prone Species. BioScience, 24:715-722. as T=) udilge iat Aah wd UNIVERSITY OF KANSAS PUBLICATIONS MUSEUM OF NATURAL.HISTORY The University of Kansas Publications, Museum of Natural History, beginning with volume 1 in 1946, was discontinued with volume 20 in 1971. Shorter research papers formerly pub- lished in the above series are now published as Occasional Papers, Museum of Natural History. The Miscellaneous Publica- tions, Museum of Natural History, began with number 1 in 1946. Longer research papers are published in that series. Monographs of the Museum of Natural History were initiated in 1970. All manuscripts are subject to critical review by intra- and extra-mural specialists; final acceptance is at the discretion of the publications committee. Institutional libraries interested in exchanging publications may obtain the Occasional Papers and Miscellaneous Publica- tions, by addressing the Exchange Librarian, The University of Kansas Library, Lawrence, Kansas 66045. Individuals may pur- chase separate numbers of all series. Prices may be obtained upon request addressed to Publications Secretary, Museum of Natural History, The University of Kansas, Lawrence, Kansas 66045. Editor: Ep Witry PRINTED BY UNIVERSITY OF KANSAS PRINTING SERVICE LAWRENCE, KANSAS ww=. COMP. ZOOL, “UNI LIBRARY sila JUL 16 970 OCCASIONAL PAPERS HARVARD 3 UNIVERSITY of the MUSEUM OF NATURAL HISTORY The University of Kansas Lawrence, Kansas NUMBER 81, PAGES 1-10 JULY 1, 1979 NOTES ON AN IMPORTANT NINETEENTH CENTURY COLLECTION OF CENTRAL AND NORTH AMERICAN BIRDS MADE BY N.- S. GOSS By MARION ANNE JENKINSON AND ROBERT M. MENGEL' In the 1870's and 1880’s Colonel Nathaniel S. Goss, of Neosho Falls, Kansas, accumulated a large collection of mounted birds (2,500 specimens are listed in his catalogues), mostly through his own collecting efforts. His aim seems to have been to prepare a lifelike mount of a male and female of each species of North American bird. To this end, Col. Goss travelled widely in the United States, to some extent in Canada, in Mexico, Honduras, British Honduras, Guatemala, Nicaragua, and Costa Rica. The fate of his specimens was varied. Most of the specimens in Goss’s catalogues are marked with a “C,”. which evidently indi- cated that the specimen was admitted to “the collection.” About 30 conventional study skins were stated to have been sent to the United States National Museum, and we have verified that all of these are at that museum now. A fair number of entries lack one or the other of these designations and we assume that Goss dis- carded these; almost without exception they are not now present. The bulk of the collection eventually became the property of the Kansas State Historical Society in Topeka (see Janes, 1964, for their history at that institution). An unknown number of specimens was given to Washburn University, Topeka, in 1949. 1The Museum of Natural History (and for RMM also The Department of Systematics and Ecology), The University of Kansas, Lawrence, Kansas 66045. 2 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY All of the Goss specimens, save the study skins, have been on exhibit for nearly 100 years and some are badly faded. In 1977, the approximately 1,350 specimens still at the Kansas State Historical Society's Museum were given to The University of Kansas Museum of Natural History. We are planning to remake these specimens (except for the type specimens) into conventional study skins insofar as possible. With the birds were three leather- bound catalogues compiled by Goss and containing, in addition to the specimen entries, nesting data, many field notes, and miscel- laneous information. Many, but not all, of the specimens housed at Washburn Uni- versity were destroyed by the great Topeka tornado. of 8 June 1966. We recently visited Washburn University and examined the approximately 135 mounted birds that remain there. Of these, § were particularly important specimens, and these were generously given to the K.U. Museum of Natural History. Goss’s catalogues indicate that he also collected a large num- ber of eggs. He gave many of these to his brother Benjamin F. Goss, whose egg collection is now housed at the Milwaukee Public Museum in Wisconsin. We do not know the whereabouts of whatever eggs Col. Goss retained; if they were given to Washburn University at some time, they are not there now. Summarizing, we have now a fairly complete conception of what Goss specimens are extant and where they are located. The identification of all of the specimens known to exist (other than eggs) has been verified either by us or by personnel at the U.S. National Museum. In 1899, D. E. Lantz published a list of birds which Goss had collected in Mexico and Central America. Various students of the Central American avifauna have referred to Lantz’s list, but seem- ingly none has examined the specimens. Lantz gave only species names and localities, not dates of collection and frequently -not sexes. The list was uncritical (Goss’s identifications were accepted without question) and contains many typographical and substan- tive errors. The large number of the latter may give the impression that Goss was not a careful worker; his catalogues and contempo- raries (see, for example, Allen, 1891) suggest otherwise. Among the 256 forms listed by Lantz, we have found only 12 errors of identification at the specific level, a very creditable performance for the times. For the species for which no specimens seem to be extant, in most cases we see no reason to doubt the identification. KANSAS Goss collected over 700 specimens in Kansas, mostly near Neosho Falls, Woodson County, and these formed the basis of his NOTES ON N. S. GOSS BIRD COLLECTION 3 book on the birds of Kansas (Goss, 1891). Most of the critical specimens from Kansas are extant and are at The University of Kansas Museum of Natural History—for example, breeding Black- capped Vireos (Vireo atricapillus), breeding Peregrine Falcons (Falco peregrinus), first state record of Sabine’s Gull (Xema sabini), ete. MEXICO Goss collected about 175 specimens in Mexico. Lantz gave only village or city names for the localities, but through study of the dates on the specimens, notes in the catalogues, and letters from Goss to his family, we have been able to identify the princi- pal collecting sites. The coordinates we give are approximate, but most sites can be found on railways. Goss was a railroad business- man, and a fair number of his collecting activities seem to have been facilitated by that fact. Some of the localities are easily found on most maps, but we mention all here for completeness. In Baja California, Goss col- lected on the Coronados Islands (32° 28’N, 117° 16’W), San Pedro Mantirlisles(28> 22)N. 2" SAW ), ua Paz (24° 30°N, 1102 27W ), and Tijuana (32° 29’N, 117° 10’W). Lantz listed a Merlin (Falco columbarius) as being from “Lower Califomia.” Goss’s catalogue indicates that this was shot by a C. H. Goodrich, on 25 March 1884, on the coast about 50 miles south of San Diego. Three collecting sites were in Sinaloa—Culiacan (24° 49’N, 107° 24’W), Limoncito (24° 46’N, 107° 45’W), and Altata (24° 38’N, 107° 57’°W); and two were in Veracruz—Coatepec (19° 29’N, 96° 59’W) and Rinconada (19° 22’N, 96° 34’W). Goss also collected briefly at Ciudad Lerdo, Durango (25° 33’N, 103° 30’W), “the end of the track on the Mexican Central R.W.” (Goss, in litt., 1883). All of these localities can be found on The American Geo- graphical Society’s “Map of Hispanic America” series. Just before his arrival at Lerdo, Goss collected at “Florido.” We think that he probably was at or near “Estacién Florido,” Chihuahua, at 27° 36’N,105° 03’W (shown on Carta Topografica G13A29, Comision de Estudios del Territorio Nacional). Goss seems to have made no errors of identification, at the species level, for the birds he collected in Mexico and the locality records represented by the specimens are not remarkable. Only one error made by Lantz is worthy of note. The specimen of the Wedge-billed Woodcreeper (Glyphorhynchus spirurus) that Lantz said was from Coatepec, Veracruz, was in fact from Costa Rica. Miller et al. (1957: 46), state that there are several records from Veracruz for that species; Goss’s specimen should no longer be considered as documenting one of those records. 4 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY The present location of the type specimens which Goss col- lected in Mexico is discussed later. GUATEMALA Goss visited Guatemala twice and collected about 200 speci- mens there. Griscom (1932: 411) stated that Lantz’s list contained “numerous obvious errors of identification” and that he accordingly “ignored all records in [that] paper referring to critical groups.” Most of the records we discuss are from Santo Tomas, “a hamlet on the east coast ten miles southeast of Livingston and due south of Puerto Barrios” (Griscom, 1932: 423). Goss misidentified the oriole Icterus dominicensis prosthemelas, common enough in Guatemala, as I. graduacauda. His specimens, a pair from Santo Tomas, had represented the only known “record” of I. graduacauda from Guatemala. The male is now K.U. 72723. The female appears to be lost, but we think it must certainly be assumed that she too was misidentified. Griscom (1932) ignored the record. It was, however, accepted by Hellmayr (1937: 143) who mistakenly stated that Santo Tomas was in “extreme n6rth- western Guatemala” (italics ours). Most later workers (e.g., Blake, 1968: 165) followed Hellmayr uncritically. Land, however (1970: 313), questioned Goss’s identification and pointed out (again) the correct location of Santo Tomas on the Caribbean coast. Although the mistake is corrected in Goss’s catalogue, Goss mislabelled a male Tangara larvata (K.U. 72822) taken at Santo Tomas as Tangara lavinia. So reported by Lantz (1899: 223), this “record” was accepted by Griscom (1932: 375) and seems to be the basis. for the inclusion by subsequent workers (e.g., Storer, 1970: 374; Land, 1970: 322; Peterson and Chalif, 1973: 230) of Guatemala in the range of T. lavinia. Correction of this error seems to leave La Ceiba, Honduras (see Monroe, 1968: 358-359), as the northernmost, reliable record for this tanager. Also from Santo Tomas are two specimens of the Bare-crowned Antbird (Gymnocichla nudiceps)—a male: (K.U. 72304) taken 2 February 1886 and a female (K.U. 72305) taken 18 January 1886. Goss misidentified these as Drymophila trifasciata (a synonym of Pyriglena leucoptera of South America). Gymnocichla nudiceps is rare in Guatemala and difficult to see in the dense brush (Land, 1970; 199) and apparently the species has been taken only a few times north of Nicaragua (see Griscom, 1932: 236; Monroe, 1968: 238; and Russell, 1964: 108). The following specimens, which were misidentified by Goss (and consequently by Lantz), are all at The University of Kansas Museum. A male and female Ramphastos sulfuratus (misidenti- fied as R. ariel) from Santo Tomas; a female Xiphorhynchus flavi- NOTES ON N. S. GOSS BIRD COLLECTION 5 gaster from Santo Tomas and a male of the same species from Naranjo (misidentified as X. guttatus and Lepidocolaptes souleyetti, respectively), a female Dendrocolaptes certhia from the Chocon River in eastern Guatemala (misidentified as X. flavigaster); a male and female Pipra mentalis (called Antilophia galeata) from Santo Tomas, and a male Oryzoborus angolensis funereus from the Cho- cén River (called Sporophila americana). (The Chocon River is at approximately 15° 45/N, 89° 00’W.) Griscom (1932: 164), commenting on the status of the Bat Falcon (Falco rufigularis) in Guatemala, stated that “it is un- doubtedly pure chance that this species has not been collected in the Caribbean lowlands of Guatemala.” This apparently was a lapsus on Griscom’s part, for he also mentioned a bird taken, by Goss, in Santo Tomas (which is in the Caribbean lowlands). This specimen (K.U. 71610) is a female taken 11 January 1886. Goss especially prized his pair of Ocellated Turkeys (Agrio- charis ocellata) from Yaxha, Guatemala (according to Goss’s cata- logue “about 40 miles west of Cayo, British Honduras”). Unfortu- nately these are now missing and may have been stolen from the collection when it was housed at the State Capitol (Janes, 1964: 11). Goss stated that he also collected two eggs from the nest of the pair, on 11 March 1887. BELIZE ( British HonpurRAS ) Several of the 35 specimens collected by Goss in British Hon- duras (now Belize) represent important records. Russell (1964) mentioned these, but because he thought the collection was no longer extant, he relied on Lantz’s list, which lacked dates of col- lection. Russell (p. 88) considered Goss’s specimen of Anthrocothorax prevostii from Isabella as being a noteworthy locality record. The specimen, a male, was taken on 23 March 1887 and is now K.U. 72282. : Goss collected a male Heliothryx barroti (K.U. 72284) at Cayo on 10 March 1887. Apparently there are only a few records for that hummingbird from the country (Russell, 1964: 90). Two specimens of the Bananaquit (Coereba flaveola) in the British Museum were, at the time Russell published, the only known extant specimens of that species from British Honduras (Russell, 1964: 149-150) and they lacked data. Russell expressed skepticism about the validity of the Goss specimens which Lantz listed. However, the two specimens are extant and were correctly identified. They are a male (K.U. 72618) taken on 6 March 1887 and a female (K.U. 72617) taken on 14 March 1887, both from Cayo. 6 NOTES ON N. S. GOSS BIRD COLLECTION HONDURAS Monroe (1968) cited many of Goss’s 20 Honduran specimens, as listed in Lantz, but he did not know the dates of collection of the specimens. Several are worthy of note. A female Bicolored Antbird (Gymnopithys bicolor) taken 31 January 1887 at Choloma was the first Honduran specimen of that species (Monroe, 1968: 239). It is now K.U. 72303. Goss misidentified a specimen of the woodcreeper Dendro- colaptes certhia as Xiphorhynchus flavigaster. The specimen (K.U. 72292) is a male taken at Choloma on 29 January 1887. Goss collected a female Scaly-throated Leafscraper (Sclerurus guatemalensis) at Choloma on 5 February 1887. Monroe (1968: 232), who regarded this species as a rare to uncommon resident in the area, mentioned this specimen but gave no date for it. It is now K.U. 72298. Monroe (p. 243) considered Schiffornis turdinus, the Thrush- like Manakin, as being decidedly uncommon in Honduras. He listed Goss’s specimen as the only- one from Choloma. The speci- men (K.U. 72320) is a male taken 30 January 1887. NICARAGUA Goss spent a few weeks in December, 1889, and January, 1890, in Nicaragua, where he collected about 20 specimens. A critical specimen which is lost was identified by Goss as Anthrocothorax nigricollis. The northern limit of that humming- bird’s known distribution is in central Panama (Wetmore, 1968: 296). The Nicaraguan specimen was from Granada, and Goss stated in his notes that it “may be an immature [Anthrocothorax prevostii], or a southern race of that species.” He had collected an adult male A. prevostii, which he correctly identified, at the same place only one day earlier. L Goss identified an antbird collected at “Los Sabalos” (which we take to be Sabalos on the southern shore of Lake Nicaragua) as Pyriglena leucoptera. This is undoubtedly an incorrect identi- fication, but the specimen is no longer extant so we cannot determine what it was. However, the two specimens from Guatemala (see earlier comments) which Goss called P. lewcoptera are Gymno- cichla nudiceps; it is likely that the Nicaraguan specimen belonged to that species also. Costa Rica In the first week of January, 1890, Goss collected about 20 specimens in what he called the “San Juan Valley, Costa Rica.” We cannot determine exactly where he was in the valley. Just prior to this he was at Sabalos, Nicaragua. Thus his specimens may have NOTES ON N. S. GOSS BIRD COLLECTION it been from the western end of the valley and possibly from Nica- ragua rather than Costa Rica. However, Goss (in litt., 1889) stated: “think before I stop [at Greytown on the Caribbean coast, I] shall run up the San Juan River about 100 miles, into the hills where the water is pure and the climate healthy,” which would be in Costa Rica. At least one specimen, a woodcreeper, from the San Juan Valley is noteworthy. A male Xiphorhynchus flavigaster (K.U. 72296) was taken on 31 December 1889. Goss had misidentified this as X. guttatus. Slud (1964: 199) indicated that X. flavigaster nor- mally occurs along the forested slopes of the Guanacaste Cordillera, between 400 and 3,000 feet above sea level, but that once he had encountered it at Los Chiles on the Caribbean slope just south of Lake Nicaragua. EXTINCT AND ENDANGERED SPECIES In 1963, Hahn listed institutional holdings of all known speci- mens of various extinct and endangered bird species. His desig- nation of the institutions in Topeka, Kansas (pp. 54-55), is some- what confusing, but in fact the specimens listed are those that were housed at the State Historical Society's Museum and at Washburn University. As a result of the events described herein, the present status of these specimens is as follows: one Passenger Pigeon (Ectopistes migratorius) Kansas, three Eskimo Curlews (Numenius borealis ) two Kansas and one Texas, one Ivory-billed Woodpecker* (Cam- pephilus principalis) Arkansas, and two Whooping Cranes (Grus americana) Kansas, have been added to The University of Kansas Museum of Natural History holdings (as listed by Hahn, 1963: 42). One Carolina Parakeet (Conuropsis carolinensis), which is data- less, is still at Washburn University; lost (probably in the tornado) are one Ivory-billed Woodpecker, two Whooping Cranes, and two Passenger Pigeons. The two Whooping Crane mounts (a male and female) and one Eskimo Curlew mount (a male) have been sacrificed so we could remove the skeletal elements. Skeletal elements for these species, especially from birds of known sex, are rare in collections. Type SPECIMENS Goss (1888) described as new species two forms of boobies. One of these was Sula gossi, now known to be a synonym of S, nebouxii. (Goss erroneously thought that his friend Robert Ridg- way had already formally designated the form as “gossi,” but that was only in manuscript.) The second was Sula brewsteri (=S. leucogaster brewsteri). Both were from San Pedro Martir Isle, for 8 NOTES ON N. S. GOSS BIRD COLLECTION which coordinates have been given earlier. In each instance, Goss’s type series consisted of a male and female mount and a male and female study skin. All specimens are extant; the four skins are at the U. S. National Museum (Deignan, 1961: 18, 20) and the four mounts are now at the University of Kansas Museum. The syntypes of Sula gossi that are at the University of Kansas are the first two listed in Goss’s table of measurements (1888: 242), a female (K.U. 71382) and a male (K.U. 71381), respectively. The third set of measurements is from the female skin (USNM 113435) ‘and the fourth from the male skin (USNM 113434). Rich- ard Zusi, Curator of Birds at the U. S. National Museum, has told us (pers. comm.) that both the USNM catalogue and the speci- men label indicate that the male (USNM 113434) is the “type” (=lectotype). However, we can find no place in the literature where a lectotype was designated from the type series. It is possi- ble that Ridgway (or some other person) intended to designate this specimen as the lectotype but failed to do so. The four specimens of Sula brewsteri listed in Goss’s table of measurements (1888: 243) are a male (K.U. 71383), a female (K.U. 71384), a male (USNM 113436), and a female (USNM 113487), respectively. Ridgway (1897: 597) designated USNM 113436 as the lectotype of S. brewsteri. Friedmann et al. (1950: 23) were mistaken in stating that the type of S. leucogaster brewsteri was at the University of Kansas; at that time none of the specimens was so located, although two paralectotypes now are. There are several errors concerning the published measure- ments of the lectotype and paralectotypes of S. brewsteri. Ridg- ways measurements (1897: 598) are not the same as those given by Goss (1888: 243) for either the male or female skins. Ridgway clearly gave incorrect measurements for the lectotype, a male. The measurements he listed are identical to those he gave for -the female in his very next paragraph, although Goss (1888: 243) stated that “the females in all cases were the largest.” Furthermore, there are three measurements published by Goss that are different from those listed in his own catalogue. The catalogue (which gives “fresh measurements”), reading down in the same order as Goss’s publication, gives “stretch of wing” figures of 56.50, 59.00, 56.50, and 59.50 inches; and “wing” measurements of 14.35, 15.60, 14.40, and 15.60 (discrepancies italicized ). ACKNOWLEDGMENTS We are deeply indebted to Mr. Edgar Langsdorf, Executive Director of the Kansas State Historical Society, and Mr. Mark A. Munt, Assistant Director of the Society's Museum whose efforts made the transfer of the Goss Collection possible. Likewise, Dr. OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 9 Donald R. Boyer, Professor of Biology at Washburn University, cooperated effectively in the matters of study and transfer of the specimens in his care. A considerable number of persons at The University of Kansas Museum of Natural History assisted in various degrees in moving the specimens; Mr. Robert R. Patterson deserves special thanks for his managing the logistics of the move. Dr. Donald W. Janes, at the University of Southern Colorado, was most helpful in answering our questions about archival ma- terial regarding N. S. Goss, and Mr. Lewis Armstrong, Curator of the University of Kansas’s Map Collection, provided much as- sistance to us. Dr. Merlin D. Tuttle, of the Milwaukee Public Museum, wrote to us about the egg collection at that museum. We especially burdened Dr. Richard Zusi with questions about the Goss materials now under his care at the U. S. National Mu- seum and thank him for his assistance. Mr. and Mrs. Robert T. Drake (Martha Swan Drake is a grand- niece of N. S. Goss), of Winnetka, Illinois, donated to The Uni- versity of Kansas all of the letters, diaries, and other documents in their care which relate to the Goss family. These now reside in Spencer Library as a part of The Kansas Collection. All persons interested in the early history of Kansas owe a large debt of grati- tude to Mr. and Mrs. Drake. Finally, we wish to thank Mr. Eugene Eisenmann, of the American Museum of Natural History, and Dr. Kenneth C. Parkes, of the Carnegie Museum, for their careful reading of the MS of this paper and their many helpful suggestions for its improvement. Taxidermic work on the specimens and storage of them at The University of Kansas are made possible by National Science Foun- dation grant number DEB78-04832. LITERATURE CITED ALLEN, J. A. 1891. [Obituary of N. S. Goss.] Auk, 8: 245-247. BuakeE, E. R. 1968. Family Icteridae. Pp. 138-202 in Check-list of birds of the world. Vol. XIV (R. A. Paynter, Jr., ed.). Mus. Comp. Zool., Cam- bridge, Massachusetts. 433 pp. Deicnan, H. G. 1961. Type specimens of birds in the United States National Museum. U.S. Natl. Mus., Bull. 221, 718 pp. FRIEDMANN, H., L. Griscom, and R. T. Moore. 1950. Distributional check- list of the birds of Mexico. Part I. Pacific Coast Avifauna, No. 29. 202 pp. Goss, N. S. 1883. [Letter to T. L. and Sarah Clark. 19 November 1883.] The University of Kansas Libraries, Kansas Collection. Goss, N. S. 1888. New and rare birds found breeding on the San Pedro Martir Isle. Auk, 5: 240-244. Goss, N. S. 1889. [Letter to T. L. and Sarah Clark. 27 November 1889.] The University of Kansas Libraries, Kansas Collection. Goss, N. S. 1891. History of the birds of Kansas. Geo. W. Crane & Co., Topeka, Kansas. 692 pp. 10 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Griscom, L. 1932. The distribution of bird-life in Guatemala. Bull. Amer. Mus. Nat. Hist., Vol. 64. 439 pp. Haun, P. 1963. Where is that vanished bird? Royal Ontario Museum, Uni- versity of Toronto. 347 pp. Hetitmayr, C. E. 1937. Catalogue of birds of the Americas and the adjacent islands. Part X, Icteridae. Zool. Ser., Field Mus. Nat. Hist. Vol. XIII. 228 pp. Janes, D. W. 1964. The Goss Ornithological Collection—a centennial. Kan- sas Om. Soc. Bull., 15: 9-12. Lanp, H. C. 1970. Birds of Guatemala. Livingston Publ. Co., Wynnewood, Pennsylvania. 381 pp. Lantz, D. E. 1899. A list of birds collected by Col. N. S. Goss in Mexico and Central America. Trans. Kansas Acad. Sci. (1897-1898), 16: 218-224. Minter, A. H., H. FRIEDMANN, L. Griscom, and R. T. Moors. 1957. Distri- butional check-list of the birds of Mexico. Part II. Pacific Coast Avi- fauna, no. 33. 436 pp. Monroe, B. L., Jr. 1968. A distributional survey of the birds of Honduras. Omithological Monographs, No. 7. 458 pp. Peterson, R. T., and E. L. Cuauir. 1973. A field guide to Mexican birds. Houghton Mifflin Co., Boston. 298 pp. Ripcway, R. 1897. Birds of the Galapagos Archipelago. Proc. U.S. Natl. Mus., Vol. XIX, No. 1116, pp. 459-670. a Russecyi, S. M. 1964. A distributional study of the birds of British Honduras. Omnithological Monographs No. 1, 195 pp. Stup, P. 1964. The birds of Costa Rica / Distribution and ecology. Bull. Amer. Mus. Nat. Hist., vol. 128, 430 pp. Storer, R. W. 1970. Subfamily Thraupinae. Pp. 246-408 in Check-list of birds of the world. Vol. XIII (R. A. Paynter, Jr., ed.). Mus. Comp. Zool., Cambridge, Massachusetts. Wetmore, A. 1968. The birds of Panama. Part 2. Smithsonian Misc. Colls., Vol. 150, Pt. 2. 605 pp. ° Unfortunately, this and several other, less significant, specimens were lost in shipment to the taxidermist. UNIVERSITY OF KANSAS PUBLICATIONS MUSEUM OF NATURAL HISTORY The University of Kansas Publications, Museum of Natural History, beginning with volume 1 in 1946, was discontinued with volume 20 in 1971. Shorter research papers formerly pub- lished in the above series are now published as Occasional Papers, Museum of Natural History. The Miscellaneous Publica- tions, Museum of Natural History, began with number 1 in 1946. Longer research papers are published in that series. Monographs of the Museum of Natural History were initiated in 1970. All manuscripts are subject to critical review by intra- and extramural specialists; final acceptance is at the discretion of the publications committee. — Institutional libraries interested in exchanging publications may obtain the Occasional Papers and Miscellaneous Publica- tions by addressing the Exchange Librarian, The University of Kansas Library, Lawrence, Kansas 66045. Individuals may pur- chase separate numbers of all series. Prices may be obtained upon request addressed to Publications Secretary, Museum of Natural History, The University of Kansas, Lawrence, Kansas 66045. Editor: Ev Witry PRINTED BY UNIVERSITY OF KANSAS PRINTING SERVICE LAWRENCE, KANSAS f Mie SS. GOVT, sae’ hea, aie8 JUL 16 1979 HARVARD. WNIY BReSity OCCASIONAL PAPERS of the MUSEUM OF NATURAL HISTORY The University of Kansas Lawrence, Kansas NUMBER 82, PAGES 1-11 (OILY &, UST® REPLICATION OF HABITAT PROFILES FOR BIRDS By RICHARD F.. JOHNSTON" ABSTRACT The habitat profile, which is based on ten quantitative vegetational varia- bles of an area 0.1 acre (0.04 hectare) in size, has become a standard for descriptive and comparative studies of avian ecology. The technique was used to gather information on habitat structure of avian species of woodland and associated plant communities in eastern Kansas in 1976 and 1977. Because precipitation was greater in 1977 than 1976 and because field work in wood- lands was proportionately more in 1977 than in 1976, the possibility that the two seasonal data sets were not wholly commensurable was examined. Uni- variate and multivariate statistical techniques were employed to seek differ- ences and similarities. In spite of the seasonal and experimental perturbations the two data sets proved to be statistically homogeneous. The technique is thus judged to be robust; it is not likely to generate two different data sets from one avian community subject only to seasonal and experimental variation. (Avian com- munities; woodland habitats; eastern Kansas; principal component, discrimi- nant, and cluster analytic techniques. ) One of the most useful techniques for the objective study of habitats of birds is that of the habitat profile (James, 1971). At one level the routine can be employed to describe important aspects of structural habitat for birds, and at another the data can be used in a comparative or analytical way (Whitmore, 1975; John- ston, 1977). The degree to which the technique can be used re- liably between seasons, or localities, has not yet been thoroughly explored, however. The present report is a contribution toward ‘Museum of Natural History and Department of Systematics and Ecology, The University of Kansas, Lawrence, KS 66045. 2 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY assessing replicatability at one locality between successive years, based on a deliberate sampling perturbation. The technique was employed at The University of Kansas Nelson Experimental Area in documenting habitat relationships of woodland breeding birds in 1976 and 1977. In 1976 353 records of structural habitat were obtained on 35 species of birds. The rec- ords were compiled in the order in which we encountered the birds in the field. No individual was omitted as a source of data, no individual was preferentially sought for any reason, and 59% were woodland birds. In 1977 213 records were taken on 39 species of birds; the species of woodland habitats were expressly sought, and some individuals were stalked for long periods of time before a profile was secured; 63% of the sample was of woodland birds. Many individuals of common woodland birds, as well as those of edge and grassland habitats, were passed over in the course of searching for the less common woodland species. The two field sessions were therefore different in their em- phases, allowing the possibility that the two data sets might be sufficiently different from one another to preclude their amalgama- tion in the simple way that had been planned. The studies de- scribed briefly beyond were undertaken to determine whether (and then how) the data sets showed significant differences from one another. MATERIALS AND METHODS Basic data matrices—Ten variables of vegetational structure within a 0.1 acre (0.04 ha) plot defined by an individual bird were recorded in both field sessions. The variables (Table 1) are mostly of the woody vegetation, since its character determines the physi- ogngmy of the whole. Localities and dates——Most records were obtained from field work at the Nelson Environmental Study Area, 8 miles NE Law- rence, in Jefferson County, Kansas. Woody vegetation occurs chiefly in ravines and along watercourses. Red oak is the most abundant tree species, but there are some 20 species that are fairly regular in occurrence (Johnston, 1977). Woodlands are sur- rounded by prairie grasslands or agricultural fields, and join these characteristically with a shrubby edge dominated by a dogwood (Cornus drummondii) and a great number of saplings of the dominant tree species. Data were recorded from 4 June to 2 July 1976 and from 13 June to 6 July 1977. Variation in sampling effort.—The sampling technique was em- ployed invariantly in the two year period, but the choice of species sampled differed. In 1976 we secured a habitat profile on every individual bird we sighted in the course of a day’s field work. In 1977 we tended to ignore common species, especially those of edge habitats, and concentrated on the less common species of wood- REPLICATION OF HABITAT PROFILES FOR BIRDS 3 TasLe 1.—Importance of Variables Defining Factor I in Principal Component Analyses of Habitat Data (Nelson Experimental Area, 1976 and 1977). Factor Loadings Variables 1976" 1977° RemGentaGroundeGover o..2... 22) kW ei es Sou i —.561 Nimler of Slnmily Sieins 1 = 347 —.409 Nima erkOtelinee sSPCClES) =1. 2.26 ee heey .908 .908 BewmmGent Canopy, (Cover 2.22 oe = gos .940 CanopyglciGhien a atin ne ONS a ae miny eas 822 .906 INumbereotinees! G-07 DBEb 222 ee a ee Will .660 INiaimloer Or IWnees G29” IDIBAl .791 692, Nuimloee ot Wiress Qa" IDIBIBI .708 .647 iNiwmmlovere Or nexes MSI” IDS) 2 53 588 Number of Trees 15” or More DBH __...... 408 519 * DBH means diameter at breast height. = INh== SS “IN| = DIB. land habitats (such as red-eyed vireo, parula warbler, and wood thrush). Variation in the avifauna.—Composition of the avifauna of the Nelson area, including species of woodland, edge, and grassland, was nearly the same between the two years. The woodland spe- cies were markedly similar (Table 2) and achieved very high between-year Spearman rank-order correlation coefficients, in which both occurrence and relative abundance were assessed. This is, however, not particularly surprising, for woodlands as distant as those of northeastern Arkansas have avifaunas that achieve a rank-order correlation of r, = 0.908 with those of woodlands in eastern Kansas. Variation in climate.—Northeastern Kansas receives an average of 35” of precipitation per year. In 1976 the total was 23”, and this was the second consecutive dry year. In 1977 the total was ol’. Rainfall totals for June are similar in the two years, but the great difference in over-all rainfall regimes was enough to make all woodland habitats wetter in 1977, than in 1976. Also, a killing frost in early May, 1976, caused shagbark hickory, black walnut, and sycamore trees to lose their leaves. New foliage was not wholly in place for sycamores until July, and tree canopies were not so complete in 1976 as 1977. Biometrical notes —The comparisons between the two samples are by means of principal component analysis (PCA), multiple discriminant analysis (MDA) (Cooley and Lohnes, 1971; Black- ith and Reyment, 1971), and cluster analysis (Sneath and Sokal, 1973). With PCA the records for each year are processed inde- pendently for each sample. The contributions of each variable to the first new axis, and the per cent of the trace of variance 4 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY TasLe 2.—Abundance of Woodland Birds in the Nelson Experimental Area in 1976 and 1977. Number of Pairs Species 1976 1977 Brown-headed Cowbird, Molothrus ater... 38 26 Bine Jay, Cyanocitta cristaig: 2 ee 20 21 Yellow-billed Cuckoo, Coccyzus americanus —... Saka kee 16 21 Cardinal, Cardinalis cardinalis DRMND EA Teeton ne ee to 20 Red-bellied Woodpecker, Centurus co a ee ll. 16 13 Black-capped Chickadee, Parus atricapillus — 17 ll Tufted Titmouse, Parus bicolor Piseee ee ee Ba eG 9 Northern Oniole; feterus ealbula =|} fe eee 13 9 White-breasted Nuthatch, Sitta carolinensis COG 14 Eastern Wood Pewee, Contopus virens _........---------..------ 10 ( Downy Woodpecker, Picoides pubescens _....................-.. 4 13 Kentucky Warbler, Oporornis formosus _.......----..-..---------------- 11 4 Crested Flycatcher, Myiarchus crinitus _2...............-.... 5 9 Mourning Dove, Zenaidura macroura .........-......---------0-----0----- 7 6 Wood Thrush, Hylocichla musteling 240 2S 4 7 Rose-breasted Grosbeak, Pheucticus ludovicianus — 6 4 Rufous-sided Towhee, Pipilo erythopthaimus —. begs ids 3 Bi Red-headed Woodpecker, Melanerpes erythrocephalus AE ae 6 Scarlet Tanager, Piranga olivacea —...... fee hee 2s Cee 3 5 Red-eyed Vireo, Vireo olitvaceus 2.222. 02--221---22--- een : 2 3 Parula Warbler, Parula americana __.........-.-----.------ ul 1 Red-tailed Hawk, Buteo jamaicensis 2. il il Hairy Woodpecker, Picoides villosus fs 0 2 Carolina Wren, Thryothorus ludovicianus il 0 American Robin, Turdus migratorius — 0 1 SPECIES TOTAL 23 24 represented in the first three axes, are assessed. With MDA the total records of individuals or of species for cach year compose the input sets to be discriminated; the level of discrimination can be judged by means of the Mahalanobis distance statistic. Similar assessments for individual species provide additional information on degree of between-year similarity. Cluster analysis was used firstly to test the degree to which grassland and woodland species could be separated from one another, and secondly, the degree to which data for 1977 resembled those for 1976. ResuLts Component analysis of the records.—Table 1 presents the factor loadings of the ten variables onto the first principal component of variation over all specimen records for each of the yearly data sets. Note the contrast in signs of the herbaceous and shrub variables with those of the tree variables is found in both samples. Table 3 shows the similarity between 1976 and 1977 in information present in PCs I-III. REPLICATION OF HABITAT PROFILES FOR BIRDS 5 TABLE 3.—Information Content of The First Three New Axes of Variation in Principal Component Analyses of Habitat Data (Nelson Experimental Area, 1976 and 1977). Per Cent of Trace of Variance Axis 1976" 1977° Pamerpaloumonent wl) = 2.2 2 ce 45.4 49.6 EnmewpalsConponent: Milly se ee Se ee ee 11.4 IEG st Principal m@onmponemtsill HU see os sr os os ye 9.6 8.7 sco tall tae eesti at oi eee Re ee 66.4 71.0 *N = 353. IN| = DNB}. Discriminant analysis of the records.—If all habitat records for each year of study are used, with all ten habitat variables employed as discriminators, the set for 1976 cannot be reliably distinguished from that for 1977 (Table 4). On the order of 64% of the records can be correctly classified back to input groups, which is only 14% more than that obtainable at random. The centroids of the two groups of records are separated by four variables, but the distance between the two is D? = 0.52 (ns.). If records for species for which at least four profiles were ac- cumulated are used to obtain mean values for their character-states, some 21 species may be used in a two-group discriminant analysis (Table 5). Around 69% of these species can be correctly classified back to input groups, but the distance between the centroids in this case is still small: D? = 1.0-{n.s.). Use of individual species does not increase the degree to which the yearly samples can be differentiated from each other (Table 6). The Bell vireo, Vireo belli, a grassland species for which yearly differences in habitat or techniques of recording variables should be minimal, proves to have the two yearly samples in fact in- separable. The same is true for the Brown-headed cowbird, Molo- TaBLe 4,—Classification Analysis of Utilized Habitat of 562 Individuals of Birds (Nelson Experimental Area, 1976 and 1977 ).' Number of Individuals Original Classified into Year Samples Per Cent Input Group 1976 1977 Correct TAS IPAS eee metre, Soe AS < eo Re eRe eee 212 141 60.1 WG ro nin A ta ere eR eR oe betel, 68 145 68.1 ‘After entry number four of a stepwise classification analysis using ten quantitative habitat variables as discriminators; the F-ratio of the U-statistic is F = 16.62 (df. 4 and 561); the variables entering are trees larger than 15” DBH, per cent canopy cover, per cent ground cover, and number of tree species. Mahalanobis distance between the centroids of the groups for 1976 apo1o77 is D* = 0.52 (nis.):. 6 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY thrus ater, a species that can be found in every subtype of habitat in which records were taken. Records for the wood thrush, Hy- locichla mustelina, a woodland species, prove to be more nearly different in the two years, but the distance between the yearly controids is D* = 2.16 (n.s.). It is worth noting that differences between years for the wood thrush data approach the significant, and the quality of woodland habitat was. likely to have been in- volved: wood thrushes were found in some places in 1977 that were too dry for them in 1976. Cluster analysis —Relational tree diagrams provide less rigorous assessments of relationships than their antecedent similarity ma- trices (or than a PC analysis), but the information is presented in a visual fashion heuristically useful for our purposes. In the present assessments the cluster elements are bird species and the variables are the ten quantitative vegetational characters. Both correlation and distance matrices for each of the sets for 1976 and 1977 were computed, and six clustering algorithms were employed to depict the relationships between species in cach sample. In each yearly sample distance matrices summarized relation- ships more adequately than correlation matrices, judged by the degree to which grassland species clustered with grassland species and woodland with woodland species. And, for both yearly sam- ples the unweighted pair-group method of clustering, using arith- metic averages (UPGMA), had the largest cophenetic correlation coefficient of the tree diagram back to the distance matrix. The minimum number of species to be clustered was dictated by sample sizes for the sample from 1977; there are 21 species with at least four habitat profiles for 1977, and these 21 were matched by the same species in the data set for 1976. The tree dia- grams are depicted in figs. 1 and 2. As can be seen immediately, there are two major clusters in each of the diagrams—one for grassland species and the other for woodland (plus edge) species. The only species that shows’ inconstancy between years is the TasLe 5.—Classification Analysis of Utilized Habitat of Twenty-one Species of Birds (Nelson Experimental Area, 1976 and 1977).' Number of Species Original Classified into Year Samples Per Cent Input Group 1976 1977 Correct L976 15 6 71 1977 7 14 67 ' After entry number two of a stepwise classification analysis using ten quantitative hapitay variables as discriminators; the F-ratio of the U-statistic is fF = 5,08 (df. 2 and 39); the variables entering are canopy cover and number of trees 12’-15” DBH. Mahalanobis distance between the centroids of the groups for 1976 and 1977 is D* = 1.0 (n.s.). REPLICATION OF HABITAT PROFILES FOR BIRDS ii TasLe 6.—Classification Analysis of Utilized Habitat of Three Species of Birds (Nelson Experimental Area, 1976 and 1977). Number of Individuals Original Classified into Year Samples Per Cent Input Groups 1976 OM Correct Wood Thrush’ MGT} ete elec: ata! eet a ee 6 0 160 TAS hco NA SeRSSETS 30s soot aaa 5 8 61.5 Belen perm Sees s ek Uy ee es be 5 fee 0 Brown-headed Cowbird at ate 0 ‘ After entry number 1 of a stepwise classification analysis using ten quanti- tative variables of habitat as discriminators; the F-ratio of the U-statistic is F = 8.39 (df. 1 and 17); the variable entering is per cent of ground cover. Mahalanobis distance between the two samples is D? =2.16 (n.s.). brown-headed cowbird, about which there are some comments beyond. The horizontal scale of distance coefficients is in each instance most statistically valid toward the right-hand side of the diagram (Sneath and Sokal, 1973); relationships defined by large d-values are appreciably less precise than small ones. Discussion The replication described here is imperfect or suboptimal in that two sources of variation—between yearly summer climate and between yearly sampling effort—are included and therefore pos- sibly confounded. Having noted this, we are nevertheless forced to conclude that the imperfection is of little consequence: there are virtually no significant differences between the samples. PC Analysis—The contrast in signs of the factor loadings pre- sented in Table 1 identifies a simple ecological pattern—the amount or density of ground cover varies inversely with the amount or density of trees and the arboreal canopy. It is important that this relationship is evident in the two samples, because something as basic as this inversity should be reflected in the habitat data. The degree to which the yearly samples parallel each other is also important; the factor loadings for each year have a Spearman rank correlation coefficient of rz = 0.988 (P << 0.01). The information content of the first three components of varia- tion is likewise similar between years, and the difference of 5% greater trace for the data from 1977 is in the expected direction. If a data set is more rather than less homogeneous, a greater pro- portion of its variation should appear in PC I; as was noted above, the data from 1977 had a more homogeneous source than those from 1976. Even so, the difference in trace of variation over the first three components is not significant between years. Discriminant Analysis—The yearly samples are impossible to 8 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY RBWP CRFC RBGB 1976 -UPGMA PWEE CPCC =0.906 ORIO KYWB WNUT WDTH SCTN BJAY SUTN REVR CARD INBN BHCB BVIR YTHR MLRK FLSP DICK GRSP 20.6 174 13.6 101 66 34 0.3 DISTANCE Fic. 1.—Tree diagram of relationships in a matrix of distance coefficients based on ten variables of structural habitat for 21 species of birds sampled in 1976. Abbreviations: RBWP, red-bellied woodpecker; CRFC, crested fly- catcher; RBGB, rose-breasted grosbeak; PWEE, eastern wood pewee; ORIO, northern oriole; KYWB, Kentucky warbler; WNUT, white-breasted nuthatch; WDTH, wood thrush; SCTN, scarlet tanager; BJAY, blue jay; SUTN, sum- mer tanager; REVR, red-eyed vireo; CARD, cardinal; INBN, indigo bunting; BHCB, brown-headed cowbird; BVIR, Bell’s vireo; YTHR, northern yellow- throat; MLRK, eastern meadowlark; FLSP, field sparrow; DICK, dickcissel; GRSP, grasshopper sparrow. distinguish by discriminant analysis. This is one of the most im- portant findings, for discriminant analysis is the most powerful analytic technique employed in the present study. Cluster Analysis.—Of the clustering algorithms examined in the course of the study, the unweighted pair-group method provided tree diagrams having the highest cophenetic correlation coefficients (CPCC; Sneath and Sokal, 1973) back to the species-by-species distance matrix. Generally, the weighted pair-group method and the unweighted centroid method also provided high CPCCs. The REPLICATION OF HABITAT PROFILES FOR BIRDS 9 1977— UPGMA CPCC=0.911 Fle 24 20 16 12 DISTANCE Fic. 2.—Tree diagram of relationships in a matrix of distance coefficients based on ten variables of structural habitat for 21 species of birds sampled in 1977. Abbreviations as in legend to Fig. 1. poorest fits of the diagrams back to the matrices were found using single-linkage and Spearman’s averaging techniques. Most of the clustering techniques agreed in placing grassland species together, reasonably well apart from the woodland and edge cluster, and the only exception to this deserves more than passing comment. In the data from 1976 the brown-headed cow- bird joined the woodland species, pairing with the edge-oriented indigo bunting and joining the others at some distance. Yet, in the data from 1977 the cowbird is unequivocally clustered with grass- land species. The habitat characteristics of the cowbird are notably labile in the breeding season (e.g., Lowther and Johnston, 1978) and have been earlier examined by means of component and dis- criminant analyses (Whitmore, 1975; Johnston, 1977). Thus, the fact that cowbird profiles place the species in a “woodland” cate- gory in 1976 and into a “grassland” category in 1977 is merely another comment on the enormously broad habitat capabilities of 10 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY the species. However, in a cluster analysis based on the data for both years, the cowbird joins with “edge” species (pairing again with the indigo bunting) to form a subcluster in the grassland species cluster. This is approximately how one would have pre- dicted the cowbird would be represented in such a cluster analysis. Conclusions—The multivariate techniques used here to assess the degree of similarity of the two sets of data have provided no basis on which to consider the samples to have fundamental dif- ferences from one another. However, two of the variables differed significantly when tested by t: per cent ground cover (5% more in 1976 profiles) and per cent canopy cover (9% more in 1976 profiles) showed statistical differences at the 0.01% level. These two variables play an important part in providing some degree of separation of the samples in discriminant analysis (see tables 4, 5, and 6), but the samples nevertheless remain similar. Habitat profiles thus provide information about habitat struc- ture of species and groups of species in a robust manner. Neither differences in the manner in which profiles are secured in-the course of field work nor in the seasonal occurrence of climatic variables have been shown to affect the overall similarity of the two data sets here examined. It is clear that the data sets comprise one statistical entity, which means the fact that they were assem- bled in two different years is irrelevant. The most important ele- ment is, therefore, the environmental continuity provided by the complex deciduous woodland and its associated marginal commu- nities; the system is a buffer against radical change. Indeed, de- partures from habitat equilibria characteristic of the birds of these communities are essentially nonexistent. ACKNOWLEDGEMENTS Fieldwork supporting this study was in connection with sum- mer session course Biology 796, Department of Systematics and Ecology, The University of Kansas. John Bucher, Parthenia Evans, Karen Hamrick, Shyurei Kinoshita, Janet’ Lee, Peter Lowther, Christopher Luecke, John Paul, Galen Pittman, Richard Racine, and Lynda Swander assisted in fieldwork and in initial processing of the data. Peter Lowther collaborated in some aspects of data analysis. Parts of the study were funded by a grant from the Gen- eral Research Fund of The University of Kansas, and computa- tions were assisted by funds from the university Computation Cen- ter. Permission to work on the Nelson Environmental Area was granted by The University of Kansas Field Facilities Committee. REPLICATION OF HABITAT PROFILES FOR BIRDS Vi LITERATURE CITED BLACKITH, R. E., REYMENT, R. A. 1971. Multivariate Morphometrics. Aca- demic Press, London and New York. Coo.ey, W. W., Lounes, P. R. 1971. Multivariate Data Analysis. John Wiley & Sons, New York. James, F. C. 1971. Ordinations of habitat relationships among breeding birds. Wilson Bull., 83:215-236. Jounston, R. F. 1977. Composition of woodland bird communities in eastem Kansas. Bull. Kansas Ornith. Soc., 28:13-18. LowrTuer, P. E., JOHNSTON, R. F. 1978. Influences of habitat on cowbird host selection. Bull. Kansas Omith. Soc., 28:36-40. SnEATH, P. H. A., SoxaL, R. R. 1973. Numerical Taxonomy. Freeman, San Francisco. Wuirmorg, R. C. 1975. Habitat ordinations of passerine birds of the Virgin River Valley, southwestern Utah. Wilson Bull., 87:65-74. UNIVERSITY OF KANSAS PUBLICATIONS MUSEUM OF NATURAL HISTORY The University of Kansas Publications, Museum of Natural History, beginning with volume 1 in 1946, was discontinued with volume 20 in 1971. Shorter research papers formerly. pub- lished in the above series are now published as Occasional Papers, Museum of Natural History. The Miscellaneous Publica- tions, Museum of Natural History, began with number 1 in 1946. Longer research papers are published in that series. Monographs of the Museum of Natural History were initiated in 1970. All manuscripts are subject to critical review by intra- and extra-mural specialists; final acceptance is at the discretion of the publications committee. Institutional libraries interested in exchanging publications may obtain the Occasional Papers and Miscellaneous Publica- tions, by addressing the Exchange Librarian, The University of Kansas Library, Lawrence, Kansas 66045. Individuals may pur- chase separate numbers of all series. Prices may be obtained upon request addressed to Publications Secretary, Museum of Natural History, The University of Kansas, Lawrence, Kansas 66045. Editor: Ep Witry PRINTED BY UNIVERSITY OF KANSAS PRINTING SERVICE LAWRENCE, KANSAS iw, SS WIVIF, ZOOL UN LIBRARY lee — AUG 3 O%c OCCASIONAL PAPERS HARVARD af the UNIVERSITY MUSEUM OF NATURAL HISTORY The University of Kansas Lawrence, Kansas NUMBER 83, PAGES 1-56 AUGUST 15, 1979 THE RELATIONSHIPS OF THE AMPHIBERINGIAN MARMOTS (MAMMALIA: SCIURIDAE) By Rosert S. HorFMAnn,! JAMEs W. KoEppt,1 AND CHARLES F’. NADLER 2 Marmots comprise a closely related Holarctic genus (Marmota) of large terrestrial squirrels. Considerable controversy exists over the number of species contained within the genus, and questions of interspecific relationships are unresolved. Several species of mar- mots inhabit arctic and alpine tundra, and to a lesser extent sub- arctic and -alpine habitats on both sides of the Bering Strait, in northeastern Siberia and northwestern North America (Fig. 1). These amphiberingian marmots are montane species primarily adapted to alpine tundra (Hoffman, 1974), although they occur at lower elevations (to near sea level) in suitable habitat at high latitudes (Bee and Hall, 1956; Cowan and Guiguet, 1956; Kapi- tonov, 1960a, 1963). The North American hoary marmot group, as defined by Howell (1915) includes three allospecies. M. caligata (Eschscholtz, 1829) occurs in the northern and central Rocky Mountains, from the Alaska Peninsula, Alaska Range and White Mountains of central Alaska and the Ogilvie Mountains in the Yukon, southward to the Beaverhead and Flint Creek Mountains of northwestern Montana, and the Salmon River Mountains of central Idaho (Davis, 1939; Hoffmann, et al., 1969; Rausch, 1953; Youngman, 1975); it also extends south in the Coast Ranges of British Columbia and Cascade Mountains of Washington (Cowan and Guiguet, 1956; Dalquest, * Museum of Natural History, and Department of Systematics and Ecology, The University of Kansas, Lawrence, Kansas 66045, U. S. A. * Department of Medicine, Northwestern University Medical School, 303 E. Chicago Avenue, Chicago, Illinois 60611, U.S. A. 2 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 1948). The other two species, M. vancouverensis and M. olympus, are montane isolates whose ranges are restricted to the mountains of Vancouver Island and the Olympic Peninsula, respectively (Fig. 1). The only other marmot in this region is the woodchuck, M. monax, not included in this study, whose range nearly spans the North American continent. The black-capped marmot (M. camtschatica Pallas, 1821) of © Siberia is found on the Kamchatka Peninsula, and in mountains east of the Lena River as far south and west as Lake Baikal (Bibikov, 1967; Kapitonov, 1960a; Zimina and Gerasimov, 1973). Marmota caligata and M. camtschatica have been thought closely related (Bikhner’, 1888; Ellerman and Morrison-Scott, 1951; Ognev, 1947; Rausch, 1953; Wehrli, 1935). Hall and Gilmore (1934) described a new marmot from the Brooks Range of northern Alaska, which they named M. caligata broweri. They commented: “. . . it might be maintained with some justice that broweri should be accorded full specific rank... . . However, we think it probable that intergradation will be found to exist between M. c. broweri and M. c. caligata.” Rausch (1953) Azimuthal Equal—Area Projection SCALE MILES 1000 M.caligats Group See Below xO AG DISTRIBUTION OF AMPHIBERINGIAN MARMOTS Fic. 1—Map showing distribution of the amphiberingian marmots, Mar- mota camtschatica and M. broweri (hatched), and the M. caligata group (stip- pled). Approximate ranges of presently recognized species or subspecies are indicated: 1—M. caligata, 2—M. c. oxytona, 3—M. c. okanagana, 4—M. c. nivaria, 5—M. c. cascadensis, B—M. olympus, 7—M. vancouverensis, 8—M. c. raceyi. After Kapitonoy (1960), Cowan and Guiguet (1956), Hoffmann and Pattie (1968), Rausch (1953), and our data. THE RELATIONSHIPS OF AMPHIBERINGIAN MARMOTS 3 noted that the range of broweri was not known to meet caligata in Alaska, but suggested the taxa might intergrade southwest of the Mackenzie River delta; he further suggested that all of the am- phiberingian taxa were conspecific with M. marmota. Later, Rausch and Rausch (1965) decided broweri and caligata were distinct species, since the former had a diploid number of 36 chromosomes, while the latter had 42. They also compared the karyology, ecology, and behavior of M. broweri with M. caligata and M. camtschatica, as well as other members of the genus (Rausch and Rausch, 1971), and concluded M. broweri is more closely related to M. caligata than to M. camtschatica: “... M. broweri [is] probably a relict North American species which be- came established in the Brooks Range during pre-Wirm time, rather than a later Pleistocene invader of middle Asian derivation or, ... a component of the late Pleistocene Amphiberingian fauna” (Rausch and Rausch, op. cit.). The purpose of the present paper is to examine the relationships of M. broweri, M. camtschatica, and the caligata group by comparing various aspects of the biology of the species, with emphasis on cranial morphology. MATERIALS AND METHODS Skins, skulls and skeletons from 719 specimens of the five species in museum collections were studied (see Acknowledgements, and Specimens Examined). Twenty-four cranial and four external body measurements, were taken for samples of each of the named taxa (Table 1). Cranial measurements were made with dial calipers to the nearest 0.1 mm, except for numbers 20-23, which were made with a ruler to the nearest 1 mm. External measurements were transcribed from specimen labels when available. Age variation was minimized by utilizing for statistical analyses only those skulls which possessed fully erupted permanent dentition, and in which sagittal crest development was initiated. The tooth eruption cri- terion is relatively free of subjective bias, and eliminated all mar- mots in their first year of life (juvenile). The second criterion, although more subjective, eliminated yearling marmots, i.e., those in their second year (Petrov, 1961; Poole, 1950). Some measure- ments were less precise than others, but were employed to assess variation in particular aspects of skull size or shape. Examples of measurements of lower precision are sagittal crest length, tym- panic bulla width and length, and anterior and posterior profile indices. Examples of precise measurements are interorbital width, zygomatic width, occipital height, mastoid breadth, and condylo- basal length. Measurements were chosen because they had been employed in previous studies of marmots (Fokanov, 1966; Galkina, 1962; Howell, 1915), thus facilitating comparison with previous 4 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY work, or because they might permit evaluation of certain characters that had been considered of systematic significance; i.e., degree of flattening of cranium (Rausch, 1953). All computations were done at the University of Kansas Com- TasLe 1.—Definitions of cranial and external body measurements employed. A 2. 3. US Length of angular process: mid-point, anterior end of tooth row to posterior end of angular process. Length of articular process: mid-point, anterior end of tooth row to posterior end of articular process. Rostral length: premaxilla to masseteric tubercule surrounding infraorbital foramen. Rostral width: greatest width at maxillary-premaxillary suture. . *Foramen magnum width: greatest width of inside diameter of foramen magnum. . °Foramen magnum height: greatest height of inside diameter of foramen magnum. . *Interorbital width: least width across frontals anterior to postorbital processes. . *Postorbital width: least width across constriction beneath and posterior to postorbital processes. . *Sagittal crest length: lambdoidal crest anterior to point where sagittal crest bifurcates. Tympanic bulla width: greatest width, margin of auditory meatus to basioccipital margin. Tympanic bulla length: greatest antero-posterior length. Maxillary tooth row length: anterior alveolar margin of P3 to posterior alveolar margin of M3. Zygomatic width: greatest width across zygomata. Nasal length: greatest antero-posterior length. Postorbital length: posterior margin of postorbital processes to lamb- doidal crest. Rostral height: greatest height at premaxillary-maxillary suture. Maxillary height: midpoint, alveolus of P4, vertically to dorsal surface of frontal at supraorbital notch. : Occipital height: ventral surface of tympanic bullae, vertically to dorsal surface of sagitto-lambdoidal crest junction. Mastoid width: greatest width across mastoid processes. Depression of nasals: vertical distance from ‘plane of dorsal surface of cranium to anterior end of nasals. Depression of occiput: vertical distance from plane of dorsal surface of cranium to ventral surface of occipital condyles. Anterior profile index: vertical distance from flat surface supporting man- dibles to anterior end of nasals. Posterior profile index: vertical distance from flat surface supporting mandibles to sagitto-lambdoidal crest junction. Condylobasal length: premaxilla to occipital condyles. Total length: recorded from label. Tail length: recorded from label. Body length: recorded from label, or calculated by subtracting tail length from total length. Hind foot length: recorded from label. ° not measured for some specimens at the start of the study. THE RELATIONSHIPS OF AMPHIBERINGIAN MARMOTS 5 putation Center using a Honeywell 66/60 computer; their pro- grams, identified by acronyms BMDP (Dixon, 1975), NTSYS- version ITI, and SFA, were employed. Sometimes an individual’s sex was unknown or certain skull characters could not be measured; in such instances we estimated the missing data. First we determined unknown sex. Preliminary analysis showed significant sexual dimorphism in most marmot groups for many cranial variables. Hence, we lumped approxi- mately equal numbers of known males and females with unknowns, and estimated missing data for this combined sample using pro- gram SFA41D which uses linear regression to predict missing data with the variable most highly correlated with the missing variable. These estimates, intermediate between males and females, were then input as a temporary, but complete, data set into program BMDP7M, a discriminant function (DF) program. This technique in most cases separated males from females. Those few specimens which could not be assigned to a sex with reasonable confidence were not used in further analysis. Having determined sex of all individuals in the sample, we made final estimates of missing character data for each sex sepa- rately, again using SFA41D. These estimates were incorporated into the data sets for analysis. Analysis of sexual dimorphism was then completed with a series of additional DF analyses using BMDP7M on the complete data sets. To study geographic variation in amphiberingian marmots and determine groups for later analysis, we would have preferred to use DF analysis on individuals from defined geographic localities, but because sample sizes for many localities were inadequate, this was not possible. Moreover, we did not wish at the outset to lump individuals from different taxa to increase sample sizes. We there- fore employed separate principal components (PC) analysis (BMDP4M) and UPGMA cluster analysis (NTSYS) on centroids of males and females calculated from individuals from each of 34 localities (Appendix A), reasoning that locality centroids would best represent typical individuals of each locality. In this sort of analysis, groups need not be specified and even a locality repre- sented by a single individual may be included. To determine sys- tematic relationships between taxa we combined locality data based on results of the PC analysis to form our final groups. We performed DF analysis using BMDP7M on males, females, and combined male and female groups. Group centroid discriminant scores were analyzed using UPGMA cluster analysis (NTSYS). Beside mensural characters, qualitative characters were evalu- ated; these included shape of the lachrymal bone, and quality of pelage. Pelage color was compared by reference to Ridgway (1912). 6 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Tape recordings (3.5 ips) made by R. L. Rausch of several calls of captive male and female M. broweri were rendered into narrow band audiospectrograms using a Kay sound spectrograph. A limited number of blood serum samples from Marmota broweri and M. caligata were available for comparison with two M. olympus, and a single captive M. c. camtschatica made avail- able courtesy of the Institute of Cytology and Genetics, Novosi- birsk, USSR. These were examined for transferrin (Tf), albumin, and leucine amino peptidase (LAP), using the electrophoretic methods previously utilized in biochemical studies of the related ground squirrels (Spermophilus) (Nadler, et al., 1974, 1975) and prairie dogs (Cynomys) (Pizzimenti, 1975). RESULTS External Morphology Pelage: Adults—Marmota broweri, M. caligata and M. cam- tschatica are similar in general appearance, being mostly black and white, but with subtle differences in pelage color and pattern. Comparisons are of fresh pelages except where noted. d. Fic. 2.—Pelage patterns of the head and neck in amphiberingian marmots: a, b, M. camtschatica and M. broweri; c, d, M. caligata and M. olympus. THE RELATIONSHIPS OF AMPHIBERINGIAN MARMOTS ve The black-capped marmot, M. camtschatica, is named for the black dorsal surface of its head, which runs uninterrupted from the tip of the nose to behind the ears, and laterally to the level of the eyes ( Fig. 2a,b), and extends back along the neck to the shoulder as an indistinct line. The ear is sandy to orange, sides of the face sandy yellow intermingled with gray, grading into orange on the throat; lips are black. The dorsal coloration of camtschatica gives an overall appearance of black-and-white or black-and-buff. This is due to the presence of three distinct bands on the long, soft, dorsal guard hairs; dark at the base, light in the middle, and dark again at the tips (see Kapitonov, 1964). The anterior dorsal third of the body is lighter just behind the head, becoming darker in overall appearance posteriorly, and with dark pelage extending back along the spine to the base of the tail. Underlying the dorsal guard hairs is a shorter, softer underfur or wool, mostly dark, but with light tips. Ventral guard hairs in camtschatica are tawny to rufous in color, with dark bases and in many cases, dark tips; underfur is lacking ventrally. Feet are sandy to buff. Tail hairs are tawny in the middle, becoming darker subterminally, with lighter tips. Geographic variation in pelage color is relatively slight in camtschatica (Table 2). Southwestern populations (M. c. doppel- mayeri) from the Barguzin Mountains east of Lake Baikal are slightly browner on the cap and dorsal guard hair tips, and become progressively blacker to the north and east (M. c. bungei). Darkest individuals are found in the nominate race on the Kamchatka Peninsula. The middle band on the dorsal guard hairs is buff in doppelmayeri, becoming lighter in bungei, and buff again in cam- tschatica. Ventral guard hairs tend toward tawny in doppelmayeri, and are more rufous in bungei, culminating in deep chestnut and bay in some camtschatica. The hoary marmot (M. caligata) while it is also largely black- and-white (hence its name), differs in that the black dorsal surface of the head is more restricted. Anteriorly it is replaced by a white patch of variable size between the eyes and across the rostrum, sometimes interrupted by a small black spot or band; the tip of the nose is white (Fig. 2c,d). Black of the cap extends posteriorly either as a dark patch behind the ear, or as an indistinct dark band down the neck to the shoulder as in camtschatica. Ears are also black and white, and this grizzled black to brown extends down the sides of the face, becoming rather abruptly nearly white just above the level of the mouth; both lips and throat are whitish (Fig. 2c). Dorsally, M. caligata usually appears lighter than M. camtschatica, although this is subject to considerable variation geo- graphically. This is due to the middle band of the tricolored dorsal guard hairs being relatively wider, and the dark terminal bands 8 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY TasLe 2.—Geographic variation in pelage color of the black-capped marmot, Marmota camtschatica, and the Brooks Range marmot, M. broweri. M.c. M.c. doppelmayeri M. c. bungei camtschatica Dorsal guard Aniline Black Black Black hair tip Dorsal guard Light Buff to Ivory to White Warm Buff; Light hair middle Cartridge Buff Yellow Buff; Ivory Yellow; Cartridge Buff Ventral guard Sanford’s Brown Sanford’s Brown; Sanford’s Brown; hair to Cinnamon Burnt Sienna; Burnt Sienna; Tawny Cinnamon Rufous Bay; Ochraceous Tawny; Chestnut M. broweri: Chandler Lake; “Pt Lay Anaktuvuk Pass Lake Peters Dorsal guard Aniline Black Black Black hair tip Dorsal guard White to Cream White toIvory White 4 hair middle Buff Yellow; Cartridge Buff Ventral guard Black Black Black hair tip Ventral guard White White White hair middle being relatively narrower; additionally, some of the dorsal guard hairs lack dark tips entirely, and are light for their entire length above a dark base. These light-tipped guard hairs are particularly numerous on the neck and across the shoulders to the upper back, giving hoary marmots a “mantled” appearance. Posteriorly, these bicolored light-tipped guard hairs, and the light middle bands of the tricolored ones, become buff as does the distal portion of the underfur, imparting a buff cast to the lower back and rump. Underfur, as in camtschatica, is dark at the base, and light distally, and in texture is fine and soft. However, guard hairs of caligata are longer and much coarser than in camtschatica and this imparts a “harshness” to the pelage in marked contrast to the soft pelage of the black-capped marmot, which supports an important fur industry in northeastern Siberia (Kapitonov, 1963). Ventrally, caligata is sparsely furred, and guard hairs are light, with no trace of the ochraceous color seen in camtschatica; most lack dark tips. Tail hairs are dark basally, light in the middle, and with a dark tip. Another difference is the black feet, the basis of the specific epithet caligata—‘booted.” Geographic variation in pelage color is considerable in M. ca- ligata (Howell, 1915). M. c. caligata, the subspecies occurring in Alaska closest to M. broweri and to Bering Strait, contains popula- THE RELATIONSHIPS OF AMPHIBERINGIAN MARMOTS 9 tions that average among the lighter of the species. Southward along the Rocky Mountains, oxytona is darker, while at the south- ern edge, nivaria is again light; okanagana is intermediate both geographically and in color. Southward in the Coast Range, raceyi, a blackish subspecies, is apparently isolated from caligata or oxy- tona, but may be contiguous southward with the browner cascaden- sis in the Cascade Mountains of Washington (Fig. 1). However the pelage patterns described above are found in all subspecies. The same basic pelage pattern is also found in M. olympus. The tendency toward brownish dorsal color, seen in the geographi- cally adjacent M. caligata cascadensis, is, in olympus, more pro- nounced. Very little pattern can be noted in M. vancouverensis, since it is essentially monochromatic—a uniform dark brown both dorsally and ventrally. Its relationship to the caligata pattern is shown by scattered white-tipped guard hairs on both back and belly, by the white lips, and white patches of variable size on nose, chin, and underparts. Pelage of vancouverensis should be inter- preted as a melanism that has become fixed in this small, isolated population. Interestingly, color differs from that of melanic indi- viduals of M. caligata, which are blacker, and more closely ap- proaches that of melanic M. monax and especially M. caudata dichrous, a melanic population of long-tailed marmots found in northwestern Afghanistan. Marmota broweri was originally described as “a Marmot of the caligata group with black face and white feet. Remainder of col- oration essentially as in Marmota caligata caligata .. .” (Hall and Gilmore, 1934). The authors correctly pointed out differences in face and foot coloration between caligata and broweri, but glossed over other differences to be discussed. As Rausch (1953) first noted, pelage of broweri is soft, rather than harsh as in caligata. Moreover, color pattern of the dorsal guard hairs in broweri— entirely tricolored, with dark basal, light medial, and dark terminal bands—resembles that of camtschatica rather than caligata. Feet are lighter than in caligata, though best described as grizzled or black-and-white rather than “white” (Rausch, 1953, termed them dark gray). The facial pattern is essentially the same in M. cam- tschatica (Fig. 2a,b). The buff to tawny ears contrast with the black cap and sides of the head, though not to the degree usually seen in camtschatica; lips are also black, as in that species, and there is no white on the nose or muzzle. Finally, tail hairs are tawny to ochraceous, becoming darker subterminally, but with a light tip, giving the appearance of a “white apical tip” described by Bee and Hall (1956). Overall, according to R. L. Rausch (pers. comm.), adults in the field appear tri-colored: gray anteriorly, black in mid-body, and red on the rump (see below). Ventrally, overall color of M. broweri appears dark gray, owing 10 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY to the combination of ventral guard hairs with dark bases, light middle bands, and dark tips. This overall ventral coloration is in sharp contrast to the ochraceous color seen in camtschatica, and nearer the grayish-white belly of M. caligata. However, the band- ing pattern of the ventral guard hairs is alike in broweri and camtschatica, and both lack the “all-light” guard hairs found in caligata. Moreover, the ventral fur is, like that of camtschatica, ~ denser than in caligata. Geographic variation in dorsal color ap- pears slight from western to eastern ends of the Brooks Range in the few specimens available (Table 2). A final point concerning pelage color requires clarification. Hall and Gilmore (1934) originally described the “greater average tawniness on hinder back and flanks” of broweri, compared to caligata, and Rausch (1953) described the rump as ranging in color “from something darker than Prout’s Brown to Warm Buff,” and the tail as having some Cinnamon Brown. Bee and Hall (1956) recognized that the “reddish” or “cinnamon color on the hinder back [was] where the older pelage seems still to be in place as it is also on the tail” in two adults captured in early August. This “redness” is due to dorsal guard hair tips fading in older hairs, presumably from exposure to sun, from black to reddish-brown. Molt is initiated in early to mid-summer (differentially depending on age and reproductive activity; see Kapitonov, 1964) on lower back or rump, and spreads anteriorly and ventrally (Davis, 1966; Kapitonov, op. cit.). In some cases, however, pelage on the rump, base of tail, and tail is not replaced during the annual molt, and persists for another year or two (Kapitonoy, op. cit.). Old, faded reddish-brown hair tips form a strong contrast throughout the summer to new black-tipped dorsal pelage anteriorly along the back in both M. broweri and M. camtschatica. In addition to the change in color, this old hair can also be recognized by its altered, coarser texture. The number of unmolted, faded guard hairs in the rump pelage probably increases with age, as a known-age 11- year-old captive male was redder than 5-6 year-old animals (R. L. Rausch, pers. comm. ). Molt in M. caligata may be less regular than in camtschatica (Howell, 1915). However, in this species group also, the rump and tail may retain some hairs. These old hairs also fade, but to a yellowish-brown, or in the case of vancouverensis, light brown color rather than reddish or cinnamon, as broweri and camtschatica. Pelage: Juveniles—Pelages of young individuals, since they differ from those of adults, may also serve as a basis for comparing taxa. Young marmots when they first emerge from the natal bur- row during the weaning are well-clothed with a dense, soft and woolly juvenile pelage. During the course of their first summer of life, this is gradually replaced by a second pelage which is longer THE RELATIONSHIPS OF AMPHIBERINGIAN MARMOTS 1 and less soft. A second set of hairs grow up through the first juve- nile pelage, some of which is shed gradually in most specimens (Kapitonov, 1964), and it is possible arbitrarily to divide juvenile pelages into three stages, here termed first, second, and third juvenile pelages for comparative purposes. The first juvenile pelage in M. camtschatica has a brownish- black to black cap, with dorsal guard hair tips of the same color. The middle band of the dorsal guard hairs is Smoke-Gray to Cartridge Buff, and ventral guard hairs are near Cinnamon Rufous, as are the lower limbs, feet, and ears. The distal four-fifths of the tail has a conspicuous blackish tip. Juveniles in this stage usually have deciduous P3-P4, p3-p4 and M1, ml in place, with M2, m2 beginning to erupt. The second juvenile pelage is a transitional one, in which guard hairs begin to be shed before the second set of hairs begins to appear. Relative scarcity of black-tipped dorsal guard hairs ex- poses more underfur, producing a lighter, “woollier” appearance, but the basic color pattern is the same. Animals in this transitional stage have M2 and m2 partially erupted, but M3 and m8 either have not begun, or are just beginning, to erupt. The third juvenile pelage is attained when the second set of juvenile hairs grow out well beyond the remaining hairs of the first set, giving the juvenile a color and pattern essentially like the adult. A notable feature is the loss of the conspicuous black tail tip of the first juvenile pelage, which is replaced by long tail hairs that are black subterminally, but orangish on the tips. Individuals in this stage have M3, m3 in eruption. Marmota caligata juveniles appear to pass through the same molt pattern as other marmots (pers. obs.; Kapitonov, 1964). No- table features of juvenile pelages are the white facial markings, black feet, lack of a black tail tip, and, in the third juvenile pelage, white-tipped guard hairs. The considerable variation in juvenile pelages (light to dark), in general, parallels that seen in adult pelages. Not surprisingly, juvenile pelage of the melanistic M. vancou- verensis is a uniformly dark brown, and does not differ from adult coloration. However, in M. olympus, the third juvenile pelage is unusual in that it has not yet achieved the adult pattern, as is found in all other species in the genus. The light-tipped guard hairs so typical of the upper back in third juvenile and adult pelage in caligata are absent in olympus, which instead more nearly re- semble the third juvenile stage seen in M. monax and M. flavi- ventris. Only a few specimens of juvenile M. broweri were available. However, in black cap and lack of white facial marking, lack of contrast between color of feet and belly, and some indication of a 12 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY black tip on the tail, they more closely resemble camtschatica than caligata. Additionally, ventral guard hairs of the first juvenile pelage are dark basally, Cartridge Buff to Ivory Yellow medially, and dark at the tips. The pattern is the same found in camtschatica, and the color of the medial band deviates from that of adult broweri in the direction of camtschatica. In contrast, ventral guard hairs of the first juvenile pelage in M. c. caligata are mostly all white medially and distally, with few having black tips (this varies geographically, other subspecies having a higher frequency of blackish or brownish tips on the ventral guard hairs). Foot Pad Morphology.—Another external feature that may serve to distinguish these marmots is foot pad morphology. All marmots have six plantar pads on the hind foot, but Marmota caligata and M. olympus (and probably vancouverensis) appear to be unique in the genus, in that the posterior two pads are both nearly circular in shape (Fig. 3a). In contrast, the posterior pair of pads on the sole of M. broweri and M. camtschatica are both elongated (Fig. 3b, c). The anterior four pads at the base of the hind toes are similar in all species. This character is difficult to evaluate on dried museum skins, but it does not seem to vary much. Number of Mammae.—Moore (1961), in his survey of sciurid reproductive characteristics, described various species of Marmota as having four, five, or six pairs of mammae. He reported M. Fic. 3.—Morphology of plantar pads on sole; A, M. c. caligata (USNM, from Howell, 1915); B, M. broweri (KU 50417); C, M. camtschatica bungei (UM 5415). THE RELATIONSHIPS OF AMPHIBERINGIAN MARMOTS 13 caligata to have five pair (N=9); of 36 M. caligata whose mammae we could count accurately, 31 had five pairs, four had four pairs plus one unpaired teat, and one possessed only four pair. One M. olympus had five pairs of mammae, and another, four pairs plus one unpaired teat. Of six lactating adult female vancouwverensis, one had four pairs, two had four pairs plus one unpaired, and three had five pairs of mammae. On the other hand, about half of the M. camtschatica had five pairs (N=9), while the remainder had six pairs (N—8); moreover, five of those with five pairs had an additional, unpaired supernumerary teat. Bee and Hall (1956) report five pair in one female M. broweri, but more must be ex- amined before the mastology of this species is determined. Body Size and Proportions—Among the amphiberingian spe- cies, M. camtschatica is of average size overall, with mean adult head-body length of 473.3 mm for males (N=29) and 458.4 mm for females (N=35) (see also Kapitonov, 1963), whereas M. caligata is among the larger with males averaging 499.8 (N=46) TABLE 3.—Body size (mean + 2 SEm) and proportion (%) in the black- capped marmot, Marmota camtschatica, and the hoary marmot group. M. camtschatica Head-body Tail Tail/head- Locality, N length, mm. length, mm. body, mean % Subspecies G2 GS eee $3 28 ne) QS Barguzin Mts. tt, AAG 440 150 140 32 32 (M. c. doppelmayeri) eraeihiyer ss sss be Dit 22; 46010 Acs>) 183:0° 1246" 28.4 28 (M. c. bungei) __. SO A, F=EG 10% =Eab Kamchatka Pen. ___. St 12 08H 496.56 16244 153-4 ao kGy 230.6 (M. c. camtschatica) ail SEISNL seilb8} Sellil tl M. caligata group Head-body ‘Tail Tail/head- N length, mm. length, mm. body, mean % OTS pene 55 amon re} 29 $d 29 Alaska/Yukon -_...... 8. 14) -AS0.0° 469.9) ©212:4—- 178.3, 429 37.7 WieC sCOMEGED So. 22,0 1919 E120) ==9.6 British Columbia __... 17 18 505.9 493.6 199.8 2055 394 41.6 M. c. oxytona __.. +24.8 +18.8 +10.7 +10.2 B. C./Alberta 4 10 489.0 481.2 212.0 210.0 4385 43.7 M. c. okanagana __..... exikeil()) aotefs) jessyel marsh?! Montana __.............- bie SOeeo21o 4282 217.6) e209 42 Om) 44.7 M. c. nivaria ________. aefe! eeiie) seit) Sees Cascade Mts. _...- 10 9 OS17.7 4776 2148 2074 41.5 43.3 M. c. cascadensis _... +11.4 +15.7 +110 +138 Vancouver Island __... 5 5 482.0 480.4 202.0 2056 42.0 42.8 M. vancouverensis __. sett}7/ Sei sy S¥/(0) ell Olympic Mts. _...... 6 8 5862 5140 2125 1910 405 406 M: olympus _____. Pe1b.6) ol Seo. ow 14 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY and females 479.3 (N=48). The smallest black-capped marmots occur in the southwestern portion of the species’ range, east of Lake Baikal. Populations farther east and north are progressively larger, the largest occurring on the Kamchatka Peninsula (Tables 3, 4) and equaling caligata in size. Variation in head-body size of M. caligata does not follow such a monotonic cline (Table 3). Specimens from Alaska, the northern Yukon, and British Columbia (M. c. caligata) are relatively small, while to the south (M. c. oxytona) they average larger, though variation is considerable. Farther south still, in southern British Columbia and Alberta, Montana, and Idaho (M. c. okanagana, M. c. nivaria Q 2), size decreases once again. Coastal populations (M. c. raceyi, M. c. cascadensis, M. olympus) are large, while island populations (M. c. vigilis, M. c. sheldoni, M. vancouveren- sis) are small (see also Howell, 1915). Geographic variation in body size has not yet been related to environmental factors as has been done with the related prairie dogs (Cynomys) (Pizzimenti, 1975). Body measurements are available for only a few specimens, but the Brooks Range marmot appears to be small; adult male head-body length was 442.0+ 6.0 mm (N=9), and the mean for eight adult females was 424.94 8.7 mm. Thus, M. broweri is smaller in size than M. camtschatica, particularly the large popu- lation geographically closest to northern Alaska on the Kamchatka Peninsula (M. c. camtschatica). Adequate sampling will probably show broweri to be about the same size as smaller populations of black-capped marmots. Body size as reflected by cranial dimen- sions such as condylobasal length, also shows this same pattern of geographic and interspecific relationships (Table 4). The only proportion that changes much between marmot spe- cies is tail length. Marmota camtschatica, like most Old World taxa, is fairly short-tailed: 140.6 mm for males and 133.9 mm for females (22 to 36% of individual body length). M. caligata, olympus, and vancouverensis in contrast have some of the longest tails in the genus (Table 3), 206.2 mm in males and 196.4 mm in females of caligata (30 to 51% of individual body length), 212.5 mm in males and 191.0 mm in females of olympus (35-45% of indi- vidual body length) and 202.0 mm in males and 205.6 mm in females of vancouverensis (33-46% of individual body length). Tail length of male M. broweri averaged 163.1+ 5.8 mm (r. 152-181; N=7), and of females, 154.3+ 3.7 mm (r. 133-164; N=8) or 31-41% of individual body length, thus falling between the other amphiberingian species. Relatively few complete skeletons were examined, but one specimen of M. camtschatica bungei had 20 caudal vertebrae, and tail length of 160 mm; two M. caligata each had 21 caudal vertebrae, and tail lengths of 230 and 240 mm; two 1 2 THE RELATIONSHIPS OF AMPHIBERINGIAN MARMOTS TABLE 4a.—Cranial dimensions of amphiberingian Marmota in mm. (mean, standard deviation ), males. Character: . L. Ang. Proc. le Art) Proce: . Rost. L. . Rost. W. . For. Mag. W. . For. Mag. H. . Int.-Orb. W. . Post-Orb. W. . Sag. Cr. L. . Tymp. Bul. W. . Tymp. Bul. L. . Max. T.R.L. . Zygom. W. . Nasal L. . Post-Orb. L. . Rost. H. Max. H. ; oe H. . Mast. W. . Depr. Nas. . Depr. Occ. . Prof, Ind. Nas. . Prof, Ind. Occ. . Cond.-Bas. L. ample size N Marmota camtschatica: doppel- mayert 44.87 1.00 42.00 1.00 25.07 0.40 19.40 0.46 3 bunget 45.13 1.94 44.39 1.54 26.26 1.21 20.36 0.72 9.58 0.67 6.83 0.65 20.95 1.19 17.26 0.51 25.42 3.20 15.49 0.57 17.84 0.69 20.99 0.53 59.03 2.01 37.30 1.72 38.29 2.24 20.93 0.82 25.37 1.08 28.64 0.67 38.97 1.51 18.83 1.60 30.04 1.05 37.80 2.45 35.91 1.90 88.92 3.29 35 camts- chatica 50.03 2.56 3.76 14 M. browert 47.83 3.28 14 Marmota caligata: ssp.* 52.79 2.69 4.24 52° nivaria 51.00 4.31 46.01 2.96 25.31 7.10 19 casca- densis 53.71 2.61 8 M. olympus 54.93 1.96 103.27 2.68 12 15 M. vancou- verensis 53.07 0.70 49.63 0.87 27.98 1.16 23.73 0.25 12.50 0.89 ‘Includes caligata (N=19), nivaria (N=4), okanagana (N=7), and oxytona (N=22). 16 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY TABLE 4b.—Cranial dimensions of amphiberingian Marmota in mm. (mean, standard deviation ), females. Marmota camtschatica: Marmota caligata: M. doppel- camts- M. casca- M. vancou- Character: mayeri bungei chatica broweri ssp.° nivaria densis olympus verensis — . L. Ang. Proc. 42.91 44.23 48.07 43.38 51.45 47.90 53.71 52.98 49.59 2.03 144 267 202 244 324 256 235 2.40 2. L. Art. Proc. 40.96 43.24 45.77 41.26 47.01 44.66 48.01 48.13 46.95 199 138 214 131 4173 254 168 1627" 249 3. Rost. L. 23.99 25.90 26.89 24.06 26.80 23.62 28.67 26.93 25.60 106 O94 140 085.149 191 137 ‘T4492 4, Rost. W. 18.27 19.87 21.42 19.33 21.28 19.91 22.03 21.69 21.78 0.76.-0.938 120 0.63 106 s7 _.1.36-- 0.98 0g 5. For.Mag.W. 11.12 11.49 11.89 11.25 12.66 12.13 12.81 12.06 12.20 0.32 024 034 059 072 068 089 0.55 # 1.00 6. For. Mag. H. 945 879 812 823 896 895 9.03 883 - 9.20 0.42 042 061 079 O68 060 068 0.62 0.23 7. Int.-Orb. W. 18.32 20.36 22.94 21.71 2348 21.91 23.36 25.76 22.30 167 “123 £66. 128 t3l 223 206 125 Sih 8. Post-Orb. W. 17.12 17.27 17.19 17.06 16.44 15.92 17.45 17.08 15.48 035 028 116 128 14] -1384 [65 Latagie 9. Sag. Cr. L. 22.03 24.49 25.66 25.17 28.66 21.75 30.15 28.86 31.25 10. Tymp. Bul, W. 14.23 14.94 1529 14.09 15.78 1648 17.11 16.27 16.36 11. Tymp. Bul. L. 16.56 17.26 17.48°1611 17.65 17.65 18.94 18.02 17.59 12. Max. T.R.L. 19.66 20.23 2148 2051 2281 22.05 2337 22.97 93.73 13. Zygom. W. 51.61 54.12 56.07 5626 62.82 59.85 65.56 62.76 60.71 14, Nasal L. 35.43 3641 3950 3530 3940 3771 4204 4198 3884 15. Post-Orb. L. 34.76 37.63 39.31 38.31 42.28 39.89 4354 44.87 41.24 16. Rost. H. 18.64 20.37 21.30 19.94 22.59 21.27 23.11 23.70 22.03 0.98 O81 007 124 %103 4146 143 2.06 0 17. Max. H. 22.13 24.81 26.33 24.03 26.17 25.38 26.78 27.55 25.88 1.30 0:64 .1.40 1.432121 L77 1:66.. d39 Sie 18. Occ. H. 25.97 27.90 28.48 27.02 30.16 28.32 31.23 29.95 29.22 lel O84 Til O70 ‘T2l 168 o.27- Lie0 eae 19. Mast. W. 35.66 38.03 39.35 38.05 43.29 42.27 46.49 44.55 42.83 155, 089 211 -1338..4177 3:16. 243 “220s 20. Depr. Nas. 18.86 18.15 19.81 15.00 15.78 11.00 16.92 19.92 15.10 212 184 269 105 221.4113 247 %198 251 21. Depr. Occ. 27.00 29.59 29.94 29.90 30.12 33.33 30.69 30.50 32.76 22. Prof. Ind. Nas. 33.29 36.49 38.56 35.40 40.23 37.92 41.74 41.08 39.50 23. Prof, Ind. Occ, 31.29 35.67 38.44 36.50 42.23 39.33 44.00 44.67 39.99 24. Cond.-Bas, L. 81.70 87.25 91.14 86.89 95.82 91.06 99.61 99.83 92.68 416 2.40 426 3.19 383 637 403 441 2.20 Sample size N ae eee ee ea ® Includes caligata (N=35), nivaria (N=5), okanagana (N=10), and oxytona (N= 19), THE RELATIONSHIPS OF AMPHIBERINGIAN MARMOTS iv M. broweri.also each had 21, and tail lengths of 143 and 190 mm. Thus, tail length is apparently independent of the number of caudal vertebrae, which in all species of Marmota examined ranged from 18 to 23, the modal value being 21. Cranial Morphology Sexual dimorphism.—The extent of sexual dimorphism in am- phiberingian marmots is considerable as revealed by DF analysis (Table 5). Approximate F statistics between males and females were not significant at the p = .05 level in M. caligata cascadensis and M. olympus, the two largest taxa, but were below the p = .001 level in the remaining seven groups tested. It is not surprising that the majority of specimens were correctly classified with four or fewer variables entered in the model as determined by the classi- fication functions shown (Table 5). Here variables are represented by a V and the subscript indicates the variable number. Our cri- terion for including the variable in the model was whether the variable entered at each step was significant at the p = .05 level. Variables which best classified individuals as to sex were different for each group but because nearly all variables served to discrimi- nate between sexes to a greater or lesser extent, in some cases including as many variables as the number of degrees of freedom permitted improved the identification. Geographic and Interspecific Variation: Univariate Compari- sons.—The most trenchant differences between camtschatica and broweri on the one hand and the caligata group on the other are as follows. In the former pair, the angular process is relatively short, and does not extend much beyond the articular process (Table 4, Fig. 4). The rostra of camtschatica and broweri are rela- tively longer than caligata. Postorbital width is greater in broweri and camtschatica than in caligata; tympanic bulla length is rela- tively less in caligata. Relative zygomatic breadth is less in cam- tschatica, and the shape of the zygoma also differ; in camtschatica the maxillary portions of the arches diverge rapidly posteriorly, and the arches then tend to become more rounded, whereas in caligata the divergence posteriorly is more gradual and the arches appear more parallel, or else continue to diverge (Fig. 4). M. broweri is intermediate in zygomatic width, but in shape is closer to camtschatica. M. camtschatica has a relatively narrower mastoid width than caligata and the depression of the nasals is greater; broweri is intermediate in these characters. Finally, the elevation of the occiput viewed in profile (profile index, occiput) is higher in caligata than in camtschatica or broweri owing largely to the deeper angular process of the mandible in caligata, and greater development of sagittal and lambdoidal crests. . SSS 'SOSSSSree ae ~ : 100°0 > = cee ‘TOO § | —— aailice 7 Fic. 4, cont—Skull and mandible of, e, M. vancouverensis, 9 (UM 13521), original. «0.5 the tooth rows to be more parallel. The basioccipital projections in olympus and vancouverensis are larger than usually found in ca- ligata, and extend well below the ventral surface of the tympanic bullae. The lachrymal bone approaches a quadrate form in olym- pus and vancouverensis, and the orbital wing extends well above its upper margin, in contrast to caligata (see above). In certain other characters, vancouverensis differs from the other two species. The coronoid process is thinner and more recurved, the lower margin of the mandible not as concave, and the anterior cusplet of the P4 is small. The proximal shape of the nasals in vancouver- ensis is also unique, in that the antero-medial margins of the fron- tals extend forward, forming a V-shaped notch at the proximal ends of the nasals. Multivariate Comparisons: Principal Component Analysis.— Both the PC and cluster analysis gave similar results, but of the two, the former seemed more useful in determining groups for later DF analysis and for revealing patterns of geographic variation. THE RELATIONSHIPS OF AMPHIBERINGIAN MARMOTS 23 In the PC analysis, males and females were treated separately and the first ten factors were extracted from their correlation ma- trices, which accounted for 97.0 and 97.5 per cent of the variation for the respective sexes. Of these ten factors, only the first four were judged interpretable. Loadings = 0.250 are marked by asterisks in Table 6. For both male and female PC analyses factor loadings are highly positive on most characters for PC I, which we interpret as a general size factor (Table 6). Postorbital width, however, exhibits an inverse relationship to general size, probably as a result of age variation (Pizzimenti, 1975:39); depression of nasals may also vary inversely, but loading values are low. PC II shows heavy positive loading on depression of nasals, rostral length, and length of sagittal crest, combined with high negative loadings on depression of occiput, and foramen magnum width, while females have negative loadings on tympanic bulla width and maxillary tooth row length. PC II thus appears to be a complex shape factor; animals with long, deep rostra and long sagittal crests tend to have somewhat shallow skulls and smaller TABLE 6.—Matrix of factor loadings for principal components. Male Female I II III IV I II III IV 1. L. Ang. Proc. 0.954* -0.038 -0.080 -0.150 0.981* -0.047 0.056 0.050 2. L. Art. Proc. 0.955* 0.183 0.047 -0.080 0.983* -0.011 0.025 0.028 3. Rost. L. 0.787* 0.495* 0.124 0.101 0.814* 0.422* 0.146 —-0.001 4. Rost. W. 0.882* 0.214 -0.075 0.053 0.901* 0.170 0.095 0.013 5. For. Mag. W. 0.837* —0.254* -0.310* 0.088 0.766* -0.133 0.381* 0.148 6. For. Mag. H. 0.717* -0.465* -—0.266* 0.103 0.004 -0.382* 0.543* 0.652* 7. Int.-Orb. W. 0.799* 0.039 -0.356* -0.042 0.867* 0.063 -0.099 -0.201 8. Post-Orb. W. -0.365* 0.210 -0.478* 0.714* -0.429* 0.190 0.541* -0.541* 9. Sag. Cr. L. 0.709* 0.315* 0.240 -0.491* 0.840* 0.314* -0.254* 0.146 10. Tymp. Bul. W. 0.613* -0.225 0.538* 0.317* 0.702* —0.492* 0.092 -0.008 11. Tymp. Bul. L. 0.507* 0.241 0.609* 0.372* 0.492* -0.243 0.572* —0.338* 12. Max. T.R.L. 0.867* —0.091 0.002 0.033 £0.862* -0.290* 0.040 -0.115 13. Zygom. W. 0.928* -0.163 -0.204 -0.018 0.946* -0.106 0.008 -0.095 14, Nasal L. 0.886* 0.255* 0.070 0.114 0.913" 0.172 0.004 0.044 15. Post-Orb. L. 0.964* 0.037 -0.000 -0.005 0.935* -0.033 -0.197 -0.048 16. Rost. H. 0.959* 0.061 -0.139 0.000 0.975* 0.124 -0.017 0.005 17. Max. H. 0.938% 0.114 -0.016 -0.022 0.890* 0.157 -0.160 0.046 18. Occ. H. 0.937* 0.004 0.009 -0.026 0.961* -0.011 0.149 0.060 19. Mast. W. 0.958* -—0.035 -0.023 0.145 0.954* -0.172 0.010 0.063 20. Depr. Nas. -—0.197 0.896" -0.037 0.006 0.127 0.866" 0.155 0.101 21. Depr. Occ. 0.504* —0.470* 0.585* 0.085 0.321" -—0.745* —0.348* -—0.164 22. Prop. Ind. Nas. 0.833* -—0.094 -0.312% -0.099 0.912% 0.101 -0.056 -0.048 23. Prop. Ind. Occ. 0.926" 0.131 0.042 -0.020 0.947% 0.006 -0.086 -0.090 24. Cond.-Bas. L. 0.978* 0.116 -0.043 0.065 0.988" 0.063 -0.023 -0.013 VP 16.064 2.042 1.780 1.111 16.113 2.328 1.422 1.003 =VP/TOTVP 66.9 75.4 82.9 87.5 67.1 76.8 82.8 86.9 * — Loadings = 0.250 24 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY foramina magna. PC III for males shows high positive load- ings for tympanic width and length and depression of occiput, contrasted with high negative loadings for inter- and post-orbital width, foramen magnum width and height, and anterior profile index. For females, PC III shows high positive loadings for post- orbital width, foramen magnum width and height, and tympanic bulla length, and high negative loadings for depression of occiput and sagittal crest length. Many of these loadings involve the orbital and anterior regions. In males the cranium is somewhat domed, with a well-developed sagittal crest; the inter- and post-oribtal re- gion is narrow and the foramen magnum small. In females, the post-orbital region is particularly broad and foramen magnum larger while the cranium lacks pronounced doming or enlarged sagittal crest. For PC IV, males again show high positive loadings on tym- panic bulla width and length, and post-orbital width, and a high negative loading for sagittal crest length. For females on this fac- tor, foramen magnum height is again strongly positive, and post- orbital width and tympanic bulla length are highly negative. This factor represents the contrast in size of the post-orbital width with the length of the sagittal crest for males, which is age-related: in older males, sagittal crest development becomes pronounced and there is a relative decrease in post-orbital width. Projection of locality scores for both male and female analyses onto the first two principal component axes are shown in Fig. 5. Locality scores are connected by a straight line (some points are unpaired because only single males and females were available from these localities). Since males and females from the same locality represent the same gene pool, the midpoints of the line segments may be regarded as the best combined estimate of a locality on the first two PC axes. The broken lines surrounding several localities denote those which were combined in later analy- ses for our groups. Points not so enclosed (6, 21, 22, 31) were not included in later analysis. In summary, PC I represents general cranial size, PC II is a complex shape factor, associating long rostra with shallow crania, and PC III and IV represent sex and age-related variation in orbit, sagittal crest, and basicranium. Multivariate Comparisons: Discriminant Function Analysis.— Results of a DF analysis (on 9 male and 9 female groups com- bined), including classification functions, classification table, sum- mary table and canonical variables evaluated at group means (too extensive to be included here) are found in Appendix B. Projec- tion of discriminant scores I, II, and III (Fig. 6) differentiate the species on both size and shape, and results in two clusters showing no overlap—the M. caligata group to the right, and M. broweri and M. camtschatica to the left. Within the caligata group, skull size and shape of both male THE RELATIONSHIPS OF AMPHIBERINGIAN MARMOTS 25 @ ¥OLOVS M.c.sheldoni 21 Mc nivarig —24 —19 —1A —098 —OA —01 +06 +14 +15 FACTOR 1 Fic. 5.—Projection of mean vector scores of separate PC analyses for males and females by localities onto the first two principal component axes extracted from the correlation matrix. Lines connect males and females from the same localities; symbols represent locality midpoints. Closed squares, M. camts- chatica; open squares, M. broweri; closed circles, M. caligata; open circle, M. vancouverensis; half-open circle, M. olympus. Localities (see also Speci- mens Examined) coded by number as follows: (M. camtschatica doppel- mayeri) 1—Buryat-Mongol A.S.S.R. (M. c. bungei) 2—vic. Verkhoyansk; 3—Adicha River; 4—lower Lena River; 5—Tuora-Yurak River. (M. c. camtschatica) 6—Magadan Obl.; 7—Kamchatka Obl. (M. broweri) 8—west- ern Brooks Range; 9, 10—central Brooks Range; 11—eastern Brooks Range. (M. c. caligata) 12—west-central Alaska; 13—southwestern Alaska; 14— Kenai Peninsula; 15—Hinchinbrook Island; 16—Mt. McKinley; 17—vic. Fair- banks; 18—upper Yukon; 19, 20—southeastern Alaska and British Columbia. (M. c. sheldoni) 21—Montague Island. (M. c. vigilis) 22—Glacier Bay. (M. c. oxytona) 23, 35—Northwest Terr. and northern British Columbia; 24—central British Columbia; 25—southermn British Columbia and Alberta. (M. c. okanagana) 26, 27—southern British Columbia and Alberta. (M. c. nivaria) 28—northwestern Montana; 29—central Montana; 30—southern Mon- tana. (M. c. raceyi) 31—west-central British Columbia. (M. c. cascadensis ) 32—western British Columbia and Washington. (M. vancouverensis) 33— Vancouver Island. (M. olympus) 34—Olympic Peninsula. 26 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY and female M. olympus is not far from M. caligata cascadensis, as previously noted by Howell (1915). However, cranial morphology does not place M. vancouverensis close to olympus (or cascaden- sis, as Howell suggested) in the three-dimensional plots, but rather, closer to M. c. nivaria which is relatively isolated from the rest of the caligata group. . | Cluster Analyses—UPGMA cluster analysis on the mean dis- tance matrix (Appendix C) from separate DF analyses on males and females of the subspecies samples resulted in the distance phenogram of Figure 7 whose cophenetic correlation (Sneath and Sokal, 1973) was 0.839. This phenogram again demonstrates that, on the basis of cranial morphology, there is a major distinction between the caligata group and camtschatica, and M. broweri clusters with camtschatica. M. olympus clusters with M. caligata ssp. and M. c. cascadensis. M. vancouverensis, in contrast, is closer to M. caligata nivaria; both of these latter taxa are somewhat iso- lated from the rest of the caligata group. » Vocalization Calls of Marmota broweri have not been described. The “alarm” call, about 0.4 seconds in duration (Fig. 8a), differs from those of both camtschatica and caligata. The dominant frequency starts at Mc doppeimayeri® Fic. 6.—Projection of discriminant scores on the first three canonical axes for amphiberingian marmots grouped by taxon. Solid symbols, males; open symbols, females. THE RELATIONSHIPS OF AMPHIBERINGIAN MARMOTS 27 about 1500 Hz, rises to about 4000 Hz, then falls slightly to about 3000 Hz, and levels off for the last half of the call; its intensity appears to be greatest in the middle. The first harmonic above the dominant starts at about 5000 Hz, and rises and falls like the dominant but disappears during the last half of the call. The attenuated alarm (Fig. 8b) has a duration of almost 0.9 seconds. It begins as does the short alarm call; the dominant frequency starts below 2000 Hz, rises to almost 4000 Hz, then falls to a drawn-out note around 3000 Hz. There is also at least one har- monic. A “squeal” or “growl” is also shown in Fig. 8c. The amorphous structure of this call is similar to comparable vocali- zations by other species of marmots and ground squirrels. Since vocal sounds made by various species of marmots vary considerably, a comparison of primary or long call, also called “alarm” or whistle motif (Barash, 1973; Nikol’skii, 1976; Taulman, 1977; Waring, 1966) of the amphiberingian species is instructive. The primary call of M. caligata is a loud, prolonged “whistle,” steady in intensity at a dominant frequency of about 3200 Hz; the first harmonic is at about: 6000 Hz. Its mean duration was 0.74 seconds (Taulman, 1977) and 0.56 seconds (Waring, 1966). There is also a short, or “descending” call of the same frequency, but from 0.3 to 0.5 seconds in duration, and with an ascending be- ginning and especially a descending end (Taulman, 1977). The primary call of M. olympus is similar in sound to that of M. caligata (Barash, 1973). The dominant is around 3000 Hz, but there is no detectable harmonic, and there is a slight initial rise and terminal fall. The duration is also less than in caligata, aver- aging 0.39 seconds. Barash also recorded a short call lasting little more than 0.1 second, but having the same structure, and perhaps slightly lower frequency. M. vancouverensis calls have not been recorded; it is said to rarely vocalize (Swarth, 1911), but has “a loud chirping call and a piercing whistle of alarm” (Hardy, 1955). - In contrast, the primary call of M. camtschatica is short, lasting M. caligata ssp. M.c.cascadensis ° 8.93 8.18 743 6.68 5.93 5.18 443 3.68 Fic. 7—UPGMA phenogram based on Euclidean distances between group centroids on the canonical axis of the DF analysis. 28 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY for only about 0.2 seconds, and has a very different harmonic structure (Nikol’skii, 1976). Its dominant begins at about 1000 Hz, rises to about 3000 Hz (M. c. c.) to 5000 Hz (M. c. d.), and then falls symmetrically to 500-1000 Hz. The first harmonic starts at 3000 Hz and rises to about 8000 Hz in M. c. camtschatica; no harmonic was detected in M. c. doppelmayeri. Nikol’skii (op. cit.) regards the “alarm” call of this species as intermediate in structure between those of other Old World species and New World marmots. The harmonic structure of the primary call of broweri com- bines the changing frequency and harmonic of camtschatica in the first half of its call with a steady frequency similar to caligata in the second half. The alarm call of M. broweri has about the same duration as the long call of M. olympus, while the attenuated alarm call of broweri is as prolonged as the long call of M. caligata. However, the vocal pattern is sufficiently distinct to be considered species-specific. < Biochemical Comparisons Of the three blood serum proteins examined, only albumin showed no differences in mobility among the four species (Fig. 9), and it appears to be monomorphic in all of the species of Marmota examined so far (Nadler, unpublished). The leucine aminopepti- dase band of M. caligata (arbitrarily designated LAP 1) migrated more rapidly than that of M. broweri (LAP 2), but showed no intraspecific variability; data for M. camtschatica and M. olympus were not obtained. Transferrins appeared as either simple (ho- mozygous) or double (heterozygous) fractions after starch-gel electrophoresis. Specimens of M. caligata exhibited homozygous patterns for Tf 2 (one specimen) and Tf 3 (eight specimens); five — specimens had heterozygous Tf 2-3 patterns (Fig. 9). Only a monomorphic Tf 2 pattern was seen in the two M. olympus studied. In contrast, eleven M. broweri from the vicinity of Mt. Wachsmuth, near Chandler Lake, Alaska, were all homozygous (Tf 7-7). The single M. camtschatica was a heterozygous individual whose two Tf bands, designated 6 and 8, ran very close to, but on each side of, Tf 7 of broweri. DiIscussIoN Distributional Evidence Although Hall and Gilmore (1934) originally suggested that intergradation between M. broweri and M. c. caligata probably existed, it has not been found. Instead, Rausch (1953) and Rausch and Rausch (1971) clarified the distribution of the two species in ‘laNOLG DOWD JO S][eo wee Jo wIeIsoNOSdsoIpny—'g “Ol SPUOIES 80 90 vo z0 SpuddIaS Zjsayo!y ADuenbe 30 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Alaska, eliminating unsubstantiated early records from the Seward Peninsula, and Fort Yukon, and demonstrating that ranges of the two species were not contiguous in Alaska, as they had previously been mapped (Anderson, 1934; Hall and Kelson, 1959). Rausch and Rausch (1971) suggested there was an area of potential con- tact between the two species at the undefined eastern edge of the range of broweri, in the Richardson Mountains west of the delta of Mackenzie River. Porsild (1945) reported a sighting of marmots on Black Mountain, in the Richardson Mountains, southwest of Aklavik (old site), Northwest Territories, but no specimens were taken, and Youngman (1975) found no evidence of marmots in the Richardson. Mountains. However, a juvenile M. broweri, la- beled “Arctic Red River” [Northwest Territories], was examined in the British Museum (Natural History). This locality is south- east of the Richardson Mountains, and the Arctic Red is a tributary of the Mackenzie River which flows north out of the Mackenzie ALB M. Olympus M.c.caligata M.c.camtchatica M. broweri Fic. 9.—Zymogram patterns from electrophoresis of blood serum samples of amphiberingian marmots. M. olympus: Tf 2-2; M. caligata: LAP 1, Tf 2-2, 2-3, and 3-3; M. broweri; LAP 2, Tf 7-7; M. camtschatica: Tf 6-8. Albumin is monomorphic. THE RELATIONSHIPS OF AMPHIBERINGIAN MARMOTS 31 Mountains. But the specimen may have come from farther west or from mountains forming the divide between the Peel and Porcupine Rivers. Hoary marmots (M. caligata) occur as far north as the Mackenzie and Olgilvie Mountains, Yukon and Northwest Terri- tories (Banfield, 1974; Hall and Kelson, 1959). The little-known country between the Porcupine and Peel Rivers is therefore the most likely area for M. broweri and M. caligata to come into con- tact, if they do anywhere (R. L. Rausch, pers. comm.; Fig. 1). Marmota olympus and M. vancouverensis are both geographi- cally isolated, relict populations whose nearest congeners are the Coast Range populations of M. caligata cascadensis and M. c. raceyji. Behavioral Evidence Denning.—Rausch and Rausch (1971) compared several as- pects of the biology of the amphiberingian species to elucidate relationships. Denning behavior is similar in broweri and cam- tschatica; in both species all individuals in a colony hibernate together in a single winter den, and this seems to be characteristic of most, if not all, of the Old World marmots. In contrast, New World M. monax and M. flaviventris colonies have several hiber- nacula, and adult males appear to hibernate singly, and breeding females with their young (Grizzell, 1955; Hamilton, 1934; deVos and Gillespie, 1960). The situation in M. caligata is unclear. Barash (1974) found this species to resemble in its behavior M. olympus. While cranial morphology and pelage of M. olympus resemble caligata (see above), the karyotype (2n = 40, FN = 66; Rausch and Rausch, 1965, 1971) resembles M. camtschatica (Lya- punova and Vorontsov, 1969) (see below). Barash (1973) found that, like camtschatica and broweri, olympus individuals hibernate together in a single burrow. Barash (1974) apparently did not de- termine this for the M. caligata colonies he studied at Logan Pass, Glacier National Park. Biesemeyer (1960), who earlier studied one of the same M. caligata colonies (Oberlin-Clements), found cir- cumstantial evidence against group hibernation. However, this colony exhibited an unusual pattern of spring movement associated with the fact that hibernation burrows were some distance from snow-free: early season foraging areas (Biesemeyer, 1960; Barash, 1974) and its hibernation behavior may not be typical of the species. Breeding.—Another characteristic shared by M. camtschatica and broweri is that “. . . breeding evidently takes place before the animals are able to leave the winter den” like most Old World species (Rausch and Rausch, 1971), and in camtschatica young are born a week or two after the adults emerge from hibernation (Kapitonov, 1969). In caligata and olympus, on the other hand, 32 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY breeding probably takes place after emergence, and young are born about a month after emergence (Biesemeyer, 1960; Barash, 1973, 1974). Evidence from Parasites Rausch and Rausch (1971) also summarized pertinent. parasite records for the amphiberingian marmots. M. camtschatica and M. broweri share the flea Oropsylla silantiewi (Kapitonov, 1960b; Rausch and Rausch, 1971), which also occurs widely on other Palearctic species. M. caligata harbors Thrassis pristinus, belonging to a Nearctic genus (Holland, 1963; Rausch and Rausch, 1971). Regarding endoparasites, caligata and broweri both are infested by the cestode Diandrya composita, which occurs in all Nearctic marmots except M. monax, and by Catenotaenia reggiae. In con- trast, no cestodes of any sort have been recorded from M. cam- tschatica (Kapitonov, 1960b). Rausch and Rausch (1971) considered that the co-occurrence of the cestodes D. composita and C. reggiae was most simply explained by the hypothesis that M. broweri and M. caligata were both derived from an ancestral form of the caligata-group, but could not then account satisfactorily for the co-occurrence of the flea O. silantiewi on M. broweri and M. camtschatica. They further reasoned that if their favored hypothesis was incorrect, and alterna- tively, if the cestodes had been transferred from M. caligata to M. broweri, transfer of the caligata flea T. pristinus presumably also would have occurred. Our suggestion for reconciling this conflicting evidence (assum- ing future research does not uncover T. pristinus on M. broweri, or the Nearctic cestodes in M. camtschatica) is as follows. Assume the marmot population ancestral to M. camtschatica and M. brow- eri inhabited the Beringian refugium during the Pleistocene (see Guthrie and Matthews, 1972; Hoffmann and Taber, 1967; Mac- pherson, 1965; Rand, 1954; Rausch amd Rausch, 1965), and har- bored a flea (O. silantiewi) but no cestodes (Rausch, 1976). Iso- lation of the two forms by the Bering Strait through rising sea level at the end of a glacial period would result in populations ancestral to M. camtschatica and M. broweri to the west and east, respec- tively, of the Strait, each with a flea but no cestode parasites. If broweri has come into limited post-Pleistocene contact with ca- ligata possessing a flea (T. pristinus) and cestodes (D. composita, C. reggiae), it seems possible that transfer of the cestodes to broweri, which lacked any, might be more probable than transfer of the flea Thrassis to Brooks Range marmots that already harbored the flea Oropsylla, since host specificity in fleas is pronounced (cf. Rausch, 1976). x THE RELATIONSHIPS OF AMPHIBERINGIAN MARMOTS 33 Chromosomal Evidence In concluding that M. broweri was more closely related to M. caligata than to M. camtschatica Rausch and Rausch (1971) appear to have been influenced by Vorontsov et al. (1969). Following conventional cytogenetic interpretations that chromosomal fusion, leading to lower diploid number and increase in number of bi- armed chromosomes, is the most likely direction of Robertsonian change, Vorontsov and his colleagues concluded that the karyotype of M. caligata and M. flaviventris (2n — 42, FN = 66) was an- cestral, and that M. broweri (2n = 36, FN = 66) was the most recently differentiated, possibly from M. monax (2n = 38, FN = 66?). Only when their own paper was already in press did Rausch and Rausch (1971) see the report of the karyotype of M. cam- tschatica (2n = 40, FN = 66) by Lyapunova and Vorontsov (1969). Direction of chromosomal evolution (fission vs. fusion) aside, fewer changes are required to remodel the karyotype of camtschatica to that of broweri, than to derive broweri from caligata. Thus, chromosomal evidence also supports the hypothesis that broweri is closer to camtschatica than to caligata. Because M. camtschatica and M. olympus share similar karyo- types (2n = 40), Vorontsov and Lyapunova (1976) suggested a direct phyletic relationship between them. Although karyologi- cally differentiated from the remainder of the caligata group (2n = 42), olympus is morphologically allied with it, rather than with camtschatica and broweri. It now seems likely that the ancestors of camtschatica were derived from a Nearctic marmot (Rausch, 1977; Vorontsov and Lyapunova, 1976) possessing 2n — 40 that is now represeted by the relict M. olympus. This ancestral popula- tion may t’ en have given rise to M. caligata (2n = 42) in western North An crica (Hoffmann and Nadler, 1968) and in eastern Eurasia to a population which ultimately differentiated into M. caudata and the M. bobac superspecies (bobac, baibacina, hima- layana, siberica, menzbieri) with 2n = 38, while M. camtschatica retained a 2n = 40. Marmota vancouverensis, on the other hand, is both karyo- logically and morphologically a member of the caligata group; its principal distinguishing feature is that genes for melanism have become fixed in this insular population. It is likely a Pleistocene isolate from caligata populations in the Coast Ranges of British Columbia which has entered a severe genetic bottleneck in terms of effective population size and decrease in suitable habitat. Biochemical Evidence Although presently limited to a few loci, biochemical evidence suggests differences between M. caligata and M. broweri greater 34 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY than usually found between closely related species of ground squir- rels (Nadler, et al., 1974). Thus, caligata and broweri differ in leucine aminopeptidase mobility, the LAP 1 of caligata also being found in M. monax and M. flaviventris; transferrin 2 of caligata and olympus is also shared with flaviventris. LAP 2 and Tf 7 of M. broweri have so far not been found in other species of Marmota; the similarity in mobility of Tf 7 to that of Tf 6 and 8 in M. camtschatica does not necessarily indicate molecular similarity. The hemoglobin of M. broweri also differs from those of M. olympus (Johnson, 1974), and Blumberg et al. (1960) described a poly- morphism in M. broweri haptoglobin without comparing their rela- tive mobilities to those of other marmots. Thus, available bio- chemical evidence does not contradict the view that broweri is closer to camtschatica than caligata. THE BERINGIAN FAUNA Although Rausch and Rausch (1965) originally suggested that M. broweri was probably a component of the late Pleistocene amphiberingian fauna, they later rejected this interpretation and suggested that it was more likely “. . . a relict North American species which became established in the Brooks Range during pre-Wiirm time .. .” (Rausch and Rausch, 1971). If this were true, then both M. broweri and M. camtschatica would have in- habited the Beringian refugium during the Wiirm-Wisconsin glacial period, since camtschatica is known from a “late Pleistocene” (= Wiirm) find in the northeastern Anadyr’ Mountains (Gromoyv, et al., 1965). Geographic sympatry between two species is rare in the genus Marmota, and where it occurs (baibacina and caudata; western Tyan Shan ranges; caligata and flaviventris, central Rocky Mountains; caligata and monax, central and northern Rocky Moun- tains), ecological niche segregation between the species is well defined (Hoffmann, 1974). Thus, we would expect two species as similar ecologically as are camtschatica and broweri to become at most parapatric in distribution, and not sympatric. If M. broweri was a pre-Wiirm—Wisconsin occupant of the Beringian refugium, derived from an ancestor of the caligata group, we expect it would have occupied all habitable environments in Beringia, and pre-empted them, thus preventing M. camtschatica from occupying the northeastern part of its present range. If, on the other hand, broweri and camtschatica were derived from a common ancester in Beringia, their isolation one or more times by the formation of Bering Strait in interglacial periods would have led to divergence and speciation. Speciation having occurred, competitive interaction between the two species would have been sufficient to keep either from occupying the other’s range; the area ~ THE RELATIONSHIPS OF AMPHIBERINGIAN MARMOTS 35 most probable for this contact to occur in, during glacial periods when the “land bridge” was present, was the area of the present- day Bering Strait. Rausch (1977) has more recently come to this view. Of those mammalian species thought to have occupied the Beringian region during the late Pleistocene ( Hoffmann, 1976; Mac- pherson, 1965; Rausch, 1963), most are considered Holarctic spe- cies, i.e., populations on either side of Bering Strait are considered conspecific. Some examples are Sorex cinereus, Sorex tundrensis, (Hoffmann, 1971; Hoffmann and Peterson, 1967; Youngman, 1975); Spermophilus parryii (Gromov, et al., 1965; Nadler, et al., 1973), Lemmus sibiricus (Rausch and Rausch, 1975), Clethrionomys rutilus and Microtus oeconomus (Nadler, et al., 1976; Zimmermann, 1942). A few species, however, appear specifically distinct; these include, in addition to the two marmots, the hares Lepus timidus and L. othus (Anderson, 1974), the narrow-skulled voles, Microtus gregalis and M. miurus (Fedyk, 1970; Rausch and Rausch, 1968), the snow sheep, Ovis nivicola and O. dalli (Chernyavskii, 1962, 1976; Korobitsyna, et al., 1974), and probably the collared lem- mings Dicrostonyx torquatus and D. stevensoni, etc. (Kozlovskii, 1974; Rausch and Rausch, 1972). The greater degree of differentiation between these species pairs, both chromosomal and morphological, than is found between Palearctic and Nearctic populations of other Beringian species, suggests that all Nearctic members of the pairs might be pre-Wiirm relicts (Chernyavskii, 1976). However, another interpretation is possible. Rates of evo- lution have differed in different taxa, and the greater differen- tiation between species pairs may be due to more rapid diver- gence than has occurred in the Holarctic species. It must be emphasized that the different biological “systems” within species populations apparently diverge at different rates. For example, there appears to be no gross or G-band structural differences in the chromosomes of the Siberian long-tailed ground squirrel (S. undulatus) and the Columbian ground squirrel (S. columbianus), but morphological differentiation is considerable (Nadler, et al., 1975; Robinson and Hoffmann, 1975). At present, it is not possible to choose between these alterna- tive hypotheses (longer isolation vs. more rapid divergence). As additional fossil material becomes available from the Beringian region, however, it may be possible to rule out one of them, as was possible in the case of O. nivicola—O. dalli (Korobitsyna, et al., 1974). CONCLUSIONS In the absence of demonstrated intergradation between M. 36 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY broweri and M. caligata, and in view of the marked differences between the species in chromosome number (Hoffmann and Nad- ler, 1968; Rausch and Rausch, 1965, 1971), external morphology (Rausch, 1953; this study), and cranial morphology (this study), we follow Rausch and Rausch (op. cit.) in regarding broweri and caligata as specifically distinct. This study has also shown that, in eel and cranial morphol- ogy, and certain biochemical characters, M. broweri resembles M. camtschatica more than the caligata group. We have also inter- preted evidence reviewed by Rausch and Rausch (1971) on den- ning and breeding behavior, and conclude that breeding behavior of broweri is more like that found in camtschatica than in caligata and olympus. Vocalizations of broweri are species-specific, as Nikol’skii (1976) found for other Marmota. Evidence from parasites is contradictory; a flea species is shared by broweri and camtschatica, but broweri and caligata share two species of cestodes. Chromosomal evidence also links M. broweri (2n = 36) more closely to M. camtschatica (2n = 40) than to M. caligata (2n = 42); all three species have the same number of aaietocal arms (F.N.a = 62) and generally similar sex chromosomes. These three species, plus the relict M. olympus (2n = 40) and M. vancouver- ensis (2n—= 42) appear to form a natural group. The existence of a Beringian (or amphiberingian) fauna during the Pleistocene is well-established (Guthrie and Matthews, 1972; Hoffmann, 1976; Macpherson, 1965; Rausch, 1963). Most of the mammalian members of this fauna are Holarctic in distribution, but some are represented by Palearctic and Nearctic species pairs, in the case of the marmots, M. camtschatica and M. broweri, re- spectively. These species pairs probably stem from ancestral popu- lations that were once a part of the Beringian theriofauna, but which diverged at an earlier time than the Holarctic species, or else diverged very rapidly in the late-Pleistocene to Holocene. Taxonomically, we propose that the above relationships be reflected by employing the superspecies concept (Amadon, 1961) —we place Marmota broweri and M. camtschatica in the super- species camtschatica, and M. caligata, M. olympus, and M. van- couverensis in the superspecies caligata (see Specimens Examined). ACKNOWLEDGEMENTS Hoffmann studied marmots in the Soviet Union on an Exchange Fellowship jointly sponsored by the U. S. National Academy of Sciences and the Academy of Sciences of the U.S.S.R. Many Soviet colleagues were particularly helpful, including I. M. Gromov, I. I. Sokolov, and N. K. Vereshchagin, of the Zoological Institute, THE RELATIONSHIPS OF AMPHIBERINGIAN MARMOTS 37 AN SSSR, Leningrad (ZIN); the late V. G. Heptner, and O. L. Rossolimo, of the Zoological Museum, Moscow State University (MSU). Other colleagues who provided assistance in collections under their care include the late K. Zimmermann, and R. Anger- mann, Humboldt Museum, Berlin (ZMHU); K. Bauer, Wien Naturhistorische Museum, Vienna (WNM); G. Corbet, R. Hay- man, and J. Hill, British Museum (Natural History), London (BMNH); S. Anderson, K. Koopman, and R. Van Gelder, Ameri- can Museum of Natural History, New York (AMNH); H. Setzer, National Museum of Natural History, Washington, D.C. (USNM); J. Moore and J. Hassinger, formerly of the Field Museum of Natural History, Chicago (FMNH); P. L. Wright, Zoological Museum, University of Montana, Missoula (UM); E. T. Hooper, Museum of Zoology, University of Michigan (UMMZ); W. Lidicker, Jr., Mu- seum of Vertebrate Zoology, University of California, Berkeley (MVZ); I. McT. Cowan, University of British Columbia, Vancou- ver (UBC); R. L. Rausch, formerly of the Arctic Health Research Center, Anchorage and Fairbanks, who also supplied a tape re- cording of M. broweri vocalization, and blood samples; D. P. Barash, University of Washington, from whom we obtained M. olympus blood samples; and N. N. Vorontsov, formerly of the Institute of Cytology and Genetics, Siberian Branch, Academy of Sciences of the U.S.S.R., who made available a blood sample. A. A. Nikol’skii, Moscow State University, made available to us sonograms of marmot vocalizations. From the University of Kansas (KU), J. W. Bee cast foot pad impressions, W. E. Duellman made available audiospectrographic equipment, Deb Bennett drew most of the figures, and Rose Etta Kurtz and Janet Jackson typed the manuscripts. Ljerka Deutsch assisted Nadler in the laboratory. The sugges- tions of Rausch, Nikol’skii, Setzer, and Vorontsov greatly improved an early draft of the manuscript. Hoffmann was supported by grants from the U. S. National Academy of Sciences, he and Koeppl, by National Science Foundation Grants GB 5428 and 11644 and The University of Kansas General Research Fund, and Nadler, by the Sprague Foundation. Computer time was provided by the University of Kansas Computation Center. SPECIMENS EXAMINED Localities of specimens examined, and numbers of specimens are list alphabetically below. Quotes denote information taken directly from labels; brackets denote inferences for specimens with incomplete information on labels. Locality data are ar- ranged hierarchically, from large to small administrative units, plus geographic names where appropriate. The following possibly un- 38 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY familiar abbreviations are used: Aut. Reg. — Autonomous Region; Kr. = Krai; Obl. = Oblast’; Nats. Ok. = Natsional’nyi Okrug; Rai. = Raion. Adjectival case endings (-skii, etc.) are omitted from Russian names; transliteration follows the systems used by the U.S. Board of Geographic Names, except that in Russian names e is always transliterated as e rather than ye, é, or yé in certain cases, and u is rendered as i rather than y. Equivalent U. S. Board of Geographic Names have been given in many cases; other main sources of place names were “The Times Atlas, Mid-Century Edi- tion” (Bartholomew, 1958), and the “Atlas S.S.S.R.” (Beloglazova, 1954). For abbreviations of museum collections, see Acknowledge- ments. Marmota {caligata] caligata (358 ) M. caligata caligata. (125) Canada. British Columbia: Bennett, 2 USNM; Cheonee Mts., 15 AMNH. Northwest Territories, Mackenzie District: Ft. Good Hope, 1 USNM. Yukon Territory: head of Coal Creek, 4 USNM; Kalzas Creek, Pelly R., 2 USNM; no exact locality, 2 FMNH. U.S.A. Alaska: Alleknagik [= Aleknagik] Lake, 3 USNM; Becharof Lake, 12 USNM; between Becharof Lake and Portage Bay [= Kanatak], 5 USNM; Cantwell [= Ne- nana?; Cantwell R. = Nenana R., Baker, 1906], 6 AMNH; Cape Elizabeth [vic. of], 8 USNM; head, Charlie [= Charley] R., 1 USNM; Charley R., near Mt. Kathyn [= Kathlyn?], 1 USNM; Chickamin R., 2 MVZ; Circle Hot Springs, 1 AMNH; Cordova [= Orca] Bay, 3 MVZ; lower Delta R., Slate Cr., Mt. Hayes reg., 2 USNM; Disenchantment Bay, 1 MVZ; Dry Bay, 1 FMNH; Fairbanks [vic. of], 1 AMNH; Fem Mine, Talkeetna Mts., 1 ZMHU; Flat, Beaver Mts., 3 USNM; Francis Creek, 5 mi N Ptarmigan Lake, 1 AMNH; Hinchinbrook Island, Nutchek Bay, 4 USNM, 1 MVZ; Juneau, 5 USNM; 2 mi S Juneau, 1 USNM; “between Juneau and Ketchikan,” 1 AMNH; Kanatak, 1 USNM; head, Ketchumstock Cr., 1 USNM; Kenai Peninsula, Funny R., 1 AMNH; Kenai Peninsula, Kenai Mts., 2 AMNH; Kenai Peninsula, Seldovia, 6 AMNH; Kenai Peninsula, no exact locality, 4 USNM, 5 MVZ; upper Little Delta R., Glacier Cr., Mt. Hayes reg., 3 USNM; Mt. McKinley, 1 USNM; Mt. Robert, 1 USNM; Richardson Hgy., 1 AMNH; Savage R., 1 USNM; Shelikoff Strait, 1 USNM; head, Toklat R., 1 USNM; White Pass glacier, 2 USNM; summit, White Pass, 2 USNM; Yakutat Bay, 2 USNM. ~ M. caligata cascadensis. (46) Canada. British Columbia: Mt. Baker range, 3 USNM; Spence’s Bridge, 1 AMNH; Vancouver [vic. of], 1 AMNH. U.S.A. Washington: Chelan Co., Hart Lake, 17 mi W Lucerne, 1 USNM; Lyman Lake, 3 USNM; Kittitas Co., “mountains near Easton,” 1 USNM; Mirror Lake, 2 UMMZ; Lewis Co., Tatoosh Mts., 1 USNM; Pierce Co., [St.] Andrews Park, 2 UMMZ; Mt. Rainier, 7 USNM (incl. type); Sunset Park, 2 UMMZ; Skagit Co., head of Cascade R., 8 USNM; Whatcom Co., “Washing- ton Territory” [probably Mt. Baker range, Howell, 1915], “NW Bd Survey,” 9 USNM; “Camp Chiloweyuck,” 2 USNM; Yakima Co., Mt. Adams, Goletren Cr., vic. Lava Sp., 1 USNM; “Puget Sound,” no exact locality, 2 USNM. M. caligata nivaria. (65) U.S.A. Idaho: Idaho Co., Clearwater R., Bear Cr., 1 USNM; Locksaw [= Lochsa] Cr., 1 USNM; Brushy Fork, 4 mi above mouth, 1 USNM; Valley Co., Elk Summit [1 mi N, 5 mi W Big Creek], 2 USNM; Montana: Beaverhead Co., Upper Miner Lake, 2 UM; Flathead Co., “3 mi up highway [Going-to-the-Sun Road] from Avalanche Creek, 1 UM; Elk Calf Mt., 1 UM; Going-to-the-Sun Road, 200 yds above Weeping Wall, 1 UM; head of Hallowat Creek, 1 UM; Lake Mt. 1 UM; % mi W Logan THE RELATIONSHIPS OF AMPHIBERINGIAN MARMOTS 39 Pass, 1 UM; Red Meadow Mt., 1 UM; 2 mi N Werner Peak, 2 UM; Filat- head or Glacier Co., Logan Pass, 8 UM, 3 ZIN, 1 ZMHU; Glacier Co., Piegan Mt., 4 UM; Upper St. Marys Lake, 9 USNM (incl. type); Granite Co., Cut- away Pass, E Fork Rock Creek, 5 UM; Goat Mountain Lake, 1 UM; Lake Co., Mt. McDonald, 4 UM; Lewis and Clark Co., Sugarloaf Mt., 1 UM; Lincoln Co., Northwest Peak, 2 UM; Pipe Creek, Kootenai R., 1 UM; Robinson Mt., 1 UM; Mineral Co., Trail Lake, 2 UM; Missoula Co., Carleton Lake No. 1, 2 UM; head of Dick Creek, 1 UM; “Rattlesnake Mountains,” 3 UM; Stuart Peak, 1 UM; Ravalli Co.[?], Bitterroot Mts., 1 AMNH. M. caligata okanagana. (30) Canada. Alberta: Banff, 1 USNM; Henry House, 1 USNM; 15 mi S Henry House, 1 USNM; Maligne Pass, 2 AMNH; The Monarch, 3 AMNH; E Branch, Rocky R., 3 AMNH; Tornado Pass, 5 AMNH; British Columbia: Glacier House [= station], 7 USNM; 6 mi S Nelson, Silver King Mine [= Toad Mt.?], 4 USNM; Spillimacheen R., 3 AMNH. M. caligata oxytona. (76) Canada. Alberta: entrance, Athabasca Ranch, 1 AMNH; upper Faulk Creek, Smoky R. drainage, 1 UM; head of Moose Pass branch, Smoky R., 2 USNM (incl. type); Pobokton R., 1 USNM; British Columbia: mts. near Babine, 3 USNM; Barkerville, 6 USNM; Cassiar Mts., Tanzilla R., 3 USNM; Cassiar Mts., no exact locality, 2 AMNH, 1 USNM; Coldfish Lake, 2 UM; mts. near Findlay R., 1 USNM; 80 mi NW Hudsons Hope [= Laurier Pass?], 1 AMNH; 102 mi NW Hudsons Hope, 1 AMNH; 108 mi NW Hudsons Hope, 5 AMNH; head, Klappan R., 1 USNM; Laurier Pass, 1 USNM; Level Mts., 1 AMNH; head, Little Klappan R., 2 USNM; McConnell Cr., Sustut Mts., 1 USNM; McDane Cr. [= Dease R.], Quartz Cr., 2 USNM; Moose Pass, Moose Forks [= N Fork, Moose R.?], 3 USNM; N fork, Moose R., 1 USNM; Needham Creek, 1 USNM; upper Parsnip R., above Bad R. Pass, 1 USNM; Robb Lake, 1 USNM; Sheslay R., 2 AMNH; head, Siccanie Cr., 25 mi SE Thudade [= Thutade] Lake, Sustut Mts., 4 USNM; Stuart Lake, 2 USNM; S end, Lake Thudade, 2 USNM; pass near head, Tset-ee-yuh R. (branch of Klappan R.), 2 USNM; head, middle branches, Wapiti R., 1 USNM; Wolverine R., 1 USNM; Yanks Peak, 1 BMNH; Northwest Territory, Mackenzie Dist.: Ft. Laird, 1 USNM; Yukon Territory: Ida Lake, 60 mi W Glacier Lake, 6 AMNH; Iron Mt., Glacier Lake, 7 AMNH, 2 ZIN. M. caligata raceyi. (5) Canada. British Columbia: Farrow Pass, 1 AMNH; Itcha Mts., Chilcotin, 1 USNM; Michel Peak, Tweedsmuir Prov. Park, 3 UM. M. caligata sheldoni. (2) U.S.A. Alaska: Montague Island, 2 USNM (incl. type). M. caligata vigilis. (9) U.S.A. Alaska: Berg Bay, 5 USNM; Glacier Bay [vic. of], 3 USNM, 1 MVZ (type). Marmota [caligata] olympus (35) U.S.A. Washington: Clallam Co., Mt. Angelus, 1 USNM; head, Cat Creek, 4 USNM; Deer Park, 2 UM; Happy Lake, 1 AMNH, 6 FMNH; Hurricane Ridge, 5 KU, 1 USNM; head of Soleduck [= Sol Duc] R., 8 USNM (incl. type); Jefferson Co., head, N fork, Elwha Basin, 1 USNM; Mt. Kimpta [= Kimta], 1 MVZ; head, N fork, Quinalt R. [= Mt. Steel?], 2 USNM; Mason Co., Lake Cushman [= Cushman Reservoir, Mt. Ellinor], 3 USNM. Marmota [caligata] vancouverensis (18) Canada. British Columbia: Vancouver Island, Mt. Douglas, 20 mi S Alberni, 8 MVZ (incl. type); Golden Eagle Mine, 1 MVZ; Green Mt., Na- naimo R., 5 UBC; King Solomon Basin, 3 MVZ; Mt. Washington, 1 UM. 40 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Marmota [camtschatica] broweri (29) Canada. Northwest Territory, Mackenzie District: “Arctic Red River” [= Richardson Mts.?], 1 BMNH. U.S.A. Alaska: Anaktuvuk Pass, 1 BMNH, 5 UM, 1 ZMHU; Chandler Lake, 1 KU, 7 R. L. Rausch, 1 USNM; Inuk- pasukruk [= Ifukpasugruk] Creek, SE of Anaktuvuk Pass, 1 R. L. Rausch; “Pt. Lay” [= near head, Kukpowruk R. (Rausch, 1953)], 3 MVZ (incl. type); Lake Peters, 2 KU, 1 R. L. Rausch; Cape Sabine, 1 MVZ; Cape ‘Thompson, 1 MVZ; 10 mi S Tolugak [= Tulugaq] Lake, 15 mi N Anaktuvuk Pass, 2 USNM; Ukukkminilagat [= Uqqumilaat], W of Chandler Lake, 1 R. L. Rausch. Marmota [camtschatica] camtschatica (279) M. camtschatica bungei. (186) U.S.S.R., Yakutsk. A.S.S.R. Bulun. Rai.: Kharaulakh. Mts., Bykov. Canal, 94 ZIN; “Nizov’e Leny” [= lower Lena R.], 1 AMNH, 1 BMNH, 3 MSU, 1 WNH, 30 ZIN; Kryazhkular Mts., Olomoi R., 3 ZIN; Adicha R. [= Kyuekh-Tas Mts.?], 3 ZIN; Oimyakom. Plateau, Tuora-Yuryak R., 1 ZIN; Sakkyryr. Rai.: Orulgan Mts., Bytantai R., 32 ZIN; Tas-Khayakha-Tau Mts., Selennyakh R., 9 ZIN; Yana R. [= Ulakhin-Chistai Mts.?], 2 ZIN; Verkhoyansk. Rai.: no exact locality, 1 AMNH, 4 ZIN; Khangala R. (not found), 1 ZIN. M. camtschatica camtschatica. (55) U.S.S.R., R.F.S.F.R. Kamchat. Obt.: “Kamchatka,” no exact locality, 3 BMNH, 32 MSU, 4 ZIN; Koryak Mts., 3 ZIN; Magadan. Obl.: Chukot. Nats. Ok., “Anadyr’,” no exact locality, 4 MSU, 7 ZIN; Mys [= Cape] Taran, 2 ZIN. M. camtschatica doppelmayeri. (38) U.S.S.R., R.F.S.F.R. Buryat-Mongol’. A.S.S.R.: Barguzin [vic. of], 11 MSU, 27 ZIN. SUMMARY Marmota camtschatica of northeastern Eurasia, and M. broweri, M. caligata, M. olympus, and M. vancouverensis of northwestern North America together comprise a group of related species with an amphiberingian distribution. Multivariate analysis of external and cranial morphology of these species revealed two distinct groups, one including Siberian M. camtschatica and Alaskan M. broweri, and the other consisting of the North American species M. caligata, M. olympus, and M. vancouverensis. This grouping is supported by behavioral, ecological, and. karyological data, and to some extent by parasitological and biochemical evidence. Mar- mota olympus may represent a relict survivor of the Nearctic marmot that gave rise both to the Paleartic species group that includes M. camtschatica, and to M. caligata and M. vancouver- ensis. In turn, M. broweri was derived more recently from M. camtschatica. LITERATURE CITED Amapon, D. 1961. The superspecies concept. Syst. Zool., 15:245-249. Anperson, H. L., Jn. 1974. Natural history and systematics of the tundra hare (Lepus othus Merriam) in western Alaska. Unpubl. M.S. thesis, Univ. Alaska. x + 106 pp. ANbERSON, R. M. 1934. Notes on the distribution of the hoary marmots. Canad. Field-Nat., 48:60-63. THE RELATIONSHIPS OF AMPHIBERINGIAN MARMOTS Al Baker, M. 1906. Geographic dictionary of Alaska. U. S. Geol. Surv. Bull. No. 299, Washington, D.C. ii + 690 pp. BANFIELD, A. W. F. 1974. The mammals of Canada. Univ. Toronto Press, Toronto, Ontario. xxv + 438 pp. BarasH, D. P. 1973. The social biology of the Olympic marmot. Anim. Behav. Monogr., 6:171-245. BarasH, D. P. 1974. The social behavior of the hoary marmot (Marmota caligata). Anim. Behav., 22:256-261. BARTHOLOMEW, J. [ed.]. 1958. The Times Atlas of the World. Mid-Century Edition. 4 vols. The Times Publ. Co., London. Bee, J. W., and E. R. Hari. 1956. Mammals of northern Alaska. Univ. Kans., Mus. Nat. Hist., Misc. Publ. No. 8. 309 pp. Betociazova, O. A. [ed.]. 1954. Atlas S.S.S.R. Glav. Uprav. Geod. Kartog., MVD S.S.S.R., Moscow, 147 pp. Brsixoy, D. I. 1967. Gornye surki Srednei Azii i Kazakhstana. Nauka, Mos- cow. 199 pp. BIESEMEYER, D. G. 1960. Research on hoary marmot. Unpublished MS. (typescript ). ii + 37 pp. BIKHNER [= Biichner], E. A. 1888. Nauchnye resul’taty puteshestvie N. M. Przhelval’skovo po Tsentral’noi Azii . . . Vol. 1, Mlekopitayuschiya. Imp. Akad. Nauk, St. Petersburg. 232 pp., 25 pl. BuiumserG, B. S., A. C. ALtison, and B. Garry. 1960. The haptoglobins, hemoglobins and serum proteins of the Alaskan fur seal, ground squirrel and marmot. J. Cell. Comp. Physiol., 55:61-71. Cuernyavsku, F. B. 1962. O sistematicheskikh vzaimootnosheniyakh i istorii snezhnikh baranov Starova i Novova Sveta. Byul. Mosk. O-va. Isp. Prir., otd. Biol., 67: 17-26. CuernyAvsku, F. B. 1976. Sistematicheskie vzaimootnosheniya nekotorykh nazemnykh mlekopitayushchikh Starovo i Novovo Sveta v svyazi s problemoi Beringii. Pp. 383-390 in Beringiya v Kainozoe, V. L. Kontrimavichus, ed. Acad. Sci. USSR., Far East Sci. Center, Vladi- vostok. Cowan, I. McT., and C. J. Guicuer. 1956. The mammals of British Colum- bia. Brit. Col. Prov. Mus., Handbook No. 11. Victoria, B.C. 413 pp. DaLouEsT, W. W. 1948. Mammals of Washington. Univ. Kansas Publ., Mus. Nat. Hist., 2:1-444. Davis, D. E. 1966. The moult of woodchucks (Marmota monax). Mam- malia, 30:640-644. Davis, W. B. 1939. The Recent mammals of Idaho. Caxton Print., Caldwell, Id. 400 pp. DeEVos, A., and D. I. Gmtespre. 1960. A study of woodchucks on an Ontario farm. Canad. Field-Nat., 74:130-145. Drxon, W. J. [ed.]. 1975. BMDP Biomedical Computer programs. Univer- sity of California Press, Berkeley. 792 pp. ELLERMAN, J. R., and T. C. S. Morrison-Scorr. 1951. Checklist of Pa- learctic and Indian mammals 1758 to 1946. Brit. Mus. (Nat. Hist.), London. 810 pp. Fepyx, S. 1970. Chromosomes of Microtus (Stenocranius) gregalis major (Ognev, 1923) and phylogenetic connections between sub-arctic rep- resentatives of the genus Microtus Schrank, 1798. Acta Theriologica, 15:143-152. 42 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Foxanov, V. A. 1966. Novye podvid surka-baibaka i zamechaniya o geo- graficheskoi izmenchivosti Marmota bobac Mull. Zool. Zhurn., 45: 1862-1866. Gauxina, L. I. 1962. K taksonomii surkov (genus Marmota) yuzhnoi Sibiri. Akad. Nauk, S.S.S.R., Sib. Otd., Trudy Biol. Inst., No. 8:138-156. GrizzELL, R. A., Jr. 1955. A study of the southem pan eo Marmota monax monax. Amer. Midl. Nat., 53:257-293. Gromovy, I. M., D. I. Brrxov, N. I. KALABuKHov, and M. N. Memr. 1965. Fauna S.S.S.R. Mlekopitayushchie. Vol. 3, No. 2. Nazemnye Belich’i (Marmotinae). Nauka, Moscow—Leningrad. ii + 466 pp. Guturiz, R. D., and J. V. Marruews, Jr. 1972. The Cape Deceit fauna— early Pleistocene mammalian assemblage from the Alaskan arctic. Quat. Res., 1 (for 1971): 474-510. Haz, E. R., and R. M. Gmmore. 1934. Marmota caligata broweri, a new marmot from northern Alaska. Canad. Field-Nat., 48:57-59. Haut, E. R., and K. R. Ketson. 1959. The mammals of North America. 2 vols. Ronald Press, New York. xxx + 546 + 79; viii + 547-1083 + 79 pp. Hamitton, W. D., Jr. 1934. The life history of the rufescent woodchuck, Marmota monax rufescens Howell. Ann. Carnegie Mus., 23:85-178. Harpy, G. A. 1955. The natural history of the Forbidden Plateau area Van- couver Island, British Columbia. Pp. B24-B63 in Prov. Mus. Nat. Hist. and Anthrop., Prov. Brit. Col. Rept. for the year 1954. Victoria, B.C. HorrMAnn, R. S. 1971. Relationships of certain Holarctic shrews, genus Sorex. Zeitsch. f. Siugetierk., 36: 193-200. HorrMann, R. S. 1974. Terrestrial vertebrates. Pp. 475-568 in Arctic and alpine environments. J. D. Ives and R. G. Barry, eds. Methuen, London. HorrMann, R. S. 1976. Ekologicheskii i zoogeograficheskii analiz migratsii zhivotnykh cherez Beringiiskii most sushi v Chetvertichnom periode. Pp. 354-367 in Beringiya v Kainozoe, V. L. Kontrimavichus, ed. Acad. Sci. USSR, Far East Sci. Center, Vladivostok. HorrMann, R. S., and C. F. NApter. 1968. Chromosomes and systematics of some North American species of the genus Marmota (Rodentia: Sciuridae). Experientia, 24:740-742. HorrMAann, R. S., and D. L. Pattie. 1968. A guide to Montana Merrit Univ. Montana Print. Serv., Missoula. x + 133 pp. HorrMann, R. S., and R. S. Pererson. 1967. Systematics and zoogeography of Sorex in the Bering Strait area. Syst. Zool., 16:127-136. HorrMann, R. S., and R. D. Taser. 1967. Origin and history of Holarctic tundra ecosystems, with special reference to their vertebrate faunas. Pp. 143-170 in Arctic and Alpine Environments, H. E. Wright, Jr., and W. H. Osburn, eds. Indiana Univ. Press, Bloomington. HorrMann, R. S., P. L. Wricut, and F. E. Newsy. 1969. The distribution of some mammals in Montana. I. Mammals other than bats. J. Mamm., 50:579-604, Hoxianp, G. P. 1963. Faunal affinities of the fleas (Siphonaptera) of Alaska; with an annotated list of species. Pp. 45-63 in Pacific Basin Biogeog- raphy, J. L. Gressitt, ed. Bishop Mus. Press, Honolulu. Howe tt, A. H. 1915. Revision of the American marmots. N. Amer. Fauna, No. 37. 80 pp., 15 pl. THE RELATIONSHIPS OF AMPHIBERINGIAN MARMOTS 43 Jounson, M. L. 1974. Mammals. Pp. 1-87 in Biochemical and immunological taxonomy of animals, C. A. Wright, ed. Acad. Press, London. Karironov, V. I. 1960a. Rasprostranenie chernoshapochnovo surka (Marmota camtschatica Pall.) Byul. Moscow O-va. Isp. Prirody, otd. Biol., 65: 5-15. Karrronov, V. I. 1960b. Parazity chernoshapochnovo surka (Marmota cam- tschatica Pall.) Zool. Zhurm., 39:1435-1437. Kaprironovy, V. I. 1963. Ecologiya chernoshapochnovo surka i perspektivy evo khozyaistvennovo ispol’zovaniya. Avtoreferat, Akad. Nauk, Kirgiz. SSR. Alma Ata. 23 pp. Kapironov, V. I. 1964. Lin’ka surkov. Akad. Nauk, Kazakh. SSR, Trudy Inst. Zool., 23:169-190. Kaprronov, V. I. 1969. Ocherk biologii chemmoshapochnovo surka (Marmota camtschatica Pall.) Zool. Zhurn., 39:448-457. Korositsyna, K. B., C. F. NADLER, N. N. Vorontsov, and R. S. HorrMANN. 1974. Chromosomes of the Siberian snow sheep, Ovis nivicola, and implications concerning the origin of amphiberingian wild sheep (sub- genus Pachyceros). Quat. Res., 4:235-245. Kozxiovsxu, A. I. 1974. [Karyological differentiation of North American sub- species of hoofed lemmings] Dokl. AN SSSR, 219(4):981-984. Lyapunova, E. A., and N. N. Vorontsov. 1969. Novye dannye o khromo- somakh Evraziiskikh surkov (Marmota, Marmotinae, Sciuridae, Roden- tia). Pp. 36-40 in Mlekopitayushchie. N. N. Vorontsov, ed. AN SSR, Sib. Otd., Novosibirsk. Macpuerson, A. H. 1965. The origin of diversity in mammals of the Cana- dian arctic tundra. Syst. Zool., 14:153-173. Moore, J. C. 1961. Geographic variation in some reproductive characteristics of diurnal squirrels. Bull. Amer. Mus. Nat. Hist., 122:1-32. Naver, C. F., R. S. HorrMann, R. I. Sukernick, and N. N. Vorontsov. 1974. A comparison of biochemical and morphological evolution in Holarctic ground squirrels (Spermophilus). Trans. Ist Intern. Theriol. Cong., Moscow, 2:13-14. Naver, C. F., E. A. Lyapunova, R. S. HorrMann, N. N. Vorontsov, and N. A. Matycina. 1975. Chromosomal evolution in Holarctic ground squirrels (Spermophilus). 1. Giemsa-band homologies in Spermophilus columbianus and S. undulatus. Zeitsch. f. Siugetierk., 40:1-7. Nanb-er, C. F., V. R. Rauscu, E. A. Lyapunova, R. S. HorrMann, and N. N. Vorontsov. 1976. Chromosomal banding patterns of the Holarctic rodents, Clethrionomys rutilus and Microtus oeconomus. Zeitsch. f. Saugetierk., 41:137-146. Naver, C. F., N. N. Vorontsov, R. S. HorrMann, I. I. Fomicuova, and C. F. Napier, Jr. 1973. Zoogeography of transferrins in arctic and long-tailed ground squirrel populations. Comp. Biochem. Physiol. 44B: 33-40. Nixou’sxu, A. A. 1976. Zvukovoi, preduprezhdayushchii ob opasnosti signal surkov (Marmota) kak vidovoipriznak. Zool. Zhur., 55:1214-1224. OcneEv, S. I. 1947. Zveri S.S.S.R. i prilezhashchik stran. Vol. 5. Gryzuny. Akad. Nauk S.S.S.R., Moscow—Leningrad. 809 pp. Petrov, B. M. 1961. K sistematike i vozrastnoi izmenchivosti surka Mensbira (Marmota menzbieri Kaschk.) Zool. Zhurn., 40:93-105. PizzIMENTI, J. J. 1975. Evolution of the prairie dog genus Cynomys. Univ. Kans. Mus. Nat. Hist., Occas. Pap. No. 39. 73 pp. Ad OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Poote, D. A. 1950. Report on Marmota flaviventer. Unpubl. MS (type- script). ii + 14 pp. Porsicp, A. E. 1945. Mammals of the Mackenzie delta. Canad. Field-Nat., 59:4-22. Ranp, A. L. 1954. The ice age and mammal speciation in North America. Arctic, 7:31-35. Rauscu, R. L. 1953. On the status of some arctic mammals. Arctic, 6:91-148. Rauscu, R. L. 1963. A review of the distribution of Holarctic Recent mam- mals. Pp. 29-43 in Pacific Basin Biogeography, J. L. Gressitt, ed. Bishop Mus. Press, Honolulu. Rauscu, R. L. 1976. The genera Paranoplocephala Liihe, 1910 and Anoplo- cephaloides Baer, 1923. Ann. Parasit. Hum. Comp., 51:513-562. RausH [= Rauscu], R. L. 1977. O zoogeografii nekotorykh Beringiiskikh mlekopitayushchikh. Pp. 162-177 in Uspekhi Sovremennoi Teriologii, V. E. Sokolov, ed. Nauka, Moscow. Rauscu, R. L., and V. R. Rauscu. 1965. Cytogenetic evidence for the specific distinction of an Alaskan marmot, Marmota broweri Hall and Gilmore (Mammalia: Sciuridae). Chromosoma, 16:618-23. Rauscu, R. L., and V. R. Rauscuw. 1968. On the biology and systematic pasi- tion of Microtus abbreviatus Miller, a vole endemic to the St. Matthew Islands, Bering Sea. Zeitsch. f. Saiugetierk., 33:65-99. Rauscu, R. L., and V. R. Rauscu. 1971. The somatic chromosomes of some North American marmots (Sciuridae), with remarks on the relationships of Marmota broweri Hall and Gilmore. Mammalia, 35:85-101. Rauscu, R. L., and V. R. Rauscu. 1972. Observations on chromosomes of Dicrostonyx torquatus stevensoni Nelson and chromosomal diversity in varying lemmings. Zeitsch. f. Siugetierk., 37:372-384. Rauscu, R. L., and V. R. Rauscw. 1975. Taxonomy and zoogeography of Lemmus spp. (Rodentia: Arvicolinae), with notes on laboratory- reared lemmings. Zeitsch. f. Siugetierk., 40:8-34. Riwcway, R. 1912. Color standards and color nomenclature. Priv. Print., Washington, D.C. iii + 43 pp., 53 col. pl. Rosinson, J. W., and R. S. HorrMann. 1975. Geographical and interspecific cranial variation in big-eared ground squirrels (Spermophilus): a multivariate study. Syst. Zool., 24:79-88. SwartH, H. S. 1911. Two new species of marmots from Northwestern America. Univ. Calif. Publ. Zool. 7:201-204. SNEATH, P. H. A., and R. R. Soxau. 1973. Numerical taxonomy. W. H. Freeman, San Francisco. xv + 573 pp. TAULMAN, J. F. 1977. Vocalization of the hoary marmot, Marmota caligata. J. Mamm., 58:681-683. Vorontsov, N. N., and E. A. Lyapunova. 1976. Genetika i problemy trans- beringiiskikh svyazei Golarkticheskikh mlekopitayushchikh. Pp. 337- 353 in Beringiya v Kainozoe, V. L. Kontrimavichus, ed. Acad. Sci. USSR, Far East Sci. Center, Vladivostok. Vorontsov, N. N., E. A. LyApunova, and N. G. ZAcGorurKo. 1969. Sravnitel’- naya kariologiya i stanovlenie izoliruyuschikh mekhanizmov v_ rode Marmota. Zool. Zhurn., 48:317-334. Wanrinc, G. H. 1966. Sounds and communications of the yellow-bellied marmot (Marmota flaviventris). Anim. Behav., 14:177-183. Wenn, H. 1935. Zur Osteologie der Gattung Marmota Blumenb, (Arctomys Schreb.). Zeitsch. f£. Siugetierk., 10: 1-32. THE RELATIONSHIPS OF AMPHIBERINGIAN MARMOTS 45 YouncMAN, P.'M. 1975. Mammals of the Yukon Territory. Nat. Mus. Canada, Publ. Zool., No. 10. 192 pp. ZiMINA, R. P., and I. P. Gerasimov. 1973. The periglacial expansion of marmots (Marmota) in middle Europe during late Pleistocene. J. Mamm., 54:327-340. ZIMMERMANN, K. 1942. Zur Kenntnis von Microtus oeconomus (Pallas) Arch. Naturgesch., N. F., 11:174-197. OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 46 , z I z I ae I Payee rl I othe wee eee LL6 SOI 986 626 196 Tes S68 ZI6 G68 C18 G6 C16 816 C68 O88 LB —————— FG Ct OOF OFF Of OOF O'Ty OOf OE SE Lee Ose SR Oe ee oof Gor. ee ger 08. CSF OTF O'Th,, ae O98 GIy SOR O'Ne TOP OTe Oe Sle S18 “fee aoe ra ote OCs S63 O09 s8'63.' 9'8c OTS LOE dO LOS SOE OLE O88 C0F S66 O20” ee ae cS: OOL QT OSt Cat ht OLT LOT SS. SSI LIicg= SET’ OF} OGL TSE Cll] eee Sor Lor rr of Str er Iss sor 6 cee cir Fir SOF OGG SSE ht =... cOe Gif T0e Qos oe cee §2e .8'8S. OLE O10 SOG 88S. .880 L8G Bet. Ole. eee SI Lig @8o 16 VlLo g'00 GOOG GE BSS CVC Sts PrLe “Os C85 BS (86 Cle 62 ee LI Leo @9c .¥8 O88 ISO SEG FG F'lS 890 LOL Ce G0c OM OW 80S GEL... eer rer Of GIF-Ocr Gcr Sar Gle cor sc Pre = i. ae b € G I AaqunN aque, Sm tee to L216 Pie 826 S88. O04 Sol 86&L sco SIs 803. ¥OG.° COC FOL = = eee Cl G10 O16 T00 LOS Yas OFF O88 Lt gs Sle 88s ses 0s Ch. [dle O'S 66h Ld Lb Toh Ss'Sh GE Oh OFF BO CSL OOF GSh Cr Bh OCP he oes 999 O38 sIg s'og s'og Tor soe Tor sp do0s OL CL Sr Bar GW = — = (L1) (91) (s1) (FT) (e1) (61) (1) (ol) (6) () (2) (9) (9) =) () (ft) sloquinn AyTeoory {DIOUUDP OVUL JO SONI[LOO] FE A0J sozs o[dures puv so109s ploquey—' [-y Xrpueddy 47 THE RELATIONSHIPS OF AMPHIBERINGIAN MARMOTS ‘SO[QUIAVA TOF T O[GQUT, 095 y BOR ITSOO 1OF G ‘AT 995 » WHS OLOM SOpUU ON _ 1 ee a | OL ee eae en a Ee OR EP i a SCor FOOT 966 L2G TG 6:96 T'IOT 9°66 O26 LOOT T'ZOT 686 O26 006 F001 046 “A £9 aap S0h OSb OCF POP Teh Bab OOF Sth Sth Sah SEP Oh OOP SLh O'fR NEE £6 Bey 68th 68 Ob UOT 60h 60h 60h «CTR SOP «CSR OCR OCS OF «COT: «COPS TP '8 pee te eae GIe. Scb (TS O86 PIS OBE CHE ORE SER OTS GTE gOS OOS O“E gis gg eee 16 4st OOf CAT OOT SST TEL Sal OAT SRT SLT TOL ODT OT OFT PET Gut ee eee aay 6SGP 6RLY 6UORP 6 6ORR UP OUP OOP OSS 6S 6 6Of'OP OCP DR 6OCTED «6 CGP 6G 61 OG GIs OTe Ges .O6F TOR STS o6s OO COE Se TIE 88S “Ro 61s gg or Peo 488 PLS SOE NRT HLS (ORD SLB Os «TRS OGLE «S'OS «(ORS «OFS «CO'ks)«= Sg LI ayo IS Gets PIS GIS PES ONS OFS GER QOS PHT SEs Bw Sts ss os 91 Oop ash Luby VTP 0h OCF Sk «Oh UORh Loh 6S OUGth USER 69ee 6UCRER 68S SI Sey 6p 6OTh 6CLE 6SOE 6TH 6OCk 6OCP 6ULGS «UCOTF UPS «CLO «CBE «OSE «Th OTR oyo 68 6989 «6809 00 TD «(0109 L'20'PO OSD LPO BZD S«'8D 8D S0)L'8Q el St5 Lys 98S TSS OBS FSS BBS RBS FB NB «PER COS «COTES: «CSS «OLS ) «6Fee ~ 81 Set wer O88 ALT GAT O'8T GBT SLT SLT O8T GST gel SST TST ssl sZt II Oot S84 SOT LOT COT GAT BOT PLT SOL LOT LOT OMT Sl SI OOL S9t ~ Ol Gite s46 GTE 98S CHS [6S VB Ves OVE PEE BIC TE C83 BLS BIE TIQ 6 Oot OFT SLL COT ORT SST [St SST OOT OOT SO GST SOT OLT 8ST TT 8 alo hs [TPS 8S HES TES GES HHS VHS LES HGS LW OV GBS LP Bz L oo, 88 @& OOF G6. 86 6. 88. Le Go. GE. O68, DE BRO Ge TE 9 Ser Rot LOL GOL «(OBL «OCT «CST «PST 6USSL «ULC. SET «UCORT 6ST: 6UCSIT €«USRT: OCTET s O'S «=—LSS 68S 90S «=—CL0S: «CLS «O'S LBS COLTS «OVS «CBS COLTS «(COWS «CPUS: «CSRS CBOS v 68S O86 S8S PPS «(OSS «68ST «HES OLS «HLS «6S «HS «OCTBS OCSLS OSHS «CGS OS'RZ 7 Ach 96h Leh Chr Sch 6TLb 68h 6LLb 6UTLb #(OGh 69Gb OL Lb 6OSLh OSPR 6OCS'GhO'OF 6 Oya, Tes Lee LOk Ber O88 BES sos Lee Tso LPS TI TTS Ter ope 9Tg I (pe) (66) (Ge) (te) (oe) (6s) (ee) ~«s) (98) ~«(9a)-«s (we) (ea) (ea) te) so) a (70d “equuinyy Ayt[Boo'T ) ————— ee ee Se eee (‘qUOD) *T-y xipueddy AequNN aqeueA OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 48 Zz eI 6 I 9 € I 91 LI é Or L OF6 96 CGS6 LSS LSS ZTSsS 6OS TI6 S88 LP T98 LIS Ow @cr Gtr OOF Se @Se @ce Fee Gosc ace Lee Cie Cir €br Gir OJ€ S9e, O86 OOS See PLE Ose cer Cee O0€ 666 966 O16 866 LES OTIS 660 T0E O86 O66 O1Z Di Sol FL O81 SE “LR Ol B6l Tel <80c 221 eer Vor Sr Str LO ese sae Loe PE Ess BEE MLE LSE Bec .660 S08 <@io G22 oc TOs CSc. Cee es Bee 9c Ops GOs 80S Gre O82 O88 £1C S0c Ss eG Bre wee Bic 2c rao 0c S00 LOI GLI Sie GOs BEI O02 BT iy ser er. ‘Ose “GRE OSS Wicc EGS O86- De Oe Ree Bue “vee J6e 066 602 Of Oe- B60) bs Oise “BSE Gee 959 Leo 989 ¢1e SIs GPS Ses Tes Srp gea seg OTS Gee Geo Le F0C LOS 0c O0C GI ¢E0c Gos 002 LEI Gn 2. 621 SIGS GSES C88 “SS S6ics Sie ome ial ss9 839 L09 969 69S G6S 66S 9F9 619 T€9 GS9 969 LT19 GE9 ¢09 DLI6S 819-2 €1 FASC emOGG PASC me SEC aghiGG O'C6en LCG SSG GS L6Galvece Veo Sic OeceUico -Ciscee ollo a eeumueeeeencE q'sT 08ST 9 LI 681 SAIL 691 SLT G8I GLI OLT L LI SLI oLIT 6 LI 6 LI TLI AL oo II CLI GOL VOL TLI 09T LOT VOL COL 991 L&T T9L 991 6ST TSI 9ST OST cr ne Or Croe 686 GIE GOE S8é6I 976 GIG GCE 666 066 Vre €66 $96 T86 696 €96 TSG OLI OLL S&T CLI COT SST 6&1 STI 671 9ST 9FI G9L O9T TLI VOT GLI Lo; = a O€G $96 €66 VEG E16 OF LIG PVG LG L&G 676 VES O86 CGSVEG T66é 9G TSG NGw aes oS akGGe ds 06 “6 « OG AcGSs- PES Ee SCPUIIGG TERE "GS Oh mG ice tis cee OFrTL OGL GGL SGI 611 CTL GL OST Tél OTL 9GL YG CGL 9ITI GGL LGI I OGG LIG SIG OG O6T L6L 006 S816 GIG L1G 966 VIG TI6 661 L0G 00¢) 60G il 696 -696 996 L8G O€6 666 8€6 986 696 GLE G66 VL6E €L6 Y9C LSE GSGe T9G°- Gly 18h OLF O8F YF Grr OS LSP GLP CLP Gor lr 89% V9r “Sor USL JOS Sale Tis oO¢€S 96rF LEG EOF SLY CLF OFS GIS é6IS LYS GGG TIS CG6r ¢0S 68F O6F aap RCS) SeZe) GEe (6S) (8) (0S) a6) (Go) 5G) 59s) (Ga) (Fa) (SG) (06) (Gl) a(S) ry z(‘}UOO “‘IaquinNy A}][eO0T ) Ke) A rt | | | | | | | AAO Ht 10 Or 0 OD (qU0D ) *Z-V xIpueddy loquinyy syqeLe, OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 50 "re LS906'LEZI- LISPS'ZOSI-— LOPGLOFET— OFILT'9IZI- LESSES9ZI- YOFIPESII- FSPLESSZI- OODGLSE9TI- 76106 CSOI— LNV.LSNOO TS8¢9'°0 PESSI [— 0€660 G— 67960 0- 6LEIT I- 9€9L0'0 S8cscI- . 88L¢c9'0- 9S6L0'°0 OTVAOO FG PELZT'O- VELIS I- SL890°0- 06976 0- 9PIEs8'0- 0G6GS' I- €6979'0- L89FL'0— OPPLL I- OONTYA €6 8E99T S- L9860'¢— €VSSS E— €86S0'€— VESGS E— TOL8E°S- O9TOT G- OTI6L I- 96°69 G- VNNIUd GE S80PGS 66L6L'7 6L8E6 7 9LOPPY'9 S6SIE'V £60682 809L¢6'9 S9EEO'L 96280'9 O00dHd 1G PIETS9 6GL88'P POSTS'9 81909 7 cPp9orT'S L869 PITPL L IPLYS L LEVEL VNOdHd 0G LL8L6'T 96S0L'S ESSE P 8560'S GLIOIG'S ccr09'0 €61SP'0 OFF9T 0 TEo0so'0- GMLSYW 6I1 SOO00L'GS 099F8'SS EVLEE' 8S O€SOS'ES 6S16E°9G 96699 FG 63C8F 9G LY89OF' LS €1T960°SS LHIOOO 8T SLLOS'S 690ST'S L6869'E 9ESCS E P6606 € GILSGE 6€960'9 C809 F SCIC6'E LIHXVW LI GE889 L- 88EP6'9- 6LZ10°6— 8SGP6 F— €3L89'S- yS9ST OI- LEOEF SI- G66EL'0I- €6SEC E1- LHLSOW 9T PSE6L PV 90EOTE . CEOLe'P ES86E'E OZGL88'°E C6E6TS COLLL'S OS88F's COL6SS VOUROT ST OFZ8S S— O89LE'9- L8TTE'9- PLEOG 9— 97066 S— SE661T'9- GEOG S- 78996 S—- 99188 F— SVNOT FI L8EEH'E GLYLLYV GL6E8'E €0L90'9 LI€96'S P8898'F 6EES9'S 60ZGo'E 9S6L0°F GMODAZ CT 6PIPG TE POLEL LE 9896 GE 996L8 TE S810 GE GLSPP 8G E1T69L OE 90LLE' LG 9SS9T'9G WLANST GI 98188'€- 61E69' F— GL6GS'E—- LGLGV F— VE89L F— VELGL E— 61GL0°€—- $9006 'G- ClPIL G- OTOGAL IT SOTE6 P €ePrso'L 90188'F P6PPS'9 SO6ET'S €96L0'P LI9VGL PV TLOGE' FP 9FE8STE GMOGAL OT PSISL F— ISTtLS'¢- 9SLIE F- VIS96 F— EE 169 F— OFELS'E— S680L'¢- CO0OLL €- LOOsS"¢—- YOVSOT 6 PEEPS 1S 6PS89'8T LLOI6 GG IGE9E SI 8669906 COTLL IG GLIPGE ES TESSL'SS COICE GS woOodaM 8 PLI88'S- PSOTL'8— 88189°8- 9SLPE 8— €S8scPr's— O9TT9'9— OFEO8' L- OFFSL L- TLLIS’8— OLNIGM L Cesl9 6I S8P8S 0G PSGES LT 9SP9E'0G 8090€°8T SCOLE'ST C6IGITST OGFSS FI 9E6FS ST LHYWOA 9 €1689° L- T1669 7— S6609'°¢—- 88868 F—- OOTZ9'S— LLIEL'9— 87690 L- €lsoFrs- Lgsee'o- GAMYWOSA S$ Se6L9' I- LEOOF'E €89€8'0- 698TL I- OGLE6G I- C60FP G— T69£0'% 99€09°0- GFFOS'O GM.LSOU F PPr60'0- IOTS3'0- O8F89'T 968PE I- PE6IO'T GPSLG'T SI8e0'e 9SL9OC'’e S9ZOT'E OWISOU € ISLPS SI LIIS€' FI 8SOIG ET 9LOFS IT S88Z0'eT SSLI6°ET 0¢S90 FI SI86S°ST £0SLo'ST LYVLOT & GELOS'0- OS90T E- 0ZS69'0- 009L6°0- €00FS' I- TG8 > oiemee "0 ee hen a Q'00T 001 Mano “W (F) a tate aac. cee. Sle ee. eee 8'€6 aw! eS! oS. oa. ll, Plena L'96 0'00T ete" Seee eee ee ae eee. ee: 0'00T th” | oes tee) oe. eee rc ne O00 O'001 ie) oes | eee) ae ye pea 0°001 aa sree. ome nodes. eee eee 1 eee O'00T =: O'00T DIYDYISUDI “9 “W (€) rasung ‘0 "‘W (3) uahvujaddop ‘9 "Ww (1) £0 OF £0 OF FO OF €O OF £O OF £O OF FO OF €O OF SO OF £0 OF €O OF £O OF ©O OF €O OF FO OF FO OF €O OF *O OF (6) (9) (2) (9) (¢) (#) (€) (3) (1) 994409 % ‘DIOULID I[eULa} puv J[eUL Jo sIsATeUR JC WoIFZ s[qv} UOHKOYIsse|Q—Z-gq xIpueddy 53 THE RELATIONSHIPS OF AMPHIBERINGIAN MARMOTS ‘T PIFBL 99S 9S6'8 6000°0 VG SOLV Le oe ee ee ¥G LOE'6 G000'0 €6 LES s = eee Sela ee ee GMNGAL OT £6 0S9'6 6000'0 6G rh acl ae Arica el ileal i shire seecyraiteain 5 SVNOT PFT GG 8866 6000'0 IG SoVIG” RRR NR To eer a an LIHXVW LI 1G 89¢'01 6000'0 06 LOSSG 905 Se a rcs VOURST ST 06 S8Z'0T €000'0 61 Ci 2 || A Ads > bene os a LYVLOT @ 61 VLI'IT £000'0 SI OL8eV “STS nN \. mien © 5 aC VNNIYd G6 81 veg IT ¥000'0 LI OoCaY ~ “irglen 7 aes... aan OONIYA £6 LI 966° TT s000°0 9T DOOD: SE Be a eee ae oar: - sLHLSOW OT 9T L8E°GT 4000°0 sil OSGGS' “Tafel y —" Sige <— eeeee YwOOddM 8 SI 6S8'°CT 60000 VI DLO. “re - eee OTVdOO F¥6 As OIG ET 6100'0 €T SECO “Sseea yy ee” ae LHIO9O 81 el VOL'ET 91000 GL LOSE. “KGS kh.” Sean See eee YOVSOT 6 rat Ivo Pl €¢00'0 IT [Ck6'S “erie Sve ey Re ONV.LOT T IT S80°ST T€00°0 Or GOGOL “Sati (55a s ere CGMLSOW F OL O86'ST 6r00'0 6 CSLE LL) eee eR ee LHVWOA 9 6 S60°LT 0900°0 8 (A013) Orr) ) Me os: 3 Gane. © Dyes GMODAZ ET 8 OfS'ST S800'0 L SEG TL “fesae |... Set + eno OOOdHCd 16 L 11961 6€10°0 9 OOSu TL “Cesare 2 SERA ane OWTSOU € 9 GEG 1G 8660°0 Sj COSC OL “ei PA. See hears GMILSVW 6I 4 SEV VG SS€0°0 Vv OGCOS 6 “Gvaeer. Gata. eae OLNIGM 2 v 918'0E 8€S0'°0 € SOSS TC eka in Aa ne GMYWOA & € VPO'9E G9TT'O G COC CG pwn! eas 0 ee emma VNOdHd 06 G SLE"6V 8PLG0 I O22) OF ..4 Sieeiats fn nae a ey ULdNAOT GI I O1}SN%3S- 7 onsyeis- pepnyouy Ia}UY 0} pe1auy Jaquin yy ejyeulrxoiddy SOTqeLIvA JO ‘ON onyeA 2 9[Geued dais ‘DIOWM A, Beit puv s[eUr Jo sisA[euL Jd astadoyg jo a[qey, ArewuIng—e-g x1pueddy OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 54 8IFF9'0- 6£780'0 SOPSS'0- 0L08S°0 C9TEF'O- F606S'E TS€88'I- COLES E a= (ST) €80%8'0- €8L89'0 99998'T TOSPE'T OLPSr'E PLOLLT- 62086 T ESSLEL =~ seca = (aa 80SL9'0- 08608'I- SL88E'0 F98L0'1- €SSSE'0- 6S988'0- ossoe’s— G68I6'S ~“iivizaeee T188¢'0 GSLPL'0- 6ESS9'0- 60IL1'0 PE9OL'0 SCS6I'T G0906'F coSscct alae, 79@80'0 8S68€'0 F669F'0 bPEFS 0- L1009'0- 9ZSTL'0- S6ESF'0- CSpor'S (FT) 31Z6S'0- 9SrIS'T 660L2'0 GOEFH'0 00SL6'I- LPE6E'O 18096°3 TEGER I alee AEE? 698CL'T T0S80'0 PZS8T'0 9€0S6'T T6LL0°0 FCLIG'0 L99S¢'T- ChCcey > Wn viel see ee 98PF8E'0- 96000'0 68CFT T- ELP8S'T 81656 0- 69661 I- €6SEL'0 COLE i ee ae 16F8F' I- SLLOL'I- TTL88'T GOLLV'T 808F6'I- TE6LE I- GO80F'T SFILS' CS "Sep eee a a 9SLTL'0- T3L6S'0 FF09G'0 PSTLS8'0 6ZL00°0 OLPOT'F E9FOL IT- SEGEL L. * "Sue eee 9L6LF‘0- 8E8so'T €110S'0 66S66'0 80006'F O09LL'T- ITSOT‘0 PISUES “Savas (8) PSEEE'O- 6FLOT I- 6L60F I- s6sss'0-. 1080 8hSF9'0- 8ESI6's- TI9I6T —_ (L) ELLET‘O €€1S6'0- 1S999°0- LLIEZ'O 9F408'0 ESPrIO'T OTSso'e ccary'S — —) saa ee ee €9092'0 L¥SLI'0 S686F'0- 9SLES'0- GOZ6E'0- €¢L09'0- 9TLOL'0- LOOBO'G "SS Fe Bea hae Oe €6909'0 TLZ10°0 82069 'T 6E8FS'T- 6SESS'0- OSTLO'T TOSTS'T ossIs's- tai (F) 999ST'0 GLL08'0- LZ869'0 6EC0I'T SZZS9'T SZr0e'0 E6ESS'S— GISLG'C— ~~ ie a GPELT'0- 69L9'0 9666'0- LOLS6'T- 83088'0 SSSET'O SI610'°0 co0cr S> err oe AS 0SS06'0 LOSE9'T- E10L6'S G2999'T— €LOES T- T€Z80'0- 60ETS'0 SLIVeS- "| ae ee IIIA IIA IA A Al Ill II I oAFPHBA [euBsUO a[qeiie A [eoruoueD ‘DIOULWI B[eULsy pue o[eUL I0F suvoUT dnois ze poayenyead safqeueA [eoluOURD—'f-g xIpueddy 55 THE RELATIONSHIPS OF AMPHIBERINGIAN MARMOTS ‘T P19G2L 99S o 9L6£0'0 6E8LT 0- €079E'0 SLOST‘0— 888080 1G060'0- 00S9E"0 ELECr 0- ec0rL0- == (si) 96016 0- ¢6990°0 9E09T 0 6LVGS'0- 9TE60 0- TV910'0- S6rrE'0 9PPITO LE0VSO= (an) L3960'0- LV89T 0- IGVL0'0 SEr6l0 61190'0- 08970 0- GOTOTO 907TE'O 67600.. feos a) crsre'0 TL9ESG 0 8GPOT 0- SPses 0- 16S90'0 L6ETI'O S6CPrr'0 6€7L00 LOS8S'0) aa (OL) VVLOL O- 887000 €LV1LE'0- TL970'0 18¢689'0- €08L¢'0 ccesct 0- V6Scr 0 COD27ci0) Seen AE) L8987'0 TVE19'0- €8E0T 0- 87909 0 SIGle oO T70S9°0 €9790 I[- 8098T 0 SSEsso; == (8) 91700 8L61Z'0 9LT8¢6 0- VeSIG0 9SL8¢'0 68626 0 €6870°0 PPSG6 0 OSK9LC- Seno ()) 7S600°0 601E0'0- 66886 0- vV8LI 0- S81EG 0 COPES 0- CPSLT0- 668860 o9rzs'0 20 CS(F9*»): 86000'0 VSIOLO OVLST'O LO8VT'0- 97L00'0- 67 10G0 vI0LG 0- VPSPS 0 go90r'0 ~~ -—s: (*:) 819900 9¢6990'0- SL66¢'0 €689T 0 EvVel O- €8LVG 0- 69761 0 LLIGGT SSO aw A VGECO 0- O€ PEO 89071 0 ZcOLTS'0 69610°0 SLIVFr 0- c6rse'0- ESISE'O- cpecgo0 0 tsé«C(82=) TE1L60'0 L06<0'0 96890 0- 9E971 0- V9SG0'0- y09S0'°0 98071 0 SOE8sG 0- SPelLlOr sw ae(@) LIL89'0- GLEGE 0O- C9SVE'0- C98ST'0 986ET T GVO8G T PLLSE T- IGL6G 0- VOGLe0. same) IIAX IAX AX AIX IIx IIx IX ».4 xI #°[GRHeA [EUs a[qeiieA [eoruoueD (30D ) "p-q xIpueddy OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 56 0 eP06e'8 ILLE09 ‘0 SIEIoL 2 0 osrsr'L cso6e's T10c0rL c6Loe'L LPPIO'9 s90FF'8 SOLES L OBISS'E 8Z9E'8 ‘0 T6zeLS 6P9LEL ‘0 =SIPPs'9 0 ‘paulquios soxos “DOWD JY 10F XLQVUE gouejsip ULaJN—'D xIpueddy LEVLV L O€ISL'8 68890'8 ELLLY OT PLLS6 L PPLVL'9 ‘0 61869'8 C8P6G 6 SLEPL'S OrIPY OL FrCOe 8 €0Ss8e's 809° V 0 CTLI9'6 TLZseOL FrI96 e90IT IT GS6F6'8 €9PET'9 oreso'9 GEIS F 0 —_——__—__————_ sndufgo "W (6) —________———- sisuasaanooupa “pw (8) ~~ sysuappasvo ‘9 “W (L) pupa “9 "Ww (9) (epeuec “exseTY ) ‘dss pypsyvo “Ww (¢) ——__—__———_ manoiq. “W (¥) DOWYDYISIUDI “9 “PW (¢) sagunq 9 “FW (@) i ES uaipujaddop ‘9 “‘W (T) UN) ie mtJs,. COMP 2 ie 6d OCCASIONAL PAPERS . Set hed SFesiry of the MUSEUM OF NATURAL HISTORY The University of Kansas Lawrence, Kansas NUMBER 84, PAGES 1-13 MARCH 27, 1980 A NEW MARSUPIAL FROG (HYLIDAE: GASTROTHECA) FROM THE ANDES OF ECUADOR WILLIAM E. DUELLMAN’ AND REBECCA A. PYLES” In a review of the marsupial frogs (Gastrotheca) of the Ecua- dorian Andes, Duellman (1974) noted the occurrence of uniform green frogs on the Amazonian slopes of the Andes. Subsequent field work resulted in the acquisition of series of these green frogs, which are similar to G. plumbea (Boulenger) on the Pacific slopes of Ecuador and to G. mertensi Cochran and Goin from the eastern slopes of the Cordillera Central in southern Colombia. The latter species produces tadpoles (Cochran and Goin, 1970), whereas G. plumbea and the frogs on the Amazonian slopes have direct de- velopment of eggs into froglets. No other populations of Andean Gastrotheca are composed of uniform green individuals, although occasional green specimens of G. marsupiata (Dumeril and Bibron) and G. monticola Barbour and Noble lack dorsal markings (Duell- man and Fritts, 1972; Duellman, 1974). Thus, we undertook an analysis of the Andean populations of green Gastrotheca in an at- tempt to determine their systematic status. ANALYSES OF POPULATIONS For purposes of analysis the frogs were assigned to three geo- graphic populations—Amazonian, Pacific, and Colombian. Sixteen morphological measurements were obtained from each specimen. 1 Curator, Division of Herpetology, Museum of Natural History; Professor, Department of Systematics and Ecology, The University of Kansas, Lawrence, Kansas 66045. * Research Assistant, Division of Herpetology, Museum of Natural History, The University of Kansas, Lawrence, Kansas 66045. 2 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Eleven of these, described in Duellman (1970), are, as follows: snout-vent length (SVL), tibia length, foot length, head length, greatest head width, eye diameter, tympanum diameter, interorbital distance, internarial distance, eyelid width, and snout length. As- sessment of the lengths of the thumb and third finger follows Duell- man (1974). Three additional measurements are: 1) orbit-jaw— the horizontal distance between the ventral margin of the orbit and the margin of the upper lip; 2) naris-jaw—the horizontal distance between the ventral margin of the external naris and the upper lip; and 3) width of disc on the third finger. All measurements were obtained to the nearest 0.1 mm with needle-tipped dial calipers. A total of 25 external, descriptive characters also was assayed. Only ten of these are applicable to this analysis; the others are consistent among the three populations. Two descriptive charac- ters—iris color and tympanum color—are consistent within popula- tions. The remaining eight characters are variable; these are—labial stripe, flank color and pattern, posterior thigh pattern, anal region pattern, dorsal skin texture, and extent of webbing on fourth and fifth toes. Data for descriptive characters were recorded in a di- chotomous fashion, i.e., presence or absence of a character or char- acter state. Data obtained in this manner enables application of multivariate statistical analyses (Maxwell, 1961; Blackith and Rey- ment, 1971). All statistical analyses were accomplished through the use of Biomedical Computer Programs (Dixon, 1975) at The University of Kansas Computation Center. Univariate statistics (BMDP2D), pairwise t-tests (BMDP3D), and one-way analysis of variance (BMDPIV) were obtained on all morphometric data. A cluster analysis of cases (BMDP2M) was used to verify group membership and to check for outliers in the data set. A stepwise discriminant analysis (BMDP7M) was then utilized to determine group sepa- ration based on morphometric and descriptive data. In univariate analyses, only adults separated by sex were used; in multivariate analyses, sexes were combined and juveniles were included. Results of the ANOVA show that only eight morphometric characters (Table 1) have means that are significantly different (P < 0.05) among the three groups. In order to discern pairwise differences, two t-tests were generated. Because Fyyax tests indi- cated the presence of heteroscedasticity in some of the data, a (separate) statistic was used to check the t values obtained from a pooled variance estimate. These results show that the means of the eight morphometric characters are significantly different (P < 0.001) between the Colombian and Amazonian populations. In the comparison of the Pacific and Colombian populations, all characters were significantly different (P < 0.001), except the width of the disc (P = 0.36). However, only five characters have significantly A NEW MARSUPIAL FROG 3 different means (P < 0.05) between the Amazonian and Pacific populations; the three characters not different are head length (P = 0.07), eye diameter (P = 0.16), and tympanum diameter (P = 0.30). The stepwise discriminant analysis is based only on those char- acters or character states that exhibited variation among the three populations. These include the morphometric characters and 33 descriptive character states. However, only two morphometric characters (tibia length and eye diameter) and eight descriptive character states (those pertaining to labial stripe, flank color, flank pattern, posterior thigh pattern, anal region pattern, dorsal skin texture, and extent of webbing on the fourth and fifth toes) con- tributed to the model. The best model for classification (jackknife classification 98.3% correct) included only labial stripe (present in Pacific population) and pale flank color (present in Amazonian populations). In the discriminant analysis (see Fig. 1), canonical axis I dis- criminates between Pacific and Amazonian populations. Canonical coefficients vary from —7.97 for flank color to +4.73 for posterior thigh pattern; canonical correlation with axis I is 0.983. Canonical axis II contribution was found to be significant (eigenvalue = 16.01, X7113,1 — 165.76, P = 0.01); correlation with axis II is 0.97. Canon- ical coefficients for axis II vary from —4.59 (extent of webbing on fourth toe) and —4.36 (flank pattern) to +8.47 (labial stripe). Canonical axis II discriminates between Pacific and Colombian populations. The classification function of the discriminant analysis misiden- tified only three of 118 specimens (Fig. 1). Two specimens of the Pacific population were identified with the Colombian sample; one (KU 178528, a juvenile) lacks a labial stripe, and one (KU 164230) has a dark pattern on the flanks. One faded specimen (AMNH 17545) from the Pacific slopes appears to lack a labial stripe and dark flanks; therefore, it was classified with the Amazonian popu- lation. Nine additional specimens of the Colombian population were obtained subsequent to this analysis and form the basis for the description of coloration. DESCRIPTION OF NEW SPECIES The green Gastrotheca on the upper Amazonian slopes of the Cordillera Oriental in Ecuador obviously is distinct from other known species of the genus. No name is available for these frogs, for which we propose an epithet derived from the Greek oros mean- ing mountain and the Greek phylax meaning guard or watchman, used in the allusion that these frogs probably were watching in 1539 when a small band of Spanish Conquistadores led by Fran- OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 4 0£°0 LT0 OF'0 s0'0 L@'0 OT'0 TILT 69°0 €P'0 OL'O Se'T | eal OL'T 98°0 Lee 9T€ ds v'9-8'S 6ass GP-GE VECE ergre L’e-6S @ LG-9'SS SESH IG GPC-6 GG T'GS-F 06 GOC-E'ee eTe-1 81 OBE-LEE £6 0€ 9°CL-S°C9 L'S9-G'9S asury L0°9 GL'S LOE PEE oL6'E o0F'E LOSS 8O°CS LG'S OT'1G C6'PE 68°66 eg’ce Sole OF'89 66°09 x (889°? P9 =u) urIquiojoy ‘DIAYIOL SD) JO SUOHL[Ndod sayy Jo s}[Npe 10F SONSHRS QJeURATU— ] AIAV 09°0 0¢'0 rS'0 sc'0 8F'0 820 LOT tol 09°T OF'0 O'S 68'T GFT 66'T VOP LL ds ‘uonendod outoeg of vsoyy woz (CO'Q < q) WeIeyIp AfuRoyrUsIs you surayy , 69-SP O'S-T'E VPE'S PEGI OSTE 6e-S'S 0'9G-£°0G 6 1G-L'ST OSG-F'6I SIGT9T O'LE-T'8S 9'0E-0'TS Cre-o Oe €'CE-9'SS O'EL-S'8S O'I9-1'GF osuey ea vE'V I0'€ 69° 86E LYE TLCS €6'°61 LV'IG 08ST 9S TE T@L] Se"ce 66°86 9T'F9 g9°9S x (8 3ST ‘? $93 = 4) oyloeg 8F'0 £60 ePr'0 LEO 0s'0 (oa) 9G 6F'T S60 OL'0 Ore 69'°T Ore POT 6S Soe ds VOEP 9S-6E O'F-CS LYS VESTS VE-FS 8Sc-LLT LTG-8'ST SESS LT COCEST 6 9E-T FS 9°LG-0'SS 0°LE-0'SS S'66-L 1% O'FL-8°0S T6S°9°LP osury ef0°S oSo'P o8LG 209°S Tee FOG 99°0G €9°8T 09861 007 8ST CP'8s 8F'SS 08°6G 67°9G FE6S PLES x (8 39¢‘? 2'6=4) ueruozeuy O+ OF ©O OF €O OF £O OF ©O OF FO OF ©O OF *O OF O+ OF ©O OF €O OF €O OF FO OF FO OF ©O OF FO OF Xa - Josuy pig ysue] prey ~ yuaa-ynoug s1aqOBIeYD A NEW MARSUPIAL FROG 5 eee @d>>)))99>---- B---- CeCe ))y) Te oe al ee SUS ty Yael x -10.0 -5.0 0.0 5.0 10,0 Fic. 1.—Multivariate plot of result of discriminant analysis of three popu- lations of Gastrotheca. Circles = Amazonian population (n = 56); squares = Pacific population (n = 50); triangles = Colombian population (n = 12); open symbols = group means, cisco de Orellana descended the Papallacta Valley on the expedition that resulted in the discovery of the Amazon. Gastrotheca orophylax new species Holotype.—University of Kansas Museum of Natural History (KU) 164243, a brooding female, from 11 km (by road) east-south- east of Papallacta, 2660 m, Provincia Napo, Ecuador, obtained on 22 March 1975 by Linda Trueb. Paratypes.—All from upper Rio Papallacta Valley, Provincia Napo, Ecuador: KU 164244, 178568 from the type locality; KU 117981 from 3 km E Papallacta, 2900 m; KU 155469-70, 164242 from 12 km E Papallacta, 2630 m; USNM 211207 from 2 km E Chalpi, 2730 m. Additional specimens, from the upper Chingual Valley, are not designated as paratypes. Diagnosis—Gastrotheca orophylax is a moderately large Gas- trotheca ( 6 ¢ attaining snout-lengths of 59.1 mm and ¢ ¢ 74.0 mm) with an uniform green dorsum and having direct development 6 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY of eggs into froglets. It is further characterized by the presence of smooth or areolate skin on the dorsum, skin not co-ossified with dermal roofing bones, uniform pale green flanks, uniform dark pos- terior surfaces of thighs, a dark bronze iris reticulated with black, a bronze-colored tympanum, and the absence of a labial stripe (Fig. 2). Gastrotheca mertensi (Fig. 3) differs from G. orophylax by having more coarsely areolate or granular skin on the dorsum, dark flanks with or without pale spots, and a yellow tympanum. Gas- trotheca plumbea differs from G. orephylax by having a pale labial stripe, uniformly brown flanks, uniformly pale posterior surfaces of the thighs, an olive-green iris, and a green tympanum. Some indi- viduals of G. plumbea have dark flecks on the anterior and posterior surfaces of the thighs (Fig. 4), and others in life have a narrow, pale dorsolateral stripe (Fig. 5). Minor differences among at least some individuals of these species also exist in the extent of webbing on the fourth and fifth toes (least- extensive in G. mertensi), color pattern of anal region (dark patch and/or pale anal stripe in G. plumbea and no pattern in others), and ventral coloration (dark with pale spots in some G. mertensi and uniformly pale in all oth- ers). Gastrotheca mertensi produces tadpoles, and G. plumbea has direct development of eggs into froglets. Some individuals of two other Andean species of Gastrotheca are uniform green dorsally. Of these, G. monticola differs from G. orophylax by having a cream labial stripe, a bronze canthal stripe, dark spots on the flanks, pale dorsolateral stripe, and dark spots on the venter. Some G. marsupiata are uniform green dorsally but dif- fer from G. orophylax by being much smaller ( ¢ ¢ attain snout- vent lengths of 41.6 mm and @? ? 46.5 mm) and by having a pale labial stripe, dark canthal stripe, and granular skin on the dorsum. Both G. marsupiata and G. monticola produce tadpoles. Description of holotype.——Adult female with eggs in brood pouch; head wider than long; snout rounded in dorsal view and in profile; canthus angular; loreal region slightly concave; lips thin, rounded; nostrils nearly terminal on snout, slightly protuberant laterally; internarial region slightly depressed; interorbital area flat, much wider than eyelid; tympanic annulus indistinct; tympanum separated from eye by distance nearly equal to diameter of eye. Body robust; limbs moderately slender; ulnar tubercles absent; palmar tubercle diffuse; calcars absent; inner tarsal fold weak, present only distally; inner metatarsal tubercle elliptical, flat, visible from above; outer metatarsal tubercle absent; subarticular tubercles large, round; supernumerary tubercles large, flat, present only prox- imally; dises large, round; webbing absent between fingers; web- bing formula of foot I2—2*111.5—3I1I3—31V3—1.5V. Skin on dor- sum areolate; skin on belly and ventral surfaces of limbs coarsely granular; anal opening directed posteriorly at upper level of thighs, A NEW MARSUPIAL FROG if lacking folds and tubercles; opening of pouch small, /\-shaped. Tongue cordiform; choanae small, round; prevomerine teeth 6-7 on abutting transverse processes between choanae. Color (in preservative) of dorsum of head, body, and limbs, and anterior and posterior surfaces of thighs dull bluish gray; belly dull grayish cream; webbing dark gray. Measurements of holotype in mm.—Snout-vent length 70.1, tibia 36.0, foot 34.4, head width 26.3, head length 23.5, interorbital dis- tance 8.4, internarial distance 4.5, eyelid width 3.7, eye diameter 5.3, tympanum diameter 2.9. Coloration in life—The dorsal surfaces of the head, body, and limbs are bright emerald green. The flanks and margins of the upper lips are paler green. A bronze postorbital stripe is diffuse above the tympanum and disappears just posterior to the tympa- num. There is a bronze suffusion on the outer edges of the forearms and feet and on the dorsal surfaces of the toes. The axilla, groin, and distal parts of the posterior surfaces of the thighs have a blue suffusion. The palmar and plantar surfaces are dark gray, and the ventral surfaces of the thighs are grayish bronze; the other ventral surfaces are pale green. In calling males the throat is bluish gray. The iris is deep bronze with black reticulations. The tongue is cream, and the lining of the mouth is pale blue. Variation—The known specimens of Gastrotheca orophylax are remarkably uniform in structure and coloration. Females are slightly larger than males (Table 1) but do not differ in coloration. The amount and intensity of blue suffusion in the groin and on the posterior surfaces of the thighs is slightly variable individually, as is the distinctness of the bronze postorbital stripe. Recently hatched young lack the blue and bronze colors and have a pale gray venter. Distribution and ecology.—Gastrotheca orophylax is known from elevations of 2620-2910 m on the Amazonian slopes of the Cordillera Oriental in northern Ecuador and extreme southern Colombia (Fig. 6). Localities of occurrence are in the upper Rio Papallacta Valley (2630-2900 m) and the upper Rio Chingual Val- ley forming the border between Colombia and Ecuador (2620-2910 m). In both valleys, the frogs were found in disturbed upper mon- tane forest that had been partially cleared for pastures. Viney bam- boos (Chusquea) are dense in ravines, and large-leafed Gunnera (Haloragaceae) are abundant at edges of clearings. One individual was found beneath a stone by day; the others were collected at dusk and at night. Most were perched on leaves (1-2.5 m above the ground) of Guwnnera and other large-leafed herbs. Males have been observed calling in January, March, and July. The call is a moderately loud “bonk-bonk-bonk” repeated at intervals of 14-20 sec. Analysis of one recording (KU Tape 1224) shows that four calls have 3-6 (x = 4.5) notes about 0.005 sec in duration with 8 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Fic. 2.—Gastrotheca orophylax, KU 164243, holotype. 2, 70 mm snout- vent length. Fic. 3.—Gastrotheca mertensi, KU 181196, 2, 75 mm snout-vent length. A NEW MARSUPIAL FROG 9 Fic. 5.—Gastrotheca plumbea, KU 142614, 6, 54 mm snout-vent length. 10 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 600 M. CONTOUR) [] OVER 3000M. rw} [6] OVER 5000mM 40100 —— KILOMETERS Fic. 6.—Distribution of Gastrotheca orophylax (dots) and G. plumbea (circles ). intervals of about 0.075 sec between notes; the fundamental fre- quency at about 500 Hz is dominant. Brooding females have been found in March, May, June and July. Ten brooding females contained 12-32 (X = 21.4) eggs hav- ing mean diameters of 5.19-7.23 (x = 6.13) mm. Two brooding females collected on 22 June 1977 gave birth to froglets on 31 August and 25 September 1977. Ten newly born young have snout- vent lengths of 15.7-17.2 (x = 16.1) mm. DISCUSSION Gastrotheca orophylax is a member of the Gastrotheca plumbea group, as defined by Duellman (1974). Of the seven species now recognized in that group, five (G. cavia, lojana, monticola, psychro- phila, and riobambae) produce tadpoles that complete their de- A NEW MARSUPIAL FROG ER velopment in ponds, whereas G. orophylax and G. plumbea produce froglets. All of these species are essentially allopatric; G. orophylax and G. riobambae were both found at E] Carmelo, Ecuador. In addition to the same mode of reproduction and the similar- ities mentioned in the diagnosis, G. orophylax and G. plumbea are alike in having a bluish white cutaneous exudate, which is absent in other members of the group. Furthermore, the mating calls are similar and differ from the calls of the other species. Analysis of one’ recording of the call of G. plumbea (KU Tape 1036) reveals that the call consists of 3-6 (x = 4.3) notes and that the call repetition rate is about 25 sec. Notes have a duration of about 0.007 sec, and the interval between notes is about 0.042 sec; the fundamental fre- quency at about 650 Hz is dominant. Biochemical investigations of microcomplement fixation of al- bumins show that the immunological distance between G. plumbea and G. orophylax is 3 units and between G. mertensi and G. oro- phylax is 7 units (Linda Maxson, pers. comm.). According to Wil- son et al. (1977), 100 units of albumin distance equals 55 million years of separation. Thus, immunological distances indicate a sepa- ration of G. plumbea and G. orophylax for about 1.6 million years. This places the separation in the Pleistocene, a time of continuing uplift of, and extreme climatic fluctuation in, the Andes (Simpson, 1975; Vuilleumier, 1971). Presumably a G. orophylax-plumbea stock existed at moderate elevations in the Andes of northern Ecuador. Either because of greater uplift of the mountains or vertical cli- matic-vegetational shifts, populations became isolated on opposite sides of the Andes. Now G. orophylax lives at elevations of 2620- 2910 m on the eastern slopes and G. plumbea at similar elevations (2010-3085 m) on the western slopes (Fig. 6). Gastrotheca mertensi apparently is more distantly related to G. orophylax and G. plum- bea; immunological evidence indicates a separation of about 3.8 million years, or late Pliocene. | ACKNOWLEDGMENTS Investigations on marsupial frogs have been conducted under grants from the National Science Foundation (DEB 74-01998 and DEB 76-09986). Many of the frogs were collected by David C. Cannatella, John D. Lynch, John E. Simmons, and Linda Trueb; we are also grateful to Eugenia del Pino for logistic support in Ecuador, Linda Maxson for information on the results of biochem- ical studies, Juan M. Renjifo for translating the Spanish summary, Debra K. Bennett for drawing figure 1, Linda Trueb for critical re- view of the manuscript, and the following curators for the loan of specimens and/or provision of facilities at their institutions: Alice G. C. Grandison, British Museum (Natural History) (BMNH); W. Ronald Heyer, National Museum of Natural History (USNM); 12 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Arnold G. Kluge, University of Michigan Museum of Zoology (UMMZ); Clarence J. McCoy, Jr., Carnegie Museum (CM); Charles W. Myers, American Museum of Natural History (AMNH); Gunther Peters, Zoologisches Museum Berlin (ZMB), and Emest E. Wil- liams, Museum of Comparative Zoology, Harvard University (MCZ). RESUMEN Tres poblaciones de Gastrotheca que tienen una coloracién dor- sal verde deprovista de manchas ocurren en los Andes de Ecuador y del sur de Colombia. Estas poblaciones difieren entre si en carac- teres morfomeéetricos, detalles del patron de color e historia natural. Las ranas en el sur de Colombia son asignadas a G. mertensi, y aquellas en las laderas del Pacifico (2010-3085 m) en Ecuador a G. plumbea. Las poblaciones en las laderas Amazoénicas (2620-2910 m) de los Andes en Ecuador y el sur colombiano han sido llamadan G. orophylax en el presente trabajo. Esta especie se diferencia de las otras por caracteres morfométricos, color de iris y timpano, ausencia de rayas labiales y dorsolaterales, presencia de piel areo- latada, flancos de color verde palida, y desarrollo directo de los huevos a ranas pequenas. Gastrotheca orophylax esta mas estrecha- mente emparentada a G. plumbea; resultados de estudios in- munolégicos muestran que las dos especies han estado separadas por aproximadamente 1.6 millones de afios. Aparentemente fueron aisladas por cambios altitudinales y climatico-vegetacionales asi como por el eontinuo Jevantamente de los Andes durante el Pleistoceno. SPECIMENS EXAMINED Gastrotheca mertensi—COLOMBIA: Cauca: Moscapin, CM 50395-96, KU 181194-202, UMMZ 121024 (holotype), 121025-28, USNM 148567-68, 159896-98. , Gastrotheca orophylax—COLOMBIA: Narifo:. La Victoria, KU 140386. ECUADOR: Carchi: El] Carmelo, KU 178569, USNM 211177-203; 5.7. km NW El Carmelo, KU 178570-84, 178585-86 (skeletons), 180347 (young). Napo: 2 km E Chapli, USNM 211207; 3 km E Papallacta, KU 117981; 11 km ESE Papallacta, KU 164243-44, 178568; 12 km ESE Papallacta, KU 155469- 70, 164242; Santa Barbara, USNM 211204-05; 1 km SW Santa Barbara, USNM 211206. Gastrotheca plumbea—ECUADOR: Azuay: Llapin, AMNH 17545; Mol- leturo, ZMB 30057. Carchi: Atal, near San Gabriel, UMMZ 83655 (not plotted on map), Cotopaxi: Pilal6, KU 132413-22, 132423 (skeleton), 142614, 178499- 528. Imbabura: 17.4 km E Apuela, MCZ 95435; Intag, BMNH 1947.2.31.19 (holotype). Pichincha: 5 km ESE Chiriboga, KU 164232; 9.5 km NW Nono, KU 164229-31; old Santo Domingo road, 1700 m, UMMZ 154406. LITERATURE CITED BuackitH, R. E., ReymMent, R. A. 1971. Multivariate morphometrics. Aca- demic Press, London, 412 p. A NEW MARSUPIAL FROG 13 CocuraNn, D. M., Gorn, C. J. 1970. Frogs of Colombia. U.S. Natl. Mus. Bull. (288): 1-655. Dixon, W. J. (ed.). 1975. Biomedical computer programs. Univ. California Press, Berkeley, 792 p. DuELLMAN, W. E. 1970. The hylid frogs of Middle America. Univ. Kansas Mus. Nat. Hist. Monogr. (1):1-753. DuELLMAN, W. E. 1974. A systematic review of the marsupial frogs (Hylidae: Gastrotheca) of the Andes of Ecuador. Univ. Kansas Mus. Nat. Hist. Occas. Pap, (22):1-2,7. DvuELLMAN, W. E., Fritts, T. TH. 1972. A taxonomic review of the southern Andean marsupial frogs (Hylidae: Gastrotheca). Ibid. (9):1-37. MaxwELL, A. E. 1961. Canonical variate analysis when the variables are di- chotomous. Educ. Psychol. Measur. 21:259-271. Simpson, B. B. 1975. Pleistocene changes in the flora of the high tropical Andes. Paleobiology 1:273-294. VuILLEUMIER, B. S. 1971. Pleistocene changes in the fauna and flora of South America. Science 173:771-780. Witson, A. S., Carutson, S. S., Wurire, T. J. 1977. Biochemical evolution. Ann. Rev. Biochem. 46:573-639. 7 ¥ ' . r. hed 7 Bh nie my i aN) 5 : i j ‘ee ; - : A 7 baAS ae eo : ‘ aie er ts air 4 r | e : / : > : ot 7 ts NAL. PAPERS | eae | cman Ph fpretail*. “dit Frere) PSxyaq f ah ees. VPM | A 8 OS sabe: dpe mas MEW rheorbaae Ele) Se: VTE AST 2 (Oe Baperei a: ad i, Aeti iY) ii iex ) i ene D Pith. Mak. At > roped | ra att Smurten = Atie* diet nies . a i Von a Micete yout Tete Bi Pvde ng yy Ani lost} iz A 20 i, TihMpe fx i eat a Leaat leis J) overran. weld te ) 3 eth bin “A Cid nh r Piel bated hon ; a TO enna == $e at ¥ i itctte ae wat) hi ra aus » jane i 7 eye leiang | — ar ine Li Denliday 7 een baht y= Fe tempor yerilsa: Low tls dereinatak ear ae \ uteafied st capcom Itooaph hd Pe aby h heat A? ey mC) Tl aay a oa? Mi Neila fey iste! ea WAP yrivurdibe 5 J —" ee “thal iste wheug 2 SH UIT LA Jem th.! pan en f emi es ab heey Cet eee | Wentinmait weet yaa! @emksthy eaahirobee ine ngnalcy/? 10 Lliatpe ils tides ites ati ARs iesreryctasoe See, PPTs ty Woe ne en toy ia @n oar iY RAERE Le hin ccninaiy uf lug | Caaivaeer biliate Ar tars li iu es? ' q c et Sere [ett ; ° néth VLA . ; : 4 yh af L cae iin lear ‘ . PM Bic ann 2 Ge apis, Joes tov: é | Paik hy? Ci Py VEL. aie A Tie " gq m AG BK ae) Te Uris ah, « et s® trl pijaae: erie, ‘) hy uy var > 7 *Se5 1 7 af Al *ae4 i cul ‘atry . PRICY st) Ve © . Cry * Dr Peres: | o] yt a oe? & 7 : : " ' 7 2 ; , 7 _ im Ds Vie Cal ufast "9 ore Gro ee ues A Ls ‘ | | .' wialy ; sii Le yl @nt (2 jile & ia a Dita 8 a Ak nt UNIVERSITY OF KANSAS PUBLICATIONS MUSEUM OF NATURAL HISTORY The University of Kansas Publications, Museum of Natural History, beginning with volume 1 in 1946, was discontinued with volume 20 in 1971. Shorter research papers formerly pub- lished in the above series are now published as Occasional Papers, Museum of Natural History. The Miscellaneous Publica- tions, Museum of Natural History, began with number 1 in 1946. Longer research papers are published in that series. Monographs of the Museum of Natural History were initiated in 1970. All manuscripts are subject to critical review by intra- and extra- mural specialists; final acceptance is at the discretion of the publications committee. Institutional libraries interested in exchanging publications may obtain the Occasional Papers and Miscellaneous Publica- tions by addressing the Exchange Librarian, University of Kan- sas Library, Lawrence, Kansas, 66045. Individuals may pur- chase separate numbers of all series. Prices may be obtained upon request addressed to Publications Secretary, Museum of Natural History, University of Kansas, Lawrence, Kansas 66045. Editor: E. O. WiLey PRINTED BY UNIVERSITY OF KANSAS PRINTING SERVICE LAWRENCE, KANSAS Se. COMe ZOO '- ON IBF 2 | au WAY | 5 4 | OCCASIONAL PAPERS of the MUSEUM OF NATURAL HISTORY The University of Kansas Lawrence, Kansas NUMBER 85, PAGES 1-14 APRIL 30, 1980 ADDITIONS TO THE KNOWLEDGE OF HIPPOCAMELUS, CTENOMYS AND MYOCASTOR FROM THE MIDDLE PLEISTOCENE OF THE TARIJA BASIN, BOLIVIA by Davip FrRAILEY', KENNETH E. CAMPBELL”, AND RonaLp G. WOLFF" The purpose of this paper is to describe and illustrate additional specimens of three species of mammals from the middle Pleistocene Tarija fauna (age follows Webb, 1974) of southern Bolivia. These specimens were collected by us in December, 1976, while visiting this classic locality in the company of Ing. Bethuel Arozqueta of the University Museum of Tarija. These specimens are more com- plete and provide better systematic comparisons than those previ- ously illustrated in monographs on the Tarija fauna (i.e., Ameghino, 1902; Boule and Thevenin, 1920). All measurements are in millimeters (mm); parentheses indicate an approximate measurement. The following acronyms are used: FMNH, Field Museum of Natural History; KUM, University of Kansas, Department of Mammalogy; KUVP, University of Kansas, Department of Vertebrate Paleontology. ACKNOWLEDGMENTS We wish to thank Dr. Larry D. Martin and Dr. Robert S. Hoff- mann of the University of Kansas and two anonymous reviewers for critical comments on this paper. A special note of appreciation is extended to Ing. Bethuel Arozqueta, Director of the University 1 Department of Systematics and Ecology and Museum of Natural History, University of Kansas, Lawrence, KS 66045 2 George C. Page Museum, 8501 Wilshire Blvd., Los Angeles, CA 90036 3 Department of Zoology, University of Florida, Gainesville, FL 32611 2 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Museum of Tarija, for courtesies and assistance extended during this visit. Our research was supported by the Saul Fund and the Claude Hibbard Memorial Fund of the University of Kansas Museum of Natural History and by NSF Grant DEB-78-03122. SYSTEMATIC DESCRIPTIONS Order Artiodactyla Owen, 1884 Family Cervidae Gray, 1821 Genus Hippocamelus Leuckart, 1816 Hippocamelus sp. Fig. 1A Material—KUVP 43063, a nearly complete antler, lacking only distal ends of tines. Locality——Pueblo Viejo area, near Tarija, Departamento de Tarija, Bolivia. Specimen was discovered near top of exposed sec- tion approximately five meters beneath the two ferruginous layers found in upper part of Horizon B of Oppenheim (1943). Description—When complete, the anterior tine was slightly over half the height of the posterior tine. The two tines become vertical in approximately the first one-third of their lengths. The posterior tine also curves posteriad for the distal third of its length. The antler base and the proximal third of each tine is strongly fluted. These flutings gradually fade and are absent on the distal third of each tine. The antler base is slightly oval in cross-section, whereas each tine is round. The antler was shed naturally. Meas- urements of the base are: length, 40; width, 33; height of saddle from burr, 42. Discussion —Ameghino (1902) first noted the presence of cer- vids in the Tarija fauna and identified three species on the basis of isolated teeth. One of these was placed in the genus Hippocamelus as H. incognitus n. sp. Two were referred to Cervus as C. tuber- culatus Gervais and Ameghino (1880) and C. percultus n. sp. Boule and Thevenin (1920) declared all three species identifications. in- determinate as they felt that only a familial designation was reason- able on the basis of isolated teeth. They left the specific identi- fication of Tarija cervids to other workers and the recovery of more diagnostic material. It is apparent from the discussion of Hoffstetter (1963) that two types of deer antlers representing two genera are known from Tarija but he did not illustrate either. One antler type is described as a robust form of Hippocamelus compatible in size with the teeth which Ameghino (1902) referred to Cervus percultus and C. tuber- culatus. The second antler type is a simple, unbranching spike of a small deer for which Hoffstetter (1963) erected the new genus Charitoceros. Hoffstetter stated that the small size of this antler is more in keeping with the size of the teeth on which Ameghino ADDITIONS—HIPPOCAMELUS, CTENOMYS AND MYOCASTOR- 3 20mm Fic. 1—A, Hippocamelus sp., KUVP 43063, left antler. B-C, Ctenomys subassentiens, KUVP 43051, skull; (B) lateral view; (C) anterior view. (1902) based Hippocamelus incognitus. It appears then that at least two deer are present in the Tarija fauna and that one is Hippo- camelus. The excellently preserved antler of Hippocamelus (KUVP 43063) substantiates the presence of this genus in this fauna. As the 4 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY antlers of Hippocamelus are known to be variable among individ- uals (Pascual, et al., 1966), and as readily identifiable specimens of Hippocamelus from Tarija are rare, it seems appropriate at this time to follow Hoffstetter (1963) in leaving the species identification indeterminate. Order Rodentia Bowdich, 1821 Family Octodontidae Waterhouse, 1839 Genus Ctenomys Blainville, 1826 Ctenomys subassentiens Ameghino, 1902 . Figs. 1B, C; 2-5 Material—KUVP 43051, skull missing the occipital region; man- dible missing the right Ms; and the angle of the right ramus; hu- merus, ulna, radius, two partial scapulae, and several ribs and vertebrae. Locality—Pueblo Viejo area, near Tarija, Departamento de Tarija, Bolivia. The specimen was found near the top of the geo- logic section in a light brown, well-sorted sand stratum (thickness 1 m) approximately one meter beneath the ferruginous beds of Horizon B of Oppenheim (1943). Comparative Recent material—Ctenomys lewisi, KUM 79399, FMNH 29056, two skulls and mandibles; C. sylvuanus, KUM 79402, 79403, two skulls and mandibles; C. mendocina, KUM 79400, 79401, two skulls and mandibles; C. torquatus, KUM 79404, 79406, two skulls and mandibles, FMNH 44813, skull, mandible, forelimb elements. Description—The skull is characterized by having (1) sturdy construction; (2) cranium with dorsal part essentially flat with slight depressions along the interparietal suture, anterior part of the inter- nasal suture, and transversely across frontals; (3) nasal-frontal, maxillary-frontal, and interfrontal sutures deeply interfingered; (4) breadth of premaxillaries as great as braincase: (5) premaxillae bowed laterally and heavily rugose at exit of incisors; (6) external nares forming a triangular opening; (7) anterior margin of the origin of M. masseter medialis, pars anterior (terminology of Woods, 1972) strongly defined as the skull narrows sharply at this point; (8) post-orbital processes large, rugose, and each extending forward to unite with the infraorbital foramen bridge; (9) sagittal crests clearly evident and lyre-shaped; (10) lambdoidal crest high, as wide as inner margins of the zygomatic arch, and at a right angle to the long axis of the skull; (11) occiput missing on the fossil, but was apparently vertical; (12) large paroccipital process; (13) zygomatic arches heavy and flaring widely with the postorbital process large and strongly triangular thus giving the base of the orbit an angular, rather than round, ventral margin; (14) infraorbital bridge of the maxillary constricted by a groove at the base of the lacrimal, and ADDITIONS—HIPPOCAMELUS, CTENOMYS AND MYOCASTOR_ 5 20mm Fic. 2.—Ctenomys subassentiens, KUVP 43051, skull: (A) dorsal view; (B) palatal view. passing as a thin extension anterior to the lacrimal until it meets the premaxillary and frontal above the lacrimal; (15) jugal deep dorso- ventrally, with deeply sculpted fossae. The upper incisors are proodont, flat anteriorly, rounded pos- teriorly, and wide. They are not in their normal position in that they are not closely appressed and meet only at the tips, with the left incisor pressed to the right distally and the right incisor vertical. The P!-M? are characteristically kidney-shaped, with P* the largest tooth and M! and M? progressively smaller. The M?® is cylindrical, with a faint indication of the kidney-shaped cross-section of the other upper teeth. The right M* was displaced in life, possibly causing the anomalous positioning of the incisors and greater wear on the left tooth rows. The mandible is characterized by having (1) very sturdy con- 6 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Fic. 3.—Ctenomys subassentiens, KUVP 43051, mandible: (A) occlusal view; (B) ventral view. struction; (2) bone around exit of incisors thick and rugose, espe- cially on dorsal surface; (3) incisors projecting approximately 14 mm, equal to upper incisors; (4) mandibular foramen positioned high on the ascending ramus near origination of incisors; (5) masseteric ridge extending anteriad to a point below anterior edge of M2. The scapula is characterized by being narrow and having (1) blade spreading gradually from the glenoid; (2) coracoid process long, flat, and parallel to long axis of the scapula; (3) spine arising near anterior border; (4) glenoid fossa egg-shaped. The humerus is characterized by having (1) greater tubercle with height less than that of head of humerus and continuous with deltoid crest; (2) deltoid crest extending distal to midpoint of shaft where it terminates in a prominent node. The distal end is damaged on specimen KUVP 43051, but enough remains to indicate that it was wide and flat. The ulna and radius are both characterized by being (1) gently ADDITIONS—HIPPOCAMELUS, CTENOMYS AND MYOCASTOR’- 7 Fic. 4.—Ctenomys subassentiens, KUVP 43051: (A-C), left scapula; (A), dorsal view; (B), glenoid fossa, anterior aspect of scapula faces up; (C), ventral view. D-H: left humerus; (D), proximal end, anterior faces down; (E), anterior view; (F), medial view; (G), posterior view; (H), lateral view. sinuous in both anterior and lateral views; and (2) each having prominent oval interosseous crests. They are appressed distally for approximately one-third the length of the ulna. Radius is grooved and ulna has a corresponding ridge to facilitate close contact in the appressed area. Proximal half of shaft of radius is triangular in cross-section, but distal half is more rectangular as a result of the aforementioned groove. Lateral styloid process and distal end of ulna are missing, and distal epiphysis of radius was unfused at death and is now lost. These postcranial elements are similar in many respects to those of Recent species of Ctenomys although they, like the skull and mandible, exhibit a robustness and large size unlike any living species. Measurements are as follows. All cranial measurements in this paper follow the guidelines of DeBlase and Martin (1974). Skull: basal length, (50); basilar length, (44); condylobasal length, (52); occipitonasal length, 46.6; greatest length of skull, 50.9; breadth of braincase, 18.6; least interorbital breadth, 14.9; mastoid breadth, 29.8; postorbital constriction, 12.8; widest rostral breadth, 17.4; zy- gomatic width, 35.0; nasal length, 17.2; nasal width, 9.9; tympanic bullae length, 15.8; tympanic bullae width, at stylomastoid process. 6.7; diastema length, 14.5; maxillary tooth row length, 12.6; palatal, 8 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY ® 8 2 0 10mm a | D A\E F Fic. 5.—Ctenomys subassentiens, KUVP 43051: (A-C), left radius and ulna: (A), medial view; (B), posterior view; (C), lateral view. (D), ulna, anterior view. E-F. radius: (E), anterior view; (F), posterior view; (G), proximal end, anterior faces up. length, 26.1; palatal width between alveoli of P*’s, 1.7; palatal width between alveoli of M®*’s, 5.2; I', width, 4.1; depth, 3.8; P‘, greatest length, 4.5; M!, greatest length, 4.0; M?, greatest length, 3.4; M?, greatest length, 1.6. Mandible: mandibular diastema, 9.6; mandible length, 38.3; P,-M, length, 12.2; I,, greatest width, 4.1; I, greatest depth, 3.6; P,, greatest length 4.5; My, greatest length, 4.2; Moe, greatest length, 3.7; Mg, greatest length, 1.2. Scapula: width (antero-posterior) of neck, 4.3; glenoid length, 6.1; glenoid width, 4.3. Humerus: length, 31.0; width of head across trochanters, 8.6; depth (antero-posterior) of head, 6.5; width of - ADDITIONS—HIPPOCAMELUS, CTENOMYS AND MYOCASTOR' 9 shaft at deltoid crest, 3.9; depth of shaft (including deltoid crest), 6.7. Ulna: length, (37); depth (antero-posterior) at interosseous tubercle, 4.3; depth (antero-posterior) at semilunar notch,'3.9. Ra- dius: length (minus distal epiphysis), 25.4. Discussion—Ameghino (1902) described three new species of Ctenomys from the Tarija Basin. Later, Boule and Thevenin (1920) synonymized these under the name C. subassentiens, commenting that a detailed comparison with living species of Ctenomys was needed in order to justify the validity of this fossil species. How- ever, productive osteological comparisons with Recent species of Ctenomys require stability in the classification of the Recent species. At present the taxonomy of Ctenomys is in a state of flux, with approximately 65 described species, many of which are thought to be local variants ( Reig, et al., 1965). As far as we know, a revision of the Ctenomyinae is not forthcoming and it appears best to retain provisionally the name C. subassentiens for the fossils at Tarija. The skull of Ctenomys subassentiens typifies the features of a tooth-burrowing rodent (as recognized by Agrawal, 1967) to a greater extent than any of the living species examined. These tooth- burrowing adaptations include proodont incisors, protruding pre- maxillae, a dorsally flattened skull with a prominent lambdoidal crest, and stout, wide zygomatic arches. In C. lewisi these features are less marked but yet clearly evident and are found in conjunction with heavy claws on the forefeet, indicating that C. lewisi is readily capable of both tooth-burrowing and forelimb-burrowing. The sturdy development of the forelimbs of C. subassentiens suggests that it was also capable of forelimb burrowing. Those species which do not possess the striking skull modifications for tooth- burrowing, such as C. torquatus, C. sylvuanus, and C. mendocino, also have longer, more slender claws than does C. lewisi. It may be that the compaction of the soil where each species lives determines to a large extent whether the species will burrow with relatively light claws or require strong claws and the use,of its incisors. Weir (1974) states that Ctenomys burrows exclusively with the feet and that the incisors are rarely used, and then only for clearing away roots which are protruding into the burrow. An opposing opinion is expressed by Walker, et al. (1975) who state that the incisors, and not the forefeet, are used to loosen the earth in burrowing. These contradictory observations are possibly based on different species, as both burrowing types are present in the genus although differing in expression among the numerous species. Family Myocastoridae Miller and Gidley, 1918 Genus Myocastor Kerr, 1792 Myocastor perditus (Ameghino, 1902) Figs. 6, 7 10 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Material—KUVP 43052, partial skull with upper incisors and right and left M*’s, partial mandible with complete lower dentition; KUVP 43053, partial skull with complete upper dentition, left M? displaced; KUVP 43054, isolated left P* and right M?. Localities —KUVP 43052: San Blas area near Tarija, Departa- mento de Tarija, Bolivia. Specimen was discovered near the top of the exposed section in a fine sandy clay stratum just below a thin (less than 4 mm) layer of hardpan and approximately 20 m above a thick (approximately 1 m) ash layer; KUVP 43053 and 43054 from approximately 5 km south of Tarija, name of the area unknown. Comparative Recent material—Myocastor coypus, KUM 59551, partial skull and mandible; KU Zool unnumbered, skull; FMNH 25257-8, skulls and mandibles; FMNH 27659, skull and mandible, juvenile. Description—The cranial specimens of Myocastor perditus from Tarija are fragmentary, but in at least one specimen, KUVP 43053, the incisors and tooth rows are in their natural positions. On this specimen a small portion of the zygomatic arch is preserved near the right P*; it is relatively large for the size of the teeth as com- pared with the living species, M. coypus. The width of the arch (measured at a distance 5 mm lateral to labial edge of M') exceeds the length of M*, whereas in M. coypus the reverse is true. The upper incisors of M. perditus originate above the P#, and their cur- vature has a greater radius than in M. coypus. The upper tooth Fic. 6.—Myocastor perditus, KUVP 43053, partial skull: (A), palatal view; (B), lateral view. ADDITIONS—HIPPOCAMELUS, CTENOMYS AND MYOCASTOR 11 rows diverge posteriorly at a greater angle than do those of M. coypus. The enamel patterns of the upper teeth are basically similar between the Recent and fossil forms, with minor variations present on KUVP 43053 but not on KUVP 43054. On KUVP 43053 the paraflexus becomes separated as a fossette first and is followed by the metaflexus. In the modern specimens and in KUVP 43054 the opposite sequence occurs. The metafossette of KUVP 43053 is strongly bent whereas that of M. coypus and KUVP 43054 have a lesser twist. — The mandible of the Tarija Myocastor (KUVP 43052) has a much heavier masseteric ridge than that of the living species; the symphyseal region parallels the incisors (there is a postero-dorsad extention of the symphysis in M. coypus); and the lower incisors, like the uppers, are less curved. The lower teeth appear to be more elongate than those of M. coypus. On the lower teeth of the Tarija specimens, five fossettids are formed on Py, and four each on M;_3. The external flexid is the last to form a fossettid. Measurements are as follows (KUVP numbers in parentheses) : T'-M? (43053), 68.0; width of I' (43053), 6.2; (43052), 5.6; depth of I! (43053), 6.5; (43052), 6.5. Length and width of: P* (43053), 5.5, 5.2; M! (43053), 6.7, 6.6; M? (43053), 7.9, 7.0; M® (43053), 8.5, 7.1; (43052), 8.2, 7.3. Palatal width at P*s (43053), 1.0; palatal width at M*’s (43053), 11.9; maxillary tooth row length (43053), 28.2. Mandibular diastema (43052), (12); Py-M, length (43052), (32.5); I, width (43052), 6.4; I, depth (43052), 5.3. Length and width of: P, (43052), 7.0, 4.2; M, (43052), 6.9, 4.8; Ms (43052), 8.0, 5.4; Ms (438052), 9.8, 5.9. Discussion.—Three isolated teeth of a rodent allied with Myo- castor were reported from the Tarija fauna by Ameghino (1902), for which he erected the name Matyoscor perditus. With a com- plete lower dentition at their disposal, Boule and Thevenin (1920) synonymized Matyoscor perditus with the living species, Myocastor coypus. Lacking additional specimens, Hoffstetter (1963) followed Boule and Thevenin. The Tarija species differs from Myocastor coypus, however, by having (1) the incisors less curved; (2) the zygomatic arch heavier; (3) the cheek teeth diverging posteriad at a greater angle; and (4) the mandible with a heavier masseteric ridge. It seems probable that the Myocastor at Tarija is in fact a distinct species, for which the species name perditus (Ameghino), should be resurrected. SUMMARY New specimens of Hippocamelus sp., Ctenomys subassentiens, and Myocastor perditus from the middle Pleistocene deposits of Tarija, Bolivia, are described. An excellently preserved antler of Hippocamelus substantiates the recognition of this genus in the 12 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY oO Fic. 7.—Myocastor perditus, KUVP 43052. (A), occlusal view of mandible; (C), lateral view of mandible; (B), partial skull with M3’s. ADDITIONS—HIPPOCAMELUS, CTENOMYS AND MYOCASTOR 13 fauna by Hoffstetter (1963), although the species is evidently not that originally described as Hippocamelus incognitus by Ameghino (1902). A complete skull and mandible associated with a partial skeleton of Ctenomys subassentiens permits a more thorough discussion of this species than was previously possible. This species has a number of characters which are associated with tooth-burrowing in rodents. Some, but apparently not all, living species of Ctenomys also have skull modifications for tooth-burrowing, but none exhibit the same degree of development of these characters as does C. subassentiens from Tarija. Two partial skulls and a mandible of Myocastor perditus indi- cate that this species is in fact distinct from the living coypu, Myocastor coypus, in contradiction to the conclusions of Boule and Thevenin (1920) and Hoffstetter (1963). The specimens referred to M. perditus have more robust cranial features, with heavier zygo- matic arches and masseteric ridges, and tooth rows which diverge at a greater angle, than those of the modern species. LITERATURE CITED AcrAwAL, V. C. 1967. Skull adaptations in fossorial rodents. Mammalia, Vol. 31(2):300-312. - AMEGHINO, F. 1902. Notas sobre algunos mamiferos fdsiles nuevos 0 poco- conocidas del Valle del Tarija. Anales del Museo Nacional de Buenos Aires, Vol. 8:225-261, 3 fig., 7 pl. Bouts, M., and A. THEvENIN. 1920. Mammiféres fossiles de Tarija. Mission Scientifique, Crequi-Montfort et E. Senechal de la Grange, Paris (Soudier). VII + 256 pp., 65 fig., 25 pl. DeBuaseg, A. F., and R. E. Martin. 1974. A Manual of Mammalogy with Keys to Families of the World. Wm. C. Brown Company, Publ., Dubuque, Iowa. 329 pp. Gervais, H., and F. AMEGHINO. 1880. Les Mammiféres fossiles de ! Amérique du Sud (y versién espafiola). In-8°, Paris et Buenos Aires: 225 pp. HorFFstTetTer, R. 1963. La faune Pléistocéne de Tarija (Bolivie), note pré- liminaire. Bulletin du Museum National d’Histoire Naturelle, Paris, Ser. 2, Vol. 35, no. 2:194-203. OprENHEIM, V. 1943. The fossiliferous basin of Tarija, Bolivia. Jour. Geol., Vol. 51:548-555. PascuaL, R., N. V. Carrot, J. C. Francis, O. Gonpar, E. OrTEGA Hinojosa, E. Tonnt, J. A. Pisano, A. B. pE RINGUELET, and J. ZeT11. 1966. Fasciculo IV, Vertebrata. In: Angel V. Borello, ed., Paleontografia Bonaerense. Provincia de Buenos Aires, Comision de Investigation Cientifica, La Plata. 202 pp. Reic, O. A., J. R. Contreras, and M. J. Pranranimwa. 1965. Contribucién a la elucidacién de la sistematica de las entidades del género Ctenomys (Rodentia, Octodontidae). I. Relaciones de parentesco entre neustras de ocho poblaciones de tuco-tuco inferidas del estudio estadistico de variables del fenotipo y su correlacién con las caracteristicas del cariotipo. Univ. Buenos Aires, Facultad de Ciencias Exactas y Naturales, Contribuciones Cientificas, Serie Zoologia, Vol. 2(6):301-352. 14 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Wa ker, E. P., F. Warnick, S. E. HAMLET, K. I. Lance, M. A. Davis, H. E. Usste, P. F. Wricut, and J. Ls Parapiso. 1975. Mammals of the World. Third Ed.. Johns Hopkins Press, 2 Vols.: 1500 pp. Wess, S. D. 1974. Pleistocene llamas of Florida, with a brief review of the Lamini. In: Pleistocene Mammals of Florida, S. D. Webb (ed.), Univ. Florida Press: 270 pp. Wer, B. J., 1974. The tuco-tuco and plains viscacha. In: I. W and B. J. Weir (eds.), The Biology of Hystricomorph Rodents of the Zoological Society of London, No. 34:113-130. Woops, C> A. 1972. Comparative myology of jaw, hyoid and pectoral appen- dicuiles regions of New and Old World hystricomorph spslantts Amer. Mus. Nat. Hist., Bull. 147:117-198. . Rowlands . Symposium Peel: f . -" curr ys + (epi iat : é iw es ali > ii Oresfric emrtsnal Veer ecu a wi AL) f J mich 74 pi AS a ARERR Un ina @ hibh oe ea | bone he ms ' ee CEA est 1 ae 7 7 2 i de lies Si RAEN Yulin. sf gOQhbared _ “at i ' Lean i matty: al 4) weiiplen wire ar a Sod fol bi dey a te UNIVERSITY OF KANSAS PUBLICATIONS MUSEUM OF NATURAL HISTORY The University of Kansas Publications, Museum of Natural History, beginning with volume 1 in 1946, was discontinued with volume 20 in 1971. Shorter research papers formerly pub- lished in the above series are now published as Occasional Papers, Museum of Natural History. The Miscellaneous Publica- tions, Museum of Natural History, began with number 1 in 1946. Longer research papers are published in that series. Monographs of the Museum of Natural History were initiated in 1970. All manuscripts are subject to critical review by intra- and extra- mural specialists; final abeeptanice is at the discretion of the publications committee. Institutional libraries interested in exchanging publications may obtain the Occasional Papers and Miscellaneous Publica- tions by addressing the Exchange Librarian, University of Kan- sas Library, Lawrence, Kansas, 66045. Individuals may pur- chase separate numbers of all series. Prices may be obtained upon request addressed to Publications Secretary, Museum of Natural History, University of Kansas, Lawrence, Kansas 66045. Editor: E. O. WiLEy PRINTED BY UNIVERSITY OF KANSAS PRINTING SERVICE LAWRENCE, KANSAS U N | : MP. ZOOL: 2/2 LIBRARY OcT 9 1990 of the UNIV ERe@IT Y MUSEUM OF NATURAL HISTORY The University of Kansas Lawrence, Kansas OCCASIONAL PAPERS NUMBER 86, PAGES 1-53 SEPTEMBER 23, 1980 REPRODUCTIVE BIOLOGY OF LIZARDS ON THE AMERICAN SAMOAN ISLANDS By TERRY D. SCHWANER? The reproductive biology of most lizards is poorly known. This is particularly true for tropical species, and most especially for island populations in the Pacific region (Brown, 1956; Inger and Green- berg, 1966; Fitch, 1970; Duellman, 1978). This study concerns certain aspects of the reproductive biology of 12 species of lizards on the tropical Pacific islands of American Samoa. There are no published studies describing lizard reproduction on these islands. The species accounts presented herein summarize basic data on male and female sizes (snout-vent lengths) at reproductive matu- rity, clutch sizes, descriptions of eggs, incubation periods, and hatchling sizes. Data on some species are sufficient to suggest re- lationships between climatic variables and monthly frequencies of fecund and ovigerous females and to indicate patterns in annual reproductive activity. Reproductive modes for these and other spe- cies are summarized from the basic observations and discussed in terms of reproductive strategies (Tinkle, et al., 1970; Duellman, 1978). Limitations of data include: (1) size at maturity for males was estimated from measurements of testes lengths and the presence of enlarged convoluted tubules in freshly preserved specimens, not from sections or smears of testicular tissue from which the presence of viable sperm can be detected; (2) small samples for a few species do not always reflect the full range of snout-vent lengths from hatchlings to adults; and, (3) samples were not taken during all months of the year, and sample sizes for any given month repre- sent the total collections from several habitats and islands. *Museum of Natural History, University of Kansas, Lawrence, KS 66045. Present address: The South Australian Museum, Adelaide 5000, South Australia. 2 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Previous studies have shown that monthly samples are essential for detecting annual reproductive cycles in temperate areas with marked fluctuations in mean monthly temperatures and rainfall (Fitch, 1954, 1956; Fitch and Greene, 1965; Mount, 1963; Tanner, 1957; Tinkle, 1961, 1967; and others). Similarly, monthly samples are necessary in equatorial tropical areas where temperatures re- main relatively constant, but where fluctuations in precipitation apparently dictate corresponding seasonality in the reproductive conditions of lizards (Sexton, et al., 1971; Fitch, 1973; Ruibal, et al., 1972; Brown and Sexton, 1978; Andrews and Rand, 1974; Gorman and Licht, 1974; Crump, 1974; and, Duellman, 1978). Alternatively, reproductive patterns in lizards are relatively aseasonal in areas where both temperatures and rainfall are constant year round (Inger and Greenberg, 1966). I will demonstrate the marked con- stancy of annual temperature and rainfall in American Samoa and present correlations among these variables and other climatic pa- rameters with the monthly frequencies of reproductive activity among certain lizard species. Following the evidence of Inger and Greenberg (1966), if annual climatic patterns are relatively con- stant, there should be little seasonality in the reproductive activity of lizards in American Samoa. In discussing the reproductive strategies of lizards in American Samoa, I recognize Wilbur, Tinkle and Collins’s (1974) list of components to a life history study. These are clearly summarized in Duellman (1978) as follows: (1) mortality schedules of juveniles and adults, (2) age at first reproduction, (3) reproductive life-span, (4) fecundity (including number and size of eggs, frequency of deposition, and proportion of females breeding), (5) the fecundity- age regression, (6) degree of parental care, and (7) reproductive effort. My data partially satisfy components 2, 4, 5, 6 and 7, but are insufficient for the others. As an index of reproductive effort I have used the ratio of clutch weight (wet weight of oviducal complement in preservative) over total, wet body weight and converted the mass ratios to caloric ratios using the formula of Vitt (1978), C./C, = 1.2905 (WW./ WW,) + 0.0640, where C./C; is the caloric ratio (clutch calories/ total body + clutch calories), WW, is wet weight of clutch, and WW, is wet weight of total body + clutch. These indices and the evidence for annual reproductive activity are compared among cer- tain coexisting species of Emoia in American Samoa. The results are discussed in terms of the two generally recognized reproductive strategies in lizards (Tinkle, et al., 1970; Duellman, 1978): (1) early maturing species with annual multiple broods, and (2) late maturing species with single broods annually. Three (non-exclusive) views attempting to account for the evolution of these strategies are (1) that early maturing species with multiple clutches are primarily REPRODUCTIVE BIOLOGY OF LIZARDS 3 tropical, oviparous taxa with smaller clutch sizes adjusted by higher intraspecific competition (Tinkle, et al., 1970), (2) that more fre- quent and smaller clutches in tropical lizards (notably Anolis) is an adaptation to their more arboreal habits (an argument extended to the family Gekkonidae, numerous arboreal species of which almost always have one or two egges per clutch, Andrews and Rand, 1974), and (3) that body shape, foraging activity, escape mechanisms and other aspects of lizard ecology predict clutch size in ecologically similar species ( Vitt and Congdon, 1978). Duellman (1978) offered a more balanced view and equated overall reproductive effort for species with low clutch sizes and multiple broods and species with high clutch sizes and single broods, depending on the overall sur- vivorship of eggs, young, and gravid females, annually, in any given environment. His view is that the two reproductive strategies have general correlations with aseasonal and seasonal environments, re- spectively, but that numerous exceptions reflect differences in com- plexity of the overall environment of the individual species, which in turn affects all of the components suggested by Wilbur, e¢ al., (1974). Data for the species in American Samoa presented herein offer interesting comparisons with these views, for, although relatively few species are found on these islands, the lizard fauna has been assembled from non-endemic, rafting colonizers with notably differ- ent ecologies (Schwaner, 1979). MATERIALS AND METHODS The distribution of lizards in American Samoa is discussed by Amerson, et al., (1978), and Schwaner (1979); major collecting localities for specimens reported herein may be found in these references. The twelve species in American Samoa are non-endemic, widespread lizards on islands throughout the Pacific area and else- where. Some species are among those lizards with the greatest known geographic distributions. A total of 2047 specimens was examined during this study; in- dividual reproductive data was obtained by dissection of preserved material. Field caught individuals were preserved in 10% formalin and later transferred to 70% ETOH. External measurements were taken with vernier calipers or a millimeter rule on preserved speci- mens Only (except for a small series of Emoia cyanura which were weighed on a 5 gm Pesola spring balance in the field prior to preservation). Snout-vent length to the nearest 0.5 mm was meas- ured ventrally as the distance from the anterior edge of the cloaca to the tip of the snout. The body cavity was opened by midventral or lateral incision and the following data recorded: (1) sex, (2) reproductive condition of ovaries and oviducts, or testes, (3) num- ber and length (greatest diameter) of all developing ova, oviducal 4 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY eggs, and testes to the nearest 0.1 mm, and (4) wet weight (to the nearest 0.1 gm) of the entire clutch complement and combined weight of clutch and body of ovigerous females. Reproductive status was judged as: (1) immature, for hatch- lings and males with small, undeveloped testes (< 2.0 mm), lack- ing convoluted tubules, or females with small whitish ova (usually less than 1.0 or 2.0 mm in size) and narrow oviducts, (2) maturing, for males with slightly enlarged testes (> 2.0 mm), lacking well developed convoluted tubules, and for females with enlarged, yel- lowish, developing ova (fecund), without noticeably distended oviducts, (3) mature, for males with enlarged testes and convoluted tubules, and females with oviducal eggs (ovigerous), and (4) old mature, for males in similar condition to (3), but with greatly enlarged and vasculated testes, and for females with greatly en- larged oviducts, and small developing ova (recently spent). These categories are rather arbitrarily defined, but, they facilitate an orderly separation of males and females into age (size) classes of recognizable reproductive condition. 3 Clutches of eggs, when found in the field, were placed in loosely tied plastic bags with some of the substrate on which they were found; if the nest material was moist, it was occasionally sprinkled with water so as to maintain moisture levels during in- cubation of the eggs. For gecko eggs the nesting substrate was usually dry, so eggs were loosely covered in dry paper toweling before placing in plastic bags. Laboratory temperatures during in- cubation approximated ambient air temperatures in shade during the day (24°-26°C), and were probably slightly lower than those in the field. Most hatchlings were photographed and preserved immediately, but some (Emoia nigra) were maintained alive on a diet of small insects for several weeks to ascertain changes in color pattern. RESULTS Emoia nigra This large diurnal skink was caught on the ground (where it forages widely), and on tree trunks (where it basks) within 3 m of the ground, At night I found two individuals, one under a rock and another in an epiphytic fern about 2 m from the forest floor. Males.—The 194 specimens have snout-vent lengths of 41-121 mm (Fig. 1). Individuals less than 80 mm generally have small, whitish testes (usually below 5.0 mm in length) lacking enlarged convoluted tubules. Testes sizes for individuals with snout-vent lengths greater than 90 mm deviate noticeable upward (exponen- tially) from a regression line fitted to individuals less than 90 mm in length (Fig. 2; Y = 0.11 X-3.90, r = 53, p < .001). Testes of most individuals above 90 mm are heavily convoluted and highly REPRODUCTIVE BIOLOGY OF LIZARDS 5 SVL (mm) ! 2 3 4 REPRODUCTIVE CONDITION Fic. 1—Reproductive condition and snout-vent length (SVL in mm) of Emoia nigra from islands of American Samoa. Reproductive conditions: (1) immature, (2) maturing, (3) mature, (4) old mature (see text for further explanation). Solid circles are males; hollow circles, females. vasculated. Males with enlarged testes were found during all month of the study indicating that they are reproductively active throughout the year. Females.—Of 223 females having snout-vent lengths of 42-114 mm, the smallest fecund female (developing ova > 5.0 mm) is 86 mm and the smallest with oviducal eggs, 89 mm (Fig. 1). Thus the size at sexual maturity for female E. nigra appears to be 86-89 mm. Mean clutch size, based on 86 fecund and 31 ovigerous females, 6 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 11.0 9.0 7.0 5.0 TESTES LENGTH (mm) 3.0 1.0 50 90 130 SVL (mm) Fic. 2.—Testes length versus snout-vent length (in mm) for male Emoia nigra from American Samoa. Dots form a size gradient representing one (smallest) to five (largest) individuals. Regression line fitted by least squares method to data points below 90 mm SVL. and three clutches found in the field, is 2.32 (range = 2-4), There is a slight, but non-significant increase in clutch size with female snout-vent length (Fig. 3A). A few large females (>> 100 mm SVL) had only one egg (all with thick leathery shells) and could have laid an egg or two just prior to capture. These data confirm Greer’s (1965) note on clutch sizes of 10 gravid E. nigra. The diameters of 84 oviducal eggs measured 13.5-22.1 mm (X = 19.0 mm) and closely approximated the length x width of freshly laid eggs (K = 19.2 x 13.5 mm, range = 18,0-20.0 x 12.5-15.0 mm, n = 7). Incubating eggs of E. nigra swell in size, presumably by absorb- ing water, and are 3-4 mm larger in length and width (X = 22.0 x REPRODUCTIVE BIOLOGY OF LIZARDS vi s«.hmUle 3 3 e e D wu 2 e ee 6 ee ee ee = : A ee ee a O 65 73 82 = ey en eee ey EN eS ee ey Eo Fo fe Se - E o gastkicgnnch -4Bi ta ing i oH. i SNOUT VENT LENGTH (mm) Fic. 3.—Clutch size versus body size for (A) Emoia nigra, (B) E. samoense, (C) E. lawesii, (D) E. adspersa, (E) E. cyanura, (F) Gehyra oceanica, (G) Crytodactylus pelagicus from islands of American Samoa. Clutch size based on ovigerous and fecund females. Dots represent individual lizards. 8 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 18.0 mm, n = 7) at hatching than at partuition. Eggs not placed on moist soil or paper toweling shriveled and failed to hatch under laboratory conditions. Three clutches of eggs were taken from a birdnest fern (Asplenium nidis, Polypodiaceae) and under rotting stumps of coco- nut trees. Times from collection to hatching are 30-65 days; as a conservative estimate, two months is probably the usual incubation time for eggs in the field. Hatchlings have snout-vent lengths 34.0-41.0 mm (X = 39.1, n = 10) and differ markedly from juveniles and adults in color pat- tern (Schwaner, 1979). In one hatchling reared in the laboratory, the pattern began to change in about two weeks and after three weeks the individual was indistinguishable in color pattern, al- though smaller in size, from most juveniles and young adults col- lected in the field. Emoia samoense This large diurnal skink was caught primarily on tree trunks and in low vegetation (where it forages and basks) from near ground level to several meters above the ground. None was observed at night, but several individuals were seen perched in trees at sunrise and sunset. Males.—The 116 specimens have snout-vent lengths of 66-118 mm (Fig. 4). Individuals less than 90 mm generally have small whitish testes (below 4.0 mm in length) lacking enlarged convo- luted tubules; specimens 90-100 mm snout-vent length have slightly enlarged testes. Reproductively active males greater than about 95 mm snout-vent length deviate noticeably upward from a regression line fitted to individuals less than 95 mm in length (Fig. 5; Y = 0.1X-4.78, r= .63, p < .001). Testes sizes for individuals above 95 mm snout-vent length cluster near 8.0-9.0 mm; thus, males probably mature about 90 mm snout-vent length. Males with enlarged testes were found during all months of the study, suggesting that repro- ductive activity probably is year round. Females.—Of 81 females having snout-vent lengths of 71-114 mm, the smallest fecund female (developing ova > 4.0 mm) is 84 mm and the smallest with oviducal eggs, 95 mm (Fig. 4). Thus, the size at sexual maturity of female E. samoense is probably 84-95 mm. Mean clutch size, based on 20 fecund and 10 ovigerous females, is 5.3 (range = 4-7). There is a significant increase in number of oviducal eggs with female snout-vent length (r = .68, p < .05, Fig. 3B). Greer (1968) examined a single gravid E. samoense with 5 oviducal eggs. The diameters of 58 oviducal eggs measured 8.6-14.5 mm (X = 12.3 mm); average egg length x width at partuition is 14.5 x 9.5mm (range = 14.0-15.0 x 9.0-10.0 mm, n = 2). These eggs also swell in size during incubation; at hatching one egg meas- REPRODUCTIVE BIOLOGY OF LIZARDS 9 SVL (mm) 1 2 3 4 REPRODUCTIVE CONDITION Fic. 4—Reproductive condition and snout-vent length (in mm) of Emoia samoense from islands of American Samoa. Half-shaded circles are juveniles; star is a hatchling. Other symbols are as in Fig. 1. ured 15.5 x 10.5 mm. Eggs of E. samoense have distinct longitudinal striations running along the external surface of the shell. Hatchlings of E. samoense have never been reported. A single E. samoense egg found with a clutch of two larger eggs of E. nigra in a birdnest fern in mangrove swamp hatched 44 days after dis- covery. This incubation time is probably only slightly less than the 10 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY true incubation period. The hatchling snout-vent length is 31.0 mm. Whether or not the one discovered egg represents single place- ment of eggs by E. samoense cannot be concluded. Captive females have laid one or two eggs in laboratory cages and withheld the rest for several days or weeks. However, none of these eggs developed, and most were thin shelled and probably laid prematurely. A field observation with anecdotal implications, was made at the time of discovery of the single E. samoense egg. Moments after the collec- tion, a gravid E. nigra was captured from an adjacent mangrove tree. Upon dissection she was found to have three oviducal eggs and three E. samoense eggs in her stomach and intestines. Perhaps { 0.0 ~ Le) TESTES LENGTH (mm) © 1.0 50 60 70 80 90 i0o0 )=—h H1O-s«*120 SVL(mm) Fic. 5.—Testes length versus snout-vent length (in mm) for male Emoia samoense from American Samoa. Symbols are as in Fig. 2. Regression line fitted by least squares method to data points below 95 mm SVL. REPRODUCTIVE BIOLOGY OF LIZARDS uu these eggs were from the clutch found in the birdnest fern. The presence of other lizard eggs in stomachs of E. nigra individuals was also recorded (Schwaner, 1979). Emoia cyanura This small diurnal skink was caught on the ground and in low vegetation, where it moves rapidly and forages widely. Individuals were disturbed from presumed resting places under rocks and leaf litter at night. Males.—The 292 specimens have snout-vent lengths of 23-58 mm (Fig. 6). Individuals less than 35 mm had small whitish testes (< 2.0 mm in length). Because there is no apparent break in the testes verses snout-vent length regression (Fig. 7), adult males were estimated from visual inspection of the testes for the presence of enlarged convoluted tubules (Fig. 6). Tubules are only moderately developed in specimens 33-42 mm snout to vent; definite convo- lutions and vasculation of enlarged testes are found in individuals 38-50 mm and larger. Thus, snout-vent length at maturity in male E. cyanura appears to be 38-40 mm. Baker (1947), in a more accurate assessment based on body weight of adult male E. cyanura in the New Hebrides, estimated weight at maturity to be 1.5 gms. Among 34 males (with complete tails) weighed prior to preser- vation during the present study, four were classified as reproduc- tively mature, weighed approximately 1.5 gms (1.5-1.7 gms) and had an average snout-vent length of 42.8 mm (range = 40-43 mm); this is only slightly higher than my estimate for minimal adult maturity based on testes size and condition. Females —Of 285 females having snout-vent lengths of 22-56 mm the smallest fecund female (developing ova > 3.0 mm) is 35 mm and the smallest with oviducal eggs, 41 mm (Fig. 6). Thus, the size at sexual maturity of female E. cyanura in American Samoa is about 40 mm. Baker (1947) found no oviducal eggs in any female weighing 1.5 gms or less, and estimated the minimum adult weight for female E. cyanura in the New Hebrides as 1.6 gms. Of 23 female E. cyanura (with complete tails), captured during the pres- ent study and weighed prior to preservation, five classified as re- productively active are below 1.6 gms (1.0-1.5 gms) and average 43.5 mm snout to vent. Two individuals weighing 1.6 and 1.7 gms have body lengths of 48 to 46 mm, respectively. This is consider- ably higher than my estimate of minimum adult size at 40.0 mm based on the size and condition of developing ova. Either females mature at smaller sizes in American Samoa, or there is some dis- crepancy in equating body size and body weight in these females with the measurements of Baker’s (1947) analysis. Mean clutch size, based on 46 fecund and 46 ovigerous females, and 25 clutches found in the field, is 1.96 (or an almost constant 2 12 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY SVL (mm) Fs) o9 un ° w ° ° go Ce $8 ee [e,e) (e) COOOO00000 0000 D [e) m ee 2 US) a le) [eo ej) Z O e f= ° G g % _— 8 fe) < ° O° m7 EOS 8 taeget 6 3: 4 = 8 2 a Hite Pe Hie: i fe hat? Fic. 6.—Reproductive condition and snout-vent length (in mm) of Emoia cyanura from islands of American Samoa. Symbols are as in Fig. 1. REPRODUCTIVE BIOLOGY OF LIZARDS 13 TESTES LENGTH (mm) 30 40 50 60 SVL (mm) . Fic. 7.—Testes length versus snout-vent length (in mm) for male Emoia cyanura from American Samoa. Symbols are as in Fig. 2. eggs per clutch). There is no significant correlation between in- creasing body size and clutch size (Fig. 3E). Greer (1968), citing Baker (1947), noted that almost always one egg is found in each oviduct of ovigerous females. Similar observations were noted by Hediger (1934) for this species. The diameters of 39 oviducal eggs measured 9.5-13.2 mm (X = 11.2 mm) and closely approximated the length x width of freshly laid eggs (KX = 12.0 x 7.5 mm, range = 10.2-13.0 X 6.2 X 9.5 mm, n = 18). Incubating eggs of E. cyanura swell in size, increasing more in width than in length from partuition to hatching; mean length x width at hatching is 12.6 X 9.5 mm (range = 12.3-12.9 X 14 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 9.2-10.1 mm, n = 6). Incubation times were 6-51 days for field collected eggs maintained in the laboratory; however, the upper bound of 40-51 days is probably the usual incubation period. Aver- age hatchling snout-vent length is 22.0 mm (range = 20.0-23.0 mm, n = 45); hatchlings have the adult color pattern. Emoia cyanura lay eggs in communal nesting sites probably selected for optimal moisture and temperature. Most eggs were taken from under garden rocks; these were usually flat coral plates lying on a sand substrate. The usual number of eggs under the rocks was about 6 to 10; however, one rock covered 70 incubating eggs and many empty egg casings. Emoia lawesii This large diurnal skink, was caught on the ground where it forages among coral rubble primarily in littoral forests. None was observed at night. Males.—The 26 specimens have snout-vent lengths of 77-106 m (Fig. 8). One male (SVL = 81 mm) has undeveloped testes; oe twe smaller individuals (SVL = 77 and 80 mm, respec- tively) have enlarged testes with convoluted tubules. A conservative estimate of the snout-vent length of reproductively active males is about 85 mm; mature testes lengths are usually above 4.0 mm (Fig. 9). Individuals were collected only during June and July, 1976, and December and January, 1977-1978. Reproductively active males were observed during June, July, December and January. Females.—Of 38 females having snout-vent lengths of 70-105 mm the smallest fecund female (developing ova > 2.0 mm) is 78 mm and the smallest with oviducal eggs, 88 mm (Fig. 8). Con- servatively, reproductive activity of female E. lawesii probably begins at snout-vent lengths 85-90 mm. Mean clutch size, based on 22 fecund and 5 ovigerous females, and 2 laboratory clutches, is 1.8. Most females have two eggs per clutch; two individuals had a single oviducal egg and three others had only one developing ovum each. There is little indication of increasing clutch size with greater body size, but too few females from the lower size range of reproductive maturity were examined (Fig. 3C). The diameters of 8 oviducal eggs measured 18.0-19.7 mm (X = 19.2 mm); at partuition the average egg length x width was 21.3 x 11.8 mm (range = 21.0-21.5 x 11.0-12.5, n = 4). Similar to the other Emoia species in this study, the eggs of E. lawesii swell dur- ing incubation reaching hatching sizes 3.7-4.7 mm longer and wider, respectively, than at partuition. Two eggs laid in the laboratory hatched 72 and 77 days after partuition. This is the longest in- cubation time for any of the scincid species observed in American REPRODUCTIVE BIOLOGY OF LIZARDS 15 The) @ ° °838 $ o8e 88 * CO0ee oe ae: 90 8 § e 5 ®@ e 4 1 2 3 REPRODUCTIVE CONDITION Fic. 8.—Reproductive condition and snout-vent length (in mm) of Emoia lawesii from islands of American Samoa. Symbols are as in Figs. 1 and 4. Samoa. Hatchling color patterns resembled those of adults; snout- vent lengths of two hatchlings were 32.4 mm and 33.4 mn, re- spectively. Emoia adspersa This medium sized diurnal skink was caught on the ground where it was observed to bask and forage. No individuals were observed at night, but three were seen entering and leaving sus- pected burrows at the bases of trees on Swains Island. Males.—The 17 specimens examined from Swains, Savaii, Nuku- nonu and Funafuti islands have snout-vent lengths of 64-84 mm 16 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 8.0 7.0 6.0 5.0 TESTES LENGTH mm 4.0 3.0 80 90 100 ite) SVL mm Fic. 9.—Testes length versus snout-vent length (in mm) for male Emoia lawesii from American Samoa. Symbols are as in Fig. 2. (Fig. 10). Only the smallest individual (64 mm) has undeveloped testes; individuals of greater size have testes (with enlarged convo- luted tubules) sizes of 3.7-5.5 mm. Conservatively, adult males are probably reproductively active at 65-70 mm _ snout-vent length. Most individuals (for which dates are available) were taken during the months of April and May, and one each were collected in REPRODUCTIVE BIOLOGY OF LIZARDS 17 SVL (mm) 1 2 3 4 REPRODUCTIVE CONDITION Fic. 10.—Reproductive condition and snout-vent length (in mm) of Emoia adspersa from islands of American Samoa, Savaii, Nukunonu, and Funafuti. Symbols are as in Fig. 1. February and September; all adult specimens have swollen testes with enlarged convoluted tubules and are presumed to have been reproductively active when perserved. Females.—Of 23 females having snout-vent lengths of 63-81 mm, the smallest fecund female (developing ova > 2.0 mm) is 64 mm and the smallest with oviducal eggs, 70 mm (Fig. 10); thus, con- servatively, most females probably are mature at 70 mm. Mean clutch size, based on 14 fecund and 5 ovigerous females, is 1.9. All but two individuals have two eggs per clutch. Clutch size does not appear to increase with body size (Fig. 3D), but too few females from the lower size range of adults makes this incon- clusive. Egg sizes for seven full term oviducal eggs (all with thick leatherly shells) average 16.4 mm (13.4-19.2 mm, n = 7) in length. 18 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Cryptoblepharis boutonii This small, diurnal skink was caught on coral and lava rocks in littoral strand habitats devoid of vegetation at coastal margins on Tau and Olosega islands. None was observed at night, but the species is apparently restricted to the littoral strand. Males.—The 26 specimens have snout-vent lengths of 25-42 mm (Fig. 11). Individuals less than 37 mm generally have small, whitish testes (< 2.0 mm in length) lacking enlarged convoluted tubules, and are probably immature (Fig. 12). Conservatively, male C. boutonii are probably reproductively mature at snout-vent lengths of 38-40 mm and testes sizes of 3.0-4.0 mm in American Samoa. Females—Of 23 females having snout-vent lengths of 35-48 mm, the smallest fecund female (developing ova > 1.0 mm) is 44 mm and the smallest with oviducal eggs, 43 mm (Fig. 11); one spent female with distended oviducts is 40 mm. The size at sexual maturity for female C. boutonii is, therefore, probably at or above 40 mm snout-vent length. Mean clutch size, based on 12 fecund and 3 ovigerous females, is 1.93; only one individual had one oviducal egg. Thus, the clutch size is probably a constant two eggs in American Samoa. No eggs were collected from field nests; Haake (1977) states 50 40 SVL (mm) 30 1 2 3 4 REPRODUCTIVE CONDITION Fic. 11.—Reproductive condition and snout-vent length (in mm) of Cryp- toblepharis boutonii from islands of American Samoa. Symbols are as in Fig. 1. REPRODUCTIVE BIOLOGY OF LIZARDS 19 TESTES LENGTH (mm) 35 45 SVL (mm) Fic. 12.—Testes length versus snout-vent length (in mm) for male Cryp- toblepharis boutonii from American Samoa. Symbols are as in Fig. 2. that the species reproduces “by means of small batches of soft- shelled eggs which are laid in moist sand.” The average length of seven oviducal eggs was 10.1 mm (range = °6.9-13.5 mm). Lipinia noctua This small, diurnal skink is secretive and almost invariably found under the bark of rotting trees, or in epiphytic vegetation. Individ- uals were not observed at night. Males.—The 15 specimens have snout-vent lengths of 25-43 mm (Fig. 13). Individuals less than 35 mm snout-vent length have 20 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 50 ° ee ie 00 i e E 8 = ace fe @ > n 30 1 2 3 4 REPRODUCTIVE CONDITION Fic. 13.—Reproductive condition and snout-vent length (in mm) of Lipinia noctua from islands of American Samoa. Symbols are as in Fig. 1. small, undeveloped testes; convoluted tubules in testes > 2.0 mm in length are present in individuals above 35 mm, which is probably the snout-vent length at reproductive maturity (Fig. 14). Females.—Of 12 females having snout-vent lengths of 39-47 mm the smallest fecund female (developing ova > 1.0 mm) is 39 mm and the smallest ovigerous female, 40 mm; because smaller individ- uals were not examined, minimal size at reproductive maturity is at least 39 mm (Fig. 13). Lipinia noctua is viviporous; nine females with developing em- bryos have two (one in each oviduct), and three females have a single embryo. Fitch (1970) reported Hediger’s (1934) observation of one embryo per female in the New Hebrides; Oliver and Shaw (1953) found two embryos in each of six females, and two females with one embryo. Similar results were noted for gravid females from the Marshall, Taumotu, and Marquesas Islands (Fitch, 1970). Gravid females were observed during January, March, June and December in American Samoa; thus reproductive activity is prob- ably year round. The average diameter of seven embryo masses is 8.9 mm (range = 7.1-10.8 mm). Four full term embryos had an average snout-vent length of 15.5 mm (range =: 15.1-16.6 mm). REPRODUCTIVE BIOLOGY OF LIZARDS 21 TESTES LENGTH (mm) 30 4O 50 SVL (mm) Fic. 14.—Testes length versus snout-vent length (in mm) for male Lipinia noctua from American Samoa. Symbols are as in Fig. 2. Gehyra oceanica This large, nocturnal gecko was caught on building walls, and on tree trunks in a variety of forested habitats. During the day in- dividuals were disturbed from presumed resting sites under the bark of rotting trees, particularly coconut trees, or from under boards and other debris, and crevices in building walls and roofs. Males.—The 25 specimens have snout-vent lengths of 36-93 mm (Fig. 15). Reproductively active individuals with enlarged testes and convoluted tubules apparently mature at snout-vent lengths greater than 70 mm in American Samoa; below this size testes are 22 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 95 75 SVL (mm) 1 2 3 4 REPRODUCTIVE CONDITION Fic. 15.—Reproductive condition and snout-vent length (in mm) of Gehyra oceanica from islands of American Samoa, Symbols are as in Figs. 1 and 4; X’s are hatchlings, REPRODUCTIVE BIOLOGY OF LIZARDS 23 small (< 3.0 mm), undifferentiated masses of whitish tissue (Fig. 16). Mature males found during several months of the study are presumably reproductively active year round. Females.—Of 34 females having snout-vent lengths of 47-67 mm, the smallest fecund female (developing ova > 3.0 mm) and the smallest with oviducal eggs, 70 mm (Fig. 15). Females 70 mm or less have small underdeveloped ova with narrow, undistended ovi- ducts and are apparently immature. Thus, snout-vent lengths of both male and female G. oceanica appear to be 70 mm at repro- ductive maturity. Mean clutch size, based on 11 fecund and 6 ovigerous females, and 22 clutches found in the field, is 1.96. Most fecund and ovigerous females have two eggs; only two females have a single oviducal egg and one fecund female has four distinctly enlarged ova. A large number of field nests (48%), however, contained a single egg, and a slightly greater percentage (52%) contained two eggs. There is a tendency for females of greater body size to have more than two eggs in a clutch (Fig. 3F), but this trend is not consistent. Most gekkonid eggs have hard, brittle shells and apparently do not swell during incubation. The diameters of six oviducal eggs measured 7.5-16.5 mm (XK = 11.4 mm); length x width of 25 eggs from field clutches averaged 12.8 x 11.6 mm (range = 12.0-13.6 mm x 11.0-12.5 mm). Minimum hatching time for G. oceanica eggs was 19 days in the laboratory; eggs hatching between 19 and 36 days were usually dark when candled, revealing the eye pigment of developing em- bryos. Most eggs that hatched beyond 40 days were yellowish or pinkish in color when candled. The longest incubation periods recorded were a rather remarkable 112 and 114 days for two eggs collected on Swains Island in mid-May, 1976. Two eggs from Ta’u Island, taken in January, 1978 took 102 and 109 days to hatch. Brown and Alcala (1957) reported on incubation periods of 70 days for the gekkonid, Cosymlotus platyurus; Fitch (1970) cited Love- ridge’s (1945) note (sensu Smith, 1935) indicating a five month in- cubation period for Ptychozoan kuklii in Java; most other gekkonid species for which data are available have incubation times between one and two months. Furthermore, the eggs from American Samoa were kept in dry plastic bags on a shelf in the laboratory and were not considered viable after 3 months, until they unexpectedly hatched. Their apparent resistance to dessication and long incu- bation times undoubtedly facilitates their transport via rafting among islands. Hatchlings have snout-vent lengths of 28.0-30.0 mm (X = 29.4 mm, n = 19). Individuals from Swains and Rose islands have a 24 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 6.0 4.0 20 TESTES LENGTH (mm 50 70 ; 90 SVL (mm) Fic. 16.—Testes length versus snout-vent length (in mm) for male Gehyra oceanica from American Samoa. Symbols are as in Fig. 2. suffusion of rose colored pigment on the underside of the tail, and a faint black midventral line. When alarmed, these hatchlings raise ° their tails vertically, presumably to display the bright color. Tail autonomy is high in gekkonid species (Pianka and Pianka, 1976), and over 80% of the G. oceanica specimens from American Samoa have regenerated tails. All but one G. oceanica egg was found under the bark of rotting coconut stumps, usually near their base; the eggs were in clutches of 1-10, in varying stages of development, indicating communal use of nesting sites. A single egg was found in the tangled roots of a fallen tree. REPRODUCTIVE BIOLOGY OF LIZARDS 25 Hemidactylus frenatus This medium sized, nocturnal gecko was found exclusively on building walls on Tutuila Island; a small series taken from build- ings on Oahu, Hawaii are included with the American Samoan specimens in the following description. No individuals were ob- served during the day. Males.—The 18 specimens have snout-vent lengths of 33-60 mm (Fig. 17). Adult males with enlarged testes and convoluted tubules are greater than 40 mm snout-vent length and probably become 60 40 SVL (mm) 20 4 2 3 REPRODUCTIVE CONDITION Fic. 17.—Reproductive condition and snout-vent length (in mm) of Hemi- dactylus frenatus from islands of American Samoa (and Hawaii). Symbols are as in Figs. 1 and 4. 26 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY reproductively active between 40-50 mm. Most males above 50 mm have testes sizes between 4.0-5.0 mm (Fig. 18). Reproductively active males were found during June and January. Church (1962) observed active reproduction year round in Java; farther north in more temperate areas this pattern became seasonal with peak activity in July and August (Fukada, 1965). Females.—Of 15 females having snout-vent lengths of 33-52 mm the smallest fecund female (developing ova > 1.0 mm) and the smallest with oviducal eggs are both 43 mm (Fig. 17). Thus, size at reproductive maturity in females in American Samoa (and Hawaii) appears to be greater than 40 mm. Mean clutch size, based on 5 fecund and 3 ovigerous females, and one field clutch, is 1.9 (or almost a constant 2 per clutch); only one female had a single oviducal egg. Oviducal eggs are only slightly smaller than those freshly laid (X = 7.5 mm, range = 6.9- 8.5mm, n = 6). Just prior to hatching the eggs are identical to their size at partuition. Two clutches (four eggs) laid in the labo- ratory hatched between 77-88 days. These hatching times corre- spond closely to the “seventy-plus days” reported by Brown and Alcala (1957) for the species in the Philippines Islands. These authors noted that two eggs of H. frenatus exposed to sea water for 50 to 168 hrs hatched in 56 days, suggesting that the extent of egg development may not be of significance in determining the effects of sea water. Eggs of American Samoan females were hatched in dry plastic bags. The widespread occurrence of H. frenatus throughout the Pacific region and elsewhere is undoubtedly aided by the high resistance of its eggs to a wide range of environmental extremes. Hatchlings measured 19.0-21.0 mm snout-vent length (KX = 20.5 mm, n = 5) and resembled the adults. One clutch of eggs was recovered from the field under a wooden box on the ground near a building wall on Tutuila Island. Hatchlings were not observed in the field. Peropus mutilatus This small, nocturnal gecko was found exclusively on darkened ° buildings not exposed to village lights. It is rare in American Samoa and never observed during the day. Males.—The 8 specimens have snout-vent lengths of 36-48 mm (Fig. 19). A single individual (31 mm, SVL) is definitely immature (testes size = 1.2 mm), and a second individual (36 mm, SVL) has only slightly larger testes (1.4mm). Five individuals, with enlarged testes and convoluted tubules, range above 40 mm _ snout-vent length; testes sizes range from 3.0-3.7 mm (Fig. 20). Thus, male P. mutilatus in American Samoa probably mature at snout-vent lengths greater than 40 mm. Males taken during July, December and Jan- REPRODUCTIVE BIOLOGY OF LIZARDS 27 TESTES LENGTH (mm) 40 60 SVL (mm) Fic. 18:—Testes length versus snout-vent length (in mm) of male Hemi- dactylus frenatus from American Samoa (and Hawaii). Symbols are as in Fig, 2. uary were reproductively active. Church (1962) reported year-round breeding for this species in Java. Females.—Of 12 females having snout-vent lengths of 33-48 mm 28 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 50 45 40 SVL (mM) 35 1 2 3 4 REPRODUCTIVE CONDITION Fic. 19.—Reproductive condition and snout-vent length (in mm) of Peropus mutilatus from islands of American Samoa. Symbols are as in Fig. 1. the smallest fecund female (developing ova >. 1.0 mm) is 40 mm and the smallest with oviducal eggs, 43 mm (Fig. 19). Thus, size at reproductive maturity in female P. mutilatus from American Samoa is the same as for males, 40 mm. Mean clutch size, based on 7 fecund and 4 ovigerous females, is 1.8. With the exception of two large females (46 mm and 43 mm, respectively ) that probably had laid eggs and were just beginning the development of a second clutch (both were fecund with a single’ enlarged ovum), all gravid females in American Samoa have two eggs per clutch. Fitch (1970) noted two eggs per clutch for this species in other areas. Three oviducal eggs measured 6.9-8.1 mm (X = 7.5mm). No eggs were found and hatchlings were not observed in the field. Cyrtodactylus pelagicus This medium sized, nocturnal gecko is largely terrestrial. Indi- viduals were usually caught on tree trunks in littoral forest, within REPRODUCTIVE BIOLOGY OF LIZARDS 29 4.0 3.0 @ 2.0 TESTES LENGTH (mm) 1.0 40 60 SVL (mm) Fic. 20.—Testes length versus snout-vent length (in mm) for male Peropus mutilatus from American Samoa. Symbols are as in Fig. 2; large dot represents two individuals. one meter of the ground; when disturbed the species invariably escaped downward among rocks and litter of the forest floor. Two individuals were captured during the day under debris on the ground in village land. No males were found in any sampled population. Females.—Of 51 females having snout-vent lengths 23-69 mm the smallest fecund female (developing ova > 1.0 mm) is 50 mm and the smallest ovigerous female, 59 mm (Fig. 21). Thus, reproductive maturity occurs between 50-60 mm snout-vent length in C. pelagi- cus from American Samoa. Mean clutch size, based on 29 fecund and four ovigerous fe- males, is 1.8; five fecund and two ovigerous females have only one developing ovum each. There is only a slight correlation between body size and clutch size in C. pelagicus individuals from American 30 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 70 30 2 cae 4 REPRODUCTIVE CONDITION Fic. 21.—Reproductive condition and snout-vent length (in mm) of Cyr- todactylus pelagicus from islands of American Samoa. Symbols are as in Fig. 1; this is an all-female population. Samoa, but more specimens from all size ranges are needed for greater clarity of this point (Fig. 3G). The diameters of six oviducal eggs measured 9.9-12.0 mm (X = 11.3 mm). No field clutches were found, but a single egg laid in a laboratory cage measured 12.2 mm in length. No hatchlings were found in the field. Two dense populations of C, pelagicus were sampled from Ta’u Island, both in littoral forests near the coast. In these habitats few other gekkonid species were found. Most importantly, no males of C. pelagicus were collected. Almost all of the females are gravid REPRODUCTIVE BIOLOGY OF LIZARDS ‘31 and there is very little apparent variation in color pattern or meas- urements of various body parts. One female isolated in the labora- tory developed and laid, one egg during a seven month period, but the egg failed to hatch. The lack of males in the collections (taken during July and January) suggests that a parthenogenic population of C. pelagicus exists on Ta’u Island. Lepidodactylus lugubris This small, nocturnal, abundant gecko was caught in a number of habitats at night, ranging from building walls to village shrubs, forest epiphytes, rotting trees, and the rocky cliffs of coastal strand. Individuals were taken under the bark of trees, in epiphytic vege- tation, under debris at the shoreline, in cracks of building walls and roofs, and in crevices between rocks in littoral strand by day. The species is a known parthenogenic form; no males were found in the populations sampled. Females.—Of 213 females having snout-vent lengths of 20-44 mm the smallest fecund female (developing ova > 1.5 mm) and the smallest ovigerous female are both 35 mm (Fig. 22). Although presumably immature individuals were recorded at higher snout- vent lengths, it seems that female L. lugubris from American Samoa mature at 35 mm. Furthermore, the modal snout-vent length for fecund, ovigerous and spent females is 40 mm. Clutch size, based on 45 fecund and 47 ovigerous females, and 35 field clutches, is essentially a constant 2 eggs. In the field, two eggs in a clutch were observed without error because L. lugubris is the only gecko on these islands which lays eggs with shells that adhere to each other upon drying; this was also noted by Cagle (1946) for the species on Tinian Island. These eggs are affixed to the substrate and are difficult to detach without breakage. It is probable that the survival of L. lugubris in rocky littoral strands is facilitated by this trait, since affixed eggs may be secured in cracks and crevices of rocks and on cliff overhangs where they es- cape predation by the numerous crabs along the rocky coastline. Mean egg size at hatching (XK = 8.3 x 7.1 mm, range = 8.0-9.0 x 6.8-7.5 mm, n = 3) is only slightly larger than the mean size of oviducal eggs. The eggs must be extremely resistant to dessication because they were not found in moist (fresh water) areas, but under dead and dying tree bark, rocks exposed to salt spray, and small dry holes in coconut palms. On Rose Island, where L. lugu- bris coexists with G. oceanica, eggs of the latter were found under dead tree bark always at the base of the tree, and eggs of the former were always affixed to bark near the top of the tree. Maximal hatching time in the laboratory was 73 days; two months appears to be the modal incubation period. Oliver and Shaw (1953) reported an incubation time of 92 days, but develop- 32 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 45 ° fo} (ore) ° lele} ooo loke) O©00000000 ©000000 00000 OCODDDN0DDO00000O0 DOOOOODDD000000 O000000000 OC00000000 ©000000000 O000000 fo) O0000000 ©0000 00000 foXe) ©00000000 fole) fefere) oo00000 lele) [e) 1@) 35 c00 O0000000 lokeyveye) [ofe} 2000000 feYeye) 00000 ©00000 ©000000 ie) E oo = | = 25 15 1 2 S55 4 REPRODUCTIVE CONDITION Fic. 22.—Reproductive condition and snout-vent length (in mm) of Lepi- dodactylus lugubris from islands of American Samoa. Symbols are as in Fig. 1; this is an all-female population. ment was perhaps delayed by cooler than normal incubation tem- peratures (Fitch, 1970). Mean snout-vent length of hatchlings is 17.1 mm (range = 14.0-20.0 mm, n = 9). Lepidodactylus lugubris is parthenogenetic (Cuellar and Kluge, 1972); no males were found in American Samoa. The species is distributed from Australia and the Indo-Pacific region, throughout islands of the tropical Pacific Ocean and on mainland Central America and Ecuador (Smith, et al, 1961; Fugler, 1966), where it is often the most locally abundant species. Kluge and Cuellar (1972) list the diploid chromosome number as 44. From this they suggested that the mode of parthenogenesis is not due to hybridi- REPRODUCTIVE BIOLOGY OF LIZARDS 33 zation with closely related species (or gynogenesis), since no con- geners or even closely related genera are found among populations of L. lugubris, wherever it is found. Undoubtedly the wide range of L. lugubris both in geography and habitat distributions in Amer- ican Samoa (Schwaner, 1979), is facilitated by its unique combi- nation of reproductive habits. DISCUSSION Patterns in Monthly Reproductive Activity To demonstrate the apparent constancy of the American Samoan islands with respect to annual patterns in temperature and rainfall, Colwell’s (1974) method was used in an analysis of predictability, constancy and contingency for monthly averages of precipitation and temperatures taken from published weather records for 17 con- secutive years (U.S. Bureau of Standards, 1960-1976). These data are cast into the matrices of Tables 1 and 2, respectively, for pre- dictability analyses, along with the computed proportions of con- stancy and contingency contributing to predictability. Rainfall (Table 1) is highest during December to April (the so called “rainy season”); drier months are from May to December, having lows in August and September. The pattern is not marked, however, and its predictability is only 54%. Constancy of the pattern is high how- ever (48%) in contrast to the contingency factor (6%); the former factor contributes 88% to the predictability of the pattern (as op- posed to only 12% for contingency). Therefore, whereas the pattern TasLe 1.—Predictability, constancy and contingency matrix for mean monthly distribution of precipitation on Tutuila Island, American Samoa, for the years 1960-1976. Rainfall is divided into logarithmically increasing classes, with the upper bound at > 64.00 cm. Amount monthly Months Tatas AAS! BOUAN=S Dee “MOA eM yy OA (cm) (oe) S i=) Ree OoNN re RPOsN Ee ome | Hw CO UL bd oO Oe mee Od eb od © i Npowbd Re -~10 ® © oo OD. De © CO UL Predictability = .54 Constancy = .48 Constancy/Predictability = .88 Contingency = .06 Contingency/Predictability = .12 34 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY TasLeE 2.—Predictability, constancy and contingency matrix for mean monthly temperature distributions on Tutuila Island, American Samoa, for the years 1960-1976. Temperature is divided into logarithmically increasing classes, with the upper bound at > 28.20°C. Mean monthly Months temperature S. O:. N DJ FF MA Mey (°C) 25.00 25.05 25.10 25.20 25.40 25.80 A iyit ge D apes: 26.60 ley alley pA 28.20 See > 28.20 Predictability = .78 Constancy = .61 Constancy/Predictability = .78 Contingency = .17 Contingency/Predictability = .22 ly }Sycrt lO as 3y Dhid4y Ta2o7 \ ee 4 5 © i DOW is highly constant (p < .001) and thus predictable (p < .001), it is not due to contingency (or seasonality, p > .95). Mean monthly temperatures (Table 2) are highly predictable (78%, p < .001). Constancy contributes a highly significant 78% (p < .001) to the predictability of temperature, but contingency accounts for a nonsignificant 22% (p > .5) of the pattern. Mean monthly temperatures vary only 1.3°C throughout the year. Thus, there is no statistical basis for recognizing seasonality in temper- ature or precipitation on American Samoan islands, and the climate is highly predictable by its constancy (at least during the years for which measurements were analyzed). Monthly collections of individuals of Emoia nigra, E. samoense and E. cyanura were sufficiently large (although unevenly distrib- uted) to detect patterns in reproductive activity. I defined three cate- gories of adult female reproductive status: (1) Ovigerous females contain oviducal eggs; (2) fecund females are of minimal adult, body size (as estimated above) and contain enlarged ova (> 3.0 mm in diameter for E. cyanura, > 5.0 mm for E. nigra, and > 4.0 mm for E. samoense); and (3) non-fecund females have minimal adult body sizes and ova sizes less than those for fecund individ- uals. Reproductively active females are both ovigerous and fecund (Tables 3 and 4). Maximal reproductive activity for Emoia cyanura was highest from November to April and lowest from May to October (Fig. 23). Reproductively active females were found in most months of the year, however, and were never below 25% frequency. When the 35 REPRODUCTIVE BIOLOGY OF LIZARDS ‘soynurur ut (ATY}UOUL) 3YysI[Aep Jo uoneing = quad ‘(ydur) paads purm A[yyUou ueayy = SAIN ‘(sIy OOTO) AHprumy eayejer A;yUOW uray = NHYW ‘(sty OOST) AWpramy eanejer [UU UvaW = YHYIN ‘IaA00 Ays ATYJUOW URI = OSWIW *(soyour) q[ezurer A[quou ueayy = NYININ (Do) einjyeredure} Ie A[yUOUI URaWy = LYWIWr SSS. SSS gs = cs = = 6€ 9% 9s ee 08 = 6F panuvhia “q cI 99 os = = oP = = 0s eV = cI ISUBOWDS “7 03 0 OF = = GG Iv L9 OF 8G = cs DIG “FT OSL COL PPL QL SOL 069 GLO 069 80L QL PPL COL ayunad 0'8 C6 O1T GOI GIT cit Oe OS GL c‘9 C9 08 SMW 98 98 cs cs cs ¥8 98 68 06 16 06 68 NHYW cL 9L OL SL PL GL LL OL OL OL OL 9L VHUIWN 6 9° VL v9 G9 89 39 oL SL La 8 18 OSININ 09¢ 69% BOE SOT CSI SLI 91% PVG ZIe G6G BSE GZe NUNN GAG OG 5 919G FAvi9c) 0I9G SC9G. Oia B8i0G ee cic Cro ceo ere LYWN d N O S Vv { { W Vv W A f setoods prez] pue Sy}UuOjy qsieyoueied o1eUWID SS ——————— ———— : ‘OL6T °F 9S6T steaA 9Y} J0F (poysiqndun ‘g/6T) ‘7o 7a ‘uosiowy WIOIF 1B BJVp OHVUN[D ‘eOUTeG UROLIOWIY JO SpURISI WOIZ UOIpUoD saNonpoidax Ur so[eUley DInuDho “y pue asuaowDs “q ‘DIsIU Mow JO syueoI10d A]yyUou pue siojoutered oneUT[O 10; eyep ATYJUOU ULeWW—'¢ AIAVY, 36 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 100 @— monthly %o O---- bimonthly % PERCENT REPRODUCTIVITY a ° 10 MONTH Fic. 23.—Monthly and bimonthly percent of reproductively active (oviger- ous and fecund) females of Emoia cyanura from islands of American Samoa. various stages of the reproductive cycle (ovigerous, fecund and spent females, and hatchlings) are considered separately (Fig. 24), the monthly percentages of spent females (relative to the total number of reproductive stages) is inversely proportional to simi- larly computed frequencies of hatchlings, with maximal divergence in December, March, and June. If eggs are laid in January (as indicated by the small peak in frequency of spent females) and hatch within about two months (as suggested by incubation times in laboratory clutches), hatchlings would emerge in March and April, as illustrated by the peak in hatchling frequencies during ‘ those months. Percentages of fecund and ovigerous females increase from October to December and January and decline with spent frequencies toward March. Large numbers of field clutches were found in February and March, 1976 (and again in June, 1976). Percentages of fecund females increase slightly from a low in March to a peak in May, and frequencies of spent females peak a month later, while the former drops to its lowest point (14%) dur- ing the months sampled. Hatchling percentages also reach a low (10%) in June. Monthly mean diameters of developing ova (numbers in paren- 37 REPRODUCTIVE BIOLOGY OF LIZARDS CZ - - - = Ue = 9% oe Se = GE —- azIs BAO UAW I 0 0 0 = II = 0 T 0 = i ssullyoye yy V 0 I I = I = IT 0 I = 8ST quedg ST 1 T I = v = g T v = LI punoayuony G Il 0 0 = T Ss 0 0 I = I punoa,7 T T I 0 i G = 0 I! G = G SNOIBZIAO QSUBOWDS °F IS 7 TE = = Ws SV VV 0s 6¢ = (4 9zIS BAO UBIO 0 = 0 = zs 0 i OT 13 Ol a Zz ssul[youeH € i V = ca L ve ial 91 I = LG quadg 9 = € z my 8 6 OT 9G I = 8 punosjsuo Nf r4 = i = = v v L 9 € = 8T punoey 9 ins 0 = z I V G L I = 61 SNOIZIAO pinuvho *7 1 4 = me = = LS OV OL 6S 9¢ = 9€ 9ZIS BAO Uvo{ 0 0 0 = = I € 0 9 I = I ssulyqovey 9G 0 0 = ze V I 0 G 0 = 1G quedg ce I € = zi GI eT G 6 S = 6E PunosyuON G = 0 a a € S Vv € 0 7 c punse p 0 3 2 = T 9 0 € z = 8 SnOIeZIAQ DIsWU DIOW d N O S V f f W V W A { quo ‘Udye} 919M vJepP OU 9}eOIpPUT (—) soYyseq “vOlUBSG UKONOUIY WOT spieZIT ploulos Jo soloads 9ae1y} 10Z seyeuley (pUNoezuOU pue) punosz Jo sozIs BAO UROU puUe SsUI[YO}eY ‘so[eUay y[Mpe yUeds pue puUNdazUOU ‘puNde} ‘snoIeSIAO Jo sioquinu A[YJUOWW—yF AIAV], 38 CCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY thesis above fecund frequencies in Fig. 24) appear to increase in size from December, peaking in January, and then decreasing in average size from January to April, just at the time when ovigerous and spent females of E. cyanura also decline in frequency. Ova sizes increase again in April and May, and decline in June, the same month during which spent females reach peak frequencies. The failure of ovigerous frequencies to increase and peak just prior to the high peak of spent females in May and June is incon- sistent with patterns seen from October to December. This may be a result of low sample sizes for those months; however the overall unevenness of samplings during all months of the study strains the credibility of any stronger conclusion. The data do seem to follow a consistent pattern which appears to result in two peaks of egg production, one from December to January, and another in June and July; critical data from August and September are entirely lacking, however, and greater sample sizes for each month are obviously needed. Maximal reproductive activity for Emoia nigra seems to peak 100 eee 50 / ‘ “spent PERCENT oviducal 10 / “yO. RS MONTH Fic. 24.—Monthly percent of the total counts for hatchlings, spent females, fecund females (ovarian) and ovigerous (oviducal) females of Emoia cyanura from islands of American Samoa. Numbers in parentheses above fecund fe- males are mean ova sizes for that month. REPRODUCTIVE BIOLOGY OF LIZARDS 39 from March to June, and again, slightly, in September and October (Fig. 25); these peaks although broadly overlapping those of E. cyanura, occur somewhat later than for the smaller species. Repro- ductively active females of E. nigra were found during each sampled month, reaching their lowest frequency (20%) in December. Spent females are at highest percentages in December, declining to zero percent in March (Fig. 26). Frequency of hatchlings, how- ever, increases from December to a peak in April, approximately three months later. Obviously, if eggs are deposited in December and January, and require from two to three months to hatch as previously suggested, hatchlings should be most abundant in March and April. Mean ovum sizes (in parenthesis above fecund frequencies in Fig. 26) show enlarging ova from December to January, decreasing in mean size in March and then markedly increasing again from April to May. These patterns generally precede and parallel those for increasing and decreasing frequencies of ovigerous females of E. nigra. Apparently eggs mature in December and January, are deposited, and hatch in May and April; a second developmental period occurs from March to May and eggs are again deposited 100 @q—_ Monthy % OQ----— bimonthly % REPRODUCTIVITY % MONTH Fic. 25.—Monthly and bimonthly percent of reproductively active (oviger- ous and fecund) females of Emoia nigra from islands of American Samoa. 40 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY (although perhaps not necessarily from the same females) in June and July. Reproductive activity in Emoia samoense exhibits much greater peaks and valleys than the species analyzed above (Fig. 27). While it may also be due to low sample sizes, assuming it is not, the pat- tern is markedly different. Maximal activity (using the bimonthly data) peaks in March and April, and again, more steeply, in De- cember. Reproductively active females were not found in May (and no specimens were collected during June due to scheduled trips during that month to islands where E. samoense was not present). Frequencies of spent females are highest from September to January and lowest from April to June (Fig. 28); eggs deposited in January presumably incubate for one to two months and would be expected to hatch in late February or early July. Hatchlings were found from December to July, and were most abundant in April. Percentages of ovigerous females drop sharply from October to January, but ova sizes (numbers in parenthesis above fecund frequencies, Fig. 28) increase from December to January and main- (7.6) 100 °. — 5 > \ (5.7) w 50 \ Ovarian oO co \ a \ a. \ s \ ss 7 \ 4 . \ (4.3) (es \/ \ @ spent \ a \ RQ . \ de IP NaS \ r SS / : Oviducal 10 va! “ehatchiing oO N D J F M A M J J MONTH Fic. 26.—Monthly percent of the total counts for hatchlings, spent females, fecund females (ovarian) and ovigerous (oviducal) females of Emoia nigra from islands of American Samoa. Numbers in parentheses above fecund fe- males are mean ova sizes for that month. REPRODUCTIVE BIOLOGY OF LIZARDS 4l tain constant sizes (3.2 to 3.5mm) through April. Ovigerous female frequencies increase in parallel with ova sizes from December to a peak in April. Although peak egg deposition clearly contrasts with low frequencies of ovigerous females, the patterns are not as well defined as those found for E. nigra and E. cyanura. The prolonged increase in ovigerous female frequencies from January to April, and the maintenance of a relatively high ovum size during the same time period, suggests that E. samoense may hold developing eggs longer and deposit them more slowly than either E. nigra or E. cyanura. Correlations Mean monthly climatic variables and monthly percents of re- productively active females of E. nigra, E. samoense and E. cyanura (Table 3) were analyzed to derive Pearson product-moment corre- lation coefficients (rs). As expected, certain climatic variables are highly correlated (Table 5). Mean monthly air temperature is posi- tively correlated with all but mean monthly wind speed (r = 0.97, p < .001), and significantly with mean monthly rainfall (r = .82, p < .01), mean monthly sky cover (r = .94, p < .001), and mean 100 @—_ monthly % O----— bimonthly % 50 % REPRODUCTIVITY MONTHS Fic. 27.—Monthly and bimonthly percent of reproductively active (oviger- ous and fecund) females of Emoia samoense from islands of American Samoa. 42 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY monthly relative humidity (0100 hrs) (r = .83, p < .01). Mean monthly relative humidity (0100 hrs) is also negatively correlated with mean monthly wind speed (r = -.64, p < .05), and signif- icantly with mean monthly sky cover (r = .91, p < .001) and du- ration of daylight (r = .82, p < .01). Mean monthly sky cover is negatively correlated with mean monthly: wind speed (-.80, p < .01), but positively with other variables and significantly with mean monthly relative humidity (0100 hrs) (r = .66, p < .05) and dura- tion of daylight (r = .70, p < .05). Mean monthly relative humid- ity (1300 hrs) is slightly negatively correlated only with mean monthly wind speed, but positively with mean monthly relative humidity (0100 hrs) (r = .67, p < .05). Mean monthly wind speed is negatively correlated with all variables and significantly with those mentioned above. These correlations characterize the maritime tropical climate of American Samoa in the Southern Hemisphere where there is only slight seasonality. Slight increases in temperature and rainfall dur- ing the months of November to April are accompanied by greater 100 spent@ (2.6) e 2 z : cs . eg @ ovarian ae ae ak B77 { Me ; PP } oviducal oN 7 a < i , ef MONTH Fic. 28.—Monthly percent of the total counts for hatchlings, spent females, fecund females (ovarian) and ovigerous (oviducal) females of Emoia samoense from islands of American Samoa. Numbers in parentheses above fecund fe- males are mean ova sizes for that month. 43 REPRODUCTIVE BIOLOGY OF LIZARDS 100 > d, ip Sd, co >d, SS COO EEE eee 00'T 8L'0 9¢°0- 29L'0- 660 00°0- 6e'0 9c'0- 8Pr'0- a OWVS 00'T 09°0- OO iam) FIO 7990 = Soi- 100 LL Oe YWOIN 00'T vo 0 69°0- 6c'0 v0'0- re'0 6o'0 sc’0 NVAOD 00'T 0€'0- 60°0 00°0- vOL'0 ao8 0 rS'0 qund 00'T 006'0- 8e°0- a08'0- e¥9'0- oF6'0- SMNW 00°T »L9°0 2990 rr'0 at8'0 NHYW 00'T rr'0 tom 0) cr'0 VHYWN 00°T 2160 oF6'0 OSWW 00°T acs'0 NYWN 00°T LVYNW a ee ee ee OWVS YOIN NVAO qYynNd SMWIWN NHYN VHYW OSWW NYNN LVYNW —ooooNeeeSqQqom————SSSSS ‘QSUBOWDS “FT = OINYS ‘Disiu “7 = YOIN “ounuvfo nowy = NVXAO ‘e FqGeZ, ur se oie SUONPIASIGGY “QLET 0} OG6T UoeoA\joq satreuUNs A[y}UOU WIOIZ poseI0Ae vyep SHCUO “BOUIS UROHOUTY WOIZ Soloeds piezi] 9a14} JO AWAT]Oe OATJONpoIder yusoIed pue So[qeUvA ONeUIT[O JO sUOTe[eLIOQN—'G AIAV I], Ad OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY sky cover, relative humidity, duration of daylight, and significantly reduced wind speeds. The Southeast Trades blow stronger from June to October when temperature and rainfall are reduced, ac- companied by generally shorter days, reduced sky cover, and lower relative humidity. The only significant correlations for Emoia nigra, E. samoense and E. cyanura with climatic variables (Table 5) are a positive one for E. nigra and mean monthly relative humidity (0100 hrs) (r = .66, p <-.05), and two negative correlations, E. samoense and dura- tion of daylight (r = —76, p < .05), and E. cyanura and mean monthly wind speed (r = -.69, p < .05). There are no significant correlations among the reproductively active periods of the three Emoia species, but the correlation between E. nigra and E. sa- moense is moderately high and positive (r = .78, 05 < p < 1). Most interesting is the moderately high, but negative correlation between the reproductive activities of E. cyanura and E. nigra (r= — 60:05 =< pr ol); » REPRODUCTIVE EFFORT Reproductive effort is defined as that proportion of the total energy budget of an organism that is devoted to reproductive proc- esses (Hirshfield and Tinkle, 1975). Three questions pertain to this concept: (1) What index actually measures reproductive effort? (2) What environmental conditions act as selective factors for high and low levels of effort? (3) How are those conditions predicted? Earlier workers (Gadgil and Bossert, 1970; Tinkle, 1969, and Tinkle, et al., 1970) used ratios of clutch to body weight as meas- ures of reproductive effort. Later workers (Ballinger and Clark, 1973; Tinkle and Hadley, 1973; Vitt, 1974; Vitt and Ohmart, 1975) emphasized the use of caloric ratios. However, Tinkle and Hadley (1975) and Hirshfield and Tinkle (1975) suggested that these ratios were inadequate, because the concept of reproductive effort, as defined by Fisher (1930), requires measurements that can be fitted to demographic models, not instantaneous values. Thus total en- ergy budgets, incorporating data on the proportional distribution of energy to growth, maintenance and reproduction, on an age-specific basis, are necessary to assess more accurately reproductive effort at the population level (Vitt, 1978). Caloric ratios facilitate the cal- culation of total energy budgets for species having well known demographic parameters (Vitt, 1978). These caloric ratios are pre- sented in Table 6 for four species of the genus Emoia coexisting in the same habitats on American Samoan islands, as a basis to which future demographic studies could be applied. For three of the four species with a clutch size of essentially two eggs, caloric values of relative clutch masses are significantly different (Table 6, p’s < .05). 45 REPRODUCTIVE BIOLOGY OF LIZARDS it ae Ls is aa a (163°-922' ) 9L6 (8S -1L3 ) ~=¥0 (63 “6ST ) €3 (ZT) O03 FE pinuvha “q (€S3'-€6T' ) ere (Ce U0) eta (9'08- €ST) TST (SI) ST + uSOMD] “FT (67'-E17 ) Iss ( €8- 8%) 62 (G'LZ- 10%) $33 Gl) eee yy, PSUBOULDS “FJ (6F'-60F' ) Hi (WR tots) as (G'9E- 9'€3) LSS (FE) Sig Sra DIsU DOW (S}luN] couepyuoD ¥c¢ ) x (asuey) x (esuey) x (asuey) x wu setoeds anjpa 9140]0) qys1am yoInD qys1am fpoq az1s YIM D appa J a. ee ‘eouleg UeolIoUry JO Spue[st Worf Drowy snuss ay} jo sajoads Bunsixeoo TOF (PQJGT BA) 3Y8IEm Apoq 0} yonJo zo sazeumTyso ouoreD ‘9 AIaVv], 46 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Vitt and Congdon (1978) proposed that relative clutch masses (in terms of the caloric conversions above) are predictable on the basis of body shapes of lizards as related to certain aspects of their ecology. They proposed that the two extremes of foraging and predator escape behaviors, sit and wait foragers versus wide for- agers (MacArthur and Pianka, 1966; Pianka, 1973; and Schoener, 1971), typified species with cryptic coloration and high relative clutch masses, and noncryptic coloration and low relative clutch masses, respectively. Species with contrasting combinations of be- haviors and coloration (or escape strategies) were termed inter- mediate. Schwaner (1979) described the ecology of the four coexisting species of Emoia in American Samoa, relative to resource partition- ing, behavior, and morphological correlations to microhabitat pref- erences. Emoia nigra is a widely foraging terrestrial species, highly visible by its black coloration, and most wary of all the species, as demonstrated by its flight at the slightest disturbance. Emoia samoense exhibits the other extreme; the species is arboreal, cryp- tically colored (green), is much less wary than E. nigra, and forages in a rather limited (horizontal) area on tree trunks and in low vegetation at the bases of trees.. Emoia lawesii is also cryptically colored, terrestrial (but confined to coral rubble substrates), and moves relatively slowly among the rocks when disturbed. Emoia cyanura is a small, widely foraging, semi-arboreal species, quick to escape when disturbed, with a bright blue tail which may be used as a decoy escape mechanism from predators; it would, therefore, correspond more closely to Vitt and Congdon’s (1978) intermediate strategy. When these ecological and behavioral traits are compared with the caloric ratios, the correlations are opposite of those described by Vitt and Congdon (1978) for the two extreme strategies. Emoia nigra the wide forager has the highest ratio, while E. samoense and E. lawesii, the cryptic, more sedentary species, have the lowest ratios. Emoia cyanura, with an intermediate strategy, has an inter- mediate caloric ratio between the two extreme groups. Thus most lizard species in American Samoa characteristically exhibit continuous reproductive activity with slight annual vari- ations (Sherbrooke, 1975; Duellman, 1978). In most species for which sufficient data are available, the general reproductive strategy appears to be multiple clutches annually, rather than the single clutches found in most temperate lizards (Tinkle, et al, 1970). Duellman (1978) stated that the reproductive strategies of lizards must be considered in terms of their phylogenetic limitations, struc- tural habitats, and interactions with other species. The basic as- sumption is that clutch mass, relative to body mass, determines the REPRODUCTIVE BIOLOGY OF LIZARDS 47 agility of a gravid female lizard, and thus, her survival (and the survival of her eggs) when threatened. In American Samoa, gekkonids exhibit a constant clutch size of two eggs, probably reflecting the inherent pattern exhibited throughout the family. Alternatively, or concommitantly, this pat- tern also may reflect the load-bearing limitations imposed by their arboreal habits, particularly for those species residing on building walls (Andrews and Rand, 1974). Schwaner (1979) suggested that the richest gekkonid communities on American Samoan islands are closely packed ecologically, perhaps at the point of limiting simi- larity (MacArthur and Levins, 1967). The reproductive modes of these species, involving low clutch size, continuous breeding activ- ity, multiple clutches annually, resistence of eggs to dissecation, prolonged incubation times, ability to affix eggs to surfaces (L. lugubris), and parthenogenetic forms (L. lugubris and, perhaps, C. pelagicus), are adaptations well suited to both colonization and maximal avoidance of competition (and resistence to predation). The presence of another parthenogenetic species, Cyrtodactylus pelagicus, in American Samoa would not be surprising; Lepido- dactylus lugubris (Cuellar and Kluge, 1972, is a known unisexual species on these islands, and two others, Gehyra variegata (Hall, 1970) and Hemidactylus garnotii (Kluge and Eckhardt, 1969) are found on nearby islands. Cuellar (1977) summarized the data con- cerning parthenogenetic species on islands and noted that enhance- ment of colonizing ability may be by prolonged sperm storage or parthenogenesis. Cuellar dismissed sperm storage as a strictly island phenomenon, since it apparently evolves on mainlands in response to other selective factors; however, prolonged incubation periods have been demonstrated for G. oceanica (see above), and parthenogenesis as another alternative seems ideal for American Samoan species. Their survival on these islands obviously depends in large part on how quickly their populations can respond to peri- odic catastrophy such as seawater innundations of lowland forests during tropical storms, volcanism, and human disturbance of low- land habitats. Populations of C. pelagicus on New Guinea and adjacent islands appear to have equal sex ratios (Herbert C. Dessauer, pers. comm.). However, if populations on American Samoa are parthenogenic, the possibility of finding a male is remote. Introduction of a male into an all female population might result in matings with disruption of clones and the production of triploid offspring: if some of these triploids are males, further inbreeding would eventually build chromosome complements so high that the whole population could eventually “crash” under the “weight” of its increasing chromosome number (Jay Cole, pers. comm. ). Parthenogenesis has been found in only about 1% of the known 48 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY lizard species. As a reproductive mode on islands, and in disclimax areas where few potentially competing species coexist, however, it appears a viable strategy which may be widespread in species on other islands in the Pacific region. Among scincids in American Samoa, small skinks exhibit a con- stant low clutch size of two eggs, but larger skinks have, maximally, two (E. lawesii and E. adspersa), three (E. nigra) and seven (E. samoense) eggs per clutch. Small skinks exhibit small clutches probably. due to the restrictions of small body size and the need for rapid escape (or protection of young in the case of L. noctua), as suggested by Duellman (1978) for small lizards in tropical South America. Although clutch sizes are low in small skinks, like the geckos, this may be compensated for by continuous breeding and multiple clutches annually. Competition for small food items perhaps limits the coexistence of these forms in the same microhabitats in American Samoa (Schwaner, 1979); but, interspecific competition would seem to have little effect on reproductive modes, clutch sizes, frequency of egg deposition, patterns in annual reproductive activity, and choice of nesting sites among species so widely divergent in microhabitats. However, predation by larger lizards, birds, the Pacific boa, Can- doia bibroni, and land crabs (feeding on eggs) may be sufficiently high to account for the differences in reproductive modes exhibited by small skinks. The secretive and viviporous habits of L. noctua probably pro- tect adults and young from predatory species. Tinkle and Gibbons (1977) reviewed the evidence supporting the evolution of viviparity in lizards. Their list of possible benefits for viviparity included: (1) protection of eggs from environmental mortality, (2) favorable thermoregulation for the developing embryos, (3) economy to the female in providing sustenance during development rather than making the entire reproductive commitment at ovulation, and (4) greater predictability in placement of newborn young in optimal sites at partuition. Not only are the secretive microhabitats of L. noctua often “foraged over” by other large and small scincid spe- cies (Schwaner, 1979), but also by the large gecko, G. oceanica,, which favors nesting sites under the bark of rotting coconut stumps. The reproductive mode of L. noctua seems, therefore, ideally suited for a microenvironment where nesting sites may be limited; its secretive habits provide maximal separation from most other com- peting or predatory species. Thermoregulation could be accom- plished simply by shifting to the sunny or shady sides of rotting stumps, under the bark, similar to the way Eumeces egregius in the Southeastern United States facilitates incubation of eggs in pocket- gopher mounds by shifting from lower to higher levels in burrows at different times of the day (Mount, 1963). Cryptoblepharis bou- REPRODUCTIVE BIOLOGY OF LIZARDS 49 tonii is perhaps protected by its isolation in littoral strand habitats not explored by the larger lizard species. But what of E. cyanura, which must forage and deposit its eggs in the same habitats with the larger, predatory species? Perhaps the staggered foraging times, as well as greater efficiency in handling smaller food items, intermediate foraging heights above ground, and, possibly, its tail coloration as a predator escape mechanism, contribute to the coexistence of E. cyanura with the larger species. The moderately high negative correlation between E. cyanura and E. nigra for their percentages of reproductively active females, could mean that E. cyanura has access to nesting sites also surveyed by E. nigra, but at different times of the year. Among the larger species, the significant correlation between reproductive activity and mean monthly relative humidity during the day seems related to the swelling of eggs during incubation. Swelling provides at least three functions (Cunningham and Huene, 1938): (1).actual use of water by the embryo, (2) protection from external pressure on the embryo by keeping the shell fully dis- tended, and (3) as a device for rupturing eggs (the latter function has been confirmed by these authors, and follows my own obser- vation for the disappearance of portions of the calcareous shell at which initial rupturing of the egg occurs). Otherwise, few climatic factors seem to affect the reproductive frequencies of these species. Like Ameiva ameiva in tropical South America, Emoia nigra may exhibit low clutch sizes due to the necessity of rapid move- ments (Simmons, 1975; Duellman, 1978); although there seem to be few predators capable of attacking an adult E. nigra on Amer- ican Samoan islands, the pattern may have evolved in other areas (e.g. New Guinea) where large predators are present. Alternatively, the low number of large eggs in clutches of E. nigra may facilitate the placement of large, early maturing hatchlings capable of suc- cessfully competing with other small lizards (Duellman, 1978), or avoiding predation by their cannibalistic parents (Schwaner, 1979). The large clutch size of E. samoense markedly contrasts with that of other large skinks on American Samoan islands. Perhaps like some arboreal species in tropical Asia (i.e., Draco; Inger and Greenberg, 1966) and South America (i.e., Enyalioides laticeps; Duellman, 1978), E. samoense is sufficiently isolated by microhabi- tat that its reproductive habits are generally not affected by the interspecific interactions that apparently limit clutch size in other, more terrestrial skinks. Thus, the positive correlation between monthly reproductive frequencies of E. nigra and E. samoense does not necessarily imply overlap in nesting sites. The predation by E. nigra on E. samoense eggs noted above may be an exceptional case, taking place in mangrove forest where nesting sites are obviously more limited for 50 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY the terrestrial species. Alternatively, if E. samoense eggs are laid in arboreal nests, these sites may be sensitive to adverse ambient weather conditions, thus, limiting their favorability to the develop- ing eggs. The eggs of E. samoense must also absorb water. This necessity may be indicated by the negative correlation between reproductive frequencies and duration of daylight, because months with shorter days are also highly correlated with maximal rainfall and low wind speed. This could assure that such potential nesting sites as epiphytic birdnest ferns are saturated with water and less prone to drying by wind. Perhaps the apparent retention, or pro- longed development, of ova by E. samoense is also related to the limitations of available nesting sites in arboreal habitats. If the eggs are laid singly (which must be determined by further study) more time would be required in finding suitable nests. The lengthy incubation period of E. lawesii eggs may simply reflect the cooler environments in which the eggs are placed, as- suming that these nests are under the layers of coral plates to which the species is apparently restricted. The markedly lower active foraging temperatures of E. lawesii than for all other scincid species in American Samoa (Schwaner, 1979) may be related to egg de- velopment at these presumably low temperatures. The general conclusion for most American Samoan lizards seems to suggest that their reproductive strategies reflect adaptations more for avoidance of interspecific interactions than for climatic variables which greatly influence other species in seasonal temperate and equatorial tropical areas. ACKNOWLEDGEMENTS I thank Environment Consultants, Inc., The United States Fish and Wildlife Service, and the Government of American Samoa for the opportunity to gather research data in American Samoa during 1976. Special thanks go to my field companions, A. Binion Amer- son, Jr., Arthur Whistler, Warren Pulich, and Kelly Irwin. For loans of specimens in their care I am indebted to William E. Duellman, University of Kansas Museum of Natural History, Richard Zweifel, American Museum of Natural History, George R. Zug, United States Natural Museum, Alan Leviton, California Academy of Sciences, Hymen Marx, Field Museum of Natural His- tory, Ernest Williams, Harvard Museum of Comparative Zoology, and Alan E. Greer, Harold Cogger and Paul Webber, Australian Museum. I am indebted to Harvey Lillywhite and the University of Kansas Animal Care Committee for providing space to house my animals during the course of this study. William E. Duellman, Michael S. Gaines and Peter D. Ashlock read the manuscript and offered suggestions for its improvement. I am particularly grateful REPRODUCTIVE BIOLOGY OF LIZARDS 51 to William E. Duellman for his close editing of this and other manuscripts, and to Linda Trueb for advice on the figures. My wife, Lila, typed all drafts and the final manuscript. With- out her support and encouragement throughout the long months of field work and writing, this study could not have been completed. LITERATURE CITED Amerson, A. B., Jr., A. C. WuHIsTLER, T. D. ScowaNner. 1978. An inventory of the wildlife and wildlife habitat of the islands of American Samoa. Vols. I, II, IJ, 813 pp. Environment Consultants, Inc., 14325 Proton Road, Dallas, Texas 75240 (unpublished ). AnpreEws, R., A. S. RAND. 1974. Reproductive effort in anoline lizards. Ecol- ogy, 55: 1317-1327. Baker, J. R. 1947. The seasons in a tropical rain-forest. Part 6. Lizards (Emoia). Jour. Linn. Soc. London, 41 (no. 279): 243-247. Bauuincer, R. E., D. R. Cuarx. 1973. Energy content of lizards eggs and the measurement of reproductive effort. J. Herpetol. 7: 129-132. Brown, K. M., O. J. Sexton. 1973. Stimulation of reproductive activity of female Anolis sagrei by moisture. Physiol. Zool. 46: 168-172. Brown, W. C. 1956. The distribution of terrestrial reptiles in the islands of the Pacific Basin. Proceedings of the Eighth Pacific Science Congress, 3A: 1479-1491. Brown, W. C., A. C. Atcata. 1957. Viability of lizard eggs exposed to sea water. Copeia, 1957, (1): 39-41. Cac.ez, F. R. 1946. A lizard population on Tinian. Copeia, 1946 (1): 4-9. Cuurcu, G. 1962.. The reproductive cycles of the Javanese house geckos, Cosymbotus platyurun, Hemidactylus frenatus, and Peropus mutilatus. Copeia, 1962 (2): 262-269. CoLweELL, R. K. 1974. Predictability, constancy, and contingency of periodic phenomena. Ecology 55(5): 1148-1153. Crump, M. L. 1974. Reproductive strategies in a tropical anuran community. Misc. Publ. Mus. Nat. Hist. Univ. Kansas, (61): 1-68. CuEtuar, O. 1977. Animal parthenogenesis. Science, 197: 836-843. CuEeuuar, O., A. G. Kiuce. 1972. Natural parthenogenesis in the gekkonid lizard Lepidodactylus lugubris. Jour. Genet. 61(1): 14-28. CuNNINGHAM, B., E. Huene. 1938. Further studies on water absorption by reptile eggs. Am. Nat. 72: 380-385. DvuELLMAN, W. E. 1978. The biology of an equatorial herpetofauna in Ama- zonian Ecuador. Misc. Publ. Mus. Nat. Hist. Univ. Kansas, (65): 1-352. FisHer, R. A. 1930. The genetical theory of natural selection. Am. Natur. 104: 1-24. Fircu, H. S. 1954. Life history and ecology of the five-lined skink, Eumeces fasciatus. Univ. Kansas Publ. Mus. Nat. Hist., 8 (1): 1-156. Fircu, H. S. 1956. An ecological study of the collared lizard (Crotaphytus collaris). Univ. Kansas Publ. Mus. Nat. Hist., 8 (3); 213-274. Fircu, H. S. 1970. Reproductive cycles in lizards and snakes. Misc. Publ. Mus. Nat. Hist. Univ. Kansas, (52): 1-247. Fircu, H. S. 1972. A field study of Costa Rican lizards. Univ. Kansas Sci. Bull., 50 (22): 39-126. Fircu, H. S., H. W. Greene. 1965. Breeding cycle in the ground skink, Lygosoma laterale. Univ. Kansas Publ. Mus. Hist., 15 (11): 565-575. Fucuer, C. M. 1966. Lepidodactylus lugubris Dumeril and Bibron in Western South America. Jour. Ohio Herp. Soc., 5: 162. 52 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Fuxapa, H. 1965. Breeding habits of some Japanese reptiles (critical review). Bull. Kyoto Gakugei Univ., Ser. B., (27): 65-82. Gaver, M., W. H. Bossert. 1970. Life historical consequences of natural selection. Amer. Natur., 104: 1-24. Gorman, G. C., P. Licur. 1974. Seasonality in ovarian cycles among tropical Anolis lizards. Ecology, 55: 360-369. Greer, A. E. 1968. Clutch size in the scincid genus Emoia. Copeia, 1968 (2): 417-418. Haake, W. D. 1977. The snake-eyed skink. African Wildlife, 31 (1): 30-31. Hatt, W. P. 1970. Three probable cases of parthenogenesis in lizards (Aga- midae, Chameleontidae, Gekkonidae). Experimentia, 26: 1271-1273. Hepicer, H. 1934. Beitrag zur herpetologic und zoogeographic Neu Britan- niens und einiger umliegender Gebiete. Zool. Jahrb. Bd. 65, Heft 516, pp. 441-582. HimsHFrEcp, M. F., D. W. Tinxxte. 1975. Natural selection and the evolution of reproductive effort. Proc. Nat. Acad. Sci. USA, 72: 2227-2231. INcER, R. F., B. GREENBERG. 1966. Ecological and competitive relations among three species of frogs (Genus Rana). Ecol. 47 (5): 746-759. Late summer. Kuiuce, A. G., M. J. Ecxarpr. 1969. Hemidactylus garnotti Dumeril and Bibron, a triploid all-female species of gekkonid lizard. Copeia 1969 (4): 651-664. LovermncE, A. 1945. Reptiles of the Pacific World. The Macmillan Co., N.Y., xii + 259 pp. MacArrtuor, R. H., R. Levins. 1967. The limiting similarity, convergence and divergence of coexisting species. Am. Nat., 101: 377-385. MacArtuor, R. H., E. R. Pranxa. 1966. On optimal use of a patchy environ- ment. Amer. Natur., 100: 603-609. Mount, R. H. 1963. Natural history of the red-tailed skink, Eumeces egregius Baird. Amer. Midl. Nat., 70 (2): 356-385. Ourver, J. A., C. E. SHaw. 1953. The amphibians and reptiles of the Hawaiian Islands. Zoologica, New York, 38: 65-95. Pranka, E. R. 1973. The structure of lizard communities. Annual Review of Ecology and Systematics, 4 (4052): 53-74. Prank, E. R., H. D. Pranxa. 1976. Comparative ecology of twelve species of nocturnal lizards (Gekkonidae) in the Western Australian Desert. Copeia, 1976 (1): 125-142. Ruma, R., R. Pamsmosran, J. L. Apxins. 1972. Reproductive cycle and growth in the lizard Anolis acutus, Copeia, 1972: 509-518. ScHoENER, T. W. 1971. Theory of feeding strategies. Annu. Rev. Ecol. Syst. 4; 53-74. Scuwaner, T. D. 1979. Biogeography, community ecology and reproductive biology of the herpetofauna of the American Samoan Islands. Unpub- lished Ph.D. dissertation, University of Kansas, 1979: 1-301. Sexton, O. J., E. P. Ontriep, L. M. Haruaway, L. M. BaAuwincer, R. E. Licut, P. Licur. 1971. Reproductive cycles of three species of anoline lizards from Isthmus of Panama. Ecology, 52: 201-215. Simmons, J. E. 1975. The female reproductive cycle of the teiid lizard Ameiva ameiva petersi Cope. Herpetologica, 31: 279-282. Smiru, H. M., C. Grant. 1961. The moming gecko in the Americas. Herpe- tologica 17: 68. Smitu, M. A. 1935. The fauna of British India. Reptilia and Amphibia. Vol. II — Sauria, xiii +- 583 pp. SoxaL, R. R., J. F. Rotr. 1969. Biometry. W. H. Freeman and Company, San Francisco: 1-776. REPRODUCTIVE BIOLOGY OF LIZARDS 53 SHERBROOKE, W. C. 1975. Reproductive cycle of a tropical lizard, Neusticurus ecpleopus Cope, in Peru. Biotropica, 7: 194-207. TANNER, W. W. 1957. A taxonomic and ecological study of the western skink (Eumeces skiltonianus). Great Basin Nat., 17 (3-4): 59-94. TinxLE, D. W. 1961. Population structure and reproduction in the lizard Uta stansburiana steinegeri. Amer. Midl. Nat., 66 (1): 206-234. TinkLE, D. W. 1969. The concept of reproductive effort and its relation to the evolution of life histories of lizards. Amer. Natur., 103: 501-516. TinxLE, D. W. 1967. The life and demography of the side-blotched lizard. Misc. Publ. Mus. Zool. Univ. Michigan, 132: 1-182. TINKLE, D. W., J. W. Grssons. 1977. The distribution and evolution of vivi- parity in reptiles. Misc. Pubs. Mus. Zool. Univ. Michigan, (154): 1-55. TINKLE, D. W., N. F. Haptey. 1973. Reproductive effort and winter activity in the viviparous montane lizard Scoloporus jarrovi. Copeia 1973: 272- PHI TinkxLe, D. W., H. M. Witpur, S. G. Tmtey. 1970. Evolutionary strategies in lizard reproduction. Evol. 24: 55-74 (Mar. 1970). U.S. DEPARTMENT OF ComMMeERCE. 1956-1976. Local climatological data. Monthly and annual summary with comparative data. Pago Pago, American Samoa. National Climatic Center, Asheville, N.C. Virt, L. J. 1974. Caloric content of lizard and snake (Reptilia) eggs and bodies and the conversion of weight to caloric data. J. Herpetol. 12: 65-72. Virt, L. J. 1974. Reproductive and energy comparisons of adults, eggs and neonates of Gerrhonotus coeruleus principis. J. Herpetol. 8: 165-168. Virt, L. J., J. D. Concpon. 1978. Body shape, reproductive effort, and rela- tive clutch mass in lizards: resolution of a paradox. Amer. Nat., 112: 595-608. Virt, L. J., R. D. Onmarr. 1975. Ecology, reproduction and reproductive effort of the iguanid lizard Urosaurus gracious on the lower Colorado River. Herpetologica, 31: 56-65. Wiser, H. M., D. W. Tink, J. P. Cotztivs. 1974. Environmental certainty, trophic level, and resource availability in life history evolution. Amer. Nat., 108: 805-817. | 2) See Se | aa aoe. all 7 ot i i I mn i ’ es Ss por Gr OSA i foxen ti TH 7 liye hy * i 7 J . 4 wa! oh f ia | ey eae Ses ‘ bowtie . ae FT 7 ‘ee rey Yipee i ¥ I Wa b » hy wage F ; | pele henge | ie del) hc” toe hae : . oF = Pin Pi deed Se ee fi shh lat ks a. oe nD ey ey ’ wnt ‘ih w= Xie i ¢ aE) fin i ioe " th es it ay a a i ye \% 1 ¢ i 4 al i nL. “nt iaoln' Iain th eee ‘gy pee " ees ee ‘ 1a gt A= ; a UNIVERSITY OF KANSAS PUBLICATIONS MUSEUM OF NATURAL. HISTORY The University of Kansas Publications, Museum of Natural History, beginning with volume 1 in 1946, was discontinued with volume 20 in 1971. Shorter research papers formerly pub- lished in the above series are now published as Occasional Papers, Museum of Natural History. The Miscellaneous Publica- tions, Museum of Natural History, began with number 1 in 1946. Longer research papers are published in that series. Monographs of the Museum of Natural History were initiated in 1970. All manuscripts are subject to critical review by intra- and extramural specialists; final acceptance is at the discretion of the publications committee. Institutional libraries interested in exchanging publications may obtain the Occasional Papers and Miscellaneous Publica- tions by addressing the Exchange Librarian, The University of Kansas Library, Lawrence, Kansas 66045. Individuals may pur- chase separate numbers of all series. Prices may be obtained upon request addressed to Publications Secretary, Museum of Natural History, The University of Kansas, Lawrence, Kansas 66045. Editor: Ep Witry Managing Editor: JoserH T. CoLuins PRINTED BY UNIVERSITY OF KANSAS PRINTING SERVICE LAWRENCE, KANSAS ON) 2/29 Mus oa OCCASIONAL PAPERS OCT $8 O 1980 of the ikem ty MUSEUM OF NATURAL HISTORY The University of Kansas Lawrence, Kansas NUMBER 87, PAGES 1-40 OCTOBER 17, 1980 A REVIEW OF THE PHYLLOMEDUSA BUCKLEYI GROUP (ANURA: HYLIDAE) By Davin C. CANNATELLA* The leaf-frogs of the subfamily Phyllomedusinae are among the most bizarre residents of the Neotropics. Currently, the group con- sists of three genera—Agalychnis, Pachymedusa, and Phyllomedusa (Duellman, 1968). Members of the genus Agalychnis inhabit humid lowland and montane forests of Middle America and northwestern South America. This genus includes eight species, most with bril- liant flash colors and well-developed interdigital webbing. The monotypic Pachymedusa has a robust body and occurs in the dry lowlands of western México. The genus Phyllomedusa—an unnat- ural assemblage of about 30 species—is primarily South American; only two species are known from Central America (see Duellman, 1977, for list of species). The more specialized species of Phyllo- medusa have highly modified, grasping feet, and demonstrate a walking, rather than the leaping gait characteristic of Agalychnis. Possibly one extreme in morphology among phyllomedusines is exemplified in Phyllomedusa sauvagei, a plain, xeric-adapted spe- cies with enormous parotoid glands and specialized grasping feet with no webbing. In contrast, Agalychnis callidryas, occurring in humid tropical lowlands, has gaudy flash colors and extensively webbed hands and feet. Between these extremes is a group of rela- tively generalized species inhabiting northwestern South America and southern Central America—the Phyllomedusa buckleyi group. The purposes of this paper are 1) to define the P. buckleyi group, 2) to diagnose the species, and 3) to present new data and to review the available data on the biology of the species. 1 Division of Herpetology, Museum of Natural History, The University of Kansas, Lawrence, Kansas 66045. 2 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY ACKNOWLEDGMENTS For the loan of specimens I am indebted to Douglas C. Cox, W. Ronald Heyer, Arnold G. Kluge, Alan E. Leviton, Hymen Marx, Charles W. Myers, Jay M. Savage, and Ernest E. Williams. This study was completed under the careful. direction of William E. Duellman, who made available facilities in the Museum of Natural History at The University of Kansas. Discussions with William E. Duellman, Linda Trueb, and Edward O. Wiley were enlightening; these persons and William L. Bloom critically read the manuscript. Charles W. Myers kindly provided transparencies and made avail- able his field notes. Linda Trueb was invaluable in assisting with the illustrations and osteological studies. John D. Lynch and Marsha C. Lynch aided in the collection of specimens in the field. Leslie Jakeway suggested the name for the new species. Compu- tations were done at the Academic Computer Center of The Uni- versity of Kansas on the Honeywell 66/60 computer. Field work in Ecuador was supported by a grant (DEB 76-09986) from the National Science Foundation (William E. Duellman, principal in- vestigator). My research was supported by a Graduate Fellowship from the National Science Foundation. MATERIALS AND METHODS The systematic studies are based on the examination of 262 preserved frogs, 21 skeletal preparations, 15 lots of tadpoles, two clutches of eggs, and 15 radiographs. Recordings of calls were analyzed with a Vibralyzer (Kay Electric Company). Calls were analyzed and measurements of external morphological characters were taken in the manner described by Duellman (1970). Osteo- logical observations and terminology were based on Trueb (1973, 1977). Statistics were computed in part with the aid of BMDP programs (Dixon, 1975). Webbing formulae were described in the manner of Savage and Heyer (1967), and tadpoles and eggs were staged according to Gosner (1960). Muscle dissections were stained as described by Bock and Shear (1972). Illustrations were executed by means of a Wild M-8 microscope with a camera lucida attach- ment. Ecological data and color notes were taken from the field notes of William E. Duellman and the author; these are filed in the Museum of Natural History at The University of Kansas. Through- out the text snout-vent length is abbreviated as SVL. Specimens are referred to by the following abbreviations: AMNH American Museum of Natural History BMNH British Museum (Natural History ) BYU Brigham Young University CAS-SU —_ California Academy of Sciences EBRG Estacién Biolégica Rancho Grande PHYLLOMEDUSA BUCKLEYI GROUP (ANURA: HYLIDAE) 3 FMNH Field Museum of Natural History ICN Instituto de Ciencias Naturales, Bogota KU The University of Kansas Museum of Natural History MCZ Museum of Comparative Zoology NHRM Naturhistoriska Riksmuseet, Stockholm UMMZ The University of Michigan Museum of Zoology USC-CRE University of Southern California USNM National Museum of Natural History SYSTEMATICS The Phyllomedusa buckleyi Group Definition —1) Sexual dimorphism in size; SVL to 44.5 mm in males, 54.7 mm in females; 2) hands and feet less than one-fourth webbed; 3) calcars present or not; 4) parotoid gland not differen- tiated; 5) white dorsal warts present or not; 6) males having thin horny nuptial excrescence; 7) males having single, median subgular vocal sac; 8) first toe shorter than second; 9) palpebrum unpig- mented; 10) dorsum uniformly green by day with reddish brown flecks, lacking pattern; 11) flanks bright orange in life, lacking spots or pattern; 12) iris creamy white, with silver to bronze cast; 13) prevomerine teeth present in some individuals of all species, situ- ated on posteromedially directed dentigerous processes; 14) fron- toparietal fontanelle exposed as a large oval; 15) quadratojugal present or not; 16) sacral diapophyses widely expanded, with con- vex edges; 17) posterior portion (pars scapularis) of depressor man- dibulae absent; 18) nektonic tadpoles having oval bodies and mod- erately deep tail fins; 19) mouths of larvae anteroventral, lacking labial papillae anteromedially; 20) tadpoles having a denticle formula of 2/3. Content.—Four species: Phyllomedusa buckleyi Boulenger, 1882; P. lemur Boulenger, 1882; P. medinai Funkhouser, 1962; P. psilopygion new species. Distribution—The combined distributions of the four species include the Cordillera de la Costa in Venezuela, the Caribbean and Pacific slopes of highlands in Panama and Costa Rica, the Chocd region of Colombia, and the Amazonian slopes of Ecuador, with an altitudinal range of 100 to 1870 m (Figs. 1 and 2). Remarks.—The frogs of the P. buckleyi group most closely re- semble some of the small generalized Phyllomedusa inhabiting the coastal ranges of southeastern Brasil—aspera, cochranae, fimbriata, guttata, and marginata. Phyllomedusa aspera is too poorly known to warrant further comparison with the buckleyi group. According to Izecksohn and da Cruz (1976), fimbriata and marginata form a natural group; P. guttata and cochranae form another species-pair on the basis of their specialized larvae (Bokermann, 1966). These OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 4 “poder ainjesey![ syuaserdax epomo uedQ ‘snwa] vsnpawopyfiyg Jo UoNNqIasIq—T “SIT SY¥3L3WO1, —_————— 002 OSI OO! O O ‘W OOO! YSAO0 f=4 YNOLNOD ‘W OOb PHYLLOMEDUSA BUCKLEYI GROUP (ANURA: HYLIDAE) 5 Fic. 2.—Distribution of Phyllomedusa buckleyi (squares), P. psilopygion (circles), and P. medinai (triangle). stream-adapted larvae, as well as the spotted pattern on the flanks of the adults, separate the guttata group from the buckleyi group. The fimbriata group (fimbriata and marginata) can be distinguished by the presence of a long, pointed calcar and tadpoles with labial papillae present anteromedially (Lutz and Lutz, 1939); in addition, the feet of fimbriata are about one-half webbed. Phyllomedusa marginata differs further from the buckleyi complex by a bicolored iris and gray to beige flanks (Izecksohn and da Cruz, 1976). Duellman (1973) referred Phyllomedusa perinesos to the buck- 6 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY leyi group. His conclusions were based on his examination of the only known specimen, the holotype. Material collected subse- quently has yielded the following data on perinesos that exclude it from the buckleyi group: (1) parotoid gland differentiated; (2) posterior portion (pars scapularis) of the depressor mandibulae pres- ent; (3) first toe equal in length to second. These three characters alone ally perinesos to the more specialized species of Phyllomedusa with grasping feet, such as tarsius and tomopterna. In members of the buckleyi group the first toe is distinctly shorter than the second; as it is in Agalychnis and Pachymedusa. In skeletal preparations of species of the buckleyi group, Metatarsal I is distinctly shorter than Metatarsal II. In species with specialized grasping feet, such as perinesos, tarsius, rohdei, and sauvagei, Metatarsal I is equal to, or greater in length than Metatarsal II. In these species the first toe is superficially equal to, or greater in length than the second toe; this condition obviously is derived. Data from Anderson (1978) demonstrate correlative differences in myology of the foot. Thus, on the basis of foot structure, as well as the other characters mentioned above, perinesos is excluded from the buckleyi group. Duellman (1970) remarked that possibly lemur and medinai are conspecific. The various lines of evidence discussed below do not support this suggestion. Duellman (1968, 1969) suggested that the buckleyi group should be accorded generic status. From my investigations of phyllo- medusines, it is clear that the buckleyi group does not belong in the genus Phyllomedusa; however, the proper generic allocation of these unspecialized frogs must await the completion of studies on other phyllomedusines now in progress. Within the buckleyi group, buckleyi and medinai are phenet- ically most similar; lemur and psilopygion closely resemble each other. As will be seen in the following sections, these similarities are based on superficial similarity and are not necessarily indicative of relationship. ANALYSIS OF CHARACTERS External Morphology Size and Proportions.—Pertinent measurements and ratios are given in Tables 1 and 2. The species of the P. buckleyi group are moderate-sized frogs. Phyllomedusa buckleyi is the largest (SVL of females to 54.7 mm); P. lemur is the smallest—females from Tapanti, Costa Rica attain a SVL of 41.6 mm. Phyllomedusa lemur also exhibits a trend for increase in size from west to east (Table 3). The largest specimen of lemur examined is a female having a SVL of 52.7 mm. For each ratio in Table 2, a Kruskal-Wallis analysis of variance 7 HYLIDAE) ee SSS —SSSs$09((w_—_—a_—_— 00 —Sm (1'S8-€'3E ) (9°L€-Z'98 ) (T'EE-T'Se ) (L'0S-0' LF ) SPE 698 VCE 8'6P Qo a eee uoizhidojisd -g ( L'VE-0'E8 ) (3: 9€-0'S8 ) (0'SE-9'FE ) (3'8t-8'9F) OFE LSet BPE eLy = nm = wuipau “gq (0%) (¥#GI-O'ZT) aera) (Sse) (GIZ-L1Z) (9°I#-S'6E) 0% Gio SEL 8'SI S'1Z 90F & re) (Gil cul) (0'II-@'6) (9'LI-8'6) (0°SI-8'6) (F8I-LFI) (LF8-T08) ISTO OL OLSOFTOL FLYOF SOL LOPOFTIT LESOFLOT OCPIFHCE SG Po inwa] ‘d (G'Z-0'S) (8'LI-T'9T ) (102-6 LT ) (#'S-8'LT) (9°83-G'Ss ) (L'PS-0'0S ) ZOLOF ES GILOFOLI SSLOFSSI YOST FEEI Bs0EFSIS BLOGT HOS L } (rst) (tsi) (CHESS) — (OS 2a) (SZS-L'6I) (G'Fr-e'6E) QLOEGT GLSOFOP O880F FSI OLFOFOST OFLOFTIG C6CT+OC (16 2 wiayyong “d ANAL GCIMH NATH LOOA aLL TAS u_——-xag setoedg ‘WIUI Ul ‘UOT}VIANp plepue}s 9UO pue UBELU A\O[Eq sosoyjueied Ul UaAIS oOsUPI PaAlEsqQ “1o}JCUIvIpP wmnueduit} ‘qWAL {yi3ue] eIqn “@IL ‘yysue, yueA-ynous “TAS ‘pM prey ‘dIMH ‘yysuey peoy ‘NATH ‘yisue] 00F “LOOA ‘suonereirqqy ‘dnois 1fajyong psnpawozjfiyg eu} 50 sotoeds jo syusuloinseaJy— | A1dV —————— EE 8 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY (Conover, 1971) was run among the species (males only). For each ratio the null hypothesis—that a ratio was the same in all species— was rejected (P < 0.001). A multiple range test for ranked data (Dunn, 1964) was performed next for each ratio (Table 4). Coloration —The group is characterized by a uniformly green dorsum (by day) with bright orange concealed surfaces of the limbs and flanks, bearing no spots or pattern. The coloration alone sets the group apart from other phyllomedusines. By night the dorsum is brown, as in some Agalychnis. The venter is creamy white with no markings. The iris is off-white with a silver to bronze cast. A clear palpebrum distinguishes the group from Pachymedusa and Agalychnis (except calcarifer and craspedopus). Some individuals of medinai, psilopygion, and buckleyi bear dorsal warts; the black outer border seen on the warts of some Agalychnis is lacking. Phyl- lomedusa lemur has no dorsal warts. There is no ontogenetic change in color in those species for which the young are known. Skin.—The dorsal skin is smooth, with a slight shagreen, in all species; the belly skin is granular. Some specimens of buckleyi develop faint tubercles on the hind limbs and loreal region; P. psilopygion and buckleyi bear small calcars. The grossly thickened and well-differentiated parotoid glands seen in more specialized Phyllomedusa are absent from this group, as they are from Pachy- medusa and all Agalychnis. Hands and Feet.—The hands and feet are less than one-fourth webbed (Table 5). Phyllomedusa medinai has the greatest amount of webbing (Fig. 3); buckleyi has slightly less (see Duellman, 1969, for illustration). Phyllomedusa lemur bears only traces of webbing, (see Duellman, 1970, for illustration) as does psilopygion (Fig. 3). The slightly enlarged prepollex bears a thin horny nuptial excres- cence in breeding males. A small, round outer metatarsal tubercle is sometimes present in buckleyi and medinai, but absent in the other species. A small, oval inner metatarsal tubercle is uniformly present. The first toe is shorter than the second. The discs on the fingers and toes are relatively larger in medinai and buckleyi than in lemur. Osteology Cranium.—The skulls are moderately broad and flat; they lack exostosis, co-ossification, and casquing (Figs. 4 and 5). Phyllo- medusa psilopygion exhibits the weakest overall ossification. In psilopygion and medinai the skull is slightly wider than it is long; in buckleyi and lemur the width is about the same as the length. Each skull is about one-third as high as long. In dorsal view, the snout is acutely rounded—that portion of the maxilla anterior to the palatines is almost straight. The skull of psilopygion is the largest in relative size. PHYLLOMEDUSA BUCKLEYI GROUP (ANURA: HYLIDAE) Fic. 3.—Top: Hand and foot of Phyllomedusa medinai, KU 167187, fe- male. Bottom: Hand and foot of P. psilopygion, KU 169612, female. Line equals 2 cm. 10 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY The nasals are moderately large and narrowly separated medi- ally. In lemur they are the largest and contact the ossified portion of the sphenethmoid in larger specimens. The nasals are slightly smaller in psilopygion; the maxillary processes are narrower than those of lemur. The nasals are smallest in buckleyi and medinai, and are notched on the posterior edge. The ossified portion of the sphenethmoid is shallow in the group; it is most poorly developed in psilopygion. Because the frontoparietals are slender, the exposed frontoparietal fontanelle is a large oval. The frontoparietals diverge slightly in medinai, result- ing in a greater relative interorbital width in that species. The pos- terior borders of the frontoparietals are obscured by the irregular ossification patterns on the exoccipital. The exoccipital is poorly ossified in psilopygion; it is slightly better developed in the other species. In all species the squamosal articulates with the distal portion of the crista parotica that is cartilaginous. The otic rami of the squamosals are of comparable length in all species. The zygo- matic ramus is shortest in psilopygion; in the other species the ramus is longer, extending about one-third the distance to the max- illary. The articulation of the squamosal and pterygoid is at approx- imately the level of the occipital condyles. Phyllomedusa medinai lacks a quadratojugal (Fig. 5); it is pres- ent but poorly ossified in the others. In buckleyi and lemur the quadratojugal may or may not overlap the maxilla. The maxillae are slender and lack postorbital processes; preorbital processes are uniformly present, and do not contact the maxillary processes of the nasals. The pars facialis is very shallow in psilopygion, deepest in lemur, and moderately developed in buckleyi and medinai. The pars palatina is narrow in all species, being alittle wider than the pars dentalis. Maxillary and premaxillary tooth counts are as follows (n = number of elements, mean in parentheses): buckleyi, 88-95 per maxilla (n = 4, 93) and 13-16 per premaxilla (n = 4, 13); lemur, 69-96 (n = 6, 79) and 11-15 (n = 6, 13); medinai, 100-102 (n = 2, 101) and 14-15 (n = 2, 15); psilopygion, 86-93 (n = 2, 90) and 12-12 (n = 2, 12). The great range in maxillary teeth in lemur re- flects the range of maxillary lengths; there is a direct relationship between the number of teeth and the length of the maxilla. The septomaxillaries are U-shaped. The alary processes of the premaxillaries are directed dorsally in lemur and psilopygion, and posterodorsally in buckleyi and medinai. In all species the length of the alary processes is about the same as the length of the pre- maxillary tooth row. The processes are notched in all species; they diverge slightly in buckleyi and medinai. The pars palatina is mod- erately developed and widens distally in the group. The prominent palatine processes are about one-half the length of the premaxillary tooth row. PHYLLOMEDUSA BUCKLEYI GROUP (ANURA: HYLIDAE) 11 The pterygoids are uniformly triradiate and weakly developed. The anterior rami terminate one-third of the orbital length from the anterior borders. The medial rami articulate synchondrotically with the anterolateral corners of the otic capsules. The parasphenoids are T-shaped, lacking ridges or odontoids. The alae are acuminate and posterolaterally directed. The cultriform processes are about twice the width of the alae and bear slightly notched termini. In medinai the cultriform process reaches almost to the level of the palatines; in the others it is slightly shorter and barely overlaps the ossified portion of the sphenethmoid. In ventral view the ossifi- cation of the sphenethmoid in all species extends to the midlevel of the orbit. The slender palatines are widened and flattened distally. Phyl- lomedusa lemur possesses thin, anteromedially directed processes that originate from the anterior border of the distal ends of the palatines. These peculiar processes appear to reinforce the antero- lateral margins of the internal nares (Fig. 5). In all species, the palatines are separated medially by a distance equal to the width of the sphenethmoid. The prevomers are small, with thin lateral processes forming the anterior and medial borders of the choanae. The dentigerous proc- esses are short and oriented posteromedially (\7); they are almost perpendicular to the midline in medinai. All specimens of lemur from localities west of Moravia de Tur- rialba, Costa Rica, lack dentigerous processes and prevomerine teeth. Specimens from Moravia de Turrialba are variable; den- tigerous processes and/or teeth are sometimes present. Specimens from localities east of Moravia de Turrialba have prevomerine teeth more often than not. The presence of prevomerine teeth and den- tigerous processes in lemur is correlated roughly with the SVL of the specimen, for there exists a trend for increase in SVL from west to east (Table 3). In buckleyi, medinai, and psilopygion, dentiger- ous processes are invariably present; teeth may be absent. The prevomerine tooth counts for the species are as follows (range, mean in parentheses, n — number of elements): buckleyi, 0-6 (4, n = 56); medinai, 0-4 (2, n = 14); lemur, 0-5 (2, n = 30); psilopygion, 0-5 (2, n = 14). The presence of prevomerine teeth in medinai and lemur has not been reported previously by other authors (Funk- houser, 1962; Duellman, 1969). The occipital condyles are widely separated and stalked. Colu- mellae are slender, and present in all species. The mandibles lack ridges or odontoids. The dentaries and angulars are moderately developed; a small coronoid process is present. The mentomeckelian elements are poorly ossified. Vertebral Column.—Eight presacral vertebrae are uniformly present. The cervical cotyles are widely separated in all species. OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 12 OS OOT OOT Oot 0S 99 GG ee LY 0 0 (%) yueseid yQ22 J, (L'3S-9'0S) LIS (GEF-O'GP) BEF OFF (S'OF-T'8E) S'6E (O'LI-L'Sh) FSF (S'9F-9'SF) SEP (Gieg—aige): Tre (L'Sh-8'68) TIP (Z'S8-0'08). SFE (9'IF-S'6E) 9'0F (L'PE-T'0E) FE TAS mOoNMOMONATMAA = SO O+ £O OF £O OF OF FO OF FO OF xoe ASL = eUNdIvIRT, OLIOD pue I[ey_ Olay :eUeURg AV9OSS6. Via ae ee vueduey Olle :eweuRg MACCS =. = = 2 = =. Inburyy ong :eweuRg M563 <= = = Se euasueyy) org :eureuRg MI6Eos oS SSS a eqeuiny, ep eiAevioyy -Bory eBySOD MASE:eS =| S b= = =| ea Se Quedey :evony eBysoD apnyisuoT AYIRIO'T ‘quasaid Y}99} suTIeWIOAGId YRIA\ U Jo asv}yUa0I0d —juUaseId YJoaT, “UIU UT ‘suo Joye sosayjueied ur sasuel “TAS 104 unwa) Dsnpawopfyg ut 422} suLIawOAeId Jo aouasaid pue azis Ul UOeLURA o1ydeis0ag—E AIAV], PHYLLOMEDUSA BUCKLEYI GROUP (ANURA: HYLIDAE) 13 The neural arches are low and non-imbricate. Poorly developed neural spines are present only on Presacrals I and II. The width of the transverse processes of Presacral III slightly exceeds the width of the sacral diapophyses in medinai, lemur and buckleyi. In psi- lopygion the widths are subequal. The transverse processes in all species are moderately wide and subequal in width. The profile of the process widths gradually narrows posteriorly. The tips of the transverse processes of Presacrals III-V bear conspicuous bits of calcified cartilage. These are noticeable in both dry and alizarin preparations. The processes of Presacrals III-IV are directed pos- terolaterally. The processes of Presacral VI are directed only slightly posteriorly; in the remaining vertebrae the processes are anterolat- erally directed, the inclination being most pronounced in lemur. The vertebral centra are uniformly procoelous. The sacral dia- pophyses of all species are widely expanded and articulate over a wide area with the ilial shaft. In psilopygion small flanges are pres- ent posteriorly at the base of the sacral diapophyses. In all species the coccyx bears no distinct ridge. A bicondylar sacro-coccygeal articulation is present in buckleyi, lemur and medinai. In all speci- mens of psilopygion the coccyx is fused to the sacrum with no trace of a suture. To my knowledge, this is the first report of a fused sacro-coccygeal articulation in the Hylidae (Trueb, 1973). Pectoral Girdle——The pectoral girdles are fully arciferal; the sterna are bifurcate and exhibit only traces of calcium deposition. The omosternum is questionably present in buckleyi and lemur as a small bit of cartilage; it is absent in medinai and psilopygion. The clavicles are slender and moderately arched; the coracoids are moderately robust and are best developed in medinai. Procoracoid cartilage is present in all species; it is most extensive in medinai. The epicoracoid cartilages are moderately curved and overlap over most of their length. The scapulae are much longer than the clav- icles and are proximally bifurcate in all species. The cleithra are bicapitate distally in buckleyi and medinai, and uncleft in the others. The suprascapular cartilages are about as long as wide; small amounts of calcium deposition can be noted. Pelvic Girdle.—The ilia of all species lack any indication of a crest on the shaft. Low dorsal protuberances are present in all species. The lateral preacetabular angle is less than 90° in the group; the interior pelvic angle is about 35°. The acetabular length is greater than the acetabular height. The ischium is moderately expanded; the pubis is poorly ossified. Myology The terminology of the following muscle descriptions follows Starrett (1968) and Tyler (1971). 14 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Fic. 4.—Dorsal (left) and ventral (right) views of skulls of Phyllomedusa psilopygion (top), KU 169613, male, paratype, and P. buckleyi (bottom), KU 179405, male. Line equals 1 cm. Depressor Musculature—The main portion of the depressor mandibulae originates from the posterior ramus of the squamosal (no fibers arise from the crista parotica) and inserts on the end of the mandible. A smaller slip of the depressor mandibulae arises from the medial aspect of the posteroventral portion of the tym- panic ring cartilage and also inserts on the end of the mandible. No slips of the depressor mandibulae arise from the dorsal fascia of the suprascapula. In Starrett’s code, the musculature can be characterized as SQat. Adductor Musculature—The adductor mandibulae externus lateralis is present in all four species, and originates from the quadratojugal and the base of the squamosal, except in medinai. The center of ossification is lost from the quadratojugal in medinai, PHYLLOMEDUSA BUCKLEYI GROUP (ANURA: HYLIDAE) 15 Fic. 5.—Dorsal (left) and ventral (right) views of skulls of Phyllomedusa lemur (top), KU 31707, male, and P. medinai (bottom), KU 179407, female. Line equals 1 cm. and the muscle originates solely from the squamosal. The lateral face of the angular receives the insertion in all species. The a.m. externus superficialis is present and well-developed in all species, and passes laterad to the mandibular branch of the tri- geminal nerve (V*). This muscle originates from anterior ramus and proximal portion of the ventral ramus of the squamosal, and inserts on the lateral face of the mandible. The a.m. anterior internus is present in all species, and takes its origin from the dorsal and lateral surfaces of the frontoparietal, becoming more tendinous where it passes laterad to the anterior ramus of the pterygoid to insert on the medial face of the angular. Present also in all species is the a.m. posterior longus. This most prominent of the adductor series arises from the dorsal surface of the crista parotica and inserts on the angular. 16 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY TABLE 4.—Multiple range comparisons of ratios between species, males only. 1 = P. buckleyi, 2 = P. lemur, 3 = P. medinai, 4 = P. psilopygion. Sample sizes as in Table 2. The species subtended by lines are not significantly different from each other at the 0.05 level. Ratio Multiple range comparison et) >; ee Se ie a Ara 1 ra rour/svi, a. 2... BOE 1 {ite wie ee pW Tah UES, =a ec ca See Eh DGS de aA Si Oe AR 4 Gon Se ia Likewise, the a.m. posterior articularis is found in all species, originating from the ventral ramus of the squamosal and inserting on the lateral face of the mandible, just posterior to the insertion of the externus superficialis. < The a.m. posterior subexternus apparently is absent in medinai and psilopygion. Only one specimen of each species was dissected however. In buckleyi and lemur, a few muscle fibers appear to originate from the anterior arm of the squamosal and pass ventrad to insert on the lateral face of the mandible. The fibers lie medial to nerve V*, and presumably represent a very reduced posterior subexternus muscle. Intermandibular Musculature —Tyler (1971) described the in- termandibular muscles for a number of phyllomedusines, including buckleyi and lemur. My dissections of the four species generally agree with Tyler’s observations on buckleyi and lemur. The sub- mentalis is small and araphic; the two halves of the principal element of the intermandibularis insert medially on a moderately wide aponeurosis; posterolateral elements of the intermandibularis are present; the interhyoideus is entire and not lobed. The afore- mentioned are typical phyllomedusine conditions (Tyler and Davies, 1978). Tyler (1971) pointed out that when the posterior subexternus and the externus superficialis are both present, the posterolateral ’ slip of the intermandibularis inserts on the posterior subexternus via a tendon, and the site of attachment is concealed by the ex- ternus superficialis. Furthermore, when the posterior subexternus is absent, the tendon attaches to the externus superficialis. My dis- sections of Pachymedusa dacnicolor, Agalychnis callidryas, and Phyllomedusa tarsius confirm Tyler’s observation. In all members of the buckleyi group, the tendon of the postero- lateral element inserts on the externus superficialis, whether the reduced posterior subexternus is present (buckleyi and lemur) or not (medinai and psilopygion). Obviously, further analysis is PHYLLOMEDUSA BUCKLEYI GROUP (ANURA: HYLIDAE) 17 Fic. 6.—Tadpoles of Phyllomedusa psilopygion (upper), KU 170213; P. lemur (center), USC-CRE 290-9; P. buckleyi (lower), KU 143553. x 2.5. needed to ascertain the extent of ontogenetic variation in the jaw musculature before its utility in systematics can be realized fully. Tadpoles Morphology——tiIn general morphology the tadpoles of the P. buckleyi group resemble those of Agalychnis and Pachymedusa (see Duellman, 1970). The body is oval-shaped; the caudal muscu- lature is slender (Fig. 6). The tail fins narrow gradually to a point. The denticle formula is 2/3; the second upper row is interrupted medially. The medial portion of the upper lip lacks papillae. The labial papillae are most dense in Jemur, and least dense in psi- lopygion (Fig. 7). 18 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY The chondrocranial elements are visible through the skin in psilopygion; this is probably due to a lack of pigment. Since the larvae were collected from a natural grotto, the condition is possibly environmentally induced. Coloration.—The body of buckleyi is grayish brown in preserva- tive; that of psilopygion is pale brown. The body of lemur is pur- plish pink with brown mottling. The tail musculature is cream in buckleyi, bluish gray in psilopygion, and pale cream with brown mottling in lemur. Dark brown pigment is present along the dorsal edge of the tail musculature in the three species; the tail fins are transparent in the group. KEY TO THE SPECIES IN THE PHYLLOMEDUSA BUCKLEYI GROUP 1. Snout sloping in lateral view; foot about one-fourth webbed —_...--- 2 Snout truncate in lateral view; foot webbing vestigial, barely visible 3 2. Calear generally present; quadratojugal present —.--._»__S P. buckleyi Calear absent;, quadratojugal ‘absent 2... P. medinai 3. Calcar generally present; sacro-coccygeal articulation fused; white dorsal warts usually present; para-anal tubercles absent _._....____. _ P. psilopygion Calcar absent; sacro-coccygeal articulation bicondylar; white dorsal warts never present; para-anal tubercles present _..---- P. lemur ACCOUNTS OF SPECIES Phyllomedusa buckleyi Boulenger Fig. 8 Phyllomedusa buckleyi Boulenger, 1882:425, pl. 29 [Holotype—-BMNH 80.12.5.230 (RR 1947.2.22.35) from Sarayacu, Provincia de Pastaza, Ec- uador]. Phyllomedusa loris Boulenger, 1912:186 [Holotype——BMNH 1912.11.1.71 (RR 1947.2.25.82) from El Topo, Rio Pastaza, Provincia de Tungurahua, Ecua- dor; synonymy fide Duellman, 1969:135]. Phyllomedusa buckleyi—Nieden, 1923:344. Phyllomedusa loris—Nieden, 1923:344. Hyla (Hylella) porifera Andersson, 1945:81 [Holotype —NHRM 1963 from Rio Pastaza watershed, Ecuador; synonymy fide Duellman, 1969:135]. Phyllomedusa buckleyi—Lutz, 1950:601,619. Phyllomedusa (Phyllomedusa) loris—Lutz, 1950:601,619. Phyllomedusa buckleyi—Funkhouser, 1957:26. Phyllomedusa loris—Funkhouser, 1957 :26. Phyllomedusa buckleyi—Duellman, 1968:6. Phyllomedusa loris—Duellman, 1968:6. Phyllomedusa buckleyi—Duellman, 1969: 135. Diagnosis.—Phyllomedusa buckleyi differs from other members of the buckleyi group by the following combination of characters: 1) modal webbing formula of foot II(2+)—(3.5)IIT(2.5)—(3.75) IV(3.5)—(2+)V; 2) calear present; 3) snout sloping in lateral view; 4) para-anal tubercles present; 5) outer metatarsal tubercle usually present; 6) white dorsal warts usually present; 7) quad- ratojugal present; 8) sacro-coccygeal articulation bicondylar. PHYLLOMEDUSA BUCKLEYI GROUP (ANURA: HYLIDAE) 19 Description.—Pertinent measurements and proportions in Tables 1 and 2. Head slightly wider than body; snout short, acutely rounded in dorsal view; in lateral view, sloping and rounded from lip to nostril in both sexes; canthus rostralis rounded, distinct; loreal region concave; lips thin and not flared; nostrils not protuberant, directed laterally; internarial region flat; eyes large and protuber- ant; pupil vertically elliptical; palpebrum clear; parotoid glands not differentiated; supratympanic fold thin, indistinct, obscuring dorsal edge of tympanum, and up to one-half of tympanum in some; fold extending posteroventrally to a point above the insertion of the arm; tympanum barely distinct, round, separated from eye by dis- tance equal to horizontal diameter of the tympanum. Axillary membrane absent; upper arm slender; forearm robust; ulnar fold low, indistinct; fingers moderately long; relative length from shortest to longest 1-2-4-3; discs of moderate size, rounded; subarticular tubercles large, subconical; distal tubercle on fourth finger bifid in most specimens; supernumerary tubercles lacking on fingers; palmar tubercle low, diffuse, rounded; prepollex slightly enlarged and bearing thin, horny nuptial excrescence in breeding males; slight rudimentary webbing between the last three fingers (Table 5). Legs moderately long and slender; small, blunt calcar on heel; a few small tubercles below calcar; inner tarsal fold barely distinct; outer tarsal fold barely distinct, appearing merely as a row of low tubercles in some; toes of moderate length; relative length from shortest to longest 1-2-3-5-4; discs rounded, smaller than those on fingers; inner metatarsal tubercle flattened, low, elliptical; outer metatarsal tubercle small and rounded; webbing absent between first and second toes; rudimentary webbing between second and third, and basal webbing between other toes (Table 5). Anal opening directed ventrally at midlevel of thighs; short anal flap present; para-anal region tubercular; skin on dorsum minutely granular; white dorsal warts present on some specimens; small granules in loreal and temporal regions; skin on throat, belly and posteroventral surface of thighs distinctly granular; skin elsewhere smooth; tongue lanceolate, notched posteriorly, free for about one- half its length; prevomerine teeth present in most; dentigerous processes of prevomers small, moderately separated medially, ori- ented posteromedially (\7) at anterior level of elliptical choanae; vocal slits present in males, short, parallel to jaw, extending from posterolateral corner of tongue to corner of mouth; vocal sac single, median, subgular. Coloration.—In life, dorsum lavender-brown by night; by day dorsum pale green to yellow-green, with purple flecks; dorsal sur- faces of toe and finger discs yellowish white with green wash; ventral surfaces of hands and feet flesh-colored; venter cream with OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 20 A (-~) (¥) = +P cle i (Fe). e's) Coe 7 hee ace (--) (%) hm aap ss ms (c’€) (GLE) c'é -p oa ace (SL’°S) € GL (€) +¢ € (-€) +¢ =a AI € = III (+2) (Sd) CZ “re GZ +3 I ‘sasayjuoied ul opour ‘wnuTUTU pue UMUTTXeUL =. =) SAS OS 7? Gor Pet cA ao 1pUuIpaut * snug] a ~— Mapyong — = noes Se Seige aie a en =_= 2 gt a ha ae —t Inwa] pueH ey oe uUIN[OO eB UI SenjeA puodss pue jsIg eYy, “dnois shazyong vsnpawopzjiyg ayy JO saroeds jo Surqqaa ul UOHeLIeA—'C¢ TIAV], PHYLLOMEDUSA BUCKLEYI GROUP (ANURA: HYLIDAE) 21 orange wash; throat off-white with pale orange wash; flanks and concealed surfaces bright orange, fading to cream on the edges; ulnar and tarsal fold off-white; anal tubercles off-white; dorsal warts pale yellow; fourth finger green dorsally, fading to pale yel- low; third finger dirty yellow; first and second fingers orange dor- sally; iris dirty cream; palpebrum clear, upper border slightly pig- mented. A newly metamorphosed individual (KU 178850) had the following coloration: dorsal surfaces of body and limbs light green with yellow wash; dorsal surface of toe pads and concealed surfaces dull orange; dorsal wart light yellow; iris off-white; venter fleshy orange; upper lip dirty yellow. In preservative, dorsal surfaces of body, forearms, thighs, shanks, tarsi, third and fourth fingers, and fourth and fifth toes bluish purple to pale blue; dorsal warts, anal tubercles, ulnar and tarsal fold, white; dark flecks present dorsally; other surfaces cream. Tadpoles——About 45 tadpoles were examined, ranging in de- velopment from Stage 25 to 41. The range of body lengths is 8.4— 24.1 mm; the range of total lengths is 22.1-58.0 mm. A represent- ative tadpole (KU 143553) is illustrated in Fig. 6. The following description is based on KU 143553 and 166219: body as wide as deep, deepest and widest at two-thirds the length of the body; top of head slightly convex; snout acutely rounded in lateral profile; in dorsal view almost truncate; nostrils dorsolateral, directed anterolaterally; internarial distance less than width of oral disc; eyes dorsolateral and directed laterally; spiracle a flap-like tube, ventral and sinistral to midline; spiracular opening at point about midlength of the body; chondrocranial elements not visible through skin; mouth anteroventral and directed anteriorly; cloacal tube short and dextral to caudal fin; slender caudal musculature tapers gradually to posterior end of fin; myomeres moderately de- veloped; at midlength of the tail the depth of the caudal muscu- lature slightly less than depth of ventral fin, but greater than that of dorsal fin; musculature extending to tip of tail; dorsal fin shallow anteriorly, not extending onto body; fin deepest at two-thirds its length from anterior; ventral fin deepest at midlength. Mouth small, with a shallow lateral fold; medial portion of up- per lip lacking papillae; elsewhere, papillae present in two to three rows along the border; a few papillae present medial to border in region of lateral fold; upper beak moderately deep, lower beak shallow and finely serrate; upper beak with about twelve coarse serrations medially, otherwise finely serrate; two upper and three lower rows of denticles; upper rows of same length; first upper row uninterrupted, second row broken narrowly medially; three lower rows of equal length, unbroken; denticles of first upper and third lower rows smaller than those of other rows (Fig. 7). In life, body olive-brown above; venter silver-yellow to creamy 22 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Fic. 7—Mouths of tadpoles. A. Phyllomedusa buckleyi, KU 166219; B. P. psilopygion, KU 170213; C. P. lemur, USC-CRE 290-9. Line equals 2 mm. green with greenish gold iridescence on sides; tail musculature cream to tan; fins transparent; iris silver to bronze; one metamorphic with emerging forelimbs has a green dorsum (KU 143553). In preservative, snout and top of head grayish brown; sides and venter dark bluish gray; caudal musculature dirty cream with dorsal edge of musculature dark brown; dorsal and ventral fins colorless and transparent; blood vessels easily distinguished. — One lot of tadpoles (KU 166218) has a slightly different color- ation in preservative: the caudal musculature is light purple fading to reddish brown anteriorly. This may be an artifact of preserva- tion. Another group has such an extremely heavy deposition of melanophores in the skin of the body and tail that its coloration in life was described as black. This is not unexpected, because tad- poles of Pachymedusa dacnicolor and Phyllomedusa trinitatus are’ known to exhibit a “tail-darkening reaction” (Bagnara, 1974). Eggs.—The only known clutch of eggs (n = 98) was deposited in a plastic bag by a female (KU 143234) in amplexus. The site of natural oviposition is yet unknown, but presumably on vegetation above water. The clutch (KU 143552) was preserved at Stage 8 (Gosner); the total egg and jelly diameter is about 4.5 mm. That of the egg alone is about 3.1 mm. In life, the yolk is green; the jelly, clear. Etymology.—The specific name is a patronym for Mr. Buckley, the collector of the holotype. PHYLLOMEDUSA BUCKLEYI GROUP (ANURA: HYLIDAE) 23 Distribution.—Phyllomedusa buckleyi is known from eight sites on the Amazonian slopes of Ecuador, in the drainages of the Rio Coca and Rio Pastaza, at elevations of 400-1870 m (Fig. 2). Ecology.—In the valley of the upper Rio Coca (Rio Quijos), P. buckleyi occurs in lower montane rain forest. The valley is some- what cut-over along the Quito-Lago Agrio road; temporary ponds are present in the cut-over areas and numerous small streams drain the slopes. Tadpoles of buckleyi have been taken from streamside pools, temporary ponds, and puddles. In the Pastaza drainage the El Topo and Abitagua localities are lower montane rainforest; the Sarayacu site is rainforest. Phyllo- medusa buckleyi is sympatric with P. perinesos in the Rio Quijos Valley; in the Rio Pastaza Valley buckleyi is sympatric with P. vaillanti and P. tarsius. Mating Call—No recordings of the mating call are available. However, from field observations the call is known to be a short “cluck.” Remarks.—The holotypes of P. buckleyi and its two junior synonyms (Phyllomedusa loris and Hyla porifera), and one other specimen (UMMZ 92101) are from the Rio Pastaza drainage. All other known specimens have been collected by University of Kansas field parties in the Rio Quijos Valley. Most individuals have been taken at night on low vegetation overhanging pools of water. Phyl- lomedusa buckleyi and perinesos have been collected syntopically in the Quijos drainage at a site where the Quito-Lago Agrio road crosses the Rio Salado. At this site twelve perinesos were collected on 6-7 October 1974 from bushes and tall grass in the vicinity of a small pond; no buckleyi were taken. On 17-19 March 1975 twelve perinesos were taken from low vegetation around the pond and cut-over forest at the same locality. Seven buckleyi were collected and low vegetation adjacent to the pond; none were taken from the cut-over forest. At this same site about 60 perinesos and five buckleyi were collected on the evening of 17 July 1977, following a light rain. All of the buckleyi were taken from low vegetation overhanging a small pond; the specimens of perinesos, including amplectant pairs, also were found around the pond, but the vast majority were in the surrounding forest clearing, several meters from water. These limited data suggest spatial segregation for the two species. Phyllomedusa lemur Boulenger Fig. 9 Phyllomedusa lemur Boulenger, 1882:425 [Holotype—BMNH 74.8.11.9 (RR 1947.2.22.36) from Costa Rica]. Agalychnis lemur—Cope, 1887:15. Agalychnis lemur—Giinther, 1901(1885-1902):291. Phyllomedusa lemur—Nieden, 1923:344. 24 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Phyllomedusa (Agalychnis) lemur—Lutz, 1950:601,619. Phyllomedusa lemur—Taylor, 1952:809. Phyllomedusa lemur—Funkhouser, 1957:31. Phyllomedusa lemur—Duellman, 1968:6. Diagnosis.—Phyllomedusa lemur differs from other members of the species group by the following combination of characters: 1) modal webbing formula of foot III(3)—(4)IV(4-)—(3-)V; 2) calear absent; 3) snout truncate in lateral view; 4) para-anal tubercles present; 5) outer metatarsal tubercle absent; 6) white dorsal warts absent; 7) quadratojugal present; 8) sacro-coccygeal articulation bicondylar. Description.—Pertinent measurements and proportions in Tables 1 and 2. Head wider than body; snout short, truncate in dorsal and lateral views of both sexes; nostrils at tip of snout, directed later- ally, not protuberant; internarial region flat; canthus distinct and rounded; loreal region barely concave; lips thin, not flared; inter- orbital region flat; eyes large and protuberant; pupil vertically ellip- tical; palpebrum clear; parotoid glands not differentiated; supra- tympanic fold barely discernible, thin and obscuring the upper edge of tympanum; tympanum round, distinct, separated from eye by distance equal to about three-fourths the diameter of tympanum. Axillary membrane absent; upper arm slender; forearm moder- ately robust; weak ulnar fold extending from elbow to tip of fourth finger; fingers moderately long with medium size round discs; web- bing, if present, very scanty (Table 5); order of fingers from short- est to longest 1-2-4-3; fingers oval in cross-section; subarticular tubercles large and round, single; supernumerary tubercles small, low and indistinct; palmar tubercle small, round and flattened; prepollex moderately enlarged, with thin horny nuptial excrescence in breeding males. Hind limbs slender and of medium length; outer tarsal fold present, barely distinct; calcar absent, very small tubercle on heel in a few; inner metatarsal tubercle low, elliptical, not visible from above; outer metatarsal tubercle absent; toes long and slender; discs round, smaller than those on fingers; rudimentary webbing, present (Table 5); order of toes from shortest to longest 1-2-3-5-4; fifth toe with slight toe fringe; subarticular tubercles large and round; supernumerary tubercles small, indistinct, and in a single row on proximal portions of each digit. Anus directed posteroventrally at midlevel of thighs; short anal sheath present; para-anal region and ventral surface of groin tuber- cular; tongue lanceolate, moderately notched from behind, free posteriorly for one-half its length; belly skin granular; skin else- where smooth; white dorsal warts absent; no osteoderms; vocal sac single, median, subgular, barely distensible; prevomerine teeth in some; dentigerous processes oriented posteromedially (\7), mod- PHYLLOMEDUSA BUCKLEYI GROUP (ANURA: HYLIDAE) 25 erately separated just posterior to the anterior border of elliptical choanae; vocal slits present in males, short, extending from postero- lateral edge of tongue to corner of mouth. Coloration.—In life, pale green by day; dorsal surfaces of body, forearm, fourth finger, hind limbs, and fourth and fifth toes pale green; rest of hind limb and dorsal surfaces of toes deep orange- yellow; dorsal surfaces of upper arms dark yellow; flanks orange; ventral surfaces of arms, hands and feet pinkish cream; belly creamy white; chin, upper and lower lips, outer edge of tarsi and forearms white. At night, dorsum reddish brown to lavender-brown; thighs and arms deep yellow; venter white; pale green flecks present on dorsum of some; iris silver-bronze with a black periphery. In preservative, dorsum light bluish-gray to lavender to reddish- brown; dark brown dorsal flecks in some; hidden surfaces of limbs and ventral surfaces creamy white. Tadpoles.—Ten tadpoles of P. lemur were examined, ranging from Stage 25 to Stage 41. The range of body lengths is 6.7-18.7 mm; the range of total lengths is 16.5-49.7 mm. A representative tadpole (USC-CRE 290-9) is illustrated in Fig. 6. The following description is based on USC-CRE 290-9, a tad- pole in Stage 32 with a body length of 16.0 mm and total length of 42.4 mm: body slightly wider than deep, being deepest and widest at a point two-thirds the length of the body; snout acutely rounded in lateral view; in dorsal view snout almost truncate; nos- trils dorsolateral, situated one-fourth distance from eye to tip of snout, and directed anterolaterally; internarial distance equal to width of oral disc; eyes dorsolateral and directed laterally; spiracle a flap-like tube, ventral and sinistral to midline, situated at mid- length of body; chondrocranial elements not visible through skin; mouth anteroventral and small, directed anteriorly; short cloacal tube, dextral to caudal fin; slender caudal musculature tapering posteriorly gradually to end of fin; myomeres moderately differenti- ated; at midlength of tail the dorsal fin, ventral fin, and caudal musculature are of equal depth; dorsal fin deepest at a point two- thirds the length of the tail, and extending onto body; ventral fin of uniform depth anteriorly, deepening slightly just posterior to the midlength of the tail and tapering gradually to the tip. Mouth small, with a slight lateral fold; median portion of upper lip lacking papillae; elsewhere mouth bordered by two to three rows of papillae; papillae densely packed in lateral fold region; upper beak moderately deep; lower beak shallower; medial portion of upper beak edge with about twelve coarse serrations; lower beak and lateral portions of upper beak finely serrate; two upper and three lower rows of denticles; first upper and third lower rows with fine denticles; remaining rows with coarse denticles; upper rows of equal length, the second being interrupted medially; lower rows 26 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY unbroken and shorter than upper rows; lower third row shorter than other two (Fig. 7). In preservative, top of head brown; sides and venter purplish pink; caudal cream; sides, venter, and tail with brown flecks and mottling; tail fins transparent, with some brown flecks in anterior portion of dorsal and ventral fins. . Eggs—No clutches of eggs were available for examination. A color transparency (KU 1417) of a clutch of presumably Phyllo- medusa lemur eggs from Tapanti, Costa Rica, demonstrates a clutch size of about 70. The eggs are arranged in a single layered mass in the central portion of the upper surface of a leaf. Van Eijsden (1977) reported clutch sizes in captivity of 23 and 25 eggs. Etymology.—tThe specific name is a noun in apposition, in refer- ence to the lemur-like locomotion of the species. Distribution—The species is found primarily on the Caribbean slopes of the Cordilleras de Tilaran, Central, and Talamanca in Costa Rica; from the Atlantic and upper Pacific slopes of the Cor- dillera de Talamanca, and the southern slopes of the Serrania de Tabasara, on Cerro Mali and Cerro Tacarcuna in Panama. The species ranges from elevations of 440 to 1600 m (Fig. 1). Ecology.—Phyllomedusa lemur is generally considered to be re- stricted to the Subtropical Life Zone of Holdridge (Savage and Heyer, 1969). In Costa Rica the biotemperature of this zone is 18-24° at altitudes of 500-1500 m. The species is part of the Carib- bean versant fauna in Costa Rica; the localities around Monteverde and La Palma are on the upper Pacific slopes, virtually on the continental divide where the biota is essentially an “overflow” from the Caribbean slopes. Collections from Tapanti, while properly south of the Cordillera Central, are nonetheless in the Atlantic drainage of the Rio Reventazén. All other localities in Costa Rica are clearly in Atlantic drainages. Records of the species in Panama are sparse and widespread, and the ecological distribution requires some explanation. Collec- tions from the western province of Bocas del Toro are from Carib- bean versant lower montane rainforest. In western Panama the: species is found also on the Pacific slopes, at the Fortuna Dam site on the upper Rio Chiriqui, Provincia de Chiriqui. The upper Chiriqui Valley can be considered Temperate Wet (Cf) under the Koppen system, as opposed to the Temperate Wet and Dry climate (Cw) that is characteristic of most of the Pacific versant, including the lower drainage of the Rio Chiriqui. The dam site receives about 4000 mm of rain annually and lacks a dry season. In addition, the herpetofauna of the upper Chiriqui reflects a greater zoogeo- graphic affinity to the Caribbean, rather than the Pacific assemblage (Charles W. Myers, pers. comm.). The exceptional climate and herpetofauna can be explained by noting that the upper Chiriqut a —— PHYLLOMEDUSA BUCKLEYI GROUP (ANURA: HYLIDAE) 27 Valley structurally divides the central cordillera, marking the west- ern end of the Serrania de Tabasara and eastern extent of the Cor- dillera de Talamanca. In this depression the continental divide drops below 1200 m. In effect, the climate and herpetofauna spill over onto the Pacific slopes. A second noteworthy occurrence of Phyllomedusa lemur is on the south slope of Cerro Campana and at El Valle, a low Pleistocene volcanic crater about 20 km SW of Cerro Campana. Again, a sea- sonally dry forest is associated with the Pacific drainages of this region. However, Cerro Campana is the eastern terminus of the central cordillera, and the absence of nearby high mountains allows the 1000 m peak and the surrounding area to intercept the north- easterly Caribbean winds, with the resultant orographic rainfall. Thus the upper Pacific slopes near El Valle and Cerro Campana support a humid subtropical cloud forest (Myers, 1969). The easternmost extent of Phyllomedusa lemur is the drainage of the Rio Tuira, on the slopes of Cerro Mali and Cerro Tacarcuna. The species is unknown from the Caribbean side of the Serrania del Darién; thus the Rio Tuira sites, in humid montane forest, are the only records of the species in an exclusively Pacific drainage. The species is not known from Colombia. Mating Call_—The mating call of Phyllomedusa lemur is a very short trill. A recording from La Palma, Costa Rica (KU Tape 67), at an air temperature of 17.8°C, consists of four calls in a period of 99 seconds. The call duration is 0.32-0.40 seconds. The funda- mental frequency is 250-300 Hertz; the dominant frequency is about 950-1000 Hertz. The call is poorly modulated; harmonics can be distinguished at 1350, 1900, and 2200 Hertz. The pulse rate is 39-41 pulses per second (Fig. 10). Remarks.—Van Ejijsden (1977) and Schulte (1977) provided observations on P. lemur in captivity. | Phyllomedusa medinai Funkhouser Fig. 9 Phyllomedusa medinae Funkhouser, 1962:588 [Holotype—EBRG (no number) from Parque Henri Pittier, Rancho Grande, Estado de Aragua, Venezuela]. Phyllomedusa medinae—Duellman, 1968:6. Diagnosis—Phyllomedusa medinai differs from other members of the species group by the following combination of characters: 1) modal webbing formula of foot I(2+)—(2.25)II(2+)—(3.5) III (2+)—(3.5)1IV(3+)—(2+)V; 2) calcar absent; 3) snout sloping in lateral view; 4) para-anal tubercles present; 5) outer metatarsal tubercle sometimes present; 6) white dorsal warts sometimes pres- ent; 7) quadratojugal absent; 8) sacro-coccygeal articulation bi- condylar. 28 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Fic. 8.—Top: Phyllomedusa psilopygion, male, holotype, SVL 42.0 mm, KU 169608. Bottom: P. buckleyi, male, SVL 43.4 mm, KU 143225. Photos by William E. Duellman. Description.—Pertinent measurements and proportions in Tables 1 and 2. Head slightly wider than body; snout short, acutely rounded in dorsal view; intermediate between truncate and sloping in lateral view in both sexes; nostrils at tip of snout, directed later- PHYLLOMEDUSA BUCKLEYI GROUP (ANURA: HYLIDAE) 29 Fic. 9.—Top: Phyllomedusa medinai, male, SVL 43.2 mm, KU 167186. Bottom: P. lemur, female, SVL 43.7 mm, AMNH 94915. Photos by William E. Duellman and Charles W. Myers, respectively. ally, not protuberant; internarial region slightly concave; canthus rounded, distinct; loreal region barely concave; lips thin, not flared; interorbital region flat; eyes large and protuberant; pupil vertically 30 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY elliptical; palpebrum clear; parotoid glands not differentiated; su- pratympanic fold thin, obscuring dorsal and posterior edge of tym- panum, extending posteroventrad to angle of jaw; tympanum round, distinct, separated from eye by distance equal to one-half diameter of tympanum. Axillary membrane absent; upper arm slender; forearm moder- ately robust, stouter in males; slight ulnar fold extending from elbow to-disc of fourth finger; fingers short; discs of moderate size, rounded; order of fingers from shortest to longest 1-2-4-3; fingers flattened in cross section; no finger fringe; distal subarticular tuber- cles oval, subconical, large and single; proximal subarticular tuber- cles more rounded and smaller; supernumerary tubercles small and few in number, in a single row on proximal portions of digits; palmar tubercle diffuse; prepollex moderately developed, with thin horny nuptial excrescence in breeding males; small amounts of webbing between fingers (Table 5). Hind limbs slender and moderately long; weak tarsal fold ex- tending from heel to pad of fourth toe; calcar absent; outer meta- tarsal tubercle present in some; inner metatarsal tubercle flat, oval, small, not visible from dorsal view; toes moderately long, with rudimentary webbing (Table 5); discs rounded, smaller than those on fingers; distal subarticular tubercles single, medium size, sub- conical; proximal subarticular tubercles smaller and less distinct; supernumerary tubercles very few in number, small and rounded; order of toes from shortest to longest 1-2-3-5-4. Anus directed ventrally at lower level of thighs, covered by short anal flap; para-anal tubercles present; skin on belly and ven- tral surface of groin finely granular; skin on anterior part of body slightly granular, with granules becoming better defined on dorsal surface of hind limbs and posterior dorsum; skin smooth elsewhere; white dorsal warts present in some specimens; no osteoderms; tongue lanceolate, slightly notched, free posteriorly for one-third to one-half of its length (tongue round, not notched, not free in CAS- SU 21828); prevomerine teeth present in about one-half of the specimens; dentigerous processes moderately separated, located midway between anterior and posterior levels of elliptical choanae, oriented posteromedially (\.7); vocal sac single, median, and sub- gular; vocal slits present in males, short, extending from corner of mouth anteriorly toward posterolateral corner of tongue. Coloration.—In life, dorsum pale green by day; at night, reddish brown with rusty red flecks; ulnar and tarsal strips creamy white; chest and throat white; belly, other ventral surfaces, upper portions of limbs, hands and feet orange; iris pale silvery bronze. In preservative, dorsal surfaces of head, body, limbs, portions of fourth finger and fourth and fifth toes dark lavender in recently preserved specimens, reddish brown to cream in faded specimens; PHYLLOMEDUSA BUCKLEYI GROUP (ANURA: HYLIDAE) 31 upper edge of palpebrum pigmented; brown flecks present on dor- sum of all specimens; white dorsal warts in most; ventral and con- cealed surfaces, and dorsal surfaces of fingers and toes creamy white. Tadpoles and Eggs.—The tadpoles and eggs mentioned in Funk- houser (1962) apparently are lost. However, Dr. Anne Funkhouser has kindly provided me with an unpublished manuscript on phyllo- medusine tadpoles from which is taken the following description, based on larvae from a single clutch of about 150 eggs: “Eggs. gelatinous coat 6 mm, egg proper 2 mm, pale blue. “Tadpoles. Newly hatched larvae (11 mm) still have long ex- ternal gills which become covered by the operculum within a day after hatching. There is large amount of yolk present, and feeding does not begin until several days after hatching. “In 18 mm tadpoles the body length is one-third or less the total length; body oval in outline when viewed from above; deeper than wide. Eyes lateral, prominent. Nostrils subterminal. Spiracle ven- tral and slightly sinistral to longitudinal body axis; opening slightly closer to base of tail than to snout. Anus median. Tail at least twice body length; fins broadest at midpoint, terminating posteriorly in a pointed tip. Dorsal fin not extending onto body. “Mouth terminal. Single papillary border around mouth incom- plete above. Beaks broadly arched; lower more well-developed than upper. Labial denticles 1/2; upper and first lower interrupted medi- ally; set on folds of skin. | “In life somewhat iridescent; uncolored except for scattered black pigment at front of head, along tail musculature and as a conspicuous inner covering of the abdominal cavity. Iridescence lost with preservation.” The labial denticle formula recorded above is associated with an obviously young tadpole. I would predict that a more advanced tadpole of P. medinai would have the normal phyllomedusine den- ticle formula of 2/3. The only apparent extant larval specimen of medinai (KU 179406) is a single metamorphic young at Stage 43 with a SVL of 23.0 mm and total length of 44.6 mm. The adult pattern of web- bing is fully developed between the fingers and toes. In preserva- tive, the dorsal surfaces are light gray-brown with several dark brown spots on the dorsum. The venter is pale gray. Funkhouser (1962) noted that the colubrid snake Leimadophis zweifeli preys on the eggs of P. medinai. Etymology.—The specific epithet is a patronym for the collector, Sr. Gonzalo Medina Padilla. Distribution —Phyllomedusa medinai is known only from the type locality, Rancho Grande, Venezuela, 1100 m (Fig. 2). Ecology.—Little is known of the habits of this cloud forest 32 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY dweller. Two specimens (KU 167186-7) were taken on vegetation 20-30 cm above ground at night. The range of temperatures at this time (2-12 August 1974) was 17.0-26.0° C; rainfall averaged 13.8 mm per day with a range of 0-54 mm. Mating Call_——The mating call of the species is unknown. Remarks.—The Estacion Biol6gica Rancho Grande is on the southern slope of the coastal (northern) arm of the Cordillera de la Costa (Caribbean range of the Andes) at an elevation of 1100 m. Within 200 m of the site is a narrow pass, Portachuelo, bordered on the east by Pico Guacamayo (1900 m), and on the west by Pico Periquito (1500 m). The station, noted simply as Rancho Grande on maps, is part of the Parque Nacional Henri Pittier, a largely undisturbed region. Rancho Grande lies in an ecologically transi- tional zone (Test, Heatwole, and Sexton, 1966). The station is at the lower limit of cloud forest on the southern face of the divide; below the cloud forest there is a warmer, drier, semi-deciduous vegetation zone. June, July, and August are the wettest months; January, February, and March have only traces of rain. Rivero (1967) suggested that since the species was named in honor of Gonzalo Medina, the. spelling of the trivial name as medinae was a lapsus calami by Funkhouser. He implied that the specific epithet should be spelled as medinai, but due to either a typographical error or lapsus calami on his part he again spelled it medinae. According to the 1961 Code—in effect at the time the species was described—a species-group name formed from the per- sonal name of a man must end in -i (Article 31). The spelling of the specific epithet as medinae is an incorrect original spelling and must be corrected (Article 32). The name is emended here to medinai. Phyllomedusa psilopygion new species Fig. 8 Holotype—KU 169608, an adult male, from 8 km W Danubio, Rio Anchicayé, Departamento de Valle, Colombia, 300 m, [03°37’N, 76°47’'W ], obtained on 12 June 1975, by William E. Duellman. Paratopotypes.—KU_ 169609-13, William E. Duellman, 12-14 June 1975; ICN 4755, Craig Downer, 3 August 1978. Paratype —AMNH_ 87873 from Quebrada Guangui, Rio Patia, in the Rio Saija drainage, Departamento de Cauca, Colombia, ec. 100 m, obtained on 29 October 1971, by Adriano Granja and Borys Malkin. Diagnosis —Phyllomedusa psilopygion differs from other mem- bers of the species group by the following combination of charac- ters: 1) modal webbing formula of foot III(3-)—(4)IV(4-)— (3-)V; 2) calcar present; 3) snout truncate in lateral view; 4) para- anal tubercles absent; 5) outer metatarsal tubercle absent; 6) white PHYLLOMEDUSA BUCKLEYI GROUP (ANURA: HYLIDAE) 33 dorsal warts sometimes present; 7) quadratojugal present; 8) sacro- coccygeal articulation fused. Description—N = 6 6 6,2 2 2; pertinent measurements and proportions in Tables 1 and 2. Head slightly wider than body; snout short, truncate in lateral view, acutely rounded in dorsal view of both sexes; nostrils at tip of snout, not protuberant, directed lat- erally; internarial region flat; canthus rounded; loreal region barely concave; lips thin and not flared; interorbital region flat; eyes large and protuberant; pupil vertically elliptical; palpebrum clear; paro- toid glands not differentiated; supratympanic fold thin, obscuring dorsal and posterior edge of tympanum, extending posteroventrad to angle of jaw; tympanum round, distinct, separated from eye by distance equal to one-third diameter of tympanum. Axillary membrane absent; upper arm slender; forearm moder- ately developed, with weakly developed ulnar fold extending from elbow to disc of fourth finger; fingers of moderate length, not webbed; discs rounded and small; fingers flattened in cross section; slight finger fringe present on either side of digits; order of fingers from shortest to longest 1-2-4-3; distal subarticular tubercles barely rounded and elevated; tubercle on fourth finger slightly bifid; super- numerary tubercles small, flattened, present proximally; palmar tu- bercle diffuse, single; prepollex moderately developed, with thin horny nuptial excrescence in breeding males. Hind limbs slender and moderately long; weak outer tarsal fold extending from small calcar and merging with toe fringe of fifth toe; inner tarsal fold very weak; outer metatarsal tubercle absent; inner metatarsal tubercle flat, elliptical, not visible from above; toes of moderate length, webbed basally (Table 5); discs rounded, smaller than those on fingers; toe fringe present, widening distally; subarticular tubercles large, rounded, flattened on the more distal segments; supernumerary tubercles small and indistinct; order of toes from shortest to longest 1-2-3-5-4. Anus directed posteroventrally at midlevel of thighs, covered by short anal flap; tubercles absent on para-anal region; skin on belly and ventral surface of groin faintly granular, smooth else- where; white dorsal warts present in some; no osteoderms; tongue lanceolate, slightly notched from behind, free posteriorly for one- third of its length; prevomerine teeth present; dentigerous processes oriented posteromedially (\“), moderately separated at level of anterior margins of elliptical choanae; vocal sac single, median, subgular; vocal slits present in males, short, extending from postero- lateral corner of tongue to corner of mouth. Coloration.—In life, at night, dorsum dull green to reddish brown with green flecks; by day, pale green with or without reddish brown flecks; dorsal warts creamy yellow; flanks, upper arms, an- terior and posterior surfaces of thighs, fingers 1-3, toes 1-4, and 34 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY ventral surfaces of hind limbs bright orange; throat, belly, anal re- gion and ventral portion of forearms white; tarsal stripe and elbow pale gray; iris creamy white; palpebrum clear. In preservative, dorsal surfaces of head, body, limbs, fourth fingers and fourth and fifth toes dark lavender brown, some with splotches of bluish gray and pink; dorsal warts white; ventral and concealed surfaces and dorsal surfaces of fingers 1-3 and toes 1-4 creamy white. Tadpoles.—A lot of fifteen tadpoles (KU 170213), all at Stage 25, was examined. The mean body length is 11.8 mm (range, 8.5- 17.9); the mean total length is 29.4 mm (range, 24.1-40.4). A typ- ical tadpole with body length of 17.9 mm and total length of 40.4 mm is illustrated in Fig. 6. Body slightly wider than deep; snout acutely rounded in lateral profile; in dorsal view snout almost truncate; body widest and deepest at two-thirds length of body; nostrils dorsolateral, directed anterolaterally, situated at end of snout; internarial distance equal to width of oral disc; eyes dorsolateral and directed laterally; spiracle a flap-like tube, ventral and sinistral to midline; spiracular opening at a point slightly more than one-third length of body from anterior end; skin transparent; chondrocranial elements visible through dorsal skin of head; mouth anteroventral and directed an- teriorly; cloacal tube short and dextral to caudal fin; caudal muscu- lature slender, tapering distally to end of fin; myomeres poorly differentiated; ventral fin deepest just posterior to body and taper- ing gradually to tail tip; dorsal fin narrow anteriorly, widening to its deepest at midlength of tail; dorsal fin never as deep as ventral fin in the same vertical transect; dorsal fin not extending onto body. Mouth small, without lateral fold; upper lip lacking papillae medially; otherwise, papillae present in one row laterally and two rows ventrally along border; very few present elsewhere; upper and lower beaks shallow, both with small uniform serrations; two upper and three lower rows of denticles; second upper row interrupted medially; upper rows of equal length; lower rows unbroken and not as long as upper rows; third lower row shorter than others; denticles on first upper and third lower rows finer than those of other rows (Fig. 7). In life, body reddish brown above with gold flecking below; tail bluish gray; smaller individuals nearly pigmentless; iris pale bronze. In preservative, dorsal surface brown; sides lighter brown, almost transparent with a slight metallic blue sheen; caudal fins transparent; larger tadpoles having dark brown pigment along dor- sal edge of caudal musculature and tail fin in the anterior half of tail; posterior one-fourth of tail fin without pigment. Eggs.—A single clutch of eggs (KU 170214), presumably of this species, is available for examination. The clutch (n = 32) is em- PHYLLOMEDUSA BUCKLEYI GROUP (ANURA: HYLIDAE) 35 bedded in a single gelatinous matrix; the eggs are at Stage 16; the diameter of the eggs averages 4.3 mm. In life the yolk is grayish tan; the jelly, clear. Etymology.—The specific epithet is derived from the Greek psilos, meaning smooth or bare, and the Greek pyge, f., meaning rump; pygion is the neuter diminuitive form. The epithet psilo- pygion is a noun in apposition, applied in reference to the lack of para-anal tubercles, which are possessed by other species in the group. Distribution—Phyllomedusa psilopygion is known only from elevations of 100 and 300 m on the Pacific lowlands of southern Colombia ( Fig. 2). Ecology.—The type series was collected in an area of steep slopes covered with rainforest. Some males were found calling from bushes and tangled vines at the mouth of a natural grotto at night. Tadpoles were collected from the same grotto by Linda Trueb, and John E. Simmons found the clutch of apparently phyl- lomedusine eggs on the underside of a rock protruding from the side of the grotto. During the period 10-14 June 1975 the range of temperatures at the locality was 21.0-32.0° C; the average rainfall was 4.5 mm with a range of 2.0-9.5 mm. KILOHERTZ TIME IN SECONDS SECTION Fic. 10.—Audiospectrograms of mating calls. A. Phyllomedusa lemur (KU Tape 67). B. P. psilopygion (KU Tape 1350). 36 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Mating Call—tThe call of P. psilopygion consists of a single chirp. At the type locality on 13 June 1975, five calls were recorded during a period of 143 seconds from KU 169611 (KU Tape 1350) at an air temperature of 23° C at 2130 h. The call duration is 0.04— 0.05 seconds. The fundamental frequency is about 250 Hertz; the dominant frequency, 1900 Hertz. The well-modulated call has dis- tinguishable harmonics at about 250, 700, 1100, 1500, 1900, 2300, 2700, 3200, 3600, 4000, and 4800 Hertz, with the most intense fre- quencies at 1500, 1900, and 2300 Hertz (Fig. 10). Remarks.—The type locality is on the grounds of the Central Hidroeléctrica de Anchicaya generating facility. One paratype (AMNH 87873) was taken in the drainage of the Rio Saija, about 120 km SW by air from the type locality (See Myers and Daly, 1976, for map). This paratype differs from the type series only by the apparent absence of a small calcar on the heel, but the limbs are strongly flexed at the heels, where the skin is stretched and possibly abraded slightly. In addition, some individuals of species normally having small calcars (e.g., P. buckleyi) do lack them. BIOGEOGRAPHY The most cogent biogeographical analyses are those that are based on a hypothetico-deductive model; such hypotheses have the potential to be tested and falsified. A choice between competing hypotheses therefore will have a logical basis. The formulation of such biogeographical hypotheses requires a cladistic analysis of the group under study (Ball, 1975). At present, the lack of data on other phyllomedusines precludes the formation of a hypothesis of relationships within the buckleyi group. It has been noted in pre- vious sections that buckleyi and medinai are superficially most similar to each other; lemur and psilopygion closely resemble each other. Overall phenetic similarity is not necessarily indicative of relationship, however. Thus, a deductive analysis of the biogeog- raphy of the group is not feasible. It is of heuristic value to de- scribe the general patterns of distribution in the group, however. The four allopatric species are isolated from one another by various segments of the Andean cordillera. A parsimonious explanation that does not invoke unnecessary dispersal is that the Andean orog- eny dissected a previously widespread species in northwestern South America. Possibly this stock was present in the early Plio- cene, when the northern Andes lay in a tropical belt with the great- est elevation not over 1000 m (Van der Hammen, 1974). The major uplift of the cordillera at the end of the Pliocene fragmented the stock; further differentiation may have occurred in Pleistocene refugia (Simpson, 1975; Haffer, 1970). It is worth noting that the presence of P. medinai in the Cordillera de la Costa of Venezuela provides a link of this region to the rest of the Andes; the herpeto- PHYLLOMEDUSA BUCKLEYI GROUP (ANURA: HYLIDAE) 37 faunal similarity between the Cordillera de la Costa and the Andes of Colombia is very low (Duellman, pers. comm.). It is significant that other groups of montane anurans display similar distribution patterns in northwestern South America. Hemi- phractus, Rhamphophryne, the Hyla bogotensis group, and the Gastrotheca cornuta group have patterns that suggest that a for- merly widespread biota was divided by the Andean orogeny. For distributional analyses of these groups see Duellman (1972) and Trueb (1971, 1974). RESUMEN El grupo buckleyi del género Phyllomedusa esta compuesto por cuatro especies. Una de ellas descrita aqui como nueva: Phyllo- medusa psilopygion, de la selva de la regién de Chocé en Colombia, en el valle del Rio Anchicaya (Departamento de Valle) y del Rio Patia (Departamento de Cauca). De las otras especies, Phyllo- medusa lemur es conocida de las laderas caribes de Panama y Costa Rica; P. buckleyi se encuentra en las laderas amazénicas del Ecua- dor, y P. medinai es conocida sdlo de Rancho Grande, Estado de Aragua, Venezuela. Las especies habitan selvas htunedas de 100 m hasta 1870 m de altitud. Miembros del group buckleyi difieren de las otras especies de la subfamilia Phyllomedusinae por Ja combi- naciOn siguiente de caracteres: 1) Membrana de manos y pies menos de un cuarto del longitud del dedo. 2) Primer dedo posterior mas corto que el segundo. 3) Glandula pardtida no diferenciada. 4) Parpado inferior sin reticulacién. 5) Flancos anaranjados (especimenes vivos) sin barras ni manchas. 6) Ramo posterior (pars scapularis) del musculo depressor man- dibulae ausente. Phyllomedusa perinesos no es miembro del grupo buckleyi. Los renacuajos del grupo tienen cuerpos ovalados, con aletas de altura moderada, boca anteroventral, dos filas superiores y tres filas inferiores de denticulos. Papilas labiales faltan en la regi6n mediana del labio superior. Los renacuajos son similares a los de Agalychnis, Pachymedusa, y algunos de Phyllomedusa. Los huevos sin pigmento son depositados en la vegetacién (0 piedras) encima de aguas lénticas. SPECIMENS EXAMINED Phyllomedusa buckleyi ECUADOR: Napo: South Slope Cordillera del Dué. Rio Coca, 1150 m [00°02’S, 77°33’W], KU 121445-9; Rio Azuela, 1740 m [00°07’S, 77°37’W], KU 143226-36, 143234 (skeleton), 143550-1 (tadpoles), 143552 (eggs), 143553 (tadpoles), 166217 (tadpoles), 164443; Rio Salado, + 1 km upstream from Rio 38 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Coca, 1410 m [00°13’S, 77°44’W], FMNH 204287, KU 166218—9 (tadpoles), 164444-8, 164449 (skeleton), 178848-50,179404—5 (cleared and stained); 4.7 km N Santa Rosa, 1870 m, KU 143549 (tadpoles); 16.5 km NNE Santa Rosa, 1700 m, KU 143225. Pastaza: Abitagua (8 km NW Mera), 1300 m [01°26’S, 78°07’W], UMMZ 92102. Phyllomedusa lemur COSTA RICA: Alajuela: Cariblanco, 830 m, MCZ 7966; Cinchona, 1600 m, FMNH 172156, KU 35896-7, 63993-8, 68575 (tadpoles); Isla Bonita, c. 1500 m, FMNH 101749, 174823; La Balsa, KU 140015; 4.8 km S Ciudad Quesada, 1000 m, USC-CRE 8077. Cartago: Bridge over Rio Grande at Tapanti, 1201 m, USC-CRE 2640; El Silencio de Sitio Mata, La Suiza, 1200 m, USC-CRE 234; 9.9 km NE (by road) Bridge over Rio Reventazén-Road from Turrialba to Peralta, Rio Chitaria, 775 m, USC-CRE 6141; Moravia de Chirripé, 1116 m, KU 31720 and 31722 (skeletons); Moravia de Turrialba, 1116 m, KU 31700-19 (31703, 31707, 31719, cleared and stained), 31774; Tapanti, 1170-1300 m USC-CRE 6295, 6309(6), MCZ 75353, UMMZ 129202(5), KU 63940-92 (63973, cleared and stained), 68626—9 (cleared and stained). Limdén: Confluence of Rios Lari and Pari about 20.8 km SW Amubri, 440 m, USC-CRE 7178; El Tigre, 9-14 km from Siquirres on Petrolera road to Turrialba, 680 m, KU 70026 (tadpoles), USC-CRE 290-9(3), 290-5 (tad- poles), 290-9(2) (tadpoles), 290-11 (tadpoles). Puntarenas: 3.6 km E Monteverde, 1550-1580 m, USC-CRE 7206; Vicinity Monteverde, Crest of Sierra de Tilaran, 1600 m, UMMZ 126616. San José: Bajo La Hondura, Trail to Rio Claro, 1150-1250 m, USC-CRE 9802-4; Confluence of Rio Claro and Rio La Hondura, 1128-1189 m, USC-CRE 7048(9), 7048(2) (tadpoles), 8081; 1.4 km S Alto La Palma, 1500 m, USC-CRE 7034(9), 7144(7); 1.1 km W La Hondura, 1128 m, USC-CRE 7035(2); La Hondura, 1245 m, MCZ 75351-2; Vicinity La Palma, 1500 m, USC-CRE 490(2), 502(5), 510, 527(4), 6208-11; La Palma, c. 1450 m, KU 63999-64013, 86506, MCZ 7921-2, UMMZ 122666(2), 122667(3), 129201, USNM 75066; Rio Claro on La Palma-Carrillo road, 1188 m, USC-CRE 6311. PANAMA: Bocas del Toro: Rio Changena, 650 m [09°00’N, 82°34’W], KU 101813-6 (101815, cleared and stained); Rio Changena, 830 m [09°00’N, 82°36’W], KU 101817; Rio Changena, 35.2 km W Almirante, 650 m, BYU 19143-4; Rio Claro, near junction with Rio Changena, 910 m [09°00’N, 82°37'W], KU 101818-24 (101818 and 101823, cleared and stained). Chi- riqui: Upper Rio Chiriqui, Fortuna Dam Site, 1000 m [08°40’N, 82°20’W], AMNH 94915-7, 94918 (cleared and stained). Coclé: El Valle, 560 m [08°36’N, 80°08’W] (literature record; Duellman, 1970). Darién: Cerro Mali, 1250 m, USNM 151079; Cerro Mali, 1380 m [08°07’N, 77°13’W], AMNH 90988-9; South Base of Cerro Tacarcuna, Rio Pucro, 640 m_ [08°06’N, 77°15’'W], AMNH 90990; Panamd: Cerro Campana, 800 m, AMNH 84910-1 (metamorphic young); Cerro Campana, 850 m [08°41’N, 79°56’W], KU 77496-8, UMMZ 131091. Phyllomedusa medinai VENEZUELA: Aragua: Estacién Biolégica Rancho Grande, 1100 m [10°22’N, 67°42’W], CAS-SU 20379 (paratype), 21827-8, KU 167186-7, 179107 (cleared and stained), 179406 (metamorphic young), MCZ 65965. Phyllomedusa psilopygion COLOMBIA: Cauca: Quebrada Guangui, Rio Patia, in Rio Saija drainage, c. 100 m, AMNH 87873. Valle: 8 km W Danubio, Rio Anchicay4é, 300 m PHYLLOMEDUSA BUCKLEYI GROUP (ANURA: HYLIDAE) 39 [03°37’N, 76°47’W], KU 169608-12, 169613 (cleared and stained), 170213 (tadpoles), 170214 (eggs), ICN 4755. LITERATURE CITED ANDERSEN, M. L. 1978. The comparative myology and osteology of the carpus and tarsus of selected anurans. Ph.D. Dissertation, University of Kansas pp. 1-301. ANDERSSON, L. G. 1945. Batrachians from east Ecuador collected 1937, 1938 by Wm. Clarke-MacIntyre and Rolf Blomberg. Arkiv Zool. 37A(2): 1-88. : Bacnanra, J. T. 1974. The tail-darkening reaction of phyllomedusine tadpoles (1). J. Exp. Zool. 187(1):149-154. Batu, I. R. 1975. Nature and formulation of biogeographical hypotheses. Syst. Zool. 24:407—430. Bock, W. J., SHear, C. R. 1972. A staining method for gross dissection of vertebrate muscles. Anat. Anz. 130:222—227. BokERMANN, W. C. A. 1966. A new Phyllomedusa from southeastern Brazil. Herpetologica 22:293-297. BouLenceEr, G. A. 1882. Catalogue of the Batrachia Salientia s. Ecaudata in the collection of the British Museum, ed. 2 London, xvi + 503 pp. BouLencer, G. A. 1912. Descriptions of new batrachians from the Andes of South America, preserved in the British Museum. Ann. Mag. Nat. Hist. (8)10:185—191. Conover, W. J. 1971. Practical nonparametric statistics. John Wiley and Sons, Inc., New York, x + 462 pp. © Corre, E. D. 1887. Catalogue of batrachians and reptiles of Central America and Mexico. Bull. U.S. Nat. Mus. 32:1-98. Dixon, W. J. (Ed.) 1975. BMDP Biomedical computer programs. Univ. Calif. Press, xi + 792 pp. DuELLMAN, W. E. 1968. The genera of phyllomedusine frogs (Anura: Hylidae). Univ. Kansas Publ. Mus. Nat. Hist. 18:1—-10. DuELLMAN, W. E. 1969. Phyllomedusa buckleyi Boulenger: Variation, dis- tribution, and synonymy. Herpetologica 25(2):134-140. DuELLMAN, W. E. 1970. The hylid frogs of Middle America. Monog. Mus. Nat. Hist. Univ. Kansas 1:xi + 753 pp. DuELLMAN, W. E. 1972. A review of the Neotropical frogs of the Hyla bogotensis group. Occas. Papers Mus. Nat. Hist. Univ. Kansas 11:1-31. DueLitman, W. E. 1973. Descriptions of new hylid frogs from Colombia and Ecuador. Herpetologica 29(3) :219-227. DuEttman, W. E. 1977. Hylidae, Centrolenidae, Pseudidae. Das Tierreich, Lief. 95, Berlin, xix + 225 pp. Dunn, O. J. 1964. Multiple comparisons using rank sums. Technometrics 6:241-252. Funkuouser, A. 1957. A review of the Neotropical tree-frogs of the genus Phyllomedusa. Occ. Pap. Nat. Hist. Mus. Stanford Univ. 5:1—90. Funxkuouser, A. 1962. A new Phyllomedusa from Venezuela. Copeia 1962 (3) :588-590. Gosner, K. L. 1960. A simplified table for staging anuran embryos and larvae with notes on identification. Herpetologica 16(3):183-190. Gintuer, A. C. L. G. 1885-1902. Reptilia and Batrachia. in Godman, Fred- erick D., and Osbert Salvin. Biologia Centrali-Americana, London, Zoology xx + 326 pp. Harrer, J. 1970. Geologic-climatic history and zoogeographic significance of the Uraba region in northwestern Colombia. Caldasia 10(5):605-636. 40 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY IzEcKsOHN, E., pA Cruz, C. A. G. 1976. Nova especie de Phyllomedusinae do Estado do Espirito Santo, Brasil (Amphibia, Anura, Hylidae). Rev. Brasil. Biol. 36 (1):257-261. Lutz, A., Lurz, B. 1939. Notes on the genus Phyllomedusa Wagler. Ann. Acad. Brasil. Sci. 11(3):219-263. Lutz, B. 1950. Anfibios anuros da colecao Adolpho Lutz. Hylidae in the Adolpho Lutz collection of the Instituto Oswaldo Cruz.’ Mem. Inst. Oswaldo Cruz 11(3) :599-637. Myers, C. W. 1969. The ecological geography of cloud forest in Panama. Amer. Mus. Novitates 2396:1-52. Myers, C. W., Day, J. W. 1976. Preliminary evaluation of skin toxins and vocalizations in taxonomic and evolutionary studies of poison-dart frogs (Dendrobatidae). Bull. Amer. Mus. Nat. Hist. 157(3):173-262. Nrepen, F. 1923. Anura I. Subordo Aglossa und Phaneroglossa, Sectio 1, Arcifera. Das Tierreich, Lief. 46. Berlin, xxxii + 584 pp. Rivero, J. A. 1967. Adiciones recientes a la fauna anfibia de Venezuela. Sep. Me. Soc. Cienc. Nat. La Salle 27(76):5-10. SavaGE, J. M., Heyer, W. R. 1967. Variation and distribution in the tree- frog genus Phyllomedusa, in Costa Rica, Central America. foe Neotrop. Fauna 5:111-131. SavaGcE, J. M., Hever, W. R. 1969. The tree-frogs (family Hylidae) of Costa Rica: Diagnosis and distribution. Rev. Biol. Trop. 16(1):1-127. ScHULTE, R. 1977. Der lemurenfrosch (Phyllomedusa lemur). Aquar. Mag. Stuttgart 11(3):98-103. Simpson, B. B. 1975. Pleistocene changes in the flora of the high tropical Andes. Paleobiology 1(3):273-294. Starrett, P. H. 1968. The phylogenetic significance of the jaw musculature in anuran amphibians. Ph.D. Dissertation, Univ. Michigan 179 pp. Taytor, E. H. 1952. A review of the frogs and toads of Costa Rica. Univ. Kansas Sci. Bull. 35(1):577-942. Test, F. H., Sexton, O. J., HeEArwoie, H. 1966. Reptiles of Rancho Grande and vicinity, Estado Aragua, Venezuela. Misc. Publ. Mus. Zool. Univ. Michigan 128:1-63. Trues, L. 1971. Phylogenetic relationships of certain neotropical toads with the description of a new genus (Bufonidae). Los Ang. Co. Mus. Cont. Sci. 216:1—40. Trues, L. 1973. Bones, frogs, and evolution. in Vial, J. L. (ed.). Evolution- ary biology of the anurans: Contemporary research on major problems. Univ. Missouri Press, Columbia: 65-132. Trues, L. 1974. Systematic relationships of Neotropical horned frogs, genus Hemiphractus (Anura: Hylidae). Occas. Papers Mus. Nat. Hist. Univ. Kansas (29):1-60. : Trues, L. 1977. Osteology and anuran systematics: Intrapopulational variation in Hyla lanciformis. Syst. Zool. 26:165-184. Tyter, M. J. 1971. The phylogenetic significance of vocal sac structure in hylid frogs. Univ. Kansas Pub]. Mus. Nat. Hist. 19:319-360. Tycer, M. J., Davies, M. 1978. Phylogenetic relationships of Australian hyline and Neotropical phyllomedusine frogs of the family Hylidae. Herpetologica 34(2):219-224. Van Dern Hammen, T. 1974. The Pleistocene changes of vegetation and climate in tropical South America. J. Biogeography 1974 (1):3-26. Van Esjspen, E. H. T. 1977. Notities over Phyllomedusa lemur. Lacerta 35(12):175-181. ra) if iad ve ‘ - tag a> => ot . a *) ie P i" ae i a , ; uw a ee SIP) eee ' Vest. qnd ca: >. aw «4 at ait + Me 7. (ee +. cage wae ie La : ee ‘cars = ry if ‘<= ; 7 - ry var | . * » - . % ‘ - has nll re LP 2 eee rar : aed. Te pay eee O22 spt P ailenemels hewn. 7 =a . Wel +a & UNIVERSITY OF KANSAS PUBLICATIONS MUSEUM OF NATURAL HISTORY The University of Kansas Publications, Museum of Natural History, beginning with volume 1 in 1946, was discontinued with volume 20 in 1971. Shorter research papers formerly pub- lished in the above series are now published as Occasional Papers, Museum of Natural History. The Miscellaneous Publica- tions, Museum of Natural History, began with number 1 in 1946. Longer research papers are published in that series. Monographs of the Museum of Natural History were initiated in 1970. All manuscripts are subject to critical review by intra- and extramural specialists; final acceptance is at the discretion of the publications committee. — Institutional libraries interested in exchanging publications may obtain the Occasional Papers and Miscellaneous Publica- tions by addressing the Exchange Librarian, The University of Kansas Library, Lawrence, Kansas 66045. Individuals may pur- chase separate numbers of all series. Prices may be obtained upon request addressed to Publications Secretary, Museum of Natural History, The University of Kansas, Lawrence, Kansas 66045. Editor: Ep WiLry Managing Editor: JosepH T. COLLINS PRINTED BY UNIVERSITY OF KANSAS PRINTING SERVICE LAWRENCE, KANSAS Dr Y eo Fee: OCCASIONAL PAPERS JAN 2.5 1981 of the HARVARD UNIVERSITY MUSEUM OF NATURAL HISTORY The University of Kansas Lawrence, Kansas NUMBER 88, PAGES 1-9 JANUARY 14, 1981 THREE NEW SPECIES OF CENTROLENID FROGS FROM THE PACIFIC VERSANT OF ECUADOR AND COLOMBIA by WILLIAM E.. DUELLMAN' Field work on the forested Pacific slopes of Ecuador and Co- lombia in 1975 resulted in large collections of centrolenid frogs, including some unnamed species of Centrolenella. The perpetually humid forested slopes of the Andes between 1000 and 2000 m are inhabited by many species of Centrolenella. Some have elevational distributions of nearly 1000 m; others are restricted to narrow ele- vational zones. A few species have rather broad geographic distri- butions, whereas others are known from only one valley. A change of a few hundred meters in elevation or to a different drainage sys- tem results in a different centrolenid fauna. As many as five species of Centrolenella can be found along a single stream on one night. Several taxonomic problems still exist with the centrolenid frogs on the Pacific versant of Ecuador and Colombia, but herein I am concerned only with three highly distinctive species in the Centro- lenella prosoblepon group as defined by Starrett and Savage (1973). My own collections have been augmented by specimens obtained by subsequent collectors. In the following descriptions, the format for the diagnoses and descriptions is the same as that used by Lynch and Duellman (1973). Centrolenella balionota new species (Fig. 1) Holotype —KU 164702, an adult male, from 3.5 km (by road) 1 Curator, Division of Herpetology, Museum of Natural History, and Pro- fessor, Department of Systematics and Ecology, The University of Kansas, Lawrence, Kansas 66045. 2 OCCASIOIAL PAPERS MUSEUM OF NATURAL HISTORY northeast of Mindo, 1540 m, Provicia de Pichincha, Ecuador (00° 01’S, 78°44’W), one of a series collected on 7 April 1975 by Wil- liam E. Duellman, Alan H. Savitzky, and John E. Simmons. Paratopotypes—KU 164701, 164703-13, 7-8 April 1975, same collectors. Referred specimen.—KU 145084 from La Costa, 800 m, De- partamento Cauca, Colombia. Diagnosis——1) prevomerine teeth absent; 2) bones pale green; 3) parietal peritoneum white; visceral peritoneum clear; 4) color in life pale green with reddish brown stripes and flecks and elevated yellow spots; in preservative, cream with reddish brown markings and white spots; 5) webbing between outer fingers III 2—2 IV; 6) webbing on foot I 1s—2 II 14—1 III 1s—3 IV 3—1% V; 7) snout truncate in dorsal and lateral profiles; 8) dorsal skin smooth; 9) arms and legs lacking dermal folds; 10) humeral spine present in males; 11) lower four-fifths of tympanum visible, directed postero- laterally with slight dorsal inclination. : Color pattern alone distinguishes C. balionota from all other known Centrolenella. In other species having yellow or gold spots, the spots are either small (C. flavopunctata, midas, siren), ocelli (C. anomala, cochranae, ocellata, ocellifera), or in combination with dark flecks (C. peristicta, pipilata, some prosoblepon). The only other species having red markings is C. grandisonae, a much larger species with small red spots and granular skin on the dorsum. Description.—Adults small; snout-vent length 20.5-22.5 mm (x = 21.1, n = 18) in males; females not known, head much wider Fic. 1. Centrolenella balionota, KU 164702, 4, 21.2 mm snout-vent length. THREE NEW SPECIES OF CENTROLENID FROGS 3 than body; width of head 32.9-34.8 percent (x = 33.8, n = 18) of snout-vent length. Snout short, high, truncate in dorsal and lateral profiles; canthus round; loreal region deeply concave; lips flared; nostrils about five-sixths distance from eye to tip of snout, slightly protuberant laterally; internarial area slightly depressed. Eye large, protuberant, directed anterolaterally. Supratympanic fold covering upper one-fifth of tympanum; tympanic annulus distinct; tympanum strongly directed posterolaterally, inclined dorsally. Prevomerine teeth and dentigerous processes absent; choanae small, triangular, widely separated medially; tongue ovoid, barely free behind, shal- lowly notched posteriorly; vocal slits extending from posterolateral base of tongue to angles of jaws; vocal sac, single, median, subgular. Humeral spine, short, blunt, tip not free of skin on upper arm; ulnar folds and tubercles absent; first finger slightly longer than second; fourth finger noticeably shorter than third; lateral fringes on fingers; webbing rudimentary between first and second fingers; webbing formula for other fingers II 2—3 III (2-2%)—2 IV; web- bing emarginate; discs round; subarticular tubercles small, round, simple; supernumerary tubercles small, indistinct, present on prox- imal segments of all fingers; palmar tubercle ovoid, simple; nuptial excrescences absent. Hind limbs moderately slender; length of tibia 54.7-59.0 percent (* = 56.8, n = 18) of snout-vent length; tarsal folds and tubercles absent; inner metatarsal tubercle small, flat, elliptical; outer metatarsal tubercle small, ovoid; subarticular tubercles small, round; supernumerary tubercles absent; feet about three-fourths webbed; webbing formula I 14—2 II 1%—(1-1%) Il (14-2) —( 24-3) IV (24-3)—1% V; discs round, slightly smaller than those on fingers. Skin on belly and proximal ventral surfaces of thighs weakly granular; skin on other surfaces smooth, elevated on dorsum at sites of white spots. Anal opening directed posteriorly at upper level of thighs; pair of small, flat tubercles below anus. Color in preservative: dorsum cream with reddish brown flecks on head, body, forearms, thighs, shanks, and feet; reddish brown interorbital marks not widely separated medially; reddish brown postorbital stripe extending posteromedially to postscapular area. Large, elevated white spot on anteromedial edge of each eyelid; white spots on dorsum of body in ten specimens—six with one spot, four with two spots; digits and ventral surfaces creamy white. Color in life: dorsum pale green with reddish brown stripes and flecks; tips of digits dull yellow; elevated dorsal spots bright yellow; vocal sac bluish green; visceral peritoneum clear; parietal peri- toneum white; heart not visible; bones pale green; iris iridescent golden copper. Distribution—The type series of C. balionota is from a small cascading stream at an elevation of 1540 m on the Pacific slopes of 4 OCCASIOIAL PAPERS MUSEUM OF NATURAL HISTORY the Andes; the stream, in disturbed lower montane rainforest, is a tributary of the Rio Mindo. The species also is known from La Costa, 800 m, Departamento de Cauca, Colombia; this locality is about 330 km airline north-northeast of the type locality. Remarks.—All individuals were calling from the upper surfaces of leaves of herbs and ferns overhanging a trickling stream; none was more than 1 m above the water. Other stream-breeding frogs at the type locality include Centrolenella grandisonae, C. griffithsi, and Hyla alytolylax. The call is a single “peep.’ Etymology.—The specific name is derived from the Greek balios meaning dappled or spotted and the Greek notos meaning back; it is used in allusion to the dorsal reddish brown dappling and yellow spots. Centrolenella heloderma new species (Fig. 2) Holotype—KU 164715, an adult male, from Quebrada Zapa- dores, 5 km east-southeast of Chiriboga, 2010 m, Provincia de Pichincha, Ecuador (00°17’S, 78°47’W), obtained on 4 April 1975 by William E. Duellman. Paratopotypes——KU 164714, 3 April 1979, William E. Duellman; MCZ 97835, 15 March 1979, Kenneth Miyata; USNM 211219-21, 15 March 1979, Roy W. McDiarmid. Referred specimens.—All Provincia de Pichincha, Ecuador: 14 km W Chiriboga, 1960 m, KU 164716-21; 13.1 km NW Nono, 2140 m, MCZ 97834, USNM 211216-17; 8.6 km SE Tandayapa, 2000 m, USNM 211218; 9 km SE Tandayapa, 2150 m, KU 167422-24. Diagnosis.—1) prevomerine teeth absent; 2) bones green; 3) parietal peritoneum white; visceral peritoneum clear; 4) color in life dark green with bluish white tubercles; in preservative laven- der; 5) webbing between outer fingers III 24—2 IV; 6) webbing on foot I 1—2 II 1—2 III 1—2 IV 2—1.V; 7) snout rounded in dorsal view, in profile depressed and sloping anterior to nares; 8) dorsal skin tubercular; 9) forearms and feet bearing narrow, scal- loped dermal folds; 10) humeral spine present in males; 11) entire tympanum visible, directed laterally with slight posterodorsal in- clination. The tubercular dorsal skin immediately distinguishes C. helo- derma from other known species of Centrolenella, which have either smooth, shagreened, or spiculate skin dorsally. Centrolenella helo- derma mostly closely resembles C. buckleyi, which differs by hav- ing fine spinules in the dorsal skin, less webbing on the hand and foot, no dermal folds on limbs, and only the lower part of the tympanum visible. Both species have humeral spines in males, the snout inclined anteriorly, and a narrow white or cream labial stripe. Description —Adults moderately large; snout-vent length 26.8— THREE NEW SPECIES OF CENTROLENID FROGS 5 Fic. 2. een tae Retadserii KU 164715, ¢, 29.0 mm snout-vent length. 31.5 mm (x = 29.0, n = 17) in males, 32.3 mm in single female; head slightly wider than body; width of head 33.2-36.8 percent (x — 34.1, n = 18) of snout-vent length. Snout moderately long, depressed, inclined anteriorly from nostrils to margin of lip, rounded in dorsal view; canthus round; loreal region barely concave; lips flared; nostrils about three-fifths distance from eye to tip of snout, barely protuberant laterally; internarial area slightly depressed. Eye moderately small, directed more laterally than anteriorly. Tym- panic annulus interrupted by tubercles dorsally, otherwise distinct; tympanum directed laterally with slight posterodorsal inclination. Prevomerine teeth and dentigerous processes absent; choanae small, round, widely separated medially; tongue broadly cordiform, deeply notched posteriorly, free behind for about half of its length; vocal slits extending from midlateral base of tongue to angles of jaws; vocal sac single, median, subgular, large, when inflated extending laterally well beyond sides of head. Humeral spine long, blunt, projecting anteriorly at angle of about 45° from humerus; forearm robust; low, scalloped dermal fold ventrolaterally on forearm; first and second fingers subequal in length; fourth finger longer than second; lateral fringes on fingers; webbing rudimentary between first and second fingers; webbing formula for other fingers II 2—3% IIT (2-2 )—( 14-2) IV; webbing emarginate; discs truncate; subarticular tubercles small, round, sim- ple; supernumerary tubercles absent; palmar tubercle low, indis- tinct; nuptial excrescences absent. Hind limbs moderately robust; length of tibia 53.0-58.2 percent (* == 55.4, n = 18) of snout-vent 6 OCCASIOIAL PAPERS MUSEUM OF NATURAL HISTORY length; row of tubercles on outer edge of tarsus; inner metatarsal fold absent; inner metatarsal tubercle low, elongate; outer meta- tarsal tubercle small, conical; subarticular tubercles small, round; supernumerary tubercles absent; feet about four-fifths webbed; webbing formula I 1—2 II 1—2 III 1—2 IV 2-~(1-1%) V; dises truncate, nearly as large as those on fingers. Skin on dorsum bearing large, round tubercles; skin on flanks, belly, and proximal posteroventral surfaces of thighs coarsely gran- ular; skin on other surfaces smooth. Anal opening directed pos- teriorly at upper level of thighs; pair of large round tubercles and many small tubercles below anal opening. Color in preservative: dorsal surfaces of head, body, forearms, thighs, and shanks dull lavender; other surfaces dull cream; ventral and lateral coloration meeting along a distinct line on upper flanks; anal tubercles and margin of upper lip white. Color in life: dorsum dark green with bluish white tubercles; margin of lip yellow; parietal peritoneum and throat pale golden yellow; anal, heel, and ulnar tubercles white; heart not visible; bones green; iris pale bronze with fine black reticulations. Distribution —Centrolenella heloderma is known from several localities near the upper limits of cloud forest (1960-2150 m) on the Pacific slopes of the Cordillera Occidental in Provincia de Pichincha, Ecuador. Remarks.—Males were calling from upper surfaces of leaves and ferns 1-4 m above streams and on cliff faces below seepages. The call is a harsh “peep.” A juvenile (KU 164714) having a snout-vent length of 20.9 mm has the dorsal tubercles as well developed as in the adults. A single female (MCZ 97835) having a snout-vent length of 32.3 mm is like the males in structure and coloration. At all localities where C. heloderma was found, C. griffithsi occurred in much greater abundance. At higher elevations ( >2300 m) on the Pacific slopes, in subparamo and paramo, C. heloderma seemingly is replaced by the widespread Andean species, C. buck- leyi. The type locality, Quebrada Zapadores, is a tributary of the Rio Saloya and is crossed by a bridge at Kilometer Post 45 on the road from Chillogallo to Santo Domingo de los Colorados via Chiriboga. Although the valley of the Rio Saloya is mostly pasture, cloud for- est exists in the narrow Quebrada Zapadores containing a stream 2-4 m wide. Centrolenella prasina new species Holotype —KU 169693, an adult male, from Rio Calima, 1.5 km (by road) west of Lago Calima, 1230 m, Departamento de Valle, THREE NEW SPECIES OF CENTROLENID FROGS i Colombia (4°00’N, 76°35’W ), obtained on 1 June 1975 by William E. Duellman. Paratopotypes.—KU 169691-92, 14 September 1974, William E. Duellman and Linda Trueb. Diagnosis.—1) prevomerine teeth 5-7; 2) bones pale green; 3) parietal peritoneum white; visceral peritoneum clear; 4) color in life uniform bright green; in preservative, uniform lavender; 5) webbing between outer fingers II 24—3 III 24—2 IV; 6) webbing on foot I 1%—2 II 1—2 III 1—2% IV 24%—1% V; 7) snout round in dorsal view, truncate in profile; 8) dorsal skin smooth; 9) arms and legs lacking dermal folds; 10) humeral spines absent in males; 11) tympanum concealed. The combination of large size (34.5 mm), smooth dorsal skin, uniform green dorsum, and absence of humeral spines in males dis- tinguishes C. prasina from all other known Centrolenella. The size is approached by C. megacheira (32.8 mm), which has small black spots on the dorsum and pustular skin. Other uniform green Cen- trolenella in western South America have humeral spines in males (C. buckleyi, C. heloderma, some C. prosoblepon) or have spiculate dorsal skin (C. spiculata). ‘ Description.—Adults large; snout—vent length 33.0-34.5 mm (x = 33.7, n = 3) in males; females unknown; head wider than body; width of head 31.6-31.8 percent ( = 31.7, n = 3) of snout-vent Fic. 3. Centrolenella prasina, KU 169693, 6, 33.6 mm snout-vent length. 8 OCCASIOIAL PAPERS MUSEUM OF NATURAL HISTORY length. Snout short, high, round in dorsal view, truncate in profile; canthus rounded; loreal region shallowly concave; lips flared; nos- trils about four-fifths distance from eye to tip of snout, distinctly protuberant laterally; internarial region depressed. Eye large, pro- tuberant, directed anterolaterally; tympanum concealed (lowermost edge of tympanic annulus evident in one specimen). Prevomerine teeth 5-6—7-7 (total teeth 11-14, x = 12.5, n = 2; one specimen lacking dentigerous processes and prevomerine teeth); dentigerous processes transverse between large, ovoid choanae, narrowly sepa- rated medially; tongue broadly cordiform, narrowly notched behind, barely free posteriorly; vocal slits extending from midlateral base of tongue to angles of jaws; vocal sac single, median, subgular. Humeral spine absent; forearm slender; ulnar folds and tuber- cles absent; first finger slightly longer than second; fourth finger much longer than second; lateral fringes on fingers; webbing absent between first two fingers; webbing formula for other fingers II (24— 22)—3% II] 2%—(2-24) IV; webbing emarginate; discs truncate; subarticular tubercles small, round simple; supernumerary tuber- cles in single row on proximal segments of each digit; palmar tuber- cle low, bifid; nuptial excresences absent. Hind limbs slender; length of tibia 58.0-60.0 percent (* = 59.6, n = 3) of snout-vent length; tarsal folds and tubercles absent; inner metatarsal tubercle elongate, elliptical; outer metatarsal tubercle absent; toes about two-thirds webbed; webbing formula I 14—2 II 1—2 III 1-14)—2% IV 24—1z V; discs round, noticeably smaller than those on fingers. Skin on dorsum and most ventral surfaces of limbs smooth; skin on belly and proximal ventral surfaces of thighs coarsely granular, on flanks and in tympanic region weakly granular. Anal opening directed posteriorly at upper level of thighs, devoid of tubercles and folds. Color in preservative: dorsal surfaces of head, body, and limbs deep lavender; other surfaces cream. Color in life: dorsum uniform dark green; margin of upper lip pale green to greenish white; flanks cream; hands, feet, and thighs yellowish green; throat pale green; pericardium white; heart not visible; bones green; iris greenish white with black reticulations. Distribution —This large species is known only from the type locality. Remarks.—All individuals were calling at night from the upper surfaces of leaves 1.5-2 m above a small (+2 m wide) stream in cloud forest, about 100-200 m upstream from its confluence with the Rio Calima. Lago Calima is just north of the Buga-Buena- ventura Highway, a road descends from the highway to the dam on the Rio Calima. The type locality is reached by going westward for 1.5 km on a narrow road that descends from the south edge of THREE NEW SPECIES OF CENTROLENID FROGS 9 the dam. The tributary is north (across the Rio Calima) from the road. Etymology.—The specific name is Latin, meaning leek green. ACKNOWLEDGMENTS I am grateful to my field companions—Alan H. Savitzky, John E. Simmons, and Linda Trueb—for their efforts in securing speci- mens of elusive centrolenids and to Roy W. McDiarmid, National Museum of Natural History (USNM), and Kenneth Miyata, Mu- seum of Comparative Zoology, Harvard University (MCZ) for mak- ing available specimens collected by them. The University of Kan- sas Museum of Natural History is abbreviated KU. Field work was supported by grants from the National Geographic Society (1304) and the National Science Foundation (DEB 74-01998). Stephen C. Ayala, Universidad del Valle in Cali, Colombia, and Eugenia del Pino, Universidad Catdlica del Ecuador in Quito, provided logistic support, and collecting permits for Colombia were issued by Jorge Hernandez C. of INDERENA. RESUMEN Tres especies nuevas del grupo prosoblepon del género Centro- lenella se nombran de las laderas pacificas del Ecuador y Colombia. Centrolenella balionota es una especie pequefia con manchas rojas y puntas amarillas en el dorso; se conocido de una quebrada a una elevacién de 1540 m en Ecuador y de una elevacién de 800 m en el Departamento de Cauca, Colombia. Centrolenella heloderma, conocido de elevaciones de 1960-2150 m en Ecuador, tiene piel tu- berculado en el dorso. Centrolenella prasina es una especie grande conocido solamente del Rio Calima (1230 m) en el Departamento de Valle, Colombia. LITERATURE CITED Lyncu, J. D., DuELLMAN, W. E. 1973. A review of the centrolenid frogs of Ecuador, with descriptions of new species. Univ. Kansas Mus. Nat. Hist. Occas. Pap. (16):1-66. STARRETT, P. H., SAvAGE, J. M. 1973. The systematic status and distribution of Costa Rican glass-frogs, genus Centrolenella (family Centrolenidae ), with description of a new species. South. California Acad. Sci. Bull. 72:57-78. fe : re , rh bet e5 ‘ is aAc5 * = | > , re Tee atl bee ee > Ds baad he, ee wis | he th, PD WwAhi megs - eG leery s r i *, a i Aree Wes) uli ' a ; “Feu ena 5 : 1s fi i S | me Peer ile yee & ay 49 ; Ti ] 492 J nie hes , : 7 ; ge a) : ye j Le f 7 ea ie im 7 ae 7 ; (ra Fs 4 iy Au i i ws tere) ei a ae (re | . Be e , mn as ot Nat pring don ob | ohn! pill i ping KE Ae om - - Sf wy ' ¢ ay Mii dew: abe Pig sae! a | 7 wine ~ ‘aPo? wie —" ay rg’ oy lat ile Ga ony ey ease UNIVERSITY OF KANSAS PUBLICATIONS MUSEUM OF NATURAL HISTORY The University of Kansas Publications, Museum of Natural History, beginning with volume 1 in 1946, was discontinued with volume 20 in 1971. Shorter research papers formerly pub- lished in the above series are now published as Occasional Papers, Museum of Natural History. The Miscellaneous Publi- cations, Museum of Natural History, began with number 1 in 1946. Longer research papers are published in that series. Monographs of the Museum of. Natural History were initiated in 1970. All manuscripts are subject to critical review by intra- and extramural specialists; final acceptance is at the discretion of the publications committee. Institutional libraries interested in exchanging publications may obtain the Occasional Papers and Miscellaneous Publica- tions by addressing the Exchange Librarian, The University of Kansas Library, Lawrence, Kansas 66045. Individuals may pur- chase separate numbers of all series. Prices may be obtained upon request addressed to Publications Secretary, Museum of Natural History, The University of Kansas, Lawrence, Kansas 66045. Editor: Ep Wi1LEY Managing Editor: JoseEpH T. CoLLins PRINTED BY UNIVERSITY OF KANSAS PRINTING SERVICE LAWRENCE, KANSAS Gaal 3 ( ec DD MUS COMP : n Z a J LIBRA ae OCCASIONAL PAPERS JAN 25 198) of the Pane e MUSEUM OF NATURAL HISTOR¥=~y The University of Kansas Lawrence, Kansas NUMBER 89, PAGES 1-10 JANUARY 20, 1981 REDESCRIPTION OF ETHEOSTOMA AUSTRALE AND A KEY FOR THE IDENTIFICATION OF MEXICAN ETHEOSTOMA (PERCIDAE ) By LAWRENCE M. Pace’ Since the original description of the Conchos darter in 1859 as Diplesion fasciatus Girard and redescription in 1889 as Etheostoma australe Jordan (fasciatus is preoccupied in Etheostoma—Collette and Knapp, 1966), little additional information on the species has been published. Jordan and Evermann (1896) and Meek (1904) expanded somewhat the earlier description by Jordan (1889). Bailey and Gosline (1955) provided vertebral counts. Collette (1965) noted an absence of breeding tubercles on specimens available for examination. Moore (1968) and Alvarez (1950, 1970) included E. australe in identification keys. Page (1977) discussed characteristics of the lateralis system. On 24 February 1979, Mr. Craig W. Ronto and I collected a se- ries of 26 E. australe (INHS 84082) in Rio Santa Isabel at General Trias, Chihuahua. Males were in bright breeding colors and tuber- culate, females were ripe, and spawning was in progress or about to ensue. Much of the following description, including the first accu- rate information on color, is based on this series of specimens. Etheostoma australe Jordan Conchos Darter Material examined.—Rivers are listed in a north-south arrange- ment. Acronyms are identified in ACKNOWLEDGMENTS. Num- bers in parentheses are numbers of specimens on which counts were 1 Taxonomist, Illinois Natural History Survey, Urbana, Illinois 61801, U.S.A. and Associate in Ichthyology, Museum of Natural History, The University of Kansas, Lawrence, Kansas 66045, U.S.A. 2 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY made. Complete collection locality data are available from the author. Rio Santa Isabel: UMMZ 182378 (10), INHS 84082 (10); Rio San Pedro: UMMZ 136124 (5), UMMZ 203013 (15); Rio Parral: TNHC 4033 (10), FMNH 3538; Rio Florido: KU 8407 (20), FMNH 3554 (4), FMNH 3527 (6). Diagnosis —Member of genus Etheostoma. One large and stiff anal spine, premaxillary frenum present, incomplete lateral line (with 23-50 pored scales), nape and breast unscaled (or nape with few scales posteriorly). Other species of Etheostoma with one anal spine (or one or two anal spines in about equal frequencies) are without a premaxillary frenum (E. nigrum, E. olmstedi, E. vitreum, E. chlorosomum), have a complete lateral line (E. trisella), a fully scaled nape and breast (E. tuscumbia), or fewer than 10 pored lateral-line scales (E. proeliare, E. fonticola, E. microperca). Description.—Nape typically unscaled but sometimes with few scales posteriorly; opercle, cheek, breast, and prepectoral area un- scaled; belly fully scaled or unscaled anteriorly. Infraorbital canal usually narrowly interrupted with four to six (usually five) pores in anterior segment and two to five (usually three or four) pores in posterior segment; sometimes uninterrupted. Supratemporal canal interrupted or uninterrupted, usually interrupted with two pores in each segment. Preoperculomandibular canal with 10-12 pores. Meristic counts were made as described by Hubbs and Lagler (1964). Lateral (lateral line) scales 51-71 (modally 55-65) (Table 1), 20-53 (35-45) pored (Table 2;) scales above lateral line 5-8; scales below lateral line 7-10; transverse (anal fin origin to first dorsal fin) scales 13-20 (Table 3); scales around caudal peduncle 24-32 (Table 3); dorsal spines 10-14 (11-12) (Table 4); dorsal rays 9-12 (10-11) (Table 4); pectoral rays 10-12 (11) (Table 5); anal spine 1; anal rays 6-9 (7) (Table 5). Breeding male red-brown with 8-11 wide, dark green vertical bars evenly spaced along side of body (Fig. 1). On some males, interspaces between bars red-brown; on others, yellow with large red blotches. Green vertical bars on side fuse into dark green sad- dles along mid-dorsum. Cheek, branchiostegal membranes, breast, and belly green. First dorsal fin red with narrow dark blue margin and blue base. Second dorsal, caudal, and pectoral fins yellow, with tinges of red on membranes. Anal and pelvic fins green basally and yellow distally. Two vertically aligned red-brown spots at base of caudal fin. Prominent, large, black suborbital bar and black post- orbital spot present. Preorbital bar black but diffuse. Female is green dorsally and yellow ventrally. Green vertical bars on side of body less prominent than on males and blend into green dorsum. First dorsal fin yellow with narrow blue marginal band, narrow red medial band, and blue base. Other fins yellow; REDESCRIPTION OF ETHEOSTOMA AUSTRALE 9°€ es OL QLT 9 ‘TI 06 OL 6S 9°S AO S60 08'S v6'T COFr 19'€ 6r'E ds £96 6'LE SLE 4 a8 + F L'6 99T OF CG 2l. €2CI-G. Sean cua: 'S VA ok ie), Sis On oe Se iG" AG GS SP 1 69 3861 “61 02 SEG. Ome ae [1S C ORO aes ae ei 9 Lg & GSa xa OG 61 8T LI 9T ST #1 ET S[BOS eSIaAUReIT, ‘ON Ge Siew? 7 AT | G6 Rese T i sal 6 Dale oe alee oe ic = One I IE I I (SONG (SS € ot TL 04 69 89 149 99 G9 79 ¢9 69 19 09 6S 8S 2G 9¢ GG Fe €S oS I¢ Se[eos “ON ‘2/D4ISND DuLozsoayyy UL saTeos [e19}e] JO sjuno— | atavy, OpHopy ory [ESE Oy Olpeg ueg ony [2qes] BUR ony J9Any OpHoTa ony [eueg ony OIpeg ueg org [2qes] ejUeS ony OpHol Ory [ence LL S| O1peg ueg ony [eqes] eJueS ong JOAN OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 4 18 Lg 8S c'9 AO $80 L9°0 190 69'0 ds sAvy Of 8 61 Or 6 9 0G rT eI tt ‘ul [euy “ON O|HON iE sero P of S870). ELL. OL S ch Loe arO” “CTL 06 y «OL Oe aU! 1G. Lb. ale c. _k1 AO ds x N cI_It Ol skey ulq [eloped “ON ‘a]D4ysnD DUOISOBYIT Ut sXevi uy [eue pue [e1oqoad jo syun0QN—g ATAVL COL of Ce eon’ oor ol G Oo 16 GOL 06 Tea 9'OT I x N Choi OleeS skey uly “ON *a]D44SND puojsoayyy Ut sAei pue sourds ug [esiop jo sjun0j— v9 oF VG 6'9 CLO) alt soo. 2 bk Ol Bate c90 OCI 06 Y “cl © 6g0 Trl 06 fink ORs& ds x N eI itl p a1av Ol opuopy OM jeued Oy [ees] Reyes ory o1ped ues org JOATY opHoja Ofd jeued OW orped URS OY jeqesy BuRS OY JOATY REDESCRIPTION OF ETHEOSTOMA AUSTRALE 5 Fic. 1—Etheostoma australe from Rio Santa Isabel, General Trias, Chi- huahua, 24 Feb. 1979. Upper, 42-mm male. Lower, 41-mm female. second dorsal and caudal fins with vague bands of small black spots. Three vertically aligned black spots at base of caudal fin. Breeding males (INHS 84082) with tubercles along entire length of rays and sometimes spines of anal and pelvic fins; and on scales of belly, along base of anal fin, and lower caudal peduncle. Less often tubercles on middle rays of pectoral fins and lower rays of caudal fin. One of five females captured on 24 February 1979 with tubercles on anal rays. Largest specimen examined is a female 50.5 mm in standard length; largest male is 49.7 mm (both INHS 84082). Available specimens are too few (and many are old and poorly preserved) for a thorough analysis of geographic variation in E. australe. Among specimens examined, scale counts and counts of dorsal spines and rays were highest in specimens from Rio San Pedro and its tributary, Rio Santa Isabel, intermediate in those from Rio Florido, and lowest in those from Rio Parral (Tables 1-4). This variation does not correspond to north-south or east-west arrange- ments of the data. Scale counts in Campostoma ornatum from Rio Santa Isabel also were the highest of any population examined (Burr 1976). Presumably information not now available on the drainage history of Rio Conchos would explain geographic variation in its fishes but any further analysis of variation in E. australe must await the accumulation of more museum specimens. 6 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Distribution and Habitat—Etheostoma australe is restricted to Rio Conchos, a major tributary of Rio Grande, in Chihuahua and Durango (Fig. 2). On 24 February 1979, E. australe was captured in a fast rubble rifle. Males were much more common (21 of 26 individuals captured) and may have established breeding stations in the riffles. Seining in pools above and below the riffles failed to yield any E. australe. Contreras-Balderas (1977) described E. australe_as a benthic insectivore that is rare overall but common locally and lives on sand or gravel in shallow, fast, cool streams. Late May and June was given as the spawning period by Meek (1904). Specimens collected in 1979 indicate spawning may begin as early as late February. RELATIONSHIPS AND OTHER MEXICAN DARTERS The presence of only one spine in the anal fin persuaded Jordan and Evermann (1896) to erect a subgenus (Torrentaria) for E. Fic. 2.—Distribution of localities from which Etheostoma australe has been collected. REDESCRIPTION OF ETHEOSTOMA AUSTRALE a australe and E. sagitta (the latter was erroneously thought to have only one spine). Jordan, Evermann and Clark (1930) treated Tor- rentaria as a genus. Hubbs (1936) noted that Torrentaria was pre- occupied in Aves and proposed Austroperca as a substitute sub- generic name, and in doing so restricted Austroperca to include only E. australe. Bailey & Gosline (1955), in their list of darter subgenera, also considered Austroperca to be a monotypic subgenus of Etheostoma. The presence of one and not two anal spines is unusual among darters; only nine of 128 described species have one spine (or one or two in about equal frequencies). However, in the subgenera Boleosoma and Vaillantia some species have one and others have two anal spines, and in E. vitreum, E. proeliare, and E. microperca some individuals have one and some have two anal spines. Because of this plasticity, the presence of one anal spine seems inadequate to warrant-subgeneric status for Austroperca. Consideration of many characteristics among all darters indicates that E. australe is most closely related to species in the subgenus Oligocephalus and easily can be placed in that subgenus (Page, in press). The only other described species of Etheostoma in Mexico are E. grahami, also endemic to the Rio Grande system but known only from tributaries downstream from Rio Conchos, and E. pottsi, which occurs in Rio Conchos (rarely syntopically with E. australe ), Rio Nazas, Rio Trujillo (upper part of Rio Aguanaval—see Conant, 1963: Fig. 1), and Rio Mezquital; the last drains into the Pacific Ocean. Percina macrolepida, the only other darter in Mexico, is known in Mexico only from Rio San Carlos, Coahuila (Stevenson, 1971). The following key is to the three described species of Etheo- stoma in Mexico and to E. lepidum, a species endemic to Texas and New Mexico and often confused with E. grahami. Although several references have been made to undescribed species of Etheostoma in Mexico (Minckley, 1969, 1977; Deacon, et al., 1979) no data have been presented on them and their validity remains to be demonstrated. 1. One anal spine; branchiostegal membranes rather broadly connected across isthmus; broad vertical dark bars on side of |550)6 bydesiceaelba areeceieys A IaAN lo Alley tee oh de OR aM: E. australe Two anal spines; branchiostegal membranes separate or nar- rowly connected across isthmus; side of body without broad SUGTEL KOSI GF tab Bae te en PM Weare oe a SOS me ee Ran 2 2. Usually 48 or more lateral scales; 21 or more scales around cau- dal peduncle; first dorsal fin with blue-green marginal band and red medial band (most distinct in the male) E. lepidum Usually 40-50 lateral scales; usually 17-20 scales around cau- dal peduncle; first dorsal fin without distinct bands (dusky 8 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Wed)! 23 xGert ple eres ee bet 3 A aes hc ee 3 3. Opercle fully or mostly scaled; branchiostegal membranes nar- rowly joined across isthmus; usually 9 or 10 dorsal spines ih, lpr ipeiiilis Sa oad Eee 8 dh on sole Phe Cale ee eee _ E. grahami Opercle unscaled; branchiostegal membranes separate, not joined across isthmus; usually 10-12 dorsal spines —-- E. pottsi ACKNOWLEDGMENTS I wish to thank R. R. Miller, University of Michigan Museum of Zoology (UMMZ), for various information on Mexican darters, P. W. Smith, Illinois Natural History Survey (INHS), and B. M. Burr, Southern Illinois University at Carbondale, for suggestions on improving the study and manuscript, and the following individuals and their institutions for loans of specimens: F. B. Cross and J. T. Collins, The University of Kansas Museum of Natural History (KU); R. K. Johnson and B. J. Peyton, Field Museum of Natural History (FMNH); R. F. Martin, University of Texas at Austin Memorial Museum (TNHC); and R. R. Miller and E. B. Koon, University of Michigan Museum of Zoology. Mexico’s Department of Fishes graciously permitted me to collect in Mexico. SUMMARY Etheostoma (Oligocephalus) australe, redescribed herein, is un- usual but not unique among darters in having only one anal spine and is not considered sufficiently divergent to warrant placement in a monotypic subgenus (Austroperca). E. australe is endemic to the Rio Conchos system in Chihuahua and Durango and apparently spawns as early as late February in rubble rifles. A key is pre- sented for the identification of the three described species of Etheo- stoma in Mexico and for E. lepidum, a form endemic to Texas and New Mexico and often confused with E. grahami. RESUMEN Se redescribe Etheostoma (Oligocephalus) australe, especie pe- culiar, aunque no unica en su género, por tener solamente una espina anal; no se considera lo suficientemente distinta para colo- carla en un subgénero monotipico, Austroperca. E. australe se en- cuentra sdlo en el sistema del Rio Conchos de Chihuahua y Du- rango, México. Aparentemente desova a finales de Febrero en aguas rapidas de fondo pedregoso. Se presenta una clave para la identificacién de las tres especies de Etheostoma descritas de Méx- ico y de E. lepidum, especie endémica de Texas y Nuevo México, la cual con frecuencia se confunde con E. grahami. REDESCRIPTION OF ETHEOSTOMA AUSTRALE 9 LITERATURE CITED Atyarez, J. 1950. Claves para la determinacién de especies en los peces de las aguas continentales Mexicanas. Secretaria de Marina, Direccion General de Pesca 3 Industrias Connexas, México. 144 p. AtvarEz, J. 1970. Peces Mexicanos (claves). Instituto Nacional de Investi- gaciones Bioldgico Pesqueras, México. Serie Investigacion Pesquera Estudio No. 1:1-166. Bartey, R. M., Gostine, W. A. 1955. Variation and systematic significance of vertebral counts in the American fishes of the family Percidae. Misc. Publ. Mus. Zool. Univ. of Michigan 93:1-44. Burr, B. M. 1976. A review of the Mexican stoneroller, Campostoma ornatum Girard (Pisces: Cyprinidae). Trans. San Diego Soc. Nat. Hist. 18(7): 127-144. CotuetTe, B. B. 1965. Systematic significance of breeding tubercles in fishes of the family Percidae. Proc. U.S. Nat. Mus. 117(3518):567-614. Cot.etTTE, B. B., Knapp, L. W. 1966. Catalog of type specimens of the dart- ers (Pisces, Percidae, Etheostomatini). Proc. U.S. Nat. Mus. 119(3550): 1-88: Conant, R. 1963. Semiaquatic snakes of the genus Thamnophis from the isolated drainage system of the Rio Nazas and adjacent areas in Mex- ico. Copeia 1963 (3):473-499. ConTRERAS-BALDERAS, S. 1977. Speciation aspects and man-made community composition changes in Chihuahuan desert fishes. IN: Wauer, R. H., Riskinp, D. H. (eds.). Transactions of the Symposium on the Bio- logical Resources of the Chihuahuan Desert Region United States and Mexico. U.S. Dept. Int. Natl. Park Service Trans. and Proc. 3:405-431. Deacon, J. E., Koseticu, G., WiiuiaMs, J. D., Contreras, S., and other members of the Endangered Species Committee of the American Fish- eries Society. 1979. Fishes of North America Endangered, Threatened, or of Special Concern: 1979. Fisheries 4(2):29-44. Grrarp, C. 1859. Ichthyological notices, 28-40. Proc. Acad. Nat. Sci. Phila- delphia 11:100-104. Husss, C. L. 1936. Austroperca, a new name to replace Torrentaria, for a genus of Mexican fishes. Occ. Pap. Mus. Zool. Univ. Michigan 341:1-3. Husss, C. L., Lacier, K. F. 1964. Fishes of the Great Lakes region. Univ. Michigan Press, Ann Arbor. 213 p. Jorpan, D. S. 1889. Descriptions of fourteen species of fresh-water fishes col- lected by the U.S. Fish Commission in the summer of 1888. Proc. U.S. Nat. Mus. 11:351-362. Jorpan, D. S., EvERMANN, B. W. 1896. The fishes of North and Middle America: a descriptive catalogue of the species of fish-like vertebrates found in the waters of North America, north of the Isthmus of Panama. U.S. Nat. Mus. Bull. 47, pt. 1. 1240 p. Jorpan, D. S., EvERMANN, B. W., CLarx, H. W. 1930. Check list of the fishes and fishlike vertebrates of North and Middle America north of the boundary of Venezuela and Colombia. Rept. U.S. Comm. Fish., App. 10. 670 p. Merk, S. E. 1904. The fresh-water fishes of Mexico north of the Isthmus of Tehuantepec. Field Columbian Mus. Publ. 93, Zool. Series 5:1-252. MincxLeEy, W. L. 1969. Environments of the Bolsén of Cuatro Ciénegas, Coahuila, Mexico, with special reference to the aquatic biota. Univ. Texas El Paso Sci. Series 2:1-65. Mincxiey, W. L. 1977. Endemic fishes of the Cuatro Ciénegas Basin, north- ern coahuila, Mexico. IN: Waver, R. H., Risxinp, D. H. (eds.). 10 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Transactions of the Symposium on the Biological Resources of the Chi- huahuan Desert Region United States and Mexico. U.S. Dept. Int. Natl. Park Service Trans. and Proc. 3:383-404. Moore, G. A. 1968. Fishes, pp. 21-165. IN: Bram, W. F., Buam, A. P., Bropkors, P., Cacie, F. R., Moore, G. A. (eds.). Vertebrates of the United States. McGraw-Hill, New York. Pace, L. M. 1977. The lateralis system of darters (Etheostomatini). Copeia 1977 (3) :472-475. Pace, L. M. In press. The genera and subgenera of darters (Percidae, Etheostomatini). Misc. Papers Mus. Nat. Hist. Univ. Kansas. StrEveNSsOoN, M. M. 1971. Percina macrolepida (Pisces, Percidae, Etheostoma- tinea), a new percid fish of the subgenus Percina from ‘Texas. South- west. Nat. 16(1):65-83. Ran” opin ai aa jen } SOME rf, he Ss ijRe 1 wep Laie * thks sgt » te ober ¥ Fei’ te a UNIVERSITY OF KANSAS PUBLICATIONS MUSEUM OF NATURAL HISTORY The University of Kansas Publications, Museum of Natural History, beginning with volume 1 in 1946, was discontinued with volume 20 in 1971. Shorter research papers formerly pub- lished in the above series are now published as Occasional Papers, Museum of Natural History. The Miscellaneous Publi- cations, Museum of Natural History, began with number 1 in 1946. Longer research papers are published in that series. Monographs of the Museum of Natural History were initiated in 1970. All manuscripts are subject to critical review by intra- and extramural specialists; final acceptance is at the discretion of the publications committee. Institutional libraries interested in exchanging publications may obtain the Occasional Papers and Miscellaneous Publica- tions by addressing the Exchange Librarian, The University of Kansas Library, Lawrence, Kansas 66045. Individuals may pur- chase separate numbers of all series. Prices may be obtained upon request addressed to Publications Secretary, Museum of Natural History, The University of Kansas, Lawrence, Kansas 66045. Editor: E> WILEY Managing Editor: JoseEpH T. CoLLins PRINTED BY UNIVERSITY OF KANSAS PRINTING SERVICE LAWRENCE, KANSAS wy) MUS, Comp LED, LIBRA} ey Ol. “UNS 1981 HARV A UNIVERS| ry OCCASIONAL PAPERS of the MUSEUM OF NATURAL HISTORY The University of Kansas Lawrence, Kansas NUMBER 90, PAGES 1-69 MAY 28, 1981 THE GENERA AND SUBGENERA OF DARTERS (PERCIDAE, ETHEOSTOMATINI) By LAWRENCE M. Pace! Since 1955 the generic and subgeneric classification of darters has been largely the concept of Reeve M. Bailey (Bailey, 1951; Bailey, Winn and Smith, 1954; Bailey and Gosline, 1955). As future studies and data will alter the classification presented herein, so are changes in Bailey’s classification offered on the following pages. All proposéd changes are subgeneric; the greater stability in nomen- clature inherent in more inclusive genera argues for the continued use of Bailey’s three-genus classification of darters rather than ele- vating some or all subgenera to generic status. More important contributions than the proposed changes are the morphological diagnoses of genera and subgenera incorporating characteristics ex- amined critically, either in the literature or during this study, on each species of darter treated herein. Acknowledgements—D. L. Swofford and N. J. Mankovich, Uni- versity of Illinois, and D. G. Buth, University of California at Los Angeles, provided invaluable assistance with computer programs. M. A. Morris, Southern Illinois University at Carbondale (SIU-C), assisted in the formation of a new subgeneric name. For loans of specimens I am indebted to the following curators and their insti- tutions: R. M. Bailey, University of Michigan Museum of Zoology (UMMZ); R. D. Suttkus, Tulane University Museum of Natural History (TU); B. B. Collette, NMFS Systematics Laboratory, Na- tional Museum of Natural Histor: (USNM); F. B. Cross, The Uni- versity of Kansas Museum of Natural History (KU); D. A. Etnier, *Taxonomist, Illinois Natural History Survey, Urbana, Illinois 61801 and Associate in Ichthyology, Museum of Natural History, The University of Kansas, Lawrence, Kansas 66045. 2 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY University of Tennessee (UT); A. P. Blair and H. Lindsay, Univer- sity of Tulsa (UTULSAC); and R. E. Jenkins (REJ), Virginia Commonwealth University. Technical assistance was provided by B. P. Sweeney, A. K. Adams, L. LeMere, D. A. Teeter, J. Sanders, D. L. Sublette, and L. Woodrum, Illinois Natural History Survey. B. H. Bauer (UT), B. M. Burr (SIU-C), B. B. Collette (USNM), D. A. Etnier (UT), J. L. Russo (USNM), P. W. Smith, [llinois Natural History Survey (INHS), and W. C. Starnes (UT) offered criticisms and suggestions on the manuscript. . METHODS Phenetic and phylogenetic (cladistic) methods were employed in the following study on 142 species of darters (Appendix A) using 52 morphological characters (Appendix B). “The fundamental dis- tinction between phenetic and phylogenetic methods is that the former group according to overall similarity while the latter unite taxa only on the basis of synapomorphy” ( Mickevich, 1978). Phenetic Method.—Phenetic similarities of 142 species of darters or some subset thereof were examined using CLUSTER, written by Dr. Richard B. Selander, University of Illinois. Phenograms were produced using three procedures: (1) data were standardized (mean = 0, standard deviation = 1), Pearson product-moment cor- relation coefficients were calculated as measures of similarity be- tween OTU’s (species), and OTU’s were clustered using the UPGMA (unweighted pair-group method using arithmetic aver- ages) (Sneath and Sokal, 1973); (2) data were transformed by ranging from 0 to 1, average taxonomic distances between OTU’s were calculated, and OTU’s were clustered using the UPGMA; (3) data were ranged, mean character differences between OTU’s were calculated, and OTU’s were clustered using UPGMA. Cladistic Method.—Unrooted WAGNER networks (Kluge and Farris, 1969; Farris, 1970) were constructed using the “rootless” algorithm of Farris (1970). Cladograms were produced by rooting the networks (Farris, 1972) at the midpoint of the longest patristic _ distance separating any pair of OTU’s or by the inclusion of Perca flavescens as an outgroup. The most parsimonious (shortest) clado- grams are ones based on synapomorphies (Farris, Kluge and Eckardt, 1970). Cophenetic correlation coefficients, which reflect the ability of a phenogram (or cladogram) to describe the distance matrix (Sneath and Sokal, 1973), are given in the figure legends. Characters——Characters selected for numerical analysis were those which varied among species but which were relatively con- stant within species. Characters exhibiting a large amount of intra- specific variation in at least some species (e.g. cheek and opercle THE GENERA AND SUBGENERA OF DARTERS 3 squamation) and characters difficult to express numerically even when coded (e.g. body and fin colors) were not used. For morpho- metric data (meristic counts and body proportions) mean (or if so indicated in Appendix B, modal) values for each species were used as character states (Appendix C). For qualitative characters the various states were grouped into classes and assigned numerical values (coded). Pectoral, pelvic and caudal fin lengths were lengths of the long- est rays. Second dorsal and anal fin lengths were from the anterior point of origin of the fin to the posterior tip of the depressed fin. The distance from the origin of the second dorsal fin to the center of the base of the caudal fin (QL) was used as a basis for comparing second dorsal fin lengths. Interpelvic fin width was the distance between pelvic fin bases. Other counts and measurements were made following Hubbs and Lagler (1958) or as modified by Page (1974). Lateralis terminology follows Page (1977). When sexual dimorphism was evident in a character, specimens of only one sex were used or the feature was divided into two characters. The biochemical characters used in the numerical analysis of Percina (Page, 1974) were not used in this study because data were missing for about one-half of the species. Vertebral counts also were not used because data are unavailable for many species. The 52 char- acters employed were distributed as follows: Ol rethiitetbivcwer estan eee as ee 33 nme MGhIShi@pe § La. sa i te 17 B. tbody proportions, + 16 Mies Ohuralttatey ewe) 2 eel amet Bory eye IP) ade) 19 fs Soll feagnYe) 1122) 110) 1 es rs eer re nee 2 Ba Soiammation, 2: Se 6 KE PeMrscellancous “s20 5. eee ii Of the 52 characters, 42 were continuously variable, multistate characters. The other 10 were two-state characters utilizing a derived character state found in one or more species. Systematics—Phenograms and cladograms were compared to the existing generic and subgeneric classification [that of Bailey and Gosline (1955) modified subsequently by Collette and Yerger (1962), Cole (1967), Page and Whitt (1973a), Page (1974), and Williams and Robison (1980) (Table 1)] and discordances con- sidered as potential errors in classification. Examination of syn- apomorphous (shared derived) characteristics as evidence of rela- tionship (sensu Hennig, 1966) sometimes involved characters not included in the numerical analysis because of extreme difficulty in coding (e.g., colors in first dorsal fin). a OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY TABLE 1.—Classifications of darters. Genera and Sub- Modifications Subsequent Classification Adopted genera of Bailey and_ to Bailey and Gosline (1955) | Herein (and Number of Gosline (1955) Described Species in Each Subgenus ) Percina Percina Hadropterus Hadropterus (4) Swainia Swainia (4) Alvordius Alvordius (8) Ericosma Ericosma (2) Odontopholis (Page, 1974) Odontopholis (1) Hypohomus Hypohomus (1) Cottogaster Cottogaster (1) Imostoma Imostoma (5) Percina Percina (5) Ammocrypta Ammocrypta Crystallaria : Crystallaria (1) Ammocrypta Ammocrypta (6) > Etheostoma Etheostoma Psychromaster Psychromaster (1) Litocara (Page and Whitt, 1973a) Litocara (2) Allohistium Allohistium (1) Etheostoma Etheostoma (13) Ulocentra Nanostoma (6) Doration (Cole, 1967) Doration (2) Boleosoma Boleosoma (5) Ioa Ioa (1) Vaillantia (Cole, 1967) Vaillantia (2) Nothonotus Nothonotus (13) Fuscatelum (1) Belophlox (3) Villora (Collette and Yerger, 1962) | Villora (1) Ozarka (Williams and Robison, 1980) Ozarka (5) Austroperca Oligocephalus Oligocephalus (15) Catonotus Catonotus (10) Hololepis Boleichthys (10) Microperca RESULTS Comparisons of Methods.—Comparisons among phenograms and cladograms of a given group of species were made by compar- ing their structures to the existing classification of darters, by exam- ining their overall ability to form clusters, and by contrasting their cophenetic correlation coefficients (CPCC) (Table 2). The objective in doing a numerical taxonomic study is to exam- ine and improved the existing classification of a group of organisms. The lack of congruence among phenograms and cladograms pre- vents the substitution of any one of them as the structure for a new classification and instead requires that they be compared with the existing classification, which generally is a product of careful study. * THE GENERA AND SUBGENERA OF DARTERS ‘eiguasqns oIdAjououl Surpnpoxy , eee eee 11e¥ 9 9€8'0 dnoisyno WIM pe}00I—YANOVAA ‘S Ie} 9 981'0 yujodprur ye p2}00I—YANOVM “F poos reat 86L'0 JUSIOYJ909 UOT}e[aLI0ON—-WIWOdQ 's 1ood 8 T8L'0 soUusIOHIp Jojovieyo URIUI—yWWOdQ °S 100d c 0230 QOURISIP OTMOUOXE} BBBIBAR—YINDGQ ‘T (satoods GPT) vwojsoay7q pure vidfisvowwy ‘puioiag “TI qey 3 P8L'0 dnoisjno YIM p23}00I—YANOVM “S ITey i L180 jUJOdpiw 3B p23}00I—YANOVM ‘F poos P €08'0 JUSIOYJe0d UOAL[elI0I— ION '€ 1ood v4 SLL'O sOUdIOYIp IoJOvIeYO URAUI—YIWDNIN °S ood € 008'0 9OULSIP OIWIOUOXE} OBVIBAR—YIVDGQ ‘T (sotoeds [Q[) Dwooay;q “TI poos € 0F8'0 dno1s3no YUM pr}00I—YANOVA ‘S poos 9 088'0 yuIOdprur 3B pI}001—YANOVM ‘Ff poos 9 868'0 PST re OTE [OWI Niro canes poos P PSL'0 DoUdIOYIp JayVIVYO URIUI—YIVOIN °Z poos c O18'0 90UP}SIP OIUIOUOXE} BSPIOAV—YWOdN ‘T (satoads FE) Duldlag 'T snuesqns ;pe193snyjo yUSTOyJI00 9Aoqe eiouesqns UOQP[A1I00 BULIO}snID ‘ON oneusydon *SpOYJOUL OIWIOUOXe} [eoHeumnu Jo suostieduloj— 'G ATdvV YT, 6 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY The “errors” in classification which are suggested most consistently by the phenograms and cladograms are ones to receive the most attention. Conversely, the numerical taxonomic method which gives - a structure most similar to the existing classification is likely to be the best in terms of information content, and is the one which a specialist on the taxonomic group under study will consider the most reliable. Among procedures used, UPGMA—correlation co- efficient gave results most congruent with darter systematics and WAGNER—rooted with an outgroup gave the least congruent re- sults (Table 2). For small groups of species (Figs. 1-4) WAGNER —rooted at midpoint generally gave results as good as those of UPGMA—correlation coefficient, but it gave poor results when all species were included (Table 2). Sneath and Sokal (1973), com- paring only phenetic methods, noted that when comparisons are made with existing classifications, structures produced from pro- cedures using correlation coefficients usually are deemed the best. P. AURANTIACA HYPOHOMUS P. CYMATOTAENIA < eS ssurydoy “J snjpyda203yO a + _ + igunyosog “q = + _— + uy [esiop }siy uo =—- WnsLopipyjDd “q _ + — + UISIVUIGNS 10 ULSIVUL pue 19}U9A aBULIO-MOT[IA 1UIZD19 “FT = + = + WySiIq YAM apeur ‘papyoads 10 papyou Apog wnzpjpnjound -J — _ — _— DIAS} FT DYIVZO _- = _ + uy [PSIOp }SIf UI SAAOI F-Z julmpa “7 url pue Apoq 19A0 paiayeos sjods par YIM aeyy DLO} A a — = te aDSOODYO “FJ + — — — aul] [¥19}e] }YSI[ SuysesQuoo pue (apsoojvyo "J aDuDUuL “J _- _ _— —_ ul pedojaAap Apyeam ) eduys yoRyq-uMoiq uinisyouf “7 peoiq YIM apis ‘uy [esiop ysiy UO UIBIeUI pay xojydojag — -- = -|- suy 10 Apoq uo si0[oo aguurdiaipd -y ( AoA 10 uses ‘anyq ‘par) }ysUq ON uinjayDoIsn J sai0od ais Ol oul WOd sjvuvo peroy [e19}| uoQeIO[OO ouIN peydnisezuy 9j9/duoouy daQounsiq ‘[eroduraqeidns = [g feywquovsyur = Oy ‘1epnqrpuewoynoiedoaid = YOU “esque Ajjepout = — ‘juasaid Ayjepowt = + ‘po1adoujsny pur ‘DLOULA ‘snoydas03yQ 0} peusisse ApiauIo0; satads jo soNsuajVieyO—}f ITAVL THE GENERA AND SUBGENERA OF DARTERS 27 dorsal fin is clear with two to four rows of discrete red spots. In E. parvipinne, for which the subgeneric name Fuscatelum is proposed, no bright colors in the fins or anywhere on the body are present. Oligocephalus is further separable from Belophlox by modally having an incomplete lateral line and by not having a wide brown- black lateral stripe and contrasting light lateral line; from Villora by the relatively short and deep caudal peduncle; from Ozarka by the absence of the bright yellow-orange venter of the breeding male; and from Fuscatelum by the presence of bright colors. E. grahami and E. pottsi appear to be specialized derivatives of, or early offshoots within, Oligocephalus. Unlike in other species of Oligocephalus, in E. grahami and E. pottsi the first dorsal fin is margined with red and has no blue band and vertical bars on the side of the body are only weakly developed. However, other char- acteristics and the general physiognomy agree with those of other Oligocephalus (female and young E. lepidum and E. grahami are often confused taxonomically) and leaving E. grahami and E. pottsi in Oligocephalus seems justified. E. edwini and E. okaloosae, both referred to Villora by Collette and Yerger (1963), do not appear to be sufficiently related to be considered consubgenerics. In general body shape and coloration, E. edwini and E. okaloosae are quite dissimilar. The arching of the lateral line is much more pronounced in E. edwini (as noted by Collette and Yerger); in E. okaloosae the curve in the lateral line is nearly identical to that in some species of Oligocephalus (e.g., E. swaini). The short villous genital papilla characteristic of breed- ing females of both E. edwini and E. okaloosae appears to be struc- turally the same as in darters as diverse as E. grahami and Percina sciera and is not indicative of common ancestry. E. australe, presumably referred previously to the monotypic subgenus Austroperca because it possesses only one anal spine, is except for not having two anal spines a typical Oligocephalus with red-over-blue bands in the first dorsal fin and strong vertical bars on the side of the body. GENERA AND SUBGENERA Genus Percina Haldeman Diagnosis——With modified (enlarged and strongly toothed) scales on breast of male; two anal spines; complete lateral line; uninterrupted head canals; opaque flesh; a distinctive electropho- retic mobility of the LDH By, isozyme (Page and Whitt, 1973b); mean body depth/standard length = 0.12-0.20. Subgenera of Percina were diagnosed by Page (1974). Subse- quent modifications in the diagnosis of the subgenus Imostoma were made by Etmier (1976) and Page (1976). 28 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Genus Ammocrypta Jordan Diagnosis——Without modified (as in Percina) scales; one anal spine; complete lateral line; uninterrupted head canals; translucent flesh (when alive); without LDH B, isozyme mobility of Percina; body long and slender, mean body depth/standard length = 0.11- 0.13. The subgenus Ammocrypta was diagnosed by Williams (1975). Crystallaria, the other subgenus of Ammocrypta (s.l.) may be sepa- rated from Ammocrypta (s.s.) as follows. Subgenus Crystallaria Jordan and Gilbert Diagnosis——Prevomer and palatine with teeth; premaxillary frenum present (narrow); caudal fin deeply forked; modally 12-15 dorsal spines, 13-14 dorsal rays, 12-15 anal rays, 45-48 vertebrae. Genus Etheostoma Rafinesque 2 Diagnosis.—Without modified (as in Percina) scales; one or two anal spines; complete or incomplete lateral line; uninterrupted or interrupted head canals; opaque flesh (except translucent in E. vitreum which often has two anal spines); without LDH By, isozyme mobility of Percina (except in E. cinereum among species exam- ined); mean body depth/standard length = 0.14-0.23. Among the characters examined in the literature on darters and in the CLUSTER and WAGNER analyses (Appendix B), the fol- lowing are the most useful in diagnosing subgenera of Etheostoma. 1. Lateral line arched or not arched upward anteriorly. In Boleichthys and Villora the lateral line characteristically is abruptly arched upward anteriorly, the apex of the arch being under the anterior half of the first dorsal fin. In some other subgenera, e.g. Oligocephalus, the line may be gently curved but not distinctly arched. 2. Lateral line complete or incomplete: Many species consid- ered to have complete lateral lines occasionally have a few unpored scales in the lateral-line row; consequently, the lateral line arbitrar- ily is termed complete in any species in which 95% or more of the | scales in the row are pored. Percentages of completeness are spe- cies means from Appendix C rounded to nearest 5%. 3. Infraorbital canal interrupted or uninterrupted. 4. Supratemporal canal interrupted or uninterrupted. 5. Number of preoperculomandibular pores. 6. Prevomer with or without teeth (data mainly from Richards, 1966). 7. Palatine with or without teeth (data mainly from Richards, 1966). 8. Number of branchiostegal rays. 9. Premaxillary frenum present or absent. THE GENERA AND SUBGENERA OF DARTERS 29 10. Number of anal spines. 11. Supraoccipital scaled (with scales) or unscaled (without scales ). 12. Flesh opaque or translucent when alive. 13. Squamation of nape (fully scaled, partly scaled, unscaled). 14. Squamation of breast (fully scaled, partly scaled, unscaled). 15. Connection of branchiostegal membranes across isthmus (separate, narrowly joined, moderately joined, broadly joined). 16. Breeding tubercles present or absent on males (data mainly from Collette, 1965). 17. Branchiostegal membranes scaled or unscaled. 18. Depth of caudal peduncle relative to standard length. Means from Appendix C. 19. Number of anal rays. 20. Body depth relative to standard length. Means from Ap- pendix C. 21. Interorbital width relative to head width. Means from Ap- pendix C. 22. Snout length relative to body depth. Means from Appen- dix C. 23. Second dorsal fin length relative to distance from front of base of second dorsal fin to caudal base (of females). Means from Appendix C. 24. Pelvic fin length relative to standard length. Means from Appendix C. 25. Gape width relative to snout length. Means from Appendix C. 26. Adult males average larger or smaller than adult females. 27. Anus surrounded or not surrounded by villi. 28. Interpelvic fin width relative to pelvic fin base length. 29. Genital papilla of breeding male long and tubular or not long and tubular. 30. General shape of genital papilla of breeding female (illustra- tions and descriptions in Collette, 1962, 1965; Cole, 1967; Howell, 1968; Burr, 1978). 31. Site of egg deposition. 32. Particular color patterns in some subgenera. (The absence of those patterns in other subgenera is understood and not stated as such in each diagnosis. ) The combination of characteristics in the initial paragraph of each of the following subgeneric diagnoses separates that subgenus from all other subgenera of Etheostoma. The second paragraph of each diagnosis describes the states of characters used in the initial paragraphs of the diagnoses of other subgenera. Unless stated otherwise in the above list of characters or in the diagnoses, all characteristics are modal conditions. 30 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY The subgenera of Etheostoma are arranged in approximate prim- itive to advanced sequence (Table 1). Numerous phyletic lines are involved in the evolutionary history of the genus and the arrange- ment is subjective. Psychromaster Jordan and Evermann INCLUDED SPECIES.—E. tuscumbia. Psychromaster, as a subgenus.of Etheostoma containing E. tuscumbia and E. trisella, was diagnosed by Bailey and Richards (1963). DIAGNOSIS.—Supraoccipital region with scales; branchioste- gal membranes with scales (variable intraspecifically); genital pa- pilla of male long and tubular; one thick and stiff anal spine; lateral line not arched upward anteriorly; pelvic fins short (mean length/ standard length — 0.16); gape wide in relation to snout length (mean width/snout length = 1.35). Lateral line incomplete (25% scales in lateral-line row pored); infraorbital canal uninterrupted; supratemporal canal interrupted; modally 10 preoperculomandibular pores; prevomer and palatine with teeth; modally six branchiostegal rays; premaxillary frenum present; flesh opaque; nape and breast fully scaled; branchiostegal membranes separate; breeding tubercles absent; caudal peduncle deep (mean depth/standard length more than 0.11); anal rays modally fewer than 11; mean body depth/standard length more than 0.145; interorbital wide (mean interorbital width/head width more than 0.22); snout not long in relation to body depth (mean snout length/body depth less than 0.53); in females, second dorsal fin not short (mean fin length/distance from front of base of second dorsal fin to caudal base more than 0.535); adult males average larger than females; anus not surrounded by villi; interpelvic fin width/pelvic fin base length less than 0.675; genital papilla of breeding female short and conical; eggs attached to plants. Litocara Bailey INCLUDED SPECIES.—E. sagitta, E. nianguae. DIAGNOSIS.—Modally 11 or 12 anal rays; base of caudal fin with two vertically aligned jet-black spots (often confluent in E. sagitta and breeding males). Lateral line not arched upward anteriorly, incomplete to com- plete (90-100% scales in lateral-line row pored); infraorbital and supratemporal canals uninterrupted; modally 10 preoperculoman- dibular pores; prevomer and palatine with teeth; modally six branchiostegal rays; premaxillary frenum present; modally two anal spines; supraoccipital unscaled; flesh opaque; nape fully scaled; breast unscaled; branchiostegal membranes separate to narrowly joined across isthmus; breeding tubercles present on body scales of male; branchiostegal membranes unscaled; caudal peduncle moder- ate in depth (mean depth/standard length = 0.08-0.11); mean body THE GENERA AND SUBGENERA OF DARTERS 31 depth/standard length more than 0.145; interorbital wide (mean interorbital width/head width more than 0.22); snout not long in relation to body depth (mean snout length/body depth less than 0.53); in females, second dorsal fin not short (mean fin length/dis- tance from front of base of second dorsal fin to caudal base more than 0.535); pelvic fins not short (mean fin length/standard length more than 0.16); gape not wide in relation to snout length (mean gape width/snout length less than 1.2); adult males average larger than females; anus not surrounded by villi; interpelvic fn width/ pelvic fin base length less than 0.675; genital papilla of breeding male not long and tubular, of breeding female a short tube; eggs buried in substrate (known for E. nianguae). Allohistium Bailey INCLUDED SPECIES.—E. cinereum. Allohistium was diag- nosed as a subgenus of Etheostoma by Bailey in Bailey and Gosline (1955). DIAGNOSIS.—Snout long in relation to body depth (mean snout length/body depth = 0.54); palatine without teeth; modally six branchiostegal rays; nape and breast unscaled (or few scales posteriorly on nape); branchiostegal membranes separate to nar- rowly joined across isthmus; midlateral row of small black rectan- gles, upper sides with thin brown horizontal lines; first dorsal fin with red-brown margin. Lateral line not arched upward anteriorly, complete (100% scales in lateral-line row pored); infraorbital and supratemporal canals uninterrupted; modally 10 preoperculomandibular pores; prevomer with teeth; premaxillary frenum present; modally two anal spines, the first thick and stiff; supraoccipital unscaled; flesh opaque; nape and breast unscaled; breeding tubercles absent; branchiostegal membranes unscaled; caudal peduncle moderate in depth (mean depth/standard length = 0.08-0.11); anal rays modally fewer than 11; mean body depth/standard length more than 0.145; interorbital wide (mean interorbital width/head width more than 0.22); in fe- males, second dorsal fin not short (mean fin length/distance from front of base of second dorsal fin to caudal base more than 0.535); pelvic fins not short (mean fin length/standard length more than 0.16); gape not wide in relation to snout length (mean gape width/ snout length less than 1.2); adult males average larger than females; anus not surrounded by villi; interpelvic fin width/pelvic fin base length less than 0.675; genital papilla of breeding male not long and tubular, of breeding female unknown; egg deposition site unknown. Etheostoma Rafinesque INCLUDED SPECIES.—E. blennioides, E. blennius, E. euzo- num, E. histrio, E. inscriptum, E. kanawhae, E. osburni, E. rupestre, E. sellare, E. swannanoa, E. tetrazonum, E. thalassinum, E. vari- 32 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY atum. Etheostoma was diagnosed as a subgenus including E. zonale by Richards (1966). DIAGNOSIS.—Branchiostegal membranes broadly joined across the isthmus (narrowly joined in E. sellare); modally six branchi- ostegal rays and 10 preoperculomandibular pores (nine in E. histrio and E. rupestre); genital papilla of breeding female long and tubu- lar (short and flat in E. sellare). E. sellare is apparently an early offshoot within subgenus Etheostoma; characteristics not strictly diagnostic for subgenus are indicative of relationship of E. sellare to certain species { E. variatum complex) of subgenus: back crossed by large and distinct dark-brown saddles which extend anteroven- trally below lateral line; expansive pectoral fins. Lateral line not arched anteriorly, complete (95-100% scales in lateral-line row pored); infraorbital and supratemporal canals un- interrupted; prevomer and palatine with or without teeth; premax- illary frenum present or absent; modally two anal spines, the first thick and stiff; supraoccipital unscaled; flesh opaque; nape and breast fully scaled to unscaled; breeding tubercles absent, present on body scales of males, or present on body scales of males and females; branchiostegal membranes unscaled; caudal peduncle moderate in depth (mean depth/standard length = 0.08-0.11); anal rays modally fewer than 11; mean body depth/standard length more than 0.145; interorbital moderate to wide (mean interorbital width/ head width more than 0.16); snout not long in relation to body depth (mean snout length/body depth less than 0.53); in females, second dorsal fin not short (mean fin length/distance from front of base of second dorsal fin to caudal base more than 0.535); pelvic fins not short (mean fin length/standard length more than 0.16); gape not wide in relation to snout length (mean gape width/snout length less than 1.2); adult males average larger than females; anus not surrounded by villi; genital papilla of breeding male not long and tubular; eggs attached to plants (E. blennioides.) Nanostoma Putnam INCLUDED SPECIES.—E. zonale, E. atripinne, E. coosae, E. duryi, E. etnieri, E. simoterum, and about 10 undescribed species. All included species except E. zonale, previously in Etheostoma (s.s.), were previously in Ulocentra, diagnosed by Bouchard (1977). DIAGNOSIS.—Branchiostegal membranes broadly joined across isthmus, modally five branchiostegal rays and nine preoperculo- mandibular pores (six and 10, respectively, in E. coosae; 10 pre- operculomandibular pores in E. zonale); genital papilla of breeding female long and tubular. Lateral line not arched anteriorly, complete (95-100% scales in lateral-line row pored); infraorbital and supratemporal canals un- interrupted; prevomer with or without teeth; palatine without teeth; THE GENERA AND SUBGENERA OF DARTERS 33 premaxillary frenum present or absent; modally two anal spines, the first thick and stiff; supraoccipital unscaled; flesh opaque; nape fully scaled; breast partly scaled to unscaled; breeding tubercles absent; branchiostegal membranes unscaled; caudal peduncle moderate in depth (mean depth/standard length = 0.08-0.11); anal rays mod- ally fewer than 11; mean body depth/standard length more than 0.145; interorbital moderate to wide (mean interorbital width/head width more than 0.16); snout not long in relation to body depth (mean snout length/ body depth less than 0.53); in females, second dor- sal fin not short (mean fin length/ distance from front of base of sec- ond dorsal fin to caudal base more than 0.535); pelvic fins not short (mean fin length/standard length more than 0.16); gape not wide in relation to snout length (mean gape width/ snout length less than 1.2); adult males average larger than females; anus not surrounded by villi; genital papilla of breeding male not long and tubular; eggs attached to plants (E. zonale) or to sides of rocks (E. atripinne, E. coosae, and two undescribed species). Doration Jordan INCLUDED SPECIES.—E. stigmaeum, E. jessiae. Doration was diagnosed as a subgenus of Etheostoma by Cole (1967). DIAGNOSIS.—First dorsal fin with blue marginal band and well-defined red band (bands conspicuous in large males); palatine without teeth; branchiostegal membranes narrowly joined across isthmus; genital papilla of breeding female long and tubular. Lateral line not arched upward anteriorly, incomplete to com- plete (60-100% scales in lateral-line row pored); infraorbital and supratemporal canals uninterrupted; modally 10 preoperculoman- dibular pores; prevomer with teeth; modally six branchiostegal rays; premaxillary frenum present or absent; modally two anal spines; supraoccipital unscaled; flesh opaque; nape fully to partly scaled; breast unscaled (sometimes partly scaled); breeding tubercles on body scales and anal and pelvic fins of males; branchiostegal mem- branes unscaled; caudal peduncle narrow (mean depth/standard length less than 0.08); anal rays modally fewer than 11; mean body depth/standard length more than 0.145; interorbital wide (mean in- terorbital width/head width more than 0.22); snout not long in rela- tion to body depth (mean snout length/ body depth less than 0.53); in females, length of second dorsal fin variable (mean fin length/dis- tance from front of base of second dorsal fin to caudal base = 0.49- 0.59); pelvic fins not short (mean fin length/standard length more than 0.16); gape not wide in relation to snout length (mean gape width/snout length less than 1.2); adult males average larger than females; anus not surrounded by villi; interpelvic fin width/ pelvic fin base length less than 0.675; genital papilla of breeding male not long and tubular; eggs buried in substrate (known for E. stigmaeum). 34 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Boleosoma DeKay INCLUDED SPECIES.—E. olmstedi, E. longimanum, E. ni- grum, E. perlongum, E. podostemone. Boleosoma was diagnosed as a subgenus of Etheostoma by Cole (1967). DIAGNOSIS.—Female with broad, flat bifurcate genital papilla (Cole, 1967: fig. 3); eggs attached to underside of stone or log (known for four of five species) premaxillary frenum absent; inter- pelvic fin width/pelvic fin base length more than 0.675. Lateral line not arched upward anteriorly, complete (95-100% scales in lateral-line row pored); infraorbital and supratemporal canals interrupted or uninterrupted; modally nine-11 preoperculo- mandibular pores; prevomer and palatine with teeth; modally six branchiostegal rays; modally one or two anal spines, thin and flex- ible; supraoccipital unscaled; flesh opaque; nape and breast fully scaled to unscaled; branchiostegal membranes narrowly to broadly joined across isthmus; breeding tubercles absent; branchiostegal membranes unscaled; caudal peduncle narrow to deep (mean depth/standard length less than 0.08 to more than 0.11); anal rays modally fewer than 11; mean body depth/standard length more than 0.145; interorbital moderate in width (mean interorbital width/ head width between 0.16 and 0.22); snout not long in relation to body depth (mean snout length/body depth less than 0.53); in females, second dorsal fin not short (mean fin length/ distance from front of base of second dorsal fin to caudal base more than 0.535); pelvic fins not short (mean fin length/standard length more than 0.16); gape not wide in relation to snout length (mean gape width/ snout length less than 1.2); adult males average larger than females; anus not surrounded by villi; genital papilla of breeding male not long and tubular. Toa Jordan and Brayton INCLUDED SPECIES.—E. vitreum. DIAGNOSIS.—Communal spawner, eggs attached to sides of rocks and logs; flesh translucent when alive; anus surrounded by villi; prevomer without teeth; snout long in relation to body depth (mean snout length/body depth = 0.56); interorbital narrow (mean interorbital width/head width = 0.15); body flattened (mean body depth/ standard length = 0.14). Lateral line not arched upward anteriorly, complete (95-100% scales in lateral-line row pored); infraorbital and supratemporal canal uninterrupted; modally 10 preoperculomandibular pores; pal- atine without teeth; modally six branchiostegal rays; premaxillary frenum absent; one or two anal spines, thin and flexible; supra- occipital unsealed; nape partly scaled; breast unscaled; branchi- ostegal membranes narrowly joined across isthmus; breeding tuber- cles present on pectoral and pelvic fins of males and females; THE GENERA AND SUBGENERA OF DARTERS 35 branchiostegal membranes unscaled; caudal peduncle narrow (mean depth/ standard length less than 0.08); anal rays modally fewer than 11; in females, second dorsal fin not short (mean fin length/ distance from front of base of second dorsal fin to caudal base more than 0.535); pelvic fins not short (mean fin length/standard length more than 0.16); gape not wide in relation to snout length (mean gape width/snout length less than 1.2); adult males average larger than females; interpelvic fin width/pelvic fin base length more than 0.675; genital papilla of breeding male not long and tubular, of breeding female short and villous. Vaillantia Jordan INCLUDED SPECIES.—E. chlorosomum, E. davisoni. Vail- lantia was diagnosed as a subgenus of Etheostoma by Cole (1967). DIAGNOSIS.—Preorbital bars extend around snout as contin- uous black bridle; premaxillary frenum absent; palatine with teeth; incomplete lateral line (55-65% scales in lateral-line row pored); second dorsal fin short, in females mean fin length/distance from front of base of second dorsal fin to caudal base = 0.52. Lateral line not arched upward anteriorly; infraorbital and supratemporal canals interrupted or uninterrupted; modally nine or 10 preoperculomandibular pores; prevomer with teeth, modally six branchiostegal rays; one or two anal spines, thin and flexible; supra- occipital unscaled; flesh opaque or semitranslucent when alive; nape partly scaled; breast partly scaled to unscaled; branchiostegal membranes narrowly to moderately joined across isthmus; breeding tubercles absent or present on anal and pelvic fins of male; branchi- ostegal membranes unscaled; caudal peduncle narrow (mean depth/ standard length less than 0.08); anal rays modally fewer than 11; mean body depth/standard length more than 0.145; interorbital wide (mean interorbital width/head width more than 0.22); snout not long in relation to body depth (mean snout length/body depth less than 0.53); pelvic fins not short (mean fin length/standard length more than 0.16); gape not wide in relation to snout length (mean gape width/snout length less than 1.2); adult males average larger than females; anus not surrounded by villi; interpelvic fin width/pelvic fin base length less than 0.675; genital papilla of breeding male not long and tubular, of female villous; eggs attached to plants (known for E. chlorosomum). Nothonotus Putnam INCLUDED SPECIES.—E. maculatum, E. acuticeps, E. aquali, E. bellum, E. camurum, E. chlorobranchium, E. jordani, E. juliae, E. microlepidum, E. moorei, E. rubrum, E. rufilineatum, E. tippe- canoe. Nothonotus was diagnosed as a subgenus of Etheostoma by Bailey (1959) and Zorach (1972). DIAGNOSIS.—Sides of body with thin alternating dark and 36 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY light lines (absent in E. tippecanoe and E. jordani); lateral line complete (95-100% scales in lateral-line row pored) (incomplete in E. tippecanoe); branchiostegal membranes separate to narrowly joined across isthmus (broadly joined in E. juliae); palatine with teeth; premaxillary frenum present; breeding tubercles absent; modally 10 preoperculomandibular pores; breast unscaled. Lateral line not arched upward anteriorly; infraorbital and supratemporal canals uninterrupted; prevomer with teeth; modally six branchiostegal rays; modally two anal spines, the first thick and stiff; supraoccipital unscaled; flesh opaque; nape fully scaled (in E. juliae) to unscaled; branchiostegal membranes unscaled; caudal peduncle deep (mean depth/standard length more than 0.11); anal rays modally fewer than 11; mean body depth/standard length more than 0.145; interorbital wide (mean interorbital width/head width more than 0.22); snout not long in relation to body depth (mean snout length/body depth less than 0.53); in females, second dorsal fin not short (mean fin length/distance from front of base of second dorsal fin to caudal base more than 0.535); pelvic fins not short (mean fin length/standard length more than 0.16); gape not wide in relation to snout length (mean gape width/snout length less than 1.2); adult males average larger than females; anus not surrounded by villi; interpelvic fin width/ pelvic fin base length less than 0.675; genital papilla of breeding male not long and tubular, of female short and conical or flat and broad but not bilobed; eggs buried in substrate (known for three species) or attached to under- side of stone (E. maculatum). Fuscatelum, new subgenus TYPE SPECIES.—Etheostoma parvipinne Gilbert and Swain 1887. INCLUDED SPECIES.—E. parvipinne. This species was for- merly in the subgenus Oligocephalus. The name, Fuscatelum (dark- colored dart), emphasizes the absence of bright colors. DIAGNOSIS.—No bright colors (red, blue, green or yellow) on body or fins; premaxillary frenum present; breeding tubercles pres- ent on anal fin of male. Lateral line not arched upward anteriorly, incomplete (90% scales in lateral-line row pored); infraorbital canal uninterrupted; supratemporal canal interrupted; modally 10 preoperculomandibu- lar pores; prevomer and palatine with teeth; modally six branchi- ostegal rays; modally two anal spines; supraoccipital unscaled; flesh opaque; nape fully scaled; breast fully to partly scaled; branchi- ostegal membranes moderately joined across isthmus; branchiostegal membranes unscaled; caudal peduncle deep (mean depth/standard length more than 0.11); anal rays modally fewer than 11; mean body depth/standard length more than 0.145; interorbital wide THE GENERA AND SUBGENERA OF DARTERS 37 (mean interorbital width/head width more than 0.22); snout not long in relation to body depth (mean snout length/body depth less than 0.53); in females, second dorsal fin not short (mean fin length/ distance from front of base of second dorsal fin to caudal base more than 0.535); pelvic fins not short (mean fin length/standard length more than 0.16); gape not wide in relation to snout length (mean gape width/snout length less than 1.2); anus not surrounded by villi; interpelvic fin width/pelvic fin base length less than 0.675; genital papilla of breeding male not long and tubular, of female conical; egg deposition site unknown. Belophlox Fowler INCLUDED SPECIES.—E. mariae, E. fricksium, E. okaloosae. These species were formerly in the subgenera Oligocephalus and Villora. DIAGNOSIS.—Lateral line complete or nearly complete (95- 100% scales in lateral-line row pored); side of body with broad dark stripe (usually weakly developed in E. okaloosae) and contrasting light lateral line; first dorsal fin with red margin (or submargin); palatine with teeth. Lateral line not arched upward anteriorly; infraorbital and supratemporal canals uninterrupted; modally nine or 10 preoper- culomandibular pores; prevomer with teeth; modally six branchi- ostegal rays; premaxillary frenum present; modally two anal spines; supraoccipital unscaled; flesh opaque; nape partly to fully scaled; breast unscaled to fully scaled; branchiostegal membranes narrowly to broadly joined across isthmus; breeding tubercles absent or pres- ent on anal fin of male; branchiostegal membranes unscaled; caudal peduncle moderate (mean depth/standard length — 0.08-0.11); anal rays modally fewer than 11; mean body depth/standard length more than 0.145; interorbital wide (mean interorbital width/head width more than 0.22); snout not long in relation to body depth (mean snout length/body depth less than 0.53); in females, second dorsal fin not short (mean fin length/distance from front of base of second dorsal fin to caudal base more than 0.535); pelvic fins not short (mean fin length/standard length more than 0.16); gape not wide in relation to snout length (mean gape width/snout length less than 1.2); adult males average larger than females; anus not surrounded by villi; narrow interpelvic width (mean width/ pelvic fin base length = 0.27-0.46); genital papilla of breeding male not long and tubular, of female conical to tubular; eggs attached to plants (E. okaloosae). Villora Hubbs and Cannon INCLUDED SPECIES.—E. edwini. Villora as a subgenus of Etheostoma containing E. edwini and E. okaloosae was diagnosed by Collette and Yerger (1962). 38 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY DIAGNOSIS.—Lateral line distinctly arched upward anteriorly; adult males average larger than females; breeding male with two to four rows of red spots in first dorsal fin and red spots scattered over side of body; nape fully scaled. Lateral line incomplete (65% scales in-lateral-line row pored); infraorbital and supratemporal canals uninterrupted; modally 10 preoperculomandibular pores; prevomer and palatine with teeth; modally six branchiostegal rays; premaxillary frenum present; mod- ally two anal spines, first thicker than second; supraoccipital un- scaled; flesh opaque; breast partly to fully scaled; branchiostegal membranes separate to narrowly joined across isthmus; breeding tubercles absent; branchiostegal membranes unscaled; caudal pe- duncle moderate (mean depth/standard length = 0.08-0.1!); anal rays modally fewer than 11; mean body depth/standard length more than 0.145; interorbital wide (mean interorbital width/head width more than 0.22); snout not long in relation to body depth (mean snout length/body depth less than 0.53); in females, second dorsal fin not short (mean fin length/distance from front of base of second dorsal fin to caudal base more than 0.535); pelvic fins not short (mean fin length/standard length more than 0.16); gape not wide in relation to snout length (mean gape width/snout length less than 1.2); anus not surrounded by villi; interpelvic fin width/ pelvic fin base length less than 0.675; genital papilla of breeding male not long and tubular, of female a short cone; eggs attached to plants. Ozarka Williams and Robison INCLUDED SPECIES.—E. punctulatum, E. boschungi, E. cra- gini, E. pallididorsum, E. trisella. These species were formerly in the subgenera Oligocephalus and Psychromaster. DIAGNOSIS.—Bright yellow-orange venter and margin (or sub- margin) of first dorsal fin of breeding male; supratemporal canal interrupted (uninterrupted in E. trisella); side of body mottled or speckled, without vertical bars; breast unscaled. Lateral line not arched upward anteriorly, incomplete (25-75% scales in lateral-line row pored) except in E. trisella which has com- plete lateral line (100% pored); infraorbital canal uninterrupted; modally 10 preoperculomandibular pores; prevomer and _ palatine with teeth; modally six branchiostegal rays; premaxillary frenum present; modally two anal spines (one in E. trisella); supraoccipital unscaled; flesh opaque; nape fully to partly scaled; branchiostegal membranes narrowly to moderately joined across isthmus; breeding tubercles present on anal and pelvic fins of male (absent in E. trisella); branchiostegal membranes unscaled; caudal peduncle moderate to deep (mean depth/standard length more than 0.08); anal rays modally fewer than 11; mean body depth/standard length THE GENERA AND SUBGENERA OF DARTERS 39 more than 0.145; interorbital wide (mean interorbital width/head width more than 0.22); snout not long in relation to body depth (mean snout length/ body depth less than 0.53); in females, second dor- sal fin not short (mean fin length/distance from front of base of sec- ond dorsal fin to caudal base more than 0.535); pelvic fins not short (mean fin length/standard length more than 0.16); gape not wide in relation to snout length (mean gape width/snout length less than 1.2); adult females average larger than males (?, known only for E. pallididorsum); anus not surrounded by villi; interpelvic fin width/pelvic fin base length less than 0.675; genital papilla of breeding male not long and tubular, of female tubular; eggs laid on vegetation in seepage water (known for E. boschungi and E. trisella). Oligocephalus Girard. INCLUDED SPECIES.—E. lepidum, E. asprigene, E. australe, E. caeruleum, E. collettei, E. ditrema, E. grahami, E. hopkinsi, E. luteovinctum, E. nuchale, E. pottsi, E. radiosum, E. spectabile, E. swaini, E. whipplei. Oligocephalus, as a subgenus of Etheostoma containing several species in addition to those above but not E. australe, was diagnosed by Bailey and Richards (1963). DIAGNOSIS.—First dorsal fin with blue (green in E. lepidum and E. hopkinsi) marginal band and well-defined red band (bands conspicuous in large males), except in E. grahami and E. pottsi; prevomer and palatine with teeth; adult males average larger than females. Lateral line not arched upward anteriorly, incomplete (40-90% scales in lateral-line row pored); infraorbital and supratemporal canals uninterrupted or interrupted; modally 10 preoperculoman- dibular pores; modally six branchiostegal rays; premaxillary frenum present; modally two anal spines (one in E. australe); supraoccipital unscaled; flesh opaque; nape and breast fully scaled to unscaled; branchiostegal membranes narrowly to moderately joined across isthmus; breeding tubercles absent or present on body scales and/or fins of male; branchiostegal membranes unscaled; caudal peduncle moderate to deep (mean depth/standard length more than 0.08); anal rays modally fewer than 11; mean body depth/standard length more than 0.145; interorbital wide (mean interorbital width/head width more than 0.22); snout not long in relation to body depth (mean snout length/body depth less than 0.53); in females, second dorsal fin not short (mean fin length/ distance from front of base of second dorsal fin to caudal base more than 0.535); pelvic fins not short (mean fin length/standard length more than 0.16); gape not wide in relation to snout length (mean gape width/snout length less than 1.2); anus not surrounded by villi; interpelvic fin width/ pelvic fin base length less than 0.675; genital papilla of breeding male not long and tubular, of female tubular or conical; eggs buried 40 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY in substrate (known for three species) or attached to plants (E. lepidum and E. grahami). Catonotus Agassiz INCLUDED SPECIES.—E. flabellare, E. barbouri, E. kenni- cotti, E. neopterum, E. obeyense, E. olivaceum, E. smithi, E. squa- miceps, E. striatulum, E. virgatum. Catonotus, as a subgenus of Etheostoma, was diagnosed by Kuehne and Small (1971); that diag- nosis was modified by Braasch and Page (1979). DIAGNOSIS.—Genital papilla of female broad and flat, not bifurcate; eggs attached to underside of stone (known for nine species); supratemporal canal interrupted (except in occasional E. flabellare ). Lateral line not arched upward anteriorly, incomplete (10-75% scales in lateral-line row pored); infraorbital canal interrupted (un- interrupted in E. neopterum and E. olivaceum); modally nine or 10 preoperculomandibular pores; prevomer and palatine with teeth; modally six branchiostegal rays; premaxillary frenum present; mod- ally two anal spines; supraoccipital unscaled; flesh opaque; nape and breast fully scaled to unscaled; branchiostegal membranes sep- arate to broadly joined across isthmus; breeding tubercles absent; branchiostegal membranes unscaled; caudal peduncle moderate to deep (mean depth/standard length more than 0.08); anal rays modally fewer than 11; mean body depth/standard length more than 0.145; interorbital wide (mean interorbital width/head width more than 0.22); snout not long in relation to body depth (mean snout length/body depth less than 0.53); in females, second dorsal fin not short (mean fin length/distance from front of base of second dorsal fin to caudal base more than 0.535); pelvic fins not short (mean fin length/standard length more than 0.16); gape not wide in relation to snout length (mean gape width/snout length less than 1.2); adult males average larger than females; anus surrounded by villi; interpelvic fin width/pelvic fin base length less than 0.675; genital papilla of breeding male not long and tubular. Boleichthys Girard INCLUDED SPECIES.—E. exile, E. collis, E. fonticola, E. fusi- forme, E. gracile, E. microperca, E. proeliare, E. saludae, E. serri- ferum, E. zoniferum. These species were formerly in the subgenera Hololepis (diagnosed as a subgenus of Etheostoma by Collette, 1962), Microperca (diagnosed as a subgenus of Etheostoma by Burr, 1978) and Oligocephalus. DIAGNOSIS.—Lateral line distinctly arched upward anteriorly or absent (not always arched in E. exile); adult females average larger than males. Lateral line incomplete (0-65% scales in lateral-line row pored); infraorbital canal interrupted or uninterrupted; supratemporal canal THE GENERA AND SUBGENERA OF DARTERS 4] interrupted or uninterrupted; modally six-10 preoperculomandibu- lar pores; prevomer and palatine with teeth; modally six branchi- ostegal rays (five in E. microperca); premaxillary frenum present; one or two anal spines (one in E. proeliare and E. fonticola); supra- occipital region unscaled to scaled; flesh opaque; nape fully scaled to unscaled; breast fully scaled to unscaled; branchiostegal mem- branes narrowly to moderately joined; breeding tubercles present on anal and pelvic fins of male (absent in E. exile); branchiostegal membranes unscaled; caudal peduncle moderate in depth (mean depth/standard length = 0.08-0.11); anal rays modally fewer than 11; mean body depth/standard length more than 0.145; interorbital wide (mean interorbital width/head width more than 0.22; snout not long in relation to body depth (mean snout length/body depth less than 0.53); in females, second dorsal fin not short (mean fin length/distance from front of base of second dorsal fin to caudal base more than 0.535); pelvic fins not short (mean fin length/ standard length more than 0.16); gape not wide in relation to snout length (mean gape width/snout length less than 1.2); anus not surrounded by villi; interpelvic fin width/ pelvic fin base length less than 0.675; genital papilla of breeding male not long and tubular, of breeding female conical, tubular or bilobed (but not flat and broad as in Boleosoma); eggs attached to plants (known for six of 10 species). PHYLOGENY The modified midventral scales of Percina are a derived charac- teristic setting apart all species in that genus from those in other darter genera. Although modified scales theoretically could have originated more than once among darters, the concurrence in dis- tribution among darters of the modified scales and the “Percina mobility” of the LDH By, isozyme establishes beyond reasonable doubt that Percina is monophyletic (Page and Whitt, 1973b). Ammocrypta and Etheostoma are not as easily diagnosed from one another as they are from Percina and the species from both genera could be included in one genus [rejecting the suggestion of Williams (1975) that Ammocrypta is derived from the subgenus Imostoma of Percina]. However, assuming that Ammocrypta di- verged from Etheostoma prior to diversification within Etheostoma the recognition of Ammocrypta is satisfactory. It is within Etheostoma that phylogenetic relationships are diff- cult to decipher. However, the subgeneric names presumably label the major avenues of diversification and certain characteristics pro- vide some additional information on intersubgeneric relationships. Five major divisions can be recognized within Etheostoma: (1) Boleichthys (and perhaps Villora), (2) Psychromaster, (3) Lito- cara, (4) Etheostoma, Nanostoma, Doration, Boleosoma, Ioa, Vail- 42 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY lantia (and perhaps Allohistium), (5) Nothonotus, Fuscatelum, Belophlox, Ozarka, Oligocephalus, Catonotus (and perhaps Villora and Allohistium). Boleichthys is set apart from all other subgenera by the combi- nation of an arched lateral line (shared with Villora) and females which average larger than males. The serrae on the margin of the preopercle of E. (Boleichthys) serriferum and some individuals of E. (B.) fusiforme, E. (B.) collis (Collette, 1962) and E. (B.) proeliare (Burr, 1978) are essentially identical to the serrae of Hadropterus, the primitive subgenus of Percina, and of Perca and Stizostedion. Other species have crenulate margins on the preopercle but among Etheostoma only species of Boleichthys have serrae. The presence of this plesiomorphous characteristic suggests that E. serriferum, as the most primitive species of Boleichthys, is near the origin of an early phyletic line, and it is not likely that Boleichthys descended from any other extant group of Etheostoma. It is remarkable that Boleichthys contains what appears to be one of the most primitive species of Etheostoma, serriferum, and one considered to be the most highly evolved, microperca (Bailey and Gosline, 1955; Burr, 1978). E. (Psychromaster) tuscumbia is the only species in Etheostoma in which scales are present on the branchiostegal membranes and in which the breeding male has a long tubular genital papilla. E. (Litocara) nianguae and sagitta share a distinctive physiog- nomy (long head and snout, fusiform body) and color pattern (two vertically aligned jet-black spots on the caudal peduncle, large U-marks on the side of the body), and are the only species of Etheostoma modally to have more than 10 anal rays. Boleosoma, Ioa and Vaillantia share a tendency toward reduc- tion to one anal spine and have similar pigmentation. These three subgenera, and Etheostoma, Nanostoma and Doration share the tendency toward reduction and loss of the premaxillary frenum. They and Allohistium contain the only species of darters to have lost the palatine teeth. The remaining subgenera (Nothonotus, Fuscatelum, Belophlox, Ozarka, Oligocephalus, Catonotus and perhaps Villora) are highly evolved. They may be interrelated to one another or they may be polyphyletic. Several of the advanced subgenera have breeding tubercles (Collette, 1965) and reductions in the lateralis system (Page, 1977); however, these characteristics have evolved independ- ently among darters many times and impart little information about intersubgeneric relationships. Further elucidation of relationships among darters must await the accumulation of data beyond those presently available. THE GENERA AND SUBGENERA OF DARTERS 43 SUMMARY In Bailey’s classification of darters (Bailey and Gosline, 1955), three genera and 22 subgenera (eight in Percina, two in Ammo- crypta, and 12 in Etheostoma) were recognized. Since 1955 one additional subgenus has been recognized in Percina (Page, 1974) and five additional subgenera have been recognized in Etheostoma (Collette and Yerger, 1962; Cole, 1967; Page and Whitt, 1973a; Williams and Robison, 1980) (Table 1). In the present study, phenetic and cladistic procedures were used to examine relationships among 142 species of darters. Among the procedures used, UPGMA—correlation coefficient clustered the OTU’s well (i.e. formed large clusters of small clusters) and gave results most congruent with the existing classification. Other meth- ods were less satisfactory. Phenograms and cladograms were com- pared to the existing classification of darters and discordances were considered potential errors in classification and subjected to further analysis. The three-genus classification of darters was supported. The following changes were made in the subgeneric classification of Etheostoma. 1. Nanostoma was recognized as a subgenus for E. zonale. 2. Ulocentra was synonymized with Nanostoma. 3. E. juliae, formerly in Oligocephalus, was transferred to Nothonotus. 4. Villora was reduced to a monotypic subgenus containing E. edwini. 5. Fuscatelum, new subgenus, was described for E. parvipinne. 6. Belophlox was recognized as a subgenus for E. mariae, E. fricksium and E. okaloosae. 7. Austroperca was synonymized with Oligocephalus. 8. Boleichthys was recognized as a subgenus for E. exile. 9. Hololepis and Microperca were synonymized with Bo- leichthys. The three genera and all subgenera for which adequate diag- noses have not been published were diagnosed. Phylogenetic re- lationships among the genera of darters and among the subgenera of Etheostoma were discussed. LITERATURE CITED Bary, R. M. 1951. A check list of the fishes of Iowa, with keys for identi- fication, pp. 185-237. In: Iowa fish and fishing, Harlan, J. R. and Speaker, E. B. eds. Iowa Conservation Commission, Des Moines. Baiwtey, R. M. 1959. Etheostoma acuticeps, a new darter from the Tennessee River system, with remarks on the subgenus Nothonotus. Occ. Pap. Mus. Zool. Univ. Mich. (603):1-10. Baitey, R. M., and Gosiine, W. A. 1955. Variation and systematic signif- Ad OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY icance of vertebral counts in the American fishes of the family Percidae. Misc. Publ. Mus. Zool. Univ. Mich. (93) :5-44. Baitey, R. M., and Ricuarps, W. J. 1963. Status of Poecilichthys hopkinsi Fowler and Etheostoma trisella, new species, percid fishes from Ala- bama, Georgia, and South Carolina. Occ. Pap. Mus. Zool. Univ. Mich. (630) :1-21. : BatLey, R. M., Winn, H. E., and Smirn, C. L. 1954. Fishes from the Escam- bia River, Alabama and Florida, with ecologic and taxonomic notes. Proc. Acad. Nat. Sci. Phila. 106:109-164. Boscuunc, H. 1979. Report on the breeding habitats of the slackwater darter (Percidae: Etheostoma boschungi) in the Cypress Creek watershed. Submitted to U.S.D.A. Soil Cons. Service, Auburn, Ala. 26 pp. Boucuarp, R. W. 1977. Etheostoma etnieri, a new percid fish from the Caney Fork (Cumberland) River system, Tennessee, with a redescription of the subgenus Ulocentra. Tulane Stud. Zool. Bot. 19:105-130. Braascn, M. E., and Pace, L. M. 1979. Systematic studies of darters of the subgenus Catonotus (Percidae), with the description of a new species from Caney Fork, Tennessee. Occ. Pap. Mus. Nat. Hist. Univ. Kan. (78):1-10. ~ Burr, B. M. 1978. Systematics of the percid fishes of the subgenus Micro- perca, genus Etheostoma. Bull. Alabama Mus. Nat. Hist. (4):1-53. Cotr, C. F. 1967. A study of the eastern johnny darter, Etheostoma olmstedi Storer (Teleostei, Percidae). Chesapeake Sci. 8:28-51. Co.t.etTeE, B. B. 1962. The swamp darters of the subgenus Hololepis (Pisces, Percidae). Tulane Stud. Zool. 9:115-211. Co.tetTe, B. B. 1965. Systematic significance of breeding tubercles in fishes of the family Percidae. Proc. U.S. Nat. Mus. (3518) :567-614. Co.tetTeE, B. B., and Yercer, R. W. 1962. The American percid fishes of the subgenus Villora. Tulane Stud. Zool. 9:213-230. Ernier, D. A. 1976. Percina (Imostoma) tanasi, a new percid fish from the Little Tennessee River, Tennessee. Proc. Biol. Soc. Wash. 88:469-488. Farris, J. S. 1970. Methods for computing wagner trees. Syst. Zool. 19:83-92. Farnis, J. S. 1972. Estimating phylogenetic trees from distance matrices. Amer. Nat. 106:645-668. Farris, J. S., Ktuce, A. G., and Ecxarpt, M. J. 1970. A numerical approach to phylogenetic systematics. Syst. Zool. 19:172-189. Hennic, W. 1966. Phylogenetic systematics. Univ. Ill. Press, Urbana. 263 pp. Howe.i, W. M. 1968. Taxonomy and distribution of the percid fish, Ethe- ostoma stigmaeum (Jordan), with the validation and redescription of Etheostoma davisoni Hay. Unpubl. PhD Diss. Univ. Alabama. 113 pp. Husps, C. L., and Lacier, K. F. 1958. Fishes of the Great Lakes Region. Cranbrook Inst. Sci. Bull. 26. 213 pp. Jenkins, R. E, 1971. Nuptial tuberculation and its systematic significance in the percid fish Etheostoma (loa) vitreum. Copeia 1971:735-738. Jornpan, D. S., and EverMann, B. W. 1896. The fishes of North and Middle America: a descriptive catalogue of the species of fish-like vertebrates found in the waters of North America north of the Isthmus of Panama. U.S. Nat. Mus. Bull. 47, pt. 1. 1240 pp. Kiuce, A. G., and Farris, J. S. 1969. Quantitative phyletics and the evolu- tion of anurans. Syst. Zool. 18:1-32. Kuenne, R. A., and Sma, J. W., Jn. 1971. Etheostoma barbouri, a new darter (Percidae, Etheostomatini) from the Green River with notes on the subgenus Catonotus. Copeia 1971:18-26. MaypveNn, R. L., and Pace, L. M. 1979. Systematics of Percina roanoka and THE GENERA AND SUBGENERA OF DARTERS 45 P. crassa, with comparisons to P. peltata and P. notogramma ( Percidae). Copeia 1979:413-426. Micxevicu, M. F. 1978. Taxonomic congruence. Syst, Zool. 27:143-158. Miuier, R. J., and Ropison, H. W. 1973. The fishes of Oklahoma. Okla. State Univ. Press, Stillwater. 246 pp. Moore, G. A. 1968. Fishes, pp. 21-165. In: Vertebrates of the United States, Blair, W. F., Blair, A. P., Brodkorb, P., Cagle, F. R., and Moore, G. A., eds. McGraw-Hill, New York. Pace, L. M. 1974. The subgenera of Percina (Percidae: Etheostomatini). Copeia 1974:66-86. Pace, L. M. 1976. The modified midventral scales of Percina (Osteichthyes; Percidae). J. Morph. 148:255-264. Pace, L. M. 1977. The lateralis system of darters (Etheostomatini). Copeia 1977:472-475. Pace, L. M., and Wuirr, G. S. 1973a. Lactate dehydrogenase isozymes, malate dehydrogenase isozymes, and tetrazolium oxidase mobilities of darters (Etheostomatini). Comp. Biochem. Physiol. 44B:611-623. Pace, L. M., and Wurrt, G. S. 1973b. Lactate dehydrogenase isozymes of darters and the inclusiveness of the genus Percina. Ill. Nat. Hist. Sur. Biol. Notes 82:1-7. Ricuarps, W. J. 1966. Systematics of the percid fishes of the Etheostoma thalassinum species groups with comments on the subgenus Etheostoma. Copeia 1966:823-838. SneaTH, P. H. A., and Soxan, R. R. 1973. Numerical taxonomy. W. H. Freeman and Co., San Francisco. 573 pp. STEVENSON, M. M. 1971. Percina macrolepida (Pisces, Percidae, Etheostoma- tinea), a new percid fish of the subgenus Percina from Texas. South- west. Nat. 16:65-83. Tsar, C., and Raney, E. C. 1974. Systematics of the banded darter, Ethe- ostoma zonale (Pisces: Percidae). Copeia 1974:1-24. Witey, E. O. 1975. Karl R. Popper, systematics, and classification: a reply to Walter Bock and other evolutionary taxonomists. Syst. Zool. 24:233- 243. WiiuraMs, J. D. 1975. Systematics of the percid fishes of the subgenus Am- mocrypta, genus Ammocrypta, with descriptions of two new species. Bull. Alabama Mus. Nat. Hist. (1):1-56. WituraMs, J. D., and Rosison, H. W. 1980. Ozarka, a new subgenus of Etheostoma (Pisces:Percidae). Brimleyana 4:149-156. Win, H. E., and Piccioto, A. R. 1960. Communal spawning of the glassy darter Etheostoma vitreum (Cope). Copeia 1960:186-192. ZoracH, T. 1972. Systematics of the percid fishes, Etheostoma camurum and E. chlorobranchium new species, with a discussion of the subgenus Nothonotus. Copeia 1972:427-447. APPENDIX A.—SPECIMENS EXAMINED To obtain data for the numerical analysis 1,379 specimens representing 142 species were examined. The number of specimens per species ranged from 6 to 10 and averaged 9.7. The following specimens were examined. Institution acronyms are identi- fied in the acknowledgements. Numbers in parentheses are numbers of speci- mens examined. Locality data are available from the author. Percina antesella: UT 91.471 (10); P. aurantiaca: UT 91.503 (4), 91.502 (6); P. aurolineata: TU 37676 (4), 40800 (2), 38197 (2); P. burtoni: TU 71983 (2), 72133 (1), 89496 (1), REJ—Copper Creek (6); P. caprodes: 46 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY INHS 75649 (10); P. carbonaria: TU 27454 (8); P. copelandi: INHS 75555 (10); P. crassa: INHS 75880 (6), NCMNH 3010 (4); P. cymatotaenia: INHS 75290 (10); P. evides: INHS 76728 (3), 76706 (7); P. lenticula: TU 60556 (2), 59560 (2), 58807 (3), 53715 (3); P. macrocephala: INHS 76641 (9), UMMZ 86386 (1); P. macrolepida: INHS 74774 (3), TU 64393 (2), 67567 (5); P. maculata: INHS 8541 (10); P. nasuta:' OAM 4586 (2), 4843 (1), UTULSAC—Lee Creek (2), INHS 74993 (1), TU 72243 (4); P. nigrofasciata: INHS 74811 (2), 74800 (2), 76789 (3), 74816 (3); P. notogramma: USNM 107746 (10); P. ouachitae: INHS 76753 (8), 76766 (2), P. oxyrhyncha: UMMZ 17663 (6); P. palmaris: INHS 76305 (5), 76808 (2), 76806 (2), 76790 (1); P. pantherina: UTULSAC—Little River (6), Mountain Fork (1), Sixmile Creek (1), OAM 2559 (1); P. peltata: INHS 76809 (3), UMMZ 147594 (7); P. phoxocephala: INHS 26922 (10); P. rex: REJ—South Fork Roanoke R. (10); P. roanoka: INHS 75884 (10); P. sciera: INHS 26923 (10); P. shumardi: INHS 76783 (10); P. squamata: INHS 76768 (10); P. tanasi: INHS 75000 (5), TU 99120 (3); P. uranidea: TU 66011 (10); P. (Alvordius) species (Conasauga River): UT 91.501 (7), INHS 76807 (1), 76788 (2); P. (Cottogaster) species (Pearl River): TU 17732 (8); P. (Odontopholis) species (Barren River): INHS 76784 (10); P. (Percina) species (Conasauga River): TU 18369 (2), 69140 (5); Ammocrypta asprella: INHS 74792 (10); A. beani: UMMZ 170650 (10); A. bifascia: INHS 74773 (9); A. clara: INHS 13390 (10); A. meridiana: UMMZ 197722 (10); A. pellucida: INHS 17061 (4), 26929 (2), 16819 (2), 16794 (2); A. vivax: INHS 74946 (6), 74618 (4); Etheostoma acuticeps: INHS 75149 (9); E. asprigene: INHS 6116 (10); E. atripinne: INHS 74923 (10); E. australe: UMMZ 182378 (10); E. barbouri: INHS 74894 (7), 74906 (3); E. bellum: INHS 74865 (4), 74850 (3), 74878 (3); E. blennioides: INHS 74837 (10). E. blennins: UMMZ 168121 (10); E. boschungi: UMMZ 197692 (2), 197693 (2), 197691 (3); E. caeruleum: INHS 7205 (10); E. camurum: INHS 11816 (6), 11598 (2), 11815 (2); E. chlorobranchium: TU 96323 (10); E. chlorosomum: INHS 74818 (10); E. cinereum: UMMZ 171557 (5), 175061 (3), 171590 (2); E. collettei: INHS 76594 (10); E. collis: TU 71881 (7); E. coosae: UMMZ 190910 (10); E. cragini: INHS 74817 (10); E. davisoni: TU 83098 (10); E. ditrema: UMMZ 187501 (5), TU 32762 (3); E. duryi: INHS 77542 (10); E. edwini: INHS 74698 (8), 76375 (1); E. etnieri: TU 83149 (8); E. eu- zonum: UMMZ 124595 (8); E. exile: INHS 26934 (10); E. flabellare: INHS 22414 (10); E. fonticola: INHS 75562 (10); E. fricksium: TU 97596 (10); E. fusiforme: INHS 74269 (10); E. gracile: INHS 26933 (10); E. grahami: USNM 189023 (10); E. histrio: INHS 9617 (4), 16735 (6); E. hopkinsi: TU 39482 (10); E. inscriptum: INHS 74495 (10); E. jessiae: TU 34964 (10); E. jordani: INHS 74586 (10); E. juliae: INHS 74214 (4), 74228 (2), 74247 (4); E. kanawhae: UMMZ 169360 (10); E. kennicotti: INHS 1592 (10); E. lepidum: INHS 75687 (10); E. longimanum: USNM 162877 (10); E. luteovinctum: INHS 74559 (7), 74566 (2), 74579 (1); E. maculatum: INHS 74596 (10); E. mariae: INHS 75194 (9); E. microlepidum: UMMZ 168392 (10); E. microperca: INHS 4760 (10); E. moorei: UMMZ 181397 (10); E. neopterum: INHS 74278 (10); E. nianguae: INHS 74395 (4), 74397 (3), 74398 (2); E. nigrum: INHS 12873 (10); E. nuchale: USNM 217856 (10); E. obeyense: USNM 204345 (10); E. okaloosae: UMMZ 178859 (8), INHS 74688 (2); E. olivaceum: INHS 75879 (2), 75878 (8); E. olm- stedi: INHS 74543 (10); E. osburni: UMMZ 165712 (9); E. pallididorsum: KU 6921 (10); E. parvipinne: INHS 74544 (10); E. perlongum: UMMZ 138476 (10); E. podostemone: USNM 162031 (10); E. pottsi: INHS 75300 (10); E. proeliare: INHS 1358 (10); E. punctulatum: UMMZ 137836 (10); E, radiosum: INHS 76580 (8), 76438 (2); E. rubrum: INHS 74325 (9); E. THE GENERA AND SUBGENERA OF DARTERS AT rufilineatum: INHS 74309 (10); E. rupestre: INHS 76339 (10); E. sagitta: UL 5566 (10); E. saludae: USNM 196377 (10); E. sellare; USNM 212147 (10); E. serriferum: UMMZ 175866 (7), INHS 74361 (2); E. simoterum: INHS 74300 (10); E. smithi: INHS 75017 (10); E. spectabile: INHS 22556 (10); E. squamiceps: INHS 26936 (10); E. stigmaeum: INHS 74292 (10); E. striatulum: INHS 75037 (10); E. swaini: INHS 74378 (10); E. swanna- noa: UMMZ 156073 (10); E. tetrazonum: INHS 76732 (10); E. thalassinum: INHS 74367 (2), 74735 (3), UMMZ 183619 (5); E. tippecanoe: INHS 74280 (10); E. trisella: TU 58963 (10); E. tuscumbia: USNM 217855 (9); E. vari- atum: UL 12459 (10); E. virgatum: NLU 10837 (10); E. vitreum: INHS 74355 (10); E. whipplei: UMMZ 177160 (9); E. zonale: INHS 76727 (10); E. zoniferum: TU 98971 (10); E. (Catonotus) species (Copper Creek, Scott Co., VA): TU 71976 (10); E. (Doration) species (Caney Fork): UMMZ 187465 (10); E. (Nanostoma) species 1 (Barren River): UMMZ 177622 (10); E. (N.) species 2 (Kentucky River): UMMZ 177854 (10); E. (N.) species 3 (Forked Deer River): UMMZ 177684 (8); E. (N.) species 4 (Sycamore Creek, Benton Co., TN): INHS 74279 (10); E. (N.) species 5 (Tallapoosa River): UMMZ 177758 (10); E. (N.) species 6 (Cumberland River): UMMZ 175042 (10); E. (N.) species 7 (Green River): UMMZ 177559 (10); E. (N.) species 8 (Black Warrior River): UAIC 3804 (10); E. (N.) species 9 (Cahaba River): UMMZ 168646 (10). APPENDIX B.—CHARACTERS USED IN NUMERICAL ANALYSIS Character 1.—Number of modified scales in single midbelly row from anus to pelvic fins (males). Character 2.—Number of lateral scales (in row from opercle to end of hypural plate). Character 3.—Percent of scales in Jateral-line row that are pored. Character 4—Number of scale rows above lateral line (or above mid- lateral row of scales if line absent). Character 5.—Number of scale rows below lateral line (or below mid- lateral row of scales if line absent). Character 6.—Number of scale rows around caudal peduncle. Character 7—Number of pored scales on caudal fin. Character 8.—Number of spines in first dorsal fin. Character 9.—Number of rays in second dorsal fin. Character 10.—Number of spines in anal fin. Character 11.—Number of rays in anal fin. Character 12.—Number of rays in pectoral fin. Character 13.—Number of transverse scale rows (from anal fin origin an- terodorsally to first dorsal fin). Character 14.—Modal number of infraorbital pores. Character 15.—Modal number of preoperculomandibular pores. Character 16.—Modal number of branchiostegal rays. Character 17.—Modal number of supratemporal pores. Character 18.—Interpelvic fin width/pelvic fin base length (IP2/P2bL). Character 19.—Head length/standard length (HL/SL). Character 20.—Head width/standard length (HW/SL). Character 21.—Body depth/standard length (BD/SL). Character 22.—Caudal peduncle depth/standard length (CPD/SL). Character 23.—Pectoral fin length/standard length (P1L/SL). Character 24.—Predorsal length/standard length (PreDL/SL). Character 25.—Snout (preorbital) length/body depth (SnL/BD). 48 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Character 26.—Length of second dorsal fin base/length of first dorsal fin base (D2bL/DI1bL). Character 27.—Interorbital width/head width (IOW/HW). Character 28.—Gape width/snout length (GW/SnL). Character 29.—Body depth/anal fin length (males) (BD/AL). Character 30.—Second dorsal fin length/distance from front of base of sec- ond dorsal fin to caudal base (males) (D2L/QL¢4 ). Character 31.—Second dorsal fin length/distance from front of base of sec- ond dorsal fin to caudal base (females) (D2L/QL@ ). Character 32.—Pelvic fin length/standard length (P2L/SL). Character 33.—Caudal fin length/standard length (CFL/SL). Character 34.—Flesh opaque or translucent when alive. 1 = opaque; 2 = translucent. Character 35.—Pigment bar on cheek as in E. virgatum (see photo in Kuehne and Small, 1971). 1 = absent; 2 = present. Character 36.—Breast squamation of males. 1 = fully scaled; 2 = reduced squamation, including scales embedded (neither fully scaled nor unscaled); 3 = unscaled. ; Character 37.—Breast squamation of females. 1 = fully scaled; 2 = re- duced squamation (neither fully scaled nor unscaled); 3 = unscaled. Character 38.—Anterior belly squamation of males. 1 = area of belly im- mediately behind pelvic fins unscaled; 2 = area of belly immediately behind pelvic fins scaled. Character 39.—Belly midline sacenibin of males. 1 = complete or in- complete rows of unmodified scales; 2 = scaleless strip(s) from pelvic fins to genital papilla; 3 = with a midbelly row of modified (enlarged and strongly toothed) scales on which the longest tooth is always less than 40% of the entire length of the modified scale; 4 = with a midbelly row of modified scales on which the teeth are sometimes longer than 40% of the entire length of the scales. Character 40.—Belly midline squamation of females. 1 = fully scaled; 2 = variable, from almost fully scaled to a naked strip extending one-half dis- tance from genital papilla to pelvic fins; 3 = variable, from almost unscaled to a naked strip extending one-half of distance from genital papilla to pelvic fins; = unscaled from pelvic fins to genital papilla. Character 41.—Uninterrupted interpelvic row of modified scales extending from approximately midpectoral area onto anterior part of belly (males). 1 = absent; 2 = present. Character 42.—Conical snout projecting well beyond premaxilla. 1 = ab- sent; 2 = present. Character 43.—Premaxillary frenum. 1 = absent or extremely narrow; 2 = present and not extremely narrow. Character 44,—Branchiostegal membrane connection. 1 = membranes not or only slightly joined; 2 = membranes narrowly to moderately joined; 3 = membranes broadly joined. Character 45.—Lateral line. 1 = straight (complete or incomplete) or slightly arched; 2 = arched distinctly upward anteriorly. Character 46.—Pelvic fins of breeding male. 1 = without large lateral flaps; 2 = with large lateral flaps. Character 47,—Maximum size (total adult length). 1 = up to 55 mm; 2 = to 80 mm; 3 = to 105 mm; 4 = to 130 mm; 5 = to 155 mm; 6 = to 180 mm; 7 = over 180 mm. Character 48.—Breeding tuberculation (including tubercular ridges). 1 = absent; 2 = present on males only; 3 = present on males and females (data mostly from Collette, 1965). Character 49.—Cephalic canals. 1 = uninterrupted; 2 = supratemporal THE GENERA AND SUBGENERA OF DARTERS 49 canal only interrupted; 3 = infraorbital canal only interrupted; 4 = both supratemporal and infraorbital canal interrupted. Character 50.—Dentition. 1 = teeth present on prevomer and palatine; 2 = present on palatine and absent on prevomer; 3 = present on prevomer and absent on palatine; 4 = absent on both prevomer and palatine (data mostly from Richards, 1966). Character 51.—Flesh over anterior half of maxillaries fused with skin over preorbitals. 1 = absent; 2 = present. Character 52.—Anus surrounded by fleshy villi. 1 = absent; 2 = present. OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 50 OLT 6ET Vil Oe ove Op £0 TGs T6 T9 0°00T oes Pel ISDUD} “qd £0€ L'€T 8 0% T6l O'eT iAv GSE €or 6 IT 0°00T 0°08 706 DypuDpnbs “gq GST Vel OTT 0@ SCI LOt L‘0 L'6G 88 £9 0°00T S"GS SL ipspUnys “qd G06 ial! sie) 0% jai esl oT T&G Cel OL 0°00T reo ost DIBI98 “d Gol -= O71 v8 0% jay 801 0 T9T v9 0's 0'O0T LLY 8 DYOUDOL d S86 TV¥t HOt Oe @ST ae v0 OEE GOT Lot 0°00T 698 9°9G xas ‘d £€6 VV ~~ €i6 0G 6 IT OTT TO L'SG TPT £6 0°00T 9°99 S16 pjpydasoxoyd “gq LT 6€T 66 0G 6CT 6 - ian Se Ur 166 = 0'9F ~=—0'0_~——sdT. “ds (pwojsounyy) -y EOL -S'L. 82. 08 Ol DT 00 “Fi es 6's 6°96 T9F = 0'0 ‘ds (uonnsog) “gq key ORT 2. 02° @Ir -i¢o 00 97% SL 9F ELS For 8600 ‘ds (snjouow9) “gq en 21 85 G2 OT oF 00 DIS 6's SPL sr 00 wniajdoau “y Bel 62. £0 Gi. Zor O16 00 613 68 re rast O6F 00 wuniafuor “J Syl hl Sk Of LIL 0% OO oor 9% Zs 8°66 esr 00 ajDuoz “J aoe Tel O8 OC EST. GOI G0 Oo ‘sat 68 cL 999 00 1ajddiyon “gq col 63l 36 O83 ESI F8 00 008 gz Lg 8°66 slg 00 winaina “J StI Stl S88 O83 Tel ¢6 OO Ber FZ so O'FE 98F 00 wnpo3i1a “7y COL “ORT 86: Of TS C2 oO eee “Taco-ok Uae ees 00 wnyoupa “J Gol. Sib 88: OL “Peres 00 €8 06 CL T'e3 TLr O00 piquinasny “y PH CKO O88 COL. «COE «Oe «~—CC ee rg O00I Flr O00 pyjasuua“y Ler Sel oS 0% SIF Sir 00 “om 279 S73 8°69 OSF 00 aouvsaddy *q BCL -Oet 8 0% ST - be 00. On -.o Sr SsdOOT,—S—s SS FF__sC*'10 wnuIsso]DYr “3 al ZI II O1 6 8 L 9 S p € Z I saad si9qoBIeYD (penunuoy) ‘sIsAjvuy [voLoUINN UT pasy) saje}g Jo;LVIVYQ—'D XIGNaday 55 THE GENERA AND SUBGENERA OF DARTERS €660 9FF0 PEO FPCO FLOO O8T0 8FI0 6LZ0 F69°0 € 9 Or 8 gsco)6d~TS'0—ssLvE'O «=66TO «66800 )«€66ETO 6CS0T0 =«6S8G0 866060 € 9 Or 8 8980 LOVO 6&¢0 €&c0 c600 GLIO PSITO [L060 S10 € 9 Or 8 Iss0 OF€O0 6180 FOGO 9800 SLIO 8sc10 SrZO0 6SL0 € 9 OT 8 csc0 6480 )=— SEO =6FPHGO ~OOTO «6610 )6SHITO 61930 6960 € 9 Or 8 L990 «1990 6180 1660 7r800 6910 LZI0 19¢0 8LL0 € 9 OT 8 Or90 es Uae - : aaa Jaen! seve nf ‘ PRA Aas os SC Manik a batty, sad cane ny wit ‘Sivep Wy twin Jenni eof wolitesdahn a 7 ceed pom, “ear: re ae Pan isi d: 3 ck “phi Nie ne. la a Pap cont ys iy att oy! . cs Seen rare TieW #2 d tags y reid OA A Fanta |, hr708 2 Saban i a oe Ms Sota he 4. 14 = 1 pean riarl a = htt a>, aris aes ay’ Mea ow: he Ari )i Faiat at? “3 rere ,; a me We Ag Wie) rey ve eae Nine Orirets, obec bw uy be SES oe eee pba e, iid ielbeahiles adh Bea ae 2° | a ® aus ati sy bln nee Bw teeeerw iy eek ice — aiiiay Feaas). Atl pee. eae wi ety Sse) (danny Porietle aie ame. ere ie

po pf=po Pf>po Pf>po->pf=po pf>po pfpo, 1.0-40, 6.0, 7.0 Ficure 3.—Interrelationship of the families of the suborder Tuditano- morpha. pf postfrontal, po postorbital. (1) Position of the jaw articulation: (1.0) at the level and (1.1) in front of the occipital condyle-atlas articulation. (2) Posterior margin of the cheek region: (2.0) straight, (2.1) inclined back- ward, (2.2) inclined forward. (3) Quadratojugal: (3.1) large, (3.1) small, (3.2) elongated. (4) Epicondylar foramen: (4.0) present, (4.1) absent. (5) Margin of the upper jaw: (5.1) posterior margin inclined, (5.2) ventral extension, (5.3) deep emargination up to tabular, (5.4) straight, (5.5) low emargination. (6) Posterior margin of skull table: (6.0) concave, (6.1) straight, (6.2) convex, triangular, (6.3) undulated, emargination at the occiput. (7) Supraoccipital: (7.0) present, (7.1) absent. (8) Number of presacral vertebrae: (8.0) below 30, (8.1) 37 and more. 10 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Thus, the postfrontal extends about as far posterior as does the postorbital in the Tuditanomorpha, whereas it is limited to the region close to the orbit in the Microbrachomorpha. The post- frontal is larger than the postorbital in primitive Tuditanomorpha, while the postfrontal is smaller than the postorbital in the Micro- brachomorpha. In both groups, these bones tend to become the same size. The postorbital becomes larger than the postfrontal in the specialized tuditanomorph Ostodolepidae. The evolutionary tendency towards enlargement of the postorbital is quite evident within the Tuditanomorpha, but the character is too transitional to be used for characterizing any family. (1) Position of the jaw articulation: The jaw articulation lies well behind the occiput in anthracosaurs, but on the level with or in front of the articulation between occiput and first vertebra in most microsaurs. Only some Tuditanidae (Tuditanus, Carroll and Gaskill, 1978, fig. 4) show the more primitive situation with the jaw articulation a little further posterior. All Pennsylvanian Tuditano- morpha and the Permian genera Saxonerpeton and Pantylus have both articulations on the same level. The anterior position of the jaw articulation is acquired in Hapsidopareiodontidae, Gymnarth- ridae, Goniorhynchidae, and Ostodolepidae and in parallel in the Microbrachomorpha. Within the Microbrachomorpha, only Hy- ploplesion has both articulations at the same level (Carroll and Gaskill, 1977, Fig. 89, E,H). (2) Posterior margin of the cheek region: Microsaurs do not possess an otic notch so that the cheek has a straight posterior margin from the skull roof down to the jaw articulation. This margin is nearly straight vertically in most microsaurs as one may expect from the relation of the posterior end of the tabular to the quadrate in anthracosaurs, considered here as the primitive sister- group of microsaurs. The occiput (thus the posterior margin of the cheek) is inclined forward in burrowing forms. This feature is more pronounced in Hapsidopareion than in Llistrofus within the Hapsidopareiontidae, and extreme in Pelodosotis within the Osto- dolepidae. In our opinion, this feature is an expression of burrow- ing adaptation rather than a synapomorphy; it occurs in the Osto- dolepidae along with a subterminal mouth. The backward inclina- tion of the margin of the cheek is a tendency within gymnarthrids and Rhynchonkos; thus, the margin is inclined forward in the largest gymnarthrid. The feature may or may not be a synapo- morphy for Rhynchonkos and some gymnarthrids. (3) Quadratojugal: The quadratojugal differs in size within microsaurs. It is large and relatively deep in all Microbracho- morpha except Odonterpeton, in Tuditanidae, and in Pantylidae. It is a large and relatively deep bone in anthracosaurs and most other Carboniferous amphibians. That seems to be the primitive situa- A NEW GYMNARTHRID MICROSAUR 11 tion. The quadratojugal becomes small in Tuditanomorpha and Microbrachomorpha (Odonterpeton), while a low elongated shape is found in gymnarthrids and goniorhynchids only; therefore, it seems to be an advanced feature for both families. (4) Entepicondylar foramen: The entepicondylar foramen is present in microsaurs as in anthracosaurs. It has, therefore, to be considered as primitive for microsaurs, and the loss as advanced. It is lost within Microbrachomorpha and within Tuditanomorpha, and it may be lost independently more than once within each suborder. Most gymnarthrids and goniorhynchids have no entepi- condylar foramen, although Carroll and Gaskill (1978) describe an isolated gymnarthrid humerus with an entepicondylar foramen. The shape of the humerus in gymnarthrids is different from other Tuditanomorpha, and resembles that of Goniorhynchidae (Carroll and Gaskill, 1978: 175). (5) Margin of the upper jaw: The ventral margin of the upper jaw is a good character to distinguish the families within the Tuditanomorpha. It changes in the extremes from a ventral ex- tension to a deep dorsal emargination. The slight ventral extension of the cheek region behind the toothed maxillary in Tuditanidae is the primitive situation judging from a similar situation in Micro- brachomorpha (Microbrachis after Steen, 1938; or Hyloplesion after Carroll and Gaskill, 1978) and anthracosaurs (Panchen, 1970). The Pantylidae evolve a broad ventral lappet, while Hapsido- pareiontidae and Ostodolepidae have an emargination instead. The emargination in both families is quite different, and very deep in Hapsidopareiontidae. Only Gymnarthridae and Goniorhynchidae seem to be comparable in the formation of the ventral margin of the cheek region: a straight margin with the tendency to be arched. (6) Posterior margin of skull table: The primitive position of the jaw articulation is posterior to the occipital region. Consequent- ly, the closing of the otic notch has imposed a concave shaped margin along the posterior edge of the skull table. In those forms in which the posterior margin of the skull table is straight, there is little difference in the configuration of the skull elements along the posterior margin when compared with the more primitive concave shape. The significant difference is the anterior placement of the quadrate in the forms with straight posterior skull table margin. Only the forms considered to be best adapted for burrowing show a different limit for attachment of the neck musculature. Thus, in Ostodolepidae (and BPI 3839, see Carroll and Gaskill, 1978), the musculture extends anterior and lateral to the occipital articulation in such a way that the posterior margin of the skull roof forms an undulating line. The margin of the skull table lies relatively far forward in Rhynchonkos, which also results from the extreme posterior position of the occipital articulation. The pos- 12 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY terior margin of the skull table in Rhynchonkos forms a shallow posterior projecting triangle with the apex nearly reaching the supraoccipital. (7) Supraoccipital: An ossified supraoccipital exists in anthra- cosaurs (Panchen, 1970), and it may be primitively present in all microsaurs even though it is not known in Microbrachomorpha. It is comparatively large in Tuditanidae, Pantylidae, and Hapsido- pareiontidae. It is reduced in Rhynchonkos and in Ostodolepidae, and missing only in the Gymnarthridae. The latter advanced feature may be an autapomorphy for Gymnarthridae distinguishing them from Rhynchonkos. (8) Number of vertebrae: A low number of trunk vertebrae (about 25) is “the normal count” (Panchen, 1970: 28) in labyrin- thodonts, although the anthracosaur Archeria has more than forty. A count below 30 is found in most microsaurs. Only Microbrachis within the Microbrachomorpha, and the Gymnarthridae, Rhyn- chonkos, and to an extreme, the Ostodolepidae within the Tudit- anomorpha exceed the number thirty. This feature is here con- sidered an advanced feature within the Tuditanomorpha. Within the microsaurs most of the above characters (1, 2, 3, 6, 7) evolve towards a reptilian state as has been observed earlier (see Carroll and Baird, 1968). These are useful characters to separate the families of Tuditanomorpha after one accepts the basic char- acters which unite the microsaurs into a monophyletic group not ancestral or closely related to reptiles. The genera of the Tudit- anidae are the most primitive ones within the suborder; the shape of the occiput, the ornamentation of the bones, the size of the quadratojugal, and the high number of marginal teeth are all characters primitive within the suborder. The Pantylidae are only a little more advanced. The ventral extension of the margin of the upper jaw is the main character of the family. The number of teeth is reduced as in no other family of the suborder (3 pre- maxillary teeth, and 8-9 maxillary teeth, contrary to 5 premaxillary and more than 16 maxillary teeth in most Tuditanomorpha) except the Gymnarthridae where the number can drop to 10 maxillary teeth in advanced forms. In addition, the configuration of the suture pattern (interdigitating sutures), the ovoid shape of the eye, and the long postorbital length justify the family. The genera of Hapsidopareiontidae are distinguished from other Tuditanomorpha by the high emargination of the cheek up to the tabular. In Osto- dolepidae, the emargination of the cheek region is not as high, and the skull is shaped quite differently. The Ostodolepidae are the most specialized forms within the Tuditanomorpha, the “extreme diggers” with pointed snout, triangular skull, and inclined occipital region. The last two families, the Goniorhynchidae and the Gym- narthridae (the Trihecatontidae are set aside because too little is A NEW GYMNARTHRID MICROSAUR 13 known about them), have many features in common (Fig. 3). Only the shape of the occiput seems to give the possibility of distinguish- ing clearly both families. The shape of the teeth and the number of teeth are characters considered to be of high value in separating both families for Carroll and Gaskill (1978: 155), but these are primitive characters (compare Saxonerpeton within the family). On the contrary, the existence of many advanced characters in both families could even be used to consider Rhynchonkos as one of the Gymnarthridae. A character obliterated many times in parallel is the loss of the pineal opening. It is lost in Pantylidae, Ostodolepidae, some Gymnarthridae (Pariotichus, Euryodus primus and E. bon- neri), and in “Micraroter’ BPI 3839 (not in the type of Micra- roter). The elongation of the trunk is a feature shown in Gonio- rhynchidae, Gymnarthridae, and Ostodolepidae. In these forms, the intercentrum is ossified also (these characters are known from the otherwise poorly known Trihecatontidae so that it can be argued at least that this family may have some relation to one of these families). Within the Gymnarthridae, Euryodus (peabodyi, primus, and dalyae) possesses strap-shaped intercentra. Vertebrae of other gymnarthrid genera are not known. The Pennsylvanian Hylerpeton shows some vertebrae formed by the pleurocentrum only so that Carroll and Gaskill (1978: 59) stated: “Their [intercentra] appar- ent absence in other forms could be attributed either to absence of ossification or postmortem disturbance. The apparent absence of intercentra in Pennsylvanian gymnarthrids . . . .” Rhynchonkos, the only genus within the Goniorhynchidae, shows small intercentra. The situation in Ostodolepidae, and in Micraroter is ambiguous. Ostodolepis itself has intercentra, but the other genus Pelodosotis within the family shows none. The type specimen of Micraroter has no intercentra, but intercentra are present in BPI 3839, a speci- men assigned to the same genus by Carroll and Gaskill (1978). Therefore, presence or absence, and shape of intercentra cannot be used to characterize families within the Tuditanomorpha. In conclusion, the six families within the Tuditanomorpha which can be defined by cranial characters are accepted by the authors. We disagree with the assignment of some genera to a particular family (see below). Gymnarthridae and Goniorhyn- chidea are the two most closely related families. The position of the seventh family, Trihecatontidae, is uncertain because of lack of data. The described specimen agrees in all its characters with the members of the Gymnarthridae. Genera Incertae Sedis Four genera within the Tuditanomorpha do not conform to the 14 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY characteristics of the families they are assigned to by Carroll and Gaskill (1978): Saxonerpeton and Ricnodon: Saxonerpeton is assigned to the Hapsidopareiontidae although it lacks the emargination of the cheek region, the most characteristic feature of the family (Carroll and Gaskill, 1978: 38). The margin of the upper jaw is straight, and the quadratojugal is elongated as in Gymnarthridae and Goniorhynchidae. The maxillary reaches the orbit as in some Gymnarthridae, and in the type of Micraroter. The arrangement of the bones of the skull roof is similar to Gymnarthridae. The postparietals seem to take more space on the skull roof (Carroll and Gaskill, 1978, fig. 19A,B, 21) than in the given restoration (Fig. 22). The prefrontal reaches very far rostrad in Gymnarth- ridae, as in Hapsidopareiontidae. A dentary forming most of the external side of the lower jaw together with a reduced angular is likewise a feature of the Gymnarthridae. The interclavicle is similar to that in the Gymnarthridae. In conclusion, the authors prefer to place this genus close to the Gymnarthridae, even though there are some primitive characters as the low number of presacral vertebrae, the high number of teeth, and an entepicondylar foramen in the humerus of Saxonerpeton (but an entepicondylar foramen is present also in an isolated gymnarthrid humerus, see Carroll and Gaskill, 1978, Fig. 41B). Carroll and Gaskill (1978) put Ricnodon very close to Sax- onerpeton, but there is very little known which could give us their proper relationship. As in Saxonerpeton, the characteristic emargi- nation of the cheek region of Hapsidopareiontidae is not known. The pelvic girdle, as in Saxonerpeton, is different from Hapsido- pareiontidae and matches with that of Rhynchonkos and the Mi- crobrachomorpha, and therefore we assume it is a primitive feature. Diagnostic features can be found in the lower jaw and the jugal only (Fritsch, 1883, Plate 42; or Carroll, 1966, Fig. 18). The sur- angular is very large; this is a primitive feature in Tuditanomorpha and found only in Tuditanidae. The ventral extension of the jugal indicates a ventral extension of the margin of the upper jaw, a diagnostic feature of Tuditanidae. If the specimens from Joggins are assigned correctly to Ricnodon (Carroll, 1966: 86-89), the two dorsal processes of the ilium, and the large pineal foramen agree with features in Tuditanidae. The position of Ricnodon is difficult to determine, but all known features would justify an assignment to the Tuditanidae; even here the existence of the entepicondylar foramen raises no problems. Trachystegos: Very little is known of this genus. It is assigned to the Pantylidae (Carroll and Gaskill, 1978) even though the extension of the margin of the upper jaw (p. 11) has not been determined. The dermal bone pattern does not resemble that of A NEW GYMNARTHRID MICROSAUR 15 Pantylus either. The prefrontal reaches far forward, the lacrimal has a dorsal process at the anterior margin of the orbit, and the maxillary reaches the ventral margin of the orbit. All these are features of the Gymnarthridae, but an entepicondylar foramen is present. The fragments of Trachystegos do not provide enough evidence to include the genus within the Gymnarthridae with cer- tainty (but see further discussion on p. 23). Micraroter: Carroll and Gaskill (1978) assign two specimens to M. erythrogeios, the type specimen described by Daly (1973) and one specimen of the Bernard Price Institute of Paleontological Research, Johannesburg (BPI 3839). BPI 3839 possesses the emargination characteristic for Ostodolepidae, but the cheek is not preserved in the type specimen. In the type specimen, the pre- frontal reaches the external naris, and the lacrimal has a high dorsal process at the anterior margin of the orbit. These are features of Gymnarthridae, and lacking in BPI 3839. It follows from these and other characters that BPI 3839 and the type of Micraroter erythro- geios do not belong to the same genus. The first may be closer to Ostodolepidae because of the main character cheek emargination, but the latter closer to the Goniorhynchidae/Gymnarthridae (Fig. 4 shows a reconstruction of the missing cheek region with a broader anterodorsal portion of the squamosal as found in Gymnarthridae and Goniorhynchidae; there is not much space left for an emargina- tion). In addition, the cervical vertebrae of the type specimen of Micraroter show similarities in the shape of the neural arch with Gymnarthridae (compare Carroll and Gaskill, 1978, Fig. 115F with 115E and C); the neural arches of BPI 3839 (Fig. 115G) are quite different. On the other hand, BPI 3839 possesses intercentra similar (Fig. 59A, but in contrast to the reconstruction in Fig. 60C) to those in Euryodus, but the type specimen of Micraroter does not show intercentra. Even though Micraroter erythrogeios and BPI 3839 belong to different genera, they show that the Ostodolepidae are more closely related to Goniorhynchidae and Gymnarthridae than to any other Tuditanomorpha. Family Gymnarthridae The Gymnarthridae are the family with the highest number of genera and species within the Tuditanomorpha, and the new form belongs in this family (see above). Carroll and Gaskill (1978) often use size to assign specimens to genera. This seems to be a weak agrument. Because of insufficient knowledge of many gym- narthrids Carroll and Gaskill (1978: 76) state that “the possible interrelationship of the various genera is open to speculation.” Contrary to Carroll and Gaskill (1978), let us speculate on the use of defined characters in the hope of attaining the means to assign the described specimen to the proper genus. A series of characters 16 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 1CM ‘ Ficure 4.—Micraroter erythrogeios Daly, skull in lateral view (after Carroll and Gaskill, 1978, Fig. 53; changed in the missing cheek region ).—art articular, j jugal, occ occipital, po postorbital, pp. postparietal, q quadrate, qj quadrato- jugal, so supraoccipital, sq squamosum, ¢ tabular. has been used to define the different genera and species within the Gymnarthidae and their relationship (Fig. 6): (1) Prefrontal: The prefrontal occupies a similar space as in anthracosaurs; it reaches far forward but does not reach the external naris. The later advanced condition is acquired in the new species and in E. dalyae, E. peabodyi, and in some specimens of E. primus. (2) Position of the jaw articulation: It has been shown above (p. —) that the jaw articulation is situated at the level of the oc- cipital articulation in primitive Tuditanomorpha. In Gymnarth- ridae, the jaw articulation is always in front of the occipital articulation, and it moves even farther anterior in more advanced species like Cardiocephalus sternbergi. (3) Size of orbit: The orbit is usually larger in juvenile speci- mens than in adults, so this criterion is often used as an indication of juvenility. Small specimens often have a larger orbit (compared to the length of the skull) than larger specimens. We find within the Tuditanomorpha that the members of the more primitive fam- ilies have very large orbits even when the specimens are large (i.e., Crinodon). There seems to be a tendency in more advanced Tuditanamorpha towards smaller orbits. The largest orbit (length of orbit to length of skull) of specimens figured in this paper can be found in the smallest (Saxonerpeton, Fig. 5A) and in the largest ( Micraroter, Fig. 4) forms. This demonstrates that there is not only a tendency towards smaller orbits, but another tendency to keep large eyes. The gymnarthrids generally have smaller orbits with a A NEW GYMNARTHRID MICROSAUR 17 length below 20 percent of the skull length with the exception of E. bonneri (Fig. 5E). (4) Size of external nares: The size of the external nares is re- lated to the size of the specimens; it is relatively large in small specimens. Thus, the external nares are proportionately larger in Ficure 5.—Skulls in lateral view of Gymnarthridae and Goniorhynchidae (after Carroll and Gaskill, 1978; changed); all x 2. (A) Saxonerpeton geinitzi (Credner) (ibid., restoration Fig. 22 changed after Fig. 21); (B) Leioce- phalikon problematicum (Dawson) (ibid., restoration Fig. 46C); (C) Pariotichus brachyops Cope (ibid., restoration Fig. 45C changed after Fig. 45A); (D) Cardiocephalus sternbergi Broili (ibid., restoration Fig. 30C changed after Fig. 30A); (E) Euryodus bonneri n. sp.; (F) Euryodus dalyae Carroll and Gaskill (ibid., restoration Fig. 43 changed after Fig. 42A); (G) Euryodus peabodyi (Carroll and Gaskill) (ibid., restoration Fig. 32 changed after Fig. 32 and KU VP 8967); (H) Euryodus primus Olson (ibid., restoration Fig. 37 changed after Figs. 36 + 38A); (J) Rhynchonkos stovalli (Olson) (ibid., restoration Fig. 64).—ang angular, art articular, de dentary, f frontal, j jugal, 1 lacrimal, mx maxillary, na nasal, occ occipital, pf postfrontal, pmx premaxillary, po postorbital, prf prefrontal, q quadrate, qj quadratojugal, sang surangular, sm septomaxillary, sq squamosum, ¢ tabular. 18 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY microsaurs than in the larger anthracosaurs. There is the tendency towards smaller external nares (compared to the length of the skull). In E. bonneri, the external nares are even more enlarged, while C. sternbergi shows the tendency towards smaller external nares as in other eda eens (5,6,13) Number of teeth on premaxillary, on maxillary, and on the margin of the lower jaw: The number of teeth is high in primitive Tuditanomorpha as in Microbrachomorpha. The number of teeth on all three bones is reduced in all families of Tuditano- morpha compared to the Tuditanidae. Seven premaxillary teeth interpreted as the primitive state occur in some members of the Tuditanidae and Microbrachomorpha. In the more advanced tuditanomorph families the number is reduced to five; in the Pantylidae to three. The later reduced number also occurs in two species of Gymnarthridae (E. peabodyi and E. primus), a syna- pomorphy of these two species and a convergence in Pantylidae assumed after the phylogenetic position of both groups. Gym- narthridae have fewer marginal teeth on the maxillary and the lower jaw. Only the Pantylidae have less teeth on the lower jaw than the Gymnarthridae. (7) Number of small teeth behind enlarged teeth on the maxil- lary: A differentiation between premaxillary teeth (“incisors”) and maxillary teeth occur in Pantylus and Gymnarthridae. In Pantylus the first maxillary tooth is enlarged (“canine”), while all maxillary teeth are enlarged in Gymnarthridae except the most posterior ones. The enlargement of maxillary teeth includes progressively more posterior teeth so that fewer and fewer small teeth are found in the posterior part of the maxillary. (8) Retroarticular process: A retroarticular process is found only in the largest representative within the Gymnarthridae. (9) Shape of lacrimal at the anterior margin of the orbit: The lacrimal extends from the external naris to the orbit in all micro- saurs. It has a ventral extension along the margin of the orbit, and it meets the rostral extension of the jugal. That is the shape of the lacrimal found in nearly all microsaurs, but the lacrimal extends dorsally along the anterior margin of the orbit in some forms (Crinodon within the Tuditanidae; Pariotichus, Euryodus, and Cardiocephalus within the Gymnarthridae; Pelodosotis within the Ostodolepidae; and the holotype of Micraroter). This shape is ac- quired in parallel in these families because their more primitive members lack this feature. The shape is very similar to the shape of the lacrimal in captorhinomorphs, another parallel evolution. The ventral extension may be reduced in Cardiocephalus sternbergi. (10) Quadratojugal: The elongated, narrow shape of the quadratojugal is one of the advanced characters of the Gymnarth- ridae and Goniorhynchidae (Fig. 3). The quadratojugal is primi- A NEW GYMNARTHRID MICROSAUR 19 tively situated between the jugal and the quadrate. It extends rostrad below the jugal in some gymnarthrids, and it reaches the maxillary in E. primus and possibly in E. peabodyi also. (11) Posterior margin of the skull table: In dorsal view straight posterior margin of the skull table is the common feature of gymnarthrids above Tuditanidae (Fig. 3). The margin becomes convex with the backward inclination of the posterior margin of the cheek région in Cardiocephalus sternbergi, and to some extent in E. peabodyi. E. primus shows a tendency towards this feature (Carroll and Gaskill, 1978, Fig. 36-38). (12) Posterior margin of the cheek region: A vertical posterior margin of the cheek region is primitive within Tuditanomorpha (Fig. 3). Burrowing forms tend to have a _ forward-inclined posterior margin of the cheek region. That seems to be the tendency in one gymnarthrid species, E. primus (after the restoration by Carroll and Gaskill, 1978, Fig. 37), but otherwise gymnarthrids tend to have a backward-inclined posterior margin of the cheek, with the most extreme example being Cardiocephalus sternbergi. Only the Permian gymnarthrid genera are sufficiently known to evaluate their interrelationships. Three genera, Cardiocephalus, Euryodus, and Pariotichus, are known from the Permian of Texas, Oklahoma and Kansas. Carroll and Gaskill (1978) placed C. cf. sternbergi as E. dalyae into the genus Euryodus because of differ- ence in size with both genera, Euryodus and Cardiocephalus, con- taining two species. There exist more differences between the two genera than simply size (Figs. 5, 6). Cardiocephalus (Fig. 5D) has a long, low skull with a large postorbital length, and the jaw articulation lies far rostrad from the occiptal condyle-atlas articula- tion (distance nearly 25% of skull length) as compared with other gymnarthrids. The posterior margin of the cheek is strongly in- clined backwards comparable to the situation in Rhynchonkos (Fig. 5J): The external naris is the smallest within the gymnarthrids. The lower jaw possesses a high coronoid process. In contrast to Cardiocephalus, Euryodus (Fig. 5F,H) shows a short distance between the occipital condyle-atlas articulation and the jaw articulation (about 12% of skull length). The occiput is nearly straight or slightly inclined posteriorly. The external naris is large even in proportion to the large orbit (up to 50% of orbit length). The teeth are enlarged, and one prominent feature which caused us difficulties at first is the long prefrontal reaching the external naris, a feature found in nearly all specimens of Euryodus and in the type specimen of Micraroter (but not in BPI 3839). The lacrimal exhibits a process dorsally on the anterior margin of the orbit as in Pariotichus, and the quadratojugal is prolonged rostrad so that it meets the maxillary in the type species of the genus. If we apply these characters on the four species assigned by Carroll 20 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY SAXONERPETON PARIOTICHUS EURYODUS CARDIOCEPHALUS LEIOCEPHALIKON bonneri dalyae peabodyi primus sternbergi pf=po pf>po pf=po Pf=po pf=po pf=po pf—po pf>po ete Autapomorphy =) Synapomorphy 6.10,72,11.0 {rF====>1 Questionable U—====4 Synapomorphy ee en aS Sar SS SSB 612,71,13.12 20,40,55,70-120 Ficure 6.—Interrelationship of genera and species of the family Gym- narthridae and Saxonerpeton. pf postfrontal, po postorbital. (1) Prefrontal: (1.0) far from external nares, (1.1) close to external nares, (1.2) reaching external nares. (2) Position of the jaw articulation: (2.0) close in front (less than 8% of skull length), (2.1) in front (between 10% and 14% of skull length), (2.2) far in front of occipital condyle-atlas articulation (24% of skull length). (3) Size of orbit: (3.0) very large (more than 25% of skull length), (3.1) large (20-25% of skull length), (3.2) medium (below 20% of skull length). (4) Size of external nares: (4.0) large, (4.1) very large, (4.2) medium. (5) Number of teeth on premaxillary: (5.5) 5, (5.4) 4, (5.3) 3. (6) Number of teeth on maxillary: (6.22) 22, (6.12) 12, (6.10) 10. (7) Number of small teeth behind enlarged teeth on the maxillary: (7.0) no enlarged teeth, (7.1) 4, (7.2) much enlarged teeth +- 3 small behind (exception, type specimen of E. primus), (7.21) one tooth very much enlarged, (7.3) 2 small behind. (8) Retroarticular process: (8.0) absent, (8.1) present. (9) Shape of lacrimal at the anterior margin of the orbit: (9.0) only ventral process, (9.1) dorsal and ventral process, (9.11) deep dorsal process, (9.2) ventral process shortened. (10) Quadratojugal: (10.0) well separated from maxillary, behind jugal, (10.1) reaching below jugal, (10.11) reaching far rostrad below jugal. and meeting maxillary. (11) Posterior margin of the skull table: (11.0) straight, oa A NEW GYMNARTHRID MICROSAUR 21 and Gaskill (1978) to Cardiocephalus and Euryodus, we have to assign only sternbergi to Cardiocephalus, and the other three species to Euryodus. The described specimen shows the charac- teristics of Euryodus, and must be assigned to Euryodus as a new species. REVISED DIAGNOSES Cardiocephalus Broili, 1904 Small gymnarthrid microsaur with long, low skull. Jaw articula- tion far anterior to the occipital condyle-atlas articulation, posterior margin of the cheek region strongly inclined posteriorly. Long postorbital length (about twice the antorbital length). Small ex- ternal nares with only the lacrimal on the posterior margin, pre- frontal reaching rostrad half the distance between orbit and external naris. Medium sized orbits. High coronoid process on lower jaw. Stapes imperforate. Exoccipital not deeply notched to incorporate vagus foramen. Rows of teeth on pterygoid, vomer, ectopterygoid and palatine, but no palatal shagreen of denticles. Type species and only species: Cardiocephalus sternbergi Broili, 1904 Euryodus Olson, 1939 Medium to large gymnarthrid microsaur. Skull elongated with widely rounded snout, jaw articulation anterior to occipital condyle- atlas articulation. Posterior margin of the cheek region straight or slightly inclined posteriorly. Large external nares, prefrontal reach- ing or nearly reaching the posterior margin of the external nares. Quadratojugal extending rostrad below the jugal. Stem of stapes perforate. Opening of vagus nerve penetrating exoccipital. Sha- green of teeth on the palate region, one row of teeth on ectoptery- goid and palatine. Type species: Euryodus primus Olson, 1939 The type species E. primus is characterized by an enlarged single tooth on the maxillary below the orbit; E. peabodyi shows the same feature to a reduced extent (KU VP 8967). The occiput of Euryodus primus is slightly inclined forward in the dorsal part. Another characteristic feature of the species is the retroarticular (11.1) convex. (12) Posterior margin of the cheek region: (12.0) straight, (12.1) inclined backward, (12.11) inclined strongly backward, (12.2) straight and dorsally inclined forward. (13) Number of teeth on the margin of the lower jaw: (13.22) 22, (13.12) 12, (13.10) 10. 22 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY process. The species most closely related to E. primus is E. pea- bodyi. Both species have the most reduced number of premaxillary teeth (3) within the gymnarthrids. In both species, the quad- ratojugal reaches rostrad below the jugal or close to the maxillary. The lacrimal extends a dorsal process high up on the rostral margin of the orbit. E. peabodyi is smaller than E. primus, the orbit being larger, and the external nares smaller in proportion to the length of the orbit. Both species are easy to distinguish by the inclination of the posterior margin of the cheek region (E. primus straight and forward, E. peabodyi backward), and the posterior margin of the skull table (straight in E. primus, convex in E. peabedyi). Nearly as large as the type species is E. dalyae. The teeth are more rounded, blunter cones than in the other two species. The pre- maxillary bears five teeth as in E. bonneri and in more primitive gymnarthrids. The maxillary bears ten teeth like other species of Euryodus and Cardiocephalus. The prefrontal reaches the external nares, and the dorsal process of the lacrimal at the anterior margin of the orbit is shorter than in the two first species. The quad- ratojugal is low, and reaches just. below the jugal. There are only ten teeth on the lower jaw as few as in C. sternbergi whereas the three other species of Euryodus have 12 or 13 teeth on the lower jaw. Characters in common in E. dalyae and E. bonneri are primitive characters within the genus and in themselves not con- clusive enough to oppose both species to E. peabodyi and E. primus. The peculiarly shaped suture between tabular and postparietal on the occiput indicates a closer relationship of both species. E. bonneri has a very narrow skull. The teeth are not blunt as in the other species; on the contrary, they are very pointed at least on the lower jaw. The prefrontal reaches the external naris broadly. The external nares are very large, with their length reaching half of the length of the very large orbit. Although the teeth are very different from the other three species of Euryodus, in all other features the described specimen differs only slightly from the other species, The third Permian gymnarthrid genus from Texas, Pariotichus, is very difficult to separate from Euryodus, because the preservation of the skull is so poor. The prefrontal does not reach the external naris, but it comes as close as in the type specimen of E. primus. The lacrimal has a dorsal process on the anterior margin of the orbit as in Euryodus. There may be an elongated quadratojugal which could have reached the maxillary. The maxillary bears more teeth (12) than are usually found in Euryodus (10). It is difficult to decide how closely P. brachyops is related to Euryodus because of lack of sufficient information. It is left as a separate genus more closely related to Euryodus. We argued above that the Permian European genus Saxonerpe- A NEW GYMNARTHRID MICROSAUR 23 ton may belong in the family Gymnarthridae. Saxonerpeton is a very small form, and like other small forms, it possesses a very large orbit compared to the length of the skull. Saxonerpeton is adapted to an aquatic life, in contrast to the tendency to terrestrial adaptation in Tuditanomorpha as indicated by the groove for a lateral line on the jugal of Saxonerpeton (Carroll and Gaskill, 1978: 38). The large mouth opening, peglike teeth, and the high number of teeth agree with this interpretation. These are primitive features within the Tuditanomorpha, and they distinguish the genus from all other genera within the family, but are not autapo- morphies for the genus itself. It is difficult to deal with the other Carboniferous genera (Sparodus, Leiocephalikon, and Hylerpeton) within the family because of scarcity of remains (Carroll and Gaskill, 1978: 72). The best known of the three genera is Leiocephalikon. The twelve teeth in the maxillary compare with Pariotichus; the twelve teeth in the lower jaw compare with three of the four Euryodus species. The number in Hylerpeton (15) and Sparodus (17) is higher. All other known features of Leiocephalikon are primitive within the Gym- narthridae. There exists a close resemblance in the sutures between the skull bones in Leiocephalikon and Trachystegos. The frontals have the same outline (compare Carroll and Gaskill, 1978, Fig. 46A with Carroll, 1966, Fig. 11A): an emargination on the lateral border, a posterolateral process of the right frontal, a posteromedial process of the left frontal in between the parietals, and a projection of the right frontal into the left. The suture between the parietals has the same course. The maxillary has a similar dorsal extension in front of the orbit in both specimens. But there is a distinct differ- ence in size: Trachystegos is nearly three times as large as Leioce- phalikon. This size difference explains the more exaggerated sculp- ture, the smaller pineal opening, and the larger, more blunt teeth in Trachystegos. There exists the possibility that Trachystegos is the adult form of Leiocephalikon, but further research is required. The differences could give us some indication about changes during ontogeny, assuming the interpretation is correct. For example, the number of teeth decreases from 12 in the lower jaw of Leiocephali- kon to 10 in the lower jaw of Trachystegos. In conclusion, Euryodus and Cardiocephalus can easily be distinguished, and their relationship can be settled. Pariotichus is close to Euryodus. The relationship of all the other genera of the Gymnarthridae (Saxonerpeton, and the Carboniferous genera) is not possible to define sufficiently because of lack of data. 24 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY SUMMARY The families within the suborder Tuditanomorpha of the order Microsauria have been arranged in a phylogenetic scheme. The Pantylidae are not much removed from -the ancestral form they share with the most primitive family, the Tuditanidae. The jaw articulation moves relatively more anteriorly to the occipital condyle-atlas articulation in the other families. Gymnarthridae and Goniorhynchidae are the two families most closely related to each other. The Ostodolepidae is the most specialized family. Ricnodon seems to be close to the Tuditanidae. Saxonerpeton is assigned to the Gymnarthridae. Trachystegos has many close simi- larities with Leiocephalikon, and may be conspecific with Leioce- phalikon as the adult form. Micraroter (the type specimen) has certain features in common with Goniorhynchidae and Gymnarth- ridae which are not shared by the second specimen (BPI 3839) attributed to the same species by Carroll and Gaskill (1978). The interrelationship of the Permian gymnarthrid genera has been established, but this was not possible for the Carboniferous genera because of lack of information. The genera Cardiocephalus and Euryodus have been redefined. Cardiocephalus encompasses only one species (sternbergi), but Euryodus four, including the new species E. bonneri. E. bonneri is the most primitive species within the genus Euryodus. The specimen of E. bonneri has been discovered from supposedly aquatic deposits of the Lower Permian of Kansas, about equal in age to the deposits where Pariotichus has been found in Texas. E. bonneri is therefore the oldest species within the genus Euryodus. LITERATURE CITED Broiut, F. 1904. Permische Stegocephalen und Reptilien aus Texas. Palaeonto- graphica 51: 1-120. Carro.i, R. L. 1966. Microsaurs from the Westphalian B of Joggins, Nova Scotia. Proc. Linn. Soc., London 177: 63-97. Carrot, R. L. AND Baro, D. 1968. The Carboniferous Amphibian Tuditanus [Eosauravus] and the Distinction between Microsaurs and Reptiles. American Mus. Novitates 2337: 1-50. Carroui, R. L. anp GAsKILL, P. 1978. The Order Microsauria. Mem. Ameri- can Philos. Soc. 126: VIII + 1-211. Case, E. C. 1910. New or Little Known Reptiles and Amphibians from the Permian (?) of Texas. Bull. American Mus. Natur. Hist. 28: 163-181. Day, E. 1973. A Lower Permian Vertebrate Fauna from Southern Oklahoma. Jour. Paleont. 47: 562-589. Dawson, J. W. 1863. Air Breathers of the Coal Period. American Jour. Sci. 36: 430-432. Frirscu, A. 1879-1883. Fauna der Gaskohle und der Kalksteine der Permfor- mation Boéhmens. 1: II + 182 pp., Prag. Garrney, E. S. 1979. Tetrapod monophyly: A phylogenetic analysis. Bull. Carnegie Mus. Natur. Hist. 13: 92-105. A NEW GYMNARTHRID MICROSAUR 25 Grecory, J. T., Peasopy, F. E. anp Price, L. I. 1956. Revision of the Gymnarthridae, American Permian Microsaurs. Bull. Peabody Mus. Natur. Hist. 10: 1-77. Hampson, G. F. 1896. The Fauna of British India, including Ceylon and Burma. Moths 4: XXVIII + 594 pp. (Taylor and Francis), London. Hennic, W. 1966. Phylogenetic systematics. Univ. Illinois Press. Oxson, E. C. 1939. The fauna of the Lysorophus pockets in the Clear Fork Permian, Baylor County, Texas. Jour. Geol. 47: 389-397. Otson, E. G. 1962. Permian Vertebrates Oklahoma and Texas. Oklahoma Geol. Surv. Circ. 59: 1-68. Oxson, E. C. 1970. New and Little known genera and species of vertebrates from the Lower Permian of Oklahoma. Fieldiana: Geology 18 (3): 359-434. Pancuen, A. L. 1970. Anthracosauria. In: Handbuch der Palaoherpetologie (ed. O. Kuhn). Teil 5A: VII + 77 pp. (G. Fischer), Stuttgart. PANCHEN, A. L. 1972. The interrelationships of the earliest tetrapods. In: Studies in vertebrate evolution (Joysey, K. A. and Kemp, T. S. eds.): 65-87, (Oliver and Boyd), Edinburgh. PANCHEN, A. L. 1977. The origin and early evolution of tetrapod vertebrae. In: Problems in Vertebrate Evolution (S. M. Andrews, R. S. Miles and A. D. Walker eds.). Linnean Soc. Symp. Ser. No. 4: 289-318. STEEN, M. C. 1938. On the Fossil Amphibia from the Gas Coal of Nyrany and other deposits in Czechoslovakia. Proc. Zool. Soc. London B, 108:265- 283. = | e),' os ry fae & we - - " ‘ ¢ aie : * — = ma t - . : : , a ree 7 a 5 LL a : 5 : iit : sath hacen, aheietiles urns ae Tame Palit od Bone eit “ i yas | hap 4 28 ie 0) ‘4: en art Pond - | i orn |e) be A te. Bn pepntinnry at ite it vient “| ae ” i ‘alee _ ; >= wipe a2" « aha ee ae o “Vb i 7 - f Fo.» alan a = Fels 7] AF D4 g uh BAsy) tH? flides + Batol = Fs * oer 4, ' hoa « -_ 4 ’ : = 4 f * ~ ‘ , } J Par - 5 P’ P ae | A } ; a ay , ii Mp E (ng) Spy bousy . iy iv ie a Pie A, . rh, ind. (cs ay? = ; Iw ‘pti ! Pi Batt. ‘Pieces i Gat 7 habe 2 fophn of C . wie the Arete uty « HP ttideninsh-2 E ‘t. ge : UNIVERSITY OF KANSAS PUBLICATIONS MUSEUM OF NATURAL HISTORY The University of Kansas Publications, Museum of Natural History, beginning with volume 1 in 1946, was discontinued with volume 20 in 1971. Shorter research papers formerly pub- lished in the above series are now published as Occasional Papers, Museum of Natural History. The Miscellaneous Publi- cations, Museum of Natural History, began with number 1 in 1946. Longer research papers are published in that series. Monographs of the Museum of Natural History were initiated in 1970. All manuscripts are subject to critical review by intra- and extramural specialists; final acceptance is at the discretion of the publications committee. Institutional libraries interested in exchanging publications may obtain the Occasional Papers and Miscellaneous Publica- tions by addressing the Exchange Librarian, The University of Kansas Library, Lawrence, Kansas 66045. Individuals may pur- chase separate numbers of all series. Prices may be obtained upon request addressed to Publications Secretary, Museum of Natural History, The University of Kansas, Lawrence, Kansas 66045. Editor: E. O. WiLEy Managing Editor: JosepH T. CoLiins PRINTED BY UNIVERSITY OF KANSAS PRINTING SERVICE LAWRENCE, KANSAS UM | MUS. CO %/128E Be ae JUN5 1981 OCCASIONAL PAPERS v2 i= = > a = HARVARED of the UNIVERSITY MUSEUM OF NATURAL HISTORY The University of Kansas Lawrence, Kansas NUMBER 92, PAGES 1-7 JUNE 9, 1981 A NEW HYLID FROG OF THE GENUS PLECTROHYLA FROM A CLOUD FOREST IN HONDURAS By JAMEs R. McCranie’ AND LARRY Davip WILSON” The hylid genus Plectrohyla is a notable component of the high- land forest herpetofauna of northern Middle America. The ten known species occur from southeastern Oaxaca, México, southward to western Honduras and EI Salvador (Duellman, 1970). Recent work in the relatively poorly known cloud forests of Honduras has produced two new frog taxa (McCranie et al., 1980), including a new species of Plectrohyla described herein. Prior to this discovery, only one species of Plectrohyla (P. guatemalensis) had been re- ported from Honduras (Duellman, 1970; Meyer and Wilson, 1971). In April of 1979 and May of 1980, we visited the cloud forest of the Sierra de Omoa in northwestern Honduras and worked in the region of Cerro Cusuco in the highest elevations of that range. Along the Quebrada Cusuco we collected several representatives of an undescribed species of Plectrohyla, which we here name Plectrohyla dasypus new species (Figure 1) Holotype.—KU 186025, an adult male from along the Quebrada Cusuco at El Cusuco (15°30’N, 88°13’W), a finca located 5.6 km WSW Buenos Aires (the latter locality about 19 km N Cofradia), 1580 m, Sierra de Omoa, Departamento de Cortés, Honduras, ob- 110770 S. W. 164th Street, Miami, Florida 33157. 2 Division of Intercurricular Studies, Miami-Dade Community College, South Campus, Miami, Florida 33176. 2 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Fic. 1.—Plectrohyla dasypus, KU 186027, ¢, 40.0 mm. tained on 14 April 1979 by James R. McCranie and Larry David Wilson. Paratypes.—KU 186026-28, James R. McCranie and Larry David Wilson, 14 April 1979; KU 186029-31, Gustavo A. Cruz Diaz, James R. McCranie, and Larry David Wilson, 12-13 April 1979; KU 186032-34, James R. McCranie and Larry David Wilson, 21-22 May 1980; all from along the Quebrada Cusuco at El Cusuco be- tween elevations of 1530-1660 m. Definition—A species of Plectrohyla distinguished from _ its congeners by the following combination of characteristics: small size (31.5-44.0 mm snout-vent length); dorsum smooth to weakly tuberculate; vocal slits present; no vertical rostral keel; prepollical spur short, flat, blunt; dorsum bronze and venter gray, without a dark lateral line or row of dark spots. Description of holotype —Adult male with snout-vent length (SVL) of 44.0 mm; tibia length 21.5 mm, 48.9 percent of SVL; foot length 19.5 mm, 44.3 percent of SVL; head length 13.2 mm, 30.0 percent of SVL; head width 14.7 mm, 33.4 percent of SVL. Snout short (4.0 mm in length), distance from anterior edge of orbit to tip of snout 80 percent of diameter of eye; snout truncate in lateral profile, in dorsal profile acuminate, lacking a vertical rostral keel; canthal ridge thickened; loreal region slightly concave; lips moderately thickened, barely flared. Nostrils protuberant, directed dorsolaterally, situated near tip of snout; internarial distance 3.2 mm; internarial area slightly depressed near point of convergence of canthal ridges; top of head flat; interorbital distance 4.8 mm, 32.7 percent of head width; diameter of eye 5.0 mm; width of eyelid A NEW HYLID FROG OF THE GENUS PLECTROHYLA 3 4.1 mm, 27.9 percent of head width. Heavy dermal fold extending posteriorly from posterior edge of orbit, barely covering upper edge of tympanum; remainder of tympanum distinct; length of tympa- num 2.2 mm, 44.0 percent of eye diameter. Arms robust, forearm slightly heavier than upper arm; distinct transverse fold on wrist. No axillary membrane. Fingers long, slender; length of fingers from shortest to longest, 1-2-4-3; disc on third finger slightly smaller than tympanum; webbing vestigial; subarticular tubercle large, subconical; terminal tubercle on fourth finger normal; supernumerary tubercles in single rows on proximal segments of fingers; prepollex short, flat, blunt, nonbifurcate, bone not protruding through skin. Heels overlapping slightly when hind limbs adpressed; no transverse dermal fold on heel; inner tarsal fold extending full length of tarsus; no outer tarsal fold; inner metatarsal tubercle ovoid, visible from above; no outer metatarsal tubercle. Toes long, slender; length of toes from shortest to longest, 1-2-5-3-4, fifth toe nearly as long as third; discs moderately large; subarticular tubercles moderately large, subconical; supernumerary tubercles small, in single row on proximal segment of each digit; toes about three-fourths webbed; webbing extending from base of penultimate phalanx of first toe to penultimate phalanx of second, from base of disc of second to base of penultimate phalanx of third, from base of disc of third to base of penultimate phalanx of fourth and on to base of disc of fifth toe. Anal opening directed posteroventrally at level of mid-thigh; anal sheath short, broad. Skin on dorsal surfaces and throat weakly tuberculate, on chest, belly, and ventral surfaces of thighs and anus granular, on ventral surfaces of shanks smooth. Tongue nearly round, shallowly notched anteriorly and posteriorly and barely free behind. Upper jaw shallowly notched medially. Maxillary-pre- maxillary teeth spatulate, 58-60; prevomerine teeth 4-4, situated on small elliptical elevations between quadrangular choanae; vocal slits present. Color in life: dorsum bronze with small scattered black spots; black stripe along canthus and above tympanum to above arm; mottled bronze and black subocular blotch present; venter and hidden leg areas dark gray; toe webbing gray; eye copper with black reticulations; chin gray with a bronze patina. Color in pre- servative: dorsum dark gray with small scattered black spots; venter gray. Variation.—Measurements and proportions of all specimens are given in Table 1. Four of the six male paratypes have snout-vent lengths of 39.0-42.5 mm and have hypertrophied forelimbs. The other two males have snout-vent lengths of 31.8 and 35.9 mm and do not have hypertrophied forelimbs. One male (KU 186029) with 4 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY TasLe 1, Variation in measurements (in millimeters) and proportions in Plectrohyla dasypus. Character abbreviations are in brackets. Means are in parentheses following ranges. Character Males (n=7) Snout-Vent length [SVL] Tibia length [TL] (TL/SVL) Foot length [FL]) (FL/SVL) Head length [HL] 31.8-44.0 (39.0) 17.3-21.5 (19.9) 48.9-54.4 (51.3) 14.5-19.5 (16.8) 41.3-45.6 (43.2) 10.0-13.2 (11.7) Females (n=3) 31.5-44.0 (37.5) 17.0-22.5 (19.3) 50.0-54.0 (51.7) 14.5-18.0 (16.1) 41.0-46:0 (43.2) 10.1-12.8 (11.6) (HL/SVL) 27.8-32.0 (30.0) 29.1-32.2 (31.1) Head width [HW] 11.5-14.7 (13.5) 11.2-15.0 (13:1) (HW/SVL) 33.4-36.5 (34.8) 34.1-35.6 (34.9) Snout length [SL] 2.5-4.0 (3.1) 2.5-3.2 (2.9) (SL/ED) 63.8-80.0 (71.8) 64.6-76.2 (68.9) Eye diameter [ED] 3.8-5.0 (4.4) 3.8-4.8 (4.3) Internarial distance [IND] 1.8=3.2 (2:3) 1.8-2.8 (2.3) Interorbital distance [IOD] 4.0-5.0 (4.5) 4.0-5.2 (4.6) (1OD/HW) 31.1-36.0 (33.3) 34.6-35.7 (35.0) Eyelid width [EW] 3.2-4.2 (3.8) 2..8-4.0 (3.6) (EW/HW) 25.0-29.6 (27.9) 25.0-30.0 (27.2) Tympanum length [TL] 1.7-2.2 (2.0) 1,.7-2.2 (1.9) (TL/ED) 42,.9-52.5 (45.7) 44,7-45.8 (45.2) a snout-vent length of 42.5 mm has minute spines on the prepollex and the median surface of the first finger. The color pattern of the male paratypes is essentially like that of the holotype. The dorsa of the smallest (KU 186028) and largest (KU 186033) females were bronze in life with small scattered black spots outlined by lime green. Other details of dorsal pattern of the females are like those of the holotype. In preservative, the venters of the three females are somewhat paler than those of the males. Description of tadpole—Three tadpoles (KU 186035) are presumed to be Plectrohyla dasypus. Two are in developmental stage 40 (Gosner, 1960), and have body lengths of 16.4 and 15.9 mm, and total lengths of 42.2 and 44.0 mm, respectively; the third is in developmental stage 38 and has a body length of 14.5 mm and a total length of 40.8 mm. Description of latter specimen is as follows: body ovoid in dorsal view; snout bluntly rounded in dorsal profile; eyes small, directed dorsolaterally; nostrils directed antero- laterally at a point slightly closer to eyes than to tip of snout; spiracle sinistral, on midline at midlength of body; cloacal tube long, dextral; caudal musculature moderately robust, extending nearly to tip of tail; fins shallow, caudal musculature at midlength of tail deeper than either dorsal or ventral fins; dorsal fin not extend- ing onto body. In preservative, body dark grayish brown; caudal musculature creamy tan, marked with dark brown flecks throughout length; fins A NEW HYLID FROG OF THE GENUS PLECTROHYLA 5 translucent, marked with dark brown spots, spots more numerous on dorsal fin. Mouth ventral, a little over half as wide as body at greatest width; oral disc not emarginate; edge of oral disc completely bordered by single row of small papillae; row of larger papillae present medially to fringing row; beaks well developed, possesing long, pointed serrations of equal length; upper beak in form of broad arch, lacking lateral processes; lower beak moderately robust, broadly V-shaped; labial tooth formula 2(2)3, upper rows long and equal in length, second upper row narrowly interrupted medially; lower rows of equal length, being only slightly shorter than upper rows. Two nearly metamorphosed individuals were collected. One (KU 186037) is in developmental stage 44 (Gosner, 1960), and has a body length of 16.7 mm and a total length of 25.8 mm. The other (KU 186036) is in developmental stage 45 and has a body length of 17.1 mm. Comparisons—Duellman (1970) placed the ten species of Plectrohyla then known into two groups. One group, including P. avia, glandulosa, guatemalensis, hartwegi, lacertosa, and pycno- chila is characterized by medium to large size (44.6-86.2 mm mean snout-vent length in adult males) and absence of vocal slits. The second group, including P. ixil, matudai, quecchi, and sagorum, is characterized by small size (33.1-42.2 mm mean snout-vent length in adult males) and presence of vocal slits. Plectrohyla dasypus has vocal slits and a mean snout-vent length of 41.0 mm in five males with hypertrophied forelimbs and thus is distinct from the six species in the first group. A vertical rostral keel is lacking in Plectrohyla dasypus, whereas it is present in P. sagorum and P. quecchi. Plectrohyla dasypus has a short, flat, blunt prepollical spine. In P. sagorum, P. quecchi, P. ixil, and P. matudai the prepollical spine is long, pointed, and distally curved. The dorsum of P. dasypus is weakly tuberculate, that of P. quecchi and P. matudai is tuberculate. The mean snout-vent length of adult males of P. dasypus is 41.0 mm. Plectrohyla ixil is slightly smaller, with a mean snout-vent length of adult males of 38.9 mm and P. matudai is a smaller species with a mean snout-vent length of adult males of 33.1 mm. Plectrohyla ixil has a broad pale lateral stripe, bordered below by a dark line separating the color of the dorsum from that of the venter; this pattern is lacking in P. dasypus. The tadpoles presumed to be Plectrohyla dasypus are distin- guishable from the other members of the sagorum group. Plectro- hyla dasypus tadpoles lack lateral processes on the upper beak, whereas P. sagorum tadpoles have moderately robust lateral proc- esses. In P. dasypus tadpoles all three lower rows of denticles are of equal length, whereas in P. quecchi, P. ixil, and P. matudai the 6 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY third lower row is shorter than the first and second. In P. dasypus tadpoles the three lower rows of denticles are only slightly shorter than the two upper rows; those tadoles of P. ixil and P. matudai are noticeably shorter than the upper rows. Plectrohyla dasypus tad- poles have the upper beak in the form of a broad arch and its serrations are of equal length. In P. matudai tadpoles the upper beak is barely arched and with one serration on either side greatly enlarged into fanglike projection. In P. ixil tadpoles there are three or four noticeably enlarged serrations on either side of the upper beak. In P. dasypus tadpoles the lower beak is moderately robust, whereas in P. matudai tadpoles the lower beak is narrow. Natural history—The specimens of Plectrohyla dasypus were collected along streams in relatively undisturbed hardwood cloud forest (Lower Montane Wet Forest formation of Holdridge, 1967). A prominent tree in the canopy layer is sweet gum (Liquidambar styraciflua). In general, the canopy trees bear a heavy burden of epiphytes, including bromeliads, mosses, ferns, orchids, and philo- dendrons. The middle story is dominated by palms and tree ferns. Various small palms and ferns, ground-dwelling bromeliads, bam- boo, and pepperomias make up much of the understory. Three individuals were in bromeliads (one on a stump, and the other two on trees) during the day. Another was swimming in the stream (Quebrada El] Cusuco) during the day, and the remainder were on vegetation along the Quebrada E] Cusuco and a tributary thereof at night. Adults of Plectrohyla guatemalensis and Ptychohyla spinipollex also were collected along these streams. Both of our visits were during the latter part of the dry season when the streams were clear and not torrential, a condition seem- ingly advantageous to the larvae of stream-breeding montane frogs. Plectrohyla tadpoles were seen in the Quebrada El Cusuco and a tributary on both trips. During the first visit one lot was collected (14 April 1979) which proved to be Plectrohyla guatemalensis. Of the lots collected (22 May 1980) during the second visit, one is referable to P. dasypus. In addition, two individuals which had nearly completed metamorphosis were found clinging to rocks just - above the stream during midmorning of 21 and 22 May 1980. The mating call of Plectrohyla dasypus is unknown. One female collected 22 May 1980 is gravid. Etymology.—The specific name is the generic name of the nine- banded armadillo, Dasypus novemcinctus, and is used as a noun in apposition. The name refers to Cusuco, the local name for the armadillo. SUMMARY A new species of tree frog Plectrohyla dasypus is described from A NEW HYLID FROG OF THE GENUS PLECTROHYLA ¢ a cloud forest in the Sierra de Omoa of Honduras. The new species belongs in the sagorum group as defined by Duellman (1970). In- formation on the natural history is presented and a description of the tadpole is included. RESUMEN Una nueva especie de rana arbérea ha sido descubierto en Honduras. La nueva especie es Plectrohyla dasypus y pertenece al grupo de los sagorum tal como los define Duellman (1970). In- formacion sobre la historia natural es presentada y una descripcién del renacuajo es incluida. ACKNOWLEDGEMENTS We thank our friends Gustavo A. Cruz Diaz of the Direccién General de Recursos Naturales Renovables and Jorge Porras Zuniga of the Banco Centroamericano de Integracién Econémica, Teguci- galpa, for help in obtaining topographic maps and permits to work in the area of Cerro Cusuco. Gustavo A. Cruz Diaz also provided valuable field assistance during a portion of our first trip to Cusuco. Also we are grateful to Sigfrido Burgos Flores, director of the Direccién General de Recursos Naturales Renovables for issuing us collecting permits. Louis Porras was instrumental in producing the prints for the illustration. William E. Duellman of the University of Kansas Museum of Natural History loaned comparative material of other members of the genus Plectrohyla and critically reviewed the manuscript. LITERATURE CITED DuELLMAN, W. E. 1970. The hylid frogs of Middle America. Monogr. Mus. _ Nat. Hist., Univ. Kansas (1): xi + 753 pp. Gosner, K. L. 1960. A simplified table for staging anuran embryos and larvae with notes on identification. Herpetologica 16: 183-190. Hoxprincg, L. R. 1967. Life zone ecology. Trop. Sci. Center, San José, Costa Rica. 124 pp. McCranie, J. R., L. D. Witson, and L. Porras. 1980. A new species of Leptodactylus from the cloud forests of Honduras. J. Herpetol. 14(4): 361-367. Meyer, J. R., and L. D. Witson. 1971. A distributional checklist of the am- phibians of Honduras. Los Angeles Co. Mus. Contrib. Sci. (218): 1-47. UNIVERSITY OF KANSAS PUBLICATIONS MUSEUM OF NATURAL HISTORY The University of Kansas Publications, Museum of Natural History, beginning with volume 1 in 1946, was discontinued with volume 20 in 1971. Shorter research papers formerly pub- lished in the above series are now published as Occasional Papers, Museum of Natural History. The Miscellaneous Publi- cations, Museum of Natural History, began with number 1 in 1946. Longer research papers are published in that series. Monographs of the Museum of Natural History were initiated in 1970. All manuscripts are subject to critical review by intra- and extramural specialists; final acceptance is at the discretion of the publications committee. Institutional libraries interested in exchanging publications may obtain the Occasional Papers and Miscellaneous Publica- tions by addressing the Exchange Librarian, The University of Kansas Library, Lawrence, Kansas 66045. Individuals may pur- chase separate numbers of all series. Prices may be obtained upon request addressed to Publications Secretary, Museum of Natural History, The University of Kansas, Lawrence, Kansas 66045. Editor: E. O. WiLEy Managing Editor: Jos—EpH T. CoLLins PRINTED BY UNIVERSITY OF KANSAS PRINTING SERVICE LAWRENCE, KANSAS pee) “Of Lo MUS. COMP. ZOOL: LIBRARY OCCASIONAL PAPERS JUL 2 1 198} of the _HARVARD MUSEUM OF NATURAL HISTORY ~ The University of Kansas Lawrence, Kansas NUMBER 93, PAGES 1-64 JULY 7, 1981 LIFE HISTORY OF THE SLENDER MADTOM, NOTURUS EXILIS, IN SOUTHERN ILLINOIS (PISCES: ICTALURIDAE) By RicHarp L. MAYDEN AND Brooks M. Burr’ The ictalurid catfish genus Noturus, commonly referred to as madtoms, presently includes 25 described (Taylor 1969, Douglas 1972, Etnier and Jenkins 1980) and one undescribed species (Jenkins 1976). Noturus is the largest of the six ictalurid genera and is endemic to eastern North America. Most of the species are small (generally reaching a maximum size of less than 100 mm standard length), primarily nocturnal and cryptic in coloration and all pos- sess toxins associated with the pectoral and dorsal fin spines (Taylor 1969). Much of the literature on Noturus has been of a faunal or taxonomic nature and contains little ecological information beyond that of habitat preferences and associated species. No in-depth ecological studies have been published for any of the 26 species; heretofore, all studies have been limited to a few aspects of life history. Taylor (1969) summarized most of the literature on Noturus prior to 1967. Since Taylors summary, a few studies on Noturus biology have been published. Carlson (1966) studied age and growth of N. flavus; Birkhead (1972) examined toxicity in seven species; Mahon (1977) studied age and fecundity of a popu- lation of N. gyrinus; Menzel and Raney (1973) discussed hybridi- zation within Noturus; and Mayden et al. (1980) presented in- formation on some life history aspects of N. albater. The paucity of ecological information on species of Noturus 1 Department of Zoology, Southern Illinois University, Carbondale, Illinois 62901. Present address (RLM): Museum of Natural History, University of Kansas, Lawrence, Kansas 66045 2 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY prompted a study of the slender madtom, Noturus exilis. Natural history studies of N. exilis (Fig. 1) have been limited to habitat preference, associated species (Taylor 1969) and a brief report on food habits (Curd 1960). The range of N. exilis extends from east- ern Kansas to east-central Tennessee and from southern Minnesota to northern Alabama (Taylor 1969). Recently, Robison and Winters (1978) discovered N. exilis, for the first time, in the Red River drain- age, Oklahoma, and attributed the range extension to a probable bait-bucket introduction. It is the purpose of this paper to elucidate the major aspects of the life history of N. exilis, based on a two-year study conducted in southern Illinois. STUDY AREA Green Creek, 7.2 km W Anna, Union County, Illinois, the pri- mary area chosen for study, is a small, spring-fed, first-order tribu- tary of Clear Creek that flows entirely through upland limestone beds of the Shawnee Hills. The substratum is predominately ‘a mixture of fairly clean, small to coarse gravel with occasional large rocks. In deep pools, the typical gravel substratum has a thin overlying layer of sand, silt and organic debris. Stream width ranges from 1.5 to 7.5 m; water depth from 4 cm in rifles to about 2m in pools. Most of the stream is less than 0.5 m deep. Stream banks are 2-4 m high and have little slope. The stream system is bordered throughout by trees, but adjacent land through which most of the stream flows is used predominately for farming; in some regions adjacent land is heavily forested. Water clarity in Green Creek is good during most months but in spring and early summer is somewhat turbid due to increased farm runoft. Water temper- ature ranges from a low of 1°C in January to a high of 32°C in July. In Green Creek, two predominant habitats are present during most of the year: pools and riffles. In the spring, during heavy rainfall, those habitats are, however, not discernible. The locations of specific habitats in the study area vary from time to time because of the shifting gravel substratum. Hutchins Creek, a second study area used for nesting obser- vations, 4.8 km ENE Wolf Lake, Union County, Illinois, is phys- ically similar to Green Creek, except that it flows mostly through heavily-forested land, banks are short and sloping, water is usually clearer and the stream appears to be more stable. Because of forested lands adjacent to Hutchins Creek, less flooding occurs in spring and summer as compared to Green Creek. The barren farm- lands bordering most of the Green Creek system and the narrower stream width and steeper banks enhance frequent flooding of the stream. The Hutchins Creek site is located about 8 km N of the Green Creek site. LIFE HISTORY OF THE SLENDER MADTOM 3 Fic. 1.—Noturus exilis from Green Creek. Above, 100.0 mm SL male in breeding condition collected 25 May 1979. Below, 81.0 mm SL gravid female collected 23 April 1979. The portions of Green and Hutchins creeks where nests of N. exilis occur are all similar in having a substratum of clean gravel, very little sand or silt, many large rocks and slow current. The 4 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY streams in these areas are partially shaded at some time of the day by overhanging trees, but are in direct sunlight at other times. MATERIALS AND METHODS Observations and minnow-seine collections were made at Green Creek between 6 November 1977 and 29 October 1979, at approxi- mately one-month intervals. During the breeding season obser- vations and collections of nests, eggs and nesting adults were made two-three times a week. In order to limit the effects: of sampling on the Green Creek population, monitoring of nesting activity was confined mostly to Hutchins Creek. Observations of nesting were made only in the summer of 1979. Some observations were made at Green Creek but no nesting adults, eggs nor larvae were taken. A total of 972 specimens was preserved in 10 percent formalin and later stored in 70 percent ethyl alcohol. All specimens are de- posited in the Southern Illinois University at Carbondale ichthyo- logical collection. Collections were made by habitat (e.g., pools and rifles), except during periods of high water when specific habitats were not discernible, and specimens from each habitat were kept separate for comparative purposes. Potential predators were col- lected for examination of stomach contents. In the laboratory, all specimens were sexed, measured and aged; adjusted body weights (specimen’s weight after removal of stomach, intestine, liver, kidneys and gonads) were recorded for a sample from each age class, representing all sizes of madtoms in the age class, of each monthly sample (N = 249). Gonads of all specimens were excised and examined under a stereomicroscope and an age and size sample (N = 243) of each sex was weighed. All specimens and gonads were weighed after blotting, on analytical and topload- ing balances and weights recorded to the nearest 0.001 gm. Stand- ard (SL) and total lengths (TL) were measured to the nearest 0.1 mm with dial calipers. The gonadosomatic index (GSI) is the gonad weight as a proportion of adjusted body weight. An age and size sample of each collection was dissected and stomach contents and endoparasites were identified. All specimens were inspected for external parasites. Relative survival of each year class was calcu- ~ lated by expressing the number of individuals in that year class as a proportion of the number of individuals in the 0 and/or next youngest year class (Ricker 1975). Age of specimens was determined by counting number of annuli on cross-sections of pectoral fin spines as outlined by Marzolf (1955) and Clugston and Cooper (1960). Preliminary attempts to remove spines, using methods outlined by Sneed (1951), yielded unsatis- factory results for the small spines of N. exilis; the spines were frequently broken. A method that lent itself to successful removal of all spines consisted of the following. First, the muscle around LIFE HISTORY OF THE SLENDER MADTOM 5 the base of the spine was cut. Then the tissue was removed from the outside of the spine and the spine was cut free from the re- mainder of the pectoral fin. The spine was then gently manipulated until it could be pushed in a postero-lateral direction to rest along the side of the abdomen. The spine could then be slowly and gently rotated in a clockwise (for left spine) direction by hand, rather than pliers, until it was removed. Pliers tended to break or crush the spines. Spines were mounted and sectioned as outlined by Clugston and Cooper (1960) except as follows: a second application of cement was helpful in holding the spines in place while sections were cut and the circular blade was reversed. Reversing the blade resulted in a smoother cut through the slide. After sections were cut, they were placed on microculture slides in immersion oil and observed by means of transmitted light under a stereomicroscope at 50X. As in Noturus insignis (Clugston and Cooper 1960), Ictalurus punctatus (Sneed 1951, Marzolf 1955) and Cyprinus carpio (Carl- ton and Jackson 1964), cross-sections of N. exilis pectoral fin spines typically had narrow, translucent rings (annuli) alternating with wider, opaque rings, representing winter and summer growth, re- spectively. For age class 0, the first annulus was first noted in March and April. Prior to those months spine sections of age 0 specimens appeared completely dark. Aging to month was done by using June, the month of greatest breeding activity, as month zero. For some comparisons males and females were divided into young (1-13 months old) and adult (14 months or older) age groups. Diel feeding activities were examined from 12 samples of five individuals each (N = 60), taken at two-hr intervals over a 24-hr period. Collections were begun at 1600 hr on 27 September 1979 and ended at 1400 hr on 28 September 1979. For the 60 individuals collected, numbers of each prey taxon and the condition of stomach contents were recorded. An index of digestion (DI) was calculated by weighting the condition of stomach contents using a scale of 0-5 as follows: 0-empty, 1-well digested, 2-moderate digestion, 3- partial digestion, 4-only minor signs of digestion, 5-no digestion (fresh material). A mean value was obtained for each hour by multiplying the number of individuals of each digestion class taken during the hour times the value of that class (0-5), summing all values, and dividing by 5. For example, if individuals were scored as all containing class 4 stomach contents, then the weighted value for that would be 4 (i.e.,5 x 4 = 20/5 = 4). Development of N. exilis was examined from six clutches of eggs, one spawned in the laboratory and five taken from Hutchins Creek. The clutch spawned in a 19 liter aquarium was fertilized at approximately 0600 hr, 13 June 1979. The ages of the five clutches 6 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY (at the time of capture) from Hutchins Creek were estimated from stages obtained from the clutch spawned and reared in the labo- ratory. All eggs and larvae were reared in charcoal-filtered tap water in 11] mm diameter culture dishes. Dishes were aerated with air stones and partially submerged in a ‘water bath maintained at 25°C. After most of the yolk sac had been absorbed, pulverized catfish chow was introduced into the dishes daily. Prior to hatching, one- two specimens were removed from a clutch every two-six hr and, after observations were made, were preserved in 10 percent forma- lin. After hatching, one-two specimens were removed and pre- served daily. Drawings of embryos and larvae were prepared with the aid of a camera lucida at 25X. Observations of live and un- stained preserved specimens were made using a stereomicroscope (6-50X). All statistical analyses were executed on a computer using Sta- tistical Analysis System (SAS). Solutions for regression coefficients were solved by the process of least squares. F-tests were used for testing group differences with respect to regression coefficients and homogeneity of regression. One-way analysis of variance tests were used to determine significant differences in means of samples de- termined by sex, age, year, month or habitat. Correlation coeffi- cients (r) were all Pearson product moment correlations. Unless otherwise stated, the significance level used for all tests was « = 0.05. The abbreviation SD = standard deviation. HABITAT The habitat that N. exilis occupied most frequently varied with age and, for mature individuals, with season (Table 1). During February and March, high water precluded the identification of specific habitats. A higher percentage of adults inhabited pools during all months except June and July, when most individuals were found at the crests of or in rifles. The shift by adults from pools to rifles in June and July is related to nest building and reproduction during those months. Young N. exilis occurred mostly in shallow riffles but were also found in shallow margins of pools (Table 1). Little seasonal vari- ation in habitat occurrence was noted for young individuals, except predominate occupation of rifles in late summer when young-of- the-year were small. Elsewhere in its range, N. exilis has been reported to occur in gravel, rubble and slab-rock riffles, of small to medium-sized streams (Taylor 1969). Within either habitat, during daylight hours, both young and adult N. exilis were usually found under large rocks and other objects or in the interstices of gravel, during all seasons. At night and during twilight hours, young and adults were more active and LIFE HISTORY OF THE SLENDER MADTOM 00S 00S V 6 09 T6& v9 = iM Snead Jequis.aq L‘9 €€6 cI LIV 68S eL <> Sa ee ee a JaqUIsAON OFG 092 GG O'SP oss OF Tb ae ea ial rr are 13q0O TTé 689 CV 69 T ve VF > A) ee Ne ee Jequia}dag OLE Goo 9S SOL GES 39S = Sse CL ae ee soon, om oy L9G 09 0'00T 0'0 GL > Gh Mapa Sele Oh ee a a Ajn{ TL9 6GE 68 00 00 0) SS ae ode ae ee ee 001 0°06 Or Leg 69 g9 ~~ ee eo ee, eee ee ae 6 SV gg9s €¢G 9'GYV VLS [9% -)cp-aere Spraioe te eee a Judy =" a gros Sie ia soa Se Sha ee ae ge a rT ae Yoel ster van ead L'S8 eri cs Te tears aa 5 caer ea area Ole ILC ONT GOV g8’s¢ €T 9°8S VIP «es TSS = Arenue{ sop sjoog pe y~eT[9D, Som sjoog p2}°2][9D UI Syixa "NT JO OBeYUDIEg Jequinny Ul syixa ‘NJ JO OBvJWOOINg Jequinyy yoy s}yOpV sunox ‘6LEL 1940}990 GF PUB LLGT JequieAoN g UseMJoq JeYQGeY Aq Yee1D UseID WOIZ syIxa sninjoNy YNpe pue ZuNOA Jo uoNNGINSIq “[T AIaV], 8 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY were rarely found hidden in or under objects. Stegman and Minckley (1959) also found N. exilis and two other riffle-inhabiting species of fishes, in Hutchins Creek, occupying interstices of gravel in an area of subsurface percolation. Although N. exilis occurs syntopically throughout most of its range with one to eight other species of Noturus (N. albater, N. elegans, N. flavater, N. flavus, N. gyrinus, N. nocturnus and N. placidus) (Taylor 1969), no other species of Noturus occurs in Green and Hutchins creeks. Fishes commonly collected with N. exilis in Green and Hutchins creeks included Campostoma anoma- lum, Notropis lutrensis, N. stramineus, N. umbratilis, Phenacobius mirabilis, Pimephales notatus, Semotilus atromaculatus, Ictalurus natalis, Fundulus olivaceus, Gambusia affinis, Lepomis cyanellus, L. macrochirus, L. megalotis, Etheostoma caeruleum, E. flabellare, E. nigrum and E. spectabile. Other species associated with N. exilis, but collected infrequently included Amia calva, Notemigonus crysoleucas, Catostomus commersoni, Moxostoma duquesnei, M. erythrurum, Ictalurus punctatus, Labidesthes sicculus, Aphredo- derus sayanus, Cottus carolinae, Micropterus punctulatus and Po- moxis annularis. AGE AND GROWTH Age Ages of the 972 specimens taken from Green Creek ranged from 1 to 59 months (Table 2), using June as a standard origin. Indi- viduals less than 12 months old were designated age class 0 or 0+, depending on the month of capture. Individuals 12-23 months old were 1 or 1+, those 24-35 months old, 2 or 2+, those 36-47 months old, 3 or 3+ and those 48-59 months old, 4 or 4+. Most (N = 967) individuals were of age classes 0, 1 and 2, with only five in age classes 3 and 4. Growth The earliest date when young-of-the-year (1 month) were cap- tured was 31 July 1978. Lengths ranged from 26.1 to 34.1 mm (x = 29.2, N = 12) and weights from 0.18 to 0.34 gm (x = 0.25, N = 12). In 1979, young were not taken until 2 months of age on 24 August and ranged from 22.3 to 34.5 mm (X == 29.4, N = 6) in length and 0.21 to 0.69 gm (x = 0.53, N = 6) in body weight. The smallest specimen captured was a 2-month-old, 19.0 mm female taken in August, 1979. Morphology and color of 1- and 2-month-old speci- mens were similar to adults. Individuals of N. exilis grew at a decreasing rate in length and at a constant rate in body weight for at least 3-++ years for females and 4-+- years for males (Table 2 and Figs. 2, 3 and 4). An average LIFE HISTORY OF THE SLENDER MADTOM 9 individual of either sex reached one-half of the first year’s standard length and body weight in about three weeks (Fig. 2) and four months (Fig. 3), respectively. Curvilinear growth equations for length did not differ signif- icantly between the sexes (Fig. 2). For the sexes combined, the relationship between standard length (L) and age (A) was L = 23.0 + 2.41A — 0.022A?, with r — 0.847. Standard lengths of males and females within one age class were not significantly different. Adjusted body weight increased at a constant rate with increas- TaBLE 2. Standard lengths and adjusted body weights of Noturus exilis from Green Creek by age (in months) between 6 November 1977 and 29 October 1979. Age N Standard Length (mm) Body Weight ( gm) Mean Range SD Mean Range SD il oF 29.2 26.1-34.1 2.47 0.25 0.18-0.34 0.067 2 56 28.6 17.4-37.6 D.20 0.56 0.21-0.69 0.159 3 44 33.9 24.7-47.6 6.24 0.81 0.20-1.58 0.454 4 50 35.0 26.6-45.7 4.98 0.93 0.29-1.53 0.378 5 73 34.1 19.5-53.0 8.25 0.92 0.47-1.49 0.351 6 64 35.1 22,.8-51.2 7.07 1.03 0.49-1.72 0.408 7 70 37.9 23.2-49.9 5.78 0.89 0.20-1.79 0.555 8 67 36.7 22..7-54.3 7.49 1.21 0.69-1.83 0.406 9 60 38.9 27.5-59.1 6.56 1.12 0.30-2.01 0.518 10 61 42.6 29.6-57.2 6.53 1.42 0.40-2.39 0.626 11 65 43.3 28.8-57.8 6.52 1.44 0.49-2.08 0.486 12 79 49.8 37.5-63.2 5.35 2.17 0.80-3.56 0.620 13 65 51.5 39.6-65.6 6.36 2.33 0.86-4.43 1.027 14 54 54.6 40.3-65.6 5.77 2.59 1.07-3.94 0.890 15 45 61.1 51.3-72.0 4.96 2.70 2.07-3.29 0.405 16 25 59.9 53.0-71.0 4.71 2.23 1.95-2.70 0.275 17 11 61.3 55.1-70.9 5.10 2.63 1.67-4.54 0.925 18 2 60.8 59.4-62.2 1.98 2.80 2.70-2.90 0.141 9 59.4 53.7-66.1 3.73 2.74 1.86-3.87 0.614 20 4 64.6 62.8-67.2 1.87 3.64 3.17-3.92 0.411 81.7 81.2-82.1 0.64 9.70 8.30-11.10 1.980 TOO aE ae a 21. 4 61.4 57.6-64.9 3.02 2.85 2.36-3.40 0.521 22 19 66.0 58.2-73.8 4.15 4.35 2.78-6.53 1.035 23 7 68.8 63.9-73.7 3.89 4.62 3.90-6.13 0.818 24 3 70.1 61.4-78.2 8.41 4.94 2.91-6.16 1.772 25 2 71.4 70.2-72.6 1.70 3.87 2.72-5.02 1.625 26 2 71.0 66.9-75.0 5.73 4.52 3.34-5.69 1.662 29 4 67.7 65.5-70.9 2.89 2.91 2.67-3.16 0.344 30 1 63.9 Zale Jat SS es 31 3 69.6 69.2-69.8 0.32 3.98 3.41-4.55 0.808 32 1 Ania ie © U ie 5.06 as ane 33 1 69.5 = = 5.06 en = 34 2 73.8 71.9-75.7 2.69 6.37 5.88-6.87 0.702 35 2 75.6 71.2-80.0 6.22 6.26 4.64-7.88 2.291 42 1 SOS ef a — 5.59 She = 43 1 78.0 ane 5.93 ne 2 1 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 10 *xaS Yyora 10F sajdures [e}0} yuasaidai soul] UOIsseIs9Yy “So[BUay 1OF suPOUT ajdures yuasaidai sapoito uado ‘sajeut 107 suveut ofduies yuasaidal sapo1l9 pljOg “Ya2ID Useig) WIZ syixa snunjoN JO ysue] prepurjs ul YWoIg—%. “Sly SHLNOW NI 39V 09 8S 9S ¥S ZS OS BF IF YH ZH OF BE IE VE ZE OL BZ 9Z PETZ OT BI IL vi ZLOL 8 9 ¥ Z O Piet at ee ee 02 fe} ‘o- ss80=4 JIWW34 z¥ S200 VSV'~T+SL7Z=1 ars 6€8'0=4 ‘J1IVW ,V2t00-VWrVe +16 77=1 , 8 OZ (WwW) HILONIT GYVGNV1LS 08 e ool 11 LIFE HISTORY OF THE SLENDER MADTOM *xaS YORE 10F so[duies [¥}0} JUase1da1 soul] UOISSOIBIY “So[BUloy OJ Suvaut ojdures yuasaidai sajo11o uado ‘so[eul 10F suvat afduies yussaidat sapo1lo pYOG “Yee1D useig WoIZ syIxa snunjoN JO yyZtIem Apoq pajsn{[pe ur YyoIg—E “1A SHLNOW NI 39V 09 8S 9S ¥S ZS OS BY 9 Hh Z OF BE IE VE ZE OF BZ 9Z HZ ZZ OT BI PL vi ZLOlLs 9 ¥ Z O 0) BV 0gc84e8 l 8054 = é 090 6 8 °5 ce : fo) Zee ° O -0> : a we S22 4 e Sh ee - Ba o @ O (e) Ea 9 a) wit @ < Ie ie a m oe e 2 Oo ‘ ae 64 Ol $ Lt 6SB'0=4'31vwad W9SLO+Ps0'0=mM SPA? 7 . e El 006'0=4 '3IVW VL6l'0+6E1'0-=M @e ee ral Sl 12 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY ing age in both males and females and was depicted best with linear equations (Fig. 3). Linear growth equations for body weight differed significantly (2 — 0.05) between the sexes. For females, the relationship between body weight (W) and age (A) was, W = 0.084 + 0.156A, with r = 0.859 (N = 169); for males, W = —0.139 + 0.197A, with r = 0.900 (N = 75). Adjusted body weights of females over 12 months of age were generally less than those of males (Fig. 3). Prior to 12 months, however, female body weights were greater than or equal to male body weights. The lighter body weight of older females probably involves the allocation of more energy into the production of large, yolk-filled eggs instead of body weight. Significant (« = 0.05) dif- ferences between the sexes were observed only for age classes 0 and 1. The lack of significance at older ages is attributed to sample size. For age class 0, body weights ranged from 0.182 to 2.393 gm (x = 1.110, N = 70) for females, and from 0.199 to 2.010 gm (xX = 0.867, N = 33) for males; for age class 1, ranged from 0.861 to 5.171 gm (x = 2.627, N = 85) for females and from 0.795 to 6.534 gm (x = 3.178, N = 37) for males. The largest male captured was 100.0 mm, weighed 11.66 gm and was 59 months old. The largest female was 81.2 mm, weighed 8.31 gm and was 46 months old. Taylor (1969) reported that most N. exilis were less than 100 mm but mentioned a 113 mm specimen from an aquarium at the University of Michigan Museum of Zool- ogy. Subsequent to Taylor's revision, four specimens of N. exilis were collected from Red River, Simpson County, Kentucky (Kentucky Fish and Wildlife cat. no. 1975), that measured 110.0- 124.0 mm. The relationship between adjusted body weight (W) and stand- ard length (L) for all specimens combined, differed significantly (2 = 0.05) between the sexes. Males maintained a greater body weight per mm standard length than females. The length-weight regression equation for females was Log W = —4.92 + 3.03 Log L, with r = 0.979 (N= 166) and for males was Log W = —4.63 + 2.88 Log L, with r = 0.982 (N = 75). . A comparison of all age classes, sexes combined, revealed no significant differences between age classes for the length-weight relationship. The relationship between W and L for both sexes and all ages combined was Log W = —4.77 + 2.94 Log L, with r = 0.979 (N = 241). For any age, specific growth rates in body weights of males and females were not significantly different because of the linear re- lationship between age and body weight (Table 3 and Fig. 5). Therefore, the sexes were combined and a mean male-female body weight was used to calculate specific growth rates. Because of variation in standard lengths of specimens of one age (Table 2 and LIFE HISTORY OF THE SLENDER MADTOM 13 Fig. 4), combined mean monthly male-female body weights (Table 3) were predicted from a mean monthly standard length at each age (Table 3 and Fig. 4) and sex specific length-weight regression equations. The observed loss in length for some ages (Fig. 4) was probably due to sampling bias and therefore, a line indicating no change in length was used in predicting body weights. The equation used to express the periodicity of growth in body weight of an average N. exilis was that by Brown (1957) and Clug- ston and Cooper (1960). For N. exilis, the time interval used was one month, since monthly samples were made. In order to calculate the rate of growth for individuals 1 month old, a recently hatched N. exilis (8.5 mm, 0.02 gm) from Hutchins Creek was used as the length and weight for age zero months. Specific growth rates of N. exilis varied with age and season (Table 3 and Fig. 5). Growth was most intense in summer months during the first two years of life, with a maximum growth rate of about 215 percent per month for individuals between 0 and 1 month of age. During winter months, growth rates for all age TasLE 3. Mean specific growth rates in adjusted body weight of male and female Noturus exilis from Green Creek. Age Mean Mean Mean Age Mean Mean Mean (mo) Standard Predicted Specific (mo) Standard Predicted Specific Length Body Growth Length Body Growth (mm) Weight Rate (mm) Weight Rate (gm) (gm) 0 8.5 0.02 Ee 24 70.1 4.78 9.0 1 22.5 0.17 214.0 95 71.0 4.96 3.7 2 29.8 0.38 80.4 26 71.0 4.96 0.0 3 33.0 0.52 31.4 27 71.0 4.96 0.0 4 34.3 0.58 10.9 28 71.0 4.96 0.0 5 34.9 0.61 5.0 29 71.0 4.96 0.0 6 35.4 0.64 48 30 71.0 4.96 0.0 7 36.3 0.68 7.5 31 71.0 4.96 0.0 8 37.3 0.74 8.5 32 71.0 4.96 0.0 9 39.0 0.90 14.0 33 72.0 5.17 4,2 10 41.4 1.01 7 34 73.3 5.45 ye" abil 44.5 1.25 Dies 35 75.0 5.83 6.7 12 48.0 1.56 52), ) 36 77.0 6.30 7.8 13 51.8 1.96 22.8 37 78.5 6.68 5.9 14 55.7 2.42 21.1 38 79.0 6.80 1.8 15 58.8 2.84 16.0 39 79.5 6.93 1.9 16 60.6 3.11 9.1 40 79.5 6.93 0.0 17 60.5 3.09 -—0.6 Al 79.5 6.93 0.0 18 60.5 3.09 0.0 42, 79.5 6.93 0.0 19 61.6 3.26 5.4 43 79.5 6.93 0.0 20 61.9 3.31 1.5 44 79.5 6.93 0.0 21 63.7 3.60 8.4 45 80.0 7.06 1.9 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 14 OO —— ‘uonoadsut Aq pany stM pue ‘ose Ul BSvaIOUT UP YIM YSUE] Ul asvo1Dep OU Surumnsse ‘y}Mo01Z powimnsaid sjusseido1 sul] peyoyeH{ *See14} FO @SeIOAR SUIAOU e Aq pany CE-TE pue gZ-0 sade usamjeq ou] Plog “UvEUT OL} FO SIOLIO PIvpURIS ¢ = yosoidei souy [eonieA pur Yy}Sue] prepur}s uvout yuasaidair sjulog ‘adv JO syyUOW QF pUe Q UsaMjoq (SaxeS peurquios ) syixa sninjon JO Yue] piepur}s Ul YMOIN—F “I SHLNOW NI JOV 09 8S 9S ¥S ZS OS BF 9 vy ZH OW BE IE VE ZE OF BZ 9% we ZZ OF BL Ll vl ZL Ol 8 9 ¥F Z O Ol 0z t+oe nn = t 3 ov z > >] 0 { pS-=. m z= 9) f 09 z 3 ON Oe Ae ou 2 Pe oe ae 08 06 ool LIFE HISTORY OF THE SLENDER MADTOM 15 SPECIFIC GROWTH RATE JUL JAN JUL JAN JUL JAN JUL JAN AGE IN MONTHS Fic. 5.—Chronological and seasonal variation in specific growth rate of adjusted body weight for the total sample of Noturus exilis between 0 and 46 months of age. Line fitted by a moving average of threes. groups decreased considerably and approached zero. Younger in- dividuals, however, grew more during colder months than older individuals. After the second summer of life, when maturity was reached, the magnitude of growth and length of the growth period were reduced considerably during warmer months (Fig. 5). In many species of fishes, a decreased growth rate during warmer months for older age classes is indicative of the onset of maturity (Nikolsky 1963, Lagler et al. 1977). Maximum growth also occurred earlier in the season for 2- and 16 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 3-year-olds than in younger age classes. The earlier peak in growth and narrower growth period for 2- and 3-year-olds were probably due to the allocation of most of their available energy to growth during the prereproductive season and then to reproductive proc- esses. Age 1 individuals, however, used a major portion of available energy for growth early in the spring as well as throughout the spawning season. Growth rates for Noturus insignis, although in daily intervals, were similar in pattern to those for N. exilis (Clugston.and Cooper 1960): a decreasing winter and summer rate-of-growth with in- creasing age and a decrease in growth rate at age of maturity. Clugston and Cooper (1960), however, noted that growth rates for mature N. insignis were greatest later in the summer and proposed an increase in energy investment for growth after the spawning season rather than before, as in N. exilis. ‘ REPRODUCTION Reproductive Cycle of the Male Testes of mature male N. exilis, like those of other ictalurid catfishes (Sneed and Clemens 1963), were elongated, whitish and opaque with long, finger-like projections of varied sizes. Testes of 1 month old, 26 mm males were similar in appearance to those of older males but were translucent and had smaller projections. Testis-weight-to-adjusted-body-weight (GSI) ratios were re- corded for September 1978 through August 1979 (N = 72). Only slight seasonal variation in testis structure and size was evident. The testes of mature and a few immature males began to thicken and the projections began to lengthen as the breeding season ap- proached (Fig. 6). For males 2 years or older, GSI increased from late fall to June when a maximum GSI of 7.12 (0.7%) was reached. After June, GSI decreased until November when an increase for the spawn in the following year was probably initiated. The decrease in GSI after July corroborates the ending of the breeding season of N. exilis in July (see Nesting). The relationship between GSI (Y) and month (X), with August = 1 and July = 12, was Log Y = 0.312 + 0.040X, with r = 0.720. For males less than 2 years of age, an increase in GSI over the winter and spring was also evident, but, unlike older males, GSI values were significantly (« = 0.05) less than those of adults and considerable variation in GSI was observed over the months (Fig. 6). The relationship between GSI (Y) and months (X), with August = 1 and July = 12, for age class 0 males was Log Y = 0.182 + 0.022X, r = 0.335 (not significant at « = 0.05). A few age 0 males displayed a steady increase in GSI over the months and reached a peak as great as that reached by older males that had LIFE HISTORY OF THE SLENDER MADTOM 17 already spawned. However, it is concluded that males do not breed until 2 years of age. The relationship between weight of testes (T) and standard length (L) of males was curvilinear (Fig. 7) and was best depicted by the equation 1000T = 15 — 0.9L + 0.014L?, with r = 0.912. Testis development was minimal until the male reached about 50 mm SL. With increasing size (or age) of the male, the rate of in- crease in testis weight increased with minimal seasonal variation. Between April and August mature males had enlarged cephalic epaxial muscles, swollen lips (Fig. 1) and swollen genital papillae (Fig. 8). Their heads broadened and enlarged and often had scratches on the dorsal surface and nape. Non-breeding mature and immature males did not have enlarged cephalic epaxial muscles and lips and were similar in appearance to females. The genital papillae of non-breeding males were not swollen but were different in structure from those of females (Fig. 8). General coloration of males was similar to that described by Taylor (1969) during all months, except as follows: the adipose fin was dark basally and light orange or yellow distally; the caudal fin was light orange or yellow medially and basally; and the venter, around the eyes and small areas anterior and posterior to the dorsal fin were light yellow. Judged by primary and secondary sexual characteristics and nesting habits, sexual maturity of male N. exilis was attained in the third summer of life (2 years old). Some males entering their second summer of life (1 year old) had characteristics of larger, older males but were not as well developed. Maturity of males of Noturus insignis is also attained at 2 years of age (Clugston and Cooper 1960). Nest-guarding males of N. albater were all 2 years of age ( Mayden et al. 1980). Reproductive Cycle of the Female As judged by gross examination of ovaries, GSI index (Figs. 9 and 10), degree of abdominal distention and condition of the geni- tal papilla (Fig. 8), females generally began to prepare for spawn- ing in April and were found in breeding condition until late July (Figs. 9 and 10). Females were similar in pattern and color to males, but did not develop swollen cephalic epaxial muscles and lips. For the two years studied, a relative increase in ovary size and weight was observed in all age classes of females in late fall and continued to the spawning period the following summer (Figs. 9 and 10). However, the rate of increase of GSI over time was sig- nificantly (« = 0.01) less in young females than in adult females. This probably reflects differences in ovarian cycles of young and adults: gradual development of ovaries in young females and, in adults, rapid growth of the ovaries with the production of large, mature, yolk-filled oocytes in preparation for spawning. For young 18 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 10 YOUNG Log Y=0.182+ 0.022X, r=0.335 ADULT Log Y= 0.312 + 0.040X, r= 0.720 e WEIGHT OF TESTES (gm) X 1000/ADJUSTED BODY WEIGHT AUG SEP OCT NOV DEC JAN FEB MAR APR MAY JUN JUL MONTHS Fic. 6.—Monthly variations in weight of testes in relation to adjusted body weight for 75 Noturus exilis males. Young males include those less than 14 months of age. Adult males are those 14 months or older. Males taken from nests containing eggs at Hutchins Creek are indicated by open circles. Solid circles are males collected from Green Creek in monthly samples. The ordinate is a logarithmic scale. 19 LIFE HISTORY OF THE SLENDER MADTOM SOL 66 €6 ‘So[BUL syixa SninjON GL UL YSUe, prepuL}s puv sajso}z JO JYSIOM UsaMJoq AdiysuonePayY— ) “OIA (ww) HION31 GYVGNVLS GL 69 £9 LS Ks SY 6€ ce ike \Z Sl 000l x (w5) saisal 40 LHOISM 7160 =4 {iv lo0+ 16'0-S| =10001 20 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Fic. 8.—Genital papillae of Noturus exilis from Green Creek. A) Non- breeding male 70.0 mm SL collected 7 January 1978. B) Breeding male 92.0 mm SL collected 25 May 1979. C) Non-breeding female 78.0 mm SL collected 7 January 1978. D) Breeding female 71.5 mm SL collected 27 June 1979. Left is anterior. females, the relationship between GSI (Y) and month (X), with August = 1 and June = 11, was Log Y = 0.699 + 0.064X, with r = 0.493 (Fig. 9). The relatively low r value reflects high vari- ability in sexual development of age class 0 females. Adult females showed a rapid increase in GSI early in the fall after spawning, suggesting the development of immature oocytes to replace the mature oocytes spawned. Little ovarian development occurred over winter months followed by resumed rapid growth in the spring (Fig. 10). The relationship between GSI (Y) and month (X) for adult females, with August = 1 and May = 10, was Log Y = 0.703 + 0.124X, with r = 0.822 (Fig. 10). The higher r value for adult females was due to the consistency in sexual development throughout the year. In addition to a greater ovarian growth rate, adult females main- tained a significantly (« = 0.05) greater mean GSI over time as compared to young females. In 18 females of all ages (Table 4) taken from April to July, GSI ratios ranged from 0.078 to 0.309, x = 0.211. GSI values for adult females were consistently higher (x = 0.256) than those for young females (x = 0.158), indicating a greater energy expenditure in egg production by adult females. Seasonal variation in ovarian development.—To determine fre- quencies of breeding females in each age class and peak activity in LIFE HISTORY OF THE SLENDER MADTOM 21 500 a MATURE 4 POTENTIALLY MATURE e SPENT Oo IMMATURE 4 >> _ So io) > 50 4 4 LOG Y=0.669+0.64X, r=0.493 (2) WEIGHT OF OVARIES X 1000/ADJUSTED BODY WEIGHT 8) AUG SEP OCT NOV DEC JAN FEB MAR APR MAY JUN JUL MONTH Fic. 9.—Monthly variations of ovarian weight in relation to adjusted body weight for 114 Noturus exilis females 1 to 13 months of age (two years com- bined). The ordinate is a logarithmic scale. spawning, females were classified on the basis of egg size and con- dition of genital papillae as being mature, spent, potentially mature or immature for April-August, when marked ovarian changes were observed. Mature females had two distinct classes of oocytes, mature and 22 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 500 4 MATURE 4 POTENTIALLY MATURE ~ e SPENT | * o NONBREEDING : LOG Y=0.703 + 0.124(X), r=0.822 , 100 50 10 WEIGHT OF OVARIES X 1000/ADJUSTED BODY WEIGHT AUG SEP OCT NOV DEC JAN FEB MAR APR MAY JUN JUL MONTH Fic. 10.—Monthly variations of ovarian weight in relation to adjusted body weight for 55 Noturus exilis females greater than 13 months of age (two years combined). The ordinate is a logarithmic scale. immature, and the tissues adjacent to their genital papillae were swollen (Fig. 8). Mature oocytes were large (1.5-3.4 mm), yolk- filled, spherical and orange; immature oocytes were smaller (0.1-1.3 mm) and either clear or white. Spent females had ovaries that contained small, white oocytes LIFE HISTORY OF THE SLENDER MADTOM 23 and fragments of orange yolk; in two females, one and two mature oocytes remained. The tissues adjacent to their genital papillae were usually swollen. Potentially mature females had ovaries that did not contain two well-defined size classes of oocytes, but rather a gradation in sizes of oocytes ranging from tiny (<0.1 mm) and clear to white ones (0.3-0.5 mm), to larger (1.1-2.0 mm), light-yellow ones. Tissues adjacent to their genital papillae were only slightly swollen and ovaries were consistently smaller in size and weight when compared to those of mature females. Immature females were sexed as early as 1 month or 19.0 mm SL. Their ovaries had well-defined, white oocytes and the tissues adjacent to their genital papillae were not swollen (Fig. 8). Non- breeding adult females had similar characteristics. In both 1978 and 1979, the breeding cycles of young and adult females differed, in that among adult females, potentially mature, mature and spent individuals appeared earlier in the April-August season than among young females (Fig. 11). Between 1978 and 1979, breeding cycles of both young and adult females also differed. In 1978, mature adult and potentially mature young females were present in April. In April, 1979, however, adult females were only potentially mature or non-breeding and young females were all immature. The earlier occurrence of ripe females in 1978 and the absence of spent females in August, 1978 indicated an earlier spawn in that year than in 1979. The appearance of young-of-the-year in July, 1978 and in August, 1979 lent support to this hypothesis. A warmer and earlier spring in 1978 probably accounted for this difference. Maturity.—The nesting habits of N. exilis precluded the capture of large numbers of nesting females; only three females were found in nesting cavities. Therefore, age of maturity was based mostly on preserved females from Green Creek. Females generally attained maturity by the third summer (2 years of age) as indicated by 18 of 19 females taken during the April-August season. The one exception collected at the beginning of the reproductive season (April 1979) probably would have matured later that season. Females in their fourth or fifth summer (3 and 4 years of age) were all mature. Thirty-five of 169 age class 0-1 females collected between April and August were possible contributors to spawning. Of the 35, five contained oocytes large enough to be spawned (Table 4) and eight were either spent or resorbing eggs. The remaining 22 were considered potentially mature and showed only slight ovarian de- velopment. Maturity for age class 0 females appeared to occur at about 50 mm in length. The 13 individuals considered mature or 24 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY BB mature VASPENT _}POTENTIALLY MATURE — [_] IMMATURE YOUNG FEMALES PERCENTAGES un oO ADULT FEMALES 100 90 80 70 60 50 40 30 20 10 PERCENTAGES AA GE_ ] ww MQ AWN al APR MAY JUN JUL AUG APR MAY JUN JUL AUG 1978 1979 Fic. 11—Monthly and yearly variation in the percentage of female Noturus exilis in the various reproductive states during the period of gonadal develop- ment from April to August. Young females include those less than 14 months. Adult females are those 14 months or older. Reproductive states refer to con- dition of ovaries and genital papillae. spent were the largest age 0 females and ranged in length from 46 to 59 mm. Potentially mature females were between 40 and 45 mm long. During the spawning months (May-July), age 0 females con- sistently produced fewer (26-53, x = 44.4) and lighter (7.1-10.6 mg, xX = 8.4) oocytes and had a lower GSI than adult females (Table 4 and Figs. 9 and 10). The number of mature oocytes in LIFE HISTORY OF THE SLENDER MADTOM 25 adult females ranged from 64 to 97, x = 85.2; weight of mature oocytes ranged from 12.0 to 16.6 mg, x = 13.8. The number of mature oocytes produced per gm of body weight, however, was similar for adult and young females (Table 4). Fecundity—In 18 mature females representing April to July, the number of mature oocytes ranged from 26 to 150, x = 83.6 (Table 4). Because differences between number of mature oocytes in each ovary of the same female are relatively great (Table 4), doubling the number from one ovary was not acceptable in estimating fecundity. The number of mature oocytes produced by a female is corre- lated with age, size and weight of the female. The oldest, largest and heaviest females produced the most mature oocytes (Table 4). For the 18 mature females taken April-July (Table 4), the relation- ship between number of mature oocytes (F) and age (A) was = 6.36 + 3.45A, with r = 0.912; between number of mature oocytes (F) and standard length (L), F = —142.2 + 3.7L, with r = 0.875; and between number of mature oocytes (F) and ad- justed body weight (W), F = 3.49 + 18.73W, with r — 0.915 (Fig. 12). The best predictors of F were body weight and age. Body weight, as the only predictor of F, accounted for 84 percent of the variance of F. Age, as the only predictor of F, accounted for 83 percent of the variance. Standard length was not a significant second- or third-order predictor of F with age and body weight. However, when standard length was used as the only predictor of F, 76.6 percent of the variance of F could be accounted for. Nesting Observations of nesting in Noturus are limited to six of the 25 described species. The breeding season usually occurs during sum- mer; however, at southern latitudes breeding may occur as early as April (Thomerson 1966, Hellier 1967, Taylor 1969). Mayden et al. (1980) reported N. albater nesting under large stones in July; Greeley (1929, 1934) and Langlois (1954) found N. flavus nests under large flat stones in June and July; Adams and Hankinson (1928), Bailey (1938), Evermann and Clark (1920) and Trautman (1948) found nests of N. gyrinus in tin cans, crayfish burrows and mud holes or under logs, boards, rooted vegetation and crockery in May, June and July; Fowler (1917) found nests of N. insignis beneath large flat stones in July; Taylor (1969) reported N. miurus nesting in tin cans in July and August and N. stigmosus in cans in July; and Trautman (1948) found N. miurus nesting under logs, boards and rocks in July. The use of cans for cover and nesting has also been observed in fishes of other genera (Breder and Rosen 1966, Kottecamp and Moyle 1972). In 1979, nesting of N. exilis at Hutchins and Green creeks OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 26 (91) e'E-S'S (9FTS) T'I-1'0> Gan OST L&G 99T'T e1s L6 VPI OGL €3 AP GLET 0G €'€-8'S (6¢°9) TI-T'0> (Or) 9°€T 096 I66'T T6I S6 PIBT GIL € AW 6LEI Or OCST (se‘9e) oI-1'0> ee) ve 19 0SG'0 oLI IL rd 8°19 a SOW «SL6T OL CSOT (17‘¢s) 6'0-1'0> na og OST 00S"0 T&% 96 9ST'V 6'F9 oo «Edy 6gL6T Il : OCLT (OF GF) 6'0-1'0> ar) Oh 8L 1d€'0 0°0G G8 0) i 4s) to dy SL6T Jit OST (€F'T9) 6'0-T'0> on) ve 6L OSE"0 PES FOT CrP yp 0'S9 co Edy SL6T er SCLT (s9°SL) Tio as VV cor 009°0 €'S6 LET 9L8°S 6 TL Fe = [Edy SL6T II SCS (€S°82) O'I-1'0> (on) v6 6LT Tes T6I I€T 698'9 LL re =NdY «S16 Or 839s (FL'9L) e1-T'0> LG 9°6 eLT 8eP'l U'8T OST 60€"8 os oF [Edy 9l6r (N) (Sur) umao (aus ) sasayjuaivg Uy (was ) (wu1) BAG esuvy pue uvayw einjeyy yorg SOLIBAO WSIIA\ Apog) ArvA_G Jog 1equinyny pYysIIM ( uur) suysey jo *(ulul) BAO JO WBIOM JO paisnipy Weigieag "BAGO ane JO Apog qwueT sqQUoy OZIg Uy osURY ane JO 91S osyeunxoiddy Isp WSO ssdq JO Iequny requiny [e}JO], pajsnipy Pprepuejg upesy yoy seax ‘pepnpur ow GLE pue SLGT YIOG AOy Burumeds 0} 9ynquy “U0d P[NOA jeY} sojvuloF oINyeUE pounsaid ATUQ ‘YoeID UdeIdg) WOIZ sya SsNinjoxy JO SaleWay BSurpseiq JO soNsuajVeYyD “fp IavV], 27 LIFE HISTORY OF THE SLENDER MADTOM 0'T-1'0> O'1-1'0> 6'0-1'0> 6'0-1'0> 6'0-1'0> 6'0-1'0> O'1-1'0> 6'0-T'0> O'1-T'0> (OT) 85-0 SS (ST) PEGS 1s (OT) 93-03 VS (OT) CaaS VS (OT) 6S-3S Cig (IT) 0'8-S'3 LS (OT) Te-L's 6% (OL) TE-9'3 6'3 (8ST) €'E-8'S o's wh 9'9T 6 GL 9'0T TL 9°€T S61 OFT 66 €1G e61 ert 006 O9T LVG 60€ G&G v61'0 €90°T PLY'0 T8¢°0 L6r'0 LEO 696'0 CIoT OSG'T VOT SOI £06 661 6ST GGG o8T TVG 6 LT ‘y4S1eM Apoq paysn[pe/OOOT X SeLteAo jo YSIOM sjenby z ‘IOAT, pue skoupry ‘ourjsozur “yorutojs ‘sotieAo Jo [BAOUIOI JO}Je WYSIOM suouTIOsds 94} st yystem Apog paysn{py 7 (OT‘9T) 9% (3E'se) 69 (61‘TE) os (02‘E) 6g (61‘83) LY (81‘83) oF (96°SP) IL 001s ST0'S 6SF'S L99°6 €87'S €F0°S 168°8 9€6'E 6S6'7 GCS OSL 8'PS VSS Ava L°6S 6'€9 qg9 GOL el GG a GL GL ae €% &G €G Ajn{ 4inf eun{ oun{ oun{ ounf ARI ARI ACN 6L6I 6L6T 6L6T 6L61 6L61 8L6T 6L6I 6L61 6L6T 28 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY F=3.49+18.73W, r=0.915 90 80 © 1 YEAR OLD @ 2 YEARS OLD 70 A 3 YEARS OLD e A 4 YEARS OLD NUMBER OF MATURE OVA 0 1 2 4 5 6 7 8 9 10 BODY WEIGHT (gms) Fic. 12.—Relationship between adjusted body weight of female and num- ber of mature oocytes produced. Data from Table 4. spanned at least 17 June to 17 July (Table.5) when 16 nests contain- ing clutches of eggs and guardian parent(s) were found. Water tem- peratures during this period ranged from 23.5° to 30°C. No nests or reproductively active adults, commonly found under rocks in pools and raceways, were found in April. In May, and before 17 June, adults had begun developing secondary sexual characteristics and were found in shallow pools with current and in crests of rifles. By late May and in early June, all large females were fully LIFE HISTORY OF THE SLENDER MADTOM 29 distended with eggs and large males had swollen epaxial muscles typical of nest-guarding males (Fig. 1). When males were located beneath overlying rocks in early June, a small cavity was present. Upon discovery, these males were more hesitant to move away than were males found in April and May. Females, however, showed no signs of having cleared the area beneath cover and readily left when rocks were lifted. After 17 July, no nests were found. Rocks at Green and Hutchins creeks used for nesting in June and July still had excavated cavities beneath them but were mostly occupied by Orconectes virilis or Campostoma anomalum. Nesting habitat—The nesting habitat of Noturus ranges from clean gravel riffles to silt-laden pools and two or more species may occupy the same nesting habitat (Trautman 1948, Taylor 1969, Menzel and Raney 1973). Nesting habitats of N. exilis included shallow pools and shallow crests and bases of riffles. Most nests (N = 12) were in pools with little current (Table 5) and at depths ranging from 12.5 to 40.0 cm (x = 19.2). Those at either bases or crests of rifles (N = 4) were in relatively shallow water (depth = 10.0-20.0 cm, x = 13.7). Nests ranged from a few cm from the bank to the center of the stream and, with the exception of one nest, all were in cavities beneath large rocks. Nest 10 (nest num- bers refer to corresponding numbers in Table 5) was beneath a large, flat sheet of tin, 0.6 m? (Table 5). Nest rocks were always the largest ones (15-30 cm long) in the area and were usually flat on the underside. Nest cavities ranged in depth from 3.8 to 15.0 cm, in length from 15.0 to 60.0 cm and in width from 10.0 to 60.0 cm (Table 5). They were usually free of silt and debris and were lined with fine gravel (Fig. 13). The number of entrances into a nest was not noted, but most nesting rocks appeared flush with the substratum. There was no significant relationship between the size of the nest cavity and the number of eggs present (r = 0.456). Nest construction.—Although evidence was not presented, Men- zel and Raney (1973) stated that, in Noturus, nest preparation is minimal. However, this statement should probably be reconsidered based on our observations of nesting of Noturus albater (Mayden et al. 1980), N. miurus, N. nocturnus and N. exilis. Nest construction in N. exilis apparently involves only adult males. Males observed in early June were usually found singly in excavated cavities be- neath rocks; there were no signs of excavation before June. A male forced to leave the rock, soon returned. Females found singly in June were never found in cleared cavities and always readily left cover when exposed. In late June, three reproductively active pairs were found in cavities beneath rocks; no eggs or hatchlings were found. In two cases, the same male had been observed in the same cavity in early June, indicating he had prepared the nest site alone. OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 30 " ~ OL t S06. O'ST‘OST?S'2 = ee [90d 86 Ant LT 9T 0090:4[9f LT GI OF v's OST O'S6:0'SS'o"L d O'OL""? 5°? [ood 86 4[nf LT ST . ord 0090:4[nf LT GL LG vg GOL SOligs'tsg"h G G99: 2 JO 3S981D 8G Aint LT PI OsEesieun[ cz g9 gg ee OST OO0T‘0'03‘E'9 iy 0s [90d Ses eunf 9% el : omy 00soO:eun! EZ IT 09 t 006 = G66'S'3s'SsI < ees oe jo oseg “ ,aunf ¢Z% a 00FO:eunl EF GI 69 t OST OOT‘O'SS‘SL g O'00T: 2 [90d “ geunf IT 0033:0unf TT 006 OL t OSGI 0°09:0°09‘0°ST a O'00T: 2 [90d LG eunf 6T Or a e; ve Ove Gel) S2aLos a T'L9: 2 [90d LG eunf 6T 6 00Es:ounf[ FT PIT 0S 0'e GLI O'ST‘00€'€'9 3 9°19: 2 [9d LG eun{ 6T 8 om O06T:9unL BT re 8 { O'OT ZT:0'0Z'¢E'9 @ FLO S°P = josey 2G (GT. 00TS:9eun, ET OFT (os 9°€ 6S = SIl:OSt'sL v9 O19: 2 [99d L@ eunf 6T 9 d09T:eunl ET 86 8€ be DLT O'ST‘S'to?S"L € 0°68: 2 [90d 6G eunf LT Ss OO8T:eunl ZT OZT gg 9°€ GLI G6s'S'3s's'¢ € 0'F6' 2 [90d 6G eunf LT FP OggT:ounl ET 86 Os ee OST O'0T‘0'03'E'9 3 O'8L' ? [90d 6G eunf LT € oma 00g0:eunf ST 19 Tg ee QoL O'ST‘'S'LG‘e'9 3 OSL? 2 Jo Pan). 6s eunf LT G O¢gT:eunf ET 86 PL t OOF SS LT:S°LG'9'S € 0°06: 2 [90d 6G eunf LT I Surumeds jo (14) Jueuldo (ur) JSON (wu ) Ainscraeeke (s1X) uet = (s)UeIprens mae oully, pue a}eq -[9A0q] ul os ezIs ysolvou WOIZ YIdoq JO YIPIM pue -pieny aq} JO (uur) -1eduiay, jequiny ejyeurxoiddy eyeunxoiddy yo n[D aouRysIq TOJVAA yysueyT ‘yideq jooesy YyysueT pue xes ywUQqeHR DIAL aed SON ‘arqu[reavun vyep sozeo “Ipur ** ‘juods sea o[eulay seyeorpur f ‘fpozefost o10M sjsou SoyOIPUT } ‘YIID UII) WOIZ SIO S}saU sa}yworpur , Uy ‘suauTTOads poivar-A1oye1oqey uo paseq av o8B JO soyeUISyY ‘syaoo suryoyN;T pus UseIg) WIZ S}sou syIxa SNunjoNy JO SONSUa—RILYD “CC AIav], LIFE HISTORY OF THE SLENDER MADTOM 31 Fic. 13.—Nesting cavities, guardian males and embryos of Noturus exilis. Above, 67.1 mm SL male from nest 9 (Table 5) in Hutchins Creek guarding a clutch of 34 embryos. Below, 100.0 mm SL male from nest 11 (Table 5) in Green Creek guarding a clutch of 62 embryos. Large nest-covering rocks removed. 32 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY The behavior of males and females and the presence of cleared cavities beneath rocks, later used for nesting, suggests that males are territorial during the breeding season and are solely involved in nest construction. Eggs and parental care—Eggs were always in the deepest part of the nest cavity and always in a single cluster (Fig. 13). Clutches were similar in appearance but varied in number and size of eggs. Most clutches were a roundish mass with each egg attached to one to five adjacent eggs (Fig. 13). Because the clutches were not anchored and were always free of debris, eggs probably had tem- porary adhesive properties upon oviposition that allowed them to adhere to each other but to no other surfaces. The spherical, amber-colored eggs ranged from 27 to 74 (x = 51.0, SD = 15.47, N = 16) per clutch (Table 5). Eggs collected from nest 14 (Table 5) were unusual in that yolks were significantly (« = 0.05) smaller and lighter in color than others. Chorion diam- eters for all clutches but the unusual one in nest 14 ranged from 3. ) to 4.5 (x = 4.11, SD = 0.151, N = 30) and yolk diameters from 3.2 to 3.6 mm (x = 3.39, SD = 0.124, N = 30). Eggs from the un- usual clutch had similar chorion diameters (3.9-4.2; ¥ = 4.06, SD == 0.114, N = 5) to other clutches but significantly smaller yolk (2.6- 2.8 mm; x = 2.66, SD = 0.089, N = 5) diameters. Taylor (1969) suggested that, since species of Noturus are mainly nocturnal, spawning would occur in relative darkness. By back calculating the age of each clutch from its time of collection, time of fertilization was estimated. Estimated ages of clutches ranged from 11 to 200 hr (Table 5). All estimated times of fertili- zation were either in early morning (0400-0600 hr) or afternoon to night (1600-2330 hr). Therefore, at least some spawning occurred during darkness. Female N. exilis may have helped care for young, but the degree of investment, after spawning, was probably minimal. In most cases (75%) when a nest was found, only an adult male was ob- served with the nest (Table 5 and Fig. 13). Of the 16 nests discovered, one was unattended, 12 were attended by a male alone and three were attended by both a male and a female. The rela- tively early stage of development of eggs in nests attended by both parents (Table 5) as compared to those nests attended by a male only, indicated that the female remains with the nest for only a short period (12-22 hr) after spawning. However, the role of the female during the short period in the nest and the reason for her early departure from the nest are unknown; multiple matings by females may occur. The period of parental care in N. exilis probably lasts until the yolk sacs are depleted and young are capable swimmers. One male N. exilis (nest 10) was found brooding his clutch of hatchlings and LIFE HISTORY OF THE SLENDER MADTOM 33 embryos. Fowler (1917) found males of N. insignis brooding either embryos or hatchlings and stated that parental care must exist after hatching. Mayden et al. (1980) found guardian males of N. albater brooding hatchlings at least two days after hatching. Menzel and Raney (1973) considered parental care in Noturus to last only up to hatching, but provided no supporting evidence. Characteristics of nesting males—The age of guardian males ranged from 2 to 5 years (Table 5); most (61.5%) were 2 years old. Standard lengths of guardian males ranged from 61.0 to 100.0 mm (x = 80.72, SD = 14.39). The two females taken from nests with eggs were both 2 years old and 67.4 and 70.0 mm SL, respectively (Table 5). Nesting data (Table 5) indicated that the number of eggs a male is capable of acquiring was significantly (r = 0.767, « = 0.05) related to the size of the male. No data are available on multiple matings by a single male. The relationship between age and num- ber of eggs (r = 0.671) was significant (2 = 0.05). Relationships between size or age of a male and size of the nest cavity were not significant. Water depth at a nesting site (Table 5) was not signif- icantly correlated with size or age of a male, clutch size or size of the nest cavity. Nest-guarding adults of N. exilis did not seem to feed while guarding nests. With the exception of one male, stomachs and in- testines of all guardian adults were empty. The one exception from nest 9 had one N. exilis egg and several ctenoid (Etheostoma sp.) scales in his stomach; the intestine was empty. Guardian males of N. albater also did not feed while guarding nests (Mayden et al. 1980). Behavior of guardian adults—Most males displayed a_ similar behavior to nest exposure; they swam around the nest cavity and over the eggs, occasionally exiting the nest for a few seconds. The distance traveled away from the nest by those males was usually only a few cm. Upon returning to the nest site, the male usually pushed the clutch around inside the nest, as if to roll the eggs over. Occasionally a male took the clutch partially into his mouth and shook the egg mass from side to side. This behavior has also been observed in the genus Ictalurus (Breder and Rosen 1966). In two cases the males guarding nests 8 and 15 (Table 5) ejected their clutches from nest cavities after consuming some of the eggs. In- cubation of the remaining eggs in the laboratory revealed that all eggs ejected had been fertilized and were developing. The few observations made indicated that the reaction of ex- posed, nest-guarding males to other species of fishes and crayfishes was minimal. On two occasions, after a nest had been located, individuals of Campostoma anomalum or Orconectes virilis were accidently scared into the nest cavity containing a male N. exilis 34 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY and egg clutch. The guardian males did not respond to the in- truder(s) and no egg predation was observed. However, the lack of response by guardian males to the intruders was probably ab- normal, as a result of nest exposure, since the intruding species have been observed preying on N. exilis eggs (pers. obs.). The distance between nest sites (Table 5) and the presence of only one male per nest may have resulted from territoriality of male N. exilis during the breeding season. Most males constructed nests from three to five m away from any other conspecific nest. The distance between nesting sites, however, may have been determined by factors other than territoriality, such as the availability or dis- tribution of suitable nesting material and the degree of intra- and interspecific competition for such material. Possibility of polyandry.—Although no detailed studies have dealt with the reproductive behavior of species of Noturus, it seems likely that they practice some degree of polyandry. Indirect evi- dence supporting this conclusion was presented by Menzel and Raney (1973) when they summarized the literature on number of mature oocytes produced by females of N. flavus, N. gyrinus and N. miurus and the clutch sizes of each species found in nests. They found that, in all three species, the number of eggs produced by a female was about twice the number found in a nest. We have made the same observation with N. exilis. Depending on age, mature N. exilis females produced from 26 to 150 mature oocytes (Table 4). In females 2 years or older, the number of mature oocytes ranged from 64 to 150 (Table 4). The number of eggs found in nests in Hutchins and Green creeks, however, ranged from 27 to 74 (Table 5). Female N. exilis may lay at least two clutches of eggs per year; each could be inseminated by a different male. Only two 2-year-old females were found in nest cavities con- taining eggs (nests 7 and 15, Table 5); one. was spent (nest 7) and the other was not examined. The number of eggs in nest 7 was 48 and the developmental stage of the eggs indicated that spawning had occurred within the previous 24 hr. From regression equations for body weight, length and age on the number of mature oocytes (Table 4), the number of mature oocytes this female would be expected to produce would be 84 to 92, about twice the number found in the nest. One 2-year-old female (73.0 mm SL.) layed two clutches of eggs with a single male in an aquarium. The first clutch contained 32 eggs, the second 38. The second clutch was not spawned until three hours after the first clutch had been removed from the tank. These observations on Noturus and those of Breder (1935), Breder and Rosen (1966) and Clemens and Sneed (1957) on the closely related genus Ictalurus indicate that Noturus and Ictalurus ‘ LIFE HISTORY OF THE SLENDER MADTOM 35 may differ in their reproductive strategies. Species of Ictalurus spawn a single clutch of eggs in one nest. Interspecific competition for nesting sites—-Throughout the breeding season of N. exilis, other species of fishes and crayfishes in Hutchins and Green creeks were frequently observed using cavi- ties beneath rocks similar to nesting rocks of N. exilis. Between May and August, cavities beneath rocks were occupied by O. virilis and breeding and non-breeding Pimephales notatus, C. anomalum and E. flabellare. In all cases when individuals of the above species were found under an object, no adult madtom was found. Nests of P. notatus were found from 6 May to 25 June, nests of E. flabellare from 1 May to 29 May. Thus, it appears that N. exilis competes with at least these species for cover and nest sites. DEVELOPMENT Fertilized eggs were similar in appearance to the amber, mature oocytes, except for their large perivitelline space. Their diameters ranged from 3.9 to 4.5 mm (x = 4.11, N = 30); yolk diameters ranged from 3.2 to 3.6 mm (x = 3.39). Clutches of fertilized eggs removed from nests and incubated in culture dishes at 25°C hatched in 187-210 hr (8-9 days). Hatching success in the laboratory varied from about 39 to 75 percent (x = 60%) for Hutchins Creek-spawned clutches. . Within a clutch, embryos usually exhibited synchrony in their development; however, some slight variations occurred. Illustra- tions representing stages of development are based on the modal form at that stage. Following is a chronological account of develop- ment of N. exilis from approximately 10 hr post-fertilization to six days posthatching. Early stages of cleavage are not included. Age 10 hr (Fig. 14A). Early high blastula stage. Blastodisc of approximately 100 cells covers about one-eighth of yolk; chorion clear, tough and rough-textured. Oil droplets not observed. Age 13 hr. Late high blastula stage. Number of cells present could not be estimated; no apparent increase in size of blastodisc. Age 24 hr (Fig. 14B). Early flat blastula stage. Diameter of blastula about twice that of 10-hr stage; large periblast cells present at margin of blastodisc. Blastocoel separating epiblast and periblast has formed. Embryos rotate slowly inside chorion. Age 29 hr. Late flat blastula stage. Further spreading of blasto- disc over yolk equal on all sides. Age 36 hr (Figs. 14C and 14D). Early gastrula stage. Cells around margin of blastoderm have thickened, forming germ ring; blastoderm covers about one-half of yolk. Embryonic shield pres- ent, rudimentary, slightly wider and more elevated than germ ring when viewed from side. Small lateral ridges and folds developing along axis of future embryo. 36 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY °° 3 9.000 %,%,° 0 Fic. 14.—Early stages of development of Noturus exilis. A) Early high blastula stage, age 10 hr. B) Flat blastula stage, age 24 hr. C and D) Side and top view, early gastrula stage, age 36 hr. E and F ) Side and top view, late gastrula stage, age 41 hr. G and H) Side and top view, 50 hr embryo, 2.4mm TL. LIFE HISTORY OF THE SLENDER MADTOM 37 Age 41 hr (Figs. 14E and 14F). Late gastrula stage. Further migration of germ ring over surface of yolk. Edge of blastoderm constricts yolk, causing bulge in uncovered portion. Embryonic shield and axis of future embryo well-developed. Axis of future embryo consists of medial groove and 2 lateral neural folds. Prosencephalon and deuteroencephalon apparent when viewed from side. Age 50 hr (Figs. 14G and 14H). Tail bud stage. Closure of blastoderm over yolk almost complete, leaving small blastopore at posterior end of embryo. Optic vesicles present, rudimentary; prosencephalon, mesencephalon and rhombencephalon present but ventricles not yet apparent; faint notochord visible. Caudal end of embryo has lengthened; embryo extends over one-half of yolk; 10-11 somites and small tail bud present. Ectoderm anterior to head separated from yolk, forming subcephalic cavity. Age 54 hr. Blastopore closed in most embryos. Lengths of those with closed blastopore are 2.8-2.9 mm (N = 5); those with some visible yolk are 2.4-2.5 mm (N = 8) long. Divisions of brain more defined, but no vesicles apparent. Unlike earlier stages, embryos maintain dorsal position on yolk; no rotation occurs. Spinal column truncates just anterior to optic vesicles and extends posterior to tail bud. Number of somites 13. Age 61 hr. Blastoderm coverage complete in nearly all embryos. Rudimentary otic vesicles forming just posterior to hindbrain. Num- ber of somites 19 to 21. Below and anterior to head, yolk is notice- ably absorbed, creating larger subcephalic cavity. Tail slightly compressed laterally, forming fin rudiment, but not free from yolk. Vertebral column terminates just anterior to flattened portion of tail. Five embryos average 3.7 mm TL. Age 66 hr (Figs. 15A and 15B). First contractions of tail and some anterior somites occasionally causing body to bend. Tail in most embryos free from yolk and slightly decurved, in some at- tached at extreme posterior end. Eye lens forming; rhombencepha- lon divided into three neuromeres. Circular otic capsule prominent. Somites extend farther dorsally and ventrally and are divided medially by a horizontal septum, forming the epaxial and hypaxial musculature. Caudal fin bud flattened and rounded. Spinal column partially segmented anteriorly. Head and trunk posterior to first somite elevated completely above yolk. Small, weakly pulsating heart in pericardial cavity beneath and anterior to otic capsule; no blood flow discernible. Age 73-79 hr (Figs. 15C and 15D). Tail longer and further separated from yolk, flexing regularly from side to side. Head flat- tened and connection to yolk restricted to small region posterior to otic vesicle. Urinary bladder forming on ventral side anterior to caudal fin. One median fin rudiment present, extending onto dor- 38 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY sum, around tail and on venter to urinary bladder. Fourth ventricle of brain differentiating. Eye lens formed completely but no retinal pigment present. Segmented vertebral column and rudiments of vertebral spines in some individuals. Fic. 15.—Early development and organogeny of Noturus exilis between ages 66 and 102 hr. Chorion removed. A and B) Side and top view, 66-hr embryo, 3.5 mm TL. C and D) Side and top view, 78-hr embryo, 4.3 mm TL. E and F) Side and top view, 90-hr embryo, 4.7 mm TL. G and H) Side and top view, 103-hr embryo, 5.1 mm TL. LIFE HISTORY OF THE SLENDER MADTOM 39 Age 84-90 hr (Figs. 15E and 15F). Small barbels present on anteroventral part of head. Small slit present in region of future mouth; branchial region visible, slightly enlarged, but no visible breathing occurs; opercula absent. Small, slightly elevated, light- red blood islands on yolk in some individuals just posterior to pos- terior connection point of embryo and yolk; no blood flow discern- ible. Pectoral fin buds visible above surface of yolk in tissue con- densations lateral to embryo. Embryos average 4.7 mm TL (N = 5), with most growth apparent in tail; tail has lost rounded shape, is slightly pointed and recurved. Age 97 hr. Pulsating sinus venosus apparent in subcephalic cavity beneath and slightly anterior to head. Clear liquid slowly travels through heart; light-red blood islands on yolk further elab- orated and fine strands of light-red blood visible. Two distinct barbels present on either side of head. Tail flips from side to side and occasionally, upon flexion, strikes branchial region and hind- brain. Age 102-104 hr (Figs. 15G and 15H). Few faint blood vessels present over surface of yolk; vitelline arteries extend perpendicularly from body just anterior to pectoral fin buds. Three barbels on each side of embryo apparent in lateral view; between first two barbels are rudiments of external nares. Opercula and primordial gills pres- ent. In region of bladder primordial urinary tract appears to leave body as tube. . Age 109-115 hr (Figs. 16A and 16B). Retinal pigment in eye; no other pigment on body or yolk. Blood flows in head, around lobes of mid- and hindbrain and in gill arches. On surface of yolk, circulation increased with several large blood islands and smaller vesicles coalescing; blood pumped into branchial region from two anterior vitelline veins by long sinus venosus, through body and out onto yolk by vitelline arteries. Mouth well-developed but no breath- ing discernible. Single, continuous median fin constricted on dor- sum near tip of tail. Posteroventral projection of tissue at tip of tail has spread into clear fin-forming caudal ray primordia. Age 120-128 hr. Entire eye heavily invested with retinal pig- ment, appearing dusky. Blood darker red with some flow in trunk region. Right vitelline artery bifurcated at point of junction with body; left vitelline artery single. Posterior to embryo, vessels on yolk appear discontinuous with other main vessels. Ten embryos are 5.8-6.0 mm TL. Age 133-138 hr (Figs. 16C and 16D). Beginning of sporadic jerks of trunk and increased tail flexion; embryos beat tip of tail against one side of head two-three successive times, then shift to beating opposite side of head a similar number of times. Pectoral fins close to body, but not attached, and with no indication of circulation. Condensed opaque tissue covers clear caudal fin where 40 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY E F Fic. 16.—Early development and organogeny of Noturus exilis between ages 113 and 144 hr. Chorion removed. A and B) Side and top view, 113-hr embryo, 5.3 mm TL. C and D) Side and top view, 134-hr embryo, 6.2 mm TL. E and F) Side and top view, 144-hr embryo, 6.4 mm TL. rays will develop. Constrictions in single median fin better de- veloped between primordial adipose and caudal fins and at urogen- ital region. Urogenital region with additional anal tube. Age 141-146 hr (Figs. 16— and 16F). Breathing occurs by end of stage; opercula moves at relatively slow, but constant, rate. Blood darker on yolk, lighter in embryos. Constriction of dorsal fin LIFE HISTORY OF THE SLENDER MADTOM Al results in formation of dorsal and adipose fins. Terminal, recurved portion of vertebral column degenerating. Sporadic jerks of em- bryos more frequent. Age 150-156 hr. Blood flow in embryo more pronounced; two main vessels extend posteriorly from vitelline arteries. Blood flow present in pectoral fins; all other fins apparently devoid of circu- lation. Embryos exhibit occasional violent spasms lasting 30-50 sec. Age 160-164 hr. Minimal melanophore differentiation on body and head posterior to eyes, and lateral to posterior part of brain. Pectoral fins occasionally flex. Age 168-175 hr (Figs. 17A and 17B). Further development of pigment on head and nape, concentrated immediately anterior and posterior to eyes in small clumps; some melanophores scattered over dorsum of head, some extend to first few myomeres of nape. Violent spasms subside in most individuals; tail flexures infrequent. Em- bryos slowly move around in chorion. Further opercular develop- ment; gills no longer exposed. Pelvic fins present, but no ray development; rays developing in caudal and anal fin regions. Age 180-187 hr. At 187 hr, three individuals from one clutch hatch. During hatching process, tail breaks through chorion first; with increased flexion of tail, embryos expand opening and later exit. Circulation visible in dorsal and adipose fins as single wave- like vessel. Age 194-197 hr (Figs. 17C and 17D). About 50 percent of embryos have hatched or in process of hatching. Hatchlings active with well-developed barbels and fins. Schooling behavior well- developed; swimming clumsy. Increased activity of embryos occurs under bright light and when dish is struck lightly. While swimming, hatchlings maintain dorsal position over yolk, but when immobile, usually roll to one side. Muscle flexures originate in thoracic region and move posterior on body to caudal fin, causing embryo to move forward. Melanophores further developed on dorsum of back; uro- genital duct and anus present; further development of rays in anal and caudal fins. Nares tube-like and extend from head. At bases of conjoined pectoral fins, small pits begin to form rudimentary axillary pores (Reed 1907, 1924; Birkhead 1967, 1972). Six hatchlings are 7.5-9.0 mm TL (x = 8.13). All embryos hatched by 210 hr. Age 1 day old (Figs. 18A and 18B). Melanophores concentrated over head and nape and lightly scattered over trunk and yolk sac adjacent to embryo. Rays differentiating in pectoral fins; pelvic fins extend away from body; branchiostegal rays developing. Many hatchlings resting in shade beneath air stones and are positively geotropic. Three specimens are 9.4-13.2 mm TL (x = 10.7). Age 3 days old (Figs. 18C and 18D). Beginning of ray develop- ment in all fins; pectoral and dorsal fin spines differentiating. Lat- eralis system consists of open canals and pits, on head only. Pig- OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 42 , "TL wur gg “earl (14-261) peyoyey Apuadcar ‘mora do} pue opig (q pue D “TL wu T['g “Buryoyey o10f7aq Aep duo oAIquIA 1Y-ggT ‘MoIA do} pue apg (g pur y ‘“SuTyo}ey -a1d Ul paAoulal UOLIOYD ‘BAIR, plo-Aep-T pue Buryo}eyoid plo-1y-gg{ usemjoq syixa snunjoxN Jo yuowdojaaeq—)T ‘o1q 43 LIFE HISTORY OF THE SLENDER MADTOM do} pur opis (q pue Q “Ty mu Fg ‘earey Pie ia} “TL Wu Q'OT ‘earey P[O-Aep-¢ ‘mara -Aep-T ‘MaIA do} pue apis (q pure vy ‘SYIXa SNAMJON JO avAIe]T— 2) ee “8ST “OT 44 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY mentation more complete over head, trunk and yolk sac adjacent to embryo; most pigment concentrated over dorsum, venter white. In eight specimens, TL ranges from 8.7 to 12.0 mm (x = 10.52). Age 6 days old (Figs. 19A and 19B). With exception of a few characters, six-day-old larvae closely resemble adults. Nearly all fin rays and spines completely developed and constrictions sepa- rating fins nearly complete. Anal and caudal fins remain slightly connected; fin rays unsegmented. Melanophores basally on some fin rays and membranes of caudal, dorsal and adipose fins. Melano- phore distribution much more concentrated over body. Pores of cephalic lateralis system connected by closed canals; few small, open, isolated canals remain. Body robust, head flattened. In six specimens, TL ranges from 11.0 to 12.6 mm (x = 12.05). After six days the morphology and behavior of larvae are essen- tially like that of adults. However, the typical black distal band on the dorsal, anal and caudal fins in adults is still not observed in embryos up to the thirty-second day. At 17 days, fry appear to begin feeding on crushed catfish chow. Teratology Of the 622 age-class-0 N. exilis collected from Green Creek, six were missing one or both pelvic fins. In all other respects these six > | a Fic. 19.—Six-day-old larva of Noturus exilis, 12.2 mm TL. A) Side view. B) Top view. LIFE HISTORY OF THE SLENDER MADTOM 45 appeared normal. In all species of Noturus, only one other case of possible teratology has been reported; Hocutt and Stauffer (1976) noted a single case of a forked nasal barbel in N. flavus. DIET Dipteran larvae and pupae, ephemeropteran naiads, trichopteran larvae and some crustaceans comprised over 95 percent of the total diet of 264 N. exilis (Tables 6-9). Among dipterans, chironomids (62.2%) and ceratopogonids (5.2%) were most frequently eaten. Ephemeroptera, including the genera Stenonema, Baetis and Caenis, were second in abundance, comprising 10.2 percent of the total diet. Trichopteran larvae, mainly of the genus Cheumatopsyche and to a lesser extent Chimarra, made up 5.3 percent of the diet. Among Crustacea, Asellus comprised 6.7 percent, Gammarus 1.3 percent and Copepoda (Calanoidea) 2.8 percent of the total diet. Infrequently ingested items included nematodes, annelids, other crustaceans, dipteran pupae, beetle (Elmidae) larvae, plecopteran naiads, fish eggs and small stones. Small stones were surely ingested accidentally during feeding. Males and females did not differ in types or quantities of food organisms consumed. In Oklahoma, the diet of N. exilis was similar to that of the Green Creek population; predominately immature aquatic insects and crustaceans were consumed (Curd 1960). However, ephem- eropterans were eaten more frequently in Oklahoma (88.2%) and comprised a slightly higher proportion of the diet than did chiro- nomids (81.2%). Number of prey taxa increased to 18 in early spring (April), after the low winter diversity (5 in February), then stabilized be- tween nine and 13 through summer (Tables 6 and 7). The greatest diversity corresponded with high growth rates in spring (Fig. 5) and preparation of adults for breeding (Figs. 6, 9, 10 and 11). The diet over spring and summer months consisted of prey of several taxa that were not present in other months (Tables 6 and 7). Fish eggs consumed in April and May were probably those of either Etheostoma spectabile or E. caeruleum, both of which spawned during that period; cyprinid eggs were consumed in July. Diversity of the diet usually increased with increasing body length (Tables 8 and 9). Dipteran larvae and pupae, ephemerop- terans, plecopterans, trichopterans and crustaceans were eaten by fish of all sizes. Predominance of these taxa in the diet varied, however, with age of the fish. Young fish fed mostly on small crustaceans and small dipteran larvae and pupae. Older fish ate mostly large crustaceans and large, immature aquatic insects, but still ate small dipterans. The percentage of the diet consisting of copepods and ceratopogonids decreased significantly (« = 0.05) OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 46 oe _ = = 0 CT 3 = a) 80 og 1 sepyndry, ae = O'T om as ze 40: c‘0 = = = epr[NUs eio}diq ares ee a ae g C0 Sus Fe apo? ee —— eg Tir og viajdopiday oe ae i role = ss ‘0 is a 7 8°0 7 = seplul gy via}doalop oe 61I OF oT lv ‘9 Pp 9°0 LG Lg Gt SET eia}doyou L, 18 LOL ‘9 36 Cll “el -¥'6 6% 6's 18 0) a viaydorsweydg 5 aa a r= FT a aE =a c'0 — OT O'OL via}dooa[g Bpesuy ae a = 0 — ---- — —= soa eS Toe = $3JIIUDIIO) epodesaq OI so G6 ST v0 sia VP ge alk <-3¥ a Le snyjasy epodos] a oT OT OLNh aes it BE a oF 8'P ‘* LT weet epodiyduy 8'P ay Cel s os 7 a: 60 £3 = 0% Lit Bpodado:) oT es; £0 cam =a sia 7s ovouRly epruyprry —: FA ps om a rom ay ms C0 as ay — vourplitfyy — =A pe a Z0 + a ek 6'0 F'0 = ar vyoRyoos IO epluuy se —— C0 ---- l'Z ce 90 6S C0 — ae = BpO Ula NY (PI) (03) (86) (98) (#3) (83) (08) (#3) (68) (€%) (91) (#8) uoxe], Poor 09q AON ~PO dag sny raf unf ARI idy Ivy qa uel pestdwuoy uoxvy, poo,y yory yey} JIG AjyUOWy Fo Juso10g ‘pouruexa syoRwo}s JO sioquinu are Sasoyjuoivd uy saindtJ *(peurquiod sived OA\}) UONda][Oo Jo YQUOUT Aq YaaIQ UaeEID WOIF syIX snunjoNy JO SJUaWOO YORUIOYS “g AIA], AT LIFE HISTORY OF THE SLENDER MADTOM TLE O'1G TL 9'€ Stl 9°6S OT 9°¢ <'9S v9 0G oT £6L 8S ST <0 VT a £09 OTL L'€9 0's VT ar 60 a. ZS s33q TU0}SO2Ta J, ea eedng peyguepiuy "Eg oeprutouolyy aagt eepruosodo}eie5 es aepLiogoryy OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 48 Qa So Ke ok. Sel RI’? G06) “en ee ae sepyndr, - == CP as ray con ne Se ye o ce a sepl[Nulls viaq}diq = oe --- se —-- = oa = eax ca = a vio}dopidaT — -- — == as CPT he = = pie aa tas oeprmyy via}doajon TL 0°0€ ost OVE oLé OLS 0'0F £8 L'0G V LI 8°89 eee viaydoypouy, Lge 007 STE 0°0S 8'OL e LL 0°09 8°06 9°L3 T6€ GLE 8°06 viaydoisuiay dg — a pe ae LOT ae a 8 VE mm 9 Lot v1ajdooa|q Basu] ahs aes cox se --- ~--- --- = oer: i = ss $3}I9U0IIC) rpodeosq VIL 0'0L 6'0P 8'0€ es OST 0'0€ Liv SP LG ~ GP Saar. epodosy] oe 0's T6 se aa OSI ue ae GLI O'eT i CPF STDUUDS) epodiydury eri ait L6G st LOT ae 7 £8 €Or ae ad | CF vpodadop eye = = a = CT = ee 69 ta 24 =e a BpOoe.QSO mal ee T6 ae oa i wa = ve 2 ps ese BIBIOPRD vaoRsnip ee S55 a ae — ---- ---- CP = — er Sa ovourly epruyorry = =e <2 — = —— ---- — re a —a = vauIpruiyy ee — -~ — CP --- -——- ES 69 — aa = PyoRYposyoO epleuuy (FT) (08) (83) (93) (63) (85) (03) (3) (63) (83) (91) (¥#) wore pooy 90q AON ~PoO das sny lyf unf{ ABI idy IR gay uef peltno9Q UOXe J, pooy YoryAA Ur SYOBULO}S FO JUBI1Og Ss ::2250—=OOmNao eee *peurluexed SYORUIO}S JO SIoquInu 918 sasoyquoivd url sornsty *(pourquioo sivad OA\}) UOKe[Oo Jo YUOUT Aq YaeID udaIg WoT Syixa SninjoN JO syUa}UOD YORUIOYS *“y ATGV], 49 LIFE HISTORY OF THE SLENDER MADTOM oP rs ss3q TULO}SONTO T, Oe SIS eae Gia Win. ayaa ot Ce a fa Sedad Peynuepiuy Se Mo SZC Gs SE) ey, Grde ai8 VCL S16 O'SL 010 ~—aeprmouony ene SOEs Sa = 8i0e Ski, a sce ence OSGi 08 Se ~~ Sepruosodoye.10,5 s as a - a 7 ea = Gor oar = = eepHogoryy 50 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY with increasing body size of madtom. A significant (« = 0.05) in- crease in consumption of Asellus and Gammarus occurred with in- creasing body size of madtom. The only change in diet noted by Curd (1960) with increasing size of fish was that larger madtoms fed on larger food items. Individuals collected from pools and rifles fed on similar types and quantities of food organisms (Table 10). However, hirudi- neans, ostracods, decapods and lepidopterans were eaten only by madtoms collected from pools and arachnids, coleopterans and simulids were eaten only by those collected from riffles. TABLE 8. Stomach contents of Noturus exilis from Green Creek by size class of madtom (two years combined). Figures in parentheses are numbers of stomachs examined. Percent of Diet that Each Food Taxon Comprised ‘ Food Taxon <25 25-40 41-55 56-70 >70 mm mm mm mm mm (8) (85) (83) (72) (16) Nematoda 2.9 0.8 1.4 1.5 ae Annelida Oligochaeta ae = 0.1 0.2 0.7 Hirudinea Zan oe a ae 0.7 Arachnida Araneae ren 0.2 fhe th ES Crustacea Cladocera cs 1.0 0.3 eae ah Ostracoda = 0.3 0.1 0.3 ae Copepoda 14.5 5.0 0.6 0.2 2.1 Amphipoda : Gammarus Se 0.3 0.6 — 30D ead Isopoda Asellus 0.6 2.1 5.9 13.9 8.5 Decapoda Orconectes a — ae 0.2 ee Insecta Plecoptera 0.6 1.0 0.6 0.3 0.7 Ephemeroptera 10.4 4.9 12.2 13.1 10.6 Trichoptera 5.0 5.6 5.0 5.7 4,2 Coleoptera Elmidae oer oe, 0.1 aoe ee Lepidoptera a, ae rte 0.2 an Diptera Simulidae wing ns ox 0.3 ae Tipulidae 1.2 ih 0.5 0.7 2.1 Chaoboridae Be roe 0.1 ner 0.7 Ceratopogonidae 11.0 6.0 4.7 3 2.8 Chironomidae 50.9 71.0 65.1 52.6 61.3 Unidentified Pupae 2.9 0.6 1.5 3.4 3.5 Teleostomi Eggs i 1.2 1.2 0.2 cae LIFE HISTORY OF THE SLENDER MADTOM 51 TaBLE 9. Stomach contents of Noturus exilis from Green Creek by size class of madtom (two years combined). Figures in parentheses are numbers of stomachs examined. Percent of Stomachs in Which Food Taxon Occurred Food Taxon <25 25-40 41-55 56-70 >70 mm mm mm mm mm (8) (85) (83) (72) (16) Nematoda 25.0 2.4 6.0 42, ein Annelida ; Oligochaeta ents fest 1.2 1.4 6.3 Hirudinea ais ae a a 6.3 Arachnida Araneae a 2.4 me a —_ Crustacea Cladocera = 5.9 1) To pak Ostracoda tse 2.4 1.2 2.8 Ses Copepoda 50.0 11.8 3.6 1.4 6.3 Amphipoda Gammarus es 1.2 48 16.7 6.3 Isopoda Asellus 12.5 15.3 36.1 45.8 43.8 Decapoda Orconectes ek a 2 1.4 pie Insecta Plecoptera 12.5 4.7 3.6 1.4 6.3 Ephemeroptera 87.5 25.9 39.8 56.9 56.3 Trichoptera 37.5 21.2 33.7 27.8 31.3 Coleoptera ; Elmidae ae —_ 112 anid OF Lepidoptera om ne = 14 a Diptera Simulidae Ei ey Shas 2.8 ae Tipulidae 12.5 Pes 4.8 5.6 18.8 Chaoboridae 23s a 12; —_ 6.3 Ceratopogonidae 50.0 ON. 18.1 18.1 18.8 Chironomidae 87.5 82.4 68.7 62.5 THO Unidentified Pupae SHAS 3.5 8.4 16.7 25.0 Teleostomi Eggs . = 1.2 2.4 1.4 — Feeding Periodicity Feeding activity, as measured by mean number of food items (MNFI) per stomach, increased markedly between dusk and dawn (Fig. 20), and was bimodal. Peak feeding activity occurred just prior to dawn (0600 hr) (Fig. 20), with a mean number of 84 food items per stomach and digestion index (DI) of 4.0. Feeding there- after decreased sharply to 1800 hr, when all five specimens exam- ined contained empty stomachs. The slight inflexion of MNFI at 1200 hr was probably due to sampling error. There was no indi- cation that fresh food had been ingested after dawn and before dusk; most prey in stomachs during this period were well digested. 52 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY TaBLE 10. Stomach contents of Noturus exilis from Green Creek by habitat (two years combined). Figures in parentheses are numbers of stomachs examined. 4 Food Taxon Percent Diet for Each Percent of Stomachs for Habitat that Each Each Habitat in Which Food Taxon Comprised Food Taxon Occurred Pools Riffles Pools Riffles (114) (104) (114) (104) Nematoda 0.8 2.1 Non a 8.7 Annelida : Oligochaeta 0.2 0.1 1.8 1.0 Hirudinea 0.1 cick 0.9 Sal Arachnida Araneae er 0.1 at 1.0 Crustacea Cladocera 0.3 0.5 2.6 2.9 Ostracoda 0.3 a 2.6 ae Copepoda 8) “3.0 8.8 CP ox Amphipoda Gammarus 1135 itil 6.1 6.7 Isopoda Asellus 7.9 6.0 36.0 35.6 Decapoda Orconectes 0.1 ss 0.9 eam Insecta Plecoptera 0.7 0.8 3.5 5.8 Ephemeroptera 7.8 11.9 39.5 42.3 Trichoptera 3.3 5.1 24.6 26.9 Coleoptera Elmidae Ties 0.1 =m 1.0 Lepidoptera 0.1 ae 0.9 _ Diptera ; Simulidae om 0.3 ales 1.9 Tipulidae 0.7 0.5 6.1 3.8 Chaoboridae 0.9 0.9 1.0 1.0 Ceratopogonidae 5.1 5.4 20.2 21.2 Chironomidae 63.0 60.1 70.2 69.2 Unidentified Pupae 3.2 0.9 16.7 6.7 Teleostomi Eggs 0.7 1.1 1.8 1.9 A second, smaller peak in feeding activity occurred shortly after dusk (2000 hr) (Fig. 20), with MNFI of 54 and DI of 3.3. There- after, and through the night until 0600 hr, feeding intensity de- creased; however, the dip in MNFI was exaggerated due to sam- pling error. From 2400 to 0400 hr, the largest specimen of N. exilis from each of the three hours sampled contained a fish in its stomach. In these madtoms, food items other than fishes were few. The in- gestion of large prey probably caused the decrease in mean feeding activity between 2400 and 0400 hr, because of the inability of these individuals to ingest additional food. The DI indicated that, though LIFE HISTORY OF THE SLENDER MADTOM 53 o ao = c 2 n . DIGESTION INDEX MEAN NUMBER FOOD ITEMS/STOMACH OR PERCENT OF INDIVIDUALS CONTAINING FOOD ITEMS uo ° 18002000 22002400 200 400 600 800 1000 1200 1400 1600 TIME OF DAY (hr) Fic. 20.—Feeding periodicity in Noturus exilis expressed as mean number of food items per stomach, percent of individuals containing food items and condition of stomach contents (digestion index). Solid line and solid circles represent mean number of food items per stomach; hatched line and triangles represent digestion index; dotted line and squares represent percent of indi- viduals containing food. Sunset and sunrise are indicated by solid vertical lines. feeding occurred throughout the night, it was most intense at 2000- 2200 hr and 0600 hr when most individuals contained fresh food. The percentage of individuals feeding over time was consistent with the hypothesis of a nocturnal feeding period (Fig. 20). The increase in activity at 1200 hr is attributed to sampling error; the 54 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY low DI at 1200 hr showed that food remaining in stomachs was well digested and that feeding must have occurred earlier. Bimodal feeding activity was noted by Darnell and Meierotto (1962) for Ictalurus melas. Both peaks occurred at night with a major peak just prior to dawn (0330 hr) and a minor peak shortly after dark (2130 hr). They also noted a decrease in activity in mid-night (2400-0200 hr) and no feeding during mid-day. Feeding activity of I. melas, like that of N. exilis, tapered off after an early morning feeding peak and little food was ingested between 1200 and 1600 hr. Dipteran and ephemeropteran larvae were the most frequently ingested food organisms, comprising over 80 percent of the diet both during most hours sampled and for the total sample (Table 11). Percentages of those taxa did, however, drop off to 60 and 70 percent at 1400 and 1600 hr, respectively. Of dipterans, chironomid larvae were eaten most frequently; ceratopogonid larvae were con-. sumed at most hours but only in small numbers. Other dipterans that were ingested but only irregularly and at low frequencies in- cluded larval simulids, tabanids and tipulids. Dipteran pupae were consumed throughout the night. A comparison of the frequency of occurrence of ephemerop- terans and chironomids over the 24-hr period (Table 11) indicated that mayflies reached a peak of abundance in stomachs of N. exilis in mid-night (0200 hr). Chironomids, however, exhibited a bimodal pattern and were more numerous than mayflies, with peaks corre- sponding to peaks of overall feeding activity (2000 and 0600 hr). Shortly after dusk and just prior to dawn, over 90 percent of the diet was chironomids. Mayflies were low in abundance then, but showed a marked increase at 0200 hr. These differences in abundance were probably caused by drift activity of the two prey types. Chironomids drift predominately just after dusk and just prior to dawn (Hynes 1970), which corre- sponds closely to their peaks in abundance in stomachs of N. exilis. Species of mayflies differ in times of drift but drift mainly in mid- night during fall months (Hynes 1970). Since N. exilis is primarily a nocturnal species and feeds principally by taste (as observed in aquaria) rather than by sight, the type of food ingested is probably highly influenced by its abundance in the habitat. Therefore, a peak in chironomids eaten at dusk and at dawn would be expected if a peak in drift occurred at those times. The peak in mayfly con- sumption at 0200 hr may be attributed to their drift, during mid- night, but may also be attributed to a decreased relative abundance of chironomids to mayflies. Among other food items found in stomachs of N. exilis, trichop- terans and plecopterans were next in abundance to chironomids and ephemeropterans. Coleopterans consisted mainly of adult and 55 _ LIFE HISTORY OF THE SLENDER MADTOM = = = bb = eer ei en he aa aor eg ore ee eer Cee ne © oa ee oT TO SYVDIQUIN * NT TUL0}SO9]9 J, my Pale SLi Lie OF 28:Ci- ROU eae mre LO er LIZ OT eedng peyquepiuy GLE TLy 818 TOL 8S G68 069 OPE OE9 Sb6L OPE ~~ OSL T9L oeprwmouollyy) ENCL OIG Gn ONL yl ClO SiGu ame ae a; ivy LIZ PI sepruosodojeiaD a fhe iOS) Os ame a LAO). eae 3 a ma a os T0 oepyndry, -s GiGi ae “o i a a as OS “Gs Faig a Lg TO oeprueqey ie ee ae ee eS ca ae a i= Sele a (StS T0 Ssepl[NUulls eioidiq sie ae ae a a ae OS Ge a Oi ts aR al e's T0 viaydolesayy a > She ads es eo an Sas » ky oagae Bae ae OT TQ 98ArIey poynuepuy Cigihm up me 2 ae Ek, SP Se On cette. Beek ee eee OST eT snuaydasg e19}do2[0D =a ig ii SiG LO O19 One OLS Col ae la oy L9G 6G eioydoyouy, 0SG €r 16 O96 L86 €8 OTL G8h LE SLI LE ~~ 0°89 OST eia}do1sweydy 5 os MGs 6 TS GO GShS ae — 9:0) L255 a: est OT v19}doo9[g eyasuy dee ee eS ne SE ret Cae ne oT T0 snyasy epodosy op i Ts ie = COs: ce as a vO == S's T0 epodedoy ee a Oe es ae ey ee Oia ie Gas alk tac on Seen a 9'T TO epooesO eaorsnin ie ian i (Q) ie aa ca ac: i oe ie rs OT To avourly epruyoery SiGe. SVG = os. a ee Be ee Ne eS Oe o's 0 BpoyeUlaN OE el ee oe ed at pee sak ear AIS) aM Qh ie perins09 uOXxe], SB S38 S38 S38 US aB as US SS SE 2B 2B“ poog yom _pasuduoy uoxey, . UI SYORUIO}S poo, yory 7eu} UOXe], poo pesuduioD uoxey, pooy yory yey Iq A[AMOFT FO JUIOIEg JO JUVIIOg JIC JO JUVOIOg SUOT}O9T[OD ATINOF, [P30.L ‘pouruexe syoeuro}s jo siaquinu oie sasayjueied ul soinsiq “Apnjs Sulpesy [oIp sulinp yooID surTyoNFT }e syixa sninjoxy JO syua}U0D YoeUIO}G *[][ AIAV, 56 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY larval Psephenus herricki. Although plecopterans comprised only a small percentage of the total diet, they were frequently eaten through the night. The two fishes eaten, Notropis umbratilis and Etheostoma flabellare, are predominately diurnal feeders and in- active at night. DEMOGRAPHY Density ; On five separate dates quantitative samples of N. exilis were taken from pools and riffes in Green Creek by repeatedly seining a measured area of stream until no more madtoms were collected. The number of specimens collected was transposed into the number per square meter (Table 12). Samples were not taken between April and July because of seasonal changes in positions of riffle and pool habitats and the difficulty in distinguishing specific habitats* during periods of high water. The density of N. exilis in riffles ranged from 0.6 individuals per m? in November to 12.9 per m? in December (Table 12) with peaks occurring in August and December (Table 12). Pools were sparsely populated during all seasons and were de- void of N. exilis in January and March. Pool densities ranged from 0.0 individuals per m? to 2.3 per m?. Noturus exilis occupied both pools and rifles in late summer and fall probably as a result of recruitment of young after the reproductive season and a decreased water level, causing riffle habitats to be too shallow and small. During some months of the year N. exilis were found in con- centrated groups. During fall and winter months seining for an extended period of time in some sections of the stream yielded only a few individuals. However, when seining efforts were concentrated on another section of the creek, many individuals were taken with a single seine haul. Aggregations of feeding madtoms were also observed during the diel feeding study. TaBLe 12. Number of Noturus exilis per square meter collected from pools and rifles in Green Creek. Pools Riffles Date Number Number per Number Number per Collected Square Collected Square Meter Meter 30 January 1978 0 0.00 15 2.69 18 March 1978 0 0.00 15 3.57 23 August 1978 45 2.30 oT 7.69 20 November 1978 22 1.89 vf 0.60 3 December 1977 4 0.50 18 12.92 Mean : DANTE) 5.49 _ LIFE HISTORY OF THE SLENDER MADTOM 57 Composition Of the 972 N. exilis collected over the two year period, 64.0 percent were up to | year of age, 33.3 percent were over 1 year and up to 2 years and 2.2 percent were over 2 years and up to 3 years of age. Individuals older than 3 years were few and com- prised only 0.5 percent of the population, with 0.4 percent of the total over 3 years but less than 4 years and only 0.1 percent over 4 years (Table 13). Sex ratios differed from the expected 1:1 for the two years com- bined (Table 13) and for 1977-1978. In the total sample (N = 972) females slightly outnumbered males 1.1 to 1.0 (x? = 4.48; P < 0.05). In 1977-1978 (N = 436) females outnumbered males 1.3 to 1.0 (x? = 7.19; P < 0.05). Monthly sex ratios did not differ signif- icantly (2 = 0.05) from the expected 1:1. A predominance of one sex in species of Noturus has been noted by several authors including Taylor (1969) who frequently found a skewed ratio for some species (e.g., N. eleutherus and N. stig- mosus). Thomerson (1966) found males of N. funebris to slightly outnumber females, especially in larger size classes. Clugston and Cooper (1960) reported a greater number of females than males in N. insignis and suggested that the difference was due to sampling bias. In collections of N. gyrinus from Lake Erie, males slightly outnumbered females in young specimens but significantly out- numbered them in older age classes (Mahon 1977). In collections of N. albater from southeastern Missouri, females outnumbered males 1.8-1.4 to 1 (Mayden et al. 1980). Survival Relative survival values (Table 14) for each year of life were calculated for males, females and total samples of N. exilis, using the data in Table 13. Calculations assumed that each age class was collected in proportion to its relative number in the population, that all individuals in one age class were subject to similar mortality rates, that the population was neither increasing nor decreasing and that the number of young-of-the-year entering the population each year was constant. Survival rates for males and females were similar (Table 14 and Fig. 21). Unlike the Type III survivorship curve suggested by TaBLeE 13. Distribution of sexes and year classes in samples of Noturus exilis from Green Creek between 6 November 1977 and 29 October 1979. Sex Number by Year Class Total 0 I+ 2+ 3+ 4+ Males 290 151 9 2 uf 453 Females 332 173 12 2 abe 519 Total 622 324 21 4 1 972 58 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY TABLE 14. Relative survival of year classes of Noturus exilis in Green Creek expressed as proportions of the 0 year class (1x1), the 1+ year class (1x2), the 2+ year class (1x*) and the 3+ year class (1x‘). Sample Year Number Sinvival Class of — = 3 Specimens 1x1 1x? 1x lx Males 0 290 1.000 eee eek ies 1+ 151 0.521 1.000 Pee pened 24 9 0.031 0.060 1.000 eA. id 3+ 2 0.007 0.013 0.222 1.000 44 i 0.003 0.007 0.111 0.500 Females 0 332 1.000 ene ish ease aes 1+ 173 0.521 1.000 awe» pes 22 24 12 0.036 0.069 1.000 eae t 3+ 2 0.006 0.012 0.167 1.000 44 OF pete is - eee he ee Total 0 622 TOO0On .. =2-5 ate’ oe Sample J+ 324 0.521 1.000 HAY, ees 2+ 21 0.034 0.065 1.000 pate) 3+ 4 0.006 0.012 0.190 1.000 44 1 0.002 0.003 0.048 0.250 Deevey (1947) as usual for fishes, survival rates for N. exilis were intermediate between Types II and III (Deevey 1947) and inter- mediate between Types III and V as outlined by Ricklefs (1973). That is, there was about 50 percent mortality early in life, lower than expected, and a relatively constant decreasing mortality with increasing age. Both sexes experienced a lower survival in their second year of life than in their first. The low survival rate at 2 years of age may, in part, be related to the onset of maturity. After 2 years, females had a slightly lower survivorship than males. This may be related to the increased stress of egg production at older ages. Females experienced their lowest survival in their fourth year of life. Males lived longer and experienced their lowest survival in their fifth year of life. Longevity The maximum life span for N. exilis in Green Creek appeared to be 5 years (Table 13). The single 59-month-old male collected May, 1979, was 100.0 mm SL. Few individuals of either sex lived longer than 2 years (Tables 13 and 14). Longevity in other species of Noturus ranges from 2 to 9 years. Noturus flavus, the largest of all Noturus (Taylor 1969), is the longest-lived species with a maximum life span of 9 years (Scott and Crossman 1973). Noturus funebris and N. insignis both live a maximum of 4 years (Thomerson 1966, Clugston and Cooper 1960); N. albater is reported to reach 3 years (Mayden et al. 1980). Pop- ulations of N. gyrinus have been reported to live to nearly 4 years ' LIFE HISTORY OF THE SLENDER MADTOM 59 500 100 50 NUMBER OF INDIVIDUALS 1 O %I+ 24+ 3+ 4+ O 1+ 2+ 3+ 4+ O I+ 2+ 3+ 4+ MALES FEMALES TOTAL YEAR CLASS Fic. 21.—Survival curves for Noturus exilis from Green Creek. Points on curves were obtained by multiplying data in the survival column 1x! of Table 14 by 1000. The ordinate is a logarithmic scale. in Lake Erie (Mahon 1977) and 2 years in Demming Lake, Minne- sota (Hooper 1949). INTERACTIONS WITH OTHER ORGANISMS Predation In addition to N. exilis egg predation by nest-guarding males only two other cases of predation were observed. Orconectes virilis, Campostoma anomalum and Etheostoma caerulewm ate embryos 60 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY after nests were exposed by observers and abandoned by guardian males. Fowler (1917) observed hatchling predation of N. insignis by a cyprinid, but only after the young had been dislodged from the nest. Notropis zonatus, Etheostoma caeruleum and Percina evides were observed eating embryos and. hatchlings of N. albater after nests had been exposed (Mayden et al. 1980). As potential predators of N. exilis, one Amia calva (420 mm), three Semotilus atromaculatus (93.0-126.4 mm), three Ictalurus natalis (34.0-90.0 mm), one Cottus carolinae (48.7 mm), four Mi- cropterus punctulatus (135.2-177.0 mm), 10 Lepomis megalotis (87.0-130.9 mm) and one Pomoxis annularis (113.0 mm) were ex- amined from Green and Hutchins creeks. No remains of N. exilis were found in their stomachs. Parasitism Noturus exilis from Green Creek were frequently parasitized by encysted nematodes; of the 361 individuals examined, 140 were infested. Cyst infestation occurred as early as 2-months-of-age (23.7 mm) and was found in all older age classes. Cysts were located in gastrointestinal mesenteries, on the exterior of the stomach or in- testine and extended, in some specimens, from the outside of the stomach or intestine by a cord. The number of cysts per individual madtom usually ranged from one to 15, but in some, an estimated 100 or more cysts were present per individual. Diameters of cysts ranged from 0.2 to 0.5 mm, x = 0.33, and length of the longest cord extending from a stomach was 2.6 mm. Infestation occurred during all months of the year but the percent of individuals infested varied each month. Infestation was highest in March and April with 57 and 52 percent infestation, respectively, and lowest in June and September with 18.2 and 17.9 percent infestation, respectively. For all other months, percent infestation ranged from 24 to 49 percent. Eighteen acanthocephalans (Leptorhyncoides) were found in intestines of 11 individuals from February, March, April and Sep- tember. There were one-two Leptorhyncoides per individual. The only ectoparasites were one piscicolid leech attached to the pectoral fin of a 57.4 mm female in November and one copepod (Ergasilis) on the first gill arch of a 74.1 mm male in February. ACKNOWLEDGMENTS We are indebted to S. L. Dewey, Department of Zoology, Southern Illinois University at Carbondale (SIU-C), for preparation of the illustrations; to R. A. Brandon, S. L. Dewey and W. D. Klimstra for suggestions on the manuscript; to S. L. Dewey for many hours of discussion on life history tactics; to R. A. Brandon, M. H. Burr, P. A. Burr, S. L. Dewey, M. S. Ellinger, A. Fitzpatrick, _ LIFE HISTORY OF THE SLENDER MADTOM 61 K. Fitzpatrick, M. A. Morris, A. Rauch, M. E. Retzer and S. J. Walsh (SIU-C) for aid in collecting specimens; to J. A. Beatty, R. W. Sites and R. L. Price (SIU-C) for help in identifying some food organisms and parasites; to K. Fitzpatrick for information about the clutches of N. exilis eggs spawned in his experimental aquarium; to Doris Sublette, Hlinois Natural History Survey, for providing information on rare literature; to Karen Schmitt, Scien- tific Photography and Illustration Facility of the Southern Illinois University at Carbondale Graduate School for assistance in the preparation of some of the illustrations; and to Dr. J. T. Mouw, Department of Guidance and Educational Psychology, and M. Chong, Academic Computing Facilities, for statistical and computer assistance. This research was supported, in part, by grants to B. M. Burr from the Southern Illinois University at Carbondale Office of Re- search Development and Administration and the U.S.D.A. Forest Service. This paper is extracted from a thesis by the first author submitted to Southern Illinois University at Carbondale in partial fulfillment of the requirements for the M.A. degree. SUMMARY Major aspects of the life history of Noturus exilis were studied from collections and observations made at Green and Hutchins creeks, Union County, Illinois, between 6 November 1977 and 29 October 1979. During the study, 972 specimens ranging in age from 1 to 59 months were collected from Green Creek and 16 nests were found in Green and Hutchins creeks. Habitat varied with age and season of the year; juveniles were usually in shallow rifles and adults in pools, except during the breeding season when adults were most frequently found in riffles. The density in pools ranged from 0.00 madtoms per m? in January and March to 2.30 per m? in August, in riffles from 0.60 in No- vember to 12.92 in December. Individual N. exilis grew in standard length at a decreasing rate and in adjusted body weight at a constant increasing rate for at least 3-++ years for females and 4+ years for males. One-half of the first years growth in length was reached in about 3 weeks; about 4 months was required to attain one-half of the first year’s weight. The maximum size attained at Green Creek was 100.0 mm standard length for males and 81.2 mm for females. Growth rates fluctuated seasonally and were highest during the first 12 months of life. Survival of males and females was similar and was least in the second year of life. In the total sample, females slightly outnum- bered males, but the difference was probably the result of sampling error. Both sexes generally matured in 2 years; however, some indi- 62 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY viduals 1 year of age had some characteristics of mature individuals. In mature and some subadult males, a gradual increase in the gonadosomatic index (GSI) was observed from early winter to May-July when a peak in relative testis weight was observed. Most adult and some subadult females had mature ova from April to late July. During the breeding season, there was sexual dimorphism in shape and size of the genital papillae and males uae swollen cephalic epaxial muscles. Adult females maintained a greater mean GSI than subadult females over the two years, and were more consistent in their ovarian cycle. During spawning (May-July), l-year-old females consistently produced fewer and lighter (i.e., mg) ovarian eggs and had a lower GSI than did breeding females 2-years-old or older. The number of mature, ovarian eggs produced by a female ranged from 26 to 150 (x = 83.6) and was correlated with length, weight and age of the female. : The nesting season spanned at least 17 June to 17 July. Nests were always found in cavities, constructed by adult males, in either pools or rifles beneath large rocks. In all cases only one clutch of eggs was found per nest. Nests were either guarded by a single male, by both a male and female or unattended. Clutch size averaged 51; eggs were spherical and amber, were 3.9-4.5 mm in diameter (yolk diameter = 3.2-3.6) and adhered to each other in a roundish mass. Etheostoma caeruleum, Campostoma anomalum and Orconectes virilis were observed eating N. exilis eggs. Clutches of eggs incubated in the laboratory at 25°C hatched in 187-210 hr. Hatching success in the laboratory ranged from 39 to 75 percent. Within a clutch, embryos generally exhibited synchrony in development. Aquatic dipteran larvae and some crustaceans were the domi- nant food items found in stomachs. The types and sizes of food items consumed varied with size of the madtom and habitat in which it was found. Feeding activity in September increased mark- edly at night and was bimodal, with peaks occurring at 2000 and 0600 hr. Individual N. exilis were frequently parasitized by encysted nematodes and less frequently by acanthocephalans, leaches and parasitic copepods. LITERATURE CITED Apams, C. C., and T. L. Hankinson. 1928. The ecology and economics of Oneida Lake fish. Roosevelt Wildl, Ann. 1:239-548. Baitey, R. M. 1938. The fishes of the Merrimack watershed. Pages 149-185 in Biological survey of the Merrimack watershed. N. H. Fish and Game Dep., Surv. Rep. No. 3. BrrkHeap, W. S. 1967. The comparative toxicity of stings of the ictalurid | LIFE HISTORY OF THE SLENDER MADTOM 63 catfish genera Ictalurus and Schilbeodes. Comp. Biochem. Physiol. 22:101-111. BirKHEAD, W. S. 1972. Toxicity of stings of ariid and ictalurid catfishes. Copeia 1972:790-807. Breper, C. M., Jr. 1935. The reproductive habits of the common catfish, Ameirus nebulosus (Le Sueur), with a discussion of their significance in ontogeny and phylogeny. Zoologica (N.Y.) 19:143-185. Breper, C. M., Jr., and D. E. Rosen. 1966. Modes of reproduction in fishes. Natural History Press, Garden City, N.Y. 941 p. Brown, M. E. 1957. The physiology of fishes. Vol. 1. Academic Press, N.Y. 447 p. Caruson, D. R. 1966. Age and growth of the stonecat, Noturus flavus Rafinesque, in the Vermillion River. Proc. S. D. Acad. Sci. 45:131-137. Car_ton, W. G., and W. B. Jackson. 1964. The use of spines for age determination of fish. Turtox News 42:282-283. CieMeEns, H. P., and K. E. SNEED. 1957. The spawning behavior of the channel catfish Ictalurus punctatus. U.S. Fish Wildl. Serv. Spec. Sci. Rep: Fish. No. 219. 11 p. Cxiucston, J. P., and E. L. Cooper. 1960. Growth of the common eastern madtom, Noturus insignis in central Pennsylvania. Copeia 1960:9-16. Curp, M. R. 1960. On the food and feeding habits of the catfish Schilbeodes exilis (Nelson) in Oklahoma. Proc. Okla. Acad. Sci. 40:26-29. DarneE.Lt, R. M., and R. R. Mezerotro. 1962. Determination of feeding chronology in fishes. Trans. Am. Fish. Soc. 91:313-320. Dervey, E. S., Jr. 1947. Life tables for natural populations of animals. Quart. Rev. Biol. 22:283-314. Dovucias, N. H. 1972. Noturus taylori, a new species of madtom (Pisces, Ictaluridae) from the Caddo River, southwest Arkansas. Copeia 1972: 785-789. Ernier, D. A., and R. E. Jenkins. 1980. Noturus stanauli, a new madtom catfish (Ictaluridae) from the Clinch and Duck rivers, Tennessee. Bull. Ala. Mus. Nat. Hist. 5:17-22. EvEeRMANN, B. W., and H. W. Crarx. 1920. Lake Maxinkuckee, a physical and biological survey. Ind. Dep. Cons. 1. 600 p. Fow.er, H. W. 1917. Some notes on the bregma habits of local catfishes. Copeia 1917:32-36. GREELEY, J. R. 1929. Fishes of the Erie- Nig watershed. Pages 150-179 in A biological survey of the Erie-Niagara system. Suppl. 18th Ann. Rep. N.Y. State Cons. Dep. 1928. GreeLey, J. R. 1934. Fishes of the Raquette watershed with annotated list. Pages 53-108 in A biological survey of the Raquette watershed. Suppl. 23rd Ann. Rep. N.Y. State Cons. Dep. 1933. Hewuer, T. R., Jr. 1967. The fishes of the Santa Fe River system. Bull. Flor. State Mus., Biol. Ser. 11. 46 p. Hocurr, C. H., and J. R. Sraurrer, Jr. 1976. Barbel anomaly in Noturus flavus Rafinesque. Proc. Penn. Acad. Sci. 50:117-118. Hoorer, F. F. 1949. Age analysis of a population of the ameiurid fish, Schilbeodes mollis (Hermann). Copeia 1949:34-38. Hynes, H. B. N. 1970. The ecology of running waters. Univ. of Toronto Press, Toronto. 555 p. Jenkins, R. E. 1976. A list of undescribed freshwater fish species of conti- nental United States and Canada, with additions to the 1970 checklist. Copeia 1976:642-644. Korrcamp, G. M., and P. B. Moye. 1972. Use of disposable beverage cans by fish in the San Joaquin valley. Trans. Am. Fish. Soc. 101:566. 64 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Lacter, K. F., J. E. Barpacu, R. R. Mizrer, and D. R. M. Passino. 1977. Ichthyology (2nd ed.). Wiley and Sons, N.Y. 506 p. Lancxois, T. H. 1954. The western end of Lake Erie and its ecology. Ed- “. wards Bros., Inc., Ann Arbor, Mich. 479 p. Manon, R. 1977. Age and fecundity of the tadpole madtom, Noturus gyrinus, on Long Point, Lake Erie. Can. Field-Nat. 91:292-294. Marzotr, R. C. 1955. Use of pectoral spines and vertebrae for determining age and rate of growth of the channel catfish. J. Wildl. Mgmt. 19: 243-249, Maypen, R. L., B. M. Burr, and S. L. Dewey. 1980. Aspects of the life history of the Ozark madtom, Noturus albater, in southeastern Missouri (Pisces:Ictaluridae). Am. Mid]. Nat. 104:335-340. MeENzEL, B. W., and E. C. Raney. 1973. Hybrid madtom catfish, Noturus gyrinus x Noturus miurus, from Cayuga Lake, New York. Am. Midl. Nat. 90:165-176. Nixotsky, G. V. 1963. The ecology of fishes. Academic Press, London. 352 p. Reep, H. D. 1907. The poison glands of Noturus and Schilbeodes. Am. Nat. 41:553-566. : Reep, H. D. 1924. The morphology and growth of the spines of siluroid fishes. J. Morph. 38:431-451. Ricker, W. E. 1975. Computation and interpretation of biological statistics of fish populations. Bull. Fish. Res. Board Canada 191. 382 p. Rickuers, R. E. 1973. Ecology (1st ed.). Chiron Press, Newton, Mass. 861 p. Rosison, H. W., and S. Winters. 1978. Occurrence of the slender madtom, Noturus exilis Nelson (Osteichthyes:Ictaluridae), in the Little River system, Arkansas. Southwest. Nat. 23:688-689. Scott, W. B., and E. J. Crossman. 1973. Freshwater fishes of Canada. Bull. Fish. Res. Board Canada 184. 966 p. SNEED, K. E. 1951. A method for calculating the growth of channel catfish, Ictalurus lacustris punctatus. Trans. Am. Fish Soc. 80:174-183. SNEED, K. E., and H. P. CLtemens. 1963. The morphology of the testes and accessory reproductive glands of the catfishes (Ictaluridae). Copeia 1963:606-611. STEGMAN, J. L., and W. L. Minckxiey. 1959. Occurrence of three species of | fishes in interstices of gravel in an area of subsurface flow. Copeia 1959:341. Taytor, W. R. 1969. A revision of the catfish genus Noturus Rafinesque, with an analysis of higher groups in the Ictaluridae. Bull. U.S. Natl. Mus. 282. 315 p. THoMerson, J. E. 1966. A collection of madtom catfish, Noturus funebris, from western Florida. Trans. Ill. State Acad. Sci. 59:397-398. Trautman, M. B. 1948. A natural hybrid catfish, Schilbeodes miurus x Schilbeodes mollis. Copeia 1948:166-174. UNIVERSITY OF KANSAS PUBLICATIONS MUSEUM OF NATURAL HISTORY The University of Kansas Publications, Museum of Natural History, beginning with volume 1 in 1946, was discontinued with volume 20 in 1971. Shorter research papers formerly pub- lished in the above series are now published as Occasional Papers, Museum of Natural History. The Miscellaneous Publi- cations, Museum of Natural History, began with number 1 in 1946. Longer research papers are published in that series. Monographs of the Museum of Natural History were initiated in 1970. All manuscripts are subject to critical review by intra- and extramural specialists; final acceptance is at the discretion of the publications committee. Institutional libraries interested in exchanging publications may obtain the Occasional Papers and Miscellaneous Publica- tions by addressing the Exchange Librarian, The University of Kansas Library, Lawrence, Kansas 66045. Individuals may pur- chase separate numbers of all series. Prices may be obtained upon request addressed to Publications Secretary, Museum of Natural History, The University of Kansas, Lawrence, Kansas 66045. Editor: E. O. WiLEy Managing Editor: JosepH T. CoLLins PRINTED BY UNIVERSITY OF KANSAS PRINTING SERVICE LAWRENCE, KANSAS Oe Fees ‘ OCCASIONAL PAPERS LIBRARY of the NOV 1 7 199) MUSEUM OF NATURAL HISTORY. ~- The University of Kansas ~™~!Y="S!Ty Lawrence, Kansas NUMBER 94, PAGES 1-48 NOVEMBER 5, 1981 AN ANNOTATED KEY TO THE LONG-TAILED SHREWS (GENUS SOREX) OF THE UNITED STATES AND CANADA, WITH NOTES ON MIDDLE AMERICAN SOREX By JANE ANN JUNGE' AND ROBERT S. HOFFMANN” The long-tailed shrews are a widespread and diverse group of small mammals living in the northern parts of both hemispheres. Many species, especially the smaller members of this Holarctic genus (Sorex), are poorly known, since individuals are difficult to capture and even more difficult to observe. The last complete revision of American Sorex (including Micro- sorex, treated here as a subgenus) was that of Jackson (1928). In the last half-century, a number of studies have clarified the systematic relationships of taxa within the genus, but these have not been compiled in a single review. Moreover, the most recent key (Hall, 1981) for the most part still reflects the species concepts of Jack- son’s revision, and depends heavily on geographic criteria. In this annotated key, species are identified on the basis of skull characters, in particular those of the maxillary toothrow, since the rostral/palatal region seems to remain patent even in specimens where the skull has been badly damaged. The practical significance of basing the key on these characters is that it will permit identi- fication of material from archeological or paleontological sites, as well as shrews damaged in capture, during preparation, or when decomposition has rendered body characters difficult or impossible to assess. The key includes new skull drawings to illustrate im- portant characters. 1 Curatorial Assistant, Museum of Natural History and Department of Sys- tematics and Ecology, The University of Kansas, Lawrence, Kansas 66045. 2 Curator of Mammals, Museum of Natural History and Professor, Depart- ment of Systematics and Ecology, The University of Kansas, Lawrence, Kansas 66045. 2 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY The notes on each species following the key describe external characteristics useful in identifying species, as well as suggestions for distinguishing among species of shrews found in the same geo- graphic area, and relevant habitat associations. Shrews that cannot be separated readily on the basis of skull characters are seldom found in the same geographic areas (there are exceptions to this in the Sorex cinereus group); therefore, distribution is a useful, though not infallible, aid in identification. The range maps for each species have been redrawn from published sources to reflect current dis- tributional records. The superspecies concept and nomenclatural concepts advocated by Amadon (1966) are employed herein. Decisions concerning the validity of species or allospecies, although based on the published literature as noted, are the responsibility of the second author. The Middle American Sorex are currently under study by Hoffmann and others and are therefore omitted from this key, pending revision. These shrews are included in the comments and provisional classi- fication. For a recent key to the Old World Sorex see Corbet (1978) and references cited therein. Key Characters Subgeneric Identification The presence of the post-mandibular foramen is used as a character to differentiate the subgenus Sorex from the subgenera Otisorex and Microsorex. The mandibular canal is present in all three subgenera; its foramen is situated on the lingual side of the ramus of the jaw, the canal running forward toward the incisors. The post-mandibular foramen appears in the general vicinity of the mandibular foramen in the subgenus Sorex, but the post-mandibular canal runs up into the ramus of the jaw. The mandibular and post- mandibular foramina may be separated on the jaw or may be con- fluent in the same depression (Fig. 1). The best way to identify the post-mandibular canal is to insert a fine probe or stiff hair into the foramen. Occasionally the post-mandibular foramen and canal are absent in individuals of the subgenus Sorex, more frequently in some species than in others; likewise a small foramen and canal may occasionally be present in Otisorex shrews, at least on one side of the jaw. A pigmented ridge runs from the apex to the cingulum on the lingual side of each unicuspid tooth (see below) of shrews in the subgenera Otisorex and Microsorex. The ridge may end in a pig- mented cusplet. The presence or absence of the ridge (Fig. 2) is not highly variable and is usually the best character to use in sub- generic determinations. In some individuals, however, pigmentation of the ridge may be faint or absent, especially in the species LONG-TAILED SHREWS (GENUS SOREX) 3 oe Fic. 1—vVentro-lateral view of posterior portion of mandible showing; a) absence of post-mandibular canal characteristic of Otisorex, and b) presence of post-mandibular canal characteristic of Sorex. Bar represents 1 mm in this and subsequent figures. fumeus, dispar, gaspensis, nanus, and longirostris. The distribution of these shrews is, fortunately, mostly outside that of any species of the subgenus Sorex. Long considered a separate genus, Microsorex was reduced to subgeneric status by Diersing (1980). He considered Microsorex to be the most specialized member of a gradational series of Otisorex shrews having increasingly larger accessory tines on the antero- medial surface of the first upper incisor; at the same time the jaws become shorter, leading to a reduction in size of the third and fifth 4 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Fic. 2.—Ventro-lateral view of rostrum, showing occlusal surface of uni- cuspid teeth; a) pigmented ridge characteristic of. Otisorex, and b) lack of pigmented ridge characteristic of Sorex. unicuspid teeth of the maxillary row. In Microsorex, U3 is com- pressed antero-posteriorally into a tiny disc which is usually not visible in lateral view. We agree with Diersing that the reduction of U3 is not a sufficient basis to warrant generic rank (see also Repenning, 1967). The two species thompsoni and hoyi recognized by Long (1972, 1974) were synonymized under Sorex hoyi by Van Zyll de Jong (1976a) and Diersing (1980). Specific Characters Accessory medial tine of the first upper incisor (11) Heptner and Dolgov (1967), Yudin (1969), Hoffmann (1971), Hennings and Hoffmann (1977), and Diersing and Hoffmeister LONG-TAILED SHREWS (GENUS SOREX) 5 (1977) used the antero-medial tine on the first upper incisors as a primary character in differentiating certain shrews. The presence or absence of this tine, its relative size, placement on the tooth, and the relationship of the pigmented area of the tine to that of the main pigmented area of the incisor are important characters used to discriminate between species in this key. The relative positions of tine and pigment remain constant throughout the life of the indi- vidual, becoming obscure only in old age when the incisors are extremely worn. There seems to be little individual variation of the tine within species, although it may be extremely small and difficult to find in some (e.g., S. trowbridgii), and absent in other species. Unicuspid teeth The five teeth in the maxillary toothrow following the first in- cisor (I1) are usually called “unicuspids” (Figs. 2, 4). The homol- ogies of these teeth are not clear (Repenning, 1967); all are char- acterized by development of a single cusp, making the tooth appear conical. The unicuspids are numbered U1 through U5; the four molariform teeth behind the unicuspids are P4 through MB. | Generally, U1 and U2 are subequal in size and larger than U3 and U4. U5 is always small. U3 and U4 are seldom the same size (except in S. dispar and S. gaspensis) and their relative size is one of the characters used in this key. Occasionally there may be indi- viduals within a species where U3 and U4 are subequal, rather than of different sizes (20% of S. longirostris; T. French, pers. comm.). With the possible exception of S. cinereus ohionensis, the size relationship is never consistently reversed within a population. The relative sizes of the unicuspids are evident throughout the life- span of the individual until advanced old age, when excessive tooth wear may obscure the relationship. In ventral view, the sides of the unicuspids may appear flat in some species (Fig. 23b), and relatively inflated (“bulbous”; see Choate, 1970) in others (Fig. 23a). Choate relates the degree of bulbousness of the teeth in Cryptotis to the hardness of food items —those species with bulbous teeth having to deal with relatively hard food items. J. S. Mellett (pers. comm.) has suggested that fibrous materials such as chitin may be more digestible if the fibers are reduced to small particle size. A more wedge-shaped tooth is more efficient in this respect. Although the relationship of tooth shape to diet in Sorex has not been studied, the character seems useful in making some distinctions in the key. Skull length Two skull length measurements are commonly used (Kirkland and Van Deusen, 1979); condylobasal length (anterior medial point 6 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY on premaxillary bones to posteriormost point on occipital condyle), and greatest length (anteriormost point on the first incisor to pos- teriormost point on the occipital condyle) (Fig. 3). Both are highly correlated with the overall mass and linear dimensions of the indi- vidual shrew, as well as with each other. Condylobasal length is the more difficult of the two to take, particularly on small shrews, since it requires that one point of the calipers or dividers be placed betweer-the incisors. Greatest length, while easy to measure, is subject to more age variation since the positions of the first incisors shift, rotating downwards as the individual ages and its teeth wear down (Diersing and Hoffmeister, 1977; Diersing, 1980). The skull length measurement employed here is condylobasal (CB) length. Skull length is useful in distinguishing between species in a single region, although interspecific geographic variation in size renders the measurement less useful as an absolute criterion. For example, where Sorex vagrans and Sorex monticolus are sympatric, the latter is the larger shrew at any one locality, even though over- all there is considerable overlap in measurement (Hennings and Hoffmann, 1977). In Sorex, age and sex do not affect condylobasal length, so shrews, regardless of sex and extent of tooth wear, may be compared. This is not to say that variation due to age does not appear, but it is slight (Findley, 1955a; Van Zyll de Jong, 1980; Diersing, 1980). Rostral/palatal size and proportions The size and shape of the anterior part of the skull is also useful in differentiating species of Sorex. This ranges from the long, nar- row rostrum of S. dispar (Figs. 5b, c) to the short, broad rostrum of S. merriami (Fig. 22a), but even superficially similar species such as S. cinereus and S. haydeni can be distinguished by their rostral/palatal proportions. Differences in proportions are best shown by either a bivariate plot or by the ratio of a width measure- ment against a length measurement, such as width across upper second molars versus unicuspid toothrow length (Van Zyll de Jong, 1980). Other measurements sometimes used are maxillary breadth, maxillary toothrow length, and palatal length (Fig. 3). Maxillary and unicuspid toothrow length are correlated with skull length, and may be used as an indication of size when other parts of the skull are damaged. Interorbital breadth While this measurement (Fig. 3) is not particularly variable between species, it may occasionally be useful. It has been used in distinguishing between S. vagrans and S. monticolus (Hennings and Hoffmann, 1977). LONG-TAILED SHREWS (GENUS SOREX) Fic. 3.—Lateral and ventral views of skull of Sorex dispar, showing meas- urements referred to in text. B-C, length of unicuspid toothrow; B-D, length of maxillary toothrow; E-D’, length of molariform toothrow; A-H, condylobasal length; K-H, greatest length of skull; K-M,M’, length of rostrum; J-J’, width of M2-M2; A-F, palatal length; G-G’, cranial breadth; N-P, cranial height; M-M’, maxillary breadth; L-L’, interorbital breadth. 8 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Cranial breadth and height Measurements of these dimensions (Fig. 3) are often frustrated by the fragile nature of the braincase of Sorex; moreover, cranial height varies with age and seasonally (Dehnel, 1949; Pucek, 1957; Dapson, 1968). Therefore, quantitative measurements of these char- acters have not been employed, although qualitative statements concerning relative flatness of the braincase are useful. Zygomatic plate The zygomatic plate comprises the outer wall of the infraorbital canal. It is bordered anteriorly by the anterior opening of the canal, posteriorly by its posterior opening and the maxillary process, and is pierced by the lacrimal foramen. The position of the anterior border of the plate and of the lacrimal foramen relative to the first and second upper molar teeth is sometimes a useful taxonomic character (Van Zyll de Jong, 1980). External measurements The external measurements usually taken on shrews are total length, length of tail, and length of hind foot. Since this key is based on cranial characters, other measurements are mentioned only in the notes on each species. Another reason for avoiding external measurements is that in such small animals, subject to rapid de- composition and deformation in the measuring process, it is difficult to make accurate measurements, particularly of total length. Since skull or toothrow (see above) lengths are correlated with external linear dimensions and can be measured more accurately, they are preferred as an indication of size. Key to the Long-tailed Shrews (Genus Sorex) of the United States and Canada 1. Usually no post-mandibular foramen (Fig. la); or, if pres- ent, small and often only on one mandible. Unicuspids with a ridge, usually pigmented, on the lingual face, run- ning from apex to cingulum, sometimes ending in a pig- mented ‘clispige (Pig way) yer 4 oe ete dee i Lave 2 la. Usually a well-developed post-mandibular foramen (Fig. lb). No pigmented ridge on lingual face of the unicuspids (Fig, 2b); SWpRENUIS: Sorex iis 3s 2 ee 22 2. Unicuspid toothrow “crowded,” only three unicuspids easily visible in lateral (buccal) view (Fig. 4a); U3 tiny, disc- like; U4 of normal size and shape; U5 minute (Fig. 4b). Accessory medial tine on anterior surface of first upper in- 2a. 3a. 4a. Ba. 6a. _ LONG-TAILED SHREWS (GENUS SOREX) 9 cisor (I1) relatively large and long (Fig. 4c). Condylo- basal (CB) length 12.7-15.8 mm. Distribution, boreal and montane (Fig. 24). Subgenus Microsorex _..-------- ge ee ee Sorex hoyi (p. 25) Four or five unicuspids visible from side; U3 equal to or larger than U4 (Fig. 5a), or if smaller, of normal shape rather than disc-like (Fig. 6a). Medial tine, if present, not as above (Figs. 6b, 7b, 9). Subgenus Otisorex ________. oS U3 and U4 usually equal in size (Fig. 5a, b), or if different, U3 slightly smaller; anterior edge of zygomatic plate pos- terior to plane separating M1 and M2 (Fig. 5a); rostrum unusually long and narrow, cranium flattened. Unicuspids widely spaced, relatively narrow (Fig. 5b, c); ridge on lingual face of unicuspids often lacking pigment (Fig. 5c). SOTe% (dispar enouppes | wineries Tt te 4 INGER OVS 5 Skull larger, CB length 16.5-18.4 mm. Distributed in up- land areas in a narrow belt running along the Appalachian Mountains from Maine to North Carolina, with an isolated population in New Brunswick, Canada (Fig. 25). SR SOE Ss 2S I on ae Sorex dispar (p. 25) Skull smaller, CB length 15.4-16.4 mm. Distribution, Gaspé Peninsula, northern New Brunswick, and Cape Breton Is. (Lie 2)5)) eet a: tls Sv oe Sorex gaspensis (p. 27) U3 usually distinctly smaller than U4 (Figs. 6a; 7a; 12a, b; il4a,,b), sometimes subequal 6 U3 usually larger than U4, sometimes subequal (Figs. 15a, 16a, 17a, 18, 19). Sorex fumeus and cinereus groups -.......- 14 Skull large, CB length usually more than 19.0 mm. -_..... il Skull medium to small, CB length usually less than 19.0 Largest North American Sorex, CB length 20.8-23.8 mm; skull and teeth robust; rostrum relatively long, broad, dis- tinctly downcurved (Fig. 6a); medial tine may be large, placed high on face of I1, pigmented; pigmented area of incisor may curve up to meet that of tine. Distribution, northwest Pacific coast (Fig. 24). Sorex bendirii (p. 28) 10 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Fic. 4.—Sorex hoyi; a) lateral (buccal) view of unicuspid toothrow, b) ventral view, showing U1, U2, U4, and tiny U3 and U5, and c) frontal view of first incisor. Ta. Medial tine absent or, if present, small and placed low on medial face of I1 (Fig. 7b); somewhat smaller, CB length 19.0-22.8 mm; skull and teeth not so robust, rostrum not COWIE GMNUE he A) se eee SS ee 8 Small medial tines present on Il (Fig. 7b); unicuspids longer than wide in ventral view (Fig. 8a). Distribution, boreo-montane (Fig. BOW paste uke tc ht i ees Be teste _ Sorex palustris (incl. alaskanus) (p. 28) . No Seat tines present on I] (as in Fig. 21b); unicuspids wider than long to approximately quadrate in ventral view (Fig. 8b). Distribution, Pacific coast from San Francisco Bay north to central Oregon (Fig. 27). Sorex pacificus (p. 28) . LONG-TAILED SHREWS (GENUS SOREX) 11 Fic. 5.—Sorex dispar; a) lateral view of skull, b) ventral view of skull, and c) ventral view showing occlusal surface of upper unicuspids. 9. Medial tine begins above main pigmented area of first in- CCUG Oe (G1 ENKEAARLS Ye of) rae ls Meee Rees Mes een Te nSeMan Pec sec Reema, Ore y: 10 9a. Medial tine contained entirely within pigmented area of THiS Gener S Olin GEMIGMNOC eho eee ee Ae oe Re, eee i 12 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY ae Fic. 6.—Sorex bendivrii; a) lateral view of skull, and b) frontal view of first incisor. fo) desu Fic. 7.—Sorex palustris; a) lateral view of skull, and b) frontal view of first incisor. ~ LONG-TAILED SHREWS (GENUS SOREX) 13 Fic. 8.—Ventral view of upper unicuspids of a) Sorex palustris, and b) Sorex pacificus. 10. Skull larger, CB length 15.5-17.5 mm. Usually distinct unpigmented gap between upper pigmented area of first incisor and pigmented medial tine (Fig. 9a). Rostrum rela- tively long and narrow, cranium relatively inflated. Unicus- pids quadrate or longer than wide; U1 and U2 more robust, bulbous; U3 definitely smaller than U4. Distribution, west- erm North America (Fig, 28), — 2. Sorex vagrans (p. 31) OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY a b ¢C Fic. 9.—Frontal view of first incisors of a) Sorex vagrans, b) Sorex longi- 10a. ih lla. 12. rostris, and c) Sorex monticolus. Skull smaller, CB length 13.8-15.6 mm (except S. I. fisheri, 15.4-16.4 mm). Usually no definite gap between pigmented areas of incisor and medial tine (Fig. 9b). Rostrum rela- tively short and broad, cranium relatively flat (Fig. 10). Unicuspids wider than long; Ul and U2 less robust, not bulbous; U3 and U4 sometimes subequal. Distribution, southeastern North America (Fig. 28). _.______ an ee es) a Sorex longirostris (p. 33) CB length 15.4 mm or more; in ventral view, U3 usually distmetly smaller than W4-(Figs: 12.132) ee 12 CB length 15.3 mm or less; U3 and U4 may sometimes appear subequal in ventral view (Fig. 13b). —.--___ 13 CB length 16.1-19.2 mm; cranium relatively inflated (Fig. lla). Unicuspids become appressed posteriorly, U5 against P4, with no noticeable gap between U5 and P4 along medial edge (Fig. 12a). Distribution, northwestern mon- tane and boreal North America (mostly north and east of Si ONnarysy) CW Oy Never ate Sorex monticolus (p. 33) Fic. 10.—Lateral view of skull of Sorex longirostris. LONG-TAILED SHREWS (GENUS SOREX) 15 12a. CB length 15.4-17.0 mm; cranium relatively flat (Fig. 11b). Unicuspid toothrow less appressed posteriorly, so that there is a distinct triangular gap between U5 and P4 along the medial edge (Figs. 12b, 13a). Distribution, California, south from the San Francisco Bay area, west of the crest of the Sierra Nevada, Santa Catalina Island, and Baja Califor- nia (mostly south and west of S. monticolus) (Fig. 24). Sorex ornatus (incl. S. sinuosus, S. willeti, S. juncensis) (p. 34) 13. CB length 13.8-14.8 mm; cranium extremely flat (Fig. 14a). Distribution, discontinuous, mostly montane, in Rocky Mountains, Colorado Plateau, and western Great Plains (Hitters 25) wets ye ie eee oie ere Sorex nanus (p. 35) 13a. CB length 14.5-15.3 mm; cranium less flat (Fig. 14b). Dis- tribution, southern Great Basin (Fig. 25). ws ne ise or on ie aS 2 Sorex tenellus (p. 35) Fic. 11.—Lateral view of skulls of a) Sorex monticolus, and b) Sorex ornatus. 16 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Fic. 12.—Ventral view of skulls of a) Sorex monticolus, and b) Sorex ornatus. 14. Skull large, CB length 17.8-19.0 mm; palate broad; maxil- lary breadth greater than 4.6 mm; cranium usually flattened to some degree, sometimes extremely so (Fig. 15a). Uni- cuspids wider than long; ridge on lingual face of unicuspid often lacking pigment (Fig. 15b). Small post-mandibular foramen sometimes present, at least on one mandible. Dis- tribution, northeastern United States and southeastern Canada. (Fig. 2l))eeee oo. SOrenmiumens (ps O09) 14a. Skull small to moderate, CB length 13.8-17.0 mm, the larger species with inflated braincase (Fig. 16a). Maxillary breadth less than 4.6 mm; unicuspids relatively narrow, quadrate to longer than wide (Fig. 16b). Sorex cinereus [2461] 3 eR eee EE. : SO AMMR OEE NOE. ean ve END Serve fr) LONG-TAILED SHREWS (GENUS SOREX) 17 Fic. 13.—Ventral view of upper unicuspids of a) Sorex ornatus, and b) Sorex nanus. | Fic. 14.—Lateral view of skulls of a) Sorex nanus, and b) Sorex tenellus. 18 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Fic. 15.—Sorex fumeus; a) lateral view of skull, and b) ventral view of skull. 15. Extremely small, CB length 13.8-14.6 mm, skull relatively flat. Unicuspid teeth closely appressed, rostrum relatively short (Fig. 17a, b). When sympatric with S. cinereus and S. haydeni (see below), this is the smaller shrew. Distri- bution, Columbia Plateau to western Great Plains (Fig. 29). ee eae aS 1. ee Sorex preblei (p. 35) 15a. Usually larger, CB length usually 14.6-17.0 mm (but to 14.1 mm in some populations, see below); unicuspids well- spaced; rostrum relatively more elongate (Figs. 16, 18)... 16 16. Found only on islands in Bering Strait or Sea. —_..---- 17 16a, Not. as: abovevei-e8 4 i ee ee eG 17. Known only from St. Paul, Pribilof Islands (Fig. 30). (Report from Unalaska probably in error.) —-.....-..--....-.- Sorex hydrodromus (incl. S. pribilofensis) (p. 36) 17a. Known ee from St. Lawrence Island ae OO) qos eee _ Sorex jacksoni (p. 36) LONG-TAILED SHREWS (GENUS SOREX) 19 Fic. 16.—Sorex cinereus; a) lateral view of skull, and b) ventral view of skull. 18. 18a. 19. 19a. 20. Usually smaller, CB length 14.1-15.6 mm; unicuspid tooth- row relatively shorter, M2-M2 width relatively larger; ratio of length of unicuspid toothrow to ee -M2 width usually lessathanmOiGe (heel Glo), 22:2 ce ee ee 5 ike) Usually larger, CB length 14.6-17.0 mm; unicuspid tooth- row relatively longer, width of M2-M2 relatively smaller; ratio of unicuspid toothrow length to M2-M2 width usually Greate msbncmny (On (bo WOW) ia ck 5 eee ts 20 Restricted to northern Great Plains area, south of 55 de- grees N Latitude (Fig. 30). Sorex haydeni (p. 36) Restricted to tundra of extreme northwestern North Amer- ica, north of 58 degrees N Latitude (Fig. 30). BabA scagllrate 2 AScmaaale telatod . Sorex cinereus (in part) (p. 39) CB length 14.6-15.2 mm; cranium relatively flat; rostrum broad, and unicuspid toothrow slightly shorter (Fig. 19). Distribution, southern Pennsylvania, Delaware, Maryland, northeastern Virginia (Fig. 30). Sorex fontinalis (p. 38) 20 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Fic. 17.—Sorex preblei; a) lateral view of skull, and b) ventral view of skull. 20a. CB length 15.0-17.0 mm; cranium inflated; rostrum narrow, and unicuspid toothrow slightly longer (Fig. 16a, b). -- 21 21. Found only in central Sierra Nevada (Fig. 30). —_----- re Va ee ee ee Sorex lyelli (p. 39) 21a. Found throughout much of northern and middle latitudes _ of North America (Fig. 30). _. Sorex cinereus (in part) (p. 39) 22. U3 usually smaller than U4 (Fig. 20a). Small medial tine high on anterior face of I1 i Fig. uke Distribution, Pacific Coast (Fig. 31). . ee _ Sorex trowbridgii (p. 40) 22a. U3 usually larger than U4 Figs Dias OF; O39) 52 2 hee ee 23 23. No medial tine on anterior face of Il (Fig. 21b). Palate unusually broad (Fig. 22a); CB length 15.0-16.6 mm. Dis- tribution, Columbia Plateau and Great Basin to western Great Plains (Fig. 31). —-................... Sorex merriami (p. 40) 23a. Medial tine on anterior face of I] (as in Fig. 20); palate not unusuallybrosnd«(PighB2b)i ie ee ey ea 24 LONG-TAILED SHREWS (GENUS SOREX) 21 Fic. 18.—Sorex haydeni; a) lateral view of skull, and b) ventral view of skull. Fic. 19.—Lateral view of skull of Sorex fontinalis. 24. CB length 16.5 mm or less, maxillary toothrow 6.0-6.4 mm (Fig. 22b). Described only from mountains of Arizona, New Mexico, and adjacent Mexico (Fig. 31). ri ea a ol eta 1 yn ee Sorex arizonae (p. 42) 24a. Skull length 17.0 mm or more; maxillary toothrow 6.0-7.8 mm. Found in upper Great Lakes, northeastern Great Plains, Canada, and Alaska (Fig. 29). Sorex arcticus group 22 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Fic. 20.—Sorex trowbridgii; a) lateral view of rostrum and unicuspids, and b) frontal view of first incisor. Fic. 21.—Sorex merriami; a) lateral view of upper unicuspids and b) frontal view of first incisor. LONG-TAILED SHREWS (GENUS SOREX) 23 25. CB length 18.3-20.3 mm; maxillary toothrow 6.8-7.8 mm. Unicuspid row appears uncrowded; unicuspids robust and appear bulbous in ventral view (Fig. 23a). Distribution, upper Great Lakes, northeastern Great Plains, and Canada, except NW Yukon and extreme NW British Columbia (Vice, SAC) |e a a Stee nee Reed WO ne Sorex arcticus (p. 42) 25a. CB length 17.0-18.5 mm; maxillary toothrow 6.0-6.9 mm. Unicuspids appear crowded, less robust, not bulbous (Fig. 23b). Distribution, Alaska, Yukon, and extreme NW Brit- isn Columioia Clie. 29). 2 22 2 Sorex tundrensis (p. 42) DISCUSSION With very few exceptions (e.g., S. fontinalis, S. lyelli, S. ari- zonae, and the insular species, S. hydrodromus and S. jacksoni), shrews occupy broad geographic areas. Certain regions, especially the Pacific Coast, may support a bewildering assortment of sym- Fic. 22.—Ventral view of skulls of a) Sorex merriami, and b) Sorex arizonae. 24 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY patric or parapatric shrews. Although most species key out entirely on the basis of skull characters, occasionally reference must be made to the distribution of a species. Distribution maps have been provided for each of the species in the key as well as for the Middle American species discussed below. As far.as possible, maps have been drawn from current sources; all are to the same scale and are bipolar oblique conic conformal projections. Three groups, S. hoyi, S. palustris, and the S. cinereus complex (cinereus, haydeni, preblei, fontinalis) are extremely wide-ranging, coming in contact in one part or other of their ranges with many other species of Sorex. The very small S. hoyi, with its unique disc- shaped U3, and the large S. palustris, with its fringed hind feet and distinctive pelage color, are readily distinguished from other Sorex Fic. 23.—Ventral view of palates of a) Sorex arcticus, and b) Sorex tundrensis. LONG-TAILED SHREWS (GENUS SOREX) 25 inhabiting the same range. Notes on how to distinguish members of the more generalized S. cinereus complex from other sympatric Sorex, as well as from each other, are given in the detailed com- ments on each species below. Generally, however, shrews which live in the same geographic area are quite different in cranial characteristics, while those of similar morphology occupy different ranges. Systematic questions involving “salt-marsh melanism” in S. vagrans and S. ornatus in the San Francisco Bay area are discussed under S. ornatus. Relationships of the Mexican and Guatemalan Sorex are unclear, while those of S. tundrensis and the Palearctic “S. arcticus’ are being examined by the authors and others. The sys- tematic and nomenclatural problems associated with Sorex pribi- lofensis are discussed under S. hydrodromus. Comments on the Species of North and Middle American Sorex Species are listed in the following section in the same order as they appear in the key, followed by the Middle American species that are not included in the key. Annotation consists of brief notes on habitat, means by which species can be distinguished from one another in areas of sympatry, and comments on systematic status. Once a specimen is keyed out, its identity can be checked by reference to these notes. Subgenus Microsorex Pygmy Shrew (Sorex hoyi). The pygmy shrew is a widespread inhabitant of the northern coniferous forest or taiga, with southern outliers in the montane forests of the Appalachian and Rocky mountains. It seems to be the smallest long-tailed shrew in any one locality, although S. nanus may be nearly as small where both occur in the Rocky Mountains of Wyoming and Colorado (Figs. 24, 25). The pygmy shrew can be distinguished from all other North Amer- ican Sorex by its small disc-like U3 and the long medial tine on I1. Formerly considered a monotypic genus, Microsorex was reduced to subgeneric status by Diersing (1980) (see above, pg. 3). Subgenus Otisorex Rock Shrew (Sorex dispar). This and the related Gaspé shrew (see below) have distinctive long unicolored tails (80-90% of head/ body length; Kirkland, 1981) and long narrow rostra. S. dispar shares the southern part of its range (Fig. 25) with the smaller S. hoyi, S. longirostris, S. fontinalis, and S. cinereus, as well as the larger S. palustris. The only species it is likely to be confused with is S. fumeus which is similar in size to dispar, though slightly larger than gaspensis, and similar to both in color, especially in winter 26 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY pelage. S. fumeus has a shorter, bicolored tail and a broad rostrum and palate, distinctly different from the long narrow rostra and palates of S. dispar and S. gaspensis. While the unicuspids of dispar and gaspensis have the characteristic Otisorex ridge from apex to cingulum, in some specimens it is only lightly pigmented or not at all, and may superficially resemble the condition of subgenus Sorex. 7, BU Museum of Natural History “t -i96i___| 14m om le Ceol Miles | 120 Fic. 24.—Distribution of Sorex hoyi, Sorex bendirii, Sorex ornatus, Sorex veraepacis, and Sorex macrodon. LONG-TAILED SHREWS (GENUS SOREX) 27 These shrews are, as their name implies, most commonly found in rocky areas such as talus slopes and along streams (Kirkland, 1981). Gaspé Shrew (Sorex gaspensis). Although very similar to S. dis- par in characters and habitat, this species is smaller, and the ranges of the two forms are not known to meet (Fig. 25). For comparisons with other species, see S. dispar. U Museum of Natcrall History Ie i ——. 1961 O90 2m 3) 400} soo Scaled Miles SUAS: nanus~7gtnt 10 Fic. 25.—Distribution of Sorex dispar, Sorex gaspensis, Sorex nanus, and Sorex tenellus. 28 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Marsh Shrew (Sorex bendirii.) This is the largest species of Sorex in North America. It is restricted to the northwestern Pacific coastal region (Fig. 24) where it inhabits the forest floor in damp to wet forests, often along the banks of streams and ponds. It enters the water freely, but does not possess the dense fringe of hairs on the margins of the hind feet that adapt the water shrew (S. palustris) to swimming and diving. Both species occur in the Olympic Moun- tains of northwestern Washington and the Cascades of southwestern British Columbia, Washington, and Oregon, but only S. bendirii is known from the Coast Ranges, south to extreme northern Cali- fornia. The marsh shrew may be distinguished from the water shrew by its generally larger size (there is some overlap), relatively longer, more robust downcurved rostrum, and large medial tine on Il (Fig. 6). The dorsal color of S. bendirii is blackish, and the belly is black to brownish gray, unlike most S. palustris. S. pacificus (see below), which also occurs from central Oregon southward, differs in possessing no medial incisor tines, and is cinnamon-brown in color. S. bendirii also occurs with S. trowbridgii, a subgenus Sorex shrew with large post-mandibular foramina; and S. vagrans, from which it may be distinguished by size and incisor tine pattern. In the northern parts of its range S. bendirii occurs with the smaller S. monticolus and S. cinereus. Water Shrew (Sorex palustris). Second only to the marsh shrew in size, this species is the largest Sorex throughout much of its range. It is boreomontane in distribution, occurring throughout most of the boreal taiga and southward into the Sierra Nevada, Rocky, and Appalachian mountains (Fig. 26). It is usually closely restricted to the vicinity of streams and ponds, and is the most aquatic member of the genus. It occurs with three subgenus Sorex shrews; the usually tricolored S. arcticus and S. tundrensis in the northern parts of its range, as well as with the concolor S. trow- bridgii on the Pacific Coast. Its large size, fringed hind feet, and distinctive color (black dorsally, with a “frosting” of light hairs; silvery-white to grayish below, except in the northeastern U.S. and on Vancouver Island, where populations with dark brownish ven- ters occur) distinguish it from all other Otisorex shrews that may occur with it. These include S. cinereus, S. hoyi, and S. monticolus in the northern part of the range; S. fumeus, S. dispar, S. gaspensis, S. fontinalis, and S. longirostris in the Appalachians; and S. nanus, S. vagrans, S. lyelli, S. preblei, and S. haydeni in the Sierra and Rockies. Provisionally included here is the Glacier Bay water shrew (S. alaskanus) (Hall, 1981), known only from two specimens taken at Point Gustavus, Glacier Bay, Alaska (Jackson, 1928). Pacific Shrew (Sorex pacificus). This large, light brown shrew is restricted to the Pacific Coast coniferous forests from the central LONG-TAILED SHREWS (GENUS SOREX) 29 Oregon coast (Siletz River) and Cascades (Crescent Lake) south to the Siskiyou Mountains of northerm California in the interior, and along the Coast ranges to San Francisco Bay (Fig. 27). Through- out this range it exists with the much smaller S. vagrans, but unlike vagrans, pacificus lacks a medial tine on Il. It may be separated from S. trowbridgii on the basis of color and the absence of the post-mandibular canal in pacificus; from S. bendirii and S. palustris cy 16.4 mm) and preblei is smaller (CB length < 14.7 mm). Individuals in a few populations of cinereus exceed 16.5 mm in skull length (S. c. streatori of the northwestern Pacific Coast, to 17.0; S. c. miscix of Labrador and S. c. acadicus of Cape Breton Island and Nova Scotia, to 16.8 mm). In most cases this does not cause problems of identification, except on Cape Breton Island and Nova Scotia where these large S. cinereus are sympatric with S. gaspensis and S. fumeus, respectively, and may be confused. Of the three, fumeus is largest, with a broad rostrum (maxillary breadth > 4.6 mm); gaspensis is smallest and has a very long tail (45-55 mm) and narrow rostrum (maxillary breadth < 4.0 mm); and S. c. acadicus is intermediate in these characters (maxillary breadth 4.0- 4.4 mm, tail < 46 mm). Other populations of S. cinereus are smaller than 15.0 mm in 40 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY skull length. S. c. ugyunak ranges in tundra habitats across. north- ern Alaska to Hudson Bay. Its skull length is 14.1-15.5 mm, but the only other shrews in this area are the much larger subgenus Sorex shrews, S. arcticus and S. tundrensis, with which S. c. ugyunak shares a “tricolor” pelage (see Bee and Hall, 1956; frontispiece). Van Zyll de Jong (1976b) reports that S. c. ugyunak is closer in cranial morphology to S. hydrodromus, S. jacksoni, and S. haydeni than it is to other populations of S. cinereus; study of the zone of potential contact between ugyunak and cinereus spp. is needed. S. cinereus along the southeastern edge of the species range (S. c. lesueurii, S. c. ohionensis) have been discussed under the account of S. haydeni. Not only are some individuals unusually small (skull length 14.6-15.9 mm), but in some ohionensis, U3 is often smaller than U4 (Bole and Molthrop, 1942). Some of these populations may eventually prove to be referable to S. haydeni, to S. fontinalis, or perhaps even to S. longirostris. Subgenus Sorex Trowbridge Shrew (Sorex trowhridgii). This species is confined to the .coniferous forests of the Pacific Coast, ranging eastward to the east slopes of the Cascade Mountains and Sierra Nevada (Fig. 31). It is a fairly large, dark colored shrew, with venter nearly as dark as dorsum, but with a sharply bicolored tail, dark above and light below. Cranially, it is distinguished from other Pacific Coast Sorex by its combination of U3 smaller than U4, small medial tine on I1, and presence of post-mandibular foramina. It occurs with S. vagrans, S. monticolus, S. pacificus, S. ornatus, S. palustris, and S. bendirii, all Otisorex shrews which lack well-developed post- mandibular foramina. In southeastern Oregon and northwestern California, its range may overlap that of another shrew of the same subgenus, S. merriami (see below), but merriami does not have a medial tine on I] and its U3 is larger than U4. The habitats of the two also differ, S. trowbridgii being primarily a woodland species, whereas S. merriami inhabits xeric steppe and desert. Merriam Shrew (Sorex merriami). This shrew is one of the most xeric-adapted of all North American Sorex. It has been taken in scattered localities in the Great Basin and Columbia Plateau, and probably inhabits sagebrush desert and shrub steppe throughout this region. It also occurs in the northern Great Plains and southem Rocky Mountains (Fig. 31). Within its range it is the only species of the subgenus Sorex, except along the east slope of the Cascades and Sierra Nevada, where it may come in contact with S. trow- bridgii (see above). In the southern Rockies, S. arizonae is known from southeastern Arizona and southwestern New Mexico, and while the two species presently are considered allopatric, their LONG-TAILED SHREWS (GENUS SOREX) Al ranges may be found to come into contact or overlap. S. arizonae (see below) has a medial tine on I], which merriami lacks, as well as a narrower palate (Diersing and Hoffmeister, 1977). The only Otisorex shrews likely to be taken in the same places as Merriam shrews are the much smaller S. nanus, S. tenellus, and S. preblei, from which S. merriami can be easily distinguished by its paler 20 im & a S.trowbridgii- \: egy, | — : er AG \ S.fumeus Se re | | } -| KU Museum of Natural |History a 1961 | o 00 2 x oO! sw Scaled Miles Fic. 31.—Distribution of Sorex fumeus, Sorex trowbridgii, Sorex merriami, Sorex arizonae, Sorex emarginatus, Sorex ventralis, and Sorex stizodon. 42 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY color, well-developed post-mandibular canal, U3 larger than U4, lack of medial tine on I1, and broad palate. Arizona Shrew (Sorex arizonae). This species was described only recently (Diersing and Hoffmeister, 1977) and is presently known only from the Huachuca, Santa Rita, and Chiricahua moun- tains of southeastern Arizona, the Animas mountains of southwest- ern New~Mexico (Conway and Schmitt, 1978), and the Sierra Madre Occidental of Chihuahua, Mexico (Caire, et al., 1978) (Fig. 31). It may occur with S. merriami (see above), and the Otisorex shrew S. monticolus, which may be distinguished by subgeneric characters and by having U3 smaller than U4. Sorex emarginatus, known from Durango, Jalisco, and Zacatecas, Mexico (see below), is also a subgenus Sorex shrew, and is closest to S. arizonae (Dier- sing and Hoffmeister, 1977). S. emarginatus is not referable to S. oreopolus (Findley, 1955b), which-is a member of the subgenus Otisorex (Diersing and Hoffmeister, op. cit.). Arctic Shrew (Sorex arcticus). This large shrew, usually dis- tinguishable by its distinctively “tricolored” coat pattern, inhabits the transcontinental northern coniferous forest or boreal taiga from Nova Scotia and Quebec westward to the central Yukon (Fig. 31). It is replaced in the western Yukon and Alaska by the similar S. tundrensis, which until recently (Youngman, 1975) was considered conspecific with S. arcticus. Throughout this entire region, the arctic shrew is the only member of the subgenus Sorex, and may be distinguished from Otisorex shrews that occur in the same area by its well-developed post-mandibular canals, lack of pigmented ridges on the unicuspids, and externally, by its light sides, which usually contrast strongly with the dark back, especially in winter pelage (juveniles in summer pelage are at best, faintly tricolored). In southern Ontario, the range of S. arcticus overlaps that of S. fumeus, in which the pigmentation of the unicuspid ridges is often weakly developed, and in which a post-mandibular canal is some- times present, at least on one side. However, S. fumeus is smaller (maxillary toothrow less than 6.6 mm, compared to more than 6.7 mm for S. arcticus), and externally it lacks the light sides of S. arcticus. In northern Canada and Alaska (Fig. 31), S. arcticus occurs with populations of S. cinereus that may have distinctly lighter sides, but these individuals are much smaller and lack post- mandibular canals. In the western part of its range, arcticus also occurs with S. monticolus, which has U3 smaller than U4. Tundra Shrew (Sorex tundrensis). This large “tricolored” shrew replaces S. arcticus in the boreal taiga of the western Yukon, ex- treme northwestern British Columbia (Nagorsen and Jones, 1981) and Alaska, and probably also occurs extensively in eastern and central Eurasia, where it is usually referred to as S. arcticus (cf. LONG-TAILED SHREWS (GENUS SOREX) 43 Corbet, 1978). The ranges of tundrensis and arcticus are nowhere in contact, as far as is now known, but in any case, the two species can be separated by size and nature of the upper unicuspid teeth (see Key; Fig. 23). The tundra shrew can be distinguished from shrews of the subgenera Microsorex and Otisorex by the same char- acters which will separate S. arcticus (see above). Middle American Sorex Although the Mexican and Guatemalan Sorex have not been included in the Key, pending revision, some discussion of these shrews is appropriate here. Subgenus Otisorex Verapaz Shrew (Sorex veraepacis) and Large-toothed Shrew (Sorex macrodon). These are very large members of the subgenus Otisorex, rivaling S. pacificus and S. fumeus in size. They occur in montane forests from central Guerrero, Puebla, and Veracruz, Mex- ico, south through the highlands of Oaxaca and Chiapas, Mexico, to southwestern Guatemala (Fig. 24). They lack post-mandibular foramina, have pigmented ridges on the unicuspids, and U3 is smaller than U4. The medial tines of the first incisors are well- developed, but not heavily pigmented, and are rather high on the face of the incisor. S. macrodon and S. veraepacis are similar in morphology and allopatric in distribution; further collecting may demonstrate that they are conspecific. Miller Shrew (Sorex milleri). This species is restricted to the Sierra Madre Oriental of Coahuila and Nuevo Leon, Mexico (Fig. 30). It is of moderate size, with a narrow rostrum and inflated braincase; the third unicuspid is larger than the fourth. The post- mandibular foramina are sometimes present, but are variable in size and occurrence; in many specimens examined, they are absent. The species is tentatively assigned to the subgenus Otisorex, in the cine- reus group, as suggested by Findley (1955a, b). Volcano Shrew (Sorex oreopolus). This medium-sized shrew has long been confused with two other Mexican taxa, emarginatus and ventralis. Findley (1955b) combined them under this oldest name, but the holotype of oreopolus is an Otisorex shrew, whereas emargi- natus and ventralis belong to the subgenus Sorex (see below) (Diersing and Hoffmeister, 1977; Hoffmann, in prep.). Sorex vagrans orizabae has also been described from the Transverse Volcanic Belt, and appears to be very similar to S. oreopolus (Hoffmann, in prep.); it is therefore provisionally placed in synonymy with the older name. The volcano shrew occurs at high elevations from extreme southwestern Jalisco (type locality of oreopolus) to eastern Puebla (type locality of orizabae) and western Veracruz (Fig. 28). S. 44 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY saussurei (see below), a larger member of the subgenus Sorex, is sympatric with oreopolus in this area, but has well-developed post- mandibular foramina. S. oreopolus is not known to be sympatric with the two smaller subgenus Sorex shrews, S. emarginatus and S. ventralis (see below). Subgenus Sorex Saussure Shrew (Sorex saussurei). This species, as presently understood, is the most widespread and geographically variable shrew in Middle America (Fig. 26). It occurs from southern Coa- huila and Durango, Mexico, south to at least central Oaxaca. South of the Isthmus of Tehuantepec, in central Chiapas, Mexico, and in southwestern Guatemala other populations occur that are provision- ally retained in S. saussurei although they may eventually be shown to be closer to S. ventralis or else specifically distinct. The Saussure shrew has a well-developed post-mandibular foramen, and the medial tines on the first incisor are well developed and pigmented. The unicuspids lack pigmented ridges, and the first and second are subequal in size, the second often somewhat larger than the first; the third and fourth are also subequal, but markedly smaller than the first two. S. saussurei is sympatric with S. oreopolus and S. veraepacis (see above), and may also contact the southern ends of the ranges of S. monticolus and S. milleri; all of these species may be distinguished as Otisorex shrews. The relationship of saussurei to the smaller subgenus Sorex shrews of Mexico (emarginatus, ven- tralis) is confused. So far, they seem separable only on the basis of size, with little or no overlap between large and small taxa (CB length; saussurei > 17.4 mm, emarginatus and ventralis < 174 mm). Moreover, each of the allopatric small taxa co-occur with the large saussurei, but not with each other (i.e., saussurei and ventralis at Huachinango, Puebla; saussurei and emarginatus near Autlan, Jalisco). Cerro San Felipe Shrew (Sorex ventralis). As discussed above, this is a small member of the subgenus Sorex, and is not conspecific with S. oreopolus. It is known to occur from central Oaxaca north to northwestern Puebla (Fig. 31). Where it is sympatric with S. saussurei and S. veraepacis, it can be distinguished on the basis of size. From S. oreopolus, if they should prove to be sympatric, it can be told by its U3 subequal to or larger than U4, and the presence of a post-mandibular foramen. Its distribution seems to be allopatric to that of the very similar S. emarginatus (see below). Zacatecas Shrew (Sorex emarginatus). Like the previous species, this one has also been erroneously considered conspecific with S. oreopolus. It is a small (CB length 16.4-16.9 mm) member of the subgenus Sorex, occurring from southwestern Durango through ~ LONG-TAILED SHREWS (GENUS SOREX) 45 Zacatecas to southwestern Jalisco (Fig. 31). S. emarginatus may eventually prove to be conspecific with S. arizonae and S. ventralis, at least in part. Sclater Shrew (Sorex sclateri). This species is known only from four specimens from the type locality, Tumbala, Chiapas, Mexico, 5000 ft (Fig. 26). It is as large as S. veraepacis, but appears to belong to the subgenus Sorex rather than Otisorex. It has well- developed post-mandibular foramina and U3 is subequal to or larger than U4. More material is required to assess its relationships. Pale-toothed Shrew (Sorex stizodon). Only the holotype of this species is known. It was collected at San Cristobal, Chiapas, Mex- ico, 9000 ft, at the same place, but 600 feet higher than another unique holotype, S. saussurei cristobalensis (Fig. 31). The two holotypes bear the same size relationship to one another that ven- tralis and saussurei do north of the Isthmus of Tehuantepec in that stizodon is smaller (CB length 17.5 mm) and relatively short-tailed (41 mm), whereas S. s. cristobalensis is larger (CB length 18.5 mm), with a longer tail (46.5 mm). S. stizodon may eventually prove to be related to ventralis. A Provisional Classification of North American Sorex Genus Sorex Subgenus Sorex Sorex stizodon, incertae sedis Sorex sclateri, incertae sedis saussurei group Sorex saussurei Sorex [ventralis] ventralis [v. ?] emarginatus [v. P] arizonae merriami group Sorex merriami trowbridgii group Sorex trowbridgii arcticus group Sorex [arcticus] arcticus [a.] tundrensis Subgenus Otisorex Sorex bendirii, incertae sedis Sorex palustris, incertae sedis Sorex longirostris, incertae sedis 46 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY fumeus group Sorex fumeus dispar group Sorex [dispar] dispar [d.] gaspensis vagrans group Sorex [veraepacis] veraepacis [v.] macrodon Sorex [pacificus]pacificus [p.] monticolus Sorex [vagrans| vagrans [v.] oreopolus ornatus group Sorex [ornatus] ornatus [o.] tenellus [o.] nanus cinereus group Sorex milleri Sorex [cinereus] cinereus [c.] lyelli [c.] fontinalis [c.] hydrodromus [c.] jacksoni Lc.] haydeni Sorex preblei Subgenus Microsorex Sorex hoyi ACKNOWLEDGEMENTS We are grateful to J. R. Choate, V. E. Diersing, T. French, D. L. Pattie, and G. L. Kirkland, Jr., for criticizing early drafts of this key. Choate, R. D. Fisher (Nat. Mus. of Natural History, Washington, D.C.), D. F. Hoffmeister (Mus. Nat. Hist., Univ. IIli- nois, Urbana), P. L. Wright (Zool. Mus., Univ. Montana, Missoula), - Sydney Anderson (American Mus. Nat. Hist., New York), Bernardo Villa-R. (Univ. Aut. Nat. Mex.), Rollin Baker (Michigan State Univ., East Lansing), Emmet Hooper (Univ. Michigan, Ann Arbor), Glen Bradley (Univ. Nevada, Las Vegas), Patricia freeman (Field Mus. Nat. Hist., Chicago), R. L. Rudd and E. D. Jameson (Univ. California, Davis), S. Kortlucke (Univ. California, Berkeley) and C. G. Van Zyll de Jong (Nat. Mus. Canada, Ottawa) generously LONG-TAILED SHREWS (GENUS SOREX) AT loaned us critical specimens. E. O. Wiley provided editorial and other valuable advice. The figures were painstakingly drawn by Carol Borg Chittenden. The maps were drafted by D. K. Bennett. LITERATURE CITED Amapon, D. 1966. The superspecies concept. Syst. Zool., 15:245-249. Bee, J. W. and E. R. Harty. 1956. Mammals of northern Alaska. Univ. Kan- sas, Mus. Nat. Hist., Misc. Publ., 8:1-309. Bote, B. P. Jr. and P. N. Morttrurop. 1942. The Ohio Recent mammal col- lection in the Cleveland Museum of Natural History. Sci. Publ. Cleve- land Mus. Nat. Hist., 5:23-181. Brown, R. J. 1970. A comparative study of the chromosomes of some Pacific Coast shrews. Unpubl. Ph.D. dissertation, Univ. California, Davis. Catre, W., J. E. VAucHN, and V. E. Diersinc. 1978. First record of Sorex arizonae (Insectivora:Soricidae) from Mexico. Southwest. Nat., 23: 532-533. CuHoaTE, J. R. 1970. Systematics and zoogeography of Middle American shrews of the genus Cryptotis. Univ. Kansas Publ., Mus. Nat. Hist., 19: 195-317. Conway, M. C. and C. G. Scumirr. 1978. Record of the Arizona shrew (Sorex arizonae) from New Mexico. J. Mamm., 59:631. Corser, G. B. 1978. The mammals of the Palaearctic region. A taxonomic review. British Museum (Natural History) and Cornell Univ. Press, London and Ithaca, N.Y. [vii] + 314 pp. Darson, R. W. 1968. Growth patterns in a post-juvenile population of short- tailed shrews (Blarina brevicauda). Amer. Mid]. Nat., 79:118-129. DEuNEL, A. 1949. Badania nadrodzajem Sorex L. [Studies on the genus Sorex L.] Ann. Univ. M. Curie-Sklowdoska, Ludin, Sec. C, 4:14:17-102 (In Polish, English summary, 75-97). Diersinc, V. E. 1980. Systematics and evolution of the pygmy shrews (sub- genus Microsorex) of North America. J. Mamm., 61:76-101. Diersinc, V. E. and D. F. HorrmMeister. 1977. Revision of the shrews Sorex merriami and a description of a new species of the subgenus Sorex. J. Mamm., 58:321-333. FINDLEY, J. S. ‘O52. Speciation of the wandering shrew. Univ. Kansas Publ., Mus. Nat. Hist., 9:1-68. Finvxey, J. S. 1955b. Taxonomy and distribution of some American shrews. Univ. Kansas Publ., Mus. Nat. Hist., 7:613-618. GreEN, M. M. 1932. Mei unrecognized shrew from New Jersey. Univ. Cali- fornia Publ. Zool., 38:387-388. Hatt, E.R. 1959. dciaonal evidence indicating that Sorex hydrodromus Dobson is a member of the Sorex arcticus group of shrews. Pp. 263-265 in Fauna of the Aleutian islands and Alaska Peninsula, by O. J. Murie. N. Amer. Fauna, 61:1-406. Hatz, E. R. 1981. The mammals of North America. 2nd ed. John Wiley and Sons, New York. Hatt, E. R. and R. M. Grrmore. 1932. New mammals from St. Lawrence Island, Bering Sea, Alaska. Univ. California Publ. Zool., 38:391-404. Hennincs, D. and R. S. HorrMann. 1977. A review of the taxonomy of the Sorex vagrans species complex from western North America. Univ. Kan- sas, Mus. Nat. Hist., Occ. Pap., 68:1-35. Heptner, V. G. and V. A. Dotcoyv. 1967. O sistematicheskom polozhenni Sorex mirabilis Ognev, 1937. [Systematic position of Sorex mirabilis 48 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Ognev, 1937 (Mammalia, Soricidae)]. Zool. Zhurn., 56:1419-1422. (In Russian, English summary. ) HorrMann, R. S. 1971. Relationships of certain Holarctic shrews, genus Sorex. Zeitsch. f. Saugetierk., 36:193-200. HorrMann, R. S. and R. D. Fisuer. 1978. Additional distributional records of Preble’s shrew (Sorex preblei). J. Mamm., 59:883-884. HoFFMANN, R. S. and J. G. Owen. 1980. Sane tenellus and Sorex nanus. seve hae Species, 131:1-4. HoFrrMann, R. S. and R. S. PETERSON. 1967. Systematics and zoogeography of Sorex in the Bering Strait area. Syst. Zool., 16:127-136. HoFrFrMann, R. S., P. L. Wricut and F. E. Newsy. 1969. The distribution of some mammals in Montana. I. Mammals other than bats. J. Mamm., 50:579-604. Jackson, H. H. T. 1928. A taxonomic review of the American long-tailed shrews. N. Amer. Fauna, 51:1-238. KiekLanp, G. L., Jn. 1977. A re-examination of the subspecific status of the Maryland shrew, Sorex cinereus fontinalis Hollister. Proc. Pennsylvania Acad. Sci., 51:43-46. KirkLanp, G. L., Jr. 1981. Sorex dispar and Sorex gaspensis. Mammalian Species, 155:1-4. KirkLAnp, G. L., Jr. and H. M. VAN Deusen. 1979. The shrews of the Sorex dispar group: Sorex dispar Batchelder and Sorex gaspensis Anthony and Goodwin. Amer. Mus. Novitates, 2675:1-21. Lone, C. A. 1972. Taxonomic revision.of the mammalian genus Microsorex Coues. Trans. Kansas Acad. Sci., 74:181-196. Lonc, C. A. 1974. Microsorex hoyi and Microsorex thompsoni. Mammalian Species, 33:1-4. MerriaM, C. H. 1895. Synopsis of the American shrews of the genus Sorex. N. Amer. Fauna, 10:57-124. NacorseEn, D. W. and D. M. Jones. 1981. First records of the tundra shrew (Sorex tundrensis) in British Columbia. Canad. Field-Nat., 95:93-94. Pucexk, Z. 1957. Histomorphologische Untersuchungen uber die Winterdepres- sion des Schadals bei Sorex L. und Neomys Kaup. Ann. Univ. M. Curie- Sklodowska, Sec. C, 9:399-428. REPENNING, C. A. 1967. Subfamilies and genera of the Soricidae. Geol. Surv. Prof. Pap., 565:1-74. Rupp, R. L. 1955. Population variation and hybridization in some California shrews. Syst. Zool., 4:21-34. VAN ZyLL DE Jonc, C. G. 1976a. Are there two species of pygmy shrews (Microsorex)? Canad. Field-Nat., 90:485-487. VAN ZYLL DE Jonc, C. G. 1976b. A comparison between woodland and cane shrews (Sorex cinereus). Canad. J. Zool., 54:963-973. VAN ZyLu DE Jonc, C. G. 1980. Systematic relationships of woodland and prairie forms of the prairie shrew, Sorex cinereus cinereus Kerr and S. c. haydeni Baird, in the Canadian prairie provinces. J. Mamm., 61:66-75. Von Boeken, J. C., Jn. 1967. The land mammals of the Southern Califormia islands, pp. 245-263 in Proc. Symp. Biol. Calif. Islands. Santa Barbara Botanic Garden. Aug. 18, 1967. Wariccey, R. E., J. E. Dubois and H. W. R. CopLtanp. 1979. Habitat, abun- dance, and distribution of six species of shrews in Manitoba. J. Mamm., 60:505-520. YOUNGMAN, P. M. 1975. Mammals of the Yukon Territory. National Museum of Natural Science (Canada), Publ. Zool., 10:1-192. Yuprn, B. S. 1969. Novie dannie po sistematike nekotorikh vidovy zemleroek (Soricidae) Palearktiki i Nearktiki. Acta Theriol., 14:21-34. “ UNIVERSITY OF KANSAS PUBLICATIONS MUSEUM OF NATURAL HISTORY The University of Kansas Publications, Museum of Natural History, beginning with volume 1 in 1946, was discontinued with volume 20 in 1971. Shorter research papers formerly pub- lished in the above series are now published as Occasional Papers, Museum of Natural History. The Miscellaneous Publi- cations, Museum of Natural History, began with number 1 in 1946. Longer research papers are published in that series. Monographs of the Museum of Natural History were initiated in 1970. All manuscripts are subject to critical review by intra- and extramural specialists; final acceptance is at the discretion of the publications committee. Institutional libraries interested in exchanging publications may obtain the Occasional Papers and Miscellaneous Publica- tions by addressing the Exchange Librarian, The University of Kansas Library, Lawrence, Kansas 66045. Individuals may pur- chase separate numbers of all series. Prices may be obtained upon request addressed to Publications Secretary, Museum of Natural History, The University of Kansas, Lawrence, Kansas 66045. Editor: E. O. WILEY Managing Editor: Jos—ErH T. CoLiins PRINTED BY UNIVERSITY OF KANSAS PRINTING SERVICE LAWRENCE, KANSAS UA | BIUO ieee OCCASIONAL PAPERS DEC 92 1 198) of the ieee YNIVERSITY MUSEUM OF NATURAL HISTORY The University of Kansas Lawrence, Kansas NUMBER 95, PAGES 1-12 NOVEMBER 25, 1981 A NEW SPECIES OF STEAMER-DUCK _(TACHYERES) FROM ARGENTINA by Puitie S. HUMPHREY! AND Max C, THOMPSON? Steamer-ducks (Tachyeres) are relatively large diving ducks oc- curring in marine coastal and freshwater environments in southern South America and the Falkland Islands. The various species of steamer-ducks are so similar in appearance and size that many authors (see review in Murphy, 1936:951-972) considered them a single species, some members of which were flightless and others not. Murphy (1936:951-972) studied extensive series of study skins of the genus and recognized a single widely distributed flying species, patachonicus (Falklands and Fuego-Patagonia), and two flightless species, one in the Falklands (brachypterus) and the other (pteneres) in Tierra del Fuego and along the Southern coast of Chile. He (Murphy 1936:958, 960, 969) believed that pteneres did not occur along the Patagonian coast and that patachonicus occurred only as far north as Latitude 48° S (Puerto Deseado). Subsequent publications (Bo 1958:39; Blake 1977:227-778; Boswall and Prytherch 1972:125; Daciuk 1977:376; Jehl, et al. 1973:61; Johnsgard 1979:452-453; Olrog 1979:50; Zapata 1967:364; Zapata 1969:23) list records of steamer-ducks identified as either pteneres or patachonicus along the coasts of Santa Cruz (Rio Gallegos, Puerto Deseado, Mazarredo), and Chubut (Isla Quintano, Punta Tombo, Punta Ninfas, Peninsula Valdez); most of these are based 1 Director, Museum of Natural History and Professor, Department of Sys- tematics and Ecology, The University of Kansas, Lawrence, Kansas 66045, U.S.A. 2 Associate Professor, Department of Biology, Southwestern College, Win- field, Kansas 67156, U.S.A. 2 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Fic. 1.—Head Plumages of White-headed Flightless Steamer Duck (Tachyeres leucocephalus). A, male definitive alternate; B, male definitive basic; C, female definitive alternate; D, female definitive basic; E, juvenal, both sexes (see text). A NEW SPECIES OF STEAMER-DUCK 3 on sight observations. Scott and Sharp (1912:487) reported a specimen of Tachyeres from the province of Rio Negro (mouth of the Rio Negro) which we have examined and identified as T. patachonicus. As far as we know, only one (Bo 1958:39) of the published records of steamer-ducks in the Province of Chubut is based on a specimen; it was identified by Bo as T. patachonicus. Maurice Rumboll and Francisco Erize (pers. comm.) have suspected for some time that steamer-ducks seen along the coast of Chubut might be either the Falkland species (brachypterus), or an undescribed form, rather than pteneres or patachonicus. Olrog (1979:50) described the distribution of T. brachypterus as “Costas de las Malvinas y eventualmente en las de Chubut.” Todd (1979: 160) stated that “Maurice Rumboll . . . is of the opinion that nearly all of the flightless steamer ducks which occur from the Lamaire Channel (between the tip of South America and Staten Island) northward to the Valdez Peninsula are either Falkland Island steamer ducks or are representatives of an undescribed species (or, at the very least, a subspecies ).” Humphrey and Thompson collected 49 specimens of steamer- ducks September through November 1979 near Ushuaia, Tierra del Fuego, and along the coasts of Chubut and Santa Cruz, Argentina, and Humphrey and Bradley C. Livezey collected 59 additional specimens of the genus at Ushuaia, Puerto Deseado (Santa Cruz), and Puerto Melo (Chubut) December 1980 through February 1981. Study of this material, which is still in progress, has raised many questions and suggests that our knowledge of the kinds and relation- ships of Tachyeres is in need of extensive revision, probably in- volving recognition of several new taxa; one of these is the abun- dant, large, flightless steamer-duck from Puerto Melo, and probably elsewhere, which is clearly distinct from all other known species of the genus. Tachyeres leucocephalus new species WHITE-HEADED FLIGHTLESS STEAMER-DUCK Holotype—Museo Argentino de Ciencias Naturales No. 52694; adult male from Puerto Melo, Provincia de Chubut, Argentina (Latitude 45° 01’ S., Longitude 65° 50’ W.), collected 24 September 1979 by Max C. Thompson and Philip $. Humphrey (PSH 1398). Diagnosis—A medium-sized, sexually dimorphic, flightless steamer-duck. Head of adult males in alternate plumage predomi- f OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY C D Fic. 2.—Humeri and sterna of Tachyeres brachypterus (A, C) and T. leucocephalus (B,D), males. Note the thick shaft characteristic of all humeri of T. leucocephalus in contrast to the slender more tapering shaft of the humerus of T. brachypterus. The xiphoid region of the sternum of T. leuco- cephalus is consistently broader and flared more laterally than in T. brachyp- terus and the posterior lateral process is much wider. A NEW SPECIES OF STEAMER-DUCK 5 TasBLeE 1.—Means and ranges of measurements of male steamer-ducks. All measurements by Humphrey; measurements of T. brachypterus from specimens not collected by the authors; those of T. pteneres and T. patachonicus are from specimens collected by the authors and by others. Exposed Nail Wing Tarsus Digit III n culmen width (arc) length length T. patachonicus 22 53 12 296 59 81 46-59 10-14 275-315 49-63 73-86 T. leucocephalus 12 55 12 279% 65 87 51-63 11-13 262-295 61-67 84-90 T. brachypterus al 56 14 270 65 90 54-61 12-15 261-290 60-70 85-95 T. pteneres 23 60 16 QBS al 96 54-70 12-18 255-289 62-78 87-104 *n = 9 5) == 2) nantly white with light gray lores and cap becoming paler pos- teriorly and patch of light chestnut on throat. Female smaller than male; head of adult in alternate plumage with dark gray crown, reddish-brown lores and cheeks, broad white postocular streak widening posteriorly to fuse with broad white collar at base of neck, throat medium cinnamon bordered with medium gray. Tachyeres leucocephalus is distinct from all known species in the genus in terms of various combinations of characters, including body weight, proportions of certain measurements, shape of hu- merus and posterior region of sternum, and coloration of the feather- ing of the head and neck in various plumages. T. leucocephalus differs from T. brachypterus as follows: 1) hu- merus broad, heavy, and with little taper; it is more slender and tapering in brachypterus (Fig. 2); 2) posterior part of sternum broadly flared; it is not flared laterally in brachypterus (Fig. 2); 3) external lateral process of sternum wide; it is much narrower in brachypterus (Fig. 2); 4) mean body weight lighter sex for sex (Table 3); 5) means of nail width and length of digit III smaller but there is considerable overlap in these and other external meas- urements (Table 1); 6) adults of both sexes lack the yellowish or golden collar at the base of the neck characteristic of brachypterus (Murphy 1936:964); 7) adult females have wide, prominent postocular streak and a more strikingly patterned head and neck. The new species differs from T. pteneres in: 1) lower body weight (Table 3) and correspondingly small skeletal elements of trunk and hindlimb; 2) lower wing loading (Table 2); 3) mean number of lamellae per cm in upper bill is 6.5 (males) and 7.0 6 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY TasLe 2.—Means and ranges of ratios of lengths of femur and humerus and wing loading for male steamer-ducks. Wing loadings n Femur/Humerus ~ n (gms/cmd2 ) T. patachonicus 27 0.59 (0.57—-0.62) 22 2.37 (2.03-2.83) T. leucocephalus 17 0.66 (0.64-0.67) 6 3.18 (2.91-3.40) T. pteneres 10 0.68 (0.67—0.70) 10 4.90 (4.35-6.01) T. brachypterus 1 0.64 0 . (females) and there is no overlap in this character with pteneres which have 5.2 (males) and 5.9 (females) per cm; 4) means of nail width and lengths of tarsus and digit III smaller but there is considerable overlap in these and other external measurements; 5) adult males have white heads most of the year in contrast to _ pteneres in which the head is light gray. T. leucocephalus differs from T. patachonicus in 1) ratio of lengths of femur and humerus 0.64—0.67; it is 0.57-0.62 in T. pata- chonicus (Table 2); 2) mean lengths of tarsus and digit III larger (Table 1); 3) higher mean weight (Table 3); 4) higher wing loadings (Table 2); 5) mean ratio of lengths of tarsus and wing arc which is 0.23 (0.21-0.24); it is 0.20 (0.18-0.22) in patachonicus; 6) adult females have wide, prominent postocular streak and a more strikingly patterned head than those of T. patachonicus. Mean measurements (mm, kg) and extremes: males (12), ex- posed culmen 55 (51-63), nail width 12 (11-13), wing (arc) 279 (262-295), tarsus 65 (61-67), digit ITI length 87 (84-90), weight (n = 11) 3.90 (2.70-4.40); females (8), exposed culmen 55 (54-57), nail width 12 (11-13), wing (arc) 271 (255-288), tarsus 63 (59- 66), digit III length 84 (81-86), weight 2.95 (2.55-3.35). Distribution Known from the vicinity of the type locality where it occurs along rocky shores of the mainland and islands, TABLE 3.—Means and ranges of weights (g) of steamer-ducks. Weights for T. brachypterus (estimated means) from Weller (1976:47). n Males n Females T. patachonicus 27 3030 28 2425 2350-3480 1950-2900 T. leucocephalus 3790 8 2950 2700-4400 2550-3350 T. brachypterus 4334 4 3383 4200-4650 3100-3580 T. pteneres 5310 6 4328 4950-5650 3800-4820 A NEW SPECIES OF STEAMER-DUCK Zi from Punta Tombo, Chubut, where it has been photographed by Jeffery Boswall, Donaldo MaclIver, and E. R. Parrinder ( Boswall and Maclver 1979a), William Conway, and Francisco Erize, and from Camarones where it has been photographed by Roberto Stranek. It probably occurs in appropriate habitat along the coast of Chubut from Bahia Bustamante north perhaps as far as Puerto Madryn and Peninsula Valdez. Description of holotype—Head mostly white, forehead and crown light gray, becoming paler posteriorly and slightly mottled with pale cinnamon where wear has exposed the bases of the feathers; lores light gray, mottled posteriorly with medium dusky brown and pearl gray feathers; upper and lower eyelids white; chin and neck white; small light chestnut patch on throat becomng slightly paler posteriorly; mantle, scapulars, back, rump, upper tail coverts, sides, and flanks predominantly medium pearl gray be- coming paler on the upper back; most of the posterior mantle feathers and scapulars with paler, silvery gray patches around the dark rachis and darker gray (sometimes dusky) posterior margins, giving a scaled appearance; upper breast medium pearl gray, be- coming predominantly dusky brown medially, with scattered light to medium chestnut where feather bases show; feathers of upper breast and sides of breast with narrow, pale smoke gray margins; feathers of the sides and flanks with a silvery or slightly metallic pale pearl gray wash, narrowly tipped with medium chestnut or pale smoke gray on the older, more worn feathers. Lower breast, ab- domen, and under-tail coverts white. Wing with carpal and meta- carpal knobs; primaries and greater upper primary coverts dark dusky brown with blackish brown shafts; rest of upper wing medium fuscous, some (newer) feathers darker and others (worn) very pale buffy brown. Secondaries (10) white, the inner vane of the innermost (eleventh) medium blackish-brown; tertials medium dusky brown, paler at the edges; innermost tertials quite worn, their frayed edges being very pale, almost whitish, buffy brown. Axillaries and central part of wing lining white; balance of under- wing coverts medium dusky brown, the greater under primary coverts paler. Rectrices dark blackish brown, with black shaft and faint mesial wash of light silvery gray. Iris brown; bill orange, becoming greenish at tomium and yellowish near tip, the nail black; legs and feet bright yellow with claws blackish. Measurements (mm) of holotype.—Wing arc 280, exposed cul- men 56.2, width of nail 12.3, tarsus 66.4; weight 3.65 kg. Specimens examined.—Skins (holotype at MACN, others depos- ited at Southwestern College) and partial skeletons: 4 males, MACN 52694 (holotype), KU 77930 (SC 3528), 77942 (SC 3533), 77943 (SC 8 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY 3534); 3 females, KU 77926 (SC 3527), 77934 (SC 3530), 77936 (SC 3531). Freshly killed specimens subsequently prepared as complete skeletons: 13 males, KU 77925, 77928, 77931, 77933, 77935, 77938, 79234, 79235, 79236, 79237, 79243 (Juvenile), 79245, 79246, 79248; 11 females, KU 77927, 77979, 77937, 77939, 77941, 79238, 79239, 79240, 79241, 79247, 79249. DISCUSSION The White-headed Flightless Steamer-Duck is an abundant spe- cies at Puerto Melo and presumably other localities with rocky shorelines along the coast of Chubut. The Flying Steamer-Duck has been collected at Puerto Melo (1 specimen) and may occur in small numbers in Atlantic coastal Chubut in sympatry with T. leucocephalus. What are undoubtedly the nest and eggs of T. leucocephalus were described by Boswall and Maclver (1979a:75) and identified by them as pertaining to T. patachonicus. They found that 6 eggs in a nest at Punta Tombo, Chubut, “all . . . fall well outside the range of Murphy’s figures (for eggs of T. patachonicus from the Fuegian region) and are much closer to the eggs of the larger Falkland Island Flightless Steamer Duck, T. brachypterus .. .” (Boswall and Maclver 1979b). The molts and plumages of steamer-ducks are poorly under- stood. We know from studies now in progress that Flying Steamer- Ducks (T. patachonicus) have three molts and plumages per annual cycle (Humphrey and Livezey, in press). From examination of spec- imens and photographs of T. leucocephalus, we judge that adult birds of both sexes go through a complete prebasic molt in summer (February ), shedding the remiges simultaneously, and then without interruption undergo a partial molt involving the head and neck and possibly other parts of the body. Adult males collected in September were white-headed as were those with worn wings collected in February before they had initiated the prebasic molt. The limited data at our disposal suggest that either adult males wear a white alternate plumage of the head and neck most of the year or that, if there is a supplemental plumage, it too is white. In what follows, we tentatively assume a two plumage cycle since we have no evidence to the contrary. The definitive alternate plumage of the head and neck of males is predominantly white with a pale to medium gray cap which terminates anterior to the nuchal region, pale brownish in the an- terior lores, and a relatively narrow patch of light cinnamon on the throat. Many males seen in bright sun in the field appear com- pletely white-headed. Figure 1A illustrates a male with vestigial A NEW SPECIES OF STEAMER-DUCK 9 bursa of Fabricius, fresh new wings and tail, and head and neck in what we judge to be definitive alternate plumage. Although we have seen no males in full definitive basic plumage, we have examined birds in both prebasic and prealternate molt that have what we judge to be mixtures of alternate and basic plumages on the head and neck. The definitive basic plumage of males is probably similar in aspect to the definitive alternate plumage of females, having an extensive dark cap, medium reddish- brown cheeks and lores, broad white postocular streak, white collar at base of neck, an extensive patch of medium cinnamon on the throat, and medium gray on the sides of throat and chin. Figure 1B illustrates a specimen of an adult male (bursa of Fabricius absent) with new, short, remiges growing in and slight molt on the head. We believe this individual has mostly basic feathers on the head and neck and is beginning a prealternate molt. The definitive alternate plumage of females has dark cap, reddish-brown cheeks and lores, prominent white postocular streak, and broad white collar at the base of the neck. Figure 1C is of a female with comparatively small bursa of Fabricius (28 X 9 mm), fresh, new wings, and head and neck in what we interpret to be definitive alternate plumage. The head and neck of females in definitive basic plumage are medium to dark purplish-gray with a large dark cinnamon patch on the throat. Since we have examined no specimens in full basic plumage we cannot tell whether it lacks a postocular streak or has a very short white one. Figure 1D is of a female with vestigial bursa of Fabricius, old, worn wings, molting tail, and moderate molt in the head and neck. The bird appears to be well along in a prebasic molt and has head and neck with mostly basic feather- ing. We believe the full basic plumage of the head and neck would show no traces of white in the cheeks and collar and the postocular streak either absent or very short. The juvenal plumage of the head and neck of both sexes (Fig- ure IE) has a dark brown cap extending to the nape, medium to dark reddish-brown cheeks and lores, lighter brown sides of neck and collar at base of neck, and an indistinct patch of medium cinnamon on the throat. The bill is blackish-gray with a broad, light bluish area bordering the black nail. The tarsus and foot are dull tan with blackish patches on the joints and blackish webs. Most of the predefinitive plumages are unknown. A young male molting into what appears to be first alternate plumage has a dark cap extending to the nape, light brown cheeks becoming darker in the lores, medium cinnamon on the throat, white collar at the base of the neck, and a white postocular streak that broadens 10 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY posterior to the auricular region. We judge that the first alternate plumage of males is similar to the alternate plumage of females. This is the first of several papers reporting the results of field and laboratory studies now in progress designed to characterize the taxa of steamer-ducks and their phylogenetic relationships and to finish testing hypotheses relating to the origins and adaptive sig- nificance of flightlessness in the genus. SUMMARY A new species of flightless steamer-duck, Tachyeres leuco- cephalus, is described from the coast of the Province of Chubut, Argentina. The new species is much smaller than T. pteneres and approximately the same size as T. brachypterus of the Falkland Islands from which it differs in aspect and certain osteological _ characteristics. RESUMEN Una nueva especie de pato vapor no volador ha sido descubierto en la costa de la Provincia de Chubut, Republica de Argentina. La nueva especie es Tachyeres leucocephalus y tiene aproximadamente el mismo tamafio de T. brachypterus de las Islas Malvinas del que se distingue por algunas caracteristicas osteologicas. ACKNOWLEDGEMENTS Our field research in Argentina would not have been possible without the generous assistance of many people and organizations. We are grateful to the following: B. Mayer, Mr. and Mrs. F. V. T. J. Fauring and associates (Patagonia Comercial); Dr. O. Kuhnemann, J. Sesti and P. Medina (Centro de Investigacién de Biologia Marina); Dr. J. M. Gallardo, A. Bockel, and Dr. J. Navas (Museo Argentino de Ciencias Naturales “Bernardino Rivadavia”); Dr. E. O. Gonzalez Ruiz and colleagues (Direccién Nacional de Fauna Sylvestre); Ing. Qca. L. O. Saigg de Chialva (Directora de Proteccién Ambiental, Provincia de Chubut); Dr. D. H. Soria (Director General de Ganaderia, Prov- incia de Santa Cruz); Dr. A. Tarak ( Direccién Nacional de Parques Nacionales); Dr. G. A. Giaroli (Subsecretario de Recursos Naturales Renovables y Ecologia de la Nacion); F. Loébbe and F. Villar (Ceremonial, Ministerio de Economia); Sr. I. Torres ( Braniff Inter- national); A. A. Gonzalez (Aerolineas Argentinas); P. Canevari, W. Conway, F. B. Cross, B. de Ferradas, F. Erize, the Gibson family, R. Hamilton, B. C. Livezey, R. M. and M. Mengel, L. A NEW SPECIES OF STEAMER-DUCK 11 Orquera, E. Piana, M. A. E. Rumboll, R. Stranek; M. L. Humphrey for her patience and enthusiastic support; R. M. Mengel for pre- paring the color plate and D. Bennet the line drawing; R. M. Mengel and anonymous reviewers for suggesting imrovements in the manu- script; the authorities of the American Museum of Natural His- tory, British Museum (Natural History), Field Museum of Nat- ural History, Museo Argentino de Ciencias Naturales “Bernardino Rivadavia” (MACN), Museo de La Plata, Museum of Comparative Zoology, National Museum of Natural History, Southwestern Col- lege Museum of Natural History (SC), and the University of Michigan Museum of Zoology for enabling us to examine speci- mens in their care; the authorities of the University of Kansas for awarding Humphrey a half-year sabbatical and subsequent re- search leave; the administration of Southwestern College for pro- viding Thompson with a half-year research leave; the Kansas Uni- versity Endowment Association, Southwestern College, Dr. and Mrs. W. Saul, Prof. M. C. Thompson, Mr. and Mrs. L. A. Osborne, Dr. R. T. Peterson, Dr. T. Mastin, Mr. R. Hamilton and the Humphrey family for funds that made the 1979 field research possible. This study was partially supported by National Science Founda- tion grant no. DEB-8012403. LITERATURE CITED Buaxe, E. R. 1977. Manual of Neotropical Birds. Vol. 1, Spheniscidae (Penguins) to Laridae (Gulls and Allies). The University of Chicago Press xix + 674. Bo, N. A. 1958. Nota sobre una coleccién de aves del este de Chubut. Rev. del Museo de La Plata (Nueva Serie). Secc. Zool., VII:35-50. Boswa.t, J. AND Maclver, D. 1979a. Nota sobre el Pato Vapor Volador (Tachyeres patachonicus). Hornero, 12:75-78. Boswatt, J., AND MaclIver, D. 1979b. Casual Notes on the Flying Steamer Duck. Original English version of Ms. published in Spanish as 1979a; Ms. available to bona fide investigators from Boswall, Birdswell, Wraxall, Bristol, BS19 1JZ, England. Boswatu, J. AND Pryruercu, R. J. 1972. Some notes on the birds of Point Tombo, Argentina. Bull. Brit. Orn. Cl., 92:118-129. Dacrux, J. 1977. Notas faunisticas y bioecologicas de Peninsula Valdez y Patagonia. VI. Observaciones sobre areas de nidificacién de la avifauna del litoral maritimo patagénica (Provincias de Chubut y Santa Cruz, Rep. Argentina). Hornero, 11:361-376. Humpurey, P. S. anp Livezey, B. C. In press. Notes on molts and plumages of flying steamer ducks. Jen, J. R. Jn, Rumpowr, M. A. E., anp Winter, J. P. 1973. Winter bird populations of Golfo San Jose Argentina. Bull. Brit. Orn. Cl., 93:56-63. Jounscarp, P. 1979. Family Anatidae, in Check-list of Birds of the World, Vol. I, Second Edition. Revision of the work of James L. Peters. Mayr, E. and Cottrell, G. W. (eds.) Museum of Comparative Zoology, Cam- bridge, Mass. Pp. xvii + 547. 12 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY Murpny, R. C. 1936. Oceanic birds of South America, Vol. II, American Mus. Nat. Hist., New York. Pp. 641-1245. Oxroc, C. 1979. Nueva lista de la avifauna Argentina. Opera Lilloana, 27:1-297. Scott, W. E. D. Anp SHaArpe, R. B. 1904-1915. Reports of the Princeton University Expeditions to Patagonia, 1896-1899. Vol. II. 1, Omithology. Pp. xii + 504. Topp, F. S. 1979. Waterfowl: Ducks, Geese and Swans of the World. Harcourt-Brace Jovanovich, New York. WELLER, M. W. 1976. Ecology and behaviour of steamer ducks. Wildfowl 27:45-53. ‘ Zapata, A. R. P. 1967. Observacidnes sobre aves de Puerto Deseado, Provincia de Santa Cruz. Hornero, 10:351-378. ZapaTA, A. R. P. 1969. Aves observadas en el Golfo San Jorge, Provincias de Chubut y Santa Cruz, Argentina. Zoologia Platense, 1:21-27. UNIVERSITY OF KANSAS PUBLICATIONS MUSEUM OF NATURAL HISTORY The University of Kansas Publications, Museum of Natural History, beginning with volume 1 in 1946, was discontinued with volume 20 in 1971. Shorter research papers formerly pub- lished in the above series are now published as Occasional Papers, Museum of Natural History. The Miscellaneous Publi- cations, Museum of Natural History, began with number 1 in 1946. Longer research papers are published in that series. Monographs of the Museum of Natural History were initiated in 1970. All manuscripts are subject to critical review by intra- and extramural specialists; final acceptance is at the discretion of the publications committee. Institutional libraries interested in exchanging publications may obtain the Occasional Papers and Miscellaneous Publications by addressing the Exchange Librarian, The University of Kansas Library, Lawrence, Kansas 66045. Individuals may purchase separate numbers of all series. Prices may be obtained upon re- quest addressed to Publications Secretary, Museum of Natural History, The University of Kansas, Lawrence, Kansas 66045. Editor: E. O. WiLEy Managing Editor: Jos—EpH T. CoLLins PRINTED BY UNIVERSITY OF KANSAS PRINTING SERVICE LAWRENCE, KANSAS Sgt 5 TNE go area at a —— SS = es e SF oh i ’ Acme Bookbinding Co., Inc. tune 3 2044 ll Il 681 oe o< fee MANES Gog eee ipo? POR e a there ae OF WAN Een cee a > Oe et WA me. Cae eee ae en Powis ny oot ee + Peers Ooo e ap ‘ pe Si tn ati oe eS © wu & =. APN EE He ca ttn ee Devt ~~ Weng a4 ee Ce ear a ae Sin ee er te FO ere ar te haere pte at ah ht ett A Wee trees BNE be ae a s@ ean