HARVARD UNIVERSITY Library of the Museum of Comparative Zoology ANNALS OF CARNEGIE MUSEUM VOLUME 46 February 3, 1976-December 9, 1977 PUBLISHED BY THE AUTHORITY OF THE BOARD OF TRUSTEES OF THE CARNEGIE INSTITUTE PITTSBURGH, PENNSYLVANIA 1977 o CO ID 3H ^ 0) -"0 • 00 rH 00 CD 0H,Q ^ 0) 13 • rd 3 to P< o I Q) rC (0 CP g 0 ro fil rd £0 O 0J (0 H O rd U m >. H g rfl 0) "O _ c 1rd rd P 3 P H O m a <* »■■■ ttr ■>*?- *?'i Publications Committee HARRY K. CLENCH C. J. McCOY, JR. DUANE A. SCHLITTER R. E. PORTEOUS, Editor V CONTENTS Contents . v New taxa . vii Author index . viii ARTICLE 1. The lower antemolar teeth of Litolestes ignotus, a Late Paleocene erinaceid (Mammalia, Insectivora). Jeffrey H. Schwartz and L. Krishtalka . 1 2. North American Nyctitheriidae (Mammalia, Insectivora). Leonard Krishtalka ... 7 3. A new species of Rhyacophila (Trichoptera, Rhyacophilidae) from Western Pennsylvania. Jan L. Sykora and John S. Weaver III . 29 4. Three new species of Sphaerodactylus (Sauria: Gekkonidae) from Hispaniola. Richard Thomas and Albert Schwartz . 33 5. The rock vole, Microtus chrotorrhinus (Miller) (Mammalia: Rodentia) in West Virginia. Gordon L. Kirkland, Jr . .45 6. Revision of Picrodontidae (Primates, Plesiadapiformes): dental homologies and relationships. Jeffrey H. Schwartz and Leonard Krishtalka . 55 7. Paleontology and geology of the Bad water Creek area, central Wyoming. Part 13. The late Eocene Insectivora and Dermoptera. Leonard Krishtalka and Takeshi Setoguchi . 71 8. A new subgenus and new species of Miocene Calliostoma (Archaeogastropoda — Trochidae). J. J. Parodiz . 101 9. Paleospecies of neotropical ampullariids and notes on other fossil non-marine South American gastropods. Kenneth J. Boss and Juan J. Parodiz . 107 10. A new petalodont (Chondrichthyes, Bradyodonti) from the Upper Mississip¬ pi of Montana. Richard Lund . 129 11. The Pleistocene (Kansan) herpetofauna of Cumberland Cave, Maryland. J. Alan Holman . 157 12. A list of the butterflies of Andros, Bahamas. Harry K. Clench . 173 13. Echinochimaera meltoni, new genus and species (Chimaeriformes), from the Mississippi of Montana. Richard Lund . 195 VI Annals of Carnegie Museum 14. Nongeographic variation in elephant shrews (genus Elephantulus Thomas and Schwann, 1906) of southern Africa. I. L. Rautenbach and Duane A. Schlitter . 223 15. Morphometric variation in the tropical pocket gopher (Geomys tropicalis). Stephen L. Williams and Hugh H. Genoways . 245 16. Butterflies of the Carnegie Museum Bahamas Expedition, 1976. Harry K. Clench . 265 17. Zoogeography of amphibians and reptiles in Cadiz Province, Spain. Stephen D. Busack . 285 vii NEW TAX A DESCRIBED IN VOL. 46 NEW SPECIES tCalliostoma (Tropidotrochus) jayae, new species. Mollusca, Gastropoda . 102 tCryptobranchus guildayi, new species. Amphibia, Caudata . 159 fEchinochimaera meltoni, new species. Chondrichthyes, Holocephali . 198 fHeteropetalus elegantulus, new species. Chondrichthyes, Elasmobranchii . 136 tLeptacodon catulus, new species. Mammalia, Insectivora . 13 fMacrocranion robinsoni, new species. Mammalia, Insectivora . 73 fNyctitherium christopheri, new species. Mammalia, Insectivora . 80 fPomacea (Effusa) pattersoni, new species. Mollusca, Gastropoda . 112 tPomacea (Pomacea) prourceus, new species. Mollusca, Gastropoda . 110 Rhyacophila pennsylvanica, new species. Insecta, Trichoptera . 30 Sphaerodactylus cryphius, new species. Reptilia, Sauria . 39 Sphaerodactylus nycteropus, new species. Reptilia, Sauria . 41 Sphaerodactylus streptophorus, new species. Reptilia, Sauria . 34 NEW GENERA AND SUBGENERA fEchinochimaera, new genus. Chondrichthyes, Holocephali . 198 fHeteropetalus, new genus. Chondrichthyes, Elasmobranchii . 135 fTropidotrochus, new subgenus. Mollusca, Gastropoda . 101 HIGHER CATEGORIES fEchinochimaeridae, new family. Chondrichthyes, Holocephali . 198 fEchinochimaeroidei, new suborder. Chondrichthyes, Holocephali . 198 fFossil taxa vm AUTHOR INDEX Boss, Kenneth J . 107 Busack, Stephen D . 285 Clench, Harry K . 173, 265 Genoways, Hugh H., co-author . 245 Holman, J. Alan . 157 Kirkland, Gordon L., Jr. . 45 Krishtalka, L.; see Krishtalka, Leonard. Krishtalka, Leonard . 7 Krishtalka, Leonard, and Takeshi Setoguchi . 71 Krishtalka, Leonard, co-author . 1, 55 Lund, Richard . 129, 195 Parodiz, J. J.; see Parodiz, Juan J. Parodiz, Juan J . 101 Parodiz, Juan J., co-author . 107 Rautenbach, I. L., and Duane A. Schlitter . 223 Schlitter, Duane A., co-author . 223 Schwartz, Albert, co-author . 33 Schwartz, Jeffrey H., and Leonard Krishtalka . 1, 55 Setoguchi, Takeshi, co-author . 71 Sykora, Jan L., and John S. Weaver III . 29 Thomas, Richard, and Albert Schwartz . 33 Weaver, John S. Ill, co-author . 29 Williams, Stephen L., and Hugh H. Genoways 245 \S-/}ft-?lLfciur 137 mm, Hall and Cockrum, 1953) revealed that males averaged larger than females for all measurements except hind foot length (Table 1). The differences were significant for total, tail, and body lengths (Stu¬ dent’s t-test, p < .05). Kellogg (1937) described the same size relation¬ ships between the sexes for total, tail, and hind foot lengths of nine male and six female West Virginia specimens. Martin (1971) found that in New Hampshire populations males averaged larger than females for total, tail, and hind foot lengths. Kellogg (1937) measured 11 skull characters in a West Virginia sample of 10 adults of each sex. He found no consistent size differences between the sexes. Males averaged larger for five characters. Females were larger for four. In the present study, the same 1 1 characters were measured on a sample of 10 adult skulls of each sex. Males averaged larger for five characters. Females averaged larger for six (Table 2). None of the differences were significant (Student’s t-test, p > .05). However, in an analysis of New Hampshire specimens, Martin (1971) found that males averaged larger than females in 14 of 15 skull char¬ acters examined, the difference being significant for six. In addition to revealing the presence of more pronounced sexual dimorphism for size in New Hampshire populations, Martin’s data also indicate that M. chrotorrhinus are larger in New Hampshire than in West Virginia. Reproduction Specimens in this study were collected in July 1974-1975 and October 1973-1975. The sex ratio of the sample was 1.066V: 1.009 9. Of 12 adult females captured in July, eight were reproductively active with six pregnant. One of four July-caught non-adult females (total length =131 mm) was pregnant with four embryos. The previous record for smallest reproductively active female was 132 mm (Martin, 1971). In the October sample of 13 adult females, only two were sexually active, including one pregnant female captured on 12 October, 1974. 48 Annals of Carnegie Museum vol. 46 Martin’s (1971) latest record of female reproductive activity was that of a lactating female on 12 October, 1950 in the Great Smoky Mountains National Park. The mean embryo count for the eight pregnant females in this study was 2.88 (Range 2-5; Mode 3). This mean value for West Virginia specimens of M. chrotorrhinus was smaller than those exhibited in populations in the northern portion of the species’ range. For example, eleven pregnant females from the province of Ontario (ROM, 8; NMC, 2; UIMNH, 1) had a mean embryo count of 3.91 (Range 2-5; Mode 4). This value was significantly greater than the mean for the eight West Virginia specimens (Student’s t = 2.71, p < .02). Martin’s (1971) exam¬ ination of embryo counts from throughout the range of M. chrotorr¬ hinus revealed a latitudinal cline for embryo counts with a general increase at higher latitudes. Nine of 1 1 July-caught males were considered to be sexually active with testes measurements ranging from 8x3 mm to 13x9 mm (X = 1 1.6 x 7.9 mm). Martin (1971) generally found active sperm in the epididymi of males when the testes length exceeded 9 mm. In October, three of 10 males appeared to be sexually active. Thus, data for both sexes reveal potential reproductive activity in October, although apparently at a reduced level compared to July. Habitat and Ecological Associations Published descriptions of the habitat of M. chrotorrhinus from throughout its range emphasize the importance of rocks, boulders, and talus as habitat components (Burt, 1957; Grimm and Whitebread, 1951; Gunderson and Beer, 1953; Hamilton, 1943; Martin, 1971; Osgood, 1938; Peterson, 1966; Roslund, 1951; and Wyman, 1923). In West Virginia, M. chrotorrhinus has been recorded in beech-maple-oak forests with mossy, rocky floors near Cranberry Glades (Kellogg, 1937). During the West Virginia Mammal Survey, they were also trapped among large moss-covered rocks in red spruce-northern hardwood forests, as well as in and near forest streams and springs (McKeever, 1952). In the present study, 23 of 73 M. chrotorrhinus were captured in natural habitats similar to those previously described for this species: characteristically cool, moist, rocky or boulder-strewn forests (Table 3). It is of special significance, however, that 50 of the specimens were trapped on clearcuts in both northern coniferous and mixed deciduous forests (Table 3). This represents the first record of M. chrotorrhinus inhabiting this disturbed habitat. In fact, at three sites in Pocahontas and Randolph Counties, M. chrotorrhinus were more than twice as abundant on recent red spruce and mixed red spruce-deciduous clear¬ cuts (< 3 yrs. old) than in adjacent uncut stands (2.23 catches/ 100 trapnights; total 1618 trapnights vs. 0.95 catches/ 100 trapnights; total 1977 The Rock Vole 49 1580 trapnights). These data indicate that M. chrotorrhinus were actively exploiting these recently disturbed sites, and therefore may benefit from the forest openings created by clearcutting. This con¬ clusion stands in marked contrast to Lowman’s (1975) management recommendations for M. chrotorrhinus in the Southern Appalachians, which include retaining forests in an undisturbed state; restricting clearcutting from mountain-top forests and from along mountain streams; and permitting mountainous forests to attain the latter stages of ecological succession, although keeping the high balds open. M. chrotorrhinus were relatively more abundant on the red spruce clearcuts, where they comprised 18.1% (44 of 243 specimens) of the small mammals collected, than on deciduous or mixed deciduous- coniferous clearcuts, where they represented only 3.0% (6 of 202 speci¬ mens) of the small mammals trapped. In general, they were more abundant at higher altitudes. Sixty-four were captured above 4000 feet in 5656 trapnights, whereas only nine were captured in 9069 trapnights at altitudes ranging from 2000-3500 feet. This greater abundance at higher altitudes, with accompanying boreal habitats, is expected considering the center of distribution of M. chrotorrhinus in eastern Canada and its generally acknowledged association with the Canadian life zone habitats. Thirteen other species of small mammals were captured on the 12 trapping grids yielding M. chrotorrhinus (Table 4). The red-backed vole, Clethrionomys gapperi (Vigors) and the cloudland deer mouse, Peromyscus maniculatus nubiterrae Rhoads, invariably were trapped on the same sites as M. chrotorrhinus. Other small mammals captured in at least 75% of the grids yielding M. chrotorrhinus were the masked shrew, Sorex cinereus Kerr, the smoky shrew, Sorex fumeus Miller, and the short-tailed shrew, Blarina brevicauda (Say). Grimm and Whitebread (1951) observed that in northeastern Pennsylvania, M. chrotorrhinus were invariably trapped in the same habitats as C. gapperi, with the big-tailed shrew, Sorex dispar Batchelder, S’, cinereus, and S. fumeus often among its associates. In addition to C. gapperi, two other microtines, the meadow vole, Microtus pennsylvanicus (Ord), and the southern bog lemming, Synaptomys cooperi Baird, were captured at the same sites as M. chrotorrhinus. At three sites, 19 M. pennsylvanicus were trapped with 1 1 M. chrotorrhinus. On four grids, M. chrotorrhinus exceeded S. cooperi in abundance, 32 to 5. The ratio of C. gapperi to M. chrotorr¬ hinus on the 12 grids was 2.83:1.00 (195:69). Martin (1971) has collected a relatively unchanging proportion of three C. gapperi for every M. chrotorrhinus in favorable rock vole habitat. In this study, | there was considerable variation in relative abundance of these two species on individual grids, ranging from 13:8 in favor of M. chro- 50 Annals of Carnegie Museum vol. 46 torrhinus to 20:1 in favor of C. gapperi. M. chrotorrhinus were more abundant than C. gapperi on three plots (33 to 23), and the abundance of C. gapperi was greater on nine (172 to 36). This variation may be a reflection of the individual species’ responses to the diverse habitat conditions prevailing on the nine grids in forest clearcuts. Specimen Localities and Literature Records Greenbrier county: Richwood, 4.2 mi NE, 2750' (McKeever, 1952). pendleton county: Circleville, 3.1 mi NW, 4700' (ibid.). Pocahontas county: Cheat Bridge (Randolph Co.), 6.0 mi S, 4400' (16); Cranberry Glades; Mill Point (Kellogg, 1937). RANDOLPH county: Cheat Bridge, 2.2 mi ENE, 4200' (1); 2.8 mi ENE, 4100' (3); 5.6 mi SW, 4600' (44); Durbin (Pocahontas Co.), 5 mi NNW 4350' (McKeever, 1952); Elkins (Martin, 1971); Parsons (Tucker Co.), 5.7 mi S, 3200' (3); 8.0 mi S, 3500' (1). TUCKER county: Parsons, 2.6 mi S, 2000' (4); 2.9 mi SE, 2650' (1). Table 1 . Univariate statistical analysis, external measurements of adult Microtus chrotorrhinus. External measurements N Mean S.E. S.D. C.V. Range t-ratio Males : Total Length 25 157.36 2.10 10.50 1.33 140-175 2.69** Tail Length 25 43.42 .90 4.43 2.08 37-51 2.43** Body Length 25 114.12 1.48 7.40 1.30 98-128 2.21* Hind Foot 25 19.90 .22 1.12 1.12 18-21.5 1.03 Ear Length 25 13.94 .31 1.53 2.19 11-17 0.42 Weight 25 35.84 1.37 6.83 3.81 25.7-44.2 0.83 Females : Total Length 28 150.41 1.57 8.30 1.04 137-165 Tail Length 28 40.46 .81 4.32 2.02 31-49 Body Length 28 109.95 1.20 6.37 1.10 100-120 Hind Foot 28 20.20 .11 .98 .92 19-22 Ear Length 28 13.77 .27 1.42 1.95 11-17 Weight 28 34.27 1.34 6.97 3.91 24.6-50.9 *Significant at p < .05 **Significant at p < .02 1977 The Rock Vole 51 Table 2. Univariate statistical analysis, 11 skull characters of adult Microtus chrotorrhinus. Skull character N Mean S.E. S.D. C.V. Range t-ratio Males : Greatest length of skull 10 26.72 .23 .74 .87 25.6-28.1 .23 Condylobasal length 10 26.24 .24 .77 .93 25.0-27.7 .22 Least interorbital breadth 10 3.91 .03 .10 .82 3.75-4.05 .91 Nasal length 10 7.62 .13 .41 1.70 7.15-8.6 .62 Palatal length 10 13.18 .15 .48 1.16 12.5-14.15 .52 Mastoid breadth 10 12.57 .13 .42 1.06 12.0-13.25 1.40 Zygomatic width 10 14.88 .17 .53 1.12 14.2-16.0 .37 Basilar length 10 23.46 .25 .78 1.05 22.2-23.7 .17 Maxillary tooth row 10 6.55 .08 .27 1.29 6.05-6.9 .67 Cranial height 10 9.74 .12 .38 1.24 9.2-10.15 .72 Mandible length 10 15.09 .12 .38 .80 14.35-15.55 1.06 Females : Greatest length of skull 10 26.79 .17 .54 .64 25.85-27.5 Condylobasal length 10 26.17 .15 .49 .60 25.25-26.75 Least interorbital breadth 10 3.86 .04 .12 .98 3.7-4. 1 Nasal length 10 7.73 .13 .41 1.69 7.05-8.45 Palatal length 10 13.08 .12 .37 .89 12.7-13.95 Mastoid breadth 10 12.85 .14 .46 1.12 12.35-13.9 Zygomatic width 10 14.96 .12 .37 .79 14.45-15.55 Basilar length 10 23.41 .16 .51 .69 22.65-24.3 Maxillary tooth row 10 6.47 .06 .20 .95 6.2-6.8 Cranial height 10 9.86 .09 .30 .96 9.5-10.25 Mandible length 10 15.32 .17 .55 1.14 14.6-16.05 Table 3. Distribution of capture, 73 Microtus chrotorrhinus in natural and disturbed habitats. Habitat Description Number Caught A. Undisturbed Habitats: 1. Uncut red spruce, Picea rubens 14 2. Uncut mixed red spruce — northern hardwoods {Be tula, Acer ) 3 3. Stream bed in mixed hardwoods with Rhododendron 3 4. Mixed deciduous forest 3 23 B. Disturbed Habitats: 1. Recent red spruce clearcuts, < 3 years old 41 2. Recent northern hardwoods clearcut, < 3 years old 5 3. Older deciduous clearcut, 6-8 years old 1 4. Older red spruce clearcut, 7 years old 3 50 52 Annals of Carnegie Museum vol. 46 Table 4. Frequency of association of other small mammals with Microtus chrotorrhinus on 12 traplines. No. No. Species Traplines Individuals Masked Shrew, Sorex cinereus Kerr 9 62 Smoky Shrew, Sorex fumeus Miller 11 47 Long-tailed Shrew, Sorex dispar Batchelder 3 7 Short-tailed Shrew, Blarina brevicauda (Say) 10 47 Chipmunk, Tamias striatus (L.) 3 3 Southern Flying Squirrel, Glaucomys volans (L). 1 1 Northern Flying Squirrel, Glaucomys sabrinus (Shaw) Cloudland Deer Mouse, Peromyscus maniculatus 1 1 nubiterrae Rhoads 12 137 Eastern Wood Rat, Neotoma floridana (Ord) 1 1 Red-backed Vole, Clethrionomys gapperi (Vigors) 12 195 Meadow Vole, Microtus pennsyhanicus (Ord) 3 19 Rock Vole, Microtus chrotorrhinus (Miller) 12 69 Southern Bog Lemming, Synaptomys cooperi Baird 4 5 Woodland Jumping Mouse, Napaeozapus insignis (Miller) 5 27 References Cited Burt, W. R. 1957. Mammals of the Great Lakes Region. The University of Michigan Press, Ann Arbor. 246 pp. DeBlase, A. F. and R. E. Martin 1974. A Manual of Mammalogy. Wm. C. Brown Co., Dubuque, Iowa. 329 pp. Grimm, W. C. and R. Whitebread 1951. Mammal Survey of Northeastern Pennsylvania. Pennsylvania Game Commis¬ sion, Harrisburg. 82 pp. Gunderson, H. L. and J. R. Beer 1953. The Mammals of Minnesota. The University of Minnesota Press, Minneapolis. 190 pp. Hall, E. R. and E. L. Cockrum 1953. A synopsis of the North American microtine rodents. Univ. of Kansas Publ., Mus. Nat. Hist., 5(27):373-498. Hall, E. R. and K. R. Kelson 1959. The Mammals of North America, Vol. I & II. Ronald Press, New York. 1083 pp. Hamilton, W. J., Jr. 1943. The Mammals of the Eastern United States. Comstock Publ. Co., Ithaca, N.Y., 432 pp. Kellogg, R. 1937. Annotated checklist of West Virginia mammals. Proc. U.S. Nat. Mus., 84(3022):443-479. Komarek, E. V. 1932. Distribution of Microtus chrotorrhinus, with description of a new subspecies. Jour. Mammalogy, 13:155-158. 1977 The Rock Vole 53 Lowman, G.E. 1975. A survey of endangered, threatened, rare, status undetermined, peripheral, and unique mammals of the Southeastern National Forests and Grasslands. U.S. Dept, of Agr., Forest Service, Atlanta, Georgia. 132 pp. Martin, R. L. 1965. The yellow nosed vole. Mount Washington Observatory News Bulletin. 2 pp. Martin, R. L. 1971. The natural history and taxonomy of the rock vole, Microtus chrotorrhinus. Ph.D. Dissertation, University of Connecticut, Storrs. 164 pp. Martin, R. L. 1972. Parasites and diseases of the rock vole. Occas. Papers, University of Connecti¬ cut, Biol. Ser., 2(8): 107-1 13. Martin, R. L. 1973a. The dentition of Microtus chrotorrhinus (Miller) and related forms. Ibid., 2(1 2): 183-201. Martin R. L. 1973b. Molting in the rock vole, Microtus chrotorrhinus. Mammalia, 37(2):342-347. McKeever, S. 1952. The Survey of West Virginia Mammals. Conservation Comm. West Virginia, Pittman-Robertson Project 22-TC, 126 pp. — mimeographed January 4. Osgood, F. L., Jr. 1938. First Vermont record for the rock vole. Jour. Mammalogy, 19:108. Peterson, R. L. 1966. The Mammals of Eastern Canada. Oxford University Press, Toronto. 465 pp. Roslund, H. R. 1951. Mammal Survey of Northcentral Pennsylvania. Pennsylvania Game Commis¬ sion, Harrisburg. 55 pp. Wilson, L. W. and J. E. Friedel 1941. A list of mammals collected in West Virginia. Proc. W. Va. Acad. Sci., 15:85-92. Wyman, L. C. 1923. Microtus chrotorrhinus in Maine. Jour. Mammalogy, 4:125-126. Back issues of many Annals of Carnegie Museum articles are available, and a few early complete volumes and parts are listed at half price. Orders and inquiries should be addressed to: Publications Secretary, Carnegie Museum, 4400 Forbes Avenue, Pittsburgh, Pa. 15213. MUS. I«*fcfl<)flS'44£fe©L, LIBRARY JUL7 1977 ANNALS of CARNEGIE MUSEOW CARNEGIE MUSEUM OF NATURAL HISTORY HARVARD TV 4400 FORBES AVENUE * PITTSBURGH, PENNSYLVANIA 15213 VOLUME 46 jUNE 24/ 1977 ARTICLE 6 REVISION OF PICRODONTIDAE (PRIMATES, PLESIADAPIFORMES): DENTAL HOMOLOGIES AND RELATIONSHIPS Jeffrey H. Schwartz1 Research Associate Leonard Krishtalka Section of Vertebrate Fossils Abstract The Paleocene (Torrejonian-Tiffanian) North American primates Picrodus and Zanyc- teris are re-studied and compared to other plesiadapiforms. It is suggested that the teeth traditionally designated as Ml/ml, M2/m2, M3/m3 in picrodontids are homolo¬ gous with the ultimate premolar (P5/p5), Ml/ml, and M2/m2, respectively, of other primates. M3/m3 has been lost. In light of recent re-interpretations of plesiadapiform and tarsiiform dentitions, the dental complement of picrodontids is identified as a canine, five premolars, and two molars. Within plesiadapiforms, Picrodus and Zanycteris are most closely related to Phenacolemur. This clade forms a sister-group with the paromomy- ines — relationships reflected in a reclassification of the Paromomyidae. Introduction Recent discoveries and descriptions of eutherian mammals with five premolars have necessitated a review of mammalian dental homologies and mammalian phylogeny. Lillegraven (1969), Clemens (1973) and McKenna (1975) have reported the presence of five premolars in the Cretaceous Asian palaeoryctid Kennalestes, the Cretaceous North American leptictid Gypsonictops, and the Jurassic pro-tribosphenic genus Peramus. Partly on the basis of these findings, McKenna’s (1975) initial re-interpretation of dental homologies among mammals has led 'Department of Anthropology, University of Pittsburgh, Pittsburgh, Pa. 15260. Submitted for publication August 17, 1976. 55 56 Annals of Carnegie Museum vol. 46 to a new recontruction of mammalian phylogeny and classification. An additional reason for re-evaluating dental homologies has been the recent analyses of mammalian dental development that question the traditional views of fixed tooth loci and transmutation of tooth mor¬ phologies (Osborn, 1973; Schwartz, MS). Subsequent work has suggested the occurrence of five premolars in various plesiadapiform and tarsiiform primates, Litolestes- like erin- aceids, Plagiomene- like dermopterans, and possibly the adapisoricid Ankylodon and the nyctitheriid Leptacodon tener (Krishtalka, 1976a, 1976b; Schwartz, MS; Schwartz and Krishtalka, 1976). Of the recog¬ nized plesiadapiform taxa, the picrodontids have not been studied in this regard. These studies have implied that the occurrence of five pre¬ molars is primitive for Tokotheria (McKenna, 1975) and thus Primates (Schwartz, MS). The dental formula that is primitive for plesiadapiform-tarsiiform primates has been identified as Cl /cl, dPl/dpl, P2/p2, dP3/p3, P4/p4, P5/p5, Ml/ ml, M2/m2, M3/m3 (Schwartz, MS). The presence of deciduous or permanent teeth at the premolar loci is, however, still uncertain. In light of these conclusions, the dentition and relationships of the picrodontid primates and other plesiadapiforms are here re¬ examined and emended. The Paleocene (Torrejonian-Tiffanian) Picrodontidae includes two genera: Picrodus Douglass, 1908, and Zanycteris Matthew, 1917. Previous descriptions of picrodontid remains (Simpson, 1937; McGrew and Patterson, 1962; Szalay, 1968, 1972) emphasized the unique occlu¬ sal patterns and extreme specialization of the first of the three molars. Indeed, the seeming occurrence of a specialized Ml/ml in picrodon¬ tids is singular among plesiadapiform-tarsiiform primates, since the ultimate premolar is the tooth usually specialized in the latter. Our re¬ examination of the available picrodontid material indicates that (1) the specialized alleged Ml /ml are homologous with the ultimate premolars (P5/p5) of other plesiadapiforms; (2) the traditionally termed Ml-3/ml- 3 of picrodontids are more properly identified as P5/p5, Ml-2/ml-2; and (3) M3/m3 in picrodontids has been lost. Description and Remarks The discussion of the lower dentition of picrodontids applies only to Picrodus. Lower teeth of Zanycteris are not known (Szalay, 1968). Fig. 1. Picrodus silberlingi, SEM stereo-micrographs. A. AMNH 35453, partial left dentary, with p5mlm2, Gidley Quarry, Montana, approx. X 16. B. AMNH 89502, left dentary with c, p5 and alveoli for dpi, p2, dp3, p4, ml, Swain Quarry, Wyoming, approx. X 15. C. AMNH 89505, partial right dentary with p4p5 and alveolus for dp3, Swain Quarry, approx. X 13. 1977 Revision of Picrodontidae 57 58 Annals of Carnegie Museum vol. 46 “ ml The trigonid on “ml” (Fig. 1) is not typically molariform, but more nearly premolariform. It is tiny, approximately one-fourth as long as the talonid, and very high relative to the latter. The three trigonid cusps are minute and crowded together. In contrast to ml of other plesiadapiforms, the paraconid is positioned anteromedially and the metaconid occurs on the posterolingual face of the protoconid, so that the trigonid is triangular, rather than square in occlusal outline. Ac¬ cordingly, Figures 1 and 2 in Szalay (1968:3) are misleading. The struc¬ ture of the talonid is also contrary to that of other plesiadapiform ml’s. The talonid is very long and tear-shaped, and the talonid basin is ex¬ tremely shallow and terminates posterolingually in a narrow trough. As a result, the posterior margin of the talonid is neither straight, nor broad, labiolingually. The opposite is true of the talonid on ml of other plesiadapiforms, e.g., that of Phenacolemur (Fig. 3). Relative to “m2” and “m3” of Picrodus, and ml of other plesiadapiforms, the tooth tra¬ ditionally designated as ml in the picrodontids is non-molariform, highly specialized, and more properly considered an ultimate premolar. A number of crenulations, of which some are enlarged and may repre¬ sent a hypoconid, mark the labial margin of the talonid. Two cuspules occur along the lingual margin: one anteriorly, at the base of the trigonid, and one posteriorly. These may (Szalay, 1968) or may not be homolog¬ ous to an entoconid and hypoconulid, respectively, As noted below, a hypoconulid does not occur on described “m2” and “m3” of Picrodus. “m2” “m3”: The trigonid on “m2” (Fig.l) is square in occlusal out¬ line, fully molariform, and resembles the trigonid of plesiadapiform ml’s, especially that of Phenacolemur. Similarly, “m3” of Picrodus closely resembles m2 of Phenacolemur and other plesiadapiforms in that the trigonid is compressed anteroposteriorly, the paraconid is closely appressed to the metaconid, and the talonid is not elongate or doubled. In contrast, m3 in plesiadapiforms is characterized by an elongate talonid with a large hypoconulid expansion. In Picrodus the trigonid on both “m2” and “m3” is much lower relative to the talonid than on “ml.” The talonids on “m2” and “m3” are approximately of equal size, broader than the trigonids, and possess relatively straight posterior margins, as on ml and m2 of Phenacolemur and other plesia¬ dapiforms. Two cuspules that occur at the entoconid region of the tal¬ onid on “m2” and “m3” may represent a divided entoconid, and as on “ml”, the hypoconid area bears a series of crenulations. A hypoconulid is absent on “m2” and “m3”, and all the cusps are small and nubbin¬ like, as on the lower molars of Phenacolemur. The teeth of picrodontids (Fig. 1) traditionally identified as ml (with a tiny, high, laterally compressed trigonid and a long, tear-shaped talonid), m2 (with a squared trigonid) and m3 (with an anteroposteriorly compressed trigonid and a non-elongate talonid), exhibit an unorthodox sequential morphology in comparison with ml -3 in Phenacolemur and 1977 Revision of Picrodontidae 59 Fig. 2. A. RP5M1M2, approx. X 13, B. LP5M1M2, approx. X 14; Zanycteris paleoconus, AMNH 17180, Mason Pocket, Colorado. C. Picrodus silberlingi, CM 18003 and CM 17999, RP5, RM1, Shotgun Locality, Wyoming, approx. X 14. SEM stereo-micrographs. 60 Annals of Carnegie Museum vol. 46 other plesiadapiforms. Compared to the latter, the last three lower teeth in picrodontids are morphologically more properly identified as a specialized ultimate premolar, followed by a first and second molar. “Ml” “M2” “M3”: The teeth in Picrodus and Zanycteris (Fig. 2) traditionally designated as Ml and M2 have been adequately described by McGrew and Patterson (1962) and by Szalay (1968). Compared to “M2”, “Ml” is “peculiar” (Szalay, 1968:20), extremely specialized, and mirrors the sequential morphological relationship between the specialized “ml” and molariform “m2” of Picrodus. “Ml” is much larger than “M2,” more nearly triangular in occlusal aspect, and bears stronger protocristae and cingula, and an expanded metastylar area. “M3” of Picrodus is not known. “M2” and “M3” of Zanycteris are similarly molariform, quadrate in occlusal outline, lack protocristae, and possess greatly reduced stylar areas. Among plesiadapiforms known to us, “M2” of Picrodus and Zanycteris and “M3” of the latter most closely resemble Ml and M2, respectively, of Phenacolemur. The trigon basins are similarly shallow, the cusps are conular nubbins rather than the bulbous cones of other plesiadapiforms (Simpson, 1955), the stylar areas are reduced, the protocristae are very weak, and a true hypocone is absent. “M3” in Zanycteris lacks the lingual postcingular expansion of M3 in Phenacolemur, but is virtually identical to the M2 of the latter in occlusal outline. As concluded for the lower dentition, the sequential morphology of the teeth traditionally referred to as M1M2M3 in picrodontids implies, in comparison to that of other plesiadapiforms (notably Phenacolemur), that they represent an ultimate upper premolar followed by a first and second molar. Thus, we suggest that the last three upper and lower teeth of picro¬ dontids are homologous to the ultimate upper and lower premolars, Ml /ml and M2/m2 of other plesiadapiforms. Indeed, this re-interpreta¬ tion brings the picrodontids into coincidence with other plesiadapiforms and tarsiiforms in which the ultimate premolar is the most specialized tooth in the dental complement. Dental Homologies The occurrence of five premolar loci (PI -5/ pi -5) is primitive for Pri¬ mates, and also Plesiadapiformes and Tarsiiformes (Schwartz, MS). Many plesiadapiforms (e.g., Plesiolestes, Paromomys, Palaechthon, Palenochtha) and tarsiiforms (e.g., omomyines, various anaptomor- phines, Tarsius upper jaw) retain six antemolar teeth: a caniniform tooth followed by five premolariform teeth (see Fig. 4). As argued elsewhere Fig. 3. Phenacolemur, redrawn from Simpson (1955). A. P. pagei, PU 14030, cp5mlm2m3. ► B. P.jepseni, AMNH 48005, P5M1M2M3. 1977 Revision of Picrodontidae 61 62 Annals of Carnegie Museum vol. 46 (Schwartz, in press) the caniniform tooth — the first tooth in the jaw — is a canine rather than an incisor, as previously suggested by Matthew (1915) for the anterior tooth of Tetonius (- Pseudotetonius Bown, 1974) and by McKenna (1963) for that of apatemyids. Similarly, the five pre- molariform teeth posterior to the canine we identify as premolars. The incisors have been lost in plesiadapiforms and tarsiiforms. These identifications of tooth types and homologies are based not only on distinctive morphology, but also on the evidence that tooth development and positioning precede and are independent of the ossi¬ fication of any associated bone and subsequent sutural differentiation (Kollar and Baird, 1971; Miller, 1971; Tonge, 1971). Therefore, identi¬ fication of tooth homologies cannot be founded solely on the tradi¬ tional concepts of fixed tooth position and occlusion, as exemplified by the following commonly accepted definitions: The upper canine occurs immediately posterior to the premaxillary-maxillary suture; the lower canine occludes in front of the upper; only incisors develop in the premaxilla. Acceptance of these generalizations as invariably appli¬ cable to all mammalian taxa may result in incorrect identification of dental homologies involving unwarranted suggestions of dental trans¬ mutations, e.g., incisors become caniniform, and incisors and canines become premolariform. As in other plesiadapiforms and tarsiiforms, the most anterior lower tooth of Picrodus (Fig. 1) is large, robust, and caniniform (Szalay, 1968, 1972) and is here identified as a canine. Similarly, we propose that of the loci posterior to this canine, the anterior five contained premolars, including the tooth usually designated as ml. As discussed above, the morphology of “Ml/ml” of picrodontids and their specialized nature relative to “M2/m2” and “M3/m3” imply that “Ml/ml” are ultimate premolars, or P5/p5, rather than molars. The comparative morphologies of “M2/m2” and “M3/m3” relative to other plesiadapiforms suggest that they are homologous to Ml /ml and M2/m2 of the latter, respec¬ tively. Thus M3/m3 are absent in Picrodus and Zanycteris. Perhaps the growth of the greatly enlarged occlusal area of P5/p5 of picrodontids may have crowded the developing molars and ultimately contributed to the inhibition of M3/m3 (Butler, 1939, 1963; Griineberg, 1952, 1963). Our study of the available material of Picrodus suggests the following lower dental formula: cdplp2dp3p4p5mlm2 (see Fig. 5). Although the lower dentition of Zanycteris is not known, the great similarity between the latter and Picrodus in known parts of the upper dentition implies a similar resemblance in the lower dentition. Indeed, recovery of addi¬ tional material may prove the two forms to be congeneric. Szalay (1968) has figured the only known palate of Zanycteris ( AMNH 17180) showing a large alveolus in a reconstructed premaxilla, a large tooth immediately behind the presumed premaxillary-maxillary suture, 1977 Revision of Picrodontidae 63 Canine 5 Premolars 3 Molars Fig. 4. Diagram of lower dentition primitive for plesiadapiform-tarsiiform primates as seen in Omomys carteri. The suggested dental homologies are cdplp2dp3p4p5mlm2m3. 5 mm T - 1 - 1 - T Fig. 5. Diagram of lower dentition of Picrodus silberlingi with suggested dental homologies. 64 Annals of Carnegie Museum vol. 46 2mm 1977 Revision of Picrodontidae 65 Fig. 7. A. Anterior part of palate of Zanvcteris paleoconus (AMNH 17180). B. Right side of palate, approx. X 14. C. Left side of palate, approx. X 14. Fig. 6. Szalay’s (A: 1968; B: 1972) reconstructions of the palate of Zanvcteris paleoconus (AMNH 17.180), redrawn to scale. 66 Annals of Carnegie Museum vol. 46 and alveoli for three double-rooted premolars anterior to P5 (Fig. 6A). In a subsequent discussion, Szalay (1972) figures a second alveolus in the reconstructed premaxilla (Fig. 6B). The large incomplete tooth crown, which Szalay figures behind the reconstructed premaxillary¬ maxillary suture, is glued atop an alveolus that appears too small to have contained a tooth of this size (Fig. 7). This crown may not be associated with this specimen. With the material at hand, it is impossible to discern with certainty the premaxillary-maxillary suture, or whether the entire premaxilla is preserved. There is no evidence of the presence of either one (Szalay, 1968) or two (Szalay, 1972) anterior alveoli or incisive foramina (Fig. 7 A). If AMNH 17180 does preserve the entire palate, the alveoli present indicate the occurrence of four teeth anterior to P5 (Fig. 7B,C). We suggest that of these tooth loci the posterior three contained premolars, possibly P2, dP3 and P4, as also seems the case in many other plesia- dapiforms (Schwartz, MS). The anteriormost tooth locus is separated from the others by a diastema and probably also contained a premolar. The edentulous part of the palate anterior to this tooth may have been covered by a cud pad, as in Lepilemur and presumed for Megaladapis (Tattersall, 1972), both of which lack teeth in the prexamilla. However, if Zanycteris did possess a dentulous premaxilla — not preserved on AMNH 17180 — it is likely that it bore a canine, as in other plesiadapi- forms (Schwartz, MS). Based on AMNH 17180, we suggest that Zanyc¬ teris possessed the following teeth in the upper jaw: ?CdPlP2dP3P4P 5M1M2. The same dental formula may also describe the upper dentition of Picrodus, although teeth anterior to P5 are unknown. Relationships Romer (1945) aligned the picrodontids with Insectivora, but later (Romer, 1966) referred this family to Dermoptera incertae sedis. Simpson (1937) and then McGrew and Patterson (1962) maintained that the affinities of picrodontids were probably with Primates or Insectivora. Van Valen (1965, 1967, 1969), as well as McKenna (1967) recognized the primate affinities of picrodontids, and most recently, Szalay (1968, 1969, 1972) aligned the latter with paromomyids — a conclusion with which we concur, but for different reasons, if the dental homologies discussed earlier are valid. The similarities between picrodontids and Phenacolemur also serve to distinguish both from other paromomyids (Simpson, 1955). On Ml /ml and M2/m2 of Phenacolemur the cusps are ill-defined nubbins that tend to be positioned marginally on the crown. The basins are broad and extremely shallow, and inter-cusp elaboration (cristae, con- ules) is minimal or non-existent. On Ml -2 of Phenacolemur the talon is greatly expanded in comparison with that of other paromomyids, and 1977 Revision of Picrodontidae 67 the teeth are less transverse in occlusal outline. On the lower molars the trigonids are more strongly compressed anteroposteriorly, and the para- conids are minute (Simpson, 1955). Although the same is characteristic of picrodontids in comparison with other paromomyids, the condition in picrodontids is more extreme in that cusp size is further diminished and inter-cusp cristae are absent. These dental features — unique among plesiadapiforms and tarsiiforms — may be shared derived or independently attained. At present, we opt for the former interpretation since the degree of uniqueness more probably implies a common ancestry for, and a sister group relationship between, picrodontids and Phenacolemur. Although they are closely related, picrodontids and Phenacolemur do possess some divergent character complexes. Picrodontids have retained the primitive pleisia- dapiform-tarsiiform condition of a canine — the first tooth in the jaw — followed by five premolars, but have lost M3/m3, perhaps as a con¬ sequence of the process of specialization of P5/p5. Phenacolemur has three molars, and, with loss of premolars from behind the lower canine Phenacolemur Zanycteris Picrodus Paromomyinae Fig. 8. Cladogram of suggested relationships among Paromomyidae. 68 Annals of Carnegie Museum vol. 46 (Simpson, 1955; Szalay, 1968, 1972), possesses a diastema. These diver¬ gent complexes could have been derived from an ancestral picrodontid- Phenacolemur morphotytpe characterized by a canine at the front of the jaw followed by five premolars and three molars — a condition that is also primitive for all, and retained in many plesiadapiform and tarsiiform primates (Schwartz, MS). Suggested relationships among Paromomyidae are depicted in Figure 8, and we propose the following classification: Infraorder Plesiadapiformes Simons and Tattersall, 19721 Superfamily Plesiadapoidea Trouessart, 1879 Family Paromomyidae Simpson, 1940 Subfamily Paromomyinae Simpson, 1940 Subfamily Picrodontinae (Simpson, 1937) Tribe Phenacolemurini new rank Tribe Picrodontini new rank Cited in Simons (1972). Acknowledgments We thank Dr. Mary R. Dawson, Carnegie Museum of Natural History, and Dr. Malcolm C. McKenna, American Museum of Natural History, for reviewing the manuscript and for their helpful comments. Dr. McKenna generously loaned us the picrodontid material in his care. We are also grateful to Ms. Nancy Perkins for preparing the draw¬ ings; Mr. Jack Capenos, Crucible Research Division, Colt Industries, Pittsburgh, for the scanning electron micrographs; and to Ms. Elizabeth Hill for typing the manuscript. This study was supported in part by an NIMH Biological Sciences Support Grant, University of Pittsburgh (to J.H.S.); and a post-doctoral fellowship, Carnegie Museum of Natural History (to L.K.). 1977 Revision of Picrodontidae 69 References Cited Bown, T. M. 1974. Notes on early Eocene anaptomorphine primates. Contrib. Geol., 13:19-26. Butler, P. M. 1939. Studies of the mammalian dentition. Differentiation of the post-canine denti¬ tion. Proc. Zool. Soc. Lond., B109: 1-36. 1963. Tooth morphology and primate evolution, pp. 1-14. In Brothwell, D., ed. Dental Anthropology. Pergamon Press; London. Clemens, W. A. Jr. 1973. Fossil mammals of the Type Lance Formation, Wyoming, Part III. Eutheria and Summary. Univ. Calif. Publ. Geol. Sci., 94:1-102. Douglass, E. 1908. Vertebrate fossils from the Fort Union beds. Ann. Carnegie Mus., 5:11-26. Gruneberg, H. 1952. The Genetics of the Mouse 2nd edition. Matinus Nijhoff; The Hague. 1963. The Pathology of Development. John Wiley & Sons, Inc.; New York, N.Y. Kollar, E. J. and G. R. Baird 1971. Tissue interactions in developing mouse tooth germs, pp. 15-29. In Dahlberg, A. A., ed., Dental Morphology and Evolution. Univ. Chicago Press, Chicago, Ill. Krishtalka, L. 1976a. Early Tertiary Adapisoricidae and Erinaceidae (Mammalia, Insectivora) of North America. Bull. Carnegie Mus. Nat. Hist., No. 1:1-40. 1976b. North American Nyctitheriidae (Mammalia, Insectivora). Ann. Carnegie Mus., 46(2):7-28. Lillegraven, J. A. 1969. Latest Cretaceous mammals of upper part of Edmonton Formation of Alberta, Canada, and review of marsupial-placental dichotomy in mammalian evolution. Univ. Kansas Paleont. Contrib., 50(Vertebrata 12): 1-122. Matthew, W. D. 1915. A revision of the Lower Eocene Wasatch and Wind River faunas. Part IV. Entelonychia, Primates, Insectivora (part). Bull. Amer. Mus. Nat. Hist., 34:429-483. 1917. A Paleocene bat. Bull. Amer. Mus. Nat. Hist., 37(20):569-571. McGrew, P. O., and B. Patterson 1962. A picrodontid insectivore (?) from the Paleocene of Wyoming. Breviora, 175:1-9. McKenna, M. C. 1963. Primitive Paleocene and Eocene Apatemyidae (Mammalia, Insectivora) and the primate-insectivore boundary. Amer. Mus. Nov., 2160:1-39. 1967. Classification, range and deployment of the prosimian primates. Coll. Int. Cent. Nat. Rech. Sci., Probs. Actuels Paleont., 163:603-610. 1975. Toward a phylogenetic classification of the Mammalia, pp. 21-46. In Luckett, W. P. and F. S. Szalay, eds., Phylogeny of the Primates, a multidisciplinary approach. Plenum Press; New York, N.Y. Miller, W. A. 1971. Early dental development in mice. pp. 3144. In Dahlberg, A. A., ed., Dental morphology and evolution. Univ. Chicago Press, Chicago, Ill. Osborn, J. W. 1973. The evolution of dentitions. Amer. Sci., 61:548-559. 70 Annals of Carnegie Museum vol. 46 Romer, A. S. 1945. Vertebrate Paleontology. 2nd edition, pp. 1-687. Univ. Chicago Press, Chicago, Ill. 1966. Vertebrate Paleontology. 3rd edition, pp. 1-468. Univ. Chicago Press, Chicago, Ill. Schwartz, J. H. [In If Tarsius is not a prosimian, is it a haplorhine? Proc. 7th Congr. Internatl. Press] Primat. Soc. [MS] Dental development, homologies and primate phylogeny. Schwartz, J. H., and L. Krishtalka 1976. The antemolar dentition of Litolestes ignotus, a late Paleocene erinaceid (Mam¬ malia, Insectivora). Ann. Carnegie Mus., 46(1): 1-6. Simons, E. L. 1972. Primate Evolution, an introduction to man’s place in nature. 322 pp. Macmillan, New York, N.Y. Simpson, G. G. 1937. The Fort Union of the Crazy Mountain field, Montana, and its mammalian faunas. Bull. U.S. Natl. Mus., 169:1-287. 1940. Studies on the earliest primates. Bull. Amer. Mus. Nat. Hist., 77:185-212. 1955. The Phenacolemuridae, new family of early Primates. Bull. Amer. Mus. Nat. Hist., 5(5):41 1-442. SZALAY, F. S. 1968. The Picrodontidae, a family of early primates. Amer. Mus. Nov., 2329:1-55. 1969. Mixodectidae, Microsyopidae, and the insectivore-primate transition. Bull. Amer. Mus. Nat. Hist., 140:1-28. 1972. Paleobiology of the earliest primates, pp. 3-35. In Tuttle, R., ed. The Functional and Evolutionary Biology of Primates. Aldine-Atherton; New York, N.Y. 1973. New Paleocene primates and a diagnosis of the new suborder Paromomyiformes. Folia Primat., 19:73-87. Tatters all, 1. 1972. The functional significance of airorhynchy in Megaladapis. Folia. Primat., 18:20-26. Tonge, C. H. 1971. The role of mesenchyme in tooth development, pp.45-58. In Dahlberg, A. A., ed. Dental morphology and evolution. Univ. Chicago Press; Chicago, Ill. Trouessart, E. L. 1879. Catalogue des mammiferes vivants et fossiles . . . Insectivores. Rev. Mag. Zool. Paris, ser. 3., 7:219-285. Van Valen, L. 1965. A middle Palaeocene primate. Nature, 207:435-436. 1967. New Paleocene insectivores and insectivore classification. Bull. Amer. Mus. Nat. Hist., 1 35(5):2 1 7-284. 1969. A classification of the Primates. Amer. J. Phys. Anthrop., 30:295-296. S - A/ fl' ' OCT 3 1977 HARVsA 6&A3-4463 UNIVERSITY ANNALS of CARNEGIE MUSEUM CARNEGIE MUSEUM OF NATURAL HISTORY 4400 FORBES AVENUE • PITTSBURGH, PENNSYLVANIA 15213 VOLUME 46 SEPTEMBER 20, 1977 ARTICLE 7 PALEONTOLOGY AND GEOLOGY OF THE BADWATER CREEK AREA, CENTRAL WYOMING Part 13. The late Eocene Insectivora and Dermoptera Leonard Krishtalka Section of Vertebrate Fossils AND Takeshi Setoguchi1 Abstract Insectivores recovered from the Bad water Uintan and Duchesnean deposits include members of the families Adapisoricidae, Erinaceidae, Nyctitheriidae, Apternodontidae and Soricidae. These deposits provide the earliest known record of Ankylodon, Domnina, two other soricids, Apternodus, Oligoryctes and the dermopteran Thylacaelurus, as well as the latest occurrence of Nyctitherium, Macrocranion and Talpavus. The insectivore and dermopteran record support previous conclusions that currently sampled Bridgerian deposits do not preserve the real diversity of middle Eocene mammals. Introduction This study is an elaboration and revision of a part of a Master’s thesis (Setoguchi, 1973) that dealt with the systematics of the insectivores and dermopterans from the late Eocene deposits of the Badwater Creek area. Recovery of additional material during the past three years from the Badwater localities (5, 5A, 5 Front, 5 Back, 6, Wood, 20) has consider¬ ably improved the insectivore record. Also, recent studies involving early Tertiary insectivores (Krishtalka, 1975, 1976a, 1976b) and Badwater faunas (Black, 1974; Krishtalka and Black, 1975) have clarified the 'Primate Research Institute, Kyoto University, Inuyama City, Aichi 484, Japan. Submitted for publication February 4, 1977. 71 72 Annals of Carnegie Museum vol. 46 systematics and relationships of some of the insectivore taxa and the stratigraphic positions of the Badwater localities. Among the latter, 5, 5A, 5 Front, 5 Back, 6 and Wood are probably Uintan, whereas 20, dated at 41 m.y., is now considered Duchesnean, or latest Eocene. The localities do not occur in the Tepee Trail Formation, as previously thought (Krishtalka and Black, 1975). A current geological review of the area should determine the formational status of the late Eocene Bad¬ water deposits. The Insectivora, as here understood, means essentially the Lipotyphla (see Butler, 1972; McKenna, 1975; Krishtalka, 1976a), and in this paper involves the erinaceomorph families Adapisoricidae and Erinaceidae, and the soricomorph families Nyctitheriidae, Geolabididae, Soricidae and Apternodontidae. The material described is housed in the collections of the Carnegie Museum of Natural History (CM). All measurements are given in milli¬ meters. All figures are stereophotographs and present occlusal views. Caps and lower case I, P, and M refer to upper and lower incisors, pre¬ molars and molars, respectively. Abbreviations in the tables are: L, length; W, width; AW, width of trigonid; PW, width of talonid. Acknowledgments We are grateful to Craig C. Black and Mary R. Dawson (Carnegie Museum of Natural History), Malcolm C. McKenna (American Museum of Natural History), and Peter Robinson (University of Colorado Mu¬ seum) for helpful discussions concerning the Badwater fauna. J. A. Lillegraven (University of Wyoming) and M. C. McKenna generously allowed the use of information from unpublished manuscripts. This study was supported in part by NSF grants GB-7801, GB-3084OX, DEB7.6- 18760, and a Carnegie Museum Post-Doctoral Fellowship (to L. K.). Order Insectivora Family Adapisoricidae (Schlosser, 1887) Macrocr anion Weitzel, 1949 Two North American species of the Eocene insectivore Entomolestes, E. nitens (Matthew, 1918) and Entomolestes cf. E. nitens (McKenna, 1960a), were recently redescribed and referred to Macrocranion, (Krish¬ talka, 1976a), an adapisoricid also known from the middle Eocene of Europe (Weitzel, 1949; Tobien, 1962; Russell, et ah, 1975). The recogni¬ tion of Macrocranion in North American Wasatchian deposits increases the already considerable evidence of a Euramerican biota during the early Eocene (Savage, 1971; Szalay and McKenna, 1971; McKenna, 1971; Dawson, et al., 1976) — a conclusion recently verified by the re¬ covery of a diverse early Eocene mammalian assemblage from Elles¬ mere Island (West and Dawson, in press). 1977 Badwater Area: Part 13 73 Macrocranion robinsoni, new species (Fig. 1; Table 1) Etymology: named after Dr. Peter Robinson for his extensive work with early Tertiary insectivores. Type: CM 18645, Rml-3, locality 6, Badwater Creek area, Wyoming, only known specimen. Diagnosis: size of M. nitens; differs from the latter in the reduction of m3 relative to ml -2. Description: Macrocranion is distinguished from all other adapi- soricids by the wide, elongate talonids on ml -2 and exodaenodont lower molars, especially, ml. CM 18645, the type and only known specimen of M. robinsoni, preserves m2-3 and a broken ml. As in other species of Macrocranion, the talonid of m2 is much longer and wider than the trigonid, the talonid basin is broadly excavated, the metaconid and protoconid are conical, bulbous cusps, and the paraconid is antero- posteriorly compressed into a low, broad crest that terminates anterior to the metaconid. The labial part of the hypoconid base is expanded labially and exhibits the exodaenodonty characteristic of ml -2 in Macro¬ cranion (Krishtalka, 1976a; Russell, et al., 1975). Although much of ml is broken away, the preserved margin of the crown indicates that the talonid was wider and much longer than the trigonid, as is the case on m2 of M. robinsoni and on ml -2 of other species of Macrocranion. Also characteristic of M. robinsoni and other adapisoricids are the high entoconid on the lower molars, the low hypoconid that becomes flat with wear, and the small median hypoconulid. In contrast with other known species of Macrocranion, m3 in M. robinsoni is reduced, in comparison with ml -2. Remarks: The Badwater material provides the latest occurrence of the genus Macrocranion, an adapisoricid also recovered from a number of Wasatchian localities but absent from Bridgerian assemblages. This disjunct distribution and its possible paleoecological interpretations are discussed at the end of this paper. Talpavus Marsh, 1872 Species of this genus are now well known from Wasatchian (Talpavus cf. T. nitidus), Bridgerian (T. nitidus) and Uintan (T. duplus) localities (Robinson, 1968c; Krishtalka, 1976a). Material described as cf. Talpavus has also been recovered from the late Paleocene deposits of the Bad¬ water Creek area (Krishtalka, et al., 1975). Among described adapisori¬ cids, Talpavus most closely resembles Scenopagus in known parts of the dentition, but can be distinguished from the latter mainly by the structure of p4: p4 of Talpavus possesses an unbasined talonid and subequal protoconid and metaconid. The paraconid on ml-3 is a low, anteroposteriorly compressed crest, and the talonid on m2 is as wide as 74 Annals of Carnegie Museum vol. 46 the trigonid. The hypoconulid on ml-2 is low and weak and occurs posterolabial to the entoconid at the lingual end of a distinct hypo- cristid. In contrast, on p4 of Scenopagus the talonid is basined and the protoconid is larger than the metaconid. Lower molars of Scenopagus are proportionately wider, the paraconid is higher than in Talpavus, the m2 talonid is narrower than the trigonid, the hypoconulid is stronger and more distinct, and the hypocristid is much weaker. Talpavus duplus Krishtalka, 1976a (Fig. 2; Tables 1,2) Referred specimens: p4: 15131, 15139, 15743; ml: 15096; m2: 15622; m3: 15649; P4: 15112; Ml: 15011, 15055, 15127, 15641, 15644, 15725, 21630, 23857; M2: 15002, 15109, 15623, 15723; Ml or M2: 15009, 15012, 15013, 15024, 15643, 15722, 18194, 21628, 23951, 23952. Localities: 5A, 5 Front, 5 Back. Known distribution: Uintan, Wyoming and Utah. Description: The lower teeth described here do not differ signifi¬ cantly from those of T. duplus from the Uinta Basin (Krishtalka, 1976a). As is characteristic of Talpavus, p4 has an unbasined talonid and sub¬ equal protoconid and metaconid. Since upper and lower molars of Talpavus have not been found in association, identification of the former is difficult. Guthrie (1967; 1971) ascribed a number of isolated upper molars from Lysite and Lost Cabin to Talpavus — an inference here considered tentative and in need of review, since part of the hypodigm was subsequently referred to Sceno¬ pagus (Krishtalka, 1976a). Similarly, the allocation here of isolated adapisoricid upper molars to T. duplus is tentative, with corroboration subject to the recovery of associated remains. However, the other insec- tivore taxa from Badwater are represented by both upper and lower molars that are distinct from those assigned here to Talpavus. Addition¬ ally, the association here of upper and lower molars of T. duplus is based on proper size, frequency distribution, and recovery from the geographi¬ cally adjacent and temporally equivalent group of localities 5A, 5 Front, and 5 Back. All but seven of the referred upper molars are lingual fragments. The crowns of the more nearly complete specimens are quadrate in occlusal view, with a shallow posterior emargination near the metaconule. As in other adapisoricids, the stylar shelf is narrow, the conules are promi¬ nent, the paracone and metacone are subequal and conical, the proto- Fig. 1. Macrocranion robinsoni, CM 18645, Rml-3, holotype; approx, x 17. Fig. 2. Talpavus duplus. (A) CM 15743, Rp4; (B) CM 15622, Lm2; (C) CM 15725, LM1: (D) CM 15002, RM2; all approx, x 21. Fig. 3. Ankylodon sp. (A) CM 29495, LM1; (B) CM 31300, Rm2; both approx, x 19. 1977 Badwater Area: Part 13 75 76 Annals of Carnegie Museum vol. 46 cone and hypocone are well developed, low and conical, and the para- cingulum and metacingulum are strong. The upper molars are most similar to those of Macrocranion nitens, among known North Ameri¬ can adapisoricids. As in M. nitens, the base of the hypocone on Ml -2 of T. duplus is as broad as the protocone, and directly posterior to the latter; the posterior margin of the hypocone swings labially and anteri¬ orly and meets the trigon below the metaconule; a weak ridge runs from the anterior face of the hypocone to the posterior face of the trigon below the postprotocrista. In contrast to T. duplus, Ml -2 of Scenopagus are much more trans¬ verse, and have much larger parastylar and metastylar areas, with deeper ectoflexi. Similarly, Ml -2 of Macrocranion nitens are slightly more transverse than those of T. duplus; their parastylar and metastylar areas are somewhat more elaborate; and the protocones are more compressed anteroposteriorly. In sum, comparison of Ml -2 of these three adapisoricids reveals a morphological progression from a nearly quadrate crown with tiny stylar areas and an uncompressed protocone (T. duplus) to a semi-trans¬ verse crown with moderately developed stylar salients and a slightly compressed protocone (Macrocranion), to a very transverse crown with large stylar areas and a comparatively highly compressed protocone (Scenopagus). Additional distinctions between Scenopagus and Macro¬ cranion are described elsewhere (Krishtalka, 1976a). The single referred P4, CM 15112, is very worn and does not reveal structural features distinct from those of Scenopagus and Macrocranion. CM 15112 is too small to be referred to M. robinsoni, but appears suit¬ able in size to be part of the upper dentition of T. duplus. As on P4 of Scenopagus and Macrocranion, the crown on CM 15112 is T-shaped, with a long (anteroposteriorly) labial segment and a narrow lingual area. The paracone is tall and dominant and is flanked anteriorly and poste¬ riorly by a small parastyle and metastyle, respectively. The parastyle is isolated from the base of the paracone, whereas the metastyle is linked to the paracone by a raised crest. The lingual area of P4 is too worn to reveal the structure of the protocone or the presence of a hypocone. Ankylodon Patterson and McGrew, 1937 Ankylodon sp. (Fig. 3; Tables 1,2) Referred specimens: m2: 31300; Ml or M2: 15018, 15702; Ml: 29495; M2:15704. Localities: 5A, 5 Front, 5 Back. Known distribution: Uintan, Wyoming. Description and remarks: The dentition of Ankylodon and the relationships of the genus have been briefly discussed elsewhere (Krish¬ talka, 1976a), and a full description of excellent Chadronian material is 1977 Badwater Area: Part 13 77 forthcoming (Lillegraven and McKenna, MS). Although the material from Badwater — an m2, one complete and three fragmentary upper molars — is too sparse for specific identification, the teeth closely resem¬ ble comparable parts of the dentition of Ankylodon. On m2 (CM 31300) the talonid is narrower than the trigonid, the cusps are high and conical, the paraconid is compressed into a flat, wide, labiolingual shelf, and the hypoconulid is large and medial. However, compared to m2 of Chad- ronian Ankylodon, the trigonid on CM 31300 is less compressed antero- posteriorly, the talonid is proportionately shorter and narrower, and the hypocristid is weaker. The single complete upper molar (CM 29495) closely resembles Ml of the Chadronian species of Ankylodon, which is represented in part by a nearly complete palate, a cast of which was kindly furnished by M. C. McKenna. Like Ml of the Chadronian Ankylodon, CM 29495 is trans¬ verse and has a strong hypocone on a wide postcingulum, a deep lingual valley between the hypocone and protocone, a tall paracone and meta¬ cone, a high postmetacrista, a labially expanded metastylar shelf, and a deep ectoflexus. Ml of the Chadronian Ankylodon is, however, more derived, in that the parastylar and metastylar (especially the latter) areas are more expanded, the valley between the hypocone and protocone is Table 1. Dimensions of lower teeth of Macro cr anion robinsoni n. sp., Talpavus duplus, Ankylodon sp. and erinaceid sp. p4 ml m2 m3 L W L AW PW L AW PW L AW PW Macrocranion robinsoni 18645 Talpavus duplus 15131 1.3 1.0 15139 1.4 1.1 15743 1.4 1.0 15096 15622 15649 Ankylodon sp. 31300 erinaceid sp. 15095 14440 15714 1.7 1.8 1.4 1.5 1.7 1.1 0.9 1.4 1.2 1.3 2.2 1.5 1.7 1.5 1.3 1.2 1.9 1.6 1.4 - 1.3 1.5 1.8 1.3 1.4 1.6 1.2 1.1 78 Annals of Carnegie Museum vol. 46 Table 2. Dimensions of upper teeth of Talpavus duplus, Ankylodon sp. and erinaceid sp. Ml M2 Ml or M2 L W L W L* W Talpavus duplus 15011 1.2* 2.0 ** 15055 1.2* 1.9 ** 15127 1.2* — 15641 1.3* 1.8** 15644 1.1* ^ g ** 15725 1.3* 2.0** 21630 1.2* 2.0** 23857 1.4* 2.0** 15002 1.4 2.1 15109 1.4 2.0 15623 1.2 2.0 15723 1.3* 2.1 15009 1.3 — 15012 1.2 — 15013 1.2 — 15024 1.2 — 15722 1.2 — 18194 1.4 — 15643 1.2 — 23951 1.3 — 23952 1.2 — Ankylodon sp. 29495 1.7 2.4 15704 1.5* 2.2 15702 1.5 — erinaceid sp. 15001 1.7* 2.8 ** * lingual length **anterior width wider, and the hypocone is higher. These features and the morphology of the Badwater Ankylodon Ml confirm an earlier diagnosis of that genus (Krishtalka, 1976a) and the described dental trends in the evolu¬ tion of McKennatherium, Scenopagus and Ankylodon: in the upper molars, increase in size of the hypocone, postcingulum, the hypocone- protocone valley, and expansion of the stylar salients. The Badwater and Chadronian material represent two species of Ankylodon. The Badwater species is significantly smaller and, at least on Ml and m2, is more primitive than the Chadronian species. The 1977 Badwater Area: Part 13 79 Fig. 4. Erinaceid sp. (A) CM 15095, Rml; (B) CM 15714, Rm2; (C) CM 15001, LM1; all approx, x 18.5. 80 Annals of Carnegie Museum vol. 46 Orellan A. annectens Patterson and McGrew, 1937 [= A. progressus Galbreath, 1953 (Lillegraven and McKenna, MS; Setoguchi, 1973)] is unfortunately known only from lower teeth. These are also smaller than the lower molars of the Chadronian species but are similar in size and crown morphology to m2 of the Badwater species. The question whether the remains from Badwater represent A. annectens or a third species of Ankylodon can be resolved when the upper dentition of A. annectens and more complete material from Badwater are recovered. Family Erinaceidae Fischer Von Waldheim, 1817 Erinaceid sp. (Fig. 4; Tables 1,2) Referred specimens: ml: 15095; m2: 14440, 15714; Ml: 15001. Localities: 5A, 5 Back, 6, Wood. Known distribution: Uintan, Wyoming. Remarks: The Badwater erinaceid is a new species also known from much more complete material from the East Fork locality, Tepee Trail Formation, Wyoming (McKenna, 1972), that is currently being described elsewhere (Krishtalka and McKenna, in prep.). The erinaceid characters of the lower molars include an m2 smaller than ml, an open trigonid, a compressed, yet cuspate, anterior paraconid, a V-shaped talonid basin, a small hypoconulid, and a high entocristid that is nearly parallel to the cristid obliqua. This species resembles and appears to be closely related to the Tiffanian erinaceid Litolestes ignotus (Krishtalka, 1976a; Schwartz and Krishtalka, 1976). Family Nyctitheriidae Simpson, 1928 Nyctitherium Marsh, 1872 The genus Nyctitherium, discussed in detail by Robinson (1968b) and Krishtalka (1976b), includes two known species, N. velox and N. sero- tinum, both Bridgerian. The Badwater occurrence of Nyctitherium extends the range of the genus to the Duchesnean. Nyctitherium christopheri, new species (Fig. 5; Table 3) Etymology: named for Christopher A. Black. Holotype: CM 27866, RP4-M3, locality 20. Referred specimens: P4: 15023; Ml or M2; 15088, 31299; M2: 16009, 23799. Localities: 5A, 5 Front, 5 Back, Wood, 20. Diagnosis: P4-M3 approximately 30% wider labiolingually than N. velox and N. serotinum; hypoconal shelf larger, flaring more posterolingually. Known distribution: Uintan, Duchesnean, Wyoming. Remarks: Only upper teeth of this new species have been recovered. Compared to N. velox and N. serotinum, P4 and the upper molars of N. 1977 Bad water Area: Part 13 81 christopheri are larger and much more transverse buccolingually, the hypoconal shelf on P4-M2 is more expanded, the precingulum is re¬ duced, and the ectoflexus on M2 is deeper. McKenna (pers. comm.) has also recovered remains of this species from the Uintan East Fork locality, Tepee Trail Formation, Wyoming (McKenna, 1972). Family Geolabididae (McKenna, 1960b) Centetodon Marsh, 1872 Lillegraven and McKenna (MS) have shown that the following genera are wholly or in part junior synonyms of Centetodon: Domnina (in part); Geolabis; Metacodon; and Hypacodon. This is a consequence of previous work by McKenna (1960b) who synonymized Domnina (in part), Herpetotherium , Embassis, and Metacodon with Geolabis, and erected a new Subfamily Geolabidinae to include Geolabis, Hypacodon, Fig. 5. Nyctitherium christopheri, CM 27866, RP4-M3, holotype; approx, x 18. 82 Annals of Carnegie Museum vol. 46 and Myolestes. Shortly thereafter, Centetodon was shown to be con¬ generic with Hypacodon, and referred to the Geolabidinae (McKenna, et al., 1962). Lillegraven and McKenna (MS) define eight species of Centetodon from Bridgerian through Arikareean deposits of North America. New species named in their manuscript are here termed Centetodon sp. B (from the Bridgerian of Wyoming), Centetodon sp. U (from the early Uintan of California), and Centetodon sp. C (from the Chadronian of Wyoming, Montana and Texas). The two known Bridger¬ ian species of Centetodon, C. pulcher and Centetodon sp. B (Lillegraven and McKenna, MS), were also described from the Powder Wash depos¬ its, Utah, by Krishtalka (1975). These species, as well as the early Uintan Centetodon sp. U from California are distinct from Oligocene species in that the upper molars have one lingual root and subequal pre- and postcingula. Oligocene species of Centetodon described by Lillegraven and McKenna (MS) have upper molars with two lingual roots and a postcingulum larger than the precingulum. In these respects, the two species of Centetodon from the late Uintan deposits of Bad water are of Oligocene aspect. Centetodon magnus (Clark, 1936) (Fig. 6; Tables 3,4) Referred specimens: ml: 29197; Ml-3: 31292; Ml: 31298; M2: 15019; Ml or M2: 23854, 31297. Localities: 5 Front, 20. Known distribution: Uintan, Duchesnean, Wyoming; Chadronian, Montana, South Dakota, Colorado; Orellan, Colorado. Description and remarks: The Badwater occurrence of C. mag¬ nus is the earliest known record of that species. C. magnus is much larger in comparative parts of the dentition than Centetodon sp. C, the other species of Centetodon recovered from the Badwater deposits, and described below. As in other species of Centetodon, ml of C. magnus has a very high trigonid and a much lower, narrower talonid. The protoconid and meta- conid are tall and subequal, the paraconid is lingual and somewhat compressed anteroposteriorly, and the trigonid is open lingually between the metaconid and paraconid. The talonid is longer in comparison with the trigonid than in other Eocene species of Centetodon. The talonid cusps are low and equal in size, with the hypoconulid closer to the ento- conid than the hypoconid. The cristid obliqua is “Straight in C. magnus Fig. 6. Centetodon magnus. (A) CM 29197, Lml; (B) CM 31292, LM 1-3; both approx, x 13. Fig. 7. Centetodon sp. C. (A) CM 16008, (B) CM 15137, (C) CM 15115, composite ^ Lml-3, approx, x 13; (D) CM 15619, RM1; (E) CM 16801, RM2; both approx, x 15. 1977 Badwater Area: Part 13 84 Annals of Carnegie Museum vol. 46 and meets the trigonid directly below the protoconid, in contrast with the labially concave cristid obliqua that strikes the trigonid more lin- gually in Centetodon pulcher, Centetodon sp. B, and Centetodon sp. U. m2 repeats the morphology of ml, except for a more nearly triangular trigonid. m3 is shorter and much narrower than ml or m2. Upper molars of C. magnus, hitherto unknown, were recovered from Badwater locality 20. In contrast to the stratigraphically older C. pul¬ cher, Centetodon sp. B and Centetodon sp. U, upper molars of C. mag¬ nus, like those of Centetodon sp. C and other known post-Eocene species of Centetodon, have two lingual roots. Otherwise, upper molars of C. magnus and other species of Centetodon are very similar. Ml -2, essentially rectangular in occlusal view, are much longer labially than lingually, and are very transverse. The stylar shelves are broad, with expanded metastylar and parastylar salients and a prominent buccal emargination of the crown. On Ml the metastylar area projects farther labially than the parastylar area, whereas the latter extends anteriorly beyond the remaining anterior margin of the crown. On M2 the para¬ stylar salient is larger than on Ml, and juts labially beyond the meta¬ stylar one. The paracone and metacone on Ml -2 are quite distant from the buccal border of the crown, and arise from a common base, so that they are united along the lower one-third of their height. The metacone is crescentic; the paracone is nearly conical. The conules are minute, with the metaconule only slightly stronger. Characteristically, a small stylocone occurs on the buccal margin of the crown, just posterior to the parastyle, and labial to the paracone. A weak preparacrista links the paracone to the stylocone. The postmetacrista is very high and slopes posterolabially from the apex of the metacone to the metastyle, and defines the posterior margin of the metastylar salient. The proto¬ cone, which forms the lingual apex of the crown, is anteroposteriorly compressed, and leans anteriorly. Pre- and postcingula occur on either side of the base of the protocone, but are not continuous lingually. On Ml -2 of C. magnus these cingula are broader than on those of Centeto¬ don sp. C and other Eocene species of Centetodon. There is no distinct hypocone on Ml -2 of C. magnus, although the postcingulum termin¬ ates lingually in a raised crest that may be interpreted as a definite cusp. Strong pre- and postprotocristae run from the protocone to the parastyle and the posterior wall of the metastylar area and demarcate a broad paracingulum and metacingulum, respectively. Of the three Eocene species of Centetodon recognized by Lillegraven and McKenna (MS), only Centetodon sp. U, from the early Uintan of California, was represented by upper molars. These have a single lingual root that shows incipient division into a wider anterior and narrower posterior column. Upper molars of the Bridgerian Centetodon sp. B recovered from Powder Wash (Krishtalka, 1975) reveal a lingual root of 1977 Badwater Area: Part 13 85 similar structure, except that on a few of the upper molars the root is partially bifurcate. The presence of two lingual roots on upper molars of both species of Centetodon from Badwater reveals a morphocline in the division of the upper molar lingual root in the evolution of Centeto¬ don during the Eocene. Centetodon sp. B seems close to the ancestry of Centetodon sp. U. Their molar morphology implies that either of these species may have been involved in the ancestry of the two species of Centetodon from Badwater. Centetodon sp. C, Lillegraven and McKenna (MS) (Fig. 7; Tables 3,4) Referred specimens: p4: 15695; ml: 15092, 15605, 16008, 16897, 18226, 23856, 23939, 29111, 29396, 29397; m2: 14547, 15004, 15060, 15064, 15120, 15137, 21631, 23863, 29398, 29399; m3: 15065, 15067, 15115, 18215, 23861; Ml: 15125, 15619, 31293, 31294; M2: 15010, 16801, 31295, 31296; M3: 15128. Localities: 5, 5A, 5 Front, 5 Back, Wood, 20. Known distribution: Uintan, Duchesnean, Wyoming; Chadronian, Wyoming, Mon¬ tana, Texas. Description: This species of Centetodon, named by Lillegraven and McKenna (MS), is based on material recovered from the early Oligocene Chadron Formation of Wyoming and Montana and the Chambers Tuff of Texas. As noted above for C. magnus, the occurrence of this species in the late Eocene deposits of Badwater extends its earliest known re¬ cord. Centetodon sp. C is somewhat smaller than the Orellan C. margin- alis, much smaller than C. magnus , and like C. magnus, differs from other Eocene species in having upper molars with two lingual roots. The morphology of the lower and upper molars of Centetodon sp. C, except for their smaller size, is virtually identical to that described above for C. magnus. The pre- and postcingula are weaker on Ml -2 of Cen¬ tetodon sp. C and do not extend as far lingually around the base of the protocone. The stylar shelf is narrower in Centetodon sp. C, and the stylar salients are less expansive. Family Apternodontidae (Matthew, 1910) Apternodus Matthew, 1903 Apternodus sp. cf. A. illifensis (Fig. 9; Tables 3,4) Referred specimens: m2-3: 27437; m3: 23868; Ml-2: 29012. Locality: 20. Known distribution: Duchesnean, Wyoming. Description and remarks: The material referred here is much larger than that of Oligoryctes sp. described below, but is similar in size and morphology to Apternodus, especially the Chadronian A. illifensis (Galbreath, 1953). Like A. illifensis, the dental remains from locality 20 are smaller that those of A. brevirostris (Schlaikjer, 1934) and A. gre- 86 Annals of Carnegie Museum vol. 46 goryi (Schlaikjer, 1933), and m3 has a longer talonid than in the latter two species and A. mediaevus (Matthew, 1903). As in A. illifensis, Ml-2 have complete cingula from the parastyle to the metastyle, but lack a definite protocone or hypocone. In contrast, the cingula on Ml-2 of A. brevirostris are not complete and, as in A. gregoryi, bear a distinguish¬ able protocone and hypocone lingually. Upper teeth of A. mediaevus are unknown. This record of Apternodus is the earliest known occurrence of the genus. Table 3. Dimensions of upper teeth of Nyctitherium christopheri n. sp., Centetodon magnus, Centetodon sp. C, Apternodus sp. cf. A. illifensis and Oligoryctes sp. Nyctitherium christopheri n. sp. 27866 2.1 15023 — 15088* 16009 23799 Centetodon magnus 31292 31298 Centetodon sp. C 15125 15619 16801 15128 Apternodus sp. cf. A. illifensis 29012 Oligoryctes sp. 23866 23942 *M1 or M2 ** Lingual length Ml M2 M3 w L W L W L W 1.7 2.4 1.6 2.6 1.3 2.6 2.4 1.6**— 1.6 2.6 1.6 2.7 2.0 2.9 1.7 2.9 1.4 2.5 2.1 — 1.6 2.4 1.7 2.4 1.5 2.3 0.7 1.5 2.5 4.0 4.9 3.9 1.0 1.6 1.0 — 1977 Badwater Area: Part 13 87 Table 4. Dimensions of lower teeth of Centetodon magnus, Centetodon sp. C, Apternodus sp. cf. A. illifensis and Oligoryctes sp. L W Centetodon magnus 29197 Centetodon sp. C 15695 1.6 1.1 15092 15605 16008 16897 18226 29111 29397 14547 15060 15120 15137 21631 29398 29399 15065 15067 15115 18215 Apternodus sp. cf. A. illifensis 27437 23868 Oligoryctes sp. 23867** 23869** 23940** 23941** *trigonid length only **ml or m2 ml m2 m3 L AW PW L AW PW L AW PW 2.2 1.5 1.3 1.7 1.2 1.0 1.7 1.2 — 1.8 1.3 1.1 1.7 1.2 1.1 1.6 1.2 1.0 1.6 1.2 1.0 1.7 1.2 1.1 1.8 1.3 1.0 1.7 1.2 1.0 1.6 1.1 0.9 1.6 1.2 1.0 1.7 1.2 1.0 1.8 1.3 1.1 1.7 1.2 1.0 1.7 1.0 0.8 1.5 1.0 0.8 1.6 1.0 0.8 1.5 1.0 0.8 2.2 1.8 — 2.0+ 1.4 — 2.1 1.3 — 0.7* 0.9 — 0.7* 1.0 — 0.9 1.0 — 0.8* 1.1 — 88 Annals of Carnegie Museum vol. 46 Oligoryctes Hough, 1956 Oligoryctes sp. (Fig. 8; Tables 3,4) Referred specimens: ml or m2: 23867, 23869, 23940, 23941; m3: TTM 2454; Ml: 23866, 23942. Localities: 5 Front, Wood, 6. Known distribution: Uintan, Wyoming. Remarks: These isolated teeth are much smaller than similar parts of the dentition of Aptemodus, but are equal in size to the Chadronian Oligoryctes earner onensis. A characteristic feature of the species — a single cusp on the posterior margin of the talonid on m3 as high as the protoconid (Hough, 1956) — is absent on the isolated m3 from Badwater. This material may be referrable to a new Uintan species of Oligoryctes presently being described by McKenna (MS). Family Soricidae (Fischer Von Waldheim, 1817) The soricid remains from the late Eocene of Badwater are the earliest known record of that family. On the basis of size and crown morphology, the material — isolated premolars and molars and two partial dentaries — is separable into three groups. One is identified as a species of Domnina; the other two are non-heterosoricine soricids. Subfamily Heterosoricinae Viret and Zapfe, 1951 Domnina Cope, 1873 Domnina sp. cf. D. gradata (Fig. 10; Tables 5, 6) Referred specimens: ml -3: 23797; ml -2: 15005; ml: 19776, 23865; m2: 151 17, 16062, 23864; m3: 16997, 18198; P4: 15014; Ml: 16996. Localities: 5A, 5 Front, Wood, 20. Known distribution: Uintan, Duchesnean, Wyoming. Description: Patterson and McGrew (1937) and Repenning (1967) have thoroughly described Domnina gradata from the Orellan of Colo¬ rado, South Dakota, and Nebraska. In the absence of preserved mandib¬ ular condyles or the antemolar dentition, Domnina, as well as all other heterosoricines, are best defined by P4 and Ml that lack both an emar- gination of the posterior border of the crown and the resultant posterior expansion of the hypoconal shelf. Instead, the posterior margin of P4 and Ml of Domnina is nearly straight, and the crown is quadrate in occlusal view. Such is the case on CM 15014 and CM 16996, isolated P4 and Ml. The labial half of P4 is formed as a high wall that tapers Fig. 8. Oligoryctes sp. (A) CM 23940, Rml or m2; (B) CM 23866, LM1; both approx, x 15. Fig. 9. Aptemodus sp. cf. A. illifensis. (A) CM 27437, Rml-2; (B) CM 29012, RM1-2; both approx, x 17. 1977 90 Annals of Carnegie Museum vol. 46 posterolabially from the paracone to the metastylar tip of the crown. A low parastyle juts anteriorly from the base of the paracone. Although the lingual part of the crown is worn on CM 15014, the protocone appears to have been very weak or absent, as is also the case on P4 of D. gradata. On Ml the paracone, metacone, and mesostyle are united to form a W-shaped ectoloph, with the paracone situated more buccally than the metacone. The lower molars referred here are similar to those of Domnina but differ from those of Trimylus in that a high entocristid joins the ento- conid to the posterior face of the metaconid and closes the talonid basin lingually. In Trimylus a deep notch isolates the entoconid from the metaconid (Repenning, 1967). The lower molars, larger than those of the Chadronian D. thompsoni (Simpson, 1941), are closer in size and crown morphology to those of D. gradata. In both the Badwater species and D. gradata, the mental foramen occurs below ml; the labial cingulid on ml is not continuous around the base of the protoconid; the hypo- flexid notch is deep; and the cristid obliqua meets the trigonid lingually, below the ventral apex of the protocristid. The ml of the Badwater form is only slightly smaller than the observed size-range of ml of D. gradata. Soricid sp. A. (Fig. 11; Tables 5, 6) Referred specimens: ml-2: 15098; ml: 15025, 15642, 15677, 18196; ml or m2: 15705; Ml: 15712, 16800; M2: 15107, 15113. Localities: 5A, 5 Front, 5 Back, Wood. Known distribution: Uintan, Wyoming. Description: The material referred here is significantly smaller than that of Domnina sp. cf. D. gradata. The upper molars differ from those of Domnina and other heterosoricines in possessing a flaring hypoconal flange that is accentuated by an emargination of the poste¬ rior margin of the crown. On the lower molars the external cingulid is much weaker than in Domnina sp. cf. D. gradata, and the cristid obliqua meets the trigonid labially, below the protoconid, so that the hypo- flexid notch is much shallower than in Domnina. The lower molars, however, resemble heterosoricines and soricines but differ from croci- durines and most limnoecines in that the hypoflexid notch emerges labially at the level of the external cingulid (Repenning, 1967). In the absence of preserved p4, m3, and a mandibular condyle it is difficult to establish the subfamilial or generic affinities of this material. The upper and lower molars are here tentatively assigned to the same species of soricid on the basis of size association and parsimony, although recovery of additional material may imply the occurrence of more than one taxon. 1977 Badwater Area: Part 13 91 Table 5. Dimensions of lower teeth of Domnina sp. cf. D. gradata and soricid sp. A. ml m2 m3 L AW PW L AW PW L AW PW Domnina sp. cf. D. gradata 23797 2.0 1.3 1.5 1.8 1.4 1.5 1.6 1.2 0.8 15005 — — 1.3+ 1.7 1.5 1.5 19776 2.0 1.4 1.6 23865 2.1 1.3 1.5 15117 — 1.3 — 16062 1.8 — — 23864 1.8 1.3 1.4 16997 1.5 1.1 0.7 18198 1.5 1.2 0.8 soricid sp. A 15098 1.7 1.3 1.4 1.7 1.4 1.4 15025 1.7 — 1.3 15642 1.7 — — 15677 1.6 1.3 1.3 18196 1.7 1.4 1.4 Table 6. Dimensions of upper teeth of Domnina sp. cf. D. gradata, soricid sp. A and soricid sp. B P4 Ml M2 L* L** W L* L** W L* L** Domnina sp. cf. D. gradata 15014 1.7 1.1 1.7 16996 2.1 2.0 2.3 soricid sp. A. 15712 1.7 1.7 2.0 16800 1.7 1.6 2.1 15107 1.6 1.5 2.1 15113 1.5 1.5 2.0 soricid sp. B 29198 2.1 1.3 2.4 * labial length ** lingual length 92 Annals of Carnegie Museum vol. 46 Soricid sp. B (Fig. 12; Table 6) Referred specimen: P4: 29198. Locality: 20. Known distribution: Duchesnean, Wyoming. Description: CM 29198, a right P4, is essentially T-shaped in occlu¬ sal view, with an elongate buccal margin and a narrow lingual area. The posterior margin is deeply excavated between the metastylar and lingual regions of the crown, so that the hypoconal shelf is accentuated. The paracone, a tall, conical cusp, forms the broader anterior end of a high crest that tapers toward the posterobuccal corner of the crown. A wide internal cingulum runs along the base of the crest. A low parastyle extends anteriorly from the base of the paracone, which occupies most of the labial half of the crown. The protocone, lower than, and antero- lingual to, the paracone, is highly compressed anteroposteriorly, and forms the anterolingual corner of the tooth. Protocristae are faint and conules are absent. The hypoconal flange runs posteriorly and dorsally from the base of the protocone and ends labially midway along the transverse width of the crown. The lingual margin of the flange is erod¬ ed, but a hypocone does not seem to have been present. Remarks: CM 29198, with a posteriorly emarginate crown, is a P4 of a non-heterosoricine shrew (Repenning, 1967), and resembles the posteriorly excavated upper molars of soricid sp. A described above. The tooth, however, cannot be referred to the latter since it is sig¬ nificantly larger than Ml of soricid sp. A, in contrast to the usual con¬ dition — smaller P4 than Ml — in soricids. CM 29198 appears to represent a third species of shrew at Badwater. The excavation of the posterior border of P4 is similar to that in crocidurines (e.g. Suncus, Crocidura) and soricines (e.g. Sorex). The lingual position of the protocone is closer to the condition on P4 of the former. P4 of limnoecines is unknown (Repenning, 1967). Prior to this description, the earliest records of soricids were Domnina thompsoni from the Chadronian Pipestone Springs locality, Montana (Simpson, 1941), and Sorex herrlingensis (Palmowski and Wachendorf, 1966) from an early Oligocene fissure fill near Wiirttemberg. The figured mandible of S. herrlingensis (Palmowski and Wachendorf, Fig. 10. Domnina sp. cf. D. gradata. (A) CM 23797, Rml-3, approx, x 17; (B) CM 15014, LP4; (C) 16996, LM1; both approx, x 12. Fig. 11. Soricid sp. A. (A) CM 15098, Rml-2; (B) CM 16800, LM1; (C) CM 15113, RM2; all approx, x 12. ^ Fig. 12. Soricid sp. B. CM 29198, RP4, approx. X 15. Fig. 13. Thylacaelurus sp. cf. T. montanus. CM 15061, RM1, approx, x 11. 1977 Badwater Area: Part 13 93 94 Annals of Carnegie Museum vol. 46 1966:234) bears pl-4ml-3. The posterior face of p4 is excavated and concave as in heterosoricines and crocidurines (Repenning, 1967). The molars more closely resemble those of heterosoricines in that the hypo- flexid notch emerges labially at the level of the external cingulid rather than more dorsally, as in crocidurines. Repenning (1967) did not discuss S. herrlingensis, but Thenius (1969) questioned its generic identifica¬ tion. Subsequently, Engesser (1975) correctly removed the species from Sorex and referred it to a new European heterosoricine genus, Quercysorex. The earliest occurrence and diversity of shrews at Badwater implies the presence of ancestral soricids in as yet unsampled early and pre- Uintan deposits, and considerably antedates the oldest known record of crocidurines (Burdigalian, Europe), limnoecines (Whitneyan, North America) and soricines (Stampian, Europe). As discussed below, the paleoecological implications of this disjunct distribution are clear: most mid-late Eocene and Oligocene localities preserve similarly re¬ stricted environmental situations with, probably, lowland, lake-margin communities (Black, 1967). These do not reflect the diversity of later Eocene and Oligocene life. Order Dermoptera. Family uncertain Thylacaelurus Russell, 1954 Thylacaelurus sp. cf. T. montanus (Fig. 13) Referred specimens: i: 15124, 15616, 15639, 18193, 18202, 27000; Ml: 15061, 15707. Localities: 5A, 5 Front, 5 Back. Known distribution: Uintan, Wyoming; (Chadronian, British Columbia, for T. montanus). Description: Except for their slightly smaller size (L, 1.7; W, 2.0) and the absence of a precingulum, the two isolated upper molars are virtually identical to Ml of Thylacaelurus montanus from the early Oligocene Kishenehn deposits of British Columbia, Canada (Russell, 1954). Characteristically, the paracone, paraconule, and protocone are aligned along the anterior margin of the crown, whereas the metacone, metaconule, and hypocone occur along the posterior border. Conse¬ quently, these sets of cusps are widely separated by a deep trigon basin that is oriented buccolingually from the labial border of the crown to the internal face of the protocone. The paracone and metacone are quite lingual, leaving a large stylar shelf. Three stylar cusps (two mesostyles and a metastyle) occur along the buccal margin of the crown. The most anterior mesostyle is posterobuccal to the paracone. The posterior meso- style is anterobuccal to the metacone. The metastyle forms the postero- labial corner of the crown. A postparacrista links the paracone to the anterior mesostyle, and pre- and postmetacristae join the metacone to 1977 Badwater Area: Part 13 95 the posterior mesostyle and metastyle, respectively. The protocone forms the anterolingual corner of the crown. The conules are large and conical but lack well developed cristae. The hypocone, almost as large as the metaconule, occurs posterolabial to the protocone and is linked to the postprotocrista by a strong crest. The referred incisors have digitate crowns, with three (one speci¬ men), four (three specimens), or five (one specimen) comb-like pro¬ jections. CM 15639, an incisor fragment with three lobes, appears to have had four or five when complete. These incisors most closely resemble i3 of the extant dermopteran Cynocephalus, and differ from the bilobate lower incisors of Plagiomene (Rose, 1973). Although we are reasonably confident in our referral of these incisors to Thylacae- lurus, they may also belong to the Badwater species of Nyctitherium, since digitate incisors have been identified in another genus of this family (Sige', 1976). Remarks: The affinities of Thylacaelurus are uncertain. Although Russell (1954) described T. montanus as a didelphid marsupial, the species was subsequently identified as a eutherian (McKenna, in Van Valen, 1965), a plagiomenid dermopteran (Van Valen, 1967) and a dimylid-like insectivore (Szalay, 1969). Ml of Thylacaelurus, with stylar cusps and connecting cristae, superficially resembles that of Elpido- phorus, Plagiomene, Planetitherium, Remiculus, Adapisoriculus, and didelphid marsupials. In contrast to Thylacaelurus, however, Ml of didelphids, Remiculus, and Adapisoriculus lack a true hypocone, have subcrescentic cusps, smaller conules with internal and external wings, and a medial rather than anterolingual protocone. The stylar cusps and possible reduction of M3 in Thylacaelurus are also reminiscent of some talpids and dimylids (see Szalay, 1969), but these resemblances also appear to be convergent when one considers the absence in Thylacae¬ lurus of an expanded hypoconal shelf, dilambdodonty, low crescentic conules, and a suite of other characters that unite talpids and dimylids among the soricomorphs (Schmidt-Kittler, 1973). Rather, the structure of Ml -2 of Thylacaelurus — the stylar shelf and cusps, the enlarged conules, the anterolingual protocone, and the align¬ ment of the cusps in labiolingual rows on either side of the trigon valley — is characteristic of plagiomenids, recent dermopterans, and the alleged mixodectid Elpidophorus. The latter may have been involved in the ancestry of the plagiomenids Planetitherium and Plagiomene (see Simpson, 1936; Szalay, 1969; Rose, 1973). Thylacaelurus appears to be a dermopteran, but not a plagiomenid, since its enlarged premolariform P4 and possible loss or reduction of M3 is apomorphic with regard to the fully molariform P4 and occurrence of M3 in known members of the family. The Badwater record of Thylacaelurus is the earliest known occurrence of the genus. 96 Annals of Carnegie Museum vol. 46 Comments Two facets of the Badwater late Eocene mammalian assemblage — the alleged relictual occurrences and the earliest known occurrences of taxa — have specific implications regarding our knowledge of middle and late Eocene mammal communities (Black, 1967). These facets are also reflected in the insectivore and dermopteran record. First, the occur¬ rence of Macrocranion and a dermopteran at Badwater and their ab¬ sence from North American middle Eocene and other late Eocene local¬ ities parallels the temporally disjunct distribution of Phenacolemur (Robinson, 1968a) and neoplagiaulacid multituberculates (Sloan, 1966; Krishtalka and Black, 1975). Second, the Badwater shrews, Oligoryctes, Apternodus, Centetodon, and a variety of rodents have not been recovered elsewhere from pre-Oligocene deposits. These records imply that the Badwater (and Shoddy Springs) sediments represent an other¬ wise unsampled ecological situation among North American Eocene post-Wasatchian localities. The latter are typically intermontane basin deposits that preserve similar lake margin and associated floodplain facies and correspondingly similar lowland communities. The Badwater assemblage may, in part, reflect an upland community, comprising, at least four groups: multituberculates, a dermopteran, Phenacolemur, and Macrocranion — small animals that during Bridgerian-Uintan time probably inhabited upland environments that were far from and not preserved in currently sampled Bridgerian-Uintan sediments. We con¬ cur with Black (1967) that assemblages recovered from the latter do not represent the actual diversity of middle and late Eocene vertebrates. References Cited Black, C. C. 1967. Middle and late Eocene mammal communities: a major discrepancy. Science, 156:62-64. 1974. Paleontology and geology of the Badwater Creek area, central Wyoming. Part 9. Additions to the cylindrodont rodents from the late Eocene. Ann. Carnegie Mus., 45(7): 151-1 60. Butler, P. M. 1972. The problem of insectivore classification, pp. 253-265. In Joysey, K. A., and T. S. Kemp, eds., Studies in Vertebrate Evolution. Winchester Press, New York. Clark, J. 1936. Diagnosis of Metacodon and description of Metacodon magnus. In Scott, W. B., and G. L. Jepsen (1936), The mammalian fauna of the White River Oligocene, Part 1, Insectivora and Carnivora. Trans. Amer. Phil. Soc., n. ser., 28:22. Cope, E. D. 1873. Third notice of extinct vertebrata from the Tertiary of the plains. Palaont. Bull., 16:1-8. 1977 Badwater Area: Part 13 97 Dawson, M. R., R. M. West, W. A. Langston and J. H. Hutchison. 1976. Paleocene terrestrial vertebrates: northernmost occurrence, Ellesmere Island, Canada. Science, 192:781-782. Engesser, B. 1975. Revision der europaischen Heterosoricinae (Insectivora, Mammalia). Eclogae Geol. Helv., 68(3):649-671. Fischer Von Waldheim, G. 1817. Adversaria zoologica. Mem. Soc. Imp. Nat., Moscow, 5:368-428. Galbreath, E. C. 1953. A contribution to the Tertiary geology and paleontology of northeastern Col¬ orado. Univ. Kansas Paleont. Contrib., Vertebrata, 4:1-120. Guthrie, D. A. 1967. The mammalian fauna of the Lysite Member, Wind River Formation (early Eocene), of Wyoming. Mem. So. Calif. Acad. Sci., 5:1-53. 1971. The mammalian fauna of the Lost Cabin Member, Wind River Formation (Lower Eocene) of Wyoming. Ann. Carnegie Mus., 43(4):47-l 13. Hough, J. 1956. A new insectivore from the Oligocene of the Wind River Basin, Wyoming, with notes on the taxonomy of the Oligocene Tenrecoidea. Jour. Paleo., 30(3):53 1-541. Krishtalka, L. 1975. Systematics and relationships of early Tertiary Lipotyphla (Mammalia, Insec¬ tivora) of North America. Doctoral Dissertation, Texas Tech Univ., Lubbock. 1976a. Early Tertiary Adapisoricidae and Erinaceidae (Mammalia, Insectivora) of North America. Bull. Carnegie Mus. Nat. Hist., 1:1-40. 1976b. North American Nyctitheriidae (Mammalia, Insectivora). Ann. Carnegie Mus., 46(2): 7-28. Krishtalka, L., and C. C. Black. 1975. Paleontology and geology of the Badwater Creek area, central Wyoming. Part 12. Description and review of late Eocene Multituberculata from Wyoming and Montana. Ann. Carnegie Mus., 45(15):287-297. Krishtalka, L., C. C. Black and D. W. Riedel. 1975. Paleontology and geology of the Badwater Creek area, central Wyoming. Part 10. A late Paleocene mammal fauna from the Shotgun Member of the Fort Union Formation. Ann. Carnegie Mus., 45(9): 179-212. Lillegraven, J. A., and M. C. McKenna. [MS] Phylogenetic history and microevolutionary mosaicism of Centetodon. Marsh, O. C. 1872. Preliminary description of new Tertiary mammals. Amer. J. Sci., Ser. 3, 4:122- 128, 202-224. Matthew, W. D. 1903. The fauna of the Titanotherium beds at Pipestone Springs, Montana. Bull. Amer. Mus. Nat. Hist., 19(6): 197-226. 1910. On the skull of Apternodus and the skeleton of a new artiodactyl. Bull. Amer. Mus. Nat. Hist., 28:33-42. 1918. A revision of the lower Eocene Wasatch and Wind River faunas. Part 5. — Insec¬ tivora (continued), Glires, Edentata. Bull. Amer. Mus. Nat. Hist., 38-565-657. McKenna, M. C. 1960a. Fossil Mammalia from the early Wasatchian Four Mile fauna, Eocene of north¬ west Colorado. Univ. Calif. Publ. Geol. Sci. 37(1): 1-1 30. 98 Annals of Carnegie Museum vol. 46 1960b. The Geolabidinae, a new subfamily of early Cenozoic erinaceoid insectivores. Univ. Calif. Publ. Geol. Sci., 37:131-164. 1971. Fossil mammals and the Eocene demise of the De Geer North Atlantic dispersal route. Geol. Soc. Amer. Abstr. with Programs, 3(7):644. 1972. Vertebrate paleontology of the Togwotee Pass area, northwestern Wyoming. In West, R. M., ed., Guidebook, Field Conference on Tertiary Biostratigraphy of Southern and Western Wyoming, pp. 80-101. 1975. Toward a phylogenetic classification of the Mammalia. In Szalay, F. S., and W. P. Luckett, eds., Phylogeny of the Primates; an interdisciplinary approach. Plenum Press, New York. [MS] Review of the apternodontid Zalambdodonta (Mammalia). McKenna, M. C., P, Robinson and D. W. Taylor. 1962. Notes on Eocene Mammalia and Mollusca from Tabernacle Butte, Wyoming. Amer. Mus. Novitates, 2102:1-33. Palmowski, J., amd H. Wachendorf. 1966. Eine unteroligozane Wirbeltierfauna aus einer Spaltenfullung in Herrlingen, Blau (Wiirtt.). Mitt. Bayer. Staatssamml. Palaont. hist. Geol., 6:229-245. Patterson, B., and P. O. McGrew. 1937. A soricid and two erinaceids from the White River Oligocene. Publ. Field Mus. Nat. Hist., geol. ser., 6:245-272. Repenning, C. A. 1967. Subfamilies and genera of the Soricidae. Geol. Surv. Prof. Paper, 565:1-74. Robinson, P. 1968a. Paleontology and geology of the Bad water Creek area, central Wyoming. Part 4. Late Eocene primates from Badwater, Wyoming, with a discussion of material from Utah. Ann. Carnegie Mus., 39(19):307-326. 1968b. Nyctitheriidae (Mammalia, Insectivora) from the Bridger Formation of Wyo¬ ming. Univ. Wyo. Contrib. Geol., 7:129-238. 1968c. Talpavus and Entomolestes (Insectivora, Adapisoricidae). Amer. Mus. Novitates, 2339:1-7. Rose, K. D. 1973. The mandibular dentition of Plagiomene (Dermoptera, Plagiomenidae). Bre- viora, 41 1:1-17. Russell, D. E., P. Louis and D. E. Savage. 1975. Les Adapisoricidae de l’Eocene Inferieur de France. Reevaluation des formes considerees affines. Bull. Mus. Natl. d’Hist. Nat., 327(45): 129-194. Russell, L. S. 1954. Mammalian fauna of the Kishenehn Formation, southeastern British Columbia. Ann. Rep. Natl. Mus. Canada for 1952-1953, Bull., 132:92-111. Savage, D. E. 1971. The Sparnacian-Wasatchian mammalian fauna, early Eocene of Europe and North America. Abh. Hess. L.-Amt Bodenforsch, 60:154-158. SCHLAIKJER, E. M. 1933. A detailed study of the structure and relationships of a new zalambdodont insec- tivore from the Middle Oligocene. Bull. Mus. Comp. Zool., 76:1-27. 1934. A new fossil zalambdodont insectivore, Amer. Mus. Novitates, 698:1-7. Schlosser, M. 1887. Die Affen, Lemuren, Chiropteran, Insectivora, Marsupialier, Creodontan und Carnivoren des europaischen Tertiars. Teil 1, Beitr. Palaont. Oesterreich- Ungarns, 6:1-224. 1977 Badwater Area: Part 13 99 Schmidt-Kittler, N. 1973. Dimyloides- Neufunde aus der oberoligozanen Spaltenfiillung “Ehrenstein 4” (Siiddeutschland) und die systematische Stellung der Dimyliden (Insectivora, Mammalia). Mitt. Bayer. Staatssamml. Palaont. hist. Geol., 13:115-139. Schwartz, J. H., and L. Krishtalka. 1976. The lower antemolar teeth of Litolestes ignotus, a late Paleocene erinaceid (Mammalia, Insectivora). Ann. Carnegie Mus., 46(1): 1-6. Setoguchi, T. 1973. The late Eocene marsupials and insectivores from the Tepee Trail Formation, Badwater, Wyoming. Masters Thesis, Dept. Geol., Texas Tech Univ., Lubbock. Sige, B. 1976. Insectivores primitifs de 1’Eocene Superieur et Oligocene Inferieur d’Europe occidentale. Nyctithe'riides. Mem. Mus. Natl. d’Hist. Nat., Ser. C., 34:1-140. Simpson, G. G. 1928. A new mammalian fauna from the Fort Union of southern Montana. Amer. Mus. Novitates, 297:1-15. 1936. A new fauna from the Fort Union of Montana. Amer. Mus. Novitates, 873:1-27. 1941. A new Oligocene insectivore. Amer. Mus. Novitates, 1150:1-3. Sloan, R. E. 1966. Paleontology and geology of the Badwater Creek area, central Wyoming. Part 2. The Badwater multituberculate. Ann. Carnegie Mus., 38:309-315. Szalay, F. S. 1969. Mixodectidae, Microsyopidae, and the insectivore-primate transition. Bull. Amer. Mus. Nat. Hist., 140(4): 193-330. Szalay, F. S., and M. C. McKenna 1971. Beginning of the age of mammals in Asia: the late Paleocene Gashato fauna, Mongolia. Bull. Amer. Mus. Nat. Hist., 144(4):269-318. Thenius, E. 1969. Stammesgeschichte der saugetiere (einschliesslich der Hominiden). Kukenthal’s Handbuch der Zoologie, 8, Lieferung 47:1-368. Walter de Gruyter, Berlin. Tobien, H. 1962. Insectivoren (Mamm.) aus dem Mitteleozan (Lutetium) von Messel bei Darm¬ stadt. Notizbl. hess. Land-esamt. Bodenforsch, 90:7-47. Van Valen, L. 1965. Paroxyclaenidae, an extinct family of Eurasian mammals. Jour. Mamm., 46(3): 388-397. 1967. New Paleocene insectivores and insectivore classification. Bull. Amer. Mus. Nat. Hist., 135(5):217-284. VlRET, J., AND H. ZAPFE. 1951. Sur quelques Soricides miocenes. Eclogae Geol. Helv., 44(2):41 1-426. Weitzel, K. 1949. Neue Wirbeltiere (Rodentia, Insectivora, Testudinata) aus den Mitteleozan von Messel bei Darmstadt. Abh. Senckenb. Naturforsch. Ges., 480:1-24. West, R. M., and M. R. Dawson. [In Mammals from the Paleogene of the Eureka Sound Formation: Ellesmere Press] Island, Arctic Canada. Geobios. Back issues of many Annals of Carnegie Museum articles are available, and a few early complete volumes and parts are listed at half price. Orders and inquiries should be addressed to: Publications Secretary, Carnegie Museum, 4400 Forbes Avenue, Pittsburgh, Pa. 15213. AJ f c rtsU/o p 6 SEP 2 9 1977 HARVARD 0097-4463 UNIVERSITY ANNALS of CARNEGIE MUSEUM CARNEGIE MUSEUM OF NATURAL HISTORY 4400 FORBES AVENUE • PITTSBURGH, PENNSYLVANIA 15213 ARTICLE 8 VOLUME 46 SEPTEMBER 27, 1977 A NEW SUBGENUS AND NEW SPECIES OF MIOCENE CALLIOSTOMA (ARCHAEOGASTROPODA— TROCHIDAE) J. J. Parodiz Curator of Invertebrates Abstract This paper describes a new subgenus of Calliostoma: (Archaeogastropoda — Trochidae), Tropidotrochus characterized by unbeaded spiral sculpture, concave whorls, and double peripheral carinae. Type-species is Calliostoma virginicum (Conrad) which, with other Miocene species of the eastern United States — C. carolinense Gardner, C. labrosum (Con¬ rad), and C. cyclum (Conrad) — differs from the living subgenera Calliostoma Swainson, 1840, s.s.; Elmer linia Clench and Turner, 1960; Kombologion Clench and Turner, 1960, all of which have convex whorls and beaded sculpture. A new species, C. (Tropidotrochus) jayae, is here described, from the Miocene of Virginia. Introduction In October, 1975, during a collecting excursion to the marine Mio¬ cene deposits in Virginia, Mrs. Jay J. Trippe obtained a unique specimen of Trochidae. Mrs. Tripp consulted the Section of Invertebrates for identification, and the specimen proved to be not only an undescribed species, but a member of a new subgenus of Calliostoma, herewith described, different from the ones now living. Genus Calliostoma Swainson, 1840. Tropidotrochus1, new subgenus. type-species: Zizyphinus virginicus Conrad, 1875. From Suffolk, Virginia, upper section of Yorktown Formation, late Miocene. This, the first species described, now belongs to the new group. 1 Tropida, Gr. = keel + trochus - top shell. Submitted for publication August 27, 1976. 101 102 Annals of Carnegie Museum vol. 46 diagnosis: Calliostoma with double peripheral keel or carinae. The two carinae, dividing the top cone of the shell from the base, are separated by a deep transverse sulcus. The sulcus, correspondingly narrower, is repeated above, at the base of each whorl. Sides of the whorls are concave in the type-species, and more so in C. (T. ) jayae. In other species like C. carolinense Gardner and C. labrosum (Conrad), the sides of the whorls are concave to a lesser degree, but never convex, as in C. (Calliostoma) or in C. (Leiotrochus). The sculpture consists of spiral cords, very regular in C. virginicum, and less regular in C. carolinense and C. jayae, but never beaded. The umbilicus is closed in adults, except in C. cyclum Dali, which might be an immature specimen because of its small size and other characteristics of its group. discussion: Tropidotrochus is apparently an extinct Miocene group. The only two known living species of Calliostoma, s.l., with strong keels are C. cubanum Clench & Aguayo, and C. schroederi C. & A. They have beaded sculpture and do not correspond to the group typified by C. virginicum. Calliostoma (Calliostoma), as defined by its type-species C. zizyphi- num (L.), is a flat-sided trochoid shell, the sculpture consisting of very fine, regular, transversal flattened cords. Other species are practically smooth, like C. conulum (L.), which by an error of Herrmannsen (1846) was also made the type-species of the genus. The conspicuously beaded species of Calliostoma were placed by Clench and Turner (1960) in the new subgenera Elmer linia and Kom- bologion. In the first, the shells are perforate, and in the second, gen¬ erally imperforate. The two are also differentiated by radular character¬ istics. Some of these species as well as others that remain in Calliostoma, s.l., without subgeneric definition, might belong to the subgenus Eucasta Dali, 1889. The subgenus Leiotrochus Conrad, 1862, is characterized by convex whorls, numerous and regular spiral cords, and an angulated, instead of keeled, periphery. All the known species are from the Miocene: C. distans Conrad (type-species), C. armillatum Conrad, C. herrisi Dali, C. nottowayense Gardner, and C. conradi Gardner. The following species and subspecies should be included in the new subgenus Tropidotrochus: Calliostoma virginicum gizehi Gardner, C. carolinense Gardner, C. cyclum Dali, C. labrosum (Conrad). Calliostoma (Tropidotrochus) jayae, new species diagnosis: Shell not umbilicate, wider than high, whorls concave, sculpture with unbeaded spiral cords, with double carinae and sulcus at the periphery and base of each whorl. Aperture produced to the right with basal lip festooned, Protoconch smooth, globular. description: Shell with almost six whorls (5-4/5 or 5 + 248°); those corresponding to the spire, concave. The protoconch has 1-1/2 globose and polished whorls, and these, up to the end of the second whorl, are iridescent. Starting at the middle of the second whorl is a spiral cord, which is doubled in the third and successively increases in number to the last whorl, which has six. The spaces between are deep and as wide as the cords. In the penultimate whorl there are also traces of thread-like spirals between the cords. The cords 1977 New Subgenus and Species of Calliostoma 103 and the rest of the shell surface are not beaded. Beginning at the third whorl, the sutural area is margined with a strong, rounded keel, very prominent at the periphery between cone and base. This keel, or carina, continues to the distal end of the outer lip. A similar carina is repeated below, leaving a wide, deep canal between them. Although the sutural area in the upper whorls is marked by the carinae, the suture is rather indistinct. The aperture is subquadrate, 1/5 wider than high, and falls obliquely to the base. The base is horizontal, but conspicuously crenulated by a festoon that corresponds to the ends of the cords on the outer and basal lips. Viewed from its base, the umbilical area is closed, and around it there are 16 spiral cords, finer toward the periphery. The last two, at the center, are very wide and strong. The columellar area is triangular, very wide above, and thick. At its basal end it has a tooth-like protuberance. dimensions of holotype: Height 9.2 mm, width 10.1 mm. Aperture almost 4 mm wide. Diameter of the last whorl above the carina, 7.5 mm (this accounts for a low spire in relation to the total width of the shell). Spire angle in relation to axis, 80°. Divergence of sutural line from base, 5°. Obliquity of left side of base in relation to basal line, 35°. type locality: Rice’s Pit, Hampton, Virginia. Collected by Mrs. Jay J. Tripp, Octo¬ ber 6, 1975, together with other fossils including Conus stearnsi Conrad and Polinices catenoides (Wood), in deposits of upper Yorktown Formation, late Miocene. holotype: Section of Invertebrates, CM 43647. remarks: The upper (conic) part of the shell is shorter, and corre¬ spondingly, the full diameter is wider than the same features in C. virginicum (Conrad). The columellar plate in C. jayae is also wider than that of C. virginicum, with a protuberance that the latter does not have. In both C. jayae and C. carolinense, the peripheral canal between the carinae is larger than that of C. virginicum, but the spiral cords in C. jayae are stronger and more numerous on the base than in C. virginicum. In C. jayae the sides of the whorls are more concave than in C. virgini¬ cum, and the apertural outline is more angulated and projected to the right. Crenulation of the peristome is regular. C. virginicum gizehi Gardner has more regular sculpture and less prominent carinae than C. jayae. The species of Tropidotrochus are closely related, and C. (T.) jayae differs in the same degree as the others do in proportions and sculpture. They suggest a certain relationship with Pliocene species from Ecuador, C. palmeri Dali and C. nonorum Pilsbry & Olsson, which have slightly concave whorls. The communication that existed between the Western Atlantic and Eastern Pacific before the formation of the isthmus (be¬ tween North and South America), during the Tertiary, may account for this relationship. C. jayae also can be compared with the distant C. iheringi Ortmann from the Patagonian Formation (lower Miocene) of South America. The inference is that this group of Calliostoma had an extensive distribution when the now colder southernmost areas of the Atlantic were subtropical. It can be observed that in the evolution of Calliostoma- like genera, sub¬ genera, or group of species, a process has been repeated in each of them (from the Triassic to Recent) in which earlier forms were generally smooth, then spirally sculptured, ending or continuing into forms with strong beaded sculpture. Viewed in this sense, Leiotrochus evolved into 104 Annals of Carnegie Museum vol. 46 Tropidotrochus, Calliostoma, s.s.; Elmerlinia; and Kombologion. This however, is a raw assumption, since the taxonomy of the subfamily Calliostomatinae needs a full neo-paleontological revision. i Figs. 1, 2. Calliostoma (Tropidotrochus) jayae, new species, holotype. 1. Front view. 2. Apical view. (Scale represents 2 mm.) 1977 New Subgenus and Species of Calliostoma 105 Figs. 3, 4. Calliostoma (Tropidotrochus) jayae, new species, holotype. 3. Dorsal view. 4. Basal view. (Scale represents 2 mm.) 106 Annals of Carnegie Museum vol. 46 References Cited Clench, W. J., and Turner, R. D. 1960. The genus Calliostoma in the Western Atlantic. Johnsonia, 4 (40): 1-79. Conrad, T. A. 1862. Description of new genera, subgenera and species of Tertiary and Recent Shells. Proc., Acad. Nat. Sci., Philadelphia, 14:284-291. Conrad, T. A., in Kerr, W. C. 1875. Report North Carolina Geol. Survey, 1 (1-325), App.A:l-28. Dall, W. H. 1889. Report on the Blake Mollusca. Gastropoda. Bull. Mus. Comp. Zool., 18:363-374. Gardner, J. 1948. Mollusca from the Miocene and Lower Pliocene of Virginia and North Carolina. U. S. Dept, of Interior, Geol. Survey, Professional Paper 199-B: 179-279. Herrmannsen, A. N. 1846. Generum Malacozoorum Primordia. Cassel. 1:1-637. SWAINSON, W. 1840. Treatise on Malacology. London. 1:1-419. Wenz, W. 1938. Handbuch der Palaozoologie. Gastropoda, 6 (2):269-287. Back issues of many Annals of Carnegie Museum articles are available, and a few early complete volumes and parts are listed at half price. Orders and inquiries should be addressed to: Publications Secretary, Carnegie Museum, 4400 Forbes Avenue, Pittsburgh, Pa. 15213. MU®. COMP. ZOOU /V/F PM^VyL Library OCT 2 5 ]QJ7 HAfWW3 UNIVERSITY AN NALS o/CARNECIE MUSEUM CARNEGIE MUSEUM OF NATURAL HISTORY 4400 FORBES AVENUE • PITTSBURGH, PENNSYLVANIA 15213 VOLUME 46 SEPTEMBER 28, 1977 ARTICLE 9 PALEOSPECIES OF NEOTROPICAL AMPULLARIIDS AND NOTES ON OTHER FOSSIL NON-MARINE SOUTH AMERICAN GASTROPODS Kenneth J. Boss 1 Juan J. Parodiz Curator of Invertebrates Abstract This paper describes recently collected Tertiary fresh-water gastropods from the late Eocene, of the Department of San Martin in northeastern Peru, of the family Ampul- ilariidae: a giant Pomacea prourceus, new species; Pomacea (Effusa) pattersoni, new species; and the presence of the Pleuroceridae Doryssa corrosensis (Pilsbry and Olsson) — previously known from Colombia and Ecuador — in the Upper Eocene of Peru. Other Pleis¬ tocene and Recent land shells, collected in the same area, are also taxonomically dis¬ cussed. These include Oleacinidae, Euglandina sacatta (Pfeiffer); Orthalicidae, Corona regalis (Hupe); Cyclophoridae, Poteria (Neocyclotus) inca (Orbigny); Megalobulimidae, Megalobulimus popeleirianus (Nyst). Introduction The Tertiary land and fresh-water mollusks of the Neotropical Region, especially of Peru, are relatively poorly known. Parodiz (1969) reviewed the known occurrences of Tertiary non-marine South American mollusks and provided the most thorough documentation on the paleogeography of this fauna. However, the Recent continental malacofauna of South America remains, for the most part, in need of systematic revision. Material collected from Peru during the last several decades has been the object of many descriptive papers (e.g. Haas, 1948; 1949; 1951; 1952; 1955; Pilsbry, 1932; 1944). The purpose of this paper is to record 'Museum of Comparative Zoology, Harvard University, Cambridge, Mass. 02138. Submitted for publication February 8, 1977. 107 108 Annals of Carnegie Museum vol. 46 specimens collected by Professor Bryan Patterson of the Museum of Comparative Zoology during the summer of 1974. These samples are from the Department of San Martin in northeastern Peru, and include both Tertiary and Recent material. The collection consists of two moie¬ ties, one of several early Tertiary steinkerns of freshwater prosobranchs (Ampullariidae and Pleuroceridae) and the other of a series of miscel¬ laneous Recent terrestrial gastropods (Oleacinidae, Strophocheilidae, Orthalicidae, and Cyclophoridae), picked up on river banks. Acknowledgments We express our gratitude to B. Patterson, who collected the speci¬ mens. The manuscript was read critically by R. D. Turner and R. I. Johnson. The photographs of Pomacea prourceus (Figs. 1-3) were taken by A. Coleman, MCZ labs, and the X-ray of P. urceus (Fig. 6) was made by R. D. Turner. The MS was typed and corrected by Mrs. G. Dent. Systematics Ampullariidae2 The Ampullariidae, or apple snails, so-called because of their relative abundance and large size, form a conspicuous element of the tropical and subtropical fresh-water malacofauna of both the Old and New Worlds (Kobelt, 1911-15). Although primarily aquatic, the family is unique in being amphibious prosobranchs that have respiratory struc¬ tures functioning both in and out of water. In one of its portions, the mantle cavity contains a ctenidium, while another part is modified to form a lung or gas-filled pulmonary sac (Troschel, 1845; Pelseneer, 1895; Andrews, 1965). Some species, like Pila globosa and Pomacea paludosa, may emerge from the normally aqueous habitat to lay clusters of aposematic eggs on land or on the stems of aquatic vegetation (Sax- ena, 1956; Snyder and Snyder, 1971; Perry, 1973), while other species, notably Pomacea urceus, burrow into the substrate and brood their eggs (Burky, Pacheco and Pereyra, 1972; Burky, 1974). Many ampullariids are capable of entering a dormant state in response to inhospitable environmental extremes, e.g., the advent of the dry season (Boss, 1974). Burrowing deeply into the substrate, in fact, to depths as great as four feet (Arkell, 1924; Ramanan, 1903; Bahl, 1928), the animals would seem more or less prone to fossilization. However, the fossil record of ampullariids, as noted in the literature, is anything but complete. Wenz (1938) averred that the group may have arisen as early as the late Paleozoic, since certain Carboniferous fossils have been referred to Ampullaria. Such an early history is doubted by Pain (1972) who suggested that even Mesozoic taxa tentatively allied 2We use Ampullariidae in preference to Pilidae in accordance with Article 40 of the International Code of Zoological Nomenclature. 1977 Paleospecies of Neotropical Ampullariids 109 to Ampullaria actually belong to genera such as Natica. With certainty, precursors of Old World taxa, especially Lanistes, have been recorded in the late Eocene and early Oligocene of Egypt (Blackenhorn, 1900: 456; 1901; Cuvillier, 1930; Said, 1962). The Recent East African Pila ovata (Olivier) is also found in the early Miocene of Kenya (Verdcourt, 1963). Additionally, subfossil opercula of Pila have been recorded from the Indian subcontinent in Kashmir (Prashad, 1925). In the New World, the ampullariids are widely distributed in the tropical, subtropical, and even somewhat temperate climates (Pain, 1972). About a dozen species are known from Argentina and contiguous territory (Ihering, 1919; Hylton Scott, 1957). Apparently there are about fifteen species in the Amazon (Pain, 1960). The ampullariids of British Guiana (Pain, 1950), Venezuela (Baker, 1930; Pain and Arias, 1958), Central America (Bequaert, 1957; Pain, 1964), Greater Antillean islands including Cuba and Jamaica (Pilsbry, 1927), and the south¬ eastern United States (Clench and Turner, 1956) have received some attention. Although the total diversity of the ampullariids has been estimated at 75 species (Alderson, 1925; Boss, 1971; Pain, 1972), the exact number of New World forms is not known. There are probably about 20-25 species. The evolutionary history of the New World ampullariids is poorly documented (Parodiz, 1969). In South America, the known fossil spe¬ cies, including those described herein, are all probable precursors of living representatives. As will be seen subsequently, the New World genus Pomacea, a name that replaces, in part, the old Ampullaria for nomen- clatorial reasons (Pain, 1972), has identifiable phyletic roots as far back as the Oligocene, if not the late Eocene. Four rather common South American Recent species are either known as fossils or have paleospecies that are distinctly ancestral. These are Pomacea canaliculata (Lamarck), a predominately southern species found in Argentina, Uruguay, Paraguay, Brazil, and Bolivia (Hylton Scott, 1957); P. aurostoma (Reeve) principally a species with a north¬ western distribution in Venezuela and Colombia (Baker, 1930); P. glauca (Linnaeus), a wide-ranging species found in Colombia, Vene¬ zuela, Trinidad and the Lesser Antilles, as well as the Guianas (Pain, 1950); and P. urceus (Muller), with a very extensive range including Trinidad, Venezuela, the Guianas, the Amazonian Basin throughout Brazil and into Peru and Ecuador (Pain, 1950; 1960). The respective fossil populations or paleospecies that antedate the Recent species are: the Pliocene P. canaliculata of the Entre Rios For¬ mation at Parana, Argentina (Ihering, 1897; 1907); P. quaduasensis (Anderson, 1928), probably of Pleistocene age in the Guuaduas beds along the Magdalena River, Colombia; P. pattersoni, n. sp., from Tertiary Red Beds in the vicinity of Yarina, Peru; P. prourceus, n. sp., from the 110 Annals of Carnegie Museum vol. 46 Tertiary Red Beds of probable mid Eocene — early Oligocene age near Chazuta, Peru; and P. manco Pilsbry (1944), from Pliocene deposits along the Rio Pachitea, Peru. A SPURIOUS AMPULLARIID The Pomacea bib liana of Marshall and Bowles (1932: 4, pi. 1, figs. 4 & 5) from Biblian, Province of Canar, Ecuador, although supposedly a fresh-water form of the Cuenca Basin (Liddle and Palmer, 1941:395 and 398; Pilsbry, 1944:143-4), has been re-identified as a prosobranch terrestrial snail. Parodiz (1969:103 and 104) treated the species as Poteria (Pseudaperostoma) bib liana in the cyclophoracean family Aperostomatidae of mesogastropod prosobranchs, averring that although the species was found with other fresh-water genera (Diplodon, Doryssa, and Potamolithoides) it is actually a terrestrial snail which probably was carried into fresh-water deposits by heavy rains. Pomacea (Pomacea) prourceus, new species Figs. 1-4 holotype: MCZ 272899, a steinkern collected by B. Patterson in 1974. type-locality: Chicoca (a single farmhouse on the east bank of the Rio Huallaga3, teste, Patterson, 1942:2), east of Chasuta [Chazuta] (6°35’S; 76°11’W), the Rio Huallaga, Department of San Martin, Peru. Early Tertiary (possibly middle or late Eocene) bed of gray lodolite within the Red Beds (Capas Rojas) sequence (see Rosenzweig, 1951: 179), at the type-locality of the pyrotherian mammal, Griphodon peruvianus Anthony (Patter¬ son, 1942: 2). description: Internal shell cast somewhat distorted by compression, large, width 136 mm, height 115 mm, roundly globose, perforate with narrow, apparently slightly open and probably deep umbilicus; whorls probably 2 1/2 to 3 1/2 (though not visible through surface deposit), rapidly expanding, the last whorl nearly twice width of penultimate whorl; body whorl broadly convex, somewhat flattened above; aperture large, broadly and convexly rounded on the palatal margin; sinuate to subsigmoid on the columellar; peristome apparently gently rising; suture fine, sharply and deeply impressed; sutural margin of body whorl descending sharply toward peristome at inclination of about 45° to vertical axis; spire low and sunken (possibly by distortion during preservation; see the X-ray of P. urceus in Fig. 6), surrounded by higher margin of the body whorl; last portion of columellar section of body whorl missing but incremental area possible corresponding to thickened columellar callosity; sculpture consisting of slightly oblique, strong, axial costulae or ridges about 3 mm in width on body whorl but more closely spaced at suture; these lamellations or possibly growth lines only evident in portion of body whorl and disappearing towards aperture. remarks. Pomacea prourceus is noteworthy not only because of its unusually large size, probably as large as any living freshwater gastro¬ pod, but also for its obvious phyletic affinities with the Recent Pomacea urceus (Muller) (see Figs. 1-6). 3Following most authorities, we have used what appears to be the preferred spelling, Rio Huallaga, rather than the variant, Rio Huallago. 1977 Paleospecies of Neotropical Ampullariids 111 The Ampularia ? [s/c] gigantea Barbosa Rodriguez is larger, mea¬ suring 200 x 188 mm, and may belong in the lineage of Ampullaria prourceus — urceus. However, it has not been adequately diagnosed, figured, or cited in the subsequent literature. Further, the type-specimen is not available (Patterson, 1936), the type-locality is imprecise, and the geologic age of late Tertiary, probably Pliocene, time is unlike the earlier Tertiary occurrence of prourceus. Lastly, gigantea Barbosa Rod¬ rigues is a homonym of gigantea Tristram 1864 (see Pain 1964). The citation for this nomen should be Ampularia ? [s/c] gigantea Barbosa Rodrigues, 1892, Vellosia, 2: 52-53 [type-locality, avec YEmys macro- coccygeana ... a la meme epoque geologique . . .; dans les ravins des environs du Rio Nanay; P. 45, Loreto-Yacu, dans l’6tage tertiaire (teste Price, 1956: 2, “probably near Loreto on the upper Amazon in Peru, above the junction with the Rio Javari”)], non Tristram 1864. Even though P. prourceus is known only by a single cast4, an ancestral relationship to the modern urceus can be implied by several features that the species share. Relatively few but rapidly expanding whorls, heavily thickened columellar callosities, and the broadly opened aper¬ tures are representative of both species. Moreover, indicative of the evolutionary relationship between prourceus and urceus is the rather strongly developed axial sculpture, crowded and oblique along the sutural margin, and widely spaced and raised along the swollen whorl (Figs. 4 and 5). Although it can only be inferred, possibly prourceus possessed the heavy, blackened dehiscent periostracum of some popula¬ tions of urceus. As the early Tertiary stock from which P. urceus arose, P. prourceus is itself distinguished by its enormous proportions and somewhat low-spired outline. The Recent Pomacea urceus is widely distributed in the upper reaches of the Amazonian drainage in eastern Peru, Ecuador, and Colombia as well as throughout Brazil, the Guianas, Venezuela, and even Trinidad. Most of the four so-called subspecies of the “Group of Pomacea urceus” as discussed by Pain (1960: 426-28) are by definition, biological impos¬ sibilities. Locality data for urceus guayanensis (in the Ucayali River), urceus urceus (in the Rio Pachitea, a tributary of the Ucayali), and urceus yatesi (in the Rio Ucayali) indicate sympatry, so that the species is not polytypic. Although we have seen no living samples of urceus from the Rio Huallaga itself, the species is, as noted above, well documented in the Ucayali, which is confluent with Rio Maranon, to which Huallaga is a tributary. Thus, not only does P. urceus exhibit an extensive range in the Recent but it also has definite phyletic roots in ancestral prourceus 4A steinkern is, of course, smaller than the original shell and the spire more depressed. In the event of a shell thickened apically, the spire may be excavated or sunken. The X-ray of P. urceus, Fig. 6, shows that a steinkern of that species would have a somewhat depressed spire. 112 Annals of Carnegie Museum vol. 46 of the early Tertiary. The fossil species itself lived within the zoogeo¬ graphic range of the modern form, and though we cannot document further occurrences for prourceus, the species apparently existed in a considerable population, for Rosenzweig (1951: 179) and Patterson (pers. comm.) indicate considerable numbers of internal casts at the type-locality. Willard (1966, pi. 61, fig. 1) illustrated a specimen of ampullariid, Supposedly of Pliocene age, which he referred to as “Ampullina sp.” from the Aguaytia valley, off the Ucayali River. From its size (about 75 mm in height), general outline, wrinkled superficial sculpture, open umbilicus, and large ovate sperture, this individual appears to be Poma- cea urceus. If the sample is of Pliocene age, a doubtful point, since the specimen appears to be hardly fossilized, then the phyletic continuity of prourceus and the modern Recent urceus is clearly established in the region of the upper Amazon. The giant among living freshwater prosobranchs, Pomacea maculata Perry, frequently cited by its more descriptive junior synonym, gigas Spix, also lives in the drainage of the Amazon. Pain (1960: 423) recorded the species far upstream at Iquitos, Peru, on the Rio Maranon, to which the Rio Huallaga is a tributary. In comparison with P. prourceus, P. maculata has a narrower aperture with the upper margin of the peris¬ tome sharply rising from the body whorl. The columellar callosity is thin and poorly developed in P. maculata, and the sutures are very deeply excavated. Additionally, the axial sculpture of P. maculata is extremely fine, in marked contrast to the relatively strong costulae in the prour- ceus-urceus lineage. Pomacea (Effusa) pattersoni, new species Figs. 7-9 holotype: MCZ 272900, a steinkern; height 27 mm; width 35 mm; collected by B. Patterson in 1974. paratype: MCZ 272918, a steinkern; height 36 mm; width 42 mm; collected by B. Patterson in 1974. type locality: Vicinity of Yarina (6°17'S; 75°17'W), upstream from Isla Navarro, close to Rio Huallaga, Department of San Martin, Peru. description: Internal shell cast, medium size, width to 42 mm, height to 36 mm, roundly globose, perforate, with rather broad, open, and deep umbilicus; whorls 4-5 with the last, or body, whorl about twice the width of the penultimate; body whorl expanded, broadly convex, somewhat flattened to gently rounded above; aperture large, arcuately convex on the palatal margin; columellar margin simply angled about 45° to vertical axis; peristome gently and arcuately descending toward the palatal margin; suture subcanalicu- late and deeply incised; gently inclined toward the columellar margin of aperture; slight angulosity along inner margin of whorls at suture; spire low and flattened, but with profile of early whorls slightly raised above margins of body whorl; apex of spire appears to be slightly depressed; no indication of columellar callosity; shell probably smooth; no indi¬ cation of surface sculpture. 1977 Paleospecies of Neotropical Ampullariids 113 Figs. 1-3. Holotype of Pomacea prourceus Boss and Parodiz. Width 136 mm, height 115 mm, MCZ 272899. Fig. 1. Apical view. Fig. 2. Aper- tural view showing declination of attachment of body whorl and probable thickness of columellar callosity. Fig. 3. Umbilical view. 114 Annals of Carnegie Museum vol. 46 Figs. 4 and 5. Fig. 4. Enlargement of holotype of P. prourceus showing disposition of axial sculpture. Fig. 5. Enlargement of P. urceus (MCZ 108960) showing development of axial sculpture. 1977 Paleospecies of Neotropical Ampullariids 115 Fig. 6. X-ray of Pomacea urceus (Muller), 1 14 mm in height, MCZ 10896, from the vicinity of Ciudad Bolivar, Orinoco River, Venezuela, showing the strength of the axial sculpture and the apically thickened shell. 116 Annals of Carnegie Museum vol. 46 remarks:' Since the specimens of P. pattersoni are steinkerns, the shells of the living animals were larger, and the spires taller. In compari¬ son with the known Recent fauna of the Amazonian drainage, P. pat¬ tersoni seems to represent an element of the fauna of questionable systematic affinities, namely a representative of the subgenus Effusa. Ampullariid snails of medium size, which are widely umbilicate and broader than high, flattened to subconic shape and globose in outline, have been referred to Effusa (see, inter alia, Wenz, 1938; Baker, 1930). The type-species of this dubious grouping, P. glauca, is, according to Baker (1930: 12) the “most variable species I have ever studied.” He divided glauca into nine “forms.” Subsequently, Pain (1950: 69) con¬ sidered four varieties of glauca in British Guiana alone and noted that it had a very wide distribution in South America and the West Indies. He maintained that it is found in Venezuela, Colombia, the Guianas, Brazil, Trinidad, some of the Lesser Antilles and even Bolivia, but repeatedly asserted that Effusa is absent from the Amazon (Pain, 1960: 431). Yet fossil that it is, P. pattersoni exhibits the conchological features of Effusa, and though larger in size, is remarkably similar to Pomacea baeri Dautzenberg (1901: 312, PI. 9, Figs. 12 and 13), a nominal species that Pain (1960) overlooked. By its broad, open umbilicus, relatively low spire and globosity, P. baeri from the Rio Mixiollo (or Mishollo), (8°0FS, 76°39'W), a tributary of the Huallaga in the Department of San Martin, Peru, is also referable to Effusa. Except for its smaller size, purplish coloration, possibly greater whorl number, and broader umbilicus, P. baeri is very similar to P. pattersoni, notably in the absence of a columellar callosity and in the development of rather deeply incised sutures. Dautzenberg (1901: 312) compared his new species, baeri, with glauca Linnaeus and luteostoma Swainson, the latter synonymous with P. (Effusa) glauca (Pain, 1950: 71). By the structural features of its shell, P. baeri is not only referable to Effusa but is probably synonymous with P. glauca. The ancient lineage of these populations in the upper Amazonian drainage is represented by P. pattersoni. The so-called P. glauca variety, crocostoma Philippi (Pain, 1950: 71, Pl. 7, Fig. 6), is especially similar to P. pattersoni in its low profile and incised sutures. Of South American fossil ampullariids, P. pattersoni must be con¬ trasted with the nearly contemporaneous P. manco Pilsbry (1944: 145, PI. 11, Figs. 33 and 32) from Quebrada Sungarillo (9°20'S; 74°56'W) on the Rio Pachitea, Department of Huanuco, Peru. Parodiz (1969: 110) placed P. manco in the subgenus Limnopomus, a group characterized by somewhat taller shells (height greater than width or diameter), with smooth to microscopically sculptured whorls, and with a columella margin having a strong callosity that covers the umbilicus5 (Dali, 5Pilsbry described manco as being umbilicate. 1977 Paleospecies of Neotropical Ampullariids Apertural view. Fig. 9. Umbilical view. 118 Annals of Carnegie Museum vol. 46 1904; Wenz, 1938). P. patter soni is thus distinguished from P. manco by its depressed, broader-than-tall proportions and its broadly rounded aperture. Certain specimens, steinkerns, and fragments from later Tertiary outcrops near Iquitos, Peru, were identified as ampullariids by Greve (1938: 77) and might well have affinities with P. patter soni. Likewise, P. pattersoni might be one of the species of Pomacea from the Capas Rojas at Chicoca referred to by Rozenzweig (1951). The Ampullaria guaduasensis of Anderson (1928: 23, PL 1, Figs. 19 and 20), from the lower part of the Guaduas beds, not far from a coal-vein horizon, at San Juan de Rfo Seco on the east border of the upper valley of the Magdalena River, Colombia, though once thought to be of lower Tertiary age and possibly contemporaneous with P. manco of the Rfo Pachitea (Pilsbry, 1944: 145), is now considered Pleistocene (Parodiz, 1969: 38). Notwithstanding disagreements as to the ages of different formations, guaduasensis bears a resemblance to Holocene species like Pomacea aurostoma (‘Lea’, Reeve) — see Baker 1930: 8 — and differs from P. pattersoni by its tall spired shell with angulated whorls and callosity-covered umbilicus. The well-known Pomacea canaliculata Lamarck (Hylton Scott, 1957: 299), which has been recorded as a Pliocene fossil in the Entre Rios of Parana, Argentina (Ihering, 1897: 334; 1907: 465; Wenz, 1928: 2498) differs from P. pattersoni in its much larger aperture, less globose body whorl, and rather sharply pointed spire. Pleuroceridae Doryssa corrosensis (Pilsbry and Olsson) Figs. 10 and 11 Hemisinus (Basistoma) corrosensis Pilsbry and Olsson, 1935. Proc. Acad. Nat. Sci. Philadelphia 87: 12, PI. 2, Figs. 8 and 9 (type-locality, Sucio River, a branch of Llano River, SE of Infantas and W of the Magdalena River, Colombia; Los Corros Formation, Upper Eocene; holotype, ANSP 13092). Doryssa corrosensis (Pilsbry and Olsson). Parodiz, 1969, Ann. Carnegie Mus. 40: 136. The series of 8 specimens, taken from the vicinity of Yarina, upstream from Isla Navara, close to the Rio Huallaga, San Martin, Peru, in an outcrop apparently of Eocene age, constitutes the second known occur¬ rence of corrosensis Pilsbry and Olsson. With the exception of one specimen (Fig. 10), all lack the early whorls (Fig. 11). They measure 16-27 mm in height and 5.5 to 10.5 mm in width. The most complete specimen (Fig. 10) has 5 whorls. As with many other freshwater gastro¬ pods found in the early Tertiary strata of Peru, Ecuador, and Colombia, the fossils are steinkerns compressed by diastrophic pressure. The cord¬ like spiral lirations number about 6-8 per whorl and the sutures are fine and rather deeply incised. The whorls are somewhat flattened and the 1977 Paleospecies of Neotropical Ampullariids 119 aperture comparatively small. The dimensions of the steinkerns, as well as their general form, sculpture, and whorl number indicate their con- specificity with corrosensis. The taxonomy of freshwater prosobranch ‘melanian’ snails is com¬ plicated and confusing. Morrison (1954) proposed that several distinct family lineages arose from certain marine families that invaded fresh water. Further, the principal fresh-water families, the Pleuroceridae and Thiaridae, have decidedly different reproductive mechanisms, the former being gonochoristic and the latter parthenogenetic or protand- ric. Lacking anatomical characteristics, fossils can be assigned to genera only on the basis of their similarity to living forms whose systematic relationships are known. The paleospecies corrosensis was originally placed in the genus Hemisinus and the subgenus Basistoma Pilsbry and Olsson, 1935), which apparently are parthenogenetic thiarids and essentially synonymous with Aylacostoma Spix (Morrison, 1954: 376; Wenz, 1938: 718). Subsequently, Parodiz (1969) transferred corrosensis to the dioecious pleurocerine Doryssa on the basis of its morphological similarity to the type-species of Doryssa, Melania atra Bruguiere. The spiral sculpture of corrosensis, its flattened whorls, comparatively small aperture, and deeply incised sutures suggest an affinity with Doryssa. However, there is no indication in the steinkerns of corrosensis of the rather strong axial ribs found in Doryssa atra. In addition, the known specimens of corrosensis are about one-half the height of typical D. atra. Compared with the closely related, possibly derivative Doryssa bib- liana (Marshall and Bowles, 1932: 3) from the Miocene of Ecuador (see Liddle and Palmer, 1941: 400, p. 6, Figs. 1-12; and Parodiz, 1969: 134, PI. 15, Fig. 12, PI. 16, Figs. 6, 8, and 12), corrosensis has less globose whorls and a few more spiral lirations per whorl. The discovery of corrosensis in Peru extends the range of the species from the type-locality in Colombia and reveals that it had a wide dis¬ tribution in the Eocene, in effect similar to that of living species of Doryssa that occur from northwestern South America to northeastern Argentina. Oleacinidae Euglandina saccata (Pfeiffer) Fig. 15. Oleacina saccata Pfeiffer, 1860, Novitates Conchologicae. 2: 161, pi. 43, Figs. 3-4(Republique d’L’Equateur); 1861, Proc. Zool. Soc. London, p. 26 {In Republicae AEquatoriali; holotype, ?Cuming Coll’n, British Museum (Natural History). Tryon, 1885, Manual of Conchology (2) 1: 39, pi. 9, Fig. 22. Euglandina saccata (Pfeiffer). Pilsbry, 1907-1908. Manual of Conchology (2) 19: 180, PI. 20, Figs. 5 and 6. 120 Annals of Carnegie Museum vol. 46 The single Recent specimen obtained in the vicinity of Yarina, up¬ stream from Isla Navarro, close to the Rio Huallaga, San Martin, Peru, measures 38 mm in length and 17 mm in width. The aperture is 20 mm by 8 mm at the widest point. The smooth nepionic whorls that form a conic, tapering outline, number 3 1/2 and occupy 7 mm in extent. The axial sculpture starts at the middle of the fourth whorl and consists of rather course lirations or riblets, which number about 3 per mm on the penultimate and body whorls (i. e., 16 per 5 mm). Spiral sculpture, of extremely fine lirations crossing the axial riblets, is obsolete and hardly discernible on the body whorl, even at magnifications of 50 x. In regard to the obsolescence of the spiral sculpture, the Pleistocene specimen differs from the description of saccata given by Pilsbry (1907-8: 180), who noted that the finely incised spiral lines covered the last whorl throughout, and that the axial sculpture consisted of 7-10 riblets in a space of 5 mm on the last whorl. Indeed, the variable features of Eug- landina prompted Strebel (1875: 29, pi. 12, Figs. 47, 47a. b) to classify saccata as a “Zwischenform” related to the group of E. subvaricosa (Albers) of western South America. Of the nominal varieties of Eug- landina, saccata extends farthest to the south in Ecuador and Peru. Megalobulimidae Megalobulimus popelairianus (Nyst)6 Fig. 12 Bulimus popelairiana Nyst, 1845, Bull. Acad. Bruxelles, pt. 12, no. 7, p. 151, PI. 3, Fig. 5 (South America; holotype, Brussels Museum Natural History, No. 2738). Strophocheilus (Megalobulimus) popelairianus (Nyst). Bequaert, 1948, Bull. Mus. Comp. Zool. 100: 98, PI. 11, Fig. 5 (complete synonymy). This individual Recent specimen of average size (height 137 mm, width 80 mm) from the vicinity of Yarina, upstream from Isla Navarro, close to the Rio Huallaga, San Martin, Peru, is referable to M. pope¬ lairianus (Bequaert, 1948: 98), and corresponds to a variety from the region of Huallaga that Martens (1876: 4, PI. 140, Fig. 1) called ‘ tham - mianus\ which typically possesses a callous tubercle on the upper parietal angle. However, since this feature is variable and does not occur allopatrically, ‘ thammianus * is considered synonymous with pope¬ lairianus, a species rather widely distributed in Colombia, Ecuador, and Peru (Bequaert, 1948). There is an evident, but yet to be established. relationship with M. maximus (Sowerby), a polytypic species that has a more southerly range than M popelairianus. Although Bequaert (1948) regarded M pope¬ lairianus as a distinct species, he was inclined to believe that eventually 6Recently, on extensive anatomical evidence, Leme (1973) has recognized Megalobuli¬ mus as the sole genus in the family, the Megalobulimidae, as distinct from Stropho¬ cheilus and the Strophocheilidae. 1977 Paleospecies of Neotropical Ampullariids 121 it would be recognized as another subspecies of M. maximus, one with larger shells and a more northerly range. However, since neither the specimen at hand nor any additional documentation has been offered to substantiate Bequaert’s suggestion, we follow him and the more recent authorities (Morretes, 1952; Leme, 1973) in recognizing M. popelairianus as a distinct species. Orthalicidae Corona regalis (Hupe) Figs. 13 and 14 Bulimus regalis Hupe in Castelnau, 1857, Animaux nouveaux ou rares . . . FAmerique du Sud . . . (Le Bresil; types, teste Dance, 1966: 289, Mus. Nat. Hist. Nat. Paris). Liguus (Corona) regalis (Hupe). Pilsbry, 1899, Manual of Conchology (2) 12: 180, PI. 35, Figs. 13, 14, 19; PI. 34, Figs. 9-12; PI. 36a. Figs. 26, 27; PI. 33, Figs. 3 and 6. Three Recent specimens of this species were found in the vicinity of Yarina, upstream from Isla Navarro, close to the Rio Huallaga, San Martin, Peru. Two of this series were large adults (Figs. 13 and 14), one dextral and one sinistral; the third was a small dextral individual measuring 20 mm in height (MCZ 272904) with the typical, somewhat keeled body whorl. As Pilsbry (1899) pointed out, sinistral specimens are usually more abundant in C. regalis, a species of the upper Amazon that may be distinguished from its lowland relative, C. regina (Ferussac), by its white peristome, folded columellar callus, and lack of a dark band on the last whorl. Cyclophoridae Poteria inca (Orbigny) Figs. 16-18. Cyclostoma inca Orbigny, 1835. Mag. Zool. Cl. 5, no. 62, p. 29 (republica Boliviana); 1837, Voyage dans FAmerique meridionale . . . Mollusques, 5: 361, PI. 46, Figs. 21-23 (la Colombie, la partie orientale des Brezil dans les bois de Caxoeira et les montagnes du versant oriental des Andes boliviennes, dans les provinces de Yungas, d’Ayupaya, de Valle grande). Cyclostoma blanchetianum Moricand, 1836. Mem. Soc. Phys. Hist. Nat. Geneve 7: 442, PI. 2, Figs. 21-23 [dans les bois de la Caxoeira (about 60 miles NW of Bahia)]. Aperostoma (Aperostoma) viridulum Haas, 1952. Fieldiana, 34 (9): 113, Fig. 16 (Machu Picchu near Cuzco, on Rio Urubamba, Peru, 2100 meters altitude; holotype, Chicago Natural History Museum 38379). Aperostoma (Aperostoma) indecisum Haas, 1952. Fieldiana, 34 (9): 114, Fig. 17 (Conta- mano, Rio Ucayali, Peru, 200 meters altitude; Chicago Natural History Museum 38376). Aperostoma (Aperostoma) schunkei Haas, 1955. Fieldiana, 34 (35): 361, Fig. 70 (Chan- chamayo Valley; altitude 1200 meters; Chicago Natural History Museum 47080). Only a single specimen (MCZ 272903) was obtained from the vicinity of Yarina, upstream from Isla Navarro, close to the Rfo Huallaga, San Martin, Peru; it measures 12 mm in diameter and 7.8 mm in height, and has a rounded aperture about 5x5 mm. This individual lacks a 122 Annals of Carnegie Museum vol. 46 periostracum, an operculum, and any indication of the presence or absence of color bands. The systematics of the Cyclophoridae are in an appalling state. Most species have been described typologically without detailed locality data or without mention of any variation, and most supraspecific taxa are defined solely on conchological characters, e. g., even on such notably inconstant features as color bands (Bartsch and Morrison, 1942). About a dozen species are listed from Peru by Bartsch and Morrison (1942: 282), but at least several additions have been made by Haas (1952; 1955). Our tentative identification is based on several lots of Poteria in the collection of the MCZ, which were studied and arranged by J. C. Bequaert. Unfortunately, he never published his opinions on the Neotropical taxa of this family, but his ideas are incorporated in the partial synonymy of this species as delineated above. Our specimen is referable to the group of Cyclostoma inca Orbigny, as detailed by Baker (1923: 30). Bartsch and Morrison (1942: 267) referred to it as Apero- stoma (Aperostoma) inca, gave a partial synonymy, and figured an illustration of the shell copied from Sowerby (1843, PI. 24, Figs. 71-72). Other nominal ‘species’ that might be incorporated into the synonymy, if enough comparative material were available, include: A. boliviense Bartsch and Morrison [1942: 260, PI. 38, Figs. 7-9, from “Bolivia” and erroneously referred to as ‘bolivianum’ by Haas (1952: \ 13)], filio lira- turn Sowerby (see Bartsch and Morrison, 1942: 267, p. 39, Figs. 14-16) from Bogota, Colombia; manabense Bartsch and Morrison, 1942: 239, PI. 34, Figs. 10-12, from between Quevedo and Calcata, Manabi, Ecuador; and peruense Bartsh and Morrison, 1942: 245, PI. 35, Figs. 10-12 from “Peru.” 1977 Paleospecies of Neotropical Ampullariids 123 Figs. 10 and 11. Doryssa corrosensis (Pilsbry and Olsson). Fig. 10. Height 16 mm. Fig. 11. Height 26.7 mm, MCZ 272901. Fig. 12. Megalobulimus popelairianus (Nyst). Height 137 mm, width 80 mm, MCZ 272902. Figs. 13 and 14. Corona regalis (Hupe). Fig. 13. Height 69 mm. Fig. 14. Height 72 mm (MCZ 272906). Fig. 15. Euglandina saccata (Pfeiffer). Height 38 mm, width 17 mm, MCZ 272905. Figs. 16-18. Poteria (Neocyclotus) c. f inca (Orbigny). Diameter 12 mm, height 7.8 mm (MCZ 272903). 124 Annals of Carnegie Museum vol. 46 References Cited Alderson, E. G. 1925. Studies in Ampullaria. Heffer and Sons, Cambridge, 102 pp., 19 pis. Anderson, F. M. 1928. Notes on Lower Tertiary deposits of Colombia and their molluscan and foraminiferal fauna. Proc. Calif. Acad. Sci., (4), 17 (1): 1-29. Andrews, E. B. 1965. The functional anatomy of the mantle cavity, kidney and blood system of some pilid gastropods (Prosobranchia). J. Zool., 146: 70-94. Arkell, A. J. 1924. Ampullaria wernei Phil. J. Conch., 17: 154-155. Bahl, K. N. 1928. On the reproduction processes and development of Pila globosa (Swain- son). Part I. Copulation and oviposition. Mem. Indian Mus. (Calcutta), 9: 1-11. Baker, H. B. 1923. The Mollusca collected by the University of Michigan-Williamson Expedi¬ tion in Venezuela. Occ. Papers Mus. Zool., Univ. Michigan, No. 137: 1-58, 5 pis. 1930. The Mollusca collected by the Michigan-Williamson expedition in Vene¬ zuela, Pt. 6. Occ. Pap. Mus; Zool., Univ. Michigan, No. 210: 1-95. Barbosa Rodrigues, J. 1892. Les reptiles fossiles de la vallee de l’Amazone. Vellosia, Contributes do Museu Botanico do Amazonas, Volume Segundo, Archeologia, Paleon- tologia, 1885-1888 (Segunda Edi$ao), Rio de Janeiro, Impressa Nacional, 1892. Bartsch, P., and J. P. E. Morrison. 1942. Part 3. The cyclophorid mollusks of the mainland of America. In Torre, C. de la, P. Bartsch, and J. P. E. Morrison. The cyclophorid operculate land mollusks of America. U. S. Nat. Mus. Bull., 181: 142-293. Bequaert, J. C. 1948. Monograph of the Strophocheilidae, a Neotropical family of terrestrial mollusks. Bull. Mjlis. Comp. Zool., 100 (1): 210 pp., 32 pis. 1957. Biological investigations in the Selva Lacandona, Chiapas, Mexico. Bull. Mus. Comp. Zool. Harvard Univ., 116: 204-207. Blanckenhorn, M. 1900. Neues zur Geologie und Palaeontologie Aegyptens. Z. deutsch. geol. Ges., 53: 307-502. 1901. Nachtrage zur Kenntniss der Palaeogene in Aegypten. Ctbl. Min. Geol. Pal., 1901: 265-276. Boss, K. J. 1971. Critical estimate of the number of Recent Mollusca. Occ. Papers Moll., Harvard Univ., 3 (40): 81-135. 1974. Oblomovism in the Mollusca. Trans. Amer. Micros. Soc., 93 (4): 460-481. Burky, A. J. 1974. Growth and biomass production of an amphibious snail, Pomacea urceus (Muller), from the Venezuelan savannah. Proc. Malac. Soc. London, 41: 127-144, text figs. 1-2. Burky, A. J., J. Pacheco, and E. Pereyra. 1972. Temperature, water and respiratory regimes of an amphibious snail, Pomacea urceus (Muller), from the Venezuelan savannah. Biol. Bull., 143 (2):304-316. 1977 Paleospecies of Neotropical Ampullariids 125 Clench , W. J., and R. D. Turner. 1956. Freshwater mollusks of Alabama, Georgia, and Florida from the Escambia to the Suwannee River. Bull. Fla. State. Mus., Biol. Sci., 1 (3): 97-239, 9 pis. CUVILLIER, J. 1930. Revision du nummulitique egyptien. Mem. Inst. Egypte, 16: 1-372. Dall, W. H. 1904. Notes on the genus Ampullaria. J. Conch., 11: 50-55. Dance, S. P. 1966. Shell Collecting: An Illustrated History. Univ. Calif. Press, Berkeley, 344 pp., 35 pis. Dautzenberg, Ph. 1901. Descriptions de coquilles nouvelles rapportdes du Peron par M. Baer. Jour, de Conch., 49: 306-313, pi. 9. Greve, Leonard de. 1938. Eine Molluskenfauna aus dem Neogen von Iquitos am Oberen Amazonas in Peru. Abhand. Schweizerischen Palaeontologischen Gesellschaft, 61: 133 pp., 10 pis., 25 text-figs. Haas, F. 1948. Three new land shells from Peru. Fieldiana, Zoology, 31: 189-193, figs. 38-40. 1949. Land and fresh-water mollusks from Peru. Fieldiana, Zoology, 31 (28): 235-250, figs. 50-59. 1951. Remarks on and descriptions of South American non-marine shells. Fiel¬ diana, Zoology, 31 (46): 503-545, figs. 97-126. 1952. South American non-marine shells: further remarks and descriptions. Fieldiana, Zoology, 34 (9): 107-132, figs. 14-26. 1955. On some small collections of inland shells from South America. Fieldiana, Zoology, 34 (35): 361-387, figs. 70-84. Hylton Scott, M. L. 1957. Estudio morfologico y taxondmico de los Ampullaridos, de la Republica Argentina. Revista Museo Argent. Cienc. Nat. ‘Bernardino Rivadavia’ Ciencias zoologicas, 3: 233-333. Ihering, H. von. 1897. Os molluscos do terrenos terciarios da Patagonia. Rev. do Mus. Paulista, 2: 217-382. 1907. Les Mollusques fossiles du Tertiaire et du Crdtac 6 Sup6rieur de 1’ Argen¬ tine. An. Mus. Nac. Buenos Aires, 14: 1-611. 1919. Las especies de Ampullaria en la Republica Argentina y la historia del Rio de la Plata. Seccion IV. Zoologia. Primera Reunidn Nacional de la Sociedad Argentina de Ciencias Naturales, Tucum£n, 1916, pp. 329-350 (printed in Buenos Aires, 1919). Kobelt, W. 1911-1915. Ampullaria. In Martini and Chemnitz, Conchy lien-Cabined, (2): 1: sect. 2, pt. 20, 236 pp., pis. 22-79. Leme, J. L. M. 1973. Anatomy and systematics of the Neotropical Strophocheiloidea (Gastro¬ poda, Pulmonata) with the description of a new family. Arquivos de Zoologia, Mus. Zool. Univ. Sao Paulo, 23 (5): 295-337, 2 pis. Liddle, R. A., and K. v. W. Palmer. 1941. The geology and paleontology of the Cuenca-Azogues-Biblian Region, Provinces of Canar and Azuay, Ecuador. Bull. Amer. Paleont., 26:357-421. 126 Annals of Carnegie Museum vol. 46 Marshall, W. B., and E. O. Bowles. 1932. New fossil fresh-water mollusks from Ecuador. Proc. U. S. Nat. Mus., 83 (5): 1-7. Martens, E. von. 1876. Die Bulimus — Arten aus den Gruppe Borus. Novitates Conchologicae Series 1. Band 5, Abteilung 1. Lieferung: 50-51, 28 pp., pis. 138-141. Morretes, F. L. de. 1952. Novas especies Brasileiras da familia Strophocheilidae. Arquivos de Zoologia do Estado de Sao Paulo, 8 (4): 109-126, 4 pis. Morrison, J. P. E. 1954. The relationships of old and new world Melanias. Proc. U. S. Nat. Mus., 103: 357-394. Pain, T. 1950. Pomacea (Ampullariidae) of British Guiana. Proc. Malac. Soc. London, 28: 63-74, pis. 6-8. 1960. Pomacea (Ampullariidae) of the Amazon River System. J. Conch., 24 (12): 421-432. 1964. Pomacea flagellata complex in Central America. J. Conch., 25 (6): 224-231. 1972. The Ampullariidae, an historical survey. J. Conch., 27 (7): 453-462. Pain, T., and S. C. Arias. 1958. Description de una especie de Pomacea de Venezuela. Noved. cient. mus. hist. nat. la Salle (Ser. Zool.), 24: 5-11, pis. 1-2. Parodiz, J. J. 1969. The Tertiary non-marine Mollusca of South America. Ann. Carnegie Mus. Pittsburgh, 40: 1-242, 19 pis., 7 maps. Patterson, B. 1936. Caiman latirostris from the Pleistocene of Argentina, and a summary of South American Cenozoic Crocodilia. Herpetologica, 1: 43-54. 1942. Two Tertiary mammals from northern South America. American Mus. Novitates, No. 1173, 7 pp. Pelseneer, P. 1895. Prosobranches aeriens et Pulmones branchiferes. Arch. Biol. Paris, 14: 351-393. Perry, M. C. 1973. Ecological studies of the apple snail at lake Woodruff National Wildlife Refuge. Fla. Sci. [formerly Quart J. Fla. Acad. Sci.], 36 (1): 22-30. Pilsbry, H. A. 1899. American Bulimulidae: North American and Antillean Drymaeus, Leio- stracus, Orthalicinae and Amphibuliminae. In Tryon-Pilsbry. Manual of Conchology, (2) 12: i-iii, 258 pp., 64 pis. 1907-8. Oleancinidae, Ferussacidae. In Tryon-Pilsbry. Manual of Conchology, (2) 19: i-xxvii, 366 pp., 52 pis. 1927. Revision of the Ampullariidae of Jamaica and Cuba. Proc. Acad. Nat. Sci. Philadelphia, 79: 247-253, pis. 21-22. 1932. South American land and freshwater mollusks. VIII — collections of the Carriker-Roberts Peruvian Expedition of 1932. Proc. Acad. Nat. Sci. Philadelphia, 84: 387-402. 1944a. Molluscan fossils from the Rio Pachitea and vicinity in Eastern Peru. Proc. Acad. Nat. Sci. Philadelphia, 96: 137-153. 1944b. Peruvian land Mollusca — II. Nautilus, 57: 118-127, pi. 11. Pilsbry, H. A., and A. A. Olsson. 1935. Tertiary fresh-water mollusks of the Magdalena Embayment, Colombia. Proc. Acad. Nat. Sci. Philadelphia, 87: 7-39. 1977 Paleospecies of Neotropical Ampullariids 127 Prashad, B. 1925. On a fossil ampullariid from Poonch, Kashmir. Rec. geol. Surv. India, 56: 210-212, pi. 15. Price, L. I. 1956. Sobre a suposta presen$a de um anomodonte triassico no Alto Rio Ama¬ zonas. Dept. Nac. Prod. Min., Div. Geol. Min., Not. prelim, estud., no. 93: 1-10. Ramanan, V. V. 1903. On the respiratory and locomotory habits of Ampullaria globosa Swainson. J. Malacol., 10: 107-113. Rosenzweig, A. 1951. Reconocimiento gelogico en el curso medio del Rio Huallaga. Bol. Soc. Geol. Peru, 23: 155-189, map. Said, R. 1962. The Geology of Egypt. Elsevier, New York, 377 pp. Saxena, B. B. 1956. Ecological behavior of the apple snail Pi la globosa. J. Bombay Nat. Hist. Soc., 53: 733-739. Snyder, N. F. R., and H. A. Snyder. 1971. Defenses of the Florida apple snail Pomacea paludosa. Behavior, 40: 175-215. SOWERBY, G. B. II. 1843. Thesaurus Conchyliorum., Vol. 1, Cyclostoma, pp. 89-156, pis. 23-31. Strebel, H. 1875. Unter besonderer Beriicksichtigung der Fauna angrenzender Gebiete. Beitrag zur Kenntniss der Fauna Mexikanischer Land- und' Susswasser- Conchylien. II Theil, 58 pp., 15 pis. Hamburg, C. J. Herbst. Tristram, H. B. 1863 [1864]. Supplemental catalogue of terrestrial and fluviatile mollusks collected in Guatemala by O. Salvin. Proc. Zool. Soc. London, pp. 411-414. Troschel, F. H. 1845. Anatomie von Ampullaria urceus und iiber die Gattung Lanistes Montf. Archiv ftir Naturgeschichte, 11: 197-216. Verdcourt, B. 1963. The Miocene non-marine mollusca of Rusinga Island, Lake Victoria and other localities in Kenya. Paleontographica, 121 (A): 37 pp. Wenz, W. 1928. Fam. Ampullariidae. In Fossilium Catalogus, I: Animalia. VIII, Gastropoda extramarina tertiaria. Parte 38: 2498-2502. 1938. Gastropoda. In Handbuch der Palaozoologie, Berlin, Vol. 6, pt. 1, 948 pp., 2764 figs. Willard, B. 1966. The Harvey Bassler Collection of Peruvian Fossils. Lehigh University, Bethlehem, Pennsylvania, 104 pp., 75 pis. Back issues of many Annals of Carnegie Museum articles are available, and a few early complete volumes and parts are listed at half price. Orders and inquiries should be addressed to: Publications Secretary, Carnegie Museum, 4400 Forbes Avenue, Pittsburgh, Pa. 15213. wius UUWIF, LIBRAMV | S "A/ //' f Y'( / OCT 3 1977 u»0y AN NALS o/CARNECIE MUSEUM CARNEGIE MUSEUM OF NATURAL HISTORY 4400 FORBES AVENUE ® PITTSBURGH, PENNSYLVANIA 15213 VOLUME 46 SEPTEMBER 30, 1977 ARTICLE 10 A NEW PETALODONT (CHONDRICHTHYES, BRADYODONTI) FROM THE UPPER MISSISS1PPIAN OF MONTANA Richard Lund1 Research Associate Abstract It is apparent, from a brief review of the systematics of some petalodont teeth and dentitions, that they may be divided into five morphologic groups: the Ctenopty chius, the Chomatodus, the Climaxodus, the Petalodus, and the Pristodus. Heteropetalus elegantulus, new genus and species, from the Upper Mississippian Namurian A of Montana, is described. It is a small, hyostylic, operculate chondrichthyan, with a unique biserial biaxial pectoral fin, a single dorsal fin with a small, superficial spine, a fusiform body, and a diphycercal tail. There are at most two teeth in each tooth position in the jaws, except that parasymphysial teeth tend to be retained after replace¬ ment. The teeth resemble those of the Ctenoptychius and Petalodus groups of the Petalodontiformes. Heteropetalus, as the first well preserved petalodont to be described, indicates that Janassa cannot be considered as a typical petalodont in morphology or habitus. The morphologic character combinations found in Heteropetalus blur the distinctions between the recognized chondrichthyan subclasses Bradyodonti and Elasmobranchii. Introduction The Bear Gulch limestone of Fergus County, Montana, is a thick sequence of cyclically bedded, high silica, offshore bank deposits of upper Chesterian (Namurian A) age (Scott, 1973). The Bear Gulch faunal assemblage contains nine species of eumalacostracan Crustacea (F. Schram, pers. comm.), two genera of Conodontochordata (Melton and Scott, 1973; Scott, 1973), diverse worms, cephalopods, above fifty genera of bony and cartilaginous fish (Lund and Zangerl, 1974; Lund, ‘Department of Biology, Adelphi University, Garden City, N.Y. 1 1530. Submitted for publication May 3, 1977. 129 130 Annals of Carnegie Museum vol. 46 1974), and faule layers containing rare pelecypods, inarticulate, and spiny productid brachiopods. Geologic factors as well as faunal diversity show that this fauna inhabited a tropical bay similar to those occurring in the modern Red Sea (J. Horner, pers. comm.). A soft ooze bottom of very fine calcium carbonate and detrital silica was deposited quickly in a limited area. Seasonal storm disturbances of the bottom, causing death and imme¬ diate burial, probably account for the fine preservation of whole cartila¬ ginous fish and soft-bodied invertebrates. Among the chondrichthyan genera from the Bear Gulch limestone are four forms identifiable, by their dentitions, as petalodonts. Fossil petalodonts are known almost exclusively by isolated teeth and dental batteries (Woodward, 1891, 1919; David, 1944). Janassa (Jaekel, 1899; Malzahn, 1968), one partially preserved batoid-like Permian holomorph, is known. The petalodonts have a bradyodont tooth structure of unique design (Radinsky, 1961) and have been included as an order within the chondrichthyan subclass Bradyodonti on the basis of this histology (Woodward, 1921; Romer, 1966; Obruchev, 1967). A recent review of the Bradyodonti and their interrelationships (Lund, in press) reassesses this placement, and attempts to delineate bradyodont and petalodont diversity. This paper is a descriptive study of the most common and best- preserved petalodont of the Bear Gulch fauna at this time. acknowledgments: Many people have helped in the excavations in the Bear Gulch limestone, but I particularly thank William G. Melton, Jr., leader of the party, and John Horner. Ranchers Gilbert Schultz and Clifford Meadors have graciously allowed us use of their land. Work has been supported by many agencies, including grants from the Univer¬ sity of Montana and Carnegie Museum of Natural History, a National Science Foundation grant to Harold C. Scott and, more recently, N.S.F. grant number BMS 75-02720 to Richard Lund. abbreviations: CM indicates Carnegie Museum of Natural His¬ tory; MV, University of Montana vertebrate collections. systematic status: The Petalodonts have been elevated in status from a family (Woodward, 1921) to an order (Patterson, 1965; Obruchev, 1967) in the subclass Bradyodonti. Within the Petalodontiformes, whole dentitions are known from Chomatodus (Bear Gulch, unpubl.), Climaxo- dus (Woodward, 1919), Janassa (Jaekel, 1899), “Pristodus” (Davis, 1883), Megactenopetalus (David, 1944) and the genus described in this paper. The body form is known only from Janassa, (Jaekel, 1899) and Heteropetalus (this paper). Partial dentitions are known from Petalor- hynchus (Davis, 1883). The taxonomy of many of the genera of the Petalodontiformes is confused and badly in need of revision, particularly among the cteno- ptychians. It is beyond the scope of this paper to untie this gordian knot. The following section is intended only as a review of tooth mor- 1977 New Petalodont from Upper Mississippian of Montana 131 phology, and the generic names associated with various morphologies, among several selected genera. The ctenoptychians seem to be central to an understanding of the basic radiation of the Petalodontiformes, an observation first introduced by Agassiz’s reference to them as “oro- donts squeezed flat” (in Davis, 1883). The low peaked, tumid cusped crowns of orodont teeth, surmounting short, rectangular roots, closely resemble not only ctenoptychian teeth but also those of early hybodonts, leading to a virtual taxonomic stalemate on tooth-based studies (Romer, 1966; Obruchev, 1967). I do not expect the following review to accom¬ plish more than reintroduce several unjustly synonymized taxa for consideration in relation to the petalodonts described herein. Genus Ctenopty chius Agassiz, 1838 description: Tooth with crown compressed and sharply peaked, with a few deep serrations whose edges may be crenulated; limited heel with a few prominent lingual ridges, and a thick, truncate root (Fig. 1 E). comments: Petalodopsis (Davis, 1883) is apparently correctly synonymized with Ctenoptychius by Woodward (1891), although the former is Lower Carboniferous, the latter Upper Carboniferous. Agassiz wrote a manuscript in 1859, based upon the large Enniskillen, collection, that was never published. It was evidently circulated among workers on the petalodonts, for it is repeatedly cited as the source of the taxonomic revision of these fish. Pertinent taxonomic changes concern¬ ing Petalodus and Ctenoptychius are reviewed below. Genus Ctenopetalus Davis, 1881 type species: Petalodus serratus Owen, 1840-1845. P. serratus: McCoy, 1855. C. serratus: Agassiz, 1859 (ms). C. serratus: Morris and Roberts, 1862 (name only). C. serratus: Davis, 1881. C. serratus: Davis, 1883. Ctenoptychius serratus: Woodward, 1891. description: Margin of crown peaked in principal teeth, gently rounded in others, compressed, with a large number of truncate, incompletely serrate denticulations, cren¬ ulated at the summit. Anterior base line of crown sharply curved, with few lingual folds. Root long and obtuse (Fig. IF). comments: Woodward (1891) synonymized Ctenopetalus with Ctenoptychius, a genus that bears little resemblance to it either in crown or root characters. Ctenopetalus does resemble Petalodus closely, differ¬ ing chiefly in the serrations of the coronal margin. A form that seems to bridge many of the differences between Ctenoptychius, Ctenopetalus, and Petalodus (Davis, 1883) is Ctenoptychius lobatus Etheridge (1875), described as having a peaked crown with about nine to thirteen dentic¬ ulations in principal teeth, denticulations more numerous and more obtuse in the lateral teeth, the baseline of the crown sharply angulated, 132 Annals of Carnegie Museum vol. 46 Fig. 1. Petalodont dentitions. A, Climaxodus, presumed upper dentition; B, Pristodus falcatus, upper and lower; C. Megactenopetalus kaibabanus, presumed upper denti¬ tion; D. Janassa bituminosa; E. Ctenoptychius apicalus, tooth; F, Ctenopetalus serra- tus, tooth; G. Petalorhynchus psittacinus, symphysial family, H. Petalodus ohioen- sis, upper and lower teeth in occlusion. A, after Woodward; B, F, G, after Davis; C, Mus. Northern Arizona PPh 569; D, after Jaekel; E. American Museum of Natural History 1858; H, after Hansen. 1977 New Petalodont from Upper Mississippian of Montana 133 and the root elongate and produced to a blunt point, as in Petalodus. Lingual folds evidently were present. Petalodus ohioesis, according to recent work by M. Hansen (1968; pers. comm.) has alternating upper and lower teeth in shearing occlu¬ sion, a symphysial tooth in only one jaw followed by an estimated seven laterals, and gradual change in size and shape of the teeth posteriorly (Fig. 1H). There is no evidence for development of a Pristodus- like beak (Fig. IB). These relationships are very similar to that of Megac- tenopetalus (Fig. 1C), Ctenopty chius (Fig. IE), and the new form de¬ scribed below (Figs. 9, 10), as well as to the more modified Climaxodus (Fig. 1A), and seem to be primitive for the Petalodontiformes. A partial dentition of Petalorhynchus psittacinus Agassiz is described by Davis (1883). Petalorhynchus teeth are compressed, peaked, with a prom¬ inent heel and 4-5 folds both labially and lingually, and show prominent shearing wear along their crests. Long, prominent roots are present, at an angle to the heel. The partial dentition consists of a median tooth family of five teeth flanked on either side by a less sharply peaked fam¬ ily containing at least three teeth, (Davis, 1883). The median family consists of worn and successively smaller teeth, evidently retained and overlain by their replacements. The form and arrangement of the teeth are similar to those of Janassa (Davis, 1883; Jaekel, 1899). The teeth of Janassa are derivable from a Petalodus- like form by extreme emphasis upon the lingual heel and ridges, resulting in a tooth with an open Z-shaped vertical section (Fig. ID). The teeth are arranged in a median, symphysial row and three lateral rows in each jaw, with seven teeth in each row (Jaekel, 1899). Each tooth forms a rectangular plate in the jaw with a low anterior cutting rim. The teeth are not shed but are overlain by their replacements. The development of a ray-like dental battery is peculiar to this highly derived form, but is anticipated in the earlier Petalorhynchus. Fissodus teeth are shaped like those of Petalorhynchus but have their crowns divided into 2 or 3 serrations (Woodward, 1891). Serrated teeth occur at every level of heel and root development, as in the following genera. Genus Harpacodus Davis, 1881. type species: Harpacodus dentatus Agassiz, 1859 (Enniskillen ms). Harpacodus dentatus Davis, 1881. H. dentatus: Davis, 1883: 514. Ctenoptychius dentatus: Woodward, 1891. description: Crown thick and strong, with an almost straight contour across the tips of the strong serrations. The teeth have a constricted separation between the crown and the prominent, tumid base. comments: While Harpacodus is related to Ctenoptychius in the common possession of deeply serrated crowns and thick, relatively short roots, they are distinctly different taxa of generic rank, separated by 134 Annals of Carnegie Museum vol. 46 about the same morphologic distance from each other as from the following genus. Genus Callopristodus Traquair, 1888. type species: Ctenoptychius pectinatus Agassiz, 1838. Callopristodus pectinatus: Traquair, 1888, p. 85. Ageleodus pectinatus: Owen, 1867. definition: Crown low, compressed with an almost straight contour across the tips of the strong serrations. No lingual heel or folds, root long, fibrous, divided. comments: An entirely different mode of tooth organization is shown by the dentition of Pristodus falcatus Agassiz (Fig. IB) placed in a fam¬ ily apart from the petalodonts by Woodward (1891). The teeth are sharply peaked and compressed, but have prominent lingual heels. The teeth decrease progressively in height from each side of the mid¬ line, there being about 5 to 7 teeth in lateral series. Davis (1883) de¬ scribes them as having no distinct base or root, but as being ankylosed to the jaw. Woodward describes them as having a short base below the labial, vertical edge. The opposing jaw of Pristodus falcatus, in artic¬ ulation inside the margin of the jaw just described and thus assuredly a lower jaw (Davis, 1883; contra Woodward, 1891) consists of one, peaked, simple tooth (Fig. IB). Both Davis and Woodward comment upon the resemblance to the lower jaw of the recent teleost Diodon (Tetraodontiformes). The lower Permian Megactenopetalus kaibabanus (David, 1944; Ossian, 1976; Hansen, pers. comm.) seems to be closely related to Pristodus, differing in its considerably larger size (Fig. 1C). The teeth on a jaw of Megactenopetalus present a sharply serrated shearing blade strongly reminiscent of the Recent characid teleost Serrasalmus, or among modern sharks, a short-jawed Carcharodon, both highly preda¬ ceous fish that shear chunks of flesh from their prey. Megactenopetalus is an unfortunate, although valid name; it has little in common with Ctenopetalus. The known teeth and dentitions of petalodonts may be divided into five morphologic groups, characterized by Ctenoptychius, Chomatodus, Climaxodus, Petalodus, and Pristodus, respectively. The Ctenoptychius group has a relatively homodont dentition whose teeth have peaked, cuspidate and slightly compressed crowns, a lingual heel, folds slight, if present, and a short root. The Chomatodus group has a heterodont dentition of few tooth positions and marked elongation of the posterior teeth. Crowns range from arcuate to cuspidate in outline anteriorly, to low and swollen posteriorly, with a strong heel and ridges surrounding the teeth, and relatively short roots. Polyrhizodus, for example, has high- arched and sharply compressed anterior teeth with long roots, while the posterior teeth are elongate, very low and wide, with multiple short, finger-like roots (Obruchev, 1967). The Climaxodus group has a homo¬ dont dentition with relatively low, swollen crowns, peaked at the sym- 1977 New Petalodont from Upper Mississippi an of Montana 135 physis, and with a lingual heel and ridges and short roots. The sym- physial teeth occur in a tooth family. The remainder of the teeth occur in single series imbricated along the jaw margins. The Petalodus group has a homodont dentition with compressed crowns peaked anteriorly, a pronounced lingual heel and ridges, and long roots emerging from the lingual aspect of the teeth. The Pristodus group has a homodont denti¬ tion of sharply compressed crowns, a lingual heel, but no lingual ridges, and short roots emerging vertically below the labial aspect of the teeth. The teeth are found in single series, along the jaw margins, and acro- donty may occur in the adult. The Ctenoptychius group, including Lisgodus, Callopristodus, and Harpacodus, is seen as relatively primitive, with Desmiodus St. John and Worthen (1875) intermediate between it and the orodonts. Cteno- petalus seems to be a link to the Petalodus group. The Climaxodus and Chomatodus groups, like the Pristodus group, are short-rooted forms with divergent trends in dentitional specialization. The foregoing scheme will be nothing more than a working hypothe¬ sis, as it has been for the past century, until more petalodont holomorphs are known than the poorly preserved Janassa and the few from the Bear Gulch limestone. It has been suggested, however, that the batoid- like body form of Janassa may be typical of the Bradyodonts (e.g., Moy- Thomas and Miles, 1971). None of the four diverse Bear Gulch forms show any trends in this direction. The diversity of dentitional types in the Petalodontiformes clearly suggests an equal or greater diversity of adaptive body forms. Systematics Subclass Bradyodonti Order Petalodontiformes Genus Heteropetalus, new genus type species: Heteropetalus elegantulus, new species. diagnosis: Small petalodonts with six tooth positions on the anterior part of each mandible, and five tooth positions on the anterior of each palatoquadrate, alternating vertically with those of the mandibular series in shearing occlusion. All teeth have a ling¬ ual heel, no lingual ridges, and a distinct narrowing between the crown and the wide, long, root. The root tapers to a feather edge. Crowns of all but the parasymphysial teeth are acuminate, with two to three low, thick cusps anterior to the peak and finer cusps pos¬ teriorly. The shapes of the crowns vary from low and arcuate parasymphysially to acumi¬ nate in increasing, then decreasing crown height on the palatoquadrate. The mandibular crowns vary from low and arcuate parasymphysially, to triangular with virtual cusp suppres¬ sion, to arcuate, to acuminate in the posterior three. Two tooth generations are known, and parasymphysial teeth are not shed when replaced. The suspensorium is hyostylic, the palatoquadrate is long and low, articulating subor- bitally and at the ethmoid region with the neurocranium. There is a hyoid operculum of short, thin, widely-spaced rays. The body is fusiform, with a diphycercal tail. There is a single dorsal fin with a superficial, short unornamented hollow spine, followed by a long, single series of articulated radials. No anal fin is present. The pectoral girdle is high, with a suprascapular enlargement. The pectoral fin is narrow-based, with two 136 Annals of Carnegie Museum vol. 46 segmented axes distal to the glenoid articulation and fringed with articulated radials; the distal radials are bifurcated. Pelvic girdles are not fused in the midline. Simple claspers in males. No squamation is present except for lateral line canal ring scales. Secondary sexual differentation in the greater height of the dorsal fin and development of cartila¬ ginous plates and hooks from the last dorsal radials of the male. range: Bear Gulch Limestone, Namurian A of Montana, U.S.A. relationships: Heteropetalus differs from Ctenoptychius and Ctenopetalus in having a long, wide, feather-edged root, in lacking lingual ridges, in the manner and degree of cusp development, and in its heterodonty. Heteropetalus elegantulus new species holotype: MV 2778, complete male. referred specimens: MV 2779, complete fish plus anterior part of another; 2788; 2801; 2840 (two specimens); 3940. CM 23661-23666, 27282, 30622. horizon and locality: Upper Mississippian, Namurian A, Bear Gulch limestone, Big Snowy Group, south of Becket, Fergus County, Montana. diagnosis: Heteropetalus elegantulus is the only known species of its genus. Meristic characters are: 30-32 trunk segments, over 30 in the tail. Dorsal fin of 28-30 rays, most commonly 30; spine located at body segment 14 or 15; dorsal margin of fin straight, low, and radials unmodified in female; margin of fin increases in height posteriorly in males, with the penultimate 4-5-rays enlarged into plates and bifurcated distal to plates; last ray with proximal plate, distal hook. Caudal fin diphycercal, starting at segment 41, and of variable length; lower lobe of the caudal originates 3-4 segments behind origin of upper. Pectoral fin of 12 rays, variable numbers of which are bifurcated once; pelvic fin of 12 rays. Adults up to 12 cm in total length. DESCRIPTION body form: Heteropetalus elegantulus is a fusiform fish, blunt ante¬ riorly and tapering from the back of the head to the tip of the tail (Fig. 2). The orbit is slightly posterior to the center of the skull and slightly less than 1/3 the length of the nuerocranium. Preservation of fish in the Bear Gulch Limestone not infrequently reveals skin outlines, melanin pigment in the choroid coat of the eye, and degradation products of various organometallic compounds, e.g., hemoglobin, in the gill lamel¬ lae. No significant areas of melanin concentration can be found in the skin of H. elegantulus and no squamation is present in the skin except for the lateral line ring scales. The endoskeleton is well calcified, with the exception of the very fine branchiostegal rays and the pectoral fin. No traces of notochordal calcifications are present. There is little variation in size among the specimens. Well-developed sexual dimor¬ phism indicates that the fossils represent adult, sexually mature individuals. neurocranium: Preorbital, orbitotemporal, and otico-occipital parts of the neurocranium of Heteropetalus are approximately equal in length. There is no outstanding preorbital process. The dorsal surface is slightly narrowed orbitally, flares abruptly to the prominent postorbital process, and narrows sharply from there as it slopes to the foramen magnum. The pineal foramen lies at the level of the preorbital processes. The 1977 New Petalodont from Upper Mississippian of Montana 137 Fig. 2. Heteropetalus elegantulus, female, reconstruction. Head from MV 2778, body from MV 2801. 138 Annals of Carnegie Museum vol. 46 Fig. 3. Heteropetalus elegantulus, neurocranium. A, Dorsal view, MV 2779. B, ventral view, with palatoquadrate in lateral view, MV 2840. Scale is 1 cm. H, articulation for hyomandibular (stippled); E, endolymphatic foramen; O, olfactory cup; S, sub¬ orbital process; AU, otic bulla. 1977 New Petalodont from Upper Mississippi an of Montana 139 endolymphatic foramen lies in the dorsal midline at the level of the anterior margin of the postorbital processes, and is followed by a deep, open dorsal slit that extends to the foramen magnum. The swollen otic bullae project posterolaterally beyond the edge of the dorsal surface (Fig. 3A). In ventral view, the neurocranium is moderately narrow anterior to the otic region and wide postorbitally (Fig. 3B). The posterior wall of the orbit is formed by the postorbital process dorsally, and ven- trally by the otic wall. The tall, narrow hyomandibular articulates with the posteroventral surface of the postorbital process (Fig. 3, 5B). The ventrolateral surface of the ethmoid region evidently bore an antero- ventral facet for the preorbital region of the palatoquadrate, while a ventrolaterally projecting process of the suborbital shelf at midorbital level received the posterior margin of the suborbital process of the pala¬ toquadrate (Figs. 3, 5B). The floor of the neurocranium projects ante¬ riorly to form the floor of the precerebral fontanelle, the walls of which are similar to those of modern sharks, and are lower than in Helodus (Moy-Thomas and Miles, 1971). The otic bullae are invariably swollen, and contain brightly colored vestibules in these fossils, occupying most of the otic region of the braincase. The vestibules lie in the anterior part of the labyrinth (Fig. 4, 5) and cannot be subdivided into saccular and utricular regions. A median foramen for the endolymphatic ducts penetrates the dorsal surface of the neurocranium at the level of the postorbital process be¬ tween the anterior margins of the vestibules. A narrow, open slit (fon¬ tanelle) in the dorsal midline extends from immediately behind the endolymphatic foramen almost to the foramen magnum. This slit evidently corresponds to the posterior dorsal groove in Tamiobatis vetus- tus (Romer, 1964) “cladodus” wildungensis (Gross, 1937) and xena- canths. The vestibules contain fine clastic sedimentary particles, many of which are bright red and foreign to the Bear Gulch matrix. Evidently otoconia were not secreted by the fish but obtained externally through the endolymphatic foramen, and from an area other than where these adult fish were found. visceral arches: The mouth is subterminal and the jaws are long, extending to the posterior border of the orbit (Fig. 4). There appears to be a dependent fleshy upper “lip” in MV 2778. No calcified labial cartilages are present. Both palatoquadrates and mandibles meet in the midline but evidence of symphysial fusion is equivocal for the palato¬ quadrates and not likely for the mandibles. The palatoquadrate is a long element with a low stout preorbital process, a short sharp suborbital process, a low dorsally rounded quadrate process, and a broad, slightly convex quadrate condyle (Fig. 3B). The posterodorsal edge of the pre¬ orbital process evidently abuts against the ethmoid region of the neuro¬ cranium, while the suborbital process fits against an articular ridge on the suborbital shelf at midorbital level (Figs. 3-5). The rear margin 140 Annals of Carnegie Museum vol. 46 Fig. 4. Heteropetalus elegantulus, heads. A, MV 2778, holotype, lateral view, B. CM 30622A, dorsal aspect in ventral view. For scale and morphology see Fig. 5. 1977 New Petalodont from Upper Mississippian of Montana 141 ® &W Vt {'/ji I Fig. 5. Heteropetalus elegantulus, heads. A, 2778, holotype, lateral view. Double broken lines are lateral line canals, restored with reference to several specimens. Scale is 1 mm. B. CM 30622A, dorsal aspect in ventral view. Scale is 2 mm. Abbreviations: AV, anterior vertical semicircular canal; BR, branchiostegal rays; C, ceratohyal; D, posterior dorsal slit; E, eye; EN, endolymphatic duct; ET, ethmoid commissure; F, foramen magnum; G, gill filaments; H, hyomandibular; HO, horizontal semicircular canal; I, infraorbital canal; M, mandibular canal; MA, mandible; O, occipital commissure; P, postorbital process; PA, palatoquadrate; PI, pineal foramen; PV, posterior vertical semicircular canal; S, supraorbital canal; V, vestibule (otic bulla). 142 Annals of Carnegie Museum vol. 46 of the palatoquadrate is held against the ventral part of the long hyo- mandibular. The jaw suspension is hyostylic, with a significant degree of potential anteroposterior motion guided principally by the suborbital and ethmoid articulations. The articular fossa of the mandible is a slightly concave, almost horizontal surface at the posterodorsal corner of the element. The dorsal margin is straight between the articular process and a near-symphysial dorsal excavation where the dentition was located (Fig. 4). There is no retroarticular process. The ventral edge of the mandible slopes downward, then strongly forward to the highest part of the jaw, just posterior to the dental excavation. The latter is present on all well preserved lower jaws and is not likely to be an artifact of preservation. The upwardly concave mandibular dental profile prob¬ ably corresponds with a convex palatal profile (Figs. 4, 5). There is remarkably little surface area on the palatoquadrate upon which adduc¬ tor mandibularis musculature might originate, particularly in view of the long and deep mandible and anteriorly located dentition. Adductor musculature probably originated upon the posterior half of the palato¬ quadrate and inserted on the lateral surface of the mandible along a posteroventrally sloping ridge from the rear of the dental excavation to the posteroventral rim of the lower jaw. Preorbitalis musculature presumably was a significant factor in jaw movement. The hyomandibular is long and of uniform width, and articulates dorsally with a facet on the ventral surface of the postorbital process (Fig. 3). The ceratohyal is poorly exposed, but bears eight thin calcified branchiostegal rays (Fig. 5). There are no pigmented areas suggestive of a large hyoidean gill. The precise details of the four gill arches cannot be determined (Fig. 4B). Vascular tissue of the hemibranchs can be seen in several specimens (Fig. 4). There are no visible gill rays, and the distal portions of the hemibranchs are not attached to interbranchial septa. No pharyngeal denticles are visible. The branchial chamber is located below and behind the neurocranium, and is clearly covered by a predominantly membra¬ neous hyoid operculum. The mobile nature of the hyostylic articula¬ tion suggests somewhat protrusible jaws, secured dorsally by the subor¬ bital articulation, probably protracted by preorbitalis musculature, and actively retracted by extrinsic palatal musculature. The opercular flap is extensive, and as indicated by the branchiostegal rays, approaches confluence with that of the opposite side between the mandibles. lateral line system: The ring-encased lateral line system of the head consists of long supraorbital and infraorbital canals, which join at the anterior end of the snout through the rostral commisure (Fig. 5A). The infraorbital lateral line canal joins the mandibular canal to form the otic branch, which ascends to join the supraorbital canal at the post¬ orbital process. Immediately posteriorly, an occipital commisure crosses the top of the head at the start of the descent of the neurocranium to 1977 New Petalodont from Upper Mississippi an of Montana 143 the foramen magnum and behind the endolymphatic foramen. The lateral line trunk canal is unbranched to the tip of the tail. Heteropetalus bears only one mandibular lateral line canal, while both the chimaeroids (Patterson, 1965) and Chlamydoselachus bear two mandibular lateral line canals. axial skeleton: Neural arches and spines are uniform and calcified from the occiput to the vanishing point in the tail. Basapophyses in the trunk are short but always calcified, and haemal arches and spines are present to the vanishing point. It is all but impossible to count accu¬ rately the number of vertebral elements in the tail because either the neurals and haemals decrease in size to obscurity, the tail tip remains covered by matrix which cannot be removed, or the nutrient-seeking rootlets of sagebrush have dissolved the elements. Sagebrush is a par¬ ticular hazard to preservation of small cartilaginous fish and soft-bodied invertebrates. In the holotype the entire trunk (from the dashed line, Fig. 6) forward to the occiput was etched away by sagebrush roots. The few specimens suitable for meristic studies reveal 30-32 abdom¬ inal and 31 or more caudal segments, or 61 or more body segments vis¬ ible. The dorsal fin originates over segments 14 or 15, ends over body segments 35 to 37 and contains a spine plus 30 jointed radials, (one specimen contains 28). The epicercal portion of the caudal fin starts at segment 41, ends at about segment 51 and has approximately 21 unjointed radials. The hypocercal portion starts at about segment 43 and also ends at about segment 51. There is no anal fin. The dorsal fin is markedly sexually dimorphic. The fin of the female is uniformly low, witf^ no specialized radials (Fig. 2). The fin of the male (Fig. 6) is higher anteriorly and increases in height posteriorly. The last six rays are specialized as a clasper to hold the body of the female dur¬ ing copulation. The distal segment of each of the first four specialized radials is expanded into a triangular plate, and the posterior two plates each bear two distal, pointed segments. The fin culminates in a large calcified hook, supported in part by the penultimate radial and in part by a broad supine plate probably formed from the ultimate radial. Although basidorsals are not calcified, it must be assumed that they were present to support the fin radials and articulations as well as radial musculature. Sexually dimorphic dorsal fins are known in a variety of fish, both chondrichthyan (Lund, 1974) and osteichthyan (Sterba, 1962). Where present, the enlarged fin of the male may be used in display and court¬ ship as well as in the act of copulation. Male Cichlidae and Anabantidae (Teleostei) use the lengthened distal end of the dorsal fin to clasp the body of the female during amplexus (Sterba, 1962). There is no internal specialization of fin structure in these teleosts corresponding to that of Heteropetalus, nor is there internal fertilization. paired fins: The pectoral girdles are high and narrow and are not 144 Annals of Carnegie Museum vol. 46 Fig. 6. Heteropetalus elegantulus, pelvic region of holotype, MV 2778, male. Scale is 1 1977 New Petalodont from Upper Mississippian of Montana 145 fused in the midline. There is a spatulate suprascapular process dorsally and a narrow-based articular process posteroventrally. The pectoral fin (Fig. 8) is not readily comparable to that of any other chondrichthyan except Tristy chius arcuatus (Woodward, 1924; Moy- Thomas, 1936a). Poor calcification of the fin of H. elegantulus prevents completely certain rendition of the articulations of the elements, but there is no doubt about the basic structure. The fin is narrowly articu¬ lated with the glenoid fossa and has radials completely fringing the elongated fin base. The proximomesial radial is thickened and articu¬ lated only at its base. The remaining radials are segmented throughout their length, and the distal radials, particularly on the lateral edge, are basally bifurcated. The pattern of bifurcation is variable among the known specimens. The base of the fin is biaxial, composed of two parallel segmented axes. The topographically lateral axis seems to bear the principal arti¬ culatory basal element. The proximal end of the principal basal articu¬ lates with the articular process of the girdle, and its mesial surface bears the articulation with the first and second basals of the mesial axis in CM 23662 (Fig. 8). That this may represent a disturbed condition is indicated by MV 3940, where only the proximomesial basal articulates with the principal basal. The mesial axis has four serial basal elements and extends farther posteriorly than the three-element lateral axis. Two to three radials are borne upon articulatory processes of each basal, and the two axes may be in contact distally, but they are not fused. The pectoral fin is therefore broad and quite flexible. It is possible that the two axes of the fin are variable in position relative to each other and to the body surface, and were capable of being moved to more or less Fig. 7. Heteropetalus elegantulus, left pelvic girdle and fin of female, CM 23661 A. Scale is 1 mm. I, iliac process. 146 Annals of Carnegie Museum vol. 46 Fig. 8. Heteropetalus elegantulus, pectoral fin, CM 23662A. Scale is 1 mm. divergent positions for greater or lesser fin-surface area. Several speci¬ mens do show folding of the fin. The pelvic girdles (Figs. 6, 7) are separate, roughly triangular ele¬ ments with short dorsolateral (“iliac”) processes. The axis of the pelvic fin of the female has two subequal proximal elements and a short ter¬ minal element. That of the male has one long, followed by six short, basals and a short, simple, pointed myxopterygium. Twelve very fine radials constitute the support of the fin web, and are supported on the posterior edge of the girdle and two basals in both sexes. dentition: Crushing and fragmentation of these small fish renders interpretation of some aspects of the dentition difficult, particularly in the symphysial region. There are clearly two teeth in most tooth families, the replacement tooth being larger and more clearly cuspidate than the Fig. 9. Heteropetalus elegantulus, dentitions. A, MV 2788; B, MV 2801; C, CM 23665B; D, CM 23663A. Anterior to the right. See Fig. 10 for explanations. 1977 New Petalodont from Upper Mississippi an of Montana 147 148 Annals of Carnegie Museum vol. 46 labial tooth. The successional teeth of the symphysial region show dis¬ tinctive differences in shape and alignment of their tooth whorls, indi¬ cating changes of diet and age changes in occlusal relationships that cannot yet be properly evaluated. Nor can it be determined whether older parasymphysial teeth are shed, or as seems most likely, are moved into a different and less active functional relationship with the remain¬ der of the dentition. Roots of all teeth are long and wide, thinning to a feather edge. Fusion of the roots of teeth can be seen only at the para¬ symphysial tooth position, however. At the remainder of the tooth positions there is evidence of only two successional teeth and none of fusion (Figs. 9, 10). The lower jaw of Heteropetalus contains six tooth families per ramus, the rear margin of the posterior tooth vertically below the anterior mar¬ gin of the eye. The palatal dentition consists of five tooth families per ramus, alternating vertically with those of the mandibular series. The rear margin of the posterior palatal tooth lies slightly anterior to the corresponding mandibular tooth. All unworn teeth are acuminate, with low, thick cusps along the crest. The cusps are fewer and thicker anterior to the apices of the' teeth, except in mandibular tooth six. The lingual margins of the bases of mandibular teeth two through six, and all palatal teeth, turn outward and sharply upward near their anterior margins, producing an asymmetrical base with clear indications of anteroposterior orientation. All teeth except the parasymphisial ones lack a labial coro¬ nal surface, (Figs. 9 A, 11) indicating a pleurodont position on the jaws. This condition appears to be unique among known petalodonts. The first, or most anterior, mandibular tooth is small and symmetri¬ cal, with the lingual edge slightly curved around a central basin formed on the heel. A singularly long neck separates the crown of this tooth from the root. In larger specimens a larger, arcuate replacement tooth with short neck and gently cuspidate crown lies lingual and slightly posterior to the central axis of this tooth. The second mandibular tooth has a high, triangular crown peaked slightly posterior to the middle of the tooth base, and a lingual heel slightly thicker and lower posteriorly than anteriorly. Mandibular tooth three posteriorly is about two-thirds the height of the second, and arcuate in crown profile. The peak of the crest is slightly anterior to the center of the base. The basal heel turns inward and sharply upward near the anterior margin of the tooth. The slight basin rimmed by the heel lingually is thus located posterior to the center of the tooth. The fourth mandibular tooth is somewhat higher crowned, is one-third longer than mandibular three, and is peaked slightly posterior to its middle. The fifth mandibular tooth, the longest Fig. 10. Heteropetalus elegantulus, dentitions, drawn from Fig. 9. A, MV 2788; B, MV 2801; C, CM 23665B; D, CM 23663A. All scales are 1 mm. P, palatal tooth; M, mandibu- ^ lar tooth; L, left; R, right; numbers indicate tooth positions. 1977 New Petalodont from Upper Mississippian of Montana 149 LM3 150 Annals of Carnegie Museum vol. 46 of the series, is peaked posterior to the center; the heel is widest ante¬ riorly and turns up sharply at its anterior margin; and the cusps are fewer and coarser anteriorly. The sixth mandibular tooth is almost symmetri¬ cal, but is more coarsely cuspidate anteriorly; the heel turns up sharply at the anterior margin; and the tooth is less than half the length of mandibular five. A stout articular process emerges from the neck of the root of mandib¬ ular tooth three and palatal teeth four and five in MV 2788. No tooth is present on this specimen in articulation with any of the processes. A thick, arcuate tooth whorl consisting of the fused bases of two teeth, is found in the region of the palatal symphysis (Figs. 9B, D; 10B, D). The labial member is small, symmetrical, blunt and basined, with a distinct neck. The lingual member is larger, asymmetrically curved, and oriented lateral to the axis of symmetry of the first tooth. Two para- symphysial whorls may be found in larger specimens, while in smaller and presumably younger individuals, only a small, approximately sym¬ metrical tooth, arched in coronal outline and basined, may be seen. The second palatal tooth is large, with its coronal ridge thickened into an approximation of two cusps anterior to, and three finer cusps posterior to, the apex. The apex lies slightly posterior to the middle of the tooth, while the heel is lower and wider anteriorly. The apex of the third tooth, the largest of the palatal teeth, is anterior to its midline. Its basin is posterior. The apex of the fourth tooth is distinctively anterior to its midline. In some specimens the fourth tooth position is occupied by a tooth whose strongest cusp is only slightly higher than the remain¬ der of the anterior portion of the tooth, while in other specimens this position is filled by a larger tooth with a more prominent apex, believed to be a replacement tooth. If this is so, replacement of teeth proceeds Fig. 11. Heteropetalus elegantulus, mandibular tooth 2, lingual view, MV 2840, scale is .5 mm. 1977 New Petalodont from Upper Mississippian of Montana 151 from the front of the mouth posteriorly, as the larger, higher peaked and more rugged tooth is rarely found even in individuals with parasym- physial whorls. The fifth palatal tooth is almost as long as the fourth, but thinner and lower, with the apex slightly posterior to its midline. Mandibular teeth occlude between, and lingual to, the positions of palatal teeth, producing noticeable wear at the basins. Wear also oblit¬ erates cusps and sharp shearing edges of the teeth, but as wear facets are not developed upon the crests, abrasion by food may be primarily responsible for coronal wear. Discussion Comparison of Heteropetalus with Janassa, the only other petalodont known by more than dentition, is difficult in view of the incompleteness of Janassa, but the two forms differ strongly in dentition (Jaekel, 1899; Obruchev, 1967; Woodward, 1891, 1919, 1921). Janassa has broad, batoid-like pectoral fins of unknown internal structure, unlike the biaxial biserial fin of Heteropetalus. The pelvic fins of Janassa are supported by a complex series of girdle elements and a long segmented axis like that found in male Heteropetalus. The pelvic girdle of Hetero¬ petalus, however, is essentially similar in design to that of Ctenacanthus costellatus, Tristy chius arcuatus, Xenacanthus sessilis, and Helodus simplex (Moy-Thomas and Miles, 1971) and possibly should be con¬ sidered the primitive pelvic pattern for the Chondrichthyes. Squama- tion is present in Janassa, but is secondarily absent in Heteropetalus. The skull of Janassa is unknown. To summarize, Heteropetalus re¬ sembles Janassa only in the possession of long rooted teeth, and there is a greatly divergent dental pattern separating these two petalodont dentitions as well. The occlusal pattern of Heteropetalus is similar to that of other known petalodonts, Janassa expected, and it may be hypothesized that the dental occlusion, as well as the tooth form of petalodonts, sets them apart from all known Chondrichthyes. Janassa apparently represents a peculiarly derived, Myliobatis- like habitus in feeding and propulsive systems, and seems quite isolated from the bulk of the petalodont radiation. This is perhaps appropriate in view of its Permian age. The peak of petalodont diversity seems to have been reached in the upper Dinantian-Namurian. A brief review of dental morphology allows the formulation of a model primitive petalodont dentition. Teeth are seen as primitively cuspidate and slightly compressed, with lingual heels but no ridges, and relatively short, truncate roots. Teeth occur in vertically alternating, shearing position, and are homodont. 152 Annals of Carnegie Museum vol. 46 Heteropetalus is relatively primitive in crown development, retaining traces of a cuspidate condition, and lacking lingual folds. Sharply com¬ pressed and serrate teeth, as in Harpacodus and Callopristodus, and the smooth-to-crenulate-edged teeth in Pristodus and Megactenopetalus show a comparable lack of lingual fold development, while the coronal profile of forms attributable to Ctenoptychius approach that of some rear teeth of Heteropetalus. All are plesiomorphous in root develop¬ ment compared to Heteropetalus, whose long thin roots, and crowns, are morphologically intermediate between these groups and the Cteno- petalus- Petalodus group. The nature of the heterodonty shown by Heteropetalus is distinct from that of the other heterodont forms, Poly- rhizodus and Chomatodus, and divergent from the simple, homodont conditions of Pristodus or Petalodus. Heteropetalus seems to have diverged from a Ctenoptychius- like form prior to the evolution of lingual ridges. Heteropetalus elegantulus has a unique combination of elasmobran- chian and bradyodont morphologic characters. A free, long, low pala- toquadrate and hyostylic jaw suspension and a neurocranium with a low, open rostrum and a long posterior dorsal slit are characters which ally H. elegantulus with elasmobranchs. The presence of an apparent hyoid operculum and several aspects of dental morphology indicate possible relationships to the orodont bradyodonts. The rostrum of H. elegantulus is readily comparable to that of elas¬ mobranchs and Helodus, in contrast to the long, solid, closed rostrum of edestoids (Bendix-Almgreen, 1968) or the highly modified rostral region of the Chimaeroidei (Obruchev, 1967). The otico-occipital portion of the braincase descends abruptly from the postorbital process to the occiput without any prominent otic elongation as in Helodus (Moy- Thomas, 1936b) and Fadenia (Bendix-Almgreen, 1968) and in contrast to the elasmobranchian condition (Schaeffer, 1967). The postorbital process does not continue ventrally as a postorbital wall in the manner of all known Bradyodonti (Bendix-Almgreen, 1968). The endolymphatic foramen is at the level of the postorbital process, farther forward than in any known chondrichthyan, but closest to the condition in the Chimaeroidei. Posterior to the endolymphatic foramen, the open poste¬ rior dorsal slit extends to the occiput, as in Tamiobatis vetustus (Romer, 1964), (,cladodus” wildungensis (Gross, 1937) and the xenacanths. There is no counterpart of this structure in any known bradyodont. Dental morphology has led previous workers to propose a relation¬ ship between the petalodonts and the orodont-edestoid group (Agassiz, in Davis 1883, Jaekel, 1899). Histological sectioning of various Bear Gulch teeth, now in progress, reveals that some orodontoids have a thin layer of pallial dentine covering relatively short, branching dentinal osteons, as in Petalodus (Radinsky, 1961). Petalodont dentitions are characterized by very few members in each tooth position at any one 1977 New Petalodont from Upper Mississippian of Montana 153 time, tendencies to retain teeth after replacement in symphysial and near symphysial tooth families, and a shearing occlusion. Agassizodus and Desmiodus (St. John and Worthen, 1875) as well as Pennsylvanian and Permian edestoids reveal an increasing multitude of replacement teeth per tooth family and a high number of tooth families, forming an extensive crushing dental battery. All presently known relevant Bear Gulch fish support this divergence. The jaw suspension of the orodontoids and edestoids is not primi¬ tively holostylic. There seems to be no hyoid involvement in the jaw suspension of one Bear Gulch orodont; there is, however, a clear, com¬ pletely ray-supported operculum. There is little convincing evidence to demonstrate the relationship of the Petalodontiformes to either currently recognized subclass of the Chondrichthyes. Dental characters are presently as equivocal as dental histology (Radinsky, 1961) but adaptive trends in the petalodonts are unique among Paleozoic chondrichthyans. If the primitively non-sus¬ pensory nature of the hyoid arch is characteristic of the Bradyodonti (see Patterson, 1965), the petalodonts cannot be placed among the bradyodonts. The possession of an operculum excludes them from rela¬ tionship with any known elasmobranch. Neurocranial characters also support a position that is intermediate between known Elasmobranchii and Bradyodonti. Conclusions The dentition of Heteropetalus shows a primary shearing occlusion, with alternation in positions and shapes of upper and lower teeth. This pattern fits what is known about teeth and dentitions of other petalo¬ donts, and may be as characteristic of the group as the tooth shape itself. The teeth resemble those of the Ctenoptychius and Petalodus groups of the petalodonts. The pectoral fin morphology of Heteropetalus is unique among Chondrichthyes; its prevalence among petalodonts is not known. Janassa is seen as a highly derived late offshoot of the Petal¬ odus group. The presence of a hyostylic, operculate chondrichthyan blurs the most critical differences between the bradyodonts and the elasmobranchs. Other cranial characteristics of Heteropetalus also blur the distinctions between these two groups. The Petalodontiformes cannot be included in either subclass. 154 Annals of Carnegie Museum vol. 46 References Cited Agassiz, L. 1838-1843. Recherches sur les poissons fossiles. Neuchatel, vol. 3. Bendix-Almgreen, S. E. 1968. The bradyodont elasmobranchs and their affinities; a discussion. Almquist and Wiksell, Stockholm, Nobel Symp. 4:153-169. David, L. T. 1944. A Permian shark from the Grand Canyon. Jour. Paleont., 18:90-93. Davis, J. W. 1881. Notes on the fish remains of the Bone Bed of Aust, near Bristol. Quart. Jour. Geol. Soc. London, 37:419-426. 1883. On the fossil fishes of the Carboniferous Limestone Series of Great Britain. Sci. Trans. Royal Dublin Soc., (2), 1125: 327-600. Etheridge, R. 1875. On some undescribed Carboniferous fossils. Geol. Mag. 2:241-245. Gross, W. 1937. Das Kopfskellett von Cladodus wildungensis Jaekel. Senckenbergiana, 19:80-107. Hansen, M. C. 1968. The upper Paleozoic genus Petalodus (Bradyodonti) from N. America. Proc. Nebraska Acad. Sci., April 26-27. Jaekel, O. 1899. Ueber die Organisation der Petalodonten. A. Deutsch. Geol. Ges., 51(2): 258-298. Lund, R. 1974. Stethacanthus altonensis (Elasmobranchii) from the Bear Gulch limestone of Montana. Ann. Carnegie Mus., 45:161-178. [In New information on the evolution of the bradyodont Chondrichthyes. Field Press] Museum of Natural History. _ AND R. ZANGERL. 1975. Squatinactis montanus, a new elasmobranch from the upper Mississippian of Montana. Ann. Carnegie Mus., 45:43-54. Malzahn, E. 1968. Ueber neue Funde von Janassa Bituminosa (Schloth) im niederrhenischen Zechstein. Geol. Jahrb., Hannover, 85:67-96. McCoy, F. 1855. Description of the British Paleozoic fossils in the Geologic Museum of the University of Cambridge. Cambridge: 1-661. Melton, W. G. 1971. The Bear Gulch Fauna from central Montana. Proc. N. Amer. Paleont. Conv., 1:1202-1207. _ and Harold W. Scott. 1973. Conodont animals from the Bear Gulch limestone, Montana. In Conodont Paleoecology, Geol. Soc. America Spec. Paper 141. Morris, J., and G. E. Roberts. 1862. On the Carboniferous limestone of Oreton and Farlow, Clee Hills, Shrop¬ shire. Quart. Jour. Geol. Soc. London, 18:94-106. Moy-Thomas, J. A. 1936a. The structure and affinities of the fossil elasmobranch fishes from the lower Carboniferous rocks of Glencartholm, Eskdale. Proc. Zool. Soc. Lond. 1936:761-788. 1936b. On the structure and affinities of the Carboniferous cochliodont Helodus simplex. Geol. Mag., 73:488-503. _ and R. S. Miles. Paleozoic Fishes. Saunders, Phila. 259 pages. 1971. 1977 New Petalodont from Upper Mississippi an of Montana 155 Obruchev, D. V. 1967. Fundamentals of Paleontology. Vol. II, Agnatha, Pisces. Isreal Progr. Sci. Transl. Jerusalem, 825 pp. Ossian, C. R. 1976. Redescription of Megactenopetalus kaibabanus David, 1944 (Chondrich- thyes, Petalodontidae) with comments on its geographic and stratigraphic distribution. Jour. Paleo. 50:392-398. Owen, R. 1867. On the mandible and mandibular teeth of Cochliodonts. Geol. Mag., 4:59-65. Patterson, C. 1965. The phylogeny of the chimaeroids. Phil. Trans. Roy. Soc. London (B), 249 (757): 101-219. Radinsky, L. 1961. Tooth histology as a taxonomic criterion for cartilaginous fishes. Jour. Morph., 109:73-92. Romer, A. S. 1964. The braincase of the Paleozoic elasmobranch Tamiobatis. Bull. Mus. Comp. Zool. Harvard University. 131:87-105. 1966. Vertebrate Paleontology. 3rd Edition. U. of Chicago Press. Schaeffer, B. 1967. Comments on elasmobranch evolution. In P. W. Gilbert, et al., Sharks, Skates and Rays. Baltimore, Johns Hopkins Press:3-35. Scott, H. 1973. New Conodontochordata from the Bear Gulch limestone (Namurian, Mon¬ tana). Mus. Paleont. Michigan State Univ., Paleont., 1:81-100. Sterba, G. 1962. Freshwater fishes of the World. Vista Books, Longacre, London, 878 pp. St. John, O. H., and A. H. Worthen. 1875. Descriptions of fossil fishes. Geol. Survey Illinois, 6:245-488. Traquair, R. H. 1888. Notes on Carboniferous selachii. Geol. Mag. (3), 5(2):81-86. 1888a. Further notes on Carboniferous selachii. Geol. Mag. (3): 101-104. Woodward, A. S. 1891. Catalog of fossil fishes in the British Museum (Natural History), Vol. II. London, Taylor & Francis, 567 pp. 1919. On the dentition of the petalodont shark Climaxodus. Quart. Jour. Geol. Soc. London, 75(1): 1-6. 1921. Observations on some extinct elasmobranch fishes. Proc. Linn. Soc. London, 133:29-32. Woodward, A. H. 1924. A hybodont shark (Tristy chius). Quart. Journ. Geol. Soc. London. 80: 338-342. Zangerl, R. 1966. A new shark of the family Edestidae, Ornithoprion hertwigi, from the Pennsylvanian Mecca and Logan Quarry shales of Indiana. Fieldiana, Geol., 16:1-43. Back issues of many Annals of Carnegie Museum articles are available, and a few early complete volumes and parts are listed at half price. Orders and inquiries should be addressed to: Publications Secretary, Carnegie Museum, 4400 Forbes Avenue, Pittsburgh, Pa. 15213. ANNALS"™" of CARNEGIE MUSEUM CARNEGIE MUSEUM OF NATURAL HISTORY 4400 FORBES AVENUE • PITTSBURGH, PENNSYLVANIA 15213 VOLUME 46 OCTOBER 25, 1977 ARTICLE 11 MUS, COMP LIBRARY H0V7 1Q77 ISSN 0097-4463 harvard THE PLEISTOCENE (KANSAN) HERPETOFAUNA OF CUMBERLAND CAVE, MARYLAND J. Alan Holman1 Abstract The Cumberland Cave fossil site (Pleistocene: Kansan) has provided the only pre- Wisconsinan view of the herpetological life of North America east of the Mississippi and north of Florida. At least 30 species of salamanders, anurans, chelonians, lizards, and snakes are represented. The modern nature of the herpetofauna is striking, as only one taxon, a distinctive new species of Cryptobranchus, represents an extinct form. Most of the animals occur in the area today, with the exception of Ambystoma tigrinum, which is absent from most of the Appalachian Plateau, and Elaphe vulpina, which occurs in grassland areas mainly in the Great Lakes States today. Clemmys muhlenbergi is reported for the first time as a fossil. Introduction Recent excavations at Cumberland Cave, Allegany County, Mary¬ land, by personnel of Carnegie Museum of Natural History have resulted in the collection of a Kansan fauna consisting of at least eight sala¬ manders, seven anurans, three turtles, two lizards, and 10 snakes. These animals are described in the present paper. References to the Pleisto¬ cene fauna of Cumberland Cave include Gidley (1920), Hay (1923), Gidley and Gazin (1938), Nicholas (1953, 1954), Hibbard (1958), Rich¬ mond (1963), and Lynch (1966), but the herpetofauna heretofore has not been discussed in detail. Cumberland Cave lies four miles northwest of Cumberland, Alle¬ gany County, Maryland, at 31° 4' 30" N., 78° 47' 15" W. The most 'Museum, Michigan State University, East Lansing, Michigan 48824. Submitted for publication November 8, 1976. 157 158 Annals of Carnegie Museum vol. 46 comprehensive report to date (Gidley and Gazin, 1938) deals mainly with medium to large-sized mammals. That report also lists some un¬ identified snake vertebrae and a “crocodile tooth.” Richmond (1963) demonstrated that this element was the enamel cap of an unerupted bear canine, thus the “crocodile tooth” should be removed from the faunal list of the cave. Lynch (1966) listed two hylid frogs from the site. John E. Guilday of Carnegie Museum of Natural History kindly lent me the amphibian and reptile bones from the Cumberland Cave site for study, and provided (in litt., Nov. 10, 1975) information about the deposit. Based on mammalian remains, the age is equivalent to the Cudahy fauna (Kansan) of southwest Kansas, about 600,000 years BP. The mammalian fauna shows “western influences.” I should here like to thank Mr. Guilday for allowing me to study these interesting fossils. Merald Clark, Christine Kulczycki, and Susan McCarthy made the drawings. Following is a check-list of the Cumberland Cave fossil herpetofauna. Class Amphibia Order Caudata Family Cryptobranchidae Crypt obranchus guildayi, new species Family Ambystomatidae Ambystoma maculatum (Shaw) Ambystoma opacum (Gravenhorst) Ambystoma tigrinum (Green) Family Salamandridae Notophthalmus cf. N. viridescens (Rafinesque) Family Plethodontidae Plethodon glutinosus (Green) Gyrinophilus cf. G. porphyriticus (Green) Pseudotriton ruber (Sonnini) Order Salientia Family Bufonidae Bufo americanus Holbrook Bufo woodhousei fowleri Hinckley Family Hylidae Hyla crucifer Wied Pseudacris triseriata (Wied) Family Ranidae Rana clamitans Latreille Rana sylvatica Le Conte Rana cf. R. pipiens Schreber Class Reptilia Order Chelonia Family Emydidae Clemmys muhlenbergi (Schoepff) Terrapene Carolina (Linnaeus) Chrysemys picta (Schneider) Order Squamata Family Iguanidae Sceloporus cf. S. undulatus (Latreille) 1977 Herpetofauna of Cumberland Cave 159 Family Scincidae Eumeces cf. E. fasciatus (Linnaeus) Family Colubridae Natrix sipedon (Linnaeus) Thamnophis sp. indet. Heterodon cf. H. platyrhinos Latreille Diadophis punctatus (Linnaeus) Carphophis amoenus (Say) Coluber constrictor Linnaeus Opheodrys vernalis (Harlan) Elaphe vulpina (Baird and Girard) Lampropeltis triangulum (Lacepede) Family Viperidae Crotalus horridus Linnaeus Following is an annotated list of the fossil amphibians and reptiles of Cumberland Cave. Distributional data for modern amphibians and reptiles are from the maps of Conant (1975). Class Amphibia Order Caudata Family Cryptobranchidae Cryptobranchus guildayi, new species diagnosis: A Cryptobranchus differing strongly from the Recent Cryptobranchus alleganiensis (Daudin) of North America in having the Meckelian groove of the dentary much longer (runs the length of 34 teeth in C. guildayi; runs the length of 24 teeth in C. alleganiensis studied), in having the dentary weakly curved (very strongly curved in C. alleganiensis), and in having a deeper groove on its labial border. holotype: A left dentary, complete except for its anterior tip (Fig. 1), CM 20470. etymology: Named for John E. Guilday in recognition of his contributions to Pleisto¬ cene paleobiology. description of holotype: In dorsal view, the dentary is weakly curved. In lateral view, the Meckelian groove is very long and occupies a distance of 34 teeth and alveolar spaces. The Meckelian groove is narrowly V-shaped, moderately deep, and has borders that are much thicker than those of Cryptobranchus alleganiensis, the Recent form. There is a total of at least 41 teeth and alveolar spaces, but the anterior tip of the bone is missing, so that it is impossible to tell the total number of teeth and alveoli. All crowns of the teeth are broken olf, thus the teeth are represented only by their pedicellar portions. The remains of at least 29 teeth are discernible. In labial view, a moderate groove runs along the length of the mental foramina. Four mental foramina are present. The first three mental foramina are elongate. The last one is rounded. The height of the dentary through the pedicellar portions of the teeth at the level of the Meckelian groove is 5.4 mm. The length of the tooth row along the extent of the Meckelian groove is 18.4 mm. remarks: Westphal (1958) considers the Recent genus Megaloba- trachus of the Old World to be a synonym of the Miocene Andrias, 160 Annals of Carnegie Museum vol. 46 Fig. 1. Holotype left dentary of Cryptobranchus guildayi, new species, from the Pleistocene (Kansan) of Cumberland Cave, Maryland, in lingual view. The line equals 2 mm. 1977 Herpetofauna of Cumberland Cave 161 which occurs in both the Old and the New World. Cryptobranchus has heretofore been reported only from the Recent of North America, although a vertebra was reported by Meszoely (1966) to have been mistakenly identified as Cryptobranchus from the Pleistocene of Pennsylvania. Fossil and Recent cryptobranchid species are now as follows: Piceoerpeton willwoodensis Meszoely, Lower Eocene of Wyoming. Andrias scheuchzeri Tschudi, Middle Oligocene to lower Pliocene of Europe. Andrias japonicus (Temminck), Late Pleistocene of Japan to Recent of Japan and China. Andrias matthewi (Cook), Miocene of North America. Cryptobranchus guildayi Holman, Pleistocene (Kansan) of North America. Cryptobranchus alleganiensis (Daudin), Recent of North America. Cryptobranchus and Andrias japonicus may be distinguished from Andrias matthewi on the basis of the dentary in that (1) the height of the tooth row from the top of the pedicel to its base is one-half the total height of the jaw at the level of the anterior end of the Meckelian groove in Cryptobranchus and Andrias japonicus, whereas this height is only one-third the total height of the jaw in Andrias matthewi; and (2) the labial surface of the dentary along the mental foramina is grooved in Cryptobranchus and Andrias japonicus, but not in Andrias mat¬ thewi. The fossil is similar to Cryptobranchus alleganiensis and differs from Andrias japonicus in being smaller and in having a deeper labial groove of the dentary. The holotype of C. guildayi is larger than comparative Recent C. alleganiensis dentaries studied, thus I believe that morphological differences between the fossil and the Recent forms are not due to larval characteristics of the fossil. The Cumberland Cave specimen is the first prehistoric record of the family Cryptobranchidae from the Potomac River watershed and the entire Atlantic drainage east of the Appalachian Mountains. Although the Recent species now occurs in the Susquehanna River and its tribu¬ taries, this may be a relatively recent phenomenon. Its remains, although common in Indian food refuse collections from the Mississippi/ Ohio drainage in western Pennsylvania, are conspicuously absent from archaeological sites in the Susquehanna basin that date from c. 5000 B.C. to 1600 A.D. (J. E. Guilday, pers. comm.). Family Ambystomatidae Ambystoma maculatum (Shaw) material: Twenty trunk vertebrae, CM 29729. remarks: These vertebrae fit into the Ambystoma (Ambystoma) maculatum group of Tihen (1958), and on the basis of size and char¬ acters are inseparable from the Recent species A. maculatum. The earliest report of A. maculatum is from the early Pliocene (Clarendon- ian) of Kansas (Holman, 1975). This is the first record of A. maculatum 162 Annals of Carnegie Museum vol. 46 from the Kansan age. This species occurs in Allegany County, Mary¬ land, today. Ambystoma opacum (Gravenhorst) material: A trunk vertebra, CM 29985. remarks: In Ambystoma opacum the vertebra is shorter than in A. maculatum and A. jeffersonianum, but it is somewhat longer than in A. tigrinum, as pointed out by Tihen (1958). I have noticed that the pos¬ terior tip of the neural arch in A. opacum is more produced and rough¬ ened than in A. jeffersonianum and most A. maculatum. The fossil appears to be identical to modern A. opacum. This is the first record of A. opacum from the Kansan age. This species occurs in the area today. Ambystoma tigrinum (Green) material: Seven trunk vertebrae, CM 29730 (Fig. 2A). remarks: These vertebrae seem unquestionably assignable to the living species A. tigrinum, based on (1) their large size, (2) vertebral proportions (Tihen, 1958), (3) the neural arch extending posterior to the ends of the zygapophyses, and (4) the upswept neural arch in lateral view. The earliest fossil record of A. tigrinum is from the early Pliocene (Clarendonian) of Kansas (Holman, 1975). Ambystoma tigrinum is a very common Pleistocene fossil. Today this species is absent from most of the Appalachian area, including Allegany County, Maryland. Family Salamandridae Notophthalmus cf. N. viridescens (Rafinesque) material: Eight trunk vertebrae, CM 29728. remarks: These vertebrae are assigned to the genus Notophthalmus on the basis of vertebral criteria pointed out by Tihen (1974, pp. 211- 212). Many of the slender and delicate processes of these vertebrae are broken or eroded. But on comparable areas, I cannot distinguish the fossils from vertebrae of Recent skeletons of N. viridescens, a species that occurs in the area today. Several extinct species of Notophthalmus are known from Miocene deposits, one each being reported from Florida, Texas, and South Dakota (Tihen, 1974), but this is the first report of Notophthalmus as a Kansan fossil. Holman (1962) has reported N. viridescens from the Arredondo, Florida, deposit, a “Rancholabrean” site (Webb, 1974). I have been unable to find a way to separate the vertebrae of the eft stage of Notophthalmus from the newt stage. The fossils are difficult to measure because of all the broken and eroded surfaces, but they are just slightly smaller than individuals of both the eft and the newt stage of -N. viridescens from Florida and Indiana. Greatest lengths of six of the vertebrae are 2. 3-2.7 mm (mean, 2.47 mm). 1977 Herpetofauna of Cumberland Cave 163 Fig. 2. A, trunk vertebra of Ambystoma tigrinum in dorsal view; B, trunk vertebra of Elaphe vulpina in lateral view. Both from the Pleistocene (Kansan) of Cumberland Cave, Maryland. Each line equals 2 mm. 164 Annals of Carnegie Museum vol. 46 Family Plethodontidae Plethodon glutinosus (Green) material: Eight vertebrae, right dentary, femur, CM 29986. remarks: See Holman (1967) for remarks on the identification of Plethodon glutinosus on the basis of vertebrae and femora. The right dentary, in size and characters, appears to be identical to that of Recent Plethodon glutinosus. This is the earliest record of the species, as it previously has been reported only from Rancholabrean deposits in Florida (Holman, 1958, 1959) and Georgia (Holman, 1967). Plethodon glutinosus occurs in Allegany County, Maryland, today. Gyrinophilus cf. G. porphyriticus (Green) material: Twelve vertebrae, CM 29987. remarks: It is evident from observations of three skeltons of Pseu¬ dotriton ruber, two skeletons of P. montanus and two skeletons of Gyrinophilus porphyriticus, that there is an excellent way to distinguish the trunk vertebrae of Pseudotriton from those of Gyrinophilus. In Pseudotriton there is a high, rounded, neural spine at the level of the transverse processes. This spine is obsolete or absent in Gyrinophilus. The fossil vertebrae are tentatively assigned to G. porphyriticus, the species that occurs in the area today, although there are some slight differences between the fossil and the Recent vertebrae. This is the earliest record of this taxon as a fossil. Pseudotriton ruber (Sonnini) material: Eight vertebrae, CM 29988. remarks: Holman (1967) discusses the identification of P. ruber on the basis of individual vertebrae. Pseudotriton ruber occurs in the area today. This is the earliest record of this taxon as a fossil. Order Salientia Family Bufonidae Bufo americanus Holbrook material: Four left and seven right ilia, CM 29732. remarks: Holman (1967) and Wilson (1975) have found that the ilia of Bufo americanus may be separated from those of B. woodhousei on the basis that the base of the dorsal protuberance is wider in B. americanus than it is in B. woodhousei. The above 11 ilia have the B. americanus condition. One of the ilia is quite large and has the posterior height of the acetabular fossa 7.9 mm and the posterior height of the ilial shaft 4.7 mm. 1977 Herpetofauna of Cumberland Cave 165 Bufo woodhousei fowleri Hinckley material: Two left and one right ilia, CM 29989. remarks: These ilia have the base of the dorsal protuberance narrow, as in modern B. woodhousei. It is of interest that these fossil B. wood¬ housei have the low ilial protuberance of the eastern subspecies rather than the much higher protuberance of the western subspecies B. w. woodhousei. These subspecies have been distinguished several times before on the basis of the height of the dorsal protuberance (Eshelman, 1975). This is the earliest record for the subspecies B. w. fowleri. The earliest record for B. w. woodhousei is from the early Pleistocene (late Blancan) of north-central Kansas (Eshelman, 1975). Family Hylidae Hyla crucifer Wied material: Two right ilia, CM 12576 and 12577. remarks: These ilia were reported by Lynch (1966), who stated they represent individuals with about a 30 mm snout-vent length. Pseudacris triseriata (Wied) material: Left ilium, CM 12578. remarks: This ilium was reported by Lynch (1966). Family Ranidae Rana clamitans Latreille material: Left ilium, CM 29984. remarks: Rana clamitans has a steeper slope of the posterior portion of its ilial blade into the dorsal acetabular expansion than does Rana pipiens or Rana sylvatica; and it may be distinguished from Rana catesbeiana in having less rugged muscle attachments. The ilium rep¬ resents an immature individual, as the area anterior to the acetabular fossae is perforated. Rana sylvatica Le Conte material: Left ilium, CM 29990. remarks: This ilium is assigned to R. sylvatica, based on criteria discussed in Holman (1967). This is the earliest fossil record of this form. The wood frog is found in the area today. Rana cf. R. pipiens Schreber material: Nineteen left and seventeen right ilia, seven sacral vertebrae, CM 29733. remarks: These ilia appear identical to those of the northern leopard frog, Rana pipiens. As far as I can determine there are no consistent differences between the ilia of any of the “leopard frog” species ( R . pipiens, R. blairi, R. berlandieri, and R. utricularia), thus I shall only tentatively assign the Cumberland Cave ilia to the northern species Rana 166 Annals of Carnegie Museum vol. 46 pipiens. Rana pipiens- like ilia, known from as early as the early middle Miocene of Florida (Holman, 1965), are common Pliocene and Pleisto¬ cene fossils. The bones of the “leopard frog” species are so similar that it may be that the fossil record will not elucidate many of their relation¬ ships. Class Reptilia Order Chelonia Family Emydidae Clemmys muhlenbergi (Schoepff) material: Right hypoplastron with a portion of the right hypoplastron attached, CM 29991. remarks: This is the first fossil record of this turtle, which may be near extinction today. The hypoplastron differs from C. insculpta and is similar to C. guttata in having the anterior border of its inguinal shield truncated, whereas it is pointed in C. insculpta. The fossil hypo¬ plastron differs from C. guttata in shape, in that it is longer just medial to the inguinal scute than it is just anterior to the apex of the inguinal scute. In C. guttata the hypoplastral bone is shorter just medial to the inguinal scute than it is just anterior to the apex of the inguinal scute. The distribution of C. muhlenbergi is very spotty today because of destruction of habitats. It occurs within the general area of Cumberland Cave today, but there appears to be no records from Allegany County, Maryland. Terrapene Carolina (Linnaeus) material: Two right and two left humeri, right ilium, several fragmentary peri¬ pherals, costals, neurals, and plastral fragments, one cervical vertebra, CM 29992. remarks: The identification of Terrapene shell and limb elements has been discussed several times in the literature. I should like to point out here that the ilium of Terrapene appears also to be quite character¬ istic when compared with other North American genera like Chry- semys, Clemmys, Graptemys, and Pseudemys. In Terrapene it is boomer¬ ang-shaped; in the other genera it is much straighter. Most of the neural scutes have well-developed keels, an important character of Terrapene Carolina as opposed to Terrapene ornata. This species is found in the area today. Chrysemys picta (Schneider) material: Nuchal bone, right epiplastron, three peripheral fragments, CM 20467. remarks: The nuchal bone is identical to that of Recent C. picta and differs from other related forms in having the dorsal surface of the bone smooth, with the cervical scute unraised and with the anterior edges of the nuchal bone serrated. The other elements seem to be identical to C. picta, a form that is common in the area today. 1977 Herpetofauna of Cumberland Cave 167 Order Squamata Family Iguanidae Sceloporus cf. S. undulatus (Latreille) material: Posterior part of a right maxilla bearing three teeth, CM 29993. remarks: This fossil is almost identical to Recent specimens of S. undulatus, except that it is just slightly larger than the largest skeletons I have at hand. The most posterior tooth in the bone is weakly tri¬ cuspid. The anterior two teeth are unicuspid. Sceloporus undulatus occurs in the area today. Family Scincidae Eumeces cf. E. fasciatus (Linnaeus) material: A partial left dentary, CM 29735. remarks: I cannot distinguish the partial dentary in any way from Eumeces fasciatus, the blue-tailed skink that occurs in the area today. Family Colubridae Natrix sipedon (Linnaeus) material: Five trunk vertebrae, CM 29743. remarks: Holman (1972) discusses the basis on which individual vertebrae of Natrix sipedon may be distinguished from other species of Natrix. Natrix sipedon occurs in the area today. Thamnophis sp. indet. material: One hundred and thirty vertebrae, CM 29744. remarks: These elongate natricine vertebrae are assigned to Thamno¬ phis rather than to Natrix, based on the criteria of Brattstrom (1967). I am not able to carry the identification of these mainly fragmentary vertebrae to the species level. Thamnophis sirtalis and T. sauritus occur in the area today. Heterodon cf. H. platyrhinos Latreille material: Three trunk vertebrae, CM 29741. remarks: These fragmentary vertebrae have the depressed neural arch and wide, flat, hemal keel of Heterodon. They represent a much larger form than Recent Heterodon simus, thus they are assigned ten¬ tatively to H. platyrhinos, the species that occurs in the area today. Diadophis punctatus (Linnaeus) material: Three trunk vertebrae, CM 29737. remarks: The vertebrae of Diadophis and Carphophis are similar, but Diadophis has a slightly higher neural spine that does not slope gently into the neural arch posteriorly as in Carphophis. Moreover, the neural arch is a little more vaulted in Diadophis than in Carphophis. I cannot distinguish the Cumberland Cave vertebrae from those of Recent D. punctatus, a species that lives in the area today. 168 Annals of Carnegie Museum vol. 46 Carphophis amoenus (Say) material: Three trunk vertebrae, CM 29736. remarks: The vertebrae are not distinguishable from those of Car¬ phophis amoenus, the worm snake that occurs in the area today. Coluber constrictor Linnaeus material: Nineteen trunk vertebrae, CM 29742. remarks: These vertebrae are indistinguishable from Coluber constrictor, a form that occurs in the area today. Based mainly on geo¬ graphic grounds, the vertebrae are assigned to C. constrictor rather than to C. flagellum, a species that has very similar vertebral character¬ istics (see Auffenberg. 1963). Opheodrys v emails (Harlan) material: One trunk vertebra, CM 29740. remarks: Opheodrys vernalis has a higher, thinner neural spine and a more vaulted neural arch than do Diadophis or Carphophis. I am unable to distinguish the fossil from O. vernalis, the species that occurs in the area today. Elaphe vulpina (Baird and Girard) material: Sixteen vertebrae, CM 29738, (Fig. 2b). remarks: This species of Elaphe is readily distinguished from others in the genus on the basis of its lower neural spine. Today this species occurs as two disjunct subspecies. One occurs in the marshy lands around portions of Lake Erie and Lake Huron. The other occurs from northwestern Indiana and western northern Michigan through Wis¬ consin and northern Illinois, southern Minnesota, Iowa, and the gla¬ ciated regions of Missouri west to southeastern Nebraska. Fossils of E. vulpina have been found in Idaho in the upper Pliocene and in south¬ ern Missouri and northern Arkansas in the late Pleistocene (Holman, 1974). The early Pleistocene record of this species considerably extends its fossil range to the east. Lampropeltis triangulum (Lacepede) material: Eighty-seven vertebrae, CM 29739. remarks: Lampropeltis triangulum has vertebrae with low neural spines and depressed neural arches, by which they may be distinguished from other related forms. The- species is found in the area today. Family Viperidae Crotalus horridus Linnaeus material: Five fangs and three hundred and forty-eight vertebrae, CM 29745. remarks: The vertebrae are assigned to this species on the basis of characters given by Holman (1967). The fangs, although less diagnostic than the vertebrae, are also assigned to this species as I cannot distin¬ guish them from Recent fangs of this form. This rattlesnake occurs in the area today. 1977 Herpetofauna of Cumberland Cave 169 Discussion The majority of the amphibian and reptile species of the Cumberland Cave herpetofauna are extant today and are characteristic of temperate deciduous woodlands and the ponds and streams that would occur in low areas in these woodlands. But two forms are typical grassland species, and one of these has a range that is to the west of Allegany County, Maryland, today. The nearest the fox snake, Elaphe vulpina, occurs to the fossil site is on the eastern border of Lake Huron and the northern border of Lake Erie, where it occurs in marshy country near these lakes. The smooth greensnake, Opheodrys vernalis, still occurs in upland grassy meadows of the Cumberland Cave area. How these grassland species were able to encroach upon an essen¬ tially deciduous woodland herpetofauna in Kansan times is interesting to speculate upon. Schmidt (1938) and Smith (1957) have discussed a Prairie Peninsula Corridor which today consists of scattered areas of temperate grasslands surrounded by forests. It is thought that in the post-Pleistocene the North American Prairie extended eastward, wedged between the northeastern coniferous forest and the southwest¬ ern deciduous forest (Fig. 3). Thus all amphibians and reptiles that have an eastward range north of the Appalachian region, but are not present in the southwest, could have had their origin in the western prairie. The range of Elaphe vulpina seems to follow the remnants of the Prairie Peninsula Corridor today, and can be envisioned as having extended even farther east in Kansan times. Blandings turtle (Emy- doidea blandingii) follows the remnants of the Prairie Peninsula east¬ ward to the Boston area today (although there is a hiatus in the range west of Boston) and may have a present range somewhat similar to that of Elaphe vulpina in Kansan times. Although the tiger sala¬ mander, Ambystoma tigrinum, is widely distributed in the east, it is presently absent from almost all of the Appalachian Plateau and from the Cumberland Cave area. Perhaps, during the warmer dryer period of the early Kansan (Holman, MS), more extensive grassy areas occurred in the Cumberland Cave area, thus accounting for the occurrence of those species that are absent in the area today. 170 Annals of Carnegie Museum vol. 46 Fig. 3. The “Prairie Peninsula Corridor. 1977 Herpetofauna of Cumberland Cave 171 References Cited Auffenberg, W. 1963. The fossil snakes of Florida. Tulane Stud. Zool. 10:131-216. Brattstrom, B. H. 1967. A succession of Pliocene and Pleistocene snake faunas from the high plains of the United States. Copeia: 188-202. CONANT, R. 1975. A field guide to reptiles and amphibians of eastern and central North America. Houghton Mifflin Co., Boston, 429 pp. Eshelman, R. E. 1975. Geology and paleontology of the early Pleistocene (late Blancan) White Rock fauna from north-central Kansas. Mus. Paleont. Univ. Michigan Paps. Paleont. no. 13:1-60. Gidley, J. W. 1920. A Pleistocene cave deposit of western Maryland. Rept. Smithsonian Inst, for 1918:281-287. Gidley, J. W. and C. L. Gazin 1938. The Pleistocene vertebrate fauna from Cumberland Cave, Maryland. U. S. Nat. Mus. Bull. 171:1-99. Hay, O. P. 1923. The Pleistocene of North America and its vertebrated animals from the states east of the Mississippi River and from the Canadian provinces east of 95 degrees. Carnegie Inst. Washington Publ. 322:1-499. Hibbard, C. W. 1958. Summary of North American Pleistocene mammalian local Faunas. Pap. Michigan Acad. Sci., Arts, Letters 43:3-32. Holman, J. A. 1958. The Pleistocene herpetofauna of Saber-tooth Cave, Citrus County, Florida. Copeia:276-280. 1959. Amphibians and reptiles from the Pleistocene (Illinoian) of Williston, Florida. Copeia:96-102. 1962. Additional records of Florida Pleistocene amphibians and reptiles. Herpetologica 18:115-119. 1965. Early Miocene anurans from Florida. Quart. Jour. Florida Acad. Sci. 28:68-82. 1967. A Pleistocene herpetofauna from Ladds, Georgia. Bull. Georgia Acad. Sci. 25:154-166. 1972. Amphibians and reptiles. In Skinner, M. F. et. al. Early Pleistocene preglacial and glacial rocks and faunas of north-central Nebraska. Bull. American Mus. Nat. Hist. 148:55-71. 1974. A late Pleistocene herpetofauna from southwestern Missouri. Jour. Herpetology 8:343-346. 1975. Herpetofauna of the WaKeeney local fauna (lower Pliocene: Clarendonian) of Trego County, Kansas. Mus. Paleont. Univ. Michigan Paps. Paleont. No. 12:49-66. Lynch, J. D. 1966. Additional treefrogs (Hylidae) from the North American Pleistocene. Annals Carnegie Mus. 38:265-271. Meszoely, C. 1966. North American fossil cryptobranchid salamanders. American Midi. Nat. 75:495-515. Nicholas, Brother G. 1953. Recent paleontological discoveries from Cumberland Cave. Sci. Monthly 76:301- 305. 1954. Pleistocene ecology of Cumberland Bone Cave. Natl. Speleol. Soc. Bull. 16:29-39. 172 Annals of Carnegie Museum vol. 46 Richmond, N. D. 1963. Evidence against the existence of crocodiles in Virginia and Maryland during the Pleistocene. Proc. Biol. Soc. Washington 76:65-68. Schmidt, K. P. 1938. Herpetological evidence for the postglacial eastward extension of the steppe in North America. Ecology 19:396-407. Smith, P. W. 1957. An analysis of post-Wisconsin biogeography of the Prairie Peninsula region, based on vertebrate populations. Ecology 38:205-218. Tihen, J. A. 1958. Comments on osteology and phylogeny of ambystomatid salamanders. Bull. Florida State Mus. 3:1-50. 1974. Two new North American Miocene salamandrids. Jour. Herpetology 8:211-218. Webb, S. D. 1974. Chronology of Florida Pleistocene mammals. In Webb, S. D., Ed., Pleistocene mammals of Florida. Univ. Florida Press, Gainesville, 270 pp. Westphal, F. 1958. Die Tertiaren und Rezenten Eurasiatischen Riesensalamander (genus Andrias, Urodela, Amphibia). Palaeontographica 110:20-92. Wilson, V. V. 1975. The systematics and paleoecology of two late Pleistocene herpetofaunas from the southeastern United States. Doctoral Dissertation, Michigan State University, 67 pp. S -A/ri PM “tSBff- NOV] ittNt©)?7*4463 JKSEfiifft, AN NALS of CARNEGIE MUSEUM CARNEGIE MUSEUM OF NATURAL HISTORY 4400 FORBES AVENUE • PITTSBURGH, PENNSYLVANIA 15213 VOLUME 46 OCTOBER 27, 1977 ARTICLE 12 A LIST OF THE BUTTERFLIES OF ANDROS, BAHAMAS Harry K. Clench Section of Insects and Spiders Abstract The 59 species now known from Andros Island (largest of the Bahamas) are enumerated, with localities, period of flight, and remarks on habits and habitats. The introduction includes a history of collecting activity and descriptions of the different habitat types. Introduction When I made my first visit to Andros in the summer of 1973 only six¬ teen species of butterflies had been recorded from there. Now, three years and five visits later, the number stands at 59. A few more species may be expected eventually, but it seems appropriate now to summarize what we know about the butterflies of this large and fascinating island. The Bahamas have never been connected by dry land to any other land area, so the entire butterfly fauna of these islands has been built up by waif dispersal — oversea immigration from the adjacent land areas of Florida, Cuba, and Hispaniola. Grand Bahama and the Bimini Islands are the major “port-of-entry” islands (first successful landfall in the Bahamas) for the species originating in Florida; Great Inagua and the Turks and Caicos Islands the major port-of-entry islands for those coming from Hispaniola. The species arriving from Cuba, the major source of Bahamian butterflies, have many islands as potential ports-of- entry, but among them the most important by far, because of both prox¬ imity and size, is Andros. Waif dispersal is a chance process so we should Submitted for publication December 24, 1976. 173 174 Annals of Carnegie Museum vol. 46 expect to find that some species have reached Andros and have gone no farther, and that still others have extended their ranges beyond Andros to just one or a few other islands. This is precisely the case. Those known to occur only on Andros in the Bahamas include Battus devilliers, Eurema larae, and Astraptes xagua. Those that are known to occur on Andros and just a few other islands include: Papilio aristodemus (also on Cat Island) , Aphrissa neleis (also on New Providence), Apodemia carteri (also on New Providence and Little San Salvador), Urbanus dorantes (also on Grand Bahama), Pyrrhocalles antiqua (also on Eleuthera), Atalopedes mesogramma (also on New Providence), and Choranthus richmondi ( also on the Exuma Cays ) . HISTORICAL BACKGROUND The first butterflies collected on Andros were probably taken by C. J. Maynard. Of them, however, the only 'survivor I know of is the type specimen of Agraulis vanillae insularis, which he took on “Andros” [somewhere between Fresh Creek and Middle Bright (Turner, 1957: 143)] in 1887. It is still preserved in the collection of the Museum of Comparative Zoology at Harvard University. At about the turn of the century there was a flurry of interest and activity in Bahamian biology. Little of it had any bearing on Andros butterflies. But in May and June of 1904, William M. Wheeler collected somewhere on Andros for the American Museum of Natural History (cf. Rindge, 1952); and in July and August of that same year Owen Bryant collected, mostly on Mangrove Cay, for the Museum of Com¬ parative Zoology. Not until 1953 was any further collecting done on Andros. In that year the Van Voast — American Museum of Natural History expedition visited many of the Bahamas and on Andros took a limited number of butterflies, mostly on Mangrove Cay (Rindge, 1955) . In 1965, Neil D. Richmond, then Curator of Reptiles and Amphibians at Carnegie Museum, traveled extensively in the Bahamas. In addition to doing his own work, he collected butterflies for the Museum. These included specimens from North Andros as well as the first butterflies ever taken on Green and Scrub Cays. In 1973 I started collecting on Andros as part of a general survey of the butterflies of all the Bahamas. My first visit was in late June 1973, and my second in late November 1973, both in the vicinity of Nicolls Town, North Andros. The Andros Beach Hotel served as base. In the light of subsequent knowledge this was a happy choice because that area seems to be the richest and most diversified on the island. At the time, I picked the area simply because it was one of the least known. On my third trip, in early June 1974, I went to the vicinity of Driggs Hill on South Andros, 1977 Butterflies of Andros, Bahamas 175 working out of the Las Palmas Hotel, 3 km south of the town. Most of my collecting was within walking distance of the hotel, but I also made an all-day trip by boat, ably guided by Daniel Sturrup, some 19 to 24 km into South Bight, to the Yeho Pineyard. This is an extensive tract of virgin pine forest, with a small area of virgin scrub pineland where two of the rarest butterflies in the Bahamas, Calisto sibylla and Apodemia carteri, were taken. My fourth visit to Andros was brief. In late February 1 976 the Carnegie Museum Bahamas Expedition, of which I was a member, began its field work with a few days’ stay at the Forfar Field Station (of the International Field Studies Program), at Stafford Creek. Rose M. Blanchard, Director of the station, placed its facilities at our disposal and took us to several distant places of interest, including Red Bays and Twin Lakes Farm. From North Andros the Expedition went southward and into South Bight, visiting the Yeho Pineyard again. We then crossed the Tongue of the Ocean and spent a few days on Green Cay. The butterflies taken on this trip (made in a 56-foot, diesel-powered Carolina Fisherman, with stops at over 20 different islands in the Bahamas) will be listed in a separate paper, but the records obtained on Andros and on Green Cay are included here. In many ways my latest visit was the most interesting. It was made in late September and early October, 1976, and included 6 days at Nicolls Town and 9 days at the Forfar Field Station. Miss Blanchard and her colleagues, particularly Colleen Herpmann, Linda Fitzimmons, and Becky Lawson, were unfailingly helpful and hospitable and made my stay there as pleasant as it was profitable. While at the station I made an all-day trip with Mrs. Herpmann to some second-growth pine forest about 13 km south of Stafford Creek, and another with Miss Blanchard and Gus Adderley, who guided us some 56 road kilometers south of Twin Lakes Farm to collect at several virgin coppices or patches of hardwood ham¬ mock forest. On her first trip to Andros in June 1974, Miss Blanchard spent some of her time collecting butterflies at various North Andros localities. More recently, Donald J. Harvey, now a graduate student at the University of Florida (Gainesville), spent about a week in August, and a little over a a week in June and July 1977, collecting at Mastic Point, Stafford Creek, the vicinity of Nicolls Town, and other North Andros localities. I have examined both of their collections and the records are included here. ACKNOWLEDGEMENTS It is a pleasure to thank Miss Blanchard and her staff at the Forfar Field Station for their hospitality and manifold assistance during my two visits; Mr. Harvey and Miss Blanchard for allowing me to study the col- 176 Annals of Carnegie Museum vol. 46 lections they had made and to retain some specimens for the Museum collection; Neil D. Richmond, Curator of Environmental Studies at Carnegie Museum of National History for the collections he made for the Museum on Andros and on Green and Scrub Cays; Dr. William T. Gillis, Hope College, Holland, Michigan, for identifying a number of plants; Daniel Sturrup, of Long Bay Cays, Arlington Bastian, of Victoria Point, and Gus Adderley, of San Andros, Androsians all, who know their land and were excellent guides and companions. Finally, I thank my wife, Dr. Mary H. Clench, Associate Curator of Birds at Carnegie Museum, who has helped in so many ways : as field companion on many of my trips (including the 1976 expedition, which she directed), in our innumerable discussions of the island, and in critically reading this manu¬ script. The Carnegie Museum 1976 Bahamas Expedition was supported by the M. Graham Netting Research Fund of Carnegie Museum of Natural History and by a grant from the World Wildlife Fund. My other visits were supported by the Museum’s Holland Fund. Andros GENERAL GEOGRAPHY Largest of the Bahama Islands, Andros (Fig. 1) is some 117 km long and about 64 km in greatest breadth, comprising an area of about 5960 km2. Just south of its middle it is crossed by a network of shallow marine channels, North, Middle, and South Bights, which divide the land into a large number of close subsidiary islands, of which North Andros is by far the largest. Andros, in reality, is an archipelago. Andros lies on the Great Bahama Bank, near the western edge of the Bahamas and somewhat north of their middle. At its nearest point it is about 145 km from Cuba to the southwest and 201 km from Florida to the northwest. The nearest major island is New Providence, which is only about 35 km away, almost due east of the northern end of Andros. In the vicinity of Nicolls Town (or Nicholl’s Town, as it is alternately spelled), on North Andros, the coast consists of a narrow band, no more than about 100 meters wide and often less, of low dune sand, much of it with groves of Coconut Palms. To the west (inland) this gives way to a solid limestone surface, extensively pitted with potholes that may range up to several meters across, usually with relatively rich soil in them. The limestone extends westward, gradually diminishing in elevation, and the extensive pine forests on it eventually are replaced by large areas of mangrove along the nearly inaccessible (and biologically unexplored) western coast. Farther south, in many places (for example, at Stafford Creek), the limestone reaches eastward to the sea, pinching out the belt of sand. On South Andros the coastal band of dune sand is somewhat narrower. 1977 Butterflies of Andros, Bahamas 177 The Coconut Palms on it are more numerous, forming a shade dense enough in many places to virtually exclude other vegetation beneath. This belt of sand extends along the east coast of the island, at least as far south as Mars Bay, except for a few places where, as on North Andros, the rock Fig. 1. Map of Andros Island, Bahamas, showing localities mentioned in the text. Drawn by Nancy Perkins. 178 Annals of Carnegie Museum vol. 46 reaches the coast. Along the westward edge of the sand the ground surface is low and supports a narrow band of variously specialized, moisture- loving vegetation; in some places low, open moist meadows, in others transient fresh or brackish ponds with cattails and sedges. HABITAT TYPES Like most of the Bahamas, Andros has been inhabited for a long time. To what extent the original vegetation has been altered in the course of this habitation is uncertain. Dense scrub and coppice (hardwood hammock forest) seem to be found on the better soils, and in many places they have been cut over, perhaps several times, for farms. The pine forest has long been a source of timber for local shipbuilding, but this use has had an almost negligible effect on the forest. On North Andros, however, much of the original pine forest was logged off in the last decade, for pulp wood. From what I have been able to see, however, it is recovering well and in time may return substantially to its original state. Not being a botanist, however, I cannot tell whether this would be true of the understory vege¬ tation as well. Nonetheless, substantial areas of original vegetation may be found almost anywhere on Andros. I use the terms “disturbed” and “undisturbed” below in these specific ways: “disturbed” to mean showing signs of obvious, recent, major human interference; and “undisturbed” to mean the lack of such signs. The habitats divide readily into two types, open areas and wooded areas, to which I add a third, ecotonal areas. I. open areas: At least half the ground area is covered, sometimes sparsely, with grasses, forbs, or both, less than about 1 meter high. A few taller shrubs may be present. 1. Lawns: Low, regularly mowed grass around hotels and private homes. Highly disturbed. Planted ornamental flowers are often present. The soil is mostly sand or fill. Typical butterflies include: Battus polydamas, Nathalis iole, Eurema daira, Ascia monuste, Anartia jatrophae, Euptoieta hegesia, Agraulis vanillae, Hylephila phyleus. 2. Fields: Abandoned, not recently cut, lawns, and infrequently cut road aprons. These areas grow up in tall grasses, some flowering forbs, and an occasional shrub. In rocky areas the vegetation may be quite sparse. Highly disturbed, but some similar¬ appearing areas may be natural and undisturbed. Typical butterflies: Precis evarete, Precis coenia, Anartia jatrophae, Euptoieta hegesia, Euptoieta claudia, Agraulis vanillae, Strymon columella, Urbanus proteus, Urbanus dorantes, Hylephila phyleus, Wallengrenia misera, Euphyes Cornelius. 3. Coastal parkland: These open areas on the coastal dunes may be mostly natural. The vegetation consists of tall, clumped grasses and some sedges, Bidens, some Agave, scattered low shrubs, occasional Coccoloba clumps (usually with Spider Lily in the understory), and typically with occasional Coconut Palms, singly or in small groves. The few butterflies include Battus polydamas, Agraulis vanillae, Leptotes cassius. 4. Coastal old field: Much like the preceding but with no Coccoloba, the grasses lower, more clumped, sparser, and with more forbs (especially Bidens ), more Agave, and more shrub invaders. This is surely disturbed land (cut-over and probably at one time planted to lawns or crops), but butterflies are fairly numerous and varied, 1977 Butterflies of Andros, Bahamas 179 including Battus polydamas, Battus devilliers (scarce), Leptotes cassius, Hemiargus thomasi, Eumaeus atala (scarce), Agraulis vanillae, Euptoieta hegesia, Phyciodes frisia (scarce). B. devilliers and E. atala may be present solely because of proximity of the pine forest, which seems to be their more typical association. 5. Farm clearings : Much like No. 4 above but interior in location, always on a limestone substrate, and entirely surrounded by dense scrub or pine forest. The vegetation consists of low grasses and forbs, some crop plants (corn, sugar cane, bananas, sweet potatoes, beans, pigeon peas, and okra among them), mostly growing in potholes, with shrub invaders few or many, depending on recency of clearing. Newly cleared areas have relatively few butterflies; older clearings usually many. I distinguish two types of these clearings : (a) Scrub clearings: In dense scrub. When flowers are numerous, butterflies are also, especially where shrub admixture is present. Among the species occurring; Nathalis iole (only in very open areas of low vegetation), Phoebis agarithe, Phoebis sennae, Dryas iulia (usually in more shrubby areas and at the borders), Agraulis vanillae, Euptoieta hegesia, Phyciodes frisia (scarce), Chlorostrymon maesites, Electrostrymon angelia (the last two only on the flowers of emergent shrubs), Strymon columella (local, on Bidens flowers where the vegetation is low and sparse), Strymon martialis (infrequent, local, in shrubby areas), Leptotes cassius, Urbanus proteus. (b) Pine forest clearings: Not per se visually distinguishable from the pre¬ ceding, but differs in being surrounded by pine forest. Even in June, butterflies are scarce; the clearing I visited most (near Nicolls Town) had an estimated density of all butterflies only about 1/100 that observed in scrub clearings. 6. Low meadow: Lush, dense, broad-bladed grass, over 1 meter high, with occasional forbs and a few shrubs, on the landward slope of the dunes, bordered by dense scrub. Probably disturbed (it resembles acacia scrub — No. 13 below — with the acacias absent). The few butterflies include Dryas iulia (near the scrub border) and Euphyes Cornelius. 7. Sedge marsh: The dense scrub and pine forest are occasionally interrupted by openings, mostly rather small, in which fresh water is usually present, and with nearly solid stands of a tall sedge, about 1.5 m high. I have seen no associated butterflies at all. 8. Mangrove ‘‘ponds”: Areas, often extremely large, of limestone, the surface even and about at sea level, often extensively potholed or eroded to leave knife-edges and needle-sharp points, nearly always covered with a growth of Red Mangrove. The mangrove is stunted, sometimes less than 0.5 m high, but characteristically of even height throughout a given pond. Salt or brackish water is normally present, at least in the holes in the rock. The ponds are individually often quite large and collectively cover an appreciable percentage of Andros, but seem to be nearly sterile as far as butterflies are concerned. 9. Salinas: Shallow ponds or lakes of salt water, usually more or less cut off from the sea but subject to periodic inundation and later evaporation. Sometimes shallow shores duplicate most of the essential features of salinas. Vegetation is usually sparse and low (rarely over 0.3 m high) and typically includes Sporobolus grass, Batis, and Sesuvium portulacastrum, the pink flowers of the latter being an important food source for the few butterflies that occur. I have seen no true salina on Andros. Some shallow-slope eastern shores have some of the features (and some of the butterflies), and I should not be surprised if true salinas do occur somewhere on the island. Typical butterflies are few, among them Panoquina panoquinoides which is found almost exclusively in such places, or in their near vicinity. Another typical inhabitant, Brephidium exilis, has not (as yet) been found on Andros, although it does occur on Green Cay. An occasional butterfly of salinas is Ascia monuste. 180 Annals of Carnegie Museum vol. 46 II. wooded areas: More than half the ground area, and usually all of it, is covered by woody shrubs or trees. 10. Pine forest: Probably the most extensive single vegetation type on Andros, evidently always on a limestone substrate. It occurs in a wide range of variants, depending on recency of cutting, burning, or other disturbances, as well as on sub¬ strate variation. Large areas of this forest on North Andros have been logged off recently, mostly for pulp. The principal variants I have seen are: (a) Undisturbed: Scattered pines (Pinus caribbea var. bahamensis), up to 8 or 10 meters high and about 8 to 10 meters apart, to perhaps 15 meters high and only 3 to 5 meters apart, trunks up to a maximum of about 30 cm diameter at breast height. The age distribution of the pines is broad, from seedlings to old trees. Trunks may or may not be fire charred. A dense, often nearly impenetrable, shrub stratum is present, about 2 meters high, with occasional still higher emergents, comprising a variety of broad-leaved shrubs (but flowering species infrequent), occasional fan palms, and a frequent, tall (up to 2 meters) fern, Pteridium caudatum. Butterflies are few, but include some of the more interesting species: Battus devilliers (scarce), Phoebis neleis, Eureme dina, Eureme larae, Eurema messalina (scarce), Hemiargus ammon (at times common, and the only blue seen there), occasional Dryas iulia, and Ephyriades brunnea. (b) Cut-over: Extremely variable in nature, depending on the recency of cutting and of fire, which is quite frequent. In recently cut-over areas the only remaining pines are seed trees, which I was told run about 5 to the acre (12 per hectare). Frequent fires leave the ground surface with only low plants, among which is the abundant orchid, Bletia purpurea. Butterflies are surprisingly diverse, although rarely numerous, and include Battus devilliers, Euremea larae, Hemiargus ammon, Eumaeus atala, Urbanus dor antes, Euptoieta hegesia, and even an occasional Euptoieta claudia. (c) Pine scrubland: Because of unfavorable ground conditions the pines are rather sparse and low, and of little or no commercial value. Concomitantly the scrub understory is also low and rather sparse, sometimes with grasses. Such scrubland is virgin and likely to remain so. I have seen it at the Yeho Pineyard on low, level limestone where sea water was only a meter or so away and perhaps only a few centimeters below the surface. Near Stafford Creek it is on rocky ridges which maybe too porous to retain much moisture. Pine scrubland is not conspicuous (it is easily mistaken for cut-over pine forest), but well repays persistent collecting as it is the principal habitat of a number of unusual species. Among these are: Calisto sibylla, Apodemia carteri, Choranthus richmondi, Pyrrhocalles antigua, Eurema larae, and an occasional Battus devilliers. 11. Low hammock forest: Comprises low trees, up to 5 or 6 meters high, near the coast (where it is probably selectively cut if not entirely second-growth), up to 10 meters or more high inland. Component trees include mahogany or madeira, horseflesh, and other valued timber trees. The forest has a closed canopy, but little or no understory, and is readily traversible on foot. Bromeliads are often present, generally within about 0.3 m of the ground, and an orchid, perhaps Epidendrum, occurs occasionally. Hammock forest never seems to be continuous, but is rather in scattered patches of varying size, called coppices, or locally, “coppits.” These coppices are not frequent, and hammock forest is probably an endangered plant formation because it grows on a rich soil that is eminently suited for agriculture. I have only a little experience collecting in this vegetation type, and it has produced nothing of interest (an occasional Papilio andraemon or Marpesia eleuchea ), but it might repay more prolonged and intensive study. 12. Dense scrub: Mixed broadleaf shrubs with no understory, often dense and nearly impenetrable, with occasional emergents. The botanical composition is diverse, 1977 Butterflies of Andros, Bahamas 181 and dense scrub probably could be divided into a number of compositional types. Physiognomically dense scrub is also capable of some subdivision, but as the grades are continuous there is little point in doing so. The major physiognomic variable is height. This decreases slightly from north to south on Andros: in the Nicolls Town area the dense scrub is about 2.5 m high on the average; at Stafford Creek it is about the same; at Fresh Creek it is perceptibly less. On South Andros it averages about 1.5 to 2 m high and it was evident, on a drive southward as far as Mars Bay, that the height decreases still more southward. In a general way, and with local exceptions, scrub height seems to be correlated with precipitation (a conclusion derived not just from Andros but from the whole of the Bahamas), so I presume that precipitation on Andros gradually decreases from north to south. On Andros, dense scrub forms a band of varying width along the eastern coast, on a limestone substrate just west of the dune belt. Still farther west it gives way to pine forest. Dense scrub is the most prevalent physiognomic type of vegetation in the Bahamas, although it varies greatly in height, in density, and of course in compo¬ sition. It is also the vegetation type with the greatest number of associated butterfly species. Among them, on Andros, are: Papilio andraemon, Papilio aristodemus, Phoebis agarithe, Eurema messalina, Eurema dina, Eurema larae, Dryas iulia, Mar- pesia eleuchea, Heliconius charitonius, Chlorostrymon maesites, Electro strymon angelia, Epargyreus zestos, Atalopedes mesogramma, Urbanus dorantes, Ephyriades brunnea. 13. Acacia scrub: This seems to form a zone between dense scrub and the vari¬ ous coastal dune communities, usually or always on a sandy substrate. It comprises a rather open stand of leguminous, acacia-like shrubs about 3 meters high, with a dense understory of tall (1-1.5 m), broad-bladed grasses. Most of the acacia scrub I have seen has been near settlements and other habitations and it may be some sort of disturbance habitat. Butterflies are scarce, mostly casuals, but Cymaenes tripunctus seems to be a close associate, somewhat local. III. ecotonal areas: These, of course, exist between any two differing vegetation types. Two of them, however, deserve special attention. 14. Pine forest — dense scrub ecotone: This ecotone is generally rather broad, perhaps as much as 50 to 100 meters, and is marked by two distinguishing features. The first is a progressively increased spacing between the pines, and their gradual reduction in height, until at length they disappear altogether. The second is the erratic fluctuation of the scrub canopy height. In the Nicolls Town area the adjacent dense scrub is quite consistently about 2.5 m high. In the pine forest the scrub under¬ story is equally consistently somewhat lower, about 2 m high. In the ecotonal area it fluctuates from place to place, from much lower to much higher than either. Butter¬ flies are scarce, but Pyrrhocalles antigua is seen here regularly. 15. Dense scrub — scrub clearing ecotone: Scrub clearings are, of course, man¬ made. With neglect they gradually grow up in tall grass and scattered shrubs, some¬ what blurring the otherwise sharp edge. Into this boundary area several butterfly species of dense scrub may extend, among them Urbanus dorantes, Chlorostrymon maesites, Electro strymon angelia, Dryas iulia, Heliconius charitonius. Most of the Epargyreus zestos I have seen were in this ecotone. See also 5(a) above. In the Nicolls Town area the major vegetation types lie in a series of bands that roughly parallel the eastern coast. First inland is the coastal parkland and coconut groves; next westward is a narrow strip of acacia scrub, patchy and not always present; then, on limestone, a broad band of dense scrub (which in places goes to the low hammock forest); and finally, the pine forest, which occupies much of the island’s interior. In the neighborhood of Stafford Creek both limestone and pine forest reach the coast. On South Andros a similar longitudinal zonation is present, but the scrub is lower and pine forest begins much farther west, several miles to the 182 Annals of Carnegie Museum vol. 46 west in the vicinity of Driggs Hill. Even where the limestone reaches the coast and pinches off the dune belt formations no pines are present. THE SEASONS Lepidopterists in temperate regions are used to “broodedness” in butter¬ flies. Different species have different flight periods, but in any one locality the number of broods and their times of flight are quite definite for each. On Andros the situation is rather different. There is some broodedness: Papilio aristodemus, Chlorostrymon maesites, Electrostrymon angelia, probably Calisto sibylla, and perhaps others, all seem to have limited and perhaps definite flight periods. Most species, however, seem to be on the wing all year long. They have fairly regular annual fluctuations in their numbers, and most species, if not brooded, do have one or more periods of relative abundance alternating with periods of relative scarcity. For most species the time of greatest numbers is in June, coinciding with the period of maximum rainfall. Numbers then decline slowly through the summer to early October, then rapidly to the winter lows. Beginning in November, these lows probably extend through March or April. During the height of summer, total butterfly density is some four or five times as high as in winter. Curiously, the Hesperiidae seem to contrast in this : in general they are much commoner in fall and winter than in summer. This is particularly true, for example, of Urbanus proteus, Hylephila phyleus, and Atalopedes mesogramma, and it may also be true of several others. A number of species change in appearance (phenotype) with the seasons. Individuals that fly in the summer look different — sometimes strikingly so — from individuals of the same species that have emerged and fly in the winter. On Andros, Eurema d. daira is the most striking of these, and the differences between its summer form (“jucunda”) and winter form (“daira”) are so great that they were long thought to be different species. Other Andros butterflies show at least a slight seasonal difference: most species of Eurema, Nathalis iole, both species of Hemiargus (Cyclargus), Pyrrhocalles antigua. In general, and so far as known, the change from summer to winter forms occurs about in early November, and the change from winter to summer forms sometime in March or April. Among the more mystifying seasonal phenomena are the changes in flower preferences. Data are still too few even to describe the situation accurately, but in the course of my several visits I have noticed that at some times the flowers of certain species of plants are extremely attractive to butterflies, while at others the same plant species are almost completely ignored. This was true, for instance, of the low, blue-flowered roadside and grassland plant, Stachytarpheta jamaicensis, which was much visited in late September — early October and especially in November, but almost ignored in June. It was true also of the white-flowered shrub, Lantana 1977 Butterflies of Andros, Bahamas 183 involucrata, usually ignored in June and November but frequently visited in late September — early October. Both of these plants and others, like Bidens, are in flower at all these times. Systematic List In the following list I have adopted the sequence and, in nearly all cases, the nomenclature of Riley ( 1975) . This new book is indispensable to anyone collecting on Andros, or anywhere else in the West Indies, but it does contain some errors and, unfortunately, they concern the Bahamas more than other regions. In most cases they are minor and obvious and need no comment. When significant, I remark on them in the proper place. For each species I first give the name and, on the same line, the page reference in Riley (1975). In a few instances I list species expected to occur on Andros but not yet known from there. Their names are enclosed in brackets. (Species doubtfully or erroneously reported from Andros are given in a “Hypothetical List” at the end of the paper. ) The first paragraph under the species name is a list of its known Andros localities, usually listed from north to south. The localities are shown on the accompanying map (Fig. 1 ) . Distances are given in miles, as originally recorded. If the species is extremely rare, or known from only one col¬ lection, the date or dates of capture may be added. In almost every instance I give only localities from which specimens have actually been collected and identified in the hand. I do not like sight records (they are too subject to error and cannot be verified) and I include only a few of them, always so specified. The second paragraph (omitted when capture dates are included with the localities) lists the months in which the species is known to occur on Andros. When a span is given, e.g., “August-November,” it means that there are records of capture in each of the included months. A comment on the probable seasonal occurrence often follows the list of months. The third paragraph includes brief remarks on abundance, habits, and habitat. If the species is known to occur on Green or Scrub cays, the fact is noted here. Finally, a few additional comments may be added, usually descriptive. Throughout the paper I have kept abbreviations to a minimum. The only ones that may give trouble are in these descriptive notes: un, underside; up, upperside; fw, fore wing; hw, hind wing; and these in combination, as upfw, unhw. Danaidae Danaus plexippus plexippus Linnaeus, 1758. Riley: 33. North Andros: Nicolls Town (sight); Stafford Creek. September, November. Unmistakable. Rare; records are few, perhaps all migrants from the U.S.A. 184 Annals of Carnegie Museum vol. 46 Danaus gilippus berenice Cramer, 1779. Riley: 34. North Andros: Conch Sound, about 2 km south of Nicolls Town, 4.vii.l977, leg. D. Harvey (2 females). These specimens were taken feeding at flowers, and were the only ones seen in the approximately 45 minutes that Mr. Harvey spent at the site. A new record for Andros. Satyridae Calisto sibylla Bates, 1934. Riley: 52. South Andros: Yeho Pineyard, ca. 12 mi SW Driggs Hill, 7.vi.l974 (1 $ 1 $ ), and again (but sight only) 29. ii. 1976. Extremely rare and local, flying low in the shade of shrubs in scrubby virgin pineland. Aside from the above records, this species is known only from a single female taken on New Providence in 1897. Riley mistakenly makes smintheus Bates, 1935, of Cuba a subspecies of this. C. sibylla is a distinct species, endemic in the Bahamas. Another species of this genus, Calisto herophile apollinis Bates, 1934, may well occur on Andros. It resembles sibylla, but is smaller, somewhat paler, and has a red patch in the cell unfw, which sibylla lacks. It should be looked for in April and May in shaded scrub areas. Apaturidae Anaea echemus Doubleday, Westwood & Hewitson, 1850, subspecies Riley: 59. South Andros: vie. Congo Town, 29-3 l.vii. 1977, leg. Cullers & Serbin (1 $ ). The single specimen is in the Allyn Museum of Entomology, Sarasota, Florida. I have seen only a photograph of it and so cannot tell whether it is nominate echemus (Cuba) or the subspecies bahamae Witt 1972, known from a number of the central islands in the Bahamas. Nymphalidae Marpesia eleuchea bahamensis Munroe, 1971. Riley: 62. North Andros: Nicolls Town; Stafford Creek (sight); ca. 50 mi SSW Stafford Creek. — Mangrove Cay. — South Andros: 2 mi S Driggs Hill. June, July, October. Distinctive in hand, but in flight looks much like the common Agraulis vanillae or Euptoieta hegesia (and hence may sometimes be overlooked). Occurs in or near dense scrub, usually in the vicinity of Ficus trees (the probable larval foodplant). Lucinia sida albomaculata Rindge, 1955. Riley: 70. North Andros: Nicolls Town. — South Andros: 2 mi S Driggs Hill. June, November. Distinctive. Scarce on North Andros, commoner on South Andros. Local; rarely seen at flowers; usually perching on shrub branches in scrub and at scrub edges. Junonia coenia coenia Hiibner, 1822. Riley: 74. North Andros: Nicolls Town; Stafford Creek; 7 mi W Stafford Creek. February, August, October. Much like /. evarete (below), and flies with it, but is less common. Distinguished from evarete by the extra dark ring around the large eyespot upfw and unfw, by white behind this eyespot upfw, and by the much enlarged anterior eyespot uphw. 1977 Butterflies of Andros, Bahamas 185 Junonia evarete zonalis C. & R. Felder, 1867. Riley: 74. North Andros: Nicolls Town; Mastic Point; Stafford Creek; Twin Lakes Farm. February, June, August-November. Common at times. Flies in open fields of low, dry grass and bare ground, less often in tall grass; often difficult to approach. Anartia jatrophae guantanamo Munroe, 1942. Riley: 75. North Andros: Morgans Bluff; Nicolls Town; Stafford Creek; Twin Lakes Farm; Fresh Creek. — South Andros: 2 mi S Driggs Hill. June, July, September-November. Common, rather local, flying low in open, often moist, areas of lawns and low grass, nearly always in the near vicinity of stands of Lippia nodiflora, the presumed local larval foodplant. Phyciodes frisia frisia Poey, 1832. Riley: 79. North Andros: Nicolls Town; Mastic Point; Fresh Creek. June, July, August. Uncommon to scarce, local in shrubby old fields. Euptoieta hegesia hegesia Cramer, 1779. Riley: 83. North Andros: Nicolls Town; Mastic Point; 6 mi E Red Bays; San Andros airport; Stafford Creek; 7 mi W Stafford Creek; ca. 50 mi SSW Stafford Creek; Fresh Creek; Andros Town. — Mangrove Cay: Lisbon Creek. — South Andros: Sand Point, 2 mi SW Driggs Hill. February, April, June-November. Probably flies all year. The commonest butterfly on North Andros and, although it favors open areas, it can be expected in almost any kind of habitat. Euptoieta claudia Cramer, 1779. Riley: 84. North Andros: Nicolls Town; 6 mi E Red Bays; Stafford Creek. February, June-August, October. Less common than E. hegesia, but flying with it in open grassy areas. Resembles hegesia, but is more golden (less reddish), often somewhat larger, and has blackish lines in the basal parts of uphw, absent in hegesia. This may be a recent arrival in the Bahamas. Earlier writers (including Riley) did not list it, but in the last few years it has been found on Grand Bahama, Great Abaco, New Providence, and Andros. Heliconiidae Heliconius charitonius ramsdeni Comstock & Brown, 1950. Riley: 85. North Andros: Red Bays; Nicolls Town; Stafford Creek. June-November. Unmistakable. Not rare, but somewhat local, flying singly in shaded scrub and at scrub edges. Dry as iulia carteri Riley, 1926. Riley: 86. North Andros: Nicolls Town; Mastic Point; Stafford Creek; Blanket Sound; Fresh Creek. — Mangrove Cay. — South Andros: 2 mi S Driggs Hill. April, June-November. Probably flies all year. The long wings will distinguish it from other orange butterflies. Common in shaded scrub and in nearby open areas; generally flies slowly and is easily captured. Agraulis vanillae insularis Maynard, 1889. Riley: 88. North Andros: Morgans Bluff; Nicolls Town; San Andros airport; Stafford Creek; 186 Annals of Carnegie Museum vol. 46 Blanket Sound; Twin Lakes Farm; Fresh Creek; High Cay, ca. 7 mi SSE Fresh Creek. — Mangrove Cay. — South Andros: Sand Point, 2 mi SW Driggs Hill; 2 mi S Driggs Hill. April, June-December. Probably flies all year. The underside is unmistakable, but on the wing this species is easily confused with Euptoieta hegesia and even Marpesia eleuchea. One of the commonest butterflies on Andros, flying in generally open areas almost everywhere. C. J. Maynard took the type specimen of this subspecies somewhere between Fresh Creek and Middle Bight on Andros (Turner, 1957:143) on 8 December 1887. Riodinidae Apodemia carteri carteri Holland, 1902. Riley: 93. North Andros: Stafford Creek, 5-1 1.x. 1976 (6 specimens). — South Andros: Yeho Pineyard, ca. 12 mi SW Driggs Hill on South Bight, 7.vi.l974 (5 specimens). Extremely rare, inconspicuous, and local, in virgin scrubby pinelands. It flies some¬ what like a hairstreak, in irregular rapid circles, landing often, wings outspread, on undersides of leaves. Feeds at flowers, especially of White Lantana, with wings simi¬ larly outspread. Outside of Andros, known from New Providence (January, April, May), and Little San Salvador (March). Only about 20 specimens, all told, are known in collections. Another subspecies, ramsdeni Skinner, 1912, is known from Cuba; it is at least as rare. Lycaenidae Eumaeus atala fiorida Rober, 1926. Riley: 98. North Andros: Nicolls Town; Stafford Creek (sight only); 7 mi W Stafford Creek; 8 mi S Stafford Creek. February, June, October. Unmistakable. Usually scarce and local in cut-over pine forest and in old fields nearby; on occasion may be locally abundant. Sometimes it strays rather widely. Its flight is usually slow and ponderous. Chlorostrymon maesites maesites Herrich-Schaffer, 1864. Riley: 100. North Andros: Nicolls Town. — South Andros: Sand Point, 2 mi SW Driggs Hill. June, July. Probably single brooded, in midsummer. Unmistakable. Uncommon, local, and inconspicuous, usually at small white flowers on shrubs and low trees, especially Flowering Almond and Sea Grape. Strymon martialis Herrich-Schaffer, 1864. Riley: 102. North Andros: Nicolls Town; Stafford Creek; 7 mi W Stafford Creek. — Mangrove Cay. — South Andros: Sand Point, 2 mi SW Driggs Hill. February, June, August-November. Probably flies all year. Infrequent, singly, in scrubby areas. The species has also been taken on Green Cay (March) and the Scrub Cays (July). Strymon acis armouri Clench, 1943. Riley: 103. South Andros: 2 mi S Driggs Hill, 2-6.vi.1974. Not uncommon, but quite local, often flying in the company of Strymon columella and visiting the flowers of Spanish Needles (Bide ns). Strymon columella cybira Hewitson, 1874. Riley: 104. North Andros: Morgans Bluff; Nicolls Town; Stafford Creek. — Mangrove Cay. — South Andros: 2 mi S Driggs Hill; Yeho Pineyard, 12 mi SW Driggs Hill. April, June-November. Probably flies all year. 1977 Butterflies of Andros, Bahamas 187 Common on South Andros, much less so on North Andros. Usually visits low flowers, especially Spanish Needles (Bidens), but also found on the flowers of taller shrubs, like Bay Cedar (Suriana maritima). Numbers seem to fluctuate seasonally: it is commonest in summer, and can be extremely scarce in winter. It has been taken on Green Cay (March) and the Scrub Cays (July). Electrostrymon angelia dowi Clench, 1941. Riley: 107. North Andros: Nicolls Town; 7 mi W Stafford Creek; Blanket Sound; Somerset Bay. — South Andros: 2 mi S Driggs Hill. June. Probably single brooded in midsummer. Common, but probably only in summer, mostly on the small white flowers of shrubs and low trees, especially Flowering Almond (Terminalia catappa). This species has also been taken on Green Cay, common in early March (when it was not flying on Andros at all!). Leptotes cassius theonus Lucas, 1857. Riley: 108. North Andros: Nicolls Town; Stafford Creek; Fresh Creek; High Cay, 7 mi SSE Fresh Creek. — Mangrove Cay: Lisbon Creek. — South Andros: 2 mi S Driggs Hill; Sand Point, 2 mi SW Driggs Hill; Yeho Pineyard, 12 mi SW Driggs Hill. April, June-November. Probably flies all year. Extremely common in summer, less so in winter, and found in nearly all habitats. It has also been taken on Green Cay (March). Hemiargus (Hemiargus) ceraunus ceraunus Fabricius, 1793. Riley: 109. (as H. hanno, in part). South Andros: 2 mi S Driggs Hill, 2-6.vi.1974. Not uncommon, but local, in dry open areas with low vegetation. Apparently ab¬ sent from North Andros. Riley (1975: 109) confused two species, hanno Stoll, 1790, and ceraunus Fab¬ ricius, 1793, under the name hanno. True hanno is found in the southern West Indies and, so far as known, does not occur in the Bahamas at all. Riley’s figures represent ceraunus. Hemiargus (Cyclargus) ammon Lucas, 1857. Riley: 109. North Andros: Nicolls Town; Stafford Creek; Blanket Sound. — Mangrove Cay. — South Andros: 2 mi S Driggs Hill. June-August, October, November. Two subspecies occur on Andros: nominate ammon on South Andros and Man¬ grove Cay (common in season, in sunlit open areas of low grass or low scrub), and a still unnamed subspecies on North Andros (less common, rather local in or near pine forest). Hemiargus (Cyclargus) thomasi thomasi Clench, 1941. Riley: 110. North Andros: Nicolls Town. — South Andros: 2 mi S Driggs Hill. June, July, October, November. Less common than ammon and more local. Found in open fields with low shrubs. This species and ammon are extremely close. H. thomasi has an extra postbasal black dot below the cell unhw, absent in ammon; ammon $ often has a pink tornal lunule uphw, always absent in thomasi; and in both sexes the tornal orange lunule unhw is larger and redder in ammon. [Brephidium exilis isophthalma Herrich-Schaffer, 1862.] Riley: 114. This, the smallest Bahamian butterfly, has not yet been found on Andros, although it should be looked for, particularly on South Andros. It occurs in salinas, often accompanying Panoquina panoquinoides. It has been taken on Green Cay (March, July). 188 Annals of Carnegie Museum vol. 46 PlERIDAE Ascia monustc eubotea Latreille, 1819. Riley: 116 (as A. m. evonima). North Andros: Nicolls Town; Blanket Sound; Fresh Creek. — Mangrove Cay: Lisbon Creek. — South Andros: 2 mi S Driggs Hill. April, June-August. Fairly common in summer, rare or absent in winter. Found mostly in open, dis¬ turbed areas. Also taken on Green Cay (March). Appias drusilla poeyi Butler, 1872. Riley: 117. North Andros: Nicolls Town, 4.vii.l977, leg. D. Harvey (2 males, 1 female). One of the males was taken at a mud patch in the road. The specimens were all found in a generally shaded area. This is a little-known species in the Bahamas. Specimens from Grand Bahama, lent to me by D. Hall, are referable to the Florida subspecies, neumoegenii Skinner, 1894. These Andros specimens, notably the female, agree with Cuban individuals and represent subspecies poeyi. Eurema larae Herrich-Schiiffer, 1862. Riley: 121. North Andros: Red Bays; Nicolls Town; Stafford Creek; 7 mi W Stafford Creek; 8 mi S Stafford Creek. February, June-November. Apparently flies all year. Not common, in open, grassy, disturbed areas, flying rather low. about 0.3 m of the ground. Quite widespread on North Andros, but not yet found anywhere else in the Bahamas. The species is easily recognized on Andros by its small size, pure yellow ground color with narrow black borders, and almost no orange. Eurema daira daira Godart, 1819. Riley: 122. North Andros: Nicolls Town; Mastic Point; San Andros airport; Fresh Creek. June, August-November. Common at times, flying low over open areas of low grass. In late September and early October 1976 it was common on the lawns of the Andros Beach Hotel at Nicolls Town. Curiously, E. daira has been found south of Mastic Point only at Fresh Creek. Eurema lisa sulphur ina Poey, 1851. Riley: 123. North Andros: Nicolls Town; Stafford Creek; Twin Lakes Farm. — South Andros: 2 mi S Driggs Hill. June, July, September, October. Not common, in open, grassy, disturbed areas, flying rather low. Eurema messalina blakei Maynard, 1891. Riley: 125. North Andros: Nicolls Town; Stafford Creek; 7 mi W Stafford Creek; ca. 50 mi SSW Stafford Creek. — South Andros: 2 mi S Driggs Hill; Yeho Pineyard, 12 mi SW Driggs Hill. February, June, September, October. Infrequent and rather local, perhaps a little commoner on South Andros. It flies in the scrub, within about 0.3 m of the ground. Males have a bar of pink color unfw along the inner margin, and another uphw near the costa. The pink color disappears a few months after capture. Eurema dina helios Bates, 1934. Riley: 126. North Andros: Red Bays; Nicolls Town; Stafford Creek; Fresh Creek. — Man¬ grove Cay. — South Andros: 2 mi S Driggs Hill. June-August, October. 1977 Butterflies of Andros, Bahamas 189 Sometimes common. Flies in the scrub, usually about 1.5-2 m above the ground. It flies fast and lands seldom: a difficult species to catch. [Eurema nicippe Cramer, 1782.] Riley: 130. North Andros: Nicolls Town (sight only, June 1973); San Andros airport (sight only, June 1973). So far no specimens of this species have been taken, although I have seen it twice. It occurs widely in the Bahamas, and I have records from Grand Bahama, New Providence, Eleuthera, Cat, Long, Acklins, and Great Inagua. Nathalis iole Boisduval, 1836. Riley: 130. North Andros: Nicolls Town; Stafford Creek. June, August-October. Common at times. Flies low over lawns and open areas of low herbaceous vege¬ tation. Kricogonia lyside Godart, 1819. Riley: 131. North Andros: Nicolls Town, 1 ,4.vii. 1 977; Stafford Creek, 2.vii.l977 (sight, D. Harvey). — South Andros: Sand Point, 2 mi SW Driggs Hill, 7.vi.l974. Only a few records. This species is far commoner in the southern Bahamas. It has been taken on Green Cay (March) and the Scrub Cays (July). Phoebis agarithe antiilia Brown, 1929. Riley: 134. North Andros: Nicolls Town; Stafford Creek; Fresh Creek. June-October. Not uncommon, but a powerful flier, seldom pausing, and usually flying high: it is far more often seen than taken. It also occurs on Green Cay (March). Phoebis sennae sennae Linnaeus, 1758. Riley: 134. North Andros: Nicolls Town; Twin Lakes Farm; Fresh Cresk. — Mangrove Cay: Lisbon Creek. April, June, August, September. Not uncommon but, like agarithe, far more often seen than captured. Found in open, often disturbed, areas. Aphrissa neleis Boisduval, 1836. Riley: 135. North Andros: Nicolls Town; Stafford Creek; Fresh Creek. June-September, November. Not uncommon, chiefly in areas of dense scrub. In flight it cannot be distinguished reliably from Phoebis sennae, and it is equally difficult to capture. Elsewhere in the Bahamas neleis is known only from New Providence, where it is rare. Papilionidae Battus polydamas lucayus Rothschild & Jordan, 1906. Riley: 140. North Andros: Nicolls Town. June-September, November. Uncommon, but more frequent during summer than at other times. It is found mostly in open, coastal, old fields, but is sometimes seen in the scrub. It should be more widely distributed than the above single locality would indicate. This is the only tailless swallowtail on Andros. Battus devilliers Godart, 1824. Riley: 142. North Andros: Nicolls Town; Stafford Creek; 7 mi W Stafford Creek (sight only). — South Andros: Yeho Pineyard, 12 mi SW Driggs Hill (sight only). February, June-October. Perhaps flying all year. 190 Annals of Carnegie Museum vol. 46 Not rare in midsummer, uncommon to scarce at other times. This handsome species is known in the Bahamas only from Andros, where it flies mostly in pine forest (virgin or cut-over) or nearby scrub. The blue upperside is conspicuous in flight. Papilio aristodemus ponceanus Schaus, 1911. Riley: 146. North Andros: Nicolls Town, 28. ix. 1976. — South Andros: 2 mi S Driggs Hill, 2-8.vi.1974. The principal flight is probably in late May and early June. The September indi¬ vidual may have been abnormal, or there may be a partial second brood. This species closely resembles P. andraemon in flight, and occurs in the same habitats. Papilio andraemon bonhotei Sharpe, 1900. Riley: 147. North Andros: Nicolls Town; Stafford Creek; Fresh Creek. — Mangrove Cay: Lisbon Creek. — South Andros: 2 mi S Driggs Hill. April, June-October. Common in midsummer, becoming scarce to absent in fall and winter. Flies in dense scrub, scrub edges, and in coastal disturbed areas. It is often seen at flowers in towns. Hesperiidae Phocides pigmalion batabano Lucas, 1857. Riley: 156. North Andros: Blanket Sound; Twin Lakes Farm. June, October. Scarce and local, mostly near Red Mangrove along the coast. It is easily taken at the flowers of Nicker Bean (Caesalpinia) , which blooms in September and October. This is the Cuban subspecies and, in the Bahamas, it seems to be limited to Andros. The subspecies batabanoides Holland, 1902, occurs on New Providence, the Biminis, and Grand Bahama, and is transitional to the Florida subspecies okeechobee Worthington, 1881. Epargyreus zestos Go, yer, 1832. Riley: 157. North Andros: Nicolls Town. — Mangrove Cay. — South Andros: 2 mi S Driggs Hill. June-September. Not uncommon in midsummer, scarce to absent at other times. It is found in dense scrub and scrub edges, local and perhaps territorial, often on flowers. Urbanus proteus domingo Scudder, 1872. Riley: 163. North Andros: Nicolls Town; Stafford Creek; Fresh Creek. — Mangrove Cay. June-November. Common at times (e.g., November) at flowers, especially in open waste places and gardens, even in towns. It is probably territorial. Urbanus dorantes Santiago Lucas, 1857. Riley: 163. North Andros: Red Bays; Nicolls Town; Stafford Creek. — Mangrove Cay. — South Andros: 2 mi S Driggs Hill. February, June-November. Probably flies all year. Common. Found sparingly in pine forest and dense scrub; most frequent at low flowers along roadsides, in gardens, etc. Probably territorial. Although common on Andros, elsewhere in the Bahamas it is found only on Grand Bahama, where it is scarce. 1977 Butterflies of Andros, Bahamas 191 Astraptes xagua Lucas, 1857. Riley: 165. North Andros: Nicolls Town. August, November. Rare, in dense scrub or acacia scrub. Known from nowhere else in the Bahamas, and only a few specimens are known. Burca braco castigata Rindge, 1955. Riley: 169. South Andros: 2 mi S Driggs Hill, 2-8.vi.1974. Not rare, but apparently absent from North Andros. Occurs in dense scrub, par¬ ticularly in areas where the scrub is lower and more open. Ephyriades brunnea brunnea Herrich-Schaffer, 1864. Riley: 175. North Andros: Nicolls Town; Mastic Point; Stafford Creek; Blanket Sound; Fresh Creek. — Mangrove Cay: Lisbon Creek. — South Andros: 2 mi S Driggs Hill; Yeho Pineyard, 12 mi SW Driggs Hill. February, April, June, August, October. Locally rather common, in dense scrub, scrub edges, and sometimes in pine forest. Possibly territorial. Pyrrhocalles antiqua eleutherae Bates, 1934. Riley: 179. North Andros: Nicolls Town; Stafford Creek. — South Andros: 2 mi S Driggs Hill. June, August-November. Uncommon, but apparently more numerous in late summer and fall than in mid¬ summer. Found chiefly in and near pine forest or the transition between pine forest and dense scrub. It is a handsome species and in the Bahamas is known only from Andros and Eleuthera. Cymaenes tripunctus tripunctus Herrich-Schaffer, 1865. Riley: 181. North Andros: Nicolls Town. — South Andros: 2 mi S Driggs Hill. June, September-November. Not rare but quite local, in shaded areas in acacia scrub. It much resembles Wallengrenia misera in the field, but it flies more slowly and lands more frequently. In the hand, the black discal area unfw will distinguish it. Wallengrenia misera Lucas, 1857. Riley: 184. North Andros: Red Bays; Nicolls Town; Stafford Creek; Blanket Sound; 8 mi W Stafford Creek (“Stalactite” Blue Hole); Twin Lakes Farm; ca. 50 mi SSW Stafford Creek. — South Andros: 2 mi S Driggs Hill. February, June-November. Probably flies all year. Rather common but numbers fluctuate. It is particularly numerous in the fall. It is found mostly in sunny, open areas of low vegetation, commonly on the flowers of Stachytarpheta jamaicensis. Hylephila phyleus phyleus Drury, 1773. Riley: 185. North Andros: Nicolls Town. September, November. Uncommon, but variable in numbers, chiefly in fall and winter. Open areas of bare ground or low grass are favored habitats. In September 1976 it was uncommon but quite regular on the lawns of the Andros Beach Hotel. Atalopedes mesogramma mesogramma Latreille, 1823. Riley: 185. North Andros: Nicolls Town; Stafford Creek; Blanket Sound; Andros Town. — South Andros: 2 mi S Driggs Hill. June, July, September-November. Uncommon, but numbers distinctly greater in fall than in summer. The species is 192 Annals of Carnegie Museum vol. 46 found in dense scrub, acacia scrub, and scrub borders, mostly at flowers, and easily taken. Choranthus richmondi Miller, 1966. Riley: 188. North Andros: Stafford Creek, 5 and 9.x. 1976 (1 $ 1 $ ). Extremely rare, in scrubby virgin pine forest. The male is somewhat like the male of C. radians (Riley, PI. 24, Fig. 4) but duller, yellower orange above, brighter below, and without pale veins unhw. The female is dark. A recent discovery on Andros, this species was previously known only from two females taken in the Exuma Cays. The two Andros specimens were the result of over 20 hours spent searching the area (the place where Apodemia carteri was also found). Euphyes Cornelius Cornelius Latreille, 1824. Riley: 190. North Andros: 6 mi E Red Bays; Nicolls Town; Mastic Point; Stafford Creek; Blanket Sound; 8 mi S Stafford Creek; ca. 50 mi SSW Stafford Creek. — South Andros: 2 mi S Driggs Hill. February, June-November. Probably flies all year. Common (more so than anywhere else in the Bahamas), widespread, mostly at flowers along roadsides and in grassy areas in scrub. Distinctive: a nearly uniform slaty black hesperiine with pale yellow “face” (palpi) and a few tiny whitish spots on the wings. Calpodes ethlius Stoll, 1782. Riley: 191. North Andros: Nicolls Town. — South Andros: 2 mi S Driggs Hill. June, November. Erratic in appearance. It is found at flowers in open scrubby areas. Panoquina panoquinoides panoquinoides Skinner, 1892. Riley: 195. North Andros: Nicolls Town. — South Andros: 2 mi S Driggs Hill. June, October. Uncommon, local, exclusively along beaches and in salinas and salt flats where the presumed larval foodplant, Sporobolus grass, grows. The skipper flies low, usually in or just over the grass, and feeds at adjacent flowers (mostly those of Sesuvium portulacastrum, which is common in such places). The species is wide¬ spread in the West Indies, in southern Florida, and along the coast of southern Texas and Mexico, but it is often overlooked, both because it is inconspicuous and because of its restricted habitat, where few other butterflies occur. Hypothetical List Eunica tatila Herrich-Schaffer, 1853. Riley: 71. In a children’s coloring book (Anon. 1974: [13]) this species is mentioned by name as being “found on Andros Island.” The drawing is crude, but it seems to rep¬ resent this species. E. tatila is not rare on Cuba and could conceivably occur on Andros, but I know of no records. It should be looked for in hammock forest (coppices). Chlorostrymon simaethis Drury, 1773. Riley: 100. When Neil Richmond returned from his Bahama trip in 1965 he mentioned a hairstreak he had seen, but could not capture, on 4 July on High Cay, a small off¬ shore islet about 14 km southeast of Fresh Creek. He described it as having a green underside, but larger than C. maesites : a description that fits only C. simaethis among reasonable candidates. I entered simaethis in my MS list of Bahama butter¬ flies as a “possible,” and sent the information to Riley, along with other records of 1977 Butterflies of Andros, Bahamas 193 Bahama butterflies, for him to use in preparing his book. Somehow the doubtful status of the record was dropped and simaethis appears in Riley’s book as “reported from Andros.” C. simaethis is widespread in the New World tropics and it occurs on Cuba, so it remains a possibility, but to the best of my knowledge it has never been taken either on Andros or anywhere else in the Bahamas. Eurema chamberlaini Butler, 1898. Riley: 121. Butler described this species from a single male taken by Neville Chamberlain somewhere in the Bahamas, the exact source of the specimen not given. This was an unfortunate omission because chamberlaini, in addition to being one of the few endemic full species of butterflies in the Bahamas, is also one of the few that show definite subspecific differences from one island to another. Munroe (1950: 178), in his review of chamberlaini, pointed out that Chamberlain had spent five years at a sisal plantation on Andros. Although Munroe knew of no authentic Andros mate¬ rial, he reasoned that this was the most probable source of Chamberlain’s specimen. Riley accepted this reasonable conclusion and listed Andros among the islands whence the species is known. Unfortunately, subsequent collecting, especially on North Andros (where Chamberlain’s sisal plantation was located), has failed to turn up the species. Furthermore, Eurema larae does occur on North Andros, a fact unknown when Munroe published his paper, and larae is closely related to chamberlaini, is similar in appearance, and occupies habitats of the same sort that chamberlaini does. E. larae seems to be an ecological replacement of chamberlaini, and it is doubtful that both species could coexist in one area. 194 Annals of Carnegie Museum vol. 46 References Cited Anonymous 1974. Bahama butterflies and moths colouring book. Nassau, Etienne Dupuch, Jr., 48 pp., ill. Bates, D. Marston 1934. New Lepidoptera from the Bahamas. Occ. Papers Boston Soc. Nat. Hist. 8: 133-138. Clench, Harry K. 1975. More on Urbanus dorantes (Hesperiidae) . J. Lepid. Soc. 29: 106-107. 1976a. Fugitive color in the males of certain Pieridae. J. Lepid. Soc. 30: 88-90. 1976b. Nathalis iole (Pieridae) in the southeastern United States and the Bahamas. J. Lepid. Soc. 30: 121-126. Holland, William J. 1902. Two new species of Bahaman Lepidoptera. Ann. Carnegie Mus. 1: 486- 489. Miller, Lee D. 1966. A review of the West Indian “Choranthus.” J. Res. Lepid. 4 (“1965”): 259-274, ill. Munroe, Eugene G. 1950. The dina group of the genus Eurema in the West Indies (Lepidoptera, Pieridae). J. New York Ent. Soc. 58: 172-191. 1950a. The systematics of Calisto (Lepidoptera, Satyrinae), with remarks on the evolutionary and zoogeographic significance of the genus. J. New York Ent. Soc. 58: 211-240. Riley, Norman D. 1975. A field guide to the butterflies of the West Indies. London, Collins, 224 pp., ill. Rindge, Frederick H. 1952. The butterflies of the Bahama Islands. American Mus. Novit., No. 1563, 18 pp. 1955. The butterflies of the Van Voast-American Museum of Natural History Expedition to the Bahama Islands, British West Indies. American Mus. Novit., No. 1715, 20 pp., ill. Sharpe, Emily Mary 1900. On a collection of butterflies from the Bahamas. Proc. Zool. Soc. London 1900: 197-203, ill. Torre y Callejas, Salvador Luis de la 1968. Revision de las especies Cubanas de la familia Satyridae (Lepidoptera, Rhopalocera), con la descripcion de una nueva especie. Ciencias (ser. 4, Cienc. Biol.), No. 3, 24 pp., ill. Turner, Ruth D. 1957. Charles Johnson Maynard and his work in Malacology. Occ. Papers on Mollusks, Mus. Comp. Zool., Harvard Univ., 2: 137-152, ill. West, B. K. 1966. Butterflies of New Providence Island, Bahamas. Ent. Record 78: 174-179, 206-210. Back issues of many Annals of Carnegie Museum articles are available, and a few early complete volumes and parts are listed at half price. Orders and inquiries should be addressed to: Publications Secretary, Carnegie Museum, 4400 Forbes Avenue, Pittsburgh, Pa. 15213. MU*. COMP. ZOOU U8RARV NOV 1 4 to?/ isIn 0097-4463 HARVARD MHiVRRHITY of CARNEGIE MUSEUM CARNEGIE MUSEUM OF NATURAL HISTORY 4400 FORBES AVENUE • PITTSBURGH, PENNSYLVANIA 15213 VOLUME 46 OCTOBER 28, 1977 ARTICLE 13 6-/1 1 ft- ANNALS ECHINOCHIMAERA MELTON I, NEW GENUS AND SPECIES (CHIMAERIFORMES), FROM THE MISSISSIPPIAN OF MONTANA Richard Lund1 Research Associate Abstract A new genus and species of chimaeriform, Echinochimaera meltoni, is described from the Late Mississippian Bear Gulch limestone of Montana. E. meltoni differs from modern chimaeroids in having a complete placoid squamation, an ornamented first dorsal fin spine with simpler synarcuum, a stenobasal second dorsal fin, a subpelvic tenaculum, and prominent dermal cranial armament. E. meltoni and Marracanthus rectus are placed in the new suborder Echinochimaeroidei, which is considered a sister group to the Chimaeroidei. The Squalorajoidei and the Myria- canthoidei, Mesozoic Bradyodonti, share few derived characters in common with the Chimaeriformes and are removed from the order. The ptyctodont arthrodires share no phyletically significant characters with the Chimaeriformes. Introduction The chondrichthyan fauna of the Bear Gulch Limestone (Namurian) of Fergus County, Montana, (Lund, 1974; 1977a; 1977b; Lund and Zangerl, 1974) includes several representatives of one species of fish related to the Chimaeroidei, an operculate, holostylic suborder within the order Chimaeriformes. The Chimaeriformes, hitherto known only by Liassic and younger fossils and three families of living fish, have recently been divided into two additional suborders, the Squalorajoidei and Myriacanthoidei, by Patterson (1965). The Chimaeriformes have been grouped with several orders of Paleozoic Chondrichthyes into the ‘Biology Department, Adelphi University, Garden City, N.Y. 11530. Submitted for publication March 16, 1977. 195 196 Annals of Carnegie Museum vol. 46 superorder Holocephali, subclass Bradyodonti (Arambourg and Bertin, 1958; Lund, 1977a). The members of the Chimaeroidei are united by a number of shared derived characters, among which are the two paired upper and one paired lower cutting and tritoral jaw plates, an elevated and compressed ethmoid region containing an ethmoid canal, extremely expanded orbits lying over the brain, a mobile first dorsal spine and fin pivoted upon a synarcuum composed of fused anterior vertebrae and neural arches and spines, multiple perichordal vertebral rings, tall iliac processes on separate pelvic girdles, and prepelvic tenaculae and a frontal clasper in males (Dean, 1906; Patterson, 1965). The Jurassic genus Squalor aja, the only member of its suborder, differs from the Chimaeroidei in several characters of phyletic impor¬ tance (see discussion) but most critically in lacking their unique ethmoid specializations. Additionally, independent skeletally supported pre¬ pelvic tenaculae seem to be lacking (Patterson, 1965). The Jurassic Myriacanthoidei resemble chimaeroids in the general configuration of the skull and the presence of a large frontal clasper and synarcuum-based first dorsal fin. The dentition of the myriacan- thoids contains three paired upper plates as well as one paired and one median symphysial lower plate (Patterson, 1965). The myriacanthoid dentition is plesiomorphous in its retention of three paired upper plates in contrast to the chimaeroid condition, probably plesiomorphous, in the retention of a mandibular symphysial element as well (Dean 1906, Lund, 1977a). An ethmoid canal is apparently absent. The myria¬ canthoid condition of dermal plates on the neurocranium and mandib¬ ular spines, however (Patterson, 1965), is apparently derived relative to the chimaeroid condition. The Squalorajoidei and Myriacanthoidei contain both plesiomorphous and autopomorphous characters relative to the Chimaeroidei (Lund, 1977a) and have been included within the Chimaeriformes on the assumption (questioned by Dean, 1906) that the Chimaeroidei origi¬ nated in the early Mesozoic (Patterson, 1965). As it will be demonstrated in this article that the chimaeroids possessed their uniquely derived characters by the Namurian, neither squalorajoids nor myriacanthoids can have more than sister group relationship to them and should not be included within the Chimaeriformes. Speculations on the relationships of the Chimaeriformes have con¬ tinued for over a century (see Dean, 1906; Moy-Thomas, 1939; de Beer and Moy-Thomas, 1935; Patterson, 1965; Orvig, 1962). Most interest focuses on the Bradyodonti, poorly known, principally marine, Paleo¬ zoic chondrichthyans presumably with holostylic jaw suspensions and few crushing toothplates of tubular dentine (Obruchev, 1967). Very few whole dentitions however, and even fewer holomorphs (Patterson, 1965; Lund, 1977a), are known. Bradyodont dentitions contain small numbers 1977 Echinochimaera meltoni, New Genus and Species 197 of plates ranging from those corresponding to the chimaeroid combina¬ tion through myriacathoid-like combinations to apparently more primitive conditions of three paired upper and three paired lower plates. Further evidence links fused and unfused anterior tooth families of some bradyodonts to tooth families of the “pleuroplax” and “helodus” types, as seen in the freshwater lower Pennsylvanian Helodus simplex (New¬ berry and Worthen, 1866; Patterson, 1968). It is possible to demonstrate a morphologic series of dentitions among Helodus and the Bradyodonti by which chimaeroid and myriacanthoid dentitions may be derived on the basis of differential reduction (Lund, 1977a). Several authors have also suggested the ptyctodont arthrodires as possible relatives of the chimaeroids, initially on the basis of the com¬ mon possession of tritors of tubular dentine (e.g. Eastman, 1907), and, subsequent to the discovery of well preserved holomorphs of Ctenu- rella, on the basis of similar habitus (Orvig, 1962; Westoll, 1962). Dean (1906) points out, however, that ptyctodonts possess only one upper and one lower pair of jaw plates, too few to be considered ancestral to any holocephalian; Radinsky (1961) illustrates the dangers of using common possession of tubular dentine as a phyletic character, and Patterson (1965) indicates basic histologic differences between ptyctodont tritors on the one hand and the morphologically similar holocephalian and bradyodont plates on the other hand. Other apparent morphologic and histologic dissimilarities between the two groups (discussed below) reinforce the argument that the ptyctodonts are simply a highly special¬ ized group of arthrodires (Denison, 1975), none of which have any synapomorphous relationship to the holocephalians (Zangerl and Case, 1973). The fish described here is the first known Paleozoic member of the Chimaeriformes. The age of this taxon, Namurian A, and its relation¬ ship to younger members of the order, introduces a rich store of new factual information into the debate on chimaeriform relationships. Systematics Order Chimaeriformes diagnosis: Holostylic chondrichthyans with two pairs of upper dental and one pair of lower dental plates, the ethmoid region high, compressed, and with a median ethmoid canal, orbital region expanded dorsally and posteriorly, a mobile spinous first dorsal fin articulated upon a synarcuum, an aspinous second dorsal fin, diphycercal caudal fin, and monobasal pectoral fin. The pelvic girdle has a high iliac process and there are well developed pelvic fins in members of both sexes. Males have prepelvic tenaculae articu¬ lated with the pelvic girdle as well as pelvic mixopterygia. Squamation may be complete and placoid or limited. Multiple vertebral calcifications per body segment are present. Suborder Chimaeroidei amended diagnosis: Chimaeriformes with an unornamented first dorsal fin spine, a long second dorsal fin supported by separate basidorsals and radials, squamation 198 Annals of Carnegie Museum vol. 46 reduced to few specialized scales or absent, lateral line canals enclosed by ring scales, prepelvic tenaculae of males articulated with the anterior margin of the pelvic girdle, and a frontal clasper present in males. Dermal cranial armor is absent. included families: Chimaeridae, Rhinochimaeridae, Callorhynchidae. Suborder Echinochimaeroidei, new diagnosis: Chimaeriformes with a tuberculated first dorsal fin spine, a stenobasal second dorsal fin supported on a single basal plate, squamation complete and placoid, lateral line canals enclosed by very small denticles, prepelvic tenaculae of males articu¬ lated with the posterior margin of the pelvic girdle, a frontal clasper absent in males. Dermal cranial armor of enlarged or fused denticles is present. included family: Echinochimaeridae, new family. Family Echinochimaeridae, new diagnosis: The family Echinochimaeridae, with the new genus Echinochimaera as its type, is distinguished from other chimaeroids in having several paired and a median enlarged compound denticle on the dorsal surface of the head. The first dorsal fin spine bears denticles along its anterior edge, and is laterally tuberculated. Spine and denticles are enlarged distally in males. First and second dorsal fins have small basal plates and long, radiating supporting elements. Genus Echinochimaera, new diagnosis: Echinochimaeroids with four large paired supraorbital denticles in males and one large median postorbital denticle in both sexes. The dorsal fin spine is straight or slightly curved forward distally in mature males, compressed, with a grooved posterior margin for the proximal 2/3 of its length, ornamented laterally by eight to nine small vertically aligned rows of tubercles, and anterodistally by a cluster of denticles that are considerably enlarged in males. Both dorsal fins are short based. A short anal fin is present immediately anterior to the caudal fin. The rostrum is short and rounded, the orbital region is expanded to almost completely cover the otic region of the braincase. The body is rounded and clearly differentiated from the narrow caudal region. The prepelvic tenaculum is long, thin, well calcified in three segments and capped by a single large hook. The pelvic mixopterygium of each side is long, thin and straight. Squamation of the body and fins is placoid, with each denticle having a stellate base. There is a row of enlarged trifid denticles on each side of the dorsal midline between the second dorsal fin and the origin of the epichordal lobe of the tail. derivation of name: Echinochimaera, a prickly monster. type species: Echinochimaera meltoni. Echinochimaera meltoni, new species. Figures 1-16 type specimen: MV1 5371 referred specimens: MV 5372, 5374, 5375, 5383, 5384. CM1 23656, 25588, 27336, 30626, 30629, 30630, 30631. horizon and locality: Mississippian, Namurian A, Bear Gulch Limestone of Fergus County, Montana. diagnosis: The only known species of the genus Echinochimaera meltoni is a small echinochimaerid, the largest specimen measuring 150 mm, total length. There are nine rays in the first dorsal fin, seven in the second dorsal fin and nine rays in the pelvic fin. 'Abbreviations: MV, University of Montana Vertebrate catalog; CM, Carnegie Museum of Natural History. 1977 Echinochimaera meltoni, New Genus and Species 199 Fig. 1. Echinochimaera meltoni, MV 5371, type specimen, male. Scale is in mm. 200 Annals of Carnegie Museum vol. 46 E E Fig. 2. Echinochimaera meltoni, CM 25588, female. Scale is in 1977 Echinochimaera meltoni, New Genus and Species 201 A maximum of 19 rays can be seen in the epichordal lobe of the caudal fin and 12 in the hypochordal lobe. One to three rays are present in the anal fin. derivation of name: Named in honor of William G. Melton, Jr., founder and leader of the Bear Gulch project. Description preservation and growth: The thirteen presently known speci¬ mens have all been flattened during preservation, and show little three dimensionality. Neither neural nor haemal spines are calcified. A com¬ plete squamation that covers the body and outlines the fin radials is present in all specimens, (Figs. 1, 2). Dark pigmentation is present only in the orbits, indicating the size and position of the choroid coat of the eye. A ring of denticles in the orbital region indicates the maximum size limit of the cornea. The relationship between total length and snout-vent length (Fig. 3) appears to -be constant over the size range of available and measurable specimens. One specimen that deviates significantly from the straight line in Fig. 3 probably lost several millimeters from the tip of the tail. The relative difficulty of accurately establishing total length has led to the use of snout-vent length for comparative purposes. Spine length graphed against snout-vent length (Fig. 4) also shows a straight-line relationship. There is no significant difference between males and females in either of these relationships. The ratio of the distance between the origins of the first and second dorsal fins to the snout-vent length, expressed as decimal fractions in Fig. 4, shows distinct differences between males and females. The smallest specimen, 13.2 mm in snout-vent length, lacks a calcified dorsal fin spine, but the leading edge of the first dorsal fin is heavily scaled. A specimen of 19.7 mm snout-vent length has a spine of 7.7 mm. The spine of this specimen is poorly calcified basally, and supports the interpretation that young are born without dorsal fin spines, in contrast to the condition of modern chimaeroids (Bigelow and Schroe- der, 1953). Development of the spine is initiated in Echinochimaera between 13.2 and 19 mm snout-vent length (Fig. 4). The shapes of the spines are sexually dimorphic from their initiation. Spines of females are slender, slightly shorter than the fin, taper gradually to the distal end, and are set anteriorly with very fine denticles (Fig. 5). Spines of immature males are also shorter than the fin, but taper little, terminate in an obtuse angle, and are set, in all but the smallest specimen, with an anterodistal clump of enlarged denticles (Fig. 6). The spines of the lar¬ gest, sexually mature, males extend past the distal end of the first dorsal radial, curve forward slightly, and bear large and well spaced denticles distally (Figs. 7, 8). Changes in spines occur between 30 mm and 40 mm snout-vent length in males. These changes are accompanied by several other mor- 202 Annals of Carnegie Museum vol. 46 phologic changes, all indicating the onset of sexual maturity. Specimens below 27.5 mm snout-vent length show no calcifications of the fin girdles or axial skeleton. The 27.5 mm specimen and all longer ones show calcification of pectoral and pelvic girdles. Specimens of 32.5 mm and longer show calcified vertebral elements and synarcuums, as does a male of uncertain body length, with a spine length of 15.3 mm (Fig. 7). Males of 40.2 mm snout-vent length and longer are the only specimens that have calcified mixopterygia and prepelvic claspers. Only the two largest males and largest female show advanced calcifica¬ tion of the basal plates and radials of the dorsal fins as well as of the preorbital region of the neurocranium. 150- 0 130- IIO - TL 90- o 70 - □ □ o o o o -i - 1 - 1 - 1 — — i - 1 - 1 - n - 1 - 1 - r~ 70 30 4°SVL 50 60 70 ,a»4© 1977 Echinochimaera meltoni, New Genus and Species .0 O <• O o .0 O K rs o n mo _ *93 s 0 O S2 O s o o * •J > m o o .0 fi o j.o 203 Fig. 4. Relationship between snout-vent length and spine length, in mm. The numbers are the ratios of the distance between the first and second dorsal fins to snout-vent length. Circles are males, squares are females. 204 Annals of Carnegie Museum vol. 46 to Figs. 5-8. Echinochimaera meltoni, dorsal fin spines. Fig. 5, CM 25588, female, right side. Fig. 6, CM 23656, immature male, left side. Fig. 7, MV 5372, intermediate male, left side. Fig. 8, MV 5371, type specimen, left side. Scale is 1 mm for Figs. 5 and 6; 2 mm for Fig. 7, 3 mm for Fig. 8. 1977 Echinochimaera meltoni, New Genus and Species 205 The growth curves and developmental information for Echinochi¬ maera lead to some remarkably clear conclusions. These small chi- maeroids are born spineless and undergo regular but sexually dimorphic growth patterns. There are no proveable sexually mature females among the specimens presently available, but the onset of sexual ma¬ turity in males occurs between 31 mm and 40 mm snout-vent length. This process is heralded by differential growth of the distal part of the first dorsal spine, calcification of the claspers and mixopterygia, and ultimately, calcification of the neurocranium and dorsal fin cartilages. The absence of a frontal clasper in the largest male can therefore only be interpreted as a real absence. Development of differentiated mixop¬ terygia in elasmobranchs and frontal claspers as well as mixopterygia in recent chimaeroids also occurs at sexual maturity (Raikow and Swierczewski, 1975; Bigelow & Schroeder, 1948, 1953; Dean, 1906). squamation: Placoid scales with thin stellate bases and tapering, ridged, hollow shafts inclined posteriorly at their tips cover the head and body of Echinochimaera. They are sparsely distributed on the opercular flap. Ridges of the shaft of each denticle pass onto the base and extend beyond the base to give the stellate appearance (Fig. 9). Denticles outline each radial of the dorsal, anal, and pelvic fins, the anterior radials of the caudal fins, and the anterodorsal margin of the epichordal lobe of the caudal fin. The bases of the denticles of these fins are closely fitted together and curve around the radials (Fig. 10), an indication that only thin webbing, rather than bulky ceratotrichia, occupied the space between radials. Radials of these fins extended almost to the fin margins. Radials of the pectoral fin, which is inserted midway up the flank, are obscured by flank squamation. They are not as closely set together as other fin denticles, and do not appear to curve around radials, indicating the possibility of some development of ceratotrichia in the pectoral fin. Squamation of the ventral surface of smaller individuals, e.g., MV 5374, contains numerous fingerlike denticles, shafts lacking bases. These are not present in the holotype, MV 5371, and are probably early stages of denticle replacement. A paired row of enlarged denticles extends along the back of the largest specimen, from the rear of the second dorsal fin to the origin of the epichordal lobe of the caudal fin. The denticles are posteriorly curved and bear three progressively larger cusps upon a single base (Fig. 11). A partially decomposed specimen with a spine length of 15.3 mm shows several trifid denticles. The base of these denticles is perfor¬ ated by a separate nutrient foramen for each cusp. There is no indication of addition of successive cusps to the denticles. Nine pairs of denticles can be counted in CM 30630. Very fine, closely set denticles mark the lateral line canals on the 206 Annals of Carnegie Museum vol. 46 body and tail. The denticles do not form rings enclosing the canals as do those of Mesozoic and Recent chimaeroids. All specimens except the smallest, CM 30626, bear some indications of a large, complex, strongly curved median postorbital spine. This spine has a broad, thin base, thin walls, a large, open pulp cavity, and many secondary denticles. Prominent thin-walled, narrower spines are found around the base (Fig. 13). The shape and curvature of the spine itself is somewhat variable. The spine projects posteriorly to parallel the long axis. One spine, on CM 23656, apparently arches downward, toward the dorsal midline. Subsidiary basal spines are smooth¬ surfaced and vary in number and length, but several approach half the length of the major spine in CM 23656 and CM 27336. Figs. 9-11. Echinochimaera meltoni, dermal denticles. Fig. 9, from the posterior flank region. Scale is .1 mm. Fig. 10, CM 23656B, from a lower ray of the first dorsal fin. Scale is .5 mm. Fig. 11, CM 30630, enlarged denticle from the dorsal midline of the caudal region. Scale is 1 mm. 1977 Echinochimaera meltoni, New Genus and Species 207 Fig. 12. Echinochimaera meltoni, CM 30630 A. Supraorbital spines of the right side in mesial view. Scale is 5 mm. Fig. 13. Echinochimaera meltoni, MV 5374, head. A, anterior plate; N, nasal capsule; P, posterior plate; PS, postorbital spine. Scale is 5 mm. 208 Annals of Carnegie Museum vol. 46 CM 30630 best displays a series of paired spines located on the sup¬ raorbital crests. They have relatively thick walls, small bases, and decrease in size forward. There are four pairs of these spines. The two largest (most posterior) have a bifid tip, and the first two are simple (Fig. 12). The smallest specimen, CM 30626, of undetermined sex, has two pairs of short, broad-based denticles on the head that probably correspond to the most posterior pairs of large males. The two females may also have enlarged denticles. They lack the well-developed paired spines seen in large males. These immobile spines, with narrow pulp cavities, are thus modified placoid denticles, as is the median postorbital spine. neurocranium and jaws: The skull of Echinochimaera differs from that of the recent chimaeroid Chimaera collei (obtained from Carolina Biological Supply Co.) in very few respects (Fig. 13). The antorbital wall, formed of a tendinous sheet in Chimaera is present as calcified cartilage in the fossil. Supraorbital cartilage is present in the fossil form, lacking in Chimaera. A sagittal crest, present in Chimaera, is absent in Echinochimaera. The ventrolateral shelf along the otic region of the braincase apparently extends back to the occiput in Echinochimaera. The shelf extends only halfway back in Chimaera de Beer and Moy-Thomas, 1935. All other discernible features and propor¬ tions are extraordinarily similar. In view of the morphologic similarity, it is reasonable to propose that rostral and labial cartilages were present, and uncalcified, in Echinochimaera. The jaw plates are usually crushed together and are rarely visible from the buccal aspect. The posterior upper plate is long and low in lateral view (Fig. 13), with a sharp, serrate rim set with a single row of fine villiform toothlike projections. Dorsal to this cutting rim a narrow tritoral platform extends mesially on the buccal aspect. The tritoral platform is excavated anteriorly, appears to extend to the midline from the shallow anterior excavation back to about one third of the length of the plate (Fig. 14), and then narrows to a restricted shelf posteriorly. The anterior upper plate is thick and short, and lies internal to the anterior end of the posterior plate in the notch at the anterior end of the tritoral platform. Consequently, the anterior plate is not visible in lateral view and is usually obscured by crushing. The tritoral platform in the smaller specimens is seen to have a superficial layer of white (interosteonal tissue) surrounding low bumps of material of the same color and texture as the remainder of the plate. Larger specimens show no white material and only a slightly bumpy surface. The mandibular plate is also sharp rimmed, with long villiform projections, and bears a tritoral platform mesially which narrows from the symphisis to the rear of the plate (Figs. 14, 15). The mandibular plates are not fused at the symphisis, but the intermandibular region is very heavily calcified. The tritoral platform of the anterior end of the upper and lower jaws 1977 Echinochimaera meltoni, New Genus and Species 209 is not exposed from a buccal aspect in any specimen. The rarity of large specimens with intact jaws precludes histologic sectioning at this time. Examination of the jaws under alcohol reveals that there are a series of what appear to be dentinal columns that radiate from the posterior aboral corner of each jaw. These columns are apparently continuous with the villiform projections at the man¬ dibular tips (Fig. 14). axial skeleton: Neural and haemal arches and spines are uncal¬ cified in all specimens. Stoutly calcified dorsal and ventral hemi- centra are present in specimens above 31 mm snout-vent length. Hemi- centra are more numerous and closely spaced than are fin radials or enlarged denticles in the caudal region, the only region of the body in which clear segmentation can be estimated. This indicates, as does correspondence of hemicentral proportions with ring centra propor¬ tions of Isehyodus (Obruchev, 1967) and Squalor aja (Patterson, 1965), that the chimaeroid polyspondylous central calcification system is Fig. 14. Echinochimaera meltoni, MV 5371, type specimen, left upper and right lower jaws. A, anterior plate; P, posterior plate; Q, quadrate process, presumed calcified cartilage stippled. Scale is 1 mm. 210 Annals of Carnegie Museum vol. 46 present in these Mississippian fish. Fusion of dorsal and ventral hemi- centra, however, is found only in the base of the synarcuum, and is accompanied by fusion of adjacent hemicentra, neural arches, and spines. There is therefore no indication of the number of segments involved in the formation of the synarcuum. The anterodorsal margin of the synarcuum extends posterodorsally at an angle of about 30° to the vertebral axis. The posterior margin extends vertically around the neural canal, then curves posteriorly, then dorsally, to the posterodorsally concave articular facet. Lateral horns of the synarcuum extend beyond the spine articulation to the level of the rear margin of the spine. The articular facet of the basal plate of the dorsal fin spine is a convex surface that projects anteriorly below the ventral margin of the spine itself. The arrangement of spine-synarcuum articulation in Echinochimaera allows post-mortem adduction of the dorsal fin spine to an angle in excess of 45° anterior to a line perpendicular to the vertebral axis, as seen in the three largest males (Fig. 1). MV 5371, where post-mortem tetany has also dislocated the skull-synarcuum joint, shows the same extreme anterior angle of adduction as the articulated MV 5374. Females and small males do not show anterior adduction of the dorsal fin spine (Fig. 2), although structural reasons for this cannot presently be found. The lengths of spines and short distance between the first and second dorsal fins seems to prevent the depression of the dorsal fins and spines to the level of the back, as is possible in the more recent chimaeroids. Fig. 15. Echinochimaera meltoni, MV 5383, posterior part of the right lower jaw plate in lingual aspect. Scale is 1 mm. 1977 Echinochimaera meltoni, New Genus and Species 211 Spines in the more recent forms cannot be adducted forward of a line perpendicular to the vertebral axis. The synarcuum of Chimaera and other Mesozoic and Recent chimae- roids has transversely expanded vertical anterior and posterior margins that are absent in Echinochimaera. Dorsal parasagittal musculature originating on the lateral surfaces of the sagittal crest inserts on the anterior margin of the synarcuum of Chimaera, some of it extending along the dorsal surface to the crests of the lateral horns posteriorly. Segmental epaxial musculature that inserts lateral to the parasagittal musculature on the neurocranium also attaches firmly to the lateral edges of the posterior margin of the synarcuum. Erector musculature of the spine originates anteroventrally upon the synarcuum of Chi¬ maera, posterior and mesial to the anterolateral margin. Depressor musculature originates along the posterior margin, mesial to the epaxial musculature. The relationship between erector musculature, synarcuum and spine of Chimaera appears to be essentially the same as that of Echinochimaera. Depressor musculature, lacking the broad, vertically oriented posterior laminae of Chimaera as an area of origin would have been shorter, thinner, and weaker in effect than the corresponding modern musculature. The relationship between head, synarcuum and epaxial musculature in Echinochimaera is significantly less complex than in the Chimaeroidei. median fins: The first dorsal fin spine, which is denticulated along its anterior edge (see PRESERVATION AND GROWTH, above, and Figs. 5-8), bears eight to nine rows of well-spaced tuberculations later¬ ally, and a shallow posterior groove to its tip in all but sexually mature males. The dorsal fin is attached to the spine, the first, short radial emerging from the groove at the tip of the spine. The internal pulp cavity extends virtually to the tip of the spine in all specimens. The basal plate of the first dorsal fin is high and narrow, articulating beyond the body margin with the fin radials. There are nine radials, the first short, the second long, and the third and fourth usually arising by branching from variable positions from the second radial. The ante¬ rior margin of the second radial itself emerges from the posterior groove of the fin spine. The remainder of the radials articulate solely with the basal plate of the fin. The second dorsal fin contains a high, narrow, basal plate that arti¬ culates above the body line with the radials. The first radial is stout, the second is thin and arises by bifurcation from the first. Radials three through six are thin and approximately equal in length. The seventh is short and thin, and arises from the sixth radial. The second dorsal fin contrasts strongly with that of chimaeroids, which have a long, low, undulatory dorsal fin of segmental nature. The low epichordal lobe of the short, small, caudal fin originates above the origin of the hypochordal lobe, with a series of nine or ten 212 Annals of Carnegie Museum vol. 46 radials of gradually increasing height followed by an approximately equal number of progressively shorter radials. The anterodorsal margin of the epichordal lobe bears a rim of denticles that extends to the high point of the fin. The hypochordal lobe closely follows an anal fin of one long, stout radial and as many as three thinner radials. The first radial of the hypochordal lobe is the longest and stoutest of the fin. The fol¬ lowing five radials are successively shorter, and although at least six more can be counted in the holotype, the precise number of these short radials might be somewhat higher than the twelve visible ones. paired fins and girdles: The pectoral girdles extend from the dorsal margin of the body, immediately behind the head, to the ventral midline, but are not evidently fused in the ventral midline. The pectoral fins articulate below the middle of the flank. The pectoral girdle is of uniform width throughout. Virtually nothing can be seen of the pectoral fin itself, except that it lay along the midflank, extended about 3/4 of the distance from pectoral to pelvic girdle, and had a disposition of covering squamation indicative of an internal structure grossly similar to that of recent chimaeroids. The pelvic girdle (Fig. 16) has a high iliac process that extends above the level of the vertebral column (Figs. 1, 2). There is an abrupt posterior widening of the girdle below its middle, to the articulation with the apex of the triangular basal plate of the pelvic fin. The posterior margin then curves gently forward toward the slightly anteriorly concave anterior margin (in the female), leaving a thin mesial bar which contacts the contralateral bar in the ventral midline. The basal plate of the fin itself (Fig. 16) has the profile of a low obtuse triangle. The base of the triangle faces mesially, the anterior angle forms the pelvic articulation, and the nine visible fin radials articulate with the apex, posteriorside, and posterior angle. The first ray is stout, the second and third diverge from a single base, and several other bifurcating radials seem to be variably present in different specimens. The pelvic girdle of sexually mature males bears a posteroventral articular facet immediately above the mesial process and below the articulation for the fin base (Fig. 16). Articulating with this facet is a long axis of two segments, followed distally by a broad based curved spine of uncertain histologic nature but lacking a pulp cavity. This struc¬ ture evidently corresponds to the prepelvic tenaculum of chimaeroids. Chimaeroid prepelvic tenacula, however, articulate with the anterior margin of the pelvic girdle, and usually bear enlarged denticles (Stahl, 1967, Fig. 10). A single short, curved segment connects the long, simple, gradually tapering distal element of the mixopterygium to a facet near the posterior end of the basal edge of the fin supporting plate. The anterior edge of the pelvic fin of the male bears a short, stout, anter¬ iorly curved first radial which cannot be found in the female. 1977 Echinochimaera meltoni, New Genus and Species 213 214 Annals of Carnegie Museum vol. 46 The prepelvic tenaculum of Echinochimaera differs strongly from that of the chimaeroidei. There are no major morphological differences between the girdle-fin-clasper complexes of Echinochimaera and the Chimaeroidei. Echinochimaeroidei incertae sedis Marr acanthus rectus St. John and Worthen, 1875 Man acanthus rectus St. John and Worthen, 1875; p. 466-467, PL 22, Figs. 7-9. “Upper Beds of the St. Louis Limestone,” Alton, Illinois. The spines named Marracanthus are extremely similar to those of Echinochimaera in growth, size, shape, denticulation, and ornamenta¬ tion. They differ in lacking a distal posterior groove for dorsal fin attach¬ ment, in having fewer rows of lateral ornamentation, and fewer antero- distal denticles. Thus, while Marracanthus and Echinochimaera are demonstrably different spines, they are probably quite closely related and should be included in the same suborder. Disscussion relationship to the chimaeroidei: Echinochimaera shares many uniquely derived structural characters with the Chimaeroidei. The shape and proportions of the skull, jaws, and jaw plates, as well as the position of the branchial basket, differ only in small details from those of the modern Chimaeroidei. The basic structure of the synarcuum, first dorsal fin and spine, the pectoral girdle, pelvic girdle, fin, and clasper are also structures unique to Echinochimaera and the chimae- roids. Finally, the nature of the vertebral column, the anal-fin-caudal- fin relationship and diphycercal tail, and the presence, but not the structure of the prepelvic tenaculum also indicate a very close phyletic relationship between these Mississippian fish and the more recent chimaeroids. Echinochimaera clearly has more primitive character states in its complete squamation of placoid denticles, its ornamented dorsal fin spine and in having only small, simple denticles framing the lateral line canals. The lack of a frontal clasper in males is a primitive character that markedly distinguishes these fishes from the chimae¬ roids. Finally, the postpelvic articulation of the prepelvic clasper is believed to be a primitive character state, under the presently untestable hypothesis that this structure may originally have been derived from anterior pelvic fin radials that articulated with the pelvic girdle. This arrangement of fin radials and girdle may be seen in the petalodont Heteropetalus elegantulus (Lund, 1977b). A complete squamation is unknown among the Chimaeroidei, al¬ though Callorhynchus does possess some modified denticles (Dean, 1906). Lateral line canals are encased in very fine ring scales in Chimae¬ roidei from the Mesozoic and later (Patterson, 1965). 1977 Echinochimaera meltoni, New Genus and Species 215 The absence of anterior transverse laminae of the synarcuum may be a primitive character state of synarcual evolution. This sustains the view that the synarcuum of chimaeroids originally evolved in connection with strong articulation of a mobile first dorsal spine (Lund, 1977a). The synarcuum of Arthrodira was developed in relation to cranial mobility (Stensio, 1945) like the synarcuum of the Batoidei, and thus bears only a convergent, coincidental relationship. Echinochimaera has several uniquely derived character states that differentiate it from the Chimaeroidei. The second dorsal fin is short based, all radials articulating with a single high basal plate. All Chimae¬ roidei (Obruchev, 1967), Helodus (Patterson, 1965), Chondrenchelys (Moy-Thomas, 1935), Heteropetalus (Lund, 1977), pleuracanthous and anacanthous elasrhobranchs (see Lund 1974) bear a serial basidorsal- radial second dorsal fin. This condition is evidently plesiomorphous for the class Chondrichthyes. Sexually dimorphic differentiation of the first dorsal fin spine, as well as lateral supraorbital denticles, is not known in the Chimaeroidei. Indeed, the lack of this spine in the youngest specimen also contrasts strongly with the modern chimaeroid condition (Dean, 1906). Further, although adequate information on size dimorphism in the modern forms is lacking, the information available indicates that mature females are usually larger than mature males. (Dean, 1906; Bigelow and Schroe- der, 1953). The fin spines of modern Chimaeroidei are free from the fin distally, grooved venomous, and thus principally defensive in nature (Evans, 1923) The anteriorly pivoting, anteriorly denticulated and enlarged fin spines of mature male Echinochimaera, together with their lateral supraorbital denticles, probably functioned chiefly in intra- specific sexual display and courtship. Ghiselin (1974) (also see Gould, 1974) discusses various strategies involved in sexual selection, one of which seems to explain some features of the occurrence of the Bear Gulch chimaeroids. Elaboration of sexual dimorphism through in¬ creased size and ornamentation of males relative to females seems related to gregariousness in some species. It is accompanied in these forms by delayed functional maturity of the males, increased male- male interaction and a high mortality rate of males versus females (Ghiselin, 1974:146-147). It is worthwhile noting that in spite of a pre¬ cariously small sample size, both the size distribution of specimens and the nature of the sex dimorphism fit convincingly within this model. The more recent chimaeroids have adopted an alternative strategy (Dean, 1906) involving smaller, more mobile males and selection for more effective copulatory devices. The presence of a complex median postorbital spine in both sexes of Echinochimaera is a derived character in relation to the Chimaeroidei, all of which tend to elimination of dermal defensive structures other 216 Annals of Carnegie Museum vol. 46 than the dorsal fin spine. The presence of defensive denticles on the head, and elsewhere, is quite common among Chondrichthyes (Patter¬ son, 1965, 1968; Bendix-Almgreen, 1968). The jaw plates of Echinochimaera agree in number and position with those of the Chimaeroidei. They differ, however, in detail, from the presence of villiform projections to the nature of the tritoral surface. Echinochimaera and the Chimaeroidei share many uniquely derived characters that differentiate them from all other Chondrichthyes. Each group, however, possesses autopomorphous characters that are plesio- morphous in the other. Two conclusions derive from this observation. The first is that the two must share a common ancestor — that is, they are sister groups. The second observation, which follows inevitably from the first, is that therefore the Chimaeroidei did not originate in the Jurassic (Patterson, 1965) but prior to the Namurian A, as suggested by Dean (1906). A brief reevaluation of information on other known Holo- cephali is necessary at this time. Squaloraja is a depressed holocephalian with a long, thick, flat rostrum and a very long frontal “clasper” (Patterson, 1965). Squaloraja shares the number of tooth plates, form of the postrostral braincase, and the structure of the vertebral column and endoskeleton of the pectoral and pelvic fins with the Chimaeroidei. Squamation is placoid and extensive. Squaloraja is divergent from the Chimaeroidei in the nature of the rostrum and “clasper,” in lacking a first dorsal spine and fin, in having a long, low, batoid-like synarcuum, and in lacking iliac processes or endoskeletally supported prepelvic or subpelvic tenacula. The tooth- plates are not of chimaeroid histology, but are composed of “alternating bands of osteonal and interosteonal tissue” (Patterson, 1965:121). They lack the peculiar localized tritors of tubular dentine and inter- tritoral cover of outer tissue (Peyer, 1968:79) of chimaeroids, and are not composed of dentinal osteons surrounded by interosteonal tissue, as in typical bradyodont tooth plates (Radinsky, 1961). This condition could conceivably have been derived either from a bradyodont plate, from a chimaeroid condition, or from a condition plesiomorphous to both. It is a totally autopomorphous character state. The lack (loss) of the first dorsal fin and spine, and possibly the absence of iliac pro¬ cesses, might be associated with a flattened body form, as may be the type of synarcuum. The absence of prepelvic claspers is less readily explained in terms of secondary loss. Furthermore, Helodus, a dis¬ tantly derived chondrichthyan, shares most characters of the endoskele¬ ton of the paired fins with Squaloraja and the chimaeroids, so that while these characters indicate some affinity, the relationship they indicate might be a distant one indeed. Squaloraja appears to share with the myriacanthoids the elongate rostrum, the absence of an ethmoid canal, and retention of an apparently open precerebral fontanelle (Patterson, 1965:121). The open, if narrow 1977 Echinochimaera meltoni, New Genus and Species 217 precerebral fontanelle is a plesiomorphous condition in relation to the highly compressed, elevated chimaeroid ethmoid region and ethmoid canal. The elevation of the ethmoid region is most clearly related to the great anterodorsal expansion of preorbitalis musculature (Raikow and Swierczewski, 1975) and may be relatively independent of expansion of the orbits, as Patterson (1965) indicates. Elongation of the rostrum in Squaloraja is thus autapomorphous relative to chimaeroids, but the chimaeroid ethmoid region is apomorphous relative to the squalora- joid condition. Finally the number of tooth plates in Squaloraja is the same as that in menaspoids and in some myriacanthoids as well as in chimaeroids (Lund, 1977a). In summary, Squaloraja shares no discernible uniquely derived char¬ acters with chimaeroids, except the numerous vertebrae, but does possess many character states in which it is either plesiomorphous or autopomorphous in relation to the chimaeroid condition, and thus is certainly a separately derived group. The myriacanthoids (Patterson, 1965) are strongly divergent from the chimaeroid condition in having a strong, long, calcified cartilage rostrum of the Squaloraja type, in the presence of tuberculated plates on the skull, mandibular spines, and, most strikingly in tooth-plate number and histology. Myriacanthus and Metopacanthus, the best known mem¬ bers of the group, have three paired upper plates and a symphysial plate, plus a pair of plates in the lower jaw. The two anterior upper pairs of plates have localized tritoral areas, but the pterygoid and all three man¬ dibular plates have an occlusal surface of tubular dentine. A synarcuum of chimaeroid type is present, associated with a mobile first dorsal fin and spine. Little else is known about these fish. The postrostral neurocranium and jaw proportions are quite similar to those of chimaeroids, but as in Squaloraja , there is a long, stout ros¬ trum, contrasting stongly with the delicate rostral cartilages of even the longest-snouted chimaeroids (Obruchev, 1967). Tuberculated dermal cranial and mandibular plates are a derived condition in comparison with chimaeroids. Upper tooth plate numbers, however, approximate the highest known cochliodont condition, while lower tooth plate num¬ bers are intermediate between that of “ Platyxystrodus” and that of menaspoids, Squaloraja, and chimaeroids. The disposition of tritoral areas on anterior plates and tubular dentine on the lower and posterior upper plates is again intermediate between the two separately derived conditions. Thus, while it is not presently possible to determine the evolutionary pathways of tooth morphology between chimaeroid, myriacanthoid, and cochliodontoid, the myriacanthoids provide a clear morphologic link between the two groups. The myriacanthoids cannot consequently be considered close to the chimaeroids phyletically. While the myria¬ canthoids clearly approach the menaspoid-cochliodont condition, analy- 218 Annals of Carnegie Museum vol. 46 sis of the cochliodont bradyodonts in addition to what I have done else¬ where (Lund, 1977a) would be futile without additional study material. Helodus simplex (Moy-Thomas, 1936; Patterson, 1965) is a flat headed, many toothed, late Carboniferous fresh-water chondrichthyan which happens to share holostyly and the form of the pectoral and dorsal fins with chimaeroids. Aside from holostyly, the low, wide neurocranium with a broadly open precerebral fontanelle shows none of the modifica¬ tions necessary to be comparable to the chimaeroid condition. The nature of the skull and dentition reveals a morphotype comparable to what may have been the plesiomorphous condition of the Bradyodonti (Lund, 1977a). The ptyctodont arthrodires have been frequently suggested as ancestors of the chimaeroids (Orvig, 1962; Westoll, 1962; Stahl, 1967) although the body form seems to be the only character the two groups share. The single paired upper and lower tooth plates of Ptyctodus, which lack either a basal layer or the complex relationship of an outer layer to osteodentine and tubular dentine characteristic of chimaeroids are not comparable to those of the latter group (Patterson, 1968; Dean, 1906). The structure of paired and median fins is distinctly divergent, the synarcuum is of typical arthrodiran type (Stensio, 1945), and squamation is absent. Finally, while elaborate pelvic fins are present in the form of a funnel, intromittent organs and pelvic fins and girdles, as known in the chondrichthyans, are absent. While there seems little reason to doubt that the pelvic fins of ptyctodonts evolved to facilitate internal fertilization, this feat must have been accomplished by cloacal apposition. The ptyctodont arthrodires represent a highly derived arthrodiran condition, but one far too specialized to be even distantly related to any known holocephalian or bradyodont. Conclusions Echinochimaera meltoni, the first described Paleozoic chimaeriform, closely resembles the geologically younger members of its order in body form, jaw, and neurocranial structure, in the structure of the pectoral girdle, pelvic girdle, pelvic fin and clasper, and in anal-caudal relation¬ ships. While the basic structure of the first dorsal fin-spine-synarcuum complex is close to that of the Chimaeroidei, structural differences are evident that strongly suggest a separate adaptive pathway. The steno- basal second dorsal fin and elaboration of armament, the absence of a frontal clasper, and a unique, sub-pelvic, axial, paired tenaculum rein¬ force the separately derived position of this group in relation to the Chimaeroidei. Large males, with elaborate growth of dermal cranial spines and first dorsal spines, smaller and less common females, and a good morphologic series indicating post-hatching spine formation, all 1977 Echinochimaera meltoni, New Genus and Species 219 seem to fit a particular sexual selection pathway which is also divergent from that of the Chimaeroidei. The Chimaeriformes are therefore interpreted as a discrete group that underwent a Paleozoic radiation apart from the Squalorajoidei. The squalorajoids share more derived characters with the Myriacanthiformes than with the Chimaeriformes. The Myriacanthiformes are seen as derived from some holocephalian bradyodont lineage, bearing a sister group relationship to the Chimaeriformes. Comparison with the ptyctodont arthrodires reveals no phyletically significant characters in common. acknowledgements: I would like to thank ranchers Clifford Swift and Gilbert Schultz, who allowed us to work on their land; the many people in the field crews, whose sharp eyes discovered these fossils; the University of Montana and Carnegie Museum of Natural History, whose faith and finances helped through the summers, and the National Science Foundation, under grant number BMS 75-02720, for the current invaluable support. 220 Annals of Carnegie Museum vol. 46 References Cited Arambourg, C., and L. Bertin. 1958. Sous-classe des Bradyodontes (Bradyodonti). Traite de Zoologie, 13:2057-2067. P.P. Grasse, ed, Masson, Paris. Beer, G. R. de, and J. A. Moy-Thomas. 1935. On the skull of Holocephali. Phil. Trans. Roy, Soc. London (B); 224: 287-312. Bendix-Almgreen, S. E. 1968. The bradyodont elasmobranchs and their affinities; a discussion. Nobel Symp. 4:153-139. Almquist and Wiksell, Stockholm. Bigelow, H. B., and W. C. Schroeder. 1948. Fishes of the western North Atlantic, part one: Sharks. Mem. Sears Foundation for Marine Research 1:59-576. 1953. Fishes of the Western North Atlantic, part two: Chimaeroids. Mem. Sears Foundation for Marine Research 1:515-562. Dean, B. 1906. Chimaeroid fishes and their development. Publ. Carnegie Inst. Washington, 32:1-195. Denison, R. H. 1975. Evolution and classification of placoderm fishes, Breviora, M.C.Z. 432:1-24. Eastman, C. R. 1907. Devonian fishes of Iowa, Iowa Geol. Survey, XVIIL29-386. Evans, H. M. 1923. The defensive spines of fishes, living and fossil, and the glandular structure in connection therewith, with observations on the nature of fish venoms. Phil. Trans. Roy. Soc. London (B) 212:1-33. Ghiselin, M. T. 1974. The economy of nature and the evolution of sex. U. of California press, Berke¬ ley, California, 346 pp. Gould, S. J. 1974. The origin and function of “bizarre” structures: antler size and skull size in the “Irish Elk,” Megaloceros giganteus. Evolution, 28 (2): 191-220. Lund, R. 1974. Stethacanthus altonensis (Elasmobranchii) from the Bear Gulch Limestone of Montana, Ann. Carnegie Mus., 45(8): 161-178. 1977a. New information on the origin of the bradyodont Chondrichthyes. Fieldiana, Geol., 33(28):52 1-539. 1977b. A new petalodont (Chondrichthyes, Bradyodonti) from the Upper Mississip¬ pi of Montana. Ann. Carnegie Mus., 46(10):00-00. Lund, R. and R. Zangerl. 1974. Squatinactis montanus, a new elasmobranch from the Upper Mississippian of Montana, Ann. Carnegie Mus., 45(4):43-54. Moy-Thomas, J. A. 1935. The structure and affinities of Chondrenchelys problematica. Trans. Proc. Zool., Soc., London, (2):391-403. 1936. On the Structure and Affinities of the Carboniferous cochliodont Helodus simplex, Geol. Mag. 73:488-503. 1939. The early evolution and relationships of the elasmobranchs. Biol. Rev. 14:1-26. Moy-Thomas, J. A. and R. S. Miles. 1971. Paleozoic Fishes, Saunders, Phila. Newberry, J. S. and A. H. Worthen. 1866. Descriptions of new species of vertebrates . . . Geol. Survey Illinois 2:9-134. Obruchev, D. V. 1967. Fundamentals of Paleontology, 11, Agnatha, Pisces., Israel Progr. Sci. Transl., Jerusalem. 1977 Echinochimaera meltoni, New Genus and Species 221 Orvig, T. 1962. Y a t-il une relation directe entre les Arthrodires Ptyctodontides et les Holoce- phales? Problemes actuels de Paleontologie (Evolution des vertebres), Colloque Intern, du CNRS no. 104:49-61. Patterson, C. 1965. The phylogeny of the chimaeroido, Phil. Trans. Roy. Soc. London B, 249 (757): 101-219. Patterson, C. 1968. Menaspis and the bradyodonts, Nobel Symp. 4:171-205, Almquist and Wiksell, Stockholm. Peyer, B. 1968. Comparative Odontology. Translated and edited by R. Zangerl. U. of Chicago Press. Radinsky, L. 1961. Tooth histology as a taxonomic criterion for cartilaginous fishes, J. Morph. 109(l):73-92. Raikow, R. J. and E. V. Swierczewski. 1975. Functional anatomy and sexual dimorphism of the cephalic clasper in the Pacific ratfish (Chimaera collei). Jour. Morph. 145:435-440. Stahl, B. S. 1967. Morphology and relations of the Holocephali with special reference to the venous system. Bull. Mus. Comp. Zool. Harvard, 13:141-213. Stensio, E. A. 1945. On the heads of certain arthrodires. 2. On the cranium and cervical joint of the Dolichothoraci. K. Svenska Vetenskakad — Handl. (3), 22:1-70. St. John, O. H. and A. H. Worthen. 1875. Descriptions of fossil fishes, Geol. Survey Illinois, 6:245-488. Westoll, T. S. 1962. Ptyctodont fishes and the ancestry of the Holocephali. Nature, London, 194:949-952. Zangerl, R. and G. R. Case. 1973. Iniopterygia, a new order of chondrichthyan fishes from the Pennsylvanian of North America, Fieldiana Geol. Mem. 6:1-67. Back issues of many Annals of Carnegie Museum articles are available, and a few early complete volumes and parts are listed at half price. Orders and inquiries should be addressed to: Publications Secretary, Carnegie Museum, 4400 Forbes Avenue, Pittsburgh, Pa. 15213. LIBRARY S-Nft- NOV 1 4 1977 UNr/EH^I17 4463 AN NALS of CARN EG E MUSEUM CARNEGIE MUSEUM OF NATURAL HISTORY 4400 FORBES AVENUE • PITTSBURGH, PENNSYLVANIA 15213 VOLUME 46 OCTOBER 31, 1977 ARTICLE 14 NONGEOGRAPHIC VARIATION IN ELEPHANT SHREWS (GENUS ELEPHANTULUS THOMAS AND SCHWANN, 1906) OF SOUTHERN AFRICA I. L. Rautenbach1 Duane A. Schlitter Section of Mammals Abstract Five species of southern African Elephantulus Thomas and Schwann, 1906, were analyzed for individual and secondary sexual variation. Twenty-one external and cranial measurements were tested for samples of E. eclwardi (A. Smith, 1839), E. branchyrhynchus (A. Smith, 1836), E. rupestris (A. Smith, 1831), E. intufi (A. Smith, 1836) and E. myurus Thomas and Schwann, 1906. The most individual variation in the 21 measurements was exhibited by E. eclwardi, while E. myurus showed the least as reflected by the values of the coefficients of variation. Three species (E. brachyr- hynchus, E. rupestris, and E. myurus) had statistically significant secondary sexual variation. Introduction The Macroscelididae (elephant shrews) is a monophyletic group of insectivores confined to Africa with no close phyletic relationship to other mammalian taxa (Corbet and Hanks, 1968:48; Butler and Green¬ wood, 1976:2). A taxonomic revision was proposed by Corbet and Hanks (1968), in which they arranged the elephant shrews into four genera, with 14 species being recognized. They considered their arrangement as nearly definite. 'Transvaal Museum, P.O. Box 413, Pretoria 0001, Republic of South Africa. Submitted for publication April 4, 1977. 223 224 Annals of Carnegie Museum vol. 46 Corbet and Neal (1965) and Corbet (1971) studied geographic varia¬ tion in monotypic Petrodromus Peters, 1846. Nine subspecies were recognized in Petrodromus tetradactylus Peters, 1846. Three species of Rhynchocyon Peters, 1847, are recognized by Corbet and Hanks (1968) and Corbet (1971). Rhynchocyon chrysopygus Gunther, 1881, is con¬ sidered to be monotypic. R. petersi Bocage, 1880, consists of two recog¬ nizable subspecies. R. cirnei Peters, 1847, contains six subspecies. Because of inadequate sample size, Corbet and Hanks (1968) could reach no clear-cut conclusions with regard to geographic variation in either Macroscelides A. Smith, 1829, considered to be montypic, or Elephantulus Thomas and Schwann, 1906, with nine species ( Nasilio Thomas and Schwann, 1906, as a synonym), and no subspecies were listed by them at that time. Meester et al. (1964) as well as Lundholm (Meester et al., 1964:2) believed that many of the subspecies of Ele¬ phantulus recognized by earlier workers (Shortridge, 1941; Roberts, 1951; Ellerman et al., 1953) would prove eventually to be invalid, and both stressed the need for a thorough taxonomic study of this group. More recently, Butler and Greenwood (1976) reported on a collection of Pleistocene elephant shrews from Kenya and South Africa. Two new fossil subspecies, belonging to Recent species, were described based primarily on qualitative, and to a lesser extent, quantitative characteristics. Material and Methods All specimens examined for this study are deposited in the collections of the United States National Museum of Natural History, Smithsonian Institution, Washington, D.C. Further details concerning these speci¬ mens, all prepared as conventional museum study specimens, are as follows, with total number of each age examined, locality of origin, and month of capture of specimens given for each species. E. edwardi: 9 males, 12 females, and 7 non-adults, from Pakhuispas, Clan William District, Cape Province, R.S.A., December 1963; 3 mi NE Fraserburg, Cape Province, R.S.A., August 1969; and 15 mi SSW Sutherland, Cape Province, R.S.A., August 1969. E. brachyrhynchus: 9 males, 6 females and 16 non-adults, from Tsau, Botswana, March 1965; 5 mi W Gaberones, Botswana, May 1966; and 10 mi W Ramatlabama, Botswana, June 1966. E. rupestris: 26 males, and 20 females, from 13 mi E Upington, Cape Province, R.S.A., June and November 1968. E. intufi: 10 males, 6 females and 17 non-adults, from 1 mi NE Tshipise, Transvaal Province, R.S.A., January 1967. E. myurus: 27 males, 13 females and 5 non-adults, from 12 mi W Petrusville, Orange Free State, R.S.A., April and May 1968. Twenty-one variables were measured by means of dial calipers, and used in the analyses of individual, secondary sexual, and age variation. All measurements are in millimeters. The following measurements were recorded (those abbreviations used in Figure 1 are given in parentheses 1977 Nongeographic Variation In Elephant Shrews 225 following each measurement): Total length (TOT); length of tail (to tip of last vertebrae) (T); length of hindfoot ( c.u .) (HFT); length of ear (from notch) (E); greatest length of skull (GLS); condylobasal length of skull (CBL); greatest zygomatic breadth (posterior) (GZB); least interorbital breadth (IOB); greatest breadth of braincase (taken above base of squa¬ mosal of the zygoma) (BBC); greatest length of nasals (along their joint suture) (GLN); height of rostrum (taken along the vertical suture between the premaxilla and maxilla) (HR); greatest length of anterior palatine foramina (LAP); greatest length of posterior palatine foramina (LPP); width of bulla (from the median border perpindicular to the anterior edge of auditory meatus) (WB); greatest alveolar length of upper tooth- row (from the first incisor to the last molar) (UTR); greatest breadth of palate (across labial edges of first upper molars) (Ml — Ml); greatest length of palate (from the anterior lip of the premaxilla to the posterior medial tip of the palatine) (PAL); height of skull (HS); greatest alveolar length of mandibular toothrow (from first incisor to the last molar) (MDTR); height of mandible (from the ventral edge of the angular process perpendicular to the tip of the condylus) (MDH); length of mandible (from the posterior edge of the condylus to the ventral edge of the alveolus of the first incisor) (MDL). The specimens of each species were divided into four cate¬ gories as follows: Juveniles: All those individuals with their deciduous teeth in various stages of eruption, but with none of the permanent teeth, including the molars, erupted. Subadults: All those individuals from when MI begins to erupt to that stage when the entire permanent dentition is fully erupted in both jaws and begins to exhibit slight wear on the occlusal surfaces. Replacement of all deciduous teeth occurs during this stage. Adults: All those individuals from when all permanent teeth are in apposition and exhibit some wear until all lingual cusps of upper cheek¬ teeth are worn away. Old Adults: All those individuals with the lingual cusps of the upper cheek-teeth worn away; these cheek-teeth appearing L-shaped in frontal view. Statistical comparisons were done by means of univariate analysis employing a program (UNIVAR) written by Powers (1970), and were performed on an IBM 170/ 158 computer. This program yields standard univariate statistics (sample size, mean, range, standard deviation, standard error of the mean, variance, and coefficient of variation). The samples for each sex were compared for significant secondary sexual variation, using a single-classification anova (F-test with significance level of 0.05) that tested for significant differences between means (Sokal and Rohlf, 1969). 226 Annals of Carnegie Museum vol. 46 Results age variation: After separation of the samples by sex and age, the only category with statistically valid sample sizes was the adult one. A few specimens that were barely assignable to the age category “Old Adult” were included in the adult category and were incorporated in the study of secondary sexual variation. Age variation for the five species, however, could not be analyzed statistically at this time because of small sample sizes of non-adults and well-described old adults. A comparison of measurements taken from a sample composed of five subadult males and females and a sample of adults of E. intufi are given in Table 1. Subadults were smaller than adults in all measurements compared. From these indications of size difference between subadults and adults of E. intufi, the same differences should be found for the other four species of Elephantulus. Table 1 Variation with age in a sample of Elephantulus intufi from 1 mi NE Tshipise, South Africa. Subadult Adult Measurement N Mean (Range) N Mean (Range) Total length 10 Length of tail 10 Length of hindfoot 10 Length of ear 10 Greatest length of skull 6 Condylobasal length 6 Greatest zygomatic breadth 8 Least interorbital breadth 9 Greatest breadth of braincase 7 Greatest length of nasals 9 Height of rostrum 10 Length of anterior palatine foramina 9 Length of posterior palatine foramina 9 Width of bulla 5 Length of upper toothrow 5 Greatest breadth of palate 9 Length of palate 8 Height of skull 4 Length of mandibular toothrow 5 Height of mandible 10 Length of mandible 10 231.8 (207-250) 11 240.1 (218-254) 114.6 (88-134) 11 119.5 (102-131) 33.0 (32-34) 16 33.2 (22-29) 22.9 (21-25) 16 23.8 (22-26) 31.7 (28.0-33.0) 16 34.0 (32.7-35.1) 29.5 (25.6-31.0) 14 31.5 (30.0-32.9) 18.3 (16.5-19.0) 13 19.4 (18.9-21.0) 6.1 (5. 9-6. 4) 16 6.6 (6. 2-7. 2) 14.1 (13.7-14.5) 15 14.6 (14.3-15.5) 11.7 (9.4-12.8) 16 13.8 (12.0-14.8) 3.3 (3.0-3.7) 16 3.7 (3. 4-3. 9) 3.1 (2.2-3. 7) 14 3.7 (3. 3-4. 3) 3.8 (3. 1-4.4) 15 4.6 (3.7-5. 2) 6.9 (6.0-7. 3) 14 7.3 (6. 7-7. 2) 16.8 (16.3-17.1) 16 17.2 (16.3-17.7) 10.4 (9.1-11.3) 18 11.3 (10.9-11.8) 17.3 (15.9-18.2) 15 19.0 (17.8-19.8) 13.0 (12.3-13.4) 12 13.6 (13.1-14.1) 15.7 (15.4-15.9) 16 15.5 (14.8-15.9) 9.1 (7.6-10.1) 15 9.6 (8.9-10.4) 24.0 (21.4-25.8) 15 25.3 (24.2-26.2) 1977 Nongeographic Variation In Elephant Shrews 227 Table 2 Univariate statistical analysis, measurements (in mm) of five species of southern African Elephantulus Species and sex N X 2 S.E. Min. Max. C.V. Total length E. edwardi male 9 249.3 9.53 230 273 5.73 female 11 250.8 8.75 230 275 5.79 E. brachyrhynchus male 9 213.8 5.40 202 225 3.79 female 4 212.0 5.72 205 218 2.70 E. rupestris male 24 269.0 4.61 239 292 4.20 female 19 274.5 4.83 255 292 3.84 E. intufi male 7 237.7 7.66 218 246 4.26 female 4 244.3 6.85 238 254 2.80 E. myurus male 21 275.1 3.40 259 288 2.83 female 13 276.3 4.33 264 292 2.83 Length of tail E. edwardi male 9 131.3 6.01 115 145 6.86 female 11 130.7 6.23 116 149 7.90 E. brachyrhynchus male 9 98.9 4.73 91 110 7.18 female 4 95.3 5.91 88 101 6.20 E. rupestris male 24 145.5 4.02 122 166 6.77 female 19 149.7 3.59 137 163 5.22 E. intufi male 7 118.3 5.96 102 128 6.67 female 4 121.8 6.29 117 131 5.17 E. myurus male 21 147.3 2.46 135 160 3.82 female 13 147.2 3.03 137 159 3.71 228 An-nals of Carnegie Museum vol. 46 Table 2 Univariate statistical analysis, measurements (in mm) of five species of southern African Elephantulus (Continued) Species and sex N X 2 S.E. Min. Max. C.V. Length of hindfoot E. edwardi male 9 35.0 0.82 33 37 3.50 female 12 34.8 0.88 33 39 4.39 E. brachyrhynchus male 9 29.9 0.52 29 31 2.62 female 6 28.8 0.80 27 30 3.41 E. rupestris male 26 36.5 0.46 34 38 3.22 female 20 37.1 0.41 35 39 2.46 E. intufi male 10 32.7 1.26 32 39 5.93 female 6 34.0 1.15 32 36 4.16 E. myurus male 24 39.5 0.45 38 42 2.79 female 13 39.3 Length 0.70 of ear 37 41 3.18 E. edwardi male 9 29.2 1.28 27 33 6.58 female 11 29.5 1.09 26 33 6.12 E. brachyrhynchus male 9 21.1 0.85 20 24 6.01 female 6 20.7 0.42 20 21 2.50 E. rupestris male 26 26.0 0.56 22 29 5.48 female 20 26.4 0.47 24 28 3.96 E. intufi male 10 23.7 0.52 22 25 3.47 female 6 24.0 0.89 23 26 4.56 E. myurus male 24 26.6 0.36 25 28 3.30 female 13 26.5 0.48 25 28 3.30 1977 Nongeographic Variation In Elephant Shrews 229 Table 2 Univariate statistical analysis, measurements (in mm) of five species of southern African Eleplnantulus (Continued) Species and sex N X 2 S.E. Min. Max. C.V. Greatest length of skull E. edwardi male 9 female 1 2 E. brachyrhynchus male 9 female 5 E. rupestris male 21 female 17 E. intufi male 10 female 6 E. myurus male 22 female 6 E. edwardi male 9 female 10 E. brachyrhunchus male 9 female 5 E. rupestris male 20 female 16 E. intufi male 8 female 6 E. myurus male 22 female 6 35.3 1.00 34.9 0.69 33.5 0.45 32.8 0.61 36.5 0.39 36.9 0.37 33.8 0.55 34.2 0.61 38.5 0.26 38.1 0.54 Condylobasal length 32.6 0.96 32.4 0.69 30.7 0.45 30.4 0.56 34.1 0.36 34.5 0.33 31.3 0.47 31.8 0.70 35.8 0.28 35.2 0.44 33.0 38.0 4.25 33.5 37.0 3.44 32.4 34.3 2.03 32.2 33.8 2.08 35.0 37.9 2.47 35.6 38.6 2.07 32.7 35.0 2.55 33.0 35.1 2.17 19.5 21.2 1.61 19.4 20.4 1.73 30.3 35.1 4.42 31.3 34.4 3.37 29.7 31.5 2.21 29.9 31.3 2.07 32.8 35.6 2.39 33.6 36.1 1.89 30.0 32.1 2.12 30.4 32.9 2.69 34.5 36.8 1.82 34.4 36.0 1.53 230 Annals of Carnegie Museum vol. 46 •'V Table 2 Univariate statistical analysis, measurements (in mm) of five species of southern African Elephantulus (Continued) Species and sex N X 2 S.E. Min. Max. C.V. Greatest zygomatic breadth E. edwardi male 8 19.6 0.30 19.1 20.5 2.20 female 12 19.3 0.30 18.7 20.3 2.69 E. brachyrhynchus male 9 18.1 0.28 17.6 18.8 2.29 female 5 17.5 0.64 16.8 18.5 4.06 E. rupestris male 24 20.0 0.17 19.2 20.9 2.05 female 16 20.0 0.24 19.1 20.9 2.38 E. intufi male 8 19.3 0.20 18.9 19.8 1.48 female 5 19.6 0.35 19.0 20.0 1.99 E. myurus male 20 20.4 0.17 19.5 21.2 1.87 female 7 20.1 0.29 19.4 20.4 1.91 Least interorbital breadth E. edwardi male 9 7.3 0.17 6.9 7.8 3.54 female 12 7.2 0.13 6.8 7.6 3.12 E. brachyrhynchus male 9 5.9 0.17 5.5 6.2 4.23 female 6 5.9 0.20 5.6 6.3 4.22 E. rupestris male 26 7.0 0.11 6.5 7.8 3.85 female 19 7.0 0.14 6.5 7.5 4.53 E. intufi male 10 6.6 0.18 6.2 7.2 4.33 female 6 6.5 0.16 6.2 6.8 2.98 E. myurus male 27 7.5 0.08 7.1 7.9 2.63 female 12 7.5 0.12 7.0 7.8 2.68 1977 Nongeographic Variation In Elephant Shrews 231 Table 2 Univariate statistical analysis, measurements (in mm) of five species of southern African Elephantulus (Continued) Species and sex N X 2 S.E. Min. Max. C.V. Greatest breadth of braincase E. edwardi male 9 15.1 0.17 14.7 15.6 1.70 female 12 15.1 0.14 14.7 15.5 1.66 E. brachyrhynchus male 9 13.8 0.21 13.8 14.1 2.29 female 5 13.5 0.22 13.2 13.9 1.85 E. rupestris male 25 15.1 0.13 14.4 15.8 2.10 female 18 15.1 0.17 14.6 15.9 2.44 E. intufi male 9 14.6 0.29 14.0 15.5 2.95 female 6 14.6 0.19 14.4 15.0 1.59 E. myurus male 25 15.8 0.11 15.2 16.3 1.77 female 9 15.6 0.20 15.1 15.9 1.90 Greatest length of nasals E. edwardi male 9 13.7 0.78 11.6 15.6 8.52 female 12 13.6 0.50 12.5 15.1 6.40 E. brachyrhynchus male 9 11.6 1.00 9.0 12.9 13.00 female 6 11.7 0.91 9.4 12.4 9.62 E. rupestris male 25 14.5 0.27 12.4 15.6 4.56 female 20 15.0 0.26 14.1 16.5 3.84 E. intufi male 10 13.5 0.42 12.0 14.1 4.90 female 6 14.2 0.61 12.8 14.8 5.25 E. myurus male 27 15.1 0.22 13.9 16.1 3.87 female 13 14.9 0.79 14.1 16.0 3.55 232 Annals of Carnegie Museum vol. 46 Table 2 Univariate statistical analysis, measurements (in mm) of five species of southern African Elephantulus (Continued) Species and sex N X 2 S.E. Min. Max. C.V. Height of rostrum E. edwardi male 9 3.9 0.09 3.7 4.1 3.66 female 12 3.8 0.11 3.5 4.1 4.90 E. brachyrhynchus male 9 3.6 0.08 3.4 3.8 3.52 female 6 3.3 0.12 3.1 3.5 4.52 E. rupestris male 26 3.6 0.09 3.2 3.9 6.20 female 20 3.6 0.07 3.3 3.9 4.37 E. intufi male 10 3.7 0.10 3.4 3.8 4.13 female 6 3.7 0.16 3.4 3.9 5.36 E. myurus male 27 3.8 0.06 3.3 4.0 4.10 female 13 3.7 0.06 Length of anterior palatine 3.6 foramina 3.9 2.77 E. edwardi male 8 4.1 0.16 3.8 4.5 5.46 female 12 4.3 0.19 3.8 4.8 7.64 E. brahyrhynchus male 9 3.7 0.26 3.3 4.5 10.35 female 6 3.9 0.20 3.5 4.2 6.26 E. rupestris male 26 4.4 0.14 3.6 5.0 8.24 female 20 4.4 0.19 3.1 5.0 9.60 E. intufi male 9 3.7 0.18 3.3 4.1 7.25 female 5 3.7 0.37 3.3 4.3 11.12 E. myurus male 27 4.3 0.12 3.7 4.9 7.26 female 13 4.5 0.25 3.6 5.1 10.07 1977 Nongeographic Variation In Elephant Shrews 233 Table 2 Univariate statistical analysis, measurements (in mm) of five species of southern African Elephantulus (Continued) Species and sex N X 2 S.E. Min. Max. C.V. Length of posterior palatine foramina E. edwardi male 9 4.2 0.31 3.7 5.1 10.98 female 12 4.4 0.17 3.9 4.9 6.74 E. brachyrhynchus male 9 4.6 0.19 4.1 4.9 6.22 female 6 4.8 0.20 4.5 5.1 5.28 E. rupestris male 24 5.2 0.15 4.5 5.8 6.93 female 19 5.2 0.22 4.5 6.4 9.42 E. intufi male 10 4.5 0.35 3.7 5.2 12.21 female 5 4.7 0.31 4.2 5.1 7.53 E. myurus male 27 5.5 0.21 3.7 6.1 9.97 female 12 5.5 0.17 5.2 6.1 5.25 Width of bulla E. edwardi male 9 7.6 0.20 7.0 8.1 4.00 female 12 7.5 0.20 6.9 8.0 4.64 E. brachyrhynchus male 9 6.1 0.25 5.5 6.5 6.11 female 5 6.0 0.28 5.5 6.3 5.25 E. rupestris male 23 7.3 0.20 6.2 8.1 6.89 female 16 7.2 0.20 6.5 7.9 5.54 E. intufi male 8 7.3 0.31 6.7 7.9 6.02 female 6 7.3 0.08 7.2 7.5 1.41 E. myurus male 22 7.5 0.13 7.1 8.5 4.16 female 5 7.5 0.07 7.4 7.6 1.12 234 Annals of Carnegie Museum vol. 46 Table 2 Univariate statistical analysis, measurements (in mm) of five species of southern African Elephantulus (Continued) Species and sex N X 2 S.E. Min. Max. C.V. Length of upper toothrow E. edwardi male 9 18.2 0.62 17.1 19.9 5.13 female 12 18.3 0.57 17.5 20.8 5.39 E. brachyrhynchus male 9 17.0 0.26 16.4 17.6 2.29 female 6 17.0 0.11 16.8 17.2 0.80 E. rupestris male 26 19.1 0.16 18.2 20.1 2.20 female 20 19.1 0.14 18.4 19.7 1.69 E. intufi male 10 17.1 0.31 16.3 17.7 2.83 female 6 17.3 0.29 16.7 17.7 2.05 E. myurus male 25 20.6 0.13 19.9 21.2 1.52 female 12 20.6 0.23 20.0 21.3 1.90 Greatest breadth of palate E. edwardi male 9 12.4 0.27 11.7 12.8 3.23 female 12 12.2 0.18 11.8 12.8 2.59 E. brachyrhynchus male 9 11.2 0.17 10.8 11.5 2.27 female 5 10.9 0.23 10.6 11.2 2.34 E. rupestris male 26 11.8 0.13 11.2 12.5 2.78 female 20 11.8 0.15 11.1 12.3 2.90 E. intufi male 10 11.3 0.15 11.0 11.8 2.12 female 4 11.3 0.30 10.9 11.6 2.65 E. myurus male 26 12.6 0.19 12.2 13.0 1.84 female 12 12.5 0.13 12.2 12.9 1.86 1977 Nongeographic Variation In Elephant Shrews 235 Table 2 Univariate statistical analysis, measurements (in mm) of five species of southern African Elephantulus (Continued) Species and sex N X 2 S.E. Min. Max. C.V. Length of palate E. edwardi male 7 19.9 0.93 18.1 21.8 6.23 female 9 19.9 0.59 18.9 21.3 4.47 E. brachyrhynchus male 9 18.9 0.31 18.2 19.5 2.46 female 6 18.5 0.30 18.0 18.9 1.96 E. rupestris male 24 21.1 0.21 20.1 22.0 2.44 female 17 21.3 0.24 20.6 22.5 2.37 E. intufi male 10 18.9 0.41 17.8 19.8 3.46 female 5 18.9 0.52 17.9 19.3 3.06 E. myurus male 20 22.2 0.18 21.5 23.2 1.84 female 8 22.5 0.38 Height of skull 21.7 23.4 2.39 E. edwardi male 9 12.8 0.14 12.4 13.0 1.59 female 11 12.9 0.12 12.5 13.1 1.49 E. brachyrhynchus male 9 13.2 0.31 12.5 14.0 3.54 female 5 12.7 0.28 12.4 13.1 2.39 E. rupestris male 22 13.8 0.16 13.1 14.4 2.80 female 16 13.9 0.16 13.4 14.5 2.32 E. intufi male 6 13.5 0.30 13.1 14.0 2.73 female 6 13.7 0.23 13.3 14.1 2.04 E. myurus male 22 13.0 0.11 12.6 13.5 1.96 female 5 13.0 0.27 12.5 13.6 3.21 236 Annals of Carnegie Museum vol. 46 Table 2 Univariate statistical analysis, measurements (in mm) of five species of southern African Elephantulus (Continued) Species and sex N X 2 S.E. Min. Max. C.V. Length of mandibular toothrow E. edwardi male 9 female 12 E. brachyrhynchus male 9 female 6 E. rupestris male 26 female 20 E. intufi male 10 female 6 E. myurus male 27 female 1 3 E. edwardi male 9 female 12 E. brachyrhynchus male 9 female 6 E. rupestris male 26 female 20 E. intufi male 10 female 5 E. myurus male 27 female 12 16.1 0.52 16.3 0.35 16.1 0.30 16.0 0.30 17.4 0.17 17.4 0.15 15.5 0.26 15.6 0.31 18.6 0.10 18.6 0.20 Height of mandible 10.1 0.21 10.3 0.23 9.6 0.23 9.5 0.28 10.2 0.15 10.4 0.15 9.6 0.28 9.6 0.23 10.2 0.12 10.1 0.24 15.1 17.4 4.90 15.5 17.2 3.77 15.4 16.5 2.78 15.6 16.6 2.30 16.7 18.4 2.42 16.9 18.1 1.93 14.8 16.0 2.62 15.0 16.1 2.43 17.8 18.9 1.44 18.1 19.2 1.94 9.4 10.5 3.19 9.8 11.2 3.81 9.3 10.4 3.57 9.1 10.1 3.58 9.4 10.9 3.73 9.8 11.1 3.32 8.9 10.4 4.64 9.2 9.9 2.73 9.8 10.9 3.07 9.5 10.6 4.13 1977 Nongeographic Variation In Elephant Shrews 237 Table 2 Univariate statistical analysis, measurements (in mm) of five species of southern African Elephantulus (Continued) Species and sex N X 2 S.E. Min. Max. C.V. Length of mandible E. edwardi male 9 26.2 0.74 24.5 28.1 4.23 female 12 25.9 0.65 24.5 27.0 4.32 E. brachyhynchus male 9 24.9 0.37 23.9 25.7 2.23 female 6 24.5 0.54 23.6 25.3 2.69 E. rupestris male 26 27.9 0.26 26.7 29.1 2.35 female 20 28.2 0.29 27.2 29.9 2.27 E. intufi male 10 25.2 0.41 24.2 26.0 2.56 female 5 25.7 0.44 24.9 26.2 1.89 E. myurus male 27 29.1 0.32 26.4 30.2 2.81 female 12 29.2 0.36 28.2 30.2 2.14 INDIVIDUAL VARIATION: : Results of the analyses of individual secondary sexual variation are given in Table 2 and in Figure 1. Initially, body weight was recorded and analyzed, but this measurement proved to be extremely variable. The distance between the upper canine and first upper premolar was also recorded, but because of the measuring technique, the measurement proved to be unsuitable for statistical analysis and was deleted. Abbreviations used in Table 2 are: N = sample size, x = arithmetic mean, 2 S.E. = 2 standard errors of the mean, min. = minimum value of range of variable, max. = maximum value of range of variable, and C.V. = coefficient of variation. Significant secondary sexual variation between Fig. 1: Comparisons of coefficients of variation of four external and 17 cranial measure¬ ments for samples of males and females of five species of southern African Elephantulus. The definitions of the abbreviations of the 22 variables are given in Materials and Methods. Samples of each species analyzed are indicated as follows: e = E. edwardi, b = E. brachyr- hynchus, r - E. rupestris, i = E. intufi and m - E. myurus. Letters on the left of the vertical line above each variable refer to the coefficient of variation for the sample of males for the respective species, while those on the right refer to the coefficient of variation for the sample of females. ^ AO 238 Annals of Carnegie Museum vol. 46 14.00 -i 13.00 - b-i 12.00 - 11.00 - 10.00 - 9.00 — 8.00 7.00- 6.00- 5.00 4.00 3.00- 2.00- 1.00 0.00 r- b-r~r b- -b r_ m-f-rn 6= m- j-r m -r nr -m -i l_i ■8=: hm b n i i - r IOB BBC GLN HR TOT HFT ^ i r GLS CBL GZB VARIABLE 1977 Nongeographic Variation In Elephant Shrews 239 -b b- r--i Fe r_ b~-m -b b- m m-*~m m-i ^ b-, -m r-~e b4-b e. m- -b “1 - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1 - T“ LAP LPP WB UTR MI-MI PAL HS MDTR MDH MDL VARIABLE 1—14.00 -13.00 -12.00 -11.00 -10.00 - 9.00 8.00 - 7.00 - 6.00 5.00 - 4.00 3.00 - 2.00 - 1.00 0.00 A'O 240 Annals of Carnegie -Museum vol. 46 means is indicated on the right of the variable by single asterisk for probability greater than 0.05, and by two asterisks for probability greater than 0.01. The majority of variables analyzed revealed a relatively low degree of individual variation in all five species of Elephantulus that were con¬ sidered. Most coefficients of variation were less than 6.00 (Table 2 and Figure 1); these values are within the limits of those expected for small mammals (Long, 1968; 1969; 1970; Genoways, 1973). The four external measurements have a slightly higher coefficient of variation than do the cranial variables. These higher values may be a result of the somewhat cruder method in which external measurements are taken, as well as the fact that numerous individuals took these mea¬ surements when the specimens originally were prepared. The lengths of the anterior and posterior palatine foramina and the nasals are the most variable cranial measurements tested. The length of the nasals shows not only the highest coefficients of variation (13.00 in male E. brachyrhynchus) but also the greatest range of values (13.00 in male E. brachyrhynchus to 3.87 in male E. myurus). The high values for each of these three variables are indicative of the variation that exists in the size and shape of these features. The two palatine foramina are individually somewhat divided with septa, thus making it difficult to measure them. In the case of the nasal bones, the posterior edge of the nasals is often difficult to establish with precision. As the anterior edge of the nasals is naturally square, it is possible to inadvertently record this measurement from specimens where the anterior tip has been cut off slightly during the process of field preparation of the specimens. Figure 1 shows those cranial measurements with low coefficients of variation. Included in these measurements are greatest length of skull, condylobasal length of skull, greatest zygomatic breadth, least inter¬ orbital breadth, greatest breadth of braincase, greatest alveolar length of mandibular toothrow, height of mandible, and length of mandible. Of the five species of Elephantulus examined, E. edwardi generally exhibited the highest coefficients of variation in both external and cra¬ nial variables (Table 2 and Figure 1). However, even though this species generally has the highest coefficients of variation for the variables tested, only the length of palate in males had a value over 6.00. In those variables with extremely high coefficients of variation E. edwardi had values in the middle to lower end of the range of coefficients of variation for the other species (Figure 1). The lowest coefficients of variation were most often found in E. myurus. secondary sexual variation: Of all the 21 variables in the five species, those for males averaged larger than those for females on 41 occasions. The reverse was true on 39 occasions. The differences be¬ tween sexes were slight for 25 variables. However, statistically signi- 1977 Nongeographic Variation In Elephant Shrews 241 ficant differences between the sexes were obtained in only six instances in three species (Table 2). In E. brachyrhynchus, females were significantly larger than males at the 0.01 significance level in the height of rostrum. Again, in the length of hindfoot and the greatest breadth of palate, females were significantly larger than males in E. brachyrhynchus at the 0.05 level of significance. The greatest breadth of braincase in females is signifi¬ cantly larger than in males of E. myurus at the 0.05 significance level. Males of E. rupestris are significantly (0.05 level) larger than females in length of nasals and greatest height of mandible. These were the only variables with significant secondary sexual variation. No variables analyzed from the samples of E. edwardi and E. intufi showed signifi¬ cant differences between the sexes. Summary An analysis of individual and secondary sexual variation in 21 external and cranial measurements in samples divided by sexes of five species of southern African Elephantulus revealed little nongeographic varia¬ tion. Of the five species studied — E. edwardi (A. Smith, 1839), E. bra¬ chyrhynchus (A. Smith, 1836), E. rupestris (A. Smith, 1831), E. intufi (A. Smith, 1836) and E. myurus Thomas and Schwann, 1906 — E. edwardi exhibited the most individual variation in the 21 variables, whereas E. myurus generally showed the least amount of individual variation as reflected in the coefficients of variation. Most variables had coefficients of variation below 6.00 and thus fell within the limits of those expected for small mammals. Three species (E. brachyrhynchus, E. rupestris and E. myurus) of the five analyzed had statistically significant secondary sexual variation. Females of E. brachyrhynchus were significantly larger than males in three variables — height of rostrum, length of hindfoot, and greatest breadth of palate. In E. myurus, females were significantly larger than males in the greatest breadth of braincase, whereas males of E. rupe¬ stris were significantly larger than females in length of nasals and greatest height of mandible. There were no statistically significant differences between sexes in E. edwardi and E. intufi. Based on these results, samples of adults of these five species of elephant shrews can be combined in further studies of geographic varia¬ tion, with the exception of those six variables in three species mentioned above. Judging from these results, elephant shrews have little individual and secondary sexual variation as a group. 242 Annals of Carnegie Museum vol. 46 Acknowledgments This study was completed while the senior author visited the Section of Mammals, Carnegie Museum of Natural History, as a Resident Museum Specialist in the Museum’s International Visitor Program. The Director and Board of Trustees of the Transvaal Museum, Pretoria, have his thanks for allowing him to participate in this program, as well as for additional financial support. Henry W. Setzer, Smithsonian Institution, suggested a long-term investigation into geographic variation in species of southern African Elephantulus and consequently made available the specimens used in this initial study. Thomas J. McIntyre helped make possible numerous trips to Washington, D.C. Hugh H. Genoways critically read the manu¬ script. Stephen L. Williams and John F. Sutton assisted with the com¬ puter analysis, and Herbert Leifer prepared Figure 1. We thank all of them for their contributions. 1977 Nongeographic Variation In Elephant Shrews 243 References Cited Butler, P.M., and M. Greenwood 1976. Elephant-shrews (Macroscelididae) from Olduvai and Makapansgat, pp. 1-56, in Fossil Vertebrates of Africa (R.J.G. Savage and S.C. Coryndon, eds.), Academic Press, London, 4: xi + 1-338. Corbet, G.B. 1971. Part 1.5 Family Macroscelididae, pp. 1-6, in The Mammals of Africa — An Identi¬ fication Manual (J. Meester and H. W. Setzer, eds.). Smithsonian Institution Press, Washington, D.C. Corbet, G.B., and J. Hanks 1968. A revision of the elephant-shrews, family Macroscelididae. Bull. British Mus. Nat. Hist. (Zool.), 16:45-111. Corbet, G. B., and B. R. Neal 1965. The taxonomy of the elephant-shrews of the Genus Petrodromus, with particular reference to the East African coast. Rev. Zool. Bot. Africa, 71:49-78. Ellerman, J. R., T. C. S. Morrison-Scott and R. W. Hayman 1953. Southern African Mammals 1758 to 1951: A Reclassification. Trustees British Museum (Natural History), London, 363 pp. Genoways, H. H. 1973. Systematics and Evolutionary Relationships of Spiny Pocket Mice, Genus Liomys. Spec. Publ. Mus., Texas Tech Univ., 5:1-368. Long, C. A. 1968. An analysis of patterns of variation in some representative Mammalia. Part I. A review of estimates of variability in selected measurements. Trans. Kansas Acad. Sci., 71:201-227. 1969. An analysis of patterns of variation in some representative Mammalia. Part II. Studies on the nature and correlation of measures of variation, pp. 289-302, in Contributions in Mammalogy (J. K. Jones, Jr. ed.), Misc. Publ. Mus. Nat. Hist., Univ. Kansas, 51:1-428. 1970. An analysis of patterns of variation in some representative Mammalia. Part III. Some equations of the nature of frequency distributions of estimated variabilities. Acta Theriol., 15:517-520. Meester, J., D. H. S. Davis and C. G. Coetzee 1964. An Interim Classification of Southern African Mammals. Raneved, distributed with the assistance of the Zoological Society of Southern Africa and the C.S.I.R., 75 pp. Powers, D. M. 1970. Geographic variation of red-winged blackbirds in central North America. Univ. Kansas Publ., Mus. Nat. Hist., 19:1-83. Roberts, A. 1951. The mammals of South Africa. Trustees, The Mammals of South Africa Book Fund, Johannesburg, xlviii + 700 pp. Shortridge, G. C. 1934. The Mammals of South West Africa. Volume I. William Heinemann Ltd., London, xxv + 437 pp., illus. SOKAL, R. R., AND F. J. ROHLF 1969. Biometry: the principles and practice of statistics in biological research. W. H. Freeman and Co., San Francisco, xiii + 776 pp. Back issues of many Annals of Carnegie Museum articles are available, and a few early complete volumes and parts are listed at half price. Orders and inquiries should be addressed to: Publications Secretary, Carnegie Museum, 4400 Forbes Avenue, Pittsburgh, Pa. 15213. MUs' COMP. ZOOL library NOV 1 4 1977 AN NALS issn 0097-4463 HARVARD university of CARNEGIE MUSEUM CARNEGIE MUSEUM OF NATURAL HISTORY 4400 FORBES AVENUE • PITTSBURGH, PENNSYLVANIA 15213 VOLUME 46 NOVEMBER 2, 1977 ARTICLE 15 MORPHOMETRIC VARIATION IN THE TROPICAL POCKET GOPHER (GEOMYS TROPICALIS) Stephen L. Williams Hugh H. Genoways Section of Mammals Abstract The tropical pocket gopher (Geomys tropicalis ), which exhibits no chromosomal or genic variation, was examined for variability at the morphometrical level. Univariate and mul¬ tivariate analyses were used to determine age, sexual, and geographical variation. Signifi¬ cant differences were found between different age classes and between sexes. The amount of individual variation was comparable with other rodents and did not exhibit the reduced variation expressed at the chromosomal and genic levels. G. tropicalis is considered to be a monotypic species. Introduction The tropical pocket gopher (Geomys tropicalis ) was originally de¬ scribed by Goldman (1915:134) as a subspecies of G. personatus based upon specimens from Altamira, Tamaulipas. Little additional informa¬ tion became available on this taxon until Alvarez (1963) reviewed its status. He concluded, based upon specimens from Altamira and 1 mi. S Altamira, that tropicalis is a distinct species of the genus Geomys. Subsequent studies of chromosomal (Davis et al., 1971; Baker and Wil¬ liams, 1974; Selander et al., 1975) and genic relationships (Selander et al., 1975) have supported the specific status of this taxon. Geomys tropicalis is currently known from only seven localities in extreme southern Tamaulipas. These localities are separated from the nearest population of the genus (G. personatus) by 250 kilometers. It has been estimated (Selander et al., 1975) that the entire geographic Submitted for publication April 13, 1977. 245 246 Annals of Carnegie Museum vol. 46 range of G. tropicalis occupies only approximately 300 square kilo¬ meters. Chromosomal and genic characteristics indicate reduced vari¬ ability in this species. Davis et al. (1971) reported the 2N of the G. tropicalis to be 38 and the FN to be 72. This diploid number is consid¬ erably lower than other members of the Geomys bursarius species-group (2N = 68 to 72),' although it is approached by Geomys pinetis of the pinetis species-group (2N = 42, FN = 80; Williams and Genoways, 1975). Davis et al (1971) stated that the reduced diploid number should reduce Mendelian variation and could be a method of adaptation to local conditions by G. tropicalis. Selander et al. (1975) found G. tropicalis to be monogenic at 34 struc¬ tural gene loci encoding enzymes and other proteins. They presented two possible explanations for this reduced variability. One is to postulate a series of founding events as ancestoral stock worked its way southward from Texas along the narrow barrier, islands of the Tamaulipas coast. Variability lost during these founding events could not be restored be¬ cause the disjunct distribution would prevent gene flow. The alternative, but not mutually exclusive, hypothesis is that the reduced variability is the result of adaptation to an extremely uniform and restricted habitat (an area of old sand dunes). Whatever the explanation, it is clear from the current data that Geomys tropicalis has reduced variability at the chromosomal and genic levels. Therefore, we have taken the opportunity to examine variability in this species at the morphometric level. The sample of 94 specimens that we have available is considerably larger than those available to previous authors (Alvarez, 1963, examined 19 specimens). We have examined morphometric variation as the result of both nongeographic and geo¬ graphic factors. Methods From all specimens, three external measurements and 13 cranial mea¬ surements were recorded. The external measurements are those of the collector; cranial measurements were taken by means of dial calipers. In order to clarify how cranial measurements were taken, we have de¬ fined each below. Numbers that precede each character correspond to those numbers used in Figure 1. All measurements in text are given in millimeters. 1. Greatest length of skull — distance between two vertical lines, one touching the posteriormost part of the skull at the (exoccipital condyles or supraoccipital) occipital bone, and the other touching the anteriormost part of the nasal bones. 2. Condylobasal length — distance on skull from a line connecting the posteriormost projections of the exoccipital condyles to a line connecting the anteriormost projections of the premaxillary bones. 3. Basal length — distance on skull from the anteriormost inferior border 1977 Tropical Pocket Gopher 247 Fig. 1. Skull of Geomys tropicalis illustrating points between which measurements de¬ scribed in text were taken. 248 Annals of Carnegie Museum vol. 46 of the foramen magnum to a line connecting the anteriormost projections of the premaxillary bones. 4. Palatal length — distance on skull from a line connecting the posterior margin of palate to a line connecting the anteriormost projections of the premaxillary bones. 5. Palatofrontal depth — least distance between two parallel planes, one touching the mid-point of the frontal bones and the other touching the ventralmost point of palatine bones between the molar teeth. 6. Length of nasals — greatest length of nasal bones. 7. Diastema — distance along a straight line between the posterior mar¬ gin of the alveolus of the incisors and anterior margin of the alveo¬ lus of the premolars. 8. Zygomatic breadth — greatest distance across zygomatic arches of cranium at right angles to long axis of skull. 9. Mastoid breadth — greatest distance between mastoid processes measured at right angles to the long axis of the skull. 10. Squamosal breadth — least distance between lateral margins of the squamosal bones measured at right angles to long axis of the skull. 11. Rostral breadth — greatest distance across rostrum. 12. Inter orbital constriction — least distance across top of the skull between the orbits. 13. Breadth across maxillaries — greatest distance between maxilliary bones measured at the level of M3. All specimens examined were assigned to one of three age groups described below. Juvenile. — gap between basisphenoid and basioccipital; juvenile palage; no sagital crest; zygomatic breadth nearly equal to or less than mas¬ toid breadth. Subadult. — basisphenoid and basioccipital connected but not fused; adult pelage; wide gap between sagital crests; sagital crests poorly devel¬ oped; zygomatic breadth usually not more than 1 mm greater than mastoid breadth. Adult. — basisphenoid and basioccipital partially or entirely fused; adult pelage; sagital crests well developed; zygomatic breadth always more than 1 mm greater than mastoid breadth. For the univariate analysis of geographic variation, adult specimens were grouped into two samples as follows: sample 1 — 2.5 mi. SSE Alta- mira, 2.4 mi. S Altamira; sample 2 — Altamira, 1 mi. S Altamira, 10 mi. NW Tampico. Statistical procedures were performed on the IBM 370 computer at Texas Tech University. Univariate analyses were performed using the program UNIVAR. This program yields standard statistics (mean, range, standard deviations, standard error of the mean, variance, and coeffi- 1977 Tropical Pocket Gopher 249 dent of variation) and employs a single-classification analysis of vari¬ ance (F-test, significance level .05) to test for significant differences between or among means (Sokal and Rohlf, 1969). When means were found to be significantly different, the Sum of Squares Simultaneous Test Procedure (SS-STP) developed by Gabriel (1964) was used to determine maximally nonsignificant subsets. Results Nongeographic Variation age variation: Table 1 shows the results of the analyses for varia¬ tion with age in males and females. In 12 measurements for males (all cranial measurements except interorbital constriction), adults, sub¬ adults, and juveniles form non-overlapping subsets. Adults and sub¬ adults form a subset that does not overlap the juvenile subset for total length and length of tail. Overlapping subsets are formed for length of hind foot and interorbital constriction. In all measurements of males, adults averaged the largest, except in interorbital constriction in which subadults were largest. Adults, subadults, and juveniles form non-overlapping subsets for all cranial measurements for females. In the three external measurements, adults and subadults form a subset that did not overlap with the juve¬ niles. Adult females averaged larger than other age classes in all mea¬ surements except length of tail and length of hind foot. Clearly, the three age classes that we recognized are morphometrically distinct. In almost all measurements, the three classes formed separate units. In the analyses of geographic variation that follow, only adults have been used. secondary sexual variation: Adult males and females were tested for secondary sexual variation. The same adult samples as used in age variation were used in this analysis (Table 1). Males averaged sig¬ nificantly larger than females in all measurements except interorbital constriction. In interorbital constriction, adult females averaged, but not significantly so, slightly larger than adult males. In subsequent analyses, males have been analyzed separate from females. individual variation: Coefficients of variation for adult males ranged from 2.6 to 12.4 for the 16 external and cranial measurements tested (Table 1). Interorbital constriction had the lowest value and palatal length had the highest. The mean coefficient of variation for the 16 measurements was 5.2. Coefficients of variation for adult females ranged from 2.1 to 8.0. Rostral breadth had the lowest coefficient of variation and length of tail had the highest. The mean coefficient of variation for the 16 measurements was 3.8. In only three measurements (length of tail, nasal length, and interorbital constriction) did adult females have a higher coefficient of variation than adult males. 250 Annals of Carnegie Museum vol. 46 Table 1. Variation with age in external and cranial measurements of Geomys tropicalis. Age classes as described in text were tested for significant differences at the .05 level. Groups of means that were found to be significantly different were tested with Sums of Squares - Simultaneous Testing Procedures to determine the maximally nonsigni¬ ficant subsets. The adult samples as listed in this table were used to test for secondary sexual variation. See text for results of this analysis. Sex and age N Mean + 2 SE (Range) CV Result SS-STP class Total length Males Adult 10 252.6+ 7.96 (231.0-270.0) 5.0 I Subadult 15 245.3 ± 6.08 (225.0-265.0) 4.8 I Juvenile 6 217.3 + 9.52 (198.0-230.0) 5.4 I Females Adult 23 231.5 ± 4.19 (212.0-250.0) 4.3 I Subadult 25 226.1 + 4.36 (210.0-250.0) 4.8 I Juvenile 10 194.6+9.39 (170.0-222.0) 7.6 I Length of tail Males Adult 10 79.8+ 3.94 (71.0-93.0) 7.8 I Subadult 15 78.6 + 4.36 (62.0-89.0) 10.8 I Juvenile 6 68.2 + 7.16 (55.0-79.0) 12.9 I Females Adult 23 74.3+2.47 (61.0-85.0) 8.0 I Subadult 25 75.0+2.59 (57.0-85.0) 8.6 I Juvenile 9 64.4+4.22 (58.0-76.0) 9.8 I Length of hind foot Males Adult 10 32.0+ 1.08 (30.0-35.0) 5.3 I Subadult 15 31.3 + 0.91 (29.0-35.0) 5.6 I I Juvenile 6 29.6 ± 1.51 (26.0-31.0) 6.3 I Females Adult 24 29.8 + 0.05 (28.0-33.0) 4.1 I Subadult 25 30.0 + 0.72 (27.0-33.0) 6.0 I Juvenile 10 28.0+ 1.03 (25.0-31.0) 5.8 I Greatest length of skull Males Adult 9 47.3 + 1.24 (44.8-50.5) 3.9 I Subadult 13 44.5 + 1.15 (41.7-48.7) 4.7 I Juvenile 5 37.9 + 2.26 (34.8-41.1) 6.7 1977 Tropical Pocket Gopher 251 Table 1. (Continued) Sex and age N Mean + 2 SE (Range) CV Result SS-STP class Greatest length of skull (Continued) Females Adult 24 42.5 ± 0.48 (40.1-44.3) 2.8 I Subadult 24 40.8 + 0.88 (37.2-44.6) 5.3 I Juvenile 7 35.6 + 0.79 (34.3-37.5) 3.0 I Condylobasal length Males Adult 10 45.8+ 1.21 (42.9-49.4) 4.2 I Subadult 14 43.2+ 1.03 (40.5-46.7) 4.5 I Juvenile 5 36.9+2.19 (33.8-40.1) 6.6 I Females Adult 26 41.3+0.45 (39.0-43.1) 2.7 I Subadult 20 39.8+0.93 (36.2-43.3) 5.2 I Juvenile 8 34.7 + 0.86 (32.9-37.0) 3.5 I Basal length Males Adult 10 43.2+ 1.16 (40.4-46.3) 4.3 I Subadult 14 40.3 + 1.04 (37.8-44.2) 4.8 I Juvenile 5 33.7 + 2.08 (30.9-36.3) 6.9 I Females Adult 26 38.6+0.42 (36.6-40.4) 2.8 I Subadult 20 36.9 + 0.95 (32.9-40.3) 5.8 I Juvenile 8 31.5 + 0.86 (29.2-33.5) 3.8 I Palatal length Males Adult 10 32.3+2.54 (28.0-39.7) 12.4 I Subadult 14 27.7 + 0.78 (25.6-30.4) 5.3 I Juvenile 6 23.0+ 1.52 (20.4-25.0) 8.1 I Females Adult 26 26.6+0.33 (25.0-28.3) 3.2 I Subadult 20 25.1 + 0.69 (22.3-27.8) 6.1 I Juvenile 8 21.2+0.55 (19.7-22.4) 3.7 I 252 Annals of Carnegie Museum vol. 46 Table 1. (Continued) Sex and age N Mean + 2 SE (Range) CV Result SS-STP class Palatofrontal depth Males Adult 10 16.9+0.45 (15.8-18.1) 4.2 I Subadult 15 15.8 + 0.35 (14.9-17.1) 4.3 I Juvenile 6 13.7 + 0.43 (12.9-14.4) 3.8 I Females Adult 26 15.5 + 0.22 (14.5-16.5) 3.6 I Subadult 25 14.7 + 0.29 (13.1-16.1) 5.0 I Juvenile 10 13.1 + 0.37 (12.4-14.3) 4.4 I Length of nasals Males Adult 9 16.6 + 0.43 (15.9-17.5) 3.9 I Subadult 13 15.4 + 0.55 (13.9-17.6) 6.4 I Juvenile 6 12.6 + 0.83 (11.0-13.5) 8.1 I Females Adult 24 14.4+0.32 (12.8-16.3) 5.4 I Subadult 24 13.5 + 0.43 (12.0-15.5) 7.8 I Juvenile 7 11.3 ±0.40 (10.4-12.0) 4.7 I Diastema Males Adult 10 16.7+0.66 (15.2-18.4) 6.3 I Subadult 15 14.7+0.50 (13.4-16.2) 6.6 I Juvenile 6 11.5 + 0.87 (10.2-12.6) 9.3 I Females Adult 26 13.9 + 0.23 (12.7-15.2) 4.3 I Subadult 25 12.9 + 0.41 (10.9-14.7) 8.0 I Juvenile 10 10.6+ 0.40 (9.5-11.7) 5.9 I Zygomatic breadth Males Adult 10 29.9+ 1.00 (28.1-32.9) 5.3 I Subadult 14 27.4+0.93 (24.7-30.5) 6.3 I Juvenile 6 22.5+ 1.41 (20.1-24.4) 7.7 I Females Adult 26 26.0+0.39 (24.2-27.8) 3.8 I Subadult 24 24.2 + 0.62 (21.6-26.8) 6.2 I Juvenile 9 20.9 + 0.66 (18.9-22.7) 4.8 I 1977 Tropical Pocket Gopher 253 Table 1. (Continued) Sex and age N Mean + 2 SE (Range) CV Result SS-STP class Squamosal breadth Males Adult 10 20.8 + 0.53 (19.8-22.3) 4.0 I Subadult 15 19.5 + 0.45 (18.1-21.5) 4.5 I Juvenile 6 18.0+0.62 (17.1-19.1) 4.2 I Females Adult 26 19.2 + 0.26 (17.4-20.1) 3.5 I Subadult 25 18.3 + 0.45 (14.3-20.0) 6.2 I Juvenile 10 17.2+0.36 (16.3-18.4) 3.3 I Mastoid breadth Males Adult 10 27.1 + 0.66 (25.5-28.9) 3.8 I Subadult 15 25.1 + 0.65 (23.1-27.3) 5.0 I Juvenile 6 21.7 ±0.89 (20.1-22.9) 5.0 I Females Adult 26 24.1 + 0.32 (22.5-25.1) 3.4 I Subadult 25 23.1 + 0.47 (21.1-25.0) 5.1 I Juvenile 9 20.2+ 0.48 (18.7-21.0) 3.5 I Rostral breadth Males Adult 10 10.3 + 0.23 (9.8-11.0) 3.5 I Subadult 15 9.6+ 0.20 (9.1-10.3) 4.1 I Juvenile 6 8.3 + 0.40 (7.6-8. 8) 5.9 I Females Adult 26 9.3+0.08 (9.0-9. 7) 2.1 I Subadult 25 8.9 + 0.16 (8.0-9.6) 4.6 I Juvenile 10 7.8 + 0.20 (7. 3-8.2) 4.0 I Interorbital constriction Males Subadult 10 6.1 + 0.11 (5. 7-6. 4) 3.5 1 Adult 15 6.1 + 0.10 (5.9-6. 3) 2.6 I I Juvenile 6 5.9+0.17 (5.6-6. 2) 3.5 1 Females Adult 26 6.1 + 0.07 (5. 7-6.4) 2.9 I Subadult 25 6.0+0.07 (5.7-6. 3) 2.9 I Juvenile 10 5.8+0.10 (5. 5-6.0) 2.8 I 254 Annals of Carnegie Museum vol. 46 Table 1. (Continued) Sex and age N Mean + 2 SE (Range) CV Result SS-STP class Breadth across maxillaries Males Adult 10 8.7 ± 0.23 (8. 2-9. 5) 4.2 I Subadult 15 8.1 + 0.18 (7. 7-8. 9) 4.2 I Juvenile 6 7.5 ± 0.30 (6. 9-7. 9) 4.9 I Females Adult 24 8.3 + 0.13 (7. 3-9.0) 3.9 I Subadult 25 7.9+0.13 (7. 3-8. 4) 4.3 I Juvenile 10 7.1 + 0.13 (6. 7-7.4) 3.0 I Geographic Variation univariate analysis: Table 2 gives the results of univariate com¬ parisons for males and females from our two geographic samples. In males, sample 2 averaged larger than sample 1 in all measurements except breadth across maxillaries. Sample 2 was significantly (.05 level) larger in total length and length of tail. In females, sample 2 averaged significantly larger than sample 1 in the following measurements: length of hind foot; palatofrontal depth; zygomatic breadth; rostral breadth. Females in sample 2 averaged larger than females in sample 1 in all measurements except interorbital constriction (averaged the same) and breadth across maxillaries (sample 1 larger). multivariate analyses: Individuals of Geomys tropicalis were subjected to multivariate analyses using the NT-SYS program to deter¬ mine if trends in geographic variation could be detected using this type of analysis. Males and females were analyzed separately. The cluster analysis of the distance matrix for females (Figure 2) Tevealed the individuals to be divided into two major subgroups sep¬ arated at 1.62. The upper cluster contains all individuals from Alta- mira (4), the one specimen from 1 mi. S Altamira, and seven specimens from 2.5 mi. SSE Altamira. The lower cluster contains two specimens from 2.4 mi. S Altamira and six specimens from 2.5 mi. SSE Altamira. Whatever the reason for the separation of females into these two clusters, it is apparently not the result of geographic variation. The cophenetic correlation coefficient for the phenogram is 63.1 percent. 1977 Tropical Pocket Gopher 255 Table 2. Geographic variation in external and cranial measurements between two sam¬ ples of Geo my s tropicalis. See text for key to sample numbers. Sex and N Mean + 2 SE (Range) CV Significance sample numbers Total length Males Sample 1 7 247.6 + 8.63 (231.0-264.0) 4.6 Sample 2 3 264.3 + 5.93 (260.0-270.0) 1.9 Females Sample 1 15 231.8+5.2 (212.0-250.0) 4.3 ns Sample 2 4 237.3 + 1.5 (235.0-238.0) 0.6 Length of tail Males Sample 1 7 77.3 + 2.92 (71.0-82.0) 5.0 * Sample 2 3 85.7+ 8.67 (78.0-93.0) 8.8 Females Sample 1 15 75.1 + 3.57 (61.0-85.0) 9.2 ns Sample 2 4 70.8+2.75 (68.0-74.0) 3.9 Length of hind foot Males Sample 1 7 31.7+ 1.36 (30.0-35.0) 5.7 ns Sample 2 3 32.7+ 1.76 (31.0-34.0) 4.7 Females Sample 1 15 29.4+0.55 (28.0-31.0) 3.6 * Sample 2 4 30.6 + 0.63 (30.0-31.5) 2.1 Greatest length of skull Males Sample 1 6 47.0+ 1.69 (44.8-50.5) 4.4 ns Sample 2 3 48.0+ 1.62 (46.6-49.4) 2.9 Females Sample 1 15 42.3+0.63 (40.1-43.9) 2.9 ns Sample 2 5 42.9+0.70 (41.9-43.8) 1.8 Condylobasal length Males Sample 1 7 45.5+ 1.60 (42.9-49.4) 4.6 ns Sample 2 3 46.7 + 1.50 (45.448.0) 2.8 256 Annals of Carnegie Museum vol. 46 Table 2. (Continued) Sex and N Mean + 2 SE (Range) CV Significance sample numbers Condylobasal length (Continued) Females Sample 1 16 41.1 + 0.58 (39.0-42.6) 2.8 ns Sample 2 5 41.7 + 0.55 (40.9-42.6) Basal length 1.5 Males Sample 1 7 42.8 + 1.45 (40.4-46.3) 4.5 ns Sample 2 3 44.3 + 1.55 (43.2-45.8) 3.0 Females Sample 1 16 38.4 + 0.53 (36.6-39.8) 2.7 ns Sample 2 5 39.1 + 0.48 (38.7-40.0) Palatal length 1.4 Males Sample 1 7 31.5 + 2.79 (28.0-39.1) 11.7 ns Sample 2 3 34.1 + 5.74 (30.2-39.7) 14.6 Females Sample 1 16 26.5 + 0.42 (25.0-27.7) 3.2 ns Sample 2 5 26.9 ± 0.49 (26.2-27.6) Palatofrontal depth 2.0 Males Sample 1 7 16.6 + 0.45 (15.8-17.8) 3.6 ns Sample 2 3 17.4 + 0.87 (16.6-18.1) 4.3 Females Sample 1 16 15.3 ± 0.21 (14.5-16.0) 2.8 * Sample 2 5 16.1 + 0.41 (15.3-16.5) 2.9 Length of nasals Males Sample 1 6 16.5 + 0.52 (15.9-17.3) 3.8 Sample 2 3 16.9 + 0.82 (16.1-17.5) 4.2 Females Sample 1 15 14.4 + 0.46 (12.8-16.3) 6.2 Sample 2 5 14.6 + 0.46 (14.1-15.4) 3.6 1977 Tropical Pocket Gopher 257 Table 2. (Continued) Sex and sample numbers N Mean + 2 SE (Range) cv Significance Diastema Males Sample 1 7 16.5 + 0.87 (15.2-18.4) 7.0 ns Sample 2 3 17.2+0.84 (16.7-18.0) 4.2 Females Sample 1 16 13.8+0.28 (12.7-14.4) 4.0 ns Sample 2 5 14.3+0.28 (13.8-14.6) Zygomatic breadth 2.2 Males Sample 1 7 29.7 ± 1.40 (28.1-32.9) 6.2 ns Sample 2 3 30.5 + 0.66 (30.0-31.1) 1.9 Females Sample 1 16 25.8 + 0.44 (24.4-27.6) 3.4 * Sample 2 5 26.6+ 0.13 (26.4-26.8) Squamosal breadth 0.6 Males Sample 1 7 20.7 + 0.71 (19.8-22.3) 4.6 ns Sample 2 3 21.0+ 0.72 (20.5-21.7) 3.0 Females Sample 1 16 19.1 + 0.28 (18.3-20.1) 2.9 ns Sample 2 5 19.6 + 0.37 (19.0-20.1) Mastoid breadth 2.1 Males Sample 1 7 26.8 ± 0.88 (25.5-28.9) 4.3 ns Sample 2 3 27.5 + 0.64 (26.9-27.9) 2.0 Females Sample 1 16 24.0+ 0.40 (22.7-25.1) 3.3 ns Sample 2 5 24.3 + 0.52 (23.7-25.0) Rostral breadth 2.4 Males Sample 1 7 10.2+0.30 (9.8-11.0) 3.9 ns Sample 2 3 10.4 + 0.35 (10.1-10.7) 2.9 258 Annals of Carnegie Museum vol. 46 Table 2. (Continued) Sex and N Mean + 2 SE (Range) CV Significance sample numbers Rostral breadth (Continued) Females Sample 1 16 9.2+ 0.90 (9.0-9. 6) 2.0 * Sample 2 5 9.5 ± 0.19 (9. 2-9. 7) Interorbital constriction 2.3 Males Sample 1 7 6.0 ± 0.12 (5.9-6. 3) 2.7 ns Sample 2 3 6.1 ± 0.18 (6.0-6.3) 2.5 Females Sample 1 16 6.1 + 0.10 (5. 7-6. 4) 3.3 ns Sample 2 5 6.1 ± 0.10 (6.0-6.3) 1.9 Breadth across maxillaries Males Sample 1 7 8.8+ 0.33 (8. 2-9.5) 5.0 ns Sample 2 3 8.6 ± 0.13 (8. 5-8. 7) 1.3 Females Sample 1 14 8.4+0.16 (7. 7-9.0) 3.7 ns Sample 2 5 8.3+ 0.10 (8. 2-8. 5) 1.3 Figure 2 shows the cluster analysis for the nine males tested. One male from 2.5 mi. SSE Altamira is widely separated (1.785) from all other specimens. The remaining eight specimens are separated into two clusters (separated at 1.418 level). The upper cluster contains three specimens from 2.4 mi. S Altamira and one from 10 mi. NW Tampico. The other cluster contains two specimens from 2.5 mi. SSE Altamira, one from 1 mi. S Altamira, and one from Altamira (holotype). There is no clear cut geographic pattern to this variation. Results of the principal component analyses are shown in Figure 3 and Table 3. The components account for the following amounts of the total variation: component I, males (54.7 percent), females (60.0 percent); component II, males (16.9 percent), females (9.5 percent); component III, males (11.1 percent), females (7.0 percent). Table 3 indicates the effect of measurements in the first three components. Overall size is 1977 Tropical Pocket Gopher 259 effecting the placement of individuals of both males and females on component I. In males, however, it should be noted that palatal length is negatively correlated with the other measurements and that length of the hind foot, breadth across maxillaries, and interorbital constriction have little effect in component I. Interorbital constriction had the small¬ est effect in component I for females. In component II for males, length of tail, length of the hind foot, interorbital constriction, and squamosal breadth are having the largest effects. The first three of these have negative values, whereas squamosal breadth has a positive value. Palatal length and breadth across maxillaries have the largest but opposite effects in component III for males. For females, interorbital constriction and rostral breadth have largest values for component II, and breadth across maxillaries for component III. 1.620 1.470 1.320 1.170 1.020 0.870 0.720 0.570 Fig. 2. Phenograms of Geomys tropicalis (males left, females right) computed from dis¬ tance matrices and clustered by unweighted pair-group method using arithmetic averages (UPGMA). Localities from which specimens originated are numbered as follows: 1, 2.5 mi. SSE Altamira; 2, 2.4 mi. S Altamira; 3, Altamira; 4, 1 mi. S Altamira; 5, 10 mi. NW Tam¬ pico. The cophenetic correlation coefficient for the phenogram for males is 0.737 and for the females, 0.631. 260 Annals of Carnegie Museum vol. 46 Examination of the upper portion of Figure 3 reveals that the three males from 2.4 mi. S Altamira are located at the left of the plot, whereas the three from 2.5 mi. SSE Altamira are located to the right. The holo- type from Altamira is positioned between a specimen from 1 mi. S Altamira and one from 2.5 mi. SSE Altamira in the right half of the plot. i i Fig. 3. Three-dimensional projections of individuals of Geomys tropicalis (upper, males; lower, females) onto the first three principal components based upon matrices of cor¬ relation among the 16 external and cranial measurements. Components I and II are indi¬ cated in the figure, and component III is represented by height. See Fig. 2 for key to localities of origin for specimens. Table 3. Factor matrix from correlation among 16 characters of Geomys tropicalis studied. 1977 Tropical Pocket Gopher 261 o Oh E o u o Oh £ o U o Oh 6 o U ON lO (N NO r- r- VO ro NO (N o M ONONNOON ddddddddddodd Oh a> ■a cs o ^ — x> M "O T3 ed cd a> o o >.— O « _ _ _ ed S .«* ed © m*; c C c <*_ « ° ° - =0 b H 6 O ed _ _ w g? 3 ed ed a* Dh Oh J Q N W ed d § S*S Rostral breadth 0.947 0.089 -0.008 0.476 0.525 -0.265 Interorbital constriction 0.342 -0.619 -0.064 0.320 0.726 0.429 Breadth across maxillaries 0.107 0.383 -0.796 0.603 -0.202 -0.589 262 Annals of Carnegie Museum vol. 46 The specimen from 10 mi. NW Tampico is located near the center of the plot with its position nearest a specimen from 2.5 mi. SSE Altamira. The females represented in the lower portion of Figure 3 show less of geographic pattern to their position on the plot. The 13 specimens from 2.5 mi. SSE Altamira cover nearly the entire range of variation in all components. The four specimens from Altamira are positioned near each other in the right half of the plot. The one specimen from 1 mi. S Alta¬ mira is at the far right of the plot, whereas the two from 2.4 mi. S. Altamira are located relatively near each other in the left portion of the plot. Discussion The above analyses reveal that Geomys tropicalis exhibits variation with age as would be expected in any mammal. The age classes that we recognized represent distinct maturational groups because in most measurements the three age classes from non-overlapping subsets. G. tropicalis shows a high degree of secondary sexual dimorphism. In 15 of the 16 measurements tested, males were significantly larger than females (interorbital constriction being the only exception). In four measurements (greatest length of skull, zygomatic breadth, mastoid breadth, and rostral breadth), there was no overlap between males and females. In two other measurements, basal length and diastema, males and females overlap at only one measurement (40.4 and 15.2, respec¬ tively). One of the most exciting findings of this study was that G. tropicalis does not exhibit reduced individual variation. Long (1968, 1969) studied coefficients of variation in a wide variety of mammals. Five of the mea¬ surements that he studied can be compared with our results (mean co¬ efficient of variation of mammals studied by Long, 1969, ± 1 SD, coeffi¬ cients of variation for male and female G. tropicalis, respectively): total length, 5.31 ± 1.96, 5.0, 4.3; greatest length of skull, 3.21 ±1.40, 3.9, 2.8; zygomatic breadth, 3.95 ± 1.34, 5.3, 3.8; mastoid breadth, 3.05 ± 1.24, 3.8, 3.4; interorbital constriction, 4.36 + 1.51, 2.6, 2.9. Long (1968) reported the following values for Thomomys talpoides and we have found the following values for G. personatus and G. arenarius (males followed by females), respectively: total length, 4.45, 6.1, 5.2, 11.1, 5.4; greatest length of skull, 3.10, 3.9, 2.9, 5.3, 2.6; zygomatic breadth, 5.88, 4.5, 3.0, 6.2, 4.9; mastoid breadth, — , 4.0, 3.9, 5.6, 3.2; interorbital con¬ striction, 5.85, 9.5, 3.8, 5.3, 5.3. Clearly, G. tropicalis does not exhibit significantly reduced individual variation in any of these measurements with the possible exception of interorbital constriction. It is clear from these data that the highly reduced variation found in this species at the genic and chromosomal levels was not found at the morphometric level. These results can be explained in two possible ways. First, G. tropicalis has the normal amount of individual variation at the 1977 Tropical Pocket Gopher 263 morphometric level but has reduced variation at the genic and chromo¬ somal levels. Therefore, whatever the evolutionary forces that caused the reduction of variation at the other two levels for which data are avail¬ able have not come into play at the morphometric level of variation. The other possibility is that the coefficient of variation is not sensitive enough to consistently detect lower levels of individual variation or other types of variation such as seasonal, ecological, or nutritional, are being grouped with individual variation. This could mean that, except in cases of uniquely homogeneous samples, coefficients of variation below 2.0 to 2.5 may not be expected. Based upon all currently available data, we believe that Alvarez (1963) was correct in considering Geomys tropicalis to be a distinct species. Our analysis of geographic variation in this species failed to reveal any significant trends. Therefore, we consider G. tropicalis to be a mono- typic species occupying a highly restricted habitat and geographic area in southeastern Tamaulipas, Mexico. The species should be treated as follows: Geomys tropicalis Goldman, 1915 Geomys personatus tropicalis Goldman, 1915:134. Geomys tropicalis, Alvarez, 1963:426. holotype: Adult male, skin and skull, USNM 92, 946, from Altamira, Tamaulipas; obtained on 18 April 1898 by E. A. Goldman, original No. 12,320. measurements of holotype: total length, 270; length of tail, 86; length of hind foot, 33; greatest length of skull, 48.1; condylobasal length, 46.6; basal length, 43.9; palatal length, 30.2; palatofrontal depth, 17.5; length of nasals 17.5; diastema, 16.7; zygomatic breadth, 30.3; mastoid breadth, 27.8; squamosal breadth, 21.7; rostral breadth, 10.4; interorbital constriction, 6.0; breadth across maxillaries, 8.5. distribution: Confined to extreme southeastern Tamaulipas in the vicinity of the towns of Altamira and Tampico. specimens examined (94): Tamaulipas: Altamira, 15 (USNM); 1 mi. S Altamira, 17 (KU); 2.5 mi. SSE Altamira, 51 (TTU); 2.5 mi. SW Altamira, 2 (TNHC); 2.4 mi. S Altamira, 6 (TTU); 10 mi. NW Tampico, 1 (KU); 1 mi. N Tampico, 2 (TTU). Acknowledgments We would like to thank the following curators for allowing us to examine specimens in their care (abbreviations used to identify specimens in text): Robert S. Hoffmann, Museum of Natural History, University of Kansas (KU); Robert J. Baker, The Museum, Texas Tech University (TTU); Robert F. Martin, Texas Natural History Collection, University of Texas at Austin (TNHC); Don E. Wilson and Clyde Jones, National Museum of Natural History (USNM). We are especially grateful to Robert J. Baker, who provided the financial support for the collection of the large series of G. tropicalis deposited at Texas Tech University. Analyses were performed on the IBM-370 computer in the Com¬ puter Center, Texas Tech University. 264 Annals of Carnegie Museum vol. 46 References Cited Alvarez, T. 1963. The Recent Mammals of Tamaulipas, Mexico. Univ. Kansas Publ., Mus. Nat. Hist., 14:363-473. Baker, R. J., and S. L. Williams. 1974. Geomys tropicalis. Mammalian Species, 35:1-4. Davis, B. L., S. L. Williams, and G. Lopez. 1971. Chromosomal studies of Geomys. J. Mamm., 52:617-620. Gabriel, K. R. 1964. A procedure for testing the homogeneity of all sets of means in analysis of variance. Biometrics, 20:459-477. Goldman, E. A. 1915. Five new mammals from Mexico, and Arizona. Proc. Biol. Soc. Washington, 28:133-137. Long, C. A. 1968. An analysis of patterns of variation in some representative Mammalia. Part I. A review of estimates of variability in selected measurements. Trans. Kansas Acad. Sci., 71:201-227. 1969. An analysis of patterns of variation in some representative Mammalia. Part II. Studies on the nature and correlation of measurements of variation. Pp. 289- 302, in Contributions in Mammalogy (J. K. Jones, Jr.), Misc. Publ. Mus. Nat. Hist., Univ. Kansas, 51:1-428. Selander, R. K., D. W. Kaufman, R. J. Baker, and S. L. Williams. 1975. Genic and chromosomal differentiation in pocket gophers of the Geomys bursarius group. Evolution, 28:557-564. SOKAL, R. R., AND F. J. ROHLF. 1969. Biometry: the principles and practice of statistics in biological research. W. H. Freeman and Co., San Francisco, xiii + 776pp. Williams S. L., and H. H. Genoways. 1975. Karyotype of Geomys pinetis (Mammalia: Geomyidae), with a discussion of the chromosomal relationships within the genus. Experientia, 31:1141-1142. Back issues of many Annals of Carnegie Museum articles are available, and a few early complete volumes and parts are listed at half price. Orders and inquiries should be addressed to: Publications Secretary, Carnegie Museum, 4400 Forbes Avenue, Pittsburgh, Pa. 15213. S'- Nfl- fMdj-oy