/' HARVARD UNIVERSITY Library of the Museum of Comparative Zoology / BREVIORA MUSEUM OF COMPARATIVE ZOOLOGY Harvard University Numbers 179-230 1963-1965 CAMBRIDGE, MASS., U.S.A. 1965 Edited By Nelda Wright CONTENTS BREVIORA Museum of Comparative Zoology Numbers 179-230 1963 No. 179, The holothurians of Clipperton Island in the eastern tropical Pacific. By Elisabeth Deichmann. 5 pp. January 16. No. 180. A new fresh-water amphipod crustacean from Oregon. By E. L. Bousfield. 6 pp. January 17. No. 181. Systematic notes on the land snails of the genus Tomo- cyclus (Cyclophoridae). By Fred G. Thompson. 11 pp., 1 pi. February 1. No. 182. Birds from Flores, Lesser Sunda Islands. By Raymond A. Paynter, Jr. 5 pp. February 1. No. 183. Australian carabid beetles XIII. Further notes on Agonini, and a genus of Licinini new to Australia. By P. J. Darlington, Jr. 10 pp. March 12. No. 184. Behavior as a taxonomic clue: Relationships of Lisso- nyderis (Chiroptera). By Barbara Lawrence and Alvin Novick. 16 pp. April 18. No. 185. Eleutherodactylus hedricki, a new species of frog from Puerto Rico (Salientia, Leptodactylidae). By Juan A. Rivero. 7 pp., 1 pi. April 18. No. 186. Notes on Hispaniolan herpetology. 8. The forms re- lated to Anolis hendersoni Cochran. By Ernest E. Williams. 13 pp. April 18. No. 187. The labyrinthodont dentition. By John Newland Chase. 13 pp. July 10. No. 188. Redescription of some cavernicolous pseudoscorpions (Arachnida, Chelonethida) in the collection of the Museum of Comparative Zoology. By William B. Muchmore. 16 pp. July 29. No. 189. Notes on amphisbaenids (Amphisbaenia; Reptilia). 10. Redescription and redefinition of Amphisbaena pericensis Noble from the mountains of northwestern Peru. By Carl Cans. 15 pp. August 29. No. 190. Characters and synonymies among the genera of ants. Part III. Some members of the tribe Ponerini (Ponerinae, Formicidae). By William L. Brown, Jr. 10 pp. September 30. No. 191. Three new species of M angora (Araneae, Argiopidae) from Central America. By Arthur M. Chickering. 11 pp. December 5. No. 192. A description of Dinopis longipes F. P.-Cambridge, 1902 (Araneae, Dinopidae). By Arthur M. Chicker- ing. 6 pp. December 5. No. 193. A Miocene toad from Colombia, South America. By Richard Estes and Richard Wassersug. 13 pp. December 5. No. 194. A new subspecies of Tropidophis greenwayi from the Caicos Bank. By Albert Schwartz. 6 pp. December 31. No. 195. Cayman Islands Tropidophis (Reptilia, Serpentes). By Richard Thomas. 8 pp. December 31. No. 196. A new genus and species of bathypelagic ophidioid fish from the western North Atlantic. By Daniel M. Cohen. 8 pp. December 31. No. 197. Anolis whitemani, new species from Hispaniola (Sauria, Iguanidae). By Ernest E. Williams. 8 pp. December 31. 1964 No. 198. Amphisbaena schmidti, a third species of the genus from Puerto Rico (Amphisbaenia: Reptilia). By Carl Cans. 11 pp. March 10. No. 199. A new subspecies of Varanus exanthematicus (Sauria, Varanidae). By R. F. Laurent. 9 pp. March 10. No. 200. An anguid lizard from the Leeward Islands. By Garth Underwood. 10 pp. April 10. No. 201. Food habits and young stages of North Atlantic Alepi- saiirus (Pisces, Iniomi). By Richard L. Hacdrich. 15 pp. April 10. No, 202. The blind snakes (Typhlops) of Haiti with descriptions of three new species. By Neil D. Richmond. 12 pp. April 10. A new capromyid rodent from the Quaternary of His- paniola. By Clayton E. Ray. 4 pp. April 10. The status of Pseudogekko shebae and observations on the geckos of the Solomon Islands. By Walter C. Brown. 8 pp. May 15. Redescription of Amyhishaena duhia Miiller (Amphis- baenia: Reptiha). By Carl Gans. 11pp. July 15. The aistopod amphibians surveyed. By Donald Baird. 17 pp. July 15. Notes on the horseshoe bats Hipposideros caffer, ruber and beatus. By Barbara Lawrence. 5 pp. July 28. Three new species of frogs (Leptodactylidae, Eleuthero- dactylus) from Hispaniola. By Albert Schwartz. 15 pp. November 12. A new skate, Raja cervigoni, from Venezuela and the Guianas. By Henry B. Bigelow and William C. Schroeder. 5 pp. November 12. The ants of the Florida Keys. By Edward 0. Wilson. 14 pp. November 12. A new species of the snake Leptotyphlops from Colombia- By Benjamin Shreve. 4 pp. December 21. A new species of freshwater gastropod mollusc of the genus Saulea from the Miocene of Kenya. By Thomas Pain and Dawn Beatty. 5 pp., 2 pis. December 30. 1965 No. 213. A hitherto overlooked Anodonta (Mollusca: Unionidae) from the Gulf drainage of Florida. By Richard I. Johnson. 7 pp., 2 pis. January 11. No. 214. A revised classification of the dendrochirote holo- thurians. By David L. Pawson and H. Barraclough Fell. 7 pp. February 15. No. 215. Two new subspecies of Amphisbaena (Amphisbaenia, Reptilia) from the Barahona Peninsula of Hispaniola. By Richard Thomas. 14 pp. February 15. No. 216. The geographical variation of the frog Hyperolius marmoratus (Family Hyperoliidae) in Rhodesia, Nyasaland and Tanganyika. By R. F. Laurent. 9 pp. February 15. No. 203. No. 204. No. 205. No. 206. No. 207. No. 208. No. 209. No. 210. No. 211. No. 212. No. 217. The auditory region of the borhyaenid marsupial Cladosictis. By Bryan Patterson. 9 pp. March 1. No. 218. New frogs of the genus Cornufer (Ranidae) from the Solomon Islands. By Walter C. Brown. 16 pp., 2 pis. May 7. No. 219. The early evolution of the Echinozoa. By H. Barra- clough Fell. 17 pp. May 7. No. 220. A new species of Eleutherodactylus from Guadeloupe, West Indies. By John D. Lynch. 7 pp. May 7. No. 221. New Melanesian ants of the genera Simopone and Amblyopone (Hymenoptera — Formicidae) of zoo- geographic significance. By Robert W. Taylor. 11 pp. May 7. No. 222. The genus Leptotyphlops in the West Indies with des- cription of a new species from Hispaniola (Serpentes, Leptotyphlopidae). By Richard Thomas. 12 pp. May 28. No. 223. A new subspecies of Clelia clelia (Serpentes: Colubridae) from the island of Grenada. By Allen E. Greer. 6 pp. May 28. No. 224. New species of land mollusks with notes on other species from the Solomon Islands. By William J. Clench. 8 pp., 2 pis. July 15. No. 225. The Asian species of Galeritula Strand (Coleoptera, Carabidae). By Hans Reichardt. 16 pp. July 15. No. 226. The larval form of the Heteromi (Pisces). By Giles W. Mead. 5 pp. July 15. No. 227. The species of Hispaniolan green anoles (Sauria, Iguanidae). By Ernest E. Williams. 16 pp. Septem- ber 10. No. 228. Relationships among Indo-Australian Zosteropidae (Aves). By Ernst Mayr. 6 pp. September 15. No. 229. The genus DarUngionia (Serpentes) in Hispaniola, in- cluding a new subspecies from the Dominican Re- public. By Albert Schwartz and Richard Thomas. 10 pp. September 15. No. 230. Notes on some non-passerine birds from eastern Ecua- dor. By David W. Norton. 11 pp. September 15. INDEX OF AUTHORS BREVIORA MUSEUM OF COMPARATIVE ZOOLOGY Numbers 179-230 1963-1965 No. Baird, Donald 206 Beatty, Dawn 212 BiGELOW, Henry B 209 BOUSFIELD, E. L 180 Brown, Walter C 204, 218 Brown, William L., Jr 190 Chase, John Newland 187 Chickering, Arthur M 191, 192 Clench, William J 224 Cohen, Daniel M 196 Darlington, P. J., Jr 183 Deichmann, Elisabeth 179 EsTES, Richard 193 Fell, H. Barraclough 214, 219 Gans, Carl . 189, 198, 205 Greer, Allen E 223 Haedrich, Richard L 201 Johnson, Richard 1 213 Laurent, R. F 199, 216 Lawrence, Barbara 184, 207 Lynch, John D 220 Mayr, Ernst 228 Mead, Giles W 226 MucHMORE, William B 188 Norton, David W 230 NovicK, Alvin 184 Pain, Thomas 212 Patterson, Bryan 217 Pawson, David L 214 Paynter, Raymond A., Jr 182 Ray, Clayton E 203 Reichardt, Hans 225 Richmond, Neil D 202 Rivero, Juan A 185 schroeder, william c 209 Schwartz, Albert 194, 208, 229 Shreve, Benjamin 211 Taylor, Robert W 221 Thomas, Richard 195, 215, 222, 229 Thompson, Fred G 181 Underwood, Garth 200 Wassersug, Richard 193 Williams, Ernest E 186, 197, 227 Wilson, Edward 0 210 BREVIORA Mmseiiioi of Comparative Zoology Cambridge, Mass. Jaxiary 16, 1963 Xtmber 179 THE HOLOTHURIANS OF CLIPPERTON ISLAND IN THE EASTERN TROPICAL PACIFIC By Elisabeth Deichmann Museum of Comparative Zoology Cambridge, Massachusetts In 1902 H. L. Clark listed two typical Iiido-West-Pacific aspidochirote holothurians from Clipperton Island, though at that time he considered one of them a new species. Later expe- ditions brought no holothurian material from this island. It was not until 1958 that Deichmann was able to include in her Hancock report some material of a third species, collected by the Scripps Institution's expedition a few years before. Finally, in 1958, the University of California Clipperton Expedition, which used more refined methods, with diving, etc., brought back four species, two of which had never been reported from any locality in the eastern tropical Pacific. Of the five species now known from Clipperton Island, only one, Microtliele difficilis (Semper) appears to be well established on the mainland of the Pacific coast of America. Two other spe- cies had previously been listed from some of the outlying islands, Semperothuria atra (Jaeger) and Mertensiothuria leucospilota (Brandt). Of the remaining two forms, Stichopus korrens Selenka represents a typical Hawaiian form, while Sempero- thuria flavomaculata (Semper) is considered rare; it was orig- inally described from Samoa and later listed from Tahiti, and (?)Batavia in the East Indies. The conclusions one can draw from this short list are that more Indo-West-Pacific forms are able to cross, at least inter- mittently, the ' ' Ekman Barrier ' ' than hitherto assumed, and that some of these forms may have escaped notice because they are living at a greater depth here than that in which they normally live in the more favorable localities of the Indo-West-Pacific. 2 BBEVIORA No. 179 STICHOPODIDAE 1. Stichopus horrens (Selenka) 1867 Stichopus horrens Selenka, 1867, p 316, pi. 18, figs. 27-29. H. L. Clark, 1922, p. 64, pi. 2, figs. 19-23. Stichopus godeffroyi Semper, 1868, p. 7.5, pi. 130, fig. 4; var. pygmaeus, p. 75, var. b, p. 246. Stichopus tropicalis Fisher, 1907, p. 676, pi. 70, figs. 1-11. The five specimens secured measure 15-20 cm. in length and agree well with Fisher's description of his material from Hawaii. All the specimens have numerous C-shaped bodies of all sizes in the skin ; this is the character which is supposed to separate tropicalis (= Semper 's var. b.) from the typical godeffroyi, but I fully agree with H. L. Clark in regarding the presence or absence of these structures as most unimportant. The typical tables — with a pointed spire — are present, although not in large numbers in some individuals, and likewise the peculiar rosettes could be located in all five individuals. The color of four of the specimens is indicated to be "mottled with orange and jiale cream," while the fifth is "mottled dull greenish," a color range which agrees well with H. L. Clark's observations on horrens. Fisher Avrites : ''dark olive green, mot- tled with deep brownish green." His material, 16 cm. long, in preserved condition, came from tide pools in Haw^aii. While most earlier records (from Samoa, Fiji and Hawaii) are from shallow water, the present material (Clipperton A- 588-6) came from 10-20 meters depth, off the edge of "the 10 fathoms terrace" on the northern side of the atoll: it was noted that the species did not occur at greater depth (40 meters) off the edge of that terrace. It was also observed, at similar depth, on the other side of the island, and the species appears therefore to be well established in this locality. HOLOTHURIIDAE 2. MiCROTHELE DiFFiciLis (Semper) Holothuria diffidlis Semper, 1868, p. 92, pi. 30, fig. 21. Microthele difficilis:—'Deichmann, 1958, p. 288, pi. 1, figs. 6-9 (list of ref- erences). Four dark brown individuals, 6-7 cm. long, were collected (Clipperton A-r)88-8) on the reef flats "from the sbore out, at least to the weak algal ridge." 1963 llOLOTIirRIANS OF CLII'PEKTON ISLAND 3 The pr('seiu'(> of this sjx'cies in Clipperton is not surprising, since this atoll was the type locality for IT. L. Clark's '"Holo- iliuria frcquentiamcnsis," described in 1902 on the basis of 17 individuals, about 4 cm. long-, of which one had immature geni- tal organs. The species ranges from the east coast of Africa to the Panamic region, where it is known from the Galapagos Islands, north- wards to Espiritu Santos Island in the Gulf of California. .■). Se:\iperotiiuria flavomaculata (Semper) nolotlturia flavomaculata Semper, 1868, pp. 87, 277, pi. 30, fig. 26. Pan- ning, 1934, pt. II, p. 42, text-fig. 35 (list of references). Scmperotliiiria flavomncuhiia : — Deichmann 1958, p. 303 (treated in the key for the genus). lentil the present study, there were only a few^ species belong- ing to the old genus llolothiiria of which the Museum of Com- parative Zoology did not possess representatives, and flavomacu- lata was one of these. When the key for the few^ members of the genus 8cmperothuria was worked out in 1958, it was necessary to rely entirely on the literature. It was consequently a satis- faction to discover two specimens of this comparatively rare species in the Clipperton material and to find that they agreed with the key. Both specimens came from Clipperton A-588-6, from the 10 fatlioms terrace, in w-ater varying from 1-20 meters, off the edge of the tidal flat. The species was not noted by divers off the edge of that terrace, in water up to 40 meters depth. The two individuals (M.C.Z. no. 3008) measure 8 and 12 cm. in length; they are dark, with few feet in scattered rows on the ventrum and more sparingly developed as papillae on the dor- sum, here with a yellow area around their base. The spicules agree completely with the earlier descriptions: a crowded layer of tables with no disk, but a base which tapers to a point, four pillars and the spire ending in a double Maltese cross. In the deeper layer are scattered short rods, covered by coarse, rough spines, or clusters of spines. The type, 11 cm. long, came from Samoa, and according to Panning the species has also been taken in Tahiti and (?)Bata- via. It may possil)ly. when contracted, have been mistaken for Ludwi(jothuria atra, or some of the other dark-skinned forms in the Indo-Pacific. It was at first assumed that the material 4 BREVIORA No. 179 represented atra, previously known from Clipperton Island, but the roughness of the skin eliminated that possibility, even before the spicules had been examined. 4. Mertensiothuria leucospilota (Brandt) Stichopus leucospilota Brandt, 1835, p. 51. Mertensiothuria leucospilota: — Deiclimann, 1958, p. 297, pi. 3, figs. 1-9 (list of references). Three large, reddish-brown individuals (Clipperton A-588-8) were collected on reef flats, identical in size and spiculation with those taken some years ago by the Scripps Clipperton Ex- pedition, and included in Deichmann's 1958 report. The species has the same wide range as Microthele difficilis, from the east coast of Africa to the Panamic region, but so far it has never been found established on the mainland. The Han- cock expeditions reported it from Galapagos, Clarion and Socorro islands. The largest individuals, 20 cm. long, in preserved con- dition, have all come from Clipperton Island. 5. LUDWIGOTHURIA ATRA (JaCgCr) Holothuriu atra Jaeger, 1833, p. 22. Panning, 1934, pt. II, p. 30, text-fig. 22 (list of literature). Ludwigothuria atra: — Deiclimann, 1958, p. 312, pi. 2, figs. 18-23. Of this species H. L. Clark reported nine specimens from Clipperton Island in 1902. The Hancock expeditions collected thirteen in the Galapagos Islands and one in Cocos Island. The species is known to range from Mozambique to Hawaii ; in the latter locality it is stated by Fisher to be "one of the common- est holothurians inhabiting Hawaiian shores." The fact that the recent expeditions to Clipperton Island did not collect this large and conspicuous species, in spite of the intensive collecting undertaken, may indicate that this is one of the species which only intermittently crosses the "Ekman Barrier" but is unable to become permanently established in the less favorable localities which it finds in the eastern Pacific. REFEEENCES Brandt, J. F. 1835. Ecliinoderniata ordo Holothurina. In Prodroaius descriptionis aninialium. . . . Petropoli, Fasc. 1, pp. 242-262. 1963 HOLOTIIURIANS OF CLIPPERTON ISLAND 5 Clabk, II. L. 1902. Papers from the Hopkins Stanford Galapagos Expedition. Vol.4, art. 12, Echinoderniata. Proc. Washington Acad. Science, vol. 4, pp. 521-531. 1922. The holothurians of the genus Stichopus. Bull. Mus. Comp. Zool., vol. 65, no. 3, pp. 39-74, pis. 1-2. Deighmann, E. 1958. The Holothurioidea collected by the Velero III and IV during the years 1932 to 1954. Part II. Aspidochirota. Allan Hancock Pacific Expeditions, vol. 11, no. 2, pp. 251-346, pis. 1-9. (Con- tains a number of new generic names for various members of the old genus Holothuria.) Fisher, W. K. 1907. The holothurians of the Hawaiian Islands. Proc. U. S. Na- tional Museum, vol. 32, pp. 637-744, pis. 66-82. Jaeger, G. F. 1833. De Holothuriis. Diss. Inaug. Turici (Torino, Italia), 40 pp., 3 pis. Panning, A. 1928- Die Gattung Eoloihuria. Parts I-V. Mitt. Staats Inst, und 1935. Zool. Museum Hamburg, vol. 44, pp. 91-138, 21 text-figs., 1 chart; vol. 45, pp. 24-50, 26 text-figs., 3 charts; ibid., pp. 65-84, 27 text-figs.; ibid., pp. 85-107, 32 text-figs.; vol. 46, pp. 1-18, 19 text-figs., and index to all five parts. Selenka, E. 1867. Beitrage zur Anatomic und Systematik der Holothurien. Zeit- schrift Wiss. Zool., vol. 17, pp. 291-374, pis. 17-20. Semper, C. 1868. Reisen im Archipel der Philippinen, II. Wissenschaftliche Re- sultate, vol. 1, Holothurien, 299 pp., 10 pis. ^^z ^U BREVIORA MmsemitTn of Comparative Zoology Cambuidge, Mass. January 17, 196;> Number 180 A NEW FRESH-WATER AMPHIPOD CRUSTACEAN FROM OREGON By E. L. BousFiELD National Museum of Canada In a recent account of the fresh-water amphipod fauna of Oregon (Bousfield, 1961), the writer listed ten known species including two species of Crangonyx, C. richmondensis occiden- talis Hubricht and Harrison and C. pseudogracilis Bousfield. Dur- ing a further examination of Crangonyx material from small alpine lakes in southwestern Oregon, a third species was identi- fied. It proved to be distinct from other North American species studied by the writer (Bousfield, 1958) and from other known species of the genus and is herewith described as Crangonyx alpinus sp. nov. The present study is part of a general survey of gammaridean amphipods in the collection of the Museum of Comparative Zoology, Harvard University, conducted in May, 1962, with the aid of a grant from the National Science Foundation. The writer wishes to thank Dr. Elisabeth Deichmann for instigating the study and for her many kindnesses and co-operation during the undertaking. In the list of specimens, MCZ refers to the Museum of Com- parative Zoology, and NMC to the National Museum of Canada. Family GAMMARIDAE Crangonyx alpinus n. sp. (Figs. 1, 2) ■ Material examined: A total of 52 specimens, collected in alpine lakes of Lane and Douglas counties, Oregon, by F. Ziesenheim during the summer of 1937. as follows : Corner L. (alt. 4800 ft.), Lane Co., July 29, MCZ No. 10027 — 3 fem., 12 juv. ; Ledge L. 2 BREVIORA No. 180 (5350 ft.). Lane Co., Aug. 5, MCZ No. 10028 — 1 ovig. fern. (TYPE), 1 jiiv.; Pork L. (4820 ft.). Lane Co.. Aug. 9, MCZ No. 10029 — 3 fern. imm. ; Plumb L.. Lane Co., Aug. 21, MCZ No. 10030 — 6 juv. : Opal L. (5480 ft.). Douglas Co., Sept. 7, MCZ No. 10031 — 8 imm.: Whig. L. (5270 ft.), Douglas Co. Sept. 24, MCZ No. 10032 — 2 imm.; Emma L. (5190 ft.). Lane Co., Sept. 25, MCZ No. 10033 — 6 fem. imm.. 4 male imm.; Easter Brook L. (5050 ft.), Lane Co.. Sept. 27, MCZ No. 10034 — 1 fem. (Br. 1), 3 juv.; Mud L. (4950 ft.). Lane Co., Sept. 29, NMC No. 10035 — 1 fem. imm. Diagnosis: A small species of the richtnonderisis group hav- ing elongate antennae, peraeopods and uropods, reduced mouth- parts, shallowly cleft telson, and powerful gnathopods. but distinguished by the acuminate abdominal side plates, sharply serrated ba.sal segments of peraeopods 3-5, singly inserted pos- terior marginal setae of segment 6 of gnathopod 1, long marginal spines of uropod 3 and telson, and by the small eye. Female (10.5 mm.). Eye small, black, irregularly oval, re- moved from anterior head margin. Antenna 1. flagellum of 25 segments; accessory flagellum shorter than 1st flagellar segment. Antenna 2. flagellum of 9 segments. Lower lip, inner lobes, distinct, rather broad. Mandil)ular palp, terminal segment slender, with 3 outer marginal setae. Maxilla 1, inner plate with only two plumose marginal setae. Maxilla 2, plates relatively narrow; inner plate with only one elongate plumose facial seta. Maxilliped, inner plate short, truncate apex with 5 slender spine-teeth ; outer plate small, outer margin convex ; dactyl of palp stout. Lower corners of coxal plates 1 and 2 rounded, each with 3-5 short marginal setae. Gnathopod 1, posterior margin of segment 2 with numerous slender setae, anterior margin nearly bare ; posterior margin of segment 5 with 3-4 clusters of long, slender, distally flexed spines, some minutely pectinate ; segment 6 (propodus) subquadrate, widest distally; palmar margin evenly convex, slightly oblique, armed with about 12 medium-small spine-teeth on each side, strongest and closely crowded near posterior angle ; posterior margin with 6-7 slender setae appear- ing singly inserted ; dactyl fairly heavy, closely fitting palm. Gnathopod 2, segment 2 distally broadening, margins with sev- eral long setae ; posterior margin of segment 5 with 5 posterior groups of setae ; segment 6 subrectangular, distally broadening ; 1963 NEW FRESH -WATER AMPIIIPOD 3 palm smoothly convex, oblique, lined with widely-spaced medium- strong spine-teeth ; posterior angle with one prominent spine and another smaller spine ; posterior margin nearly two-thirds the anterior ; inner face of propodus with 4 groups of superior lateral setae. 1-3 setae per cluster ; dactyl rather slender. Peraeopods 1 and 2 slender, subequal ; posterior margins of segments 4-6 moderately spinose. Coxal plate of peraeopod 2 nearly as broad as deep, proximally emarginate behind. Peraeo- pods 3-5 long and slender, 4th longest. Basal segments of peraeopods are similar in size and shape ; posterior margins gently convex, with about 7-9 rather deep serrations, distally sharpest ; length of dactyls y^ to ^ the propods. Brood plates moderately large ; setae numerous, elongate, minutely cleft at tip. Coxal gills present on segments 2-7, paired sternal gills on segment 6, two pairs on segment 7. Pleopods powerful, 1st strongest ; rami about twice the length of the peduncle, inner ramus somewhat longer than outer. Peduncles of pleopods each with a few simple marginal setae and 3-4 coupling spines of at least two types. Proximal setae of inner ramus are non-plumose and bifid at the tip. Abdominal side plates, posterior margin shallowly incurved distally, corners sharply acuminate and produced posteriorly, most strongly in side plate 2. Uropods 1 and 2 rather long and slender, lateral margins of both rami armed with short spines. Uropod 3, outer ramus slender, about twice the peduncle, lateral margins armed with about 4 groups of longish spines ; inner ramus with 1 sub-apical spine ; peduncle with 1 or 2 lateral marginal spines. Telson about as broad as long, shallowly emarginate, each lobe terminated by 2 longish spines. Remarks: Mature males were not present in material at hand. The principal breeding period is probably June and early July following which the adults die off. References BOUSFIELD, E. L. 1958. Fresh-water amphipod crustaceans of glaciated North America. Can. Field-Nat. 72(2) : 55-113. 1961. New Eecords of fresh-water amphipod crustaceans from Oregon. Nat. Mus. Can. Nat. Hist. Paper No. 12, 7 pp. BREVIORA No. 180 Abbreviations for the Figures A 1 — Antenna 1 A2 — Antenna 2 LL — lower lip Lit Md — Left mandible Rt Md —Right mandible Mxl — Maxilla 1 Mx2 — Maxilla 2 Mxpd — Maxilliped Gnl — Gnathopod 1 (jn2 — (Juathopod 2 I'l — peraeopod 1 P2 — peraeopod 2 P3- P4- P5- PLl peraeopod 3 peraeopod 4 peraeopod 5 — pleopod 1 PL3 — pleopod 3 Ul — uropod 1 U2 — uropod 2 U3 — uropod 3 T — telson EPl — abdominal side plate 1 EP2 — abdominal side plate 2 EPS — abdominal side plate 3 1963 NEW FKESII-WATER AMIMIIPOD Figure 1. Crangonyx alpinus n. sp. Ledge L., Lane Co., Oregon, August 5, 1937, female (TYPE) 10.5 mm. 6 BltEVIORA No. 180 Figure 2. Crangonyx alpinnfi ii. sp. Ledge L., Lane Co., Oi'egon, August 5, U);37, fomnlo (TYPE) 10.5 mm. BREVIORA MiasemtM of Coimpsirative Zoology (\\MBRinGi-:, Mass. Kkhkiaky 1, 1963 Numbkr ISl SYSTEMATIC NOTES ()X THE LAND SXAILS OK THE (JEXCS TOMOC'YCLrs (CVCLOIMIORIDAE) By Fkki) (i. Tiio.Mi'sox Depiiitiiieiit of Zoology University of Mi;mii The x, suture^ moderately im- pressed; whorls crossed by fine, ])oorly developed, ])osteriorly arched jirowth wriidvles. which lie about 0.5 nun. ajiart ; wrinkles faintly discernible, di. linctness and s])acin.o' of wrinkles stead- ily increasinp' on last whorl ; wrinkles continuin!])()sed by Morelet has l)een properly applied by subse((uent authors. Cyclostonia copanense Sowerby was described as being smaller than simulacrum. Later authors found that the two forms were not as distinct as Sowerby be- lieved, and Fischer and Crosse (1886: 121), von Martens (1890: 10) and Kolx'lt (lf)()2: 271) considered copaxcusc as a variet\" BREVIORA No. 181 of simulacruin. Recently, Bartseh and Morrison (1942: I-IT-MS) reallocated copancnse to specific status. Specimens under 27 mm. in total length were assigned to copancnse, and specimens over 35 nnn. in total length were assigned to siniuJacnon. As is demonstrated in the diagram (Text-figure 1) no signifi- cant diflPerence in size will distinguish the two forms. Different specimens in various lots are represented near both extremes 35 30 H X UJ X 25 . 20 • • • • • • • • • 10 WIDTH 15 Text-figure 1 Tliis (liagr.'iiii ilenionstrntes the relationsship between height nnd widtli of the shell of Tinnuciiclus .siiiiiihicrioii (Morelet). Meiisui-enieuts of heiglit were iiiiide to iiiclufle tlie l;ist t'oui- reinaiiiing whoi-ls. .Mc'isnri'iiieuts of widtli are slaiidaid. .Ml iiicasiiit'iiu'iils are in millimeters. 1968 NOTES 0.\ TU.MOCVCLUS 7 of the same <;rai)h, and the rcinaiiiinji' spcciincns in these same lots are re])resented at many intermediate loei. Since siicli vari- ability of size eommonly oeeiirs within a sin«il(> lot, copnnensc cannot be recognized even as a variety of siniulaci-Kiii. Migaloinasioma siiniilacniiii var. iniittts was [)roi)osed by von Martens as a synonym of copatirsi . This is clear by his nse of the two names. Mcgalo))Kist()ma siiniildcfion var. (jfdcilis is also in- distinguishable from simulacnDii, for tlie slij>ht ditlterence of shape used to separate the two forms is hiohly variable and lots containing' only a few sj^ecimens cannot be satisfactorily sorted into two groups by use of this character. Tomocychis siplto)iis Bartseh and Morrison was distinguished from sitnuIacrKiit by the presence of a closed pseudosiphon as op posed to an open one. Comparison of tiie type o\' sipJuniis (TSXIM 162511) with many specimens of sinuilacnon shows that this character varies from a completely enclosed hole connected to the outside by a narrow slit, to a broad open crescent. The dis- tinction between a closed and an open pseudosiphon is obscured by many intermediate stages, thus preventing the differentiation l)etween sipJionis and simulacrun}. Toniocyclus co7ist7-ictits Bartseh and Morrison was recognized on the basis of a closed pseudosiphon and a relatively deeply impressed suture. As shown above, the natui-e of the ]isendo- siphon is an unreliable charactei-. The degree of impression of the suture is also variable. Several lots examined contain speci- mens that completely bridge the diflFerence from a deep suture with strongly rounded whorls, to a shallow suture with slightly convex whorls. Occasional sp(\-imens have a deep suture between the early whorls, and a shallow sutui-e between the lat, pis. 1-44. 10 BREVIORA No. 181 Plate I II()loty])e of Tonmeiichis tistulo.sii.s', new ^spe^'ies (UM^rZ 194095) ; liigli lainfoiest nt Valleiitine Camp, 50 miles Southwest of Cayo, British ITon- iluias. Fig. 1. Frontal view. Fig. 2. Lateral view. Fig. 3. Basal view. lOO.'S NOTES OX TOMOCYCLrS 11 C)'0<-^ BREVIORA Miiseiim of Comparative Zoology Cambridge, Mass. I'EBRrARY 1, 1963 Number 182 BIRDS FRO.M FLORES. LESSER Sl'NDA ISLANDS Bv RAY.MONTn A. Paynter, Jr. During recent years tlie Musenni of Comparative Zoolooy has received from the Rev. J. A. J. \'erheijen, S.V.D., several collec- tions of birds from the western part of Flores Island. Without using a gun, Father \'erheijen and his local assistants have se- cured representatives of 70 species. Eleven of these have not been previously recorded from the island and several others are of taxonomic importance. It is the purpose of this paper to note briefly these more Intercast ing species. Nycticorax caledonicus subsp. A bird in partial adult plumage was snared at Tjantjar, Ra- hong, on 19 May 1957. LTntil 1940, when an example of N. cale- donicus was found paired with one of N. nycticorax in western Java (Hoogerwerf, 1952), this night-heron was not known west of Timor. This is the only record from Flores. Dendrocygna javanica A wing has been preserved of a bird collected in 1956. Sum- bawa is the farthest east it had been recorded previously. Elanus caeeuleus subsp. The easternmost localities for this hawk were Sumba and Celebes. A nestling, just about to fledge, was taken near Ruteng, at about 1000 meters, in July, 1957. Rallus pectoralis exsul Three birds, the first known since the type specimen was se- cured, were collected in 1958 and 1959. One is an adult; un- fortunately it was not sexed. The other two are juvenals which 2 BREVIORA No. 182 have little barring on the ventrum, lack the rufescent head, and have no bright olive margins to the feathers of the back. T have examined two specimens of Rallus mirificus Parkes and Amadon (1959) from Luzon, Philippine Islands, and believe that this form is a race of R. pecioralis. From Australia through New Guinea to Flores, R. pcctoralis exhibits a fairly orderly cline of decreasing dorsal streaking, darkening head color, and shortening of the dorsal feathers. The characters of mi7-ificus are the culmination of these trends and are those one might have predicted at the end of a cline extending from Australia to the Philippines. The differences between the Philippine bird and the Australian races are marked, but the subspecies from Flores is almost perfectly intermediate. Additional races may well be discovered between Luzon and Flores, showing that the cline is less disjunctive than it now appears. Support for this prediction may be found in the knowledge that R. p. mirificus remained un- known until only four years ago, in spite of its presence within 50 miles of Manila, long a center of ornithological research. PORZANA PUSILLA PUSILLA Three specimens were collected : two at Tjara in late April and mid-May, 1957, and one at Wangjung in late April, 1956. One is a male, and two were not sexed, but on plumage characters these seem to be a male and a female. The species apparently breeds on Flores, although there are no prior records of its presence. It may be that these specimens represent an undescribed race. However, our birds are too poorly preserved and few in number to be certain that the apparent racial characters are real. The specimens are close to nominate pusilla but are somewhat darker dorsally, with more extensive black centers to the feathers, par- ticularly on the tail, and have heavier, but not longer, bills. The pale area on the chin and upper throat of the males seems whiter and more sharply demarcated than in the nominate form. The bills of two birds are black with very narrow yellowish markings on the anterior edges of both the maxilla and mandible. The third specimen is similar but also has a small pale area near the tip of the mandible. In a series of 26 s])ecimens of P. p. pusillo one has a bill similarly colored to the third Flores bird, and none of the series resembles the other two birds. The dried legs of the P'lores specimens are considerably dai-ker tiian any of P. p. pusilla. 1963 BIRDS FROM FLORES 3 The three birds from Flores are easily distinguished from mayri of New (iuinea and from palustris of Australia by their larger size (wing 84-86 mm.). T have not seen an exami)le of inird, from East Borneo, but the original description (Riley, 1938) notes that the bill is olive-yellow and bronze, which seems vastly different from the Flores birds. POLIOLIMNAS CINEREUS CINEREUS This widespread rail is new for Flores. Gallicrex cinerea A specimen was collected at Tjantjar on 14 May 1957. The species apparently has not been found before on the Malay Archipelago east of Java, where it seems to be a winter visitor (Kuroda. 1936). Its presence on Flores in mid-May suggests that it may breed there but, unfortunately, the gonads of our specimen were not examined. Gallinula chloropus orientalis Nine specimens were collected, one of which contained an egg in mid-May. No prior records from Flores exist. KOSTRATULA BEN'GIIALENSIS BENGHALENSIS This is another bird not reported from the island before and which represents an eastward range extension. A specimen col- lected in July was breeding. Gallinago stenura Three specimens establish this snipe as a winter visitor on Flores. Gallinago megala A long series, collected over several winters, represents a new record for the island. Otus (bakkamoena) silvicola This endemic owl is probably a giant geographical representa- tive of 0. hakkarnodia. It is similar to 0. b. semitorquts of Japan, but it is considerably larger and lacks the buff crescent on the upper back. This was noted by Hartert (1897), who also 4 BKEVIORA No. 182 remarked on the similarity of silvicola to 0. b. whiteheadi of Luzon. The Philippine form has a white hindneck and is darker below than silvicola, but in size more nearly approaches the Plores bird than any other race of 0. bakkamoena. Compared to the o'eographieally closer Javan and Bornean populations, silvi- cola is markedly larger. This is suggestive of the sitviation in the Philippines, where whiteheadi is considerably larger than any nearby races. Tyto capensis subsp. A specimen obtained in March, 1956. is the tirst record of the species from the Suncla Islands, thus confirming Hartert's sug- gestion (1929) that it might some day be found there. I am following Amadon and Jewett (1946) in treating longi- memhris and capensis as conspecific. In view of the variability of the species and the paucity of material, particularly from the southwest Pacific, I am reluctant to assign this specimen to any race. Amadon (1959) tentatively placed two individuals from Celebes and Kalidupa with T. c. ivallcri (type locality Queensland); the Flores bird may also belong here. CORACINA DOHERTYI An immature bird was collected in early March. Rensch (1931) recorded several specimens from the island, but Flores was inadvertently omitted from the range of the species as given in Volume 9 of the "Check-list of Birds of the World" (Mayr and Greenway, 1960) . ACKNOWLEDGEMENTS 1 am grateful to James C. Greenway, Jr. and Ernst Mayr for assistance' with several problems, and to Father Verheijen for depositing his collection in this museum. Dean Amadon gener- ously lent specimens from the collections of the American Mu- seum of Natural llistor}-. LITERATURE CITED Amadon, 1). IStf)'). Remarks on the subspecies of the Grass Owl, I'l/lo eapensLs. .Jouni. liombav Nat. Hist. Sue. 56: ;U4-:54(i. 1963 BIRDS FROM FLORES 5 Amadon, D. and S. G. Jewktt, Jr. 1946. Notes on Philippine birds. Auk, 63 : 541-559. Habtert, E. 1897. On the birds collected by Mr. Everett in Soiih Flores. Novit. Zool., 4: 513-528. 1929. On various forms of the genus Tyto. Novit. Zool., 35: 93-104. HOOQERWERF, A. 1952. Voorkomeu van Nycticorax caledonicus subsp. in W.- Java. Limosa, 25: 29-32. KURODA, N. 1936. Birds of the Island of Java. Vol. 2. Tokyo, vi + pp. 371-794. Mayr, E. and J. 0. Grbenvtay, Jr., editors 1960. Check-list of birds of the world. Vol. 9. Cambridge, Mass. Museum of Comparative Zoology, xii + 506 pp. Parkes, K. C. and D. Amadon 1959. A new species of rail from the Philippine Islands. Wilson Bull., 71: 303-306. Peters, J. L. 1934. Check-list of birds of the world. Vol. 2. Cambridge, Mass. Harvard Univ. Press, xvii + *01 PP- Rensch, B. 1931. Die Vogelwelt von Lombok, Sumbawa und Flores. Mitt. Zool. Mus. Berlin, 17: 451-637. Riley, J. H. 1938. Three new birds from Banka and Borneo. Proc. Biol. Sot. Washington, 51 : 95-96. vu€ of the branch in this position using all four limbs. Thus, wild Rousettus use their wrists for walking and their thumbs for climbing, roosting, and moving along the underside of branches or fruits. Captive Rouseftus would roost in all of the positions desr-ribed above. When they were exposed to light, they would fre(ni(Mitly crawl into a corner of the cage floor or into their feeding disli where they would ])r('ss tliciuselves as closely as possible against theii- sui-roundings, avoiding th(> light on Iheir eyes. Crawling is a not uncommon mode of locomotion for Rousettus. whcthei' 1[)68 RELATIOXSIIIPS OK LISSON YCTERIS 3 rcstrictod to a cajic oi' allowed to move freely about a vooiii. While feedinji', foi- example, tliey would crawl between the liands of a stem of bananas. In addition, they oeeasionally used their wrists and tliumbs as pushers to reailjust morsels of fruit in their mouths and R. avgyptiacus once was seen using the claws of one hind foot to manipulate the food in its mouth. Lissonyctcris never use their win^s for locomotion other than flight. Tn the wild, and in captivity in a large flight room, IAssly in their orienta- tion. Lisson!irt( ris oi-ient entirely by vision (Novielv, 1958). The captive hat was helph'ss in the dark and refused to fly in the dark oi- when hlindfoUh'd. When it was forced to fly witliout vision. I)y Ijein""' tlirown into the air. it always crashed into the first ol)staele it met. Roiisdfus anip1( .ricanddl IIS, R. sciniiiiidiis, and R. argyptiacus all orient visually and acoustically (Mohres and Kulzer, 1956; Kulzer, 1956, 1958; Novick, 1958; Griffin, Noviek, and Korn- tield, 1958). The sin^ile R. acgyptiacus ol)served for over nine months in captivity oriented largely visually in strong light but oriented acoustically in the dark, in dim light, when avoiding intricate obstacles, and when landing or taking off. Such acoustic orientation obviously demands specializations for sound ])roduc- tion, emission, reception, and interpretation. The brain, in particular, must be highly specialized for handling acoustic information. In the pteropids, acoustic orientation has so far been found oidy in Rousettus, scusii stricto, and not in Eidolon, Pteropus, Cynoptcrus, PtcHocJiiriis, Eonycteris, Macroglossus, or Lissonyc- teris (Novick, 1958). The orientation of Ste^wnycteris, a sub- genus of Roiiscttiis, has not been observed. The distinctness of Rouscitus, not only from Lissonycteris but from a fair sample of other pteropids as well, is clearly establisiied by its acoustic orientation and the behavioral, physiological, and anatomical features associated with it. Summary To summarize, Lissoiiycfi ris differ from Rouscitus, scnsu stricto, in their roosting posture, in their non-flight locomotion, and in their iiuU)ility to orient acoustically. Comparing the roosting and locomotory behavior of those bats observed alive, we find that Rouscttus resembles Eidolon, Pteropus, and Eonyc- teris, while Lissonycteris resembles Cynopterus most closely. Brief ol)servati()n of live epomophorine bats and examination of pictures and preserved specimens of epomophorine bats and of Myonycteris suggest that in behavioi- Lisso7iycteris resembles these bats as well. 6 BREVIORA No. 184 MORlMIOLOdiCAL C0.MPAK180NS The behavioral differenees between tlie live bats are reinforced by less speetaeular, l)iit equally definite, characters of the more conventional taxononiic sort. These make it clear that Lis- soin/rteris is far closer to Myonycteris, a supposed intermediate genus between the rousettine and cynopterine "roups, than to Rousettus. They also show that Lissonycteris and Myonycteris form a natural gi-oup probably intermediate between the rouset- tine and epomophorine groups, and one whose resemblances to the cynopterines are more apparent than real. The rousettine and cynopterine sections of the Pteropinae are chiefly distinguished by the following characters possessed by the latter and not by the former: The rostrum is shortened and the facial axis is not deflected, Ms and usually M- are lost, the eyes are larger, and there is a tendency to form tubular nostrils. The differences in the ridges of the soft palate described by Andersen (1912, pp. Ix, 485, 591, figs. 29A, 50) are also characteristic. On the basis, largely, of these characters, Andersen points out that Myonycteris is somewhat intermediate between the cynop- terine and rousettine sections. He says (1912, pp. Ivi-lvii) : "Myonycteris . . . has in many respects remained on the Rouset- tine level of development, while in others it exhibits modifications approaching those of Cynopterus. The general appearance, the dental formula, and the palate ridges are qviite or nearly as in Botisettus, but the rostrum is conspicuously shortened, the facial axis less deflected, mn and m- (last lower and upper molar) reduced almost to rudiments, the orbits larger, the nostrils more prominent and the calcar weaker," and later (p. Ixi) : "The fact is that this genus has retained many characters of Rousettus, while in jn-actically all the features in which it differs from Rousdtus it more or less closely approaches to Cynopterus. Whether a genus exhibiting characters of this description ought, in a linear arrangement, to be classed at or near the end of the Rousettine section oi' as the ■opening' genus of the Cynopterine section, must necessarily remain a matter of opinion." The resemblance of Lisso)iycieris to Myonycteris in all the above traits, except the shortening of the rosti'uiii and i-eduction of the dentition, was not noted by Andersen when he classified tile foniici-asan ahcrrjint form related to but {)ossibly generically distinct from /tth of tlie anterior part of the skull (measured Irom behind the postorbital processes to th(^ tips of the premaxil- laries) as compared with both leniith of the brain ease (measured from behind the postorlatal jiroeesses) and its bulk. Typically, also, the bi-aiii case in M i/oiii/ch rIs is elon-. In all three of these genera, the lateral i)air of sinuses end in front of the postorbital foramina. In Cynopterns, on the other hand, the well inflated lateral ]iair extend at least to the le\('l of the postorbital foramina which are thus crowded outwai'ds into the postorbital ])rocesses, and the much smallci'. medial pair show scarcely any inflation. Dentition In past classifications, the i-eduction of the last molai's has been one of the most im])oi-tant reasons for grouping Myonyctcris with Cyni)pt( riis. while the numl)er of teeth has been used to relate the former to R()iis(ftiis. Actually, the shape of the an- terior teeth is fai' more important than eitlu'i- of these in showing generic relationships and is one of the most clearcut characters relating Lissoiiyrtt ris and M yonych ris. \'ai-iation at the end of the molar row is, on the other hand, an impoi'tant differential charactei- separatiinj these latter two. Ty|)ically. the cyno])terine 1963 RELATIONSHIPS OF LISSONYCTERIS 11 bats have lost tlic last u|)|)('r and lower molars, while these teeth are retained and well di^veloped in the rousettine j^roup. Both Lissonycteris and M i/oinirtf ris have the rousettine formula hut, in the latter, the last molars are tiny and elearly obsolescent, while in tiie former they are tlie oi)posite. Reduetion of the molars in M i/nuifctcris is no moi-e r-eason than their stronger development in Lissoiitfcfcri.s for jj^rouping- the one with Cy- tiopterus and the other with Houscttus. The structure of the teeth, their spacing, and the occlusion pattern are remarlval)ly similar in Mijonycteris and Lissonycteris and differ equally from both Cynopierus and Rousettus. The most striking feature of the tooth rows of Lissonycteris and Myonycteris is the wide si)aeing of the short, broad teeth; in addition, the anterior teeth have the main external cusps placed well back on the outer margin so that in i)rotile the teeth look more equilaterally triangular. The teeth in Cynopterus differ sharply in being larger and more crowded, and in the arrange- ment of the cusps, with the main external cusp more anterior and a deeper sulcus separating the inner and outer halves of the teeth. In the upper jaw of Cynopterus, F'-' has a well marked ridge or internal cusp, whereas in Lissonycteris and Myonycteris this tooth, though thick, is sim])le. In the lower jaw of Cynop- terus, Pj is relatively large and P- compared to P4 is conspicu- ously higher, almost as broad, but shorter with a less well de- veloped heel. In Lissonycteris and Myonycteris P^ is tiny and F^ reduced, much narrower and only slightly higher than P4, wdth no trace of a heel ; P4 is relatively large with a high, almost central, main cusp, bilobed in Lissonycteris and in Myonycteris with sometimes a suggestion of a heel. In genei-al, these pre- molars are more similar to those of Rousettus ; in the latter. P-''. Po, and P4 are simple, high, and differ from Lissonycteris and Myonycteris chiefly in having the single cusp set farther forward with a trace of a heel posteriorly. P"* has the more usual pteropid elongated shape with the main cusps forward and the inner margin sometimes bilobed. In the lower jaw, neither of the first two premolars of Rousdtus is reduced, as in the former genera, while the molars are longei' and narrower. The wdde spacing and simjile shape of the teeth in Ljissonycteris and Myonycteris relate to a very characteristic occlusion pattern in which the P;{ and P-' alternate with each other, P4 barely occludes with the posterior edge of P"\ and there is scarcely any more contact between P"^ and the posterior edge of Pj. In Rouset- tus, the overlapping of the teeth is more pronounced, correlated 12 BREVIORA No. 184 with the ))('- nycttns of those teeth which ai'e lacking in the foi'nier. Ex- ternally, the long-fingered wiiiii, in which the metacarpal and V)G'] RELATIONS! I IPS Ol' MSSON •^(•TI•:HIS 13 first phalanx of (li<>'it five is a little loiij-cr fliaii the forearm in Kpomops, and the attaclmient of the \vinhly Aariahle from species to sjx'eies. in one form at least of Epomops the di- vided fourth and undivided, straight first three rid<>'es are not unlike those in Lissotijich vis. The epomophorine affinities of Liss()ii!icf< ris and M i/oin/ctt ris ai'e further indicated by Benedict (IDf)?. p. L'!)2) : "Strikingly, the scale form of //. diif/oh )isis [Lissonifcicris aiKjolcnsis] is more similar to that of the Epomopliorus section than to the Roiisfffiis section to which AndersiMi assigns it." and "The scales of Mfjoiiycfcris are virtnally indistinguishable from those of the Epomophorns section. Analyzing the ditl^'erences l)et^\■een the ei)omo])horines and myonycterines. we fiml that the most conspicnoiis ones iisnally are found in the regions of the skull where, Avithin the epomo- l)horine group, there is nnich generic variation, or that these differences are more extreme developments of traits which Lisaonyctcris and M]ion}ict< ris already show. This suggests that the epomophorine is a specialized branch of the myonycterine group which in turn is fairly close to an early Eousitins-M^e stock. These annectant forms, the myonycterines, help to confirm Andersen's (1912, liii-lvi) supposition that the origin of the epomo})horine group was from a primitive and Bonscttus-Y\k(' stock. The cynopterine bats are clearly (juite different. Resem- blances between these and Myonycfcris such as tooth formula, shortening of the rostrum, slight elongation of the back of the skull, reduction of the tail, and d(»vel()])ment of the Aving as well as roosting and locomotory behavior are probably parallel developments in groups Avhich diverged early. Andersen also supposes that the cynopterine bats arose from early RoKsrftus- like ancestors, but. with M yonyctcris eliminated as an inter- mediate form, the two groups are more sharply distinct than was previously su]iposed. Geograi)hically, the j)icture pi-esented by the morphological evidence is a very logical one. The cynopterine bats become exclusively an Oriental and Austro-Malayan group, the epomo- phorines are restricted to Africa where they center in the westei-n forc^stcnl I'egions, and in this same g(Mieral area we find the rather more primitive but fairly closely related myonycterine 14 BREVIORA No. 184 group. Common to l)oth regions is the rousettine group con- sidered, on the basis of morphological characters, to be the most primitive and therefore })rol)ably the most direct descendant of the ancestral stock from which all arose. The occurrence of a highly developed s(niar system in the otherwise not particularly specialized RoKSfftus suggests, however, that they are farther removed from the ancestral stock than was previously supposed. ACKNOWLEDGMENTS This work was supported in jiart by the Office of Naval Research^, the United States Public Health Service, the Sigma Xi-RESA Fund, the Belgian American Educational Foundation, and Harvard T^niversity. During jiart of this period, Novick was a Fellow of the National Institute^ of Neurological Diseases and Blindness. For help in capturing and keeping these bats, we are grateful to the personnel of the Sangley Point Naval Air Station and Clark Air Base in the Philippines, to the Naval Attaches and other personnel of the American Embassies in the Philijipines and Ceylon, to Mr. Pablo Tanciojo and the Negritos of the Zambales Mountains of Luzon, to Capt. J. A. Simon, T^SAF, to Major A. Weinman of the Dehiwela Zoo, Ceylon, and to ]\Ir. and Mrs. W. W. A. Phillips formerly of Namunukula, Ceylon, to Dr. Louis van den Berghe. Messrs. J. Moureau, P. Pirlot, and (i. Marlier of the histitiff pour la Recherche 8cietififi(/U( ( n Afri(j}i( ('eufrnh . T^wiro, Belgian Congo, and to countless others who made these studies possible by giving of their time or experience. We are most deeply in- debted to Dr. D. R. Griffin of Harvard University for help of every description. LITERATl^BE CITED Allen, J. A., H. T.ans, and J. P. riiapin 1917. The American Museum Congo Exjieilition colleition of bats. Bull. Amer. Mus. Xat. Hist., 2^^Am-:^^^?,, pis. XLIV-LV, figs. 1-26. Andersen, K. 1912. Catalogue of tlic Cliiro])tera in the collection of the British Museum, ^'ol. 1. Megachiroi)tera. London, British Museum: i-ei, 1-854, figs. 1-79. Benedict, F. A. 1957. Flair structure as a generic (diaracter in Iiats. Uuiv. ('alif.. Paid. Zool., 59 (8):285-548, pis. -lA-?,'!, Mgs. 1-4. 1 Rpitroflnctiiui of this jinpcr in wlioli' or in pnrt is iifrniittfil for .my i)nrposc of the ITnitt'd Stati-s (Jovcrnnicnt. 1963 RELATIOXSIITI'S OF USSONYCTKIUS 15 Grittiii, I). 1{., A. Xovi.k, .-iiid M. Kornlii'M 1958. Till' sensitivitv of ccliohM-jitidii in the I'niit li;it, RousetUis. Hiol. I'.iill., 115:1, niid 4. Lismui i/cl cris a inioh iisis li.-iinlliii^' iiicccs (if li;iii;iii,i. The use (if tlic fddt ill Ji.-iiiijiiiii;' fddii .■iml tlic piistnri' of tlic wiii^is wiiiic doing so is t-lciirly siiown. Figure ~). Li.s.son!jcteri.'< (tiifiolnisis resting, Imiigiiig on ;i (■;iii\;is \v;ill .-it its junction witli tlie ceiling. Xote tiie position of the feet wiiicli c;in hold onto ;i I'oost equally well in this or the oiijiosite orientation. The wings are being held well o\'erla|iiieii, sulistnnt ially enwiapping tlie body. If the bat had not liecn disturbed by the photographer, its chiii would luive heen tucked against its chest and its eyes closed. BREVIORA Meseemi of Coimiparsitive Zoology Cambridge, Mass. April IN. l<)(;;i NlIMHER 185 ELEUTHERODACTVLUS IIEDinCKh A NEW SPP]CIES OF FROG FKO.M PFFHTO KK^O (SALIENTIA, LEPTODACTYLIDAE) By Juan A. Rivero Department of Biology University of Puerto Rieo Mayngiiez, Puerto Rico The finding of a new spet'ies of frog in Puerto Rieo has seemed very iniprobalile. The island has been thoroughly searched by a ininiber of herpetologists, and since the calls of the described forms are well known, the recognition of a new voice is not diffi- cult. Yet the author has traveled miles and miles in the island, at night, and no unfamiliar voice was ever heard in recent years. He was almost convinced that a new species was out of the ques- tion, but it appears that this conclusion may be ])remature in almost any country, no matter how well the fauna of the ])lace is known. The reason is clear: there is a tendency to look for frogs in places where the known species occur, and heiu'c many potential habitats remain uninvestigated. Eleutherodactylus hedricki was heard wliih' collecting E. karl- schmidti in a mountain stream at El Yuncjue (El Verde). It was immediately evident that the voice heard was that of a new" species, and the animal was desperately hunted, until at last col- lected. It is interesting to note that this frog was recognized as new before being seen — a proof of the usefulness of voice, especially Avhere the fauna is well known, in the collecting of frogs new to science. The name hedricki has been given in honor of Hedrick J. Rivero. aged 9, who has declared himself assistant to his father and who follows him through creeks, caves and mountains dur- ing any time of the day or night. The author also wishes to express his appreciation to his wife, Eneida, and to his son, i BREVIORA No. 185 Jiiaii Jr., 16, who always join him and Hedriek in th(> search for frogs. Iloraeio Mayorga, research assistant to tlie author, has also been (wtreniely useful, not only in field work, but also in the tedious curatorial and laboratory work that usually fol- lows. Dr. John Randall kindly took tlie photouraplis of E. hedricki, and the Galiiiaues family of San Juan made their El Verde house available as a (MMiter of activities for the colleetiuff party. To all these jx'opic, the autlioi- feels (iee])ly indebted. Elet'tiierodactylus iiKnRrcKi s]). n. Type. Museum of Coniparative Zoology Xo. ;}()!)()8, c^ from El Verde, west flank of El Yuiuinc Puerto Hico. I.IOO ft. Coll. J. A. Kivero, 11 Aug-. I!)(i2. Paratypes. University of Puerto Rico (]\Iaya belly, lower flanks and buttocks granular; male with a large subgular vocal sac. 196."] i:i,i:i"nii:i{()i)A("rYi,i's iiiiuKicKr 3 (\tl())-. Above, liray, willi viiric^at ions and vcrniicnlaTions of a liion, two shorter and more diffuse marking's of the same color; a wliitish transverse blotch with irreo'ular margins posterior to the sacral region; seat blackish with a whitish blotch on each side; thighs cinnamon brown, with a longitudinal series of whitish blotches from base to knee; two whitish, irregularly margined, transverse bars on the tibiae, and two on the tarsii; throat of male intensely infuscated; chest and belly slightly infuscated. Measuremenis. ( c^ , in nnn ) Snout-vent 34.8 ; head length ]1.8; head breadth 13; femur l-t.5 ; tibia 16; foot 21. Description of paratypcs. UPRM No. 1182 is a 6 specimen with the same data as the type and with the following measure- ments: Snout-vent 34.1; head length 12; head breadth 13.1; femur 14.19; tibia 15; foot 20.8. The snout has a more rounded appearance than in the type, because the ridge at the tip is more indistinct ; the vomerine odontoids are rounded, set well apart and considerably behind the ehoanae ; the tongue is indented behind, and the tympanum is fairly distinct. The color above is blackish all over, with iiulistinct darker areas behind the nape and on the sacrum; the interorbital bars and externally concave dorsolateral lines are not evident, but there are three, small, light-colored areas that appear as if the skin had been ero(h'd fi'om those areas. The whitish thigh blotches are not loo distinct and there is only one broad, dark band on the tibia and another on the tarsus. The venter is nuu*h darker than in the type, but still lighter than its own throat, which is (hM'ply infuscated. This specimen (no. 1132) and the following were ])i'eserved during the daytime, while the type and paratype No. 1136 were preserved at night. This may account for the dai'k(>r coloration of 1132 and 1133. rPRM No. 1133, is also a 6 with the same data and the fol- lowing measurements: Snout-vent 3.3; head length 11; head breadth 12.3; femur 13.3; tibia 14; foot 20. This animal is almost as dark above as UPKM 1132, but the dorsolateral lines are still evident and tlie liglit blotches on the 4 BREVIORA No. 185 thighs are present, altlioujili not as distinctly as in the type. The axillae have eolorless areas which appear like spots; there ai'c two fine, liyht-colored lines marg'inino- a dark band on the tibia and the same ^itnation is repeated on the tarsns. The belly is ninch liiihtei- than in the specirnen just described. CPRM Xos. n;}4 and 11.35 are juvenile specimens just coming- out of the eggs. The dorsolateral markings are present in both examples, but one is almost smooth abov(% while the other has a number of small warts and tubercles. The liml) l)lotches are also more distinct in the warty example (UPK^r 11:54). They are 7 and 8 nnn in snout-vent length. UPRM Xo. 1186 is another i. from El Yunque, 10.9 km. be- tween Palmer and La Mina, elevation 1765 ft., 2 Sept. 1962. Coll. Rivero and Mayorga. Its measurements are as follows : Snout-vent 35.3; head length 12.1; head breadth 14.8: femur 15.5; tibia 16; foot 22.2. The snout in this specimen looks rounder and higher than in the type and the snout ridge is only shown by a small ])ro- tuberance. In coloration it looks more like the type than any of the other examples, but it is more jirofusely reticulated and there is a ferruginous tinge on the dorsum, especially on the area of the dorsolat(M'al lines. The colorless area of the axillae is very distinct, and the whitish thigh spots are transverse and well indicated, but the seat is not as dark as in the type. The venter is of a light color whih^ the throat is as infuscated as in the other examples. T^PRM No. 1137 is a 7 mm juvenile from the egg clutch of the above paratype. It siiows the dorsolateral nuirkings (tlu)Ugh much broader than in the adults) but the thiaiis are moiv or less marbled and llie ventei- is pin-])ointed with black. UPRM No. 1138 is also from the egg clutch of 1136. It is 6.2 mm in length and has a small tail, 'AS) nnn in length. Appai'entjy it had just come out of the ('g cxamino only a singlf paratype of hfihanicoinKis. 2 Thcsp terms (for present purposes sufficiently defined liy the data witliin the parentheses ahove) are diM'ived from the doctoral thesis of Kicliarcl Kl heiiclLTc at the University of Michigan (availahle on microfilm). 1963 ANOLIS HENDERSONI 3 granular flank scales. Ventral scales smooth, polygonal, sub- imbricate. Digital dilations moderate. About 19-21 lamellae under phalanges ii and iii of fourth toe. Tail not compressed nor with crest. Verticils obscure. Sexually dimorphic in size and color pattern. Anterior head scales smooth. Supraorbital semi- circles separated by one scale row. Interparietal scale separated from semicircles by 4-6 scales. Loreal rows 6-7, canthals 6-7, supralabials to center of eye 6-7, suboculars in contact with supralabials. Mentals much longer than wide, throat scales medially deeply inserted between them. Three populations con- form to this definition but differ strongly in male color pattern (and in one instance in the relative head length of the adult male ) . TAXONOMIC DESCRIPTION AnOLIS HENDERSONI HENDERSONI Cochran^ Anolis hendersoni Cochran 1923. Jour. Washington Acad. Sci. 13:25. (Type locality: Petionville, Haiti) — Cochran 1941, p. 181. Specimen list. Haiti. Departement du Nord : Citadelle MCZ 25484, 25486, W. J. Eyerdam, 1927. Departement du Quest : Port- au-Prince MCZ 13792, 13794, 13795, 13797, G. M. Allen, 1919. Petionville USNM 59210 (type), J. B. Henderson and P. Bartsch, 1917. Boutillier Road MCZ 59951-7, E. Williams and A. S. Rand, 1959; MCZ 62956-9, A. S. Rand and J. Lazell, 1960. Morne Decay ette MCZ 62960-8, A. S. Rand and J. Lazell, 1960; MCZ 62969-75, 63443, L. Whiteman, I960; MCZ 65635-46, L. White- man, 1961. Diqiiini MCZ 64824-40, L. Whiteman, 1961. Below Kenskoff MCZ 59950, L. Bonfil, 1959. Penault MCZ 63437-42, L. Whiteman, 1960. Furcij MCZ 64823, L. Whiteman, 1961. Mar- bial, 21 km NE Jacmel MCZ 65170-8, CM 3812-17, L. Whiteman, 1961. Croix Joseph, Marhial, 21 km NE Jacmel MCZ 65183-202, CM 37818 (16), L. Whiteman, 1961. Source Fleury, Mayerre, 8 km E Jacmel MCZ 65179-82, CM 37819-22, L. Whiteman, 1961. Departement du Sud : Butete near Miragoane MCZ 66029-62, CM 37919 (32), L. Whiteman coll. 13-viii-61. Mingrette near Miragoane MCZ 66063-79, CM 37920 (15), L. Whiteman coll. 1 The following abbreviations have been used for the museums or collections from which specimens of anoles of this complex have been examined : AMNH. American Museum of Natural History ; CM, Carnegie Museum ; MCZ, Museum of Comparative Zoologv ; UMMZ. Universitv of Michigan Museum of Zoology : USNM, United States National Museum ; YPM, Yale Peabody Museum ; AS-X, Albert Schwartz, personal collection, BREVIORA No. 186 1961. Risque near Miragoanc MCZ 66080-85, L. Whiteman coll. 1961. Diagnosis. Head length in adult male about 33 per cent of snout-vent length. Male coloration distinctive. (Head brown. Nape vermiculate, lighter on darker brown. Dorsum brown anteriorly, greenish posteriorly, without transverse saddles or other markings. Flank stripe extending to groin bordered above by intense black and below at sides of belly by black vermicula- tion. Belly bluish.) Figure 1. Anolis hendersoni hendersoni, MCZ 59949. Lateral and dorsal views. N. Strekalovsky del. Color in life. Alive, hendersoni is quite spectacularly beautiful — especially the male. The differences between the two sexes are best shown by a tabular comparison of the descriptions of two specimens. I add a few remarks on color variation. $ from below Kenskoff Head brown and anterior back brown merging into green at sacral region. 9 from Boutillier Eoad Head, nape and center of back dark brown, head somewhat mottled. The central dorsal area bordered by a cream line with an irregular bound- ary. Nape with vermiculation behind interparietal which is white. Nape without vermiculation. parietal white. Inter- 1963 ANOLIS HENDERSONI Blue spots on upper and lower la- bials. Limbs red brown. Tail with greenish tinge, becoming black posteriorly. Upper and lower labials obscurely marked with darker. Jjinibs and tail brown, obscurely mottled. A broad blue-black band from be- hind eye over shoulder, becoming less well defined posteriorly. Below this a light line from upper labials to groin, white below eye, yellow from shoulder % of way to hind leg, beyond this with greenish tinge. Flanks I)elow light line lioldly mot- tled with black. Throat yellow green. A broad brown band from snout through eye along flanks, lighter in center, dark edged above and below. A light line from upijer labials above shoulder to groin, nearly white be- low eye, yellow green in front of slioulder, purplish and indistinct on Hanks. Flanks below light line darker, not mottled. Throat yellow green with two narrow dark longitudinal lines. Belly blue green, some orange under jBelly and underside of tail pale base of tail. Jgreenish. Comments. The female, as the table indicates, is basically similar to the male in pattern but with duller colors and with frequently a light dorsal longitudinal band. The light flank stripe of the male is very brightly colored and brought into bold relief by the black mottling below it and the blue-black of the dark band above it. The same stripe is always obscure or absent posteriorly in females and may be very little evident anteriorly. The green of the hind quarters of the male is absent in the female and so also is the vermiculation on the nape. On the other hand, the two narrow black lines on the sides of the throat may be present or absent regardless of sex. There is a striking consistency in pattern in the animals from well-separated localities (i.e. Diquini and Mayerre). Such varia- tion as exists seems to be individual only. In the males this appears to be a matter of clarity of expression of various elements of the pattern ; this has undoubtedly been influenced by vagaries of preservation. In the females, along with this same variability in the boldness or obscurity of pattern, there seems to be also real pattern variation in the dorsal zone from interparietal to sacral region. In the live specimen described above the impression is of a wide dark middorsal stripe bordered by a narrow light area on 6 BREVIORA No. 186 each side. In numerous other females, however, the middorsal dark stripe is narrow, broken or irregular with usually a narrow light center and bordered on each side by a wide light area. This appears to be the more frequent of the two conditions. This long-headed and slender form has until now been con- sidered very rare. While nowhere as abundant as the ubiquitous species cy botes and distichus or even the two common green anoles, it is, as is often true of "rare" species, not really un- common in certain restricted situations. It seems to be a bush anole of middle elevations and associated especially with certain bushy thickets. Its habits are thus similar to those which Noble and Hassler (1933, p. 14) reported for haharucoensis. Hender- soni is, however, apparently less restricted to humid situations than Noble and Hassler believed haharucoensis to be. A. S. Rand made field notes (July 30, 1961) for this species at Morne de Cayette. "A. hendersoni were on stems and branches close to the ground, three or four feet up at most. They were crawling about in the bushes, jumping from branch to branch, seldom coming to the ground, though one did do so to catch an insect. They are very shy and escape by dodging away through the stems, neither climbing nor hiding. Relatively slow moving normally. ' ' James Lazell, Jr. provides in Figure 4 sketches from life of female hendersoni from Boutillier Road (August 9, 1961). These show in excellent fashion the characteristic postures and attitudes of the species. Anolis hendersoni haharucoensis Noble and Hassler Anolis haharucoensis Noble and Hassler 1933. Amer. Mus. Novi- tates No. 652: 12. (Type locality: "Valley of Polo, Barahona Province, D. R.")— Cochran 1941: 184. Specimen list. Haiti. Departement de Quest : Road to Sal Trou, on south side of range of Mt. La Selle AMNH 50096, W. G. Hassler, 1935. Caroije near Sal Trou MCZ 68693-714, TJSNM 146616-8, YPM 3704-13. UMMZ 123372-81, CM 38497 (11), G. Whiteman, 1962. ^ Doimnican Republic. Barahona Province: Vicinity of Polo AMNH 49516. VaJle de Polo AMNH 51081-106, 51108-19, 51123-27, 51128 (type), 51129-52, 51154-56, MCZ 43822, 45952-53, 56138. Polo AMNH 50317, 50322-23. Palomino 1 The recorrts listeil above for the vifinity of Sal Trou are the first for Haiti. At Caroy^, in addition to .4. bahanicotuKis, Georjre Whiteman obtained also the rare form, Chinndcliiioropis icvtmorti. 1963 ANOLIS HENDERSONI Springs near Barahona AMNH 49840-42, 49884-85, MCZ 43827, 56137. Barahona AMNH 50263.1 Diagnosis. Head not more than 33 per cent of snout-vent length. Coloration of male distinctive. (Head green. Nape green spotted with brown. Dorsum with broad brown saddles on a green ground. Flank stripe extending to groin, not sharply defined, sometimes broken, bordered irregularly above and be- low. No black vermiculations on sides of belly. Belly cream tinged with brown and green.) Color in life. Noble and Hassler (1933, p. 13) : "In life the male is an extremely beautiful lizard. In its usual and brightest phase the dorsal surface of the head is an olive green ; the neck a lighter green, spotted with brown. The back is a bluer green, Figure 2. A^wll.s hendersoni halianicoeiisis, MCZ 68917. Lateral and dorsal views. N. Strekalovsky del. 1 Sonip confusion exists regardintf the labelling of American Museum of Natural History specimens. Seventy-three paratypes of bahdnicoeitsis — AMNH 51081-127, .">1129-r)0. all apparently from Valle de Polo, are listed above. (The specimens missiuf,' from the series are all presunuibly exchanged.) However, Noble and Hassler's text mentions specimens from five additional localities, only two of which are now accounted for by specimens so labelled in the American Museum collections. Thus, unrepresented in the extant collections, though mentioned by Noble and Hassler. is material from Maniel Viejo, from the coffee finca of Senor Luis E. Del Monte near Barahona, and from the property of Mr. G. HeiTnann near Paradis. In addition, some of the specimens now catalogued in the American Museum as baharucoensis from Polo are pulchcllus, presumably from Puerto Rico or the Virgin Islands. Doubtless, as is known to have occurred with Hassler's and other material at this period, some of the lizards were kept alive for a period and only later and rather randomly catalogued. Certain of the MCZ material (MCZ 43822, 43827) at present catalogued as paratypes, is listeu as received in exchange from J. C. Armstrong and presumably was never at the American Museum and thus is not, in fact, paratypic. 8 BREVIORA No. 186 while the four broad saddles or cross bars are a tone of burnt umber. The dorsal surface of the tail, for the anterior two-thirds, is a yellowish green with several dark brown bars. The posterior end of the tail is brown. The side of the head is greenish brown back to the eye. The upper eyelid may be vivid golden yellow. The posterior corner of the lower lid may be blue or purplish. Just posterior to the eye is a narrow patch of dark brown, fol- lowed by a crescent of light blue or white. Posterior to this the side of the head is a brownish green merging into the lighter green of the side of the body which is peppered and veined with brown. Extending from a i)oint on the upper labials anterior to the eye along the sides of tlie body nearly to tiie hind leg is a slightly broken white or cream-colored line edged with brown and suffused in the region above the front leg with yellowish green. The legs are light brown above with slightly darker bars. They are nearly white beneath. The ventral surface of the ab- domen is cream-colored, faintly tinged with brown and green. The throat is the same color with several rows of very faint brown spots along the side. "This species changes color rapidly and to a marked degree. When the lizard is caught or frightened, these colors almost instantly become darker, the green changing to gray or dark brown with the brown cross bars growing darker and almost black-edged. The liead l)ecomes dark l)rown and the labials green- ish. The pineal region becomes white and very conspicuous. The ventral surfaces turn greenish or yellowish and the spots become more distinct." AnOLIS IIENDERSONI DOIilCHOCEPIIALUS Sul)Sl). U. Type: M('Z 64510, adult male. Place Negre near Jeremie, Depart emeiit du Sud, Haiti. Luc and George Whiteman coll. 12-xii-G(). Paratypes: Haiti. Departement du Sud. Place Negre near Jeremie MCZ 64507-9, 64511-36, Luc and George Whiteman coll. 12-xii-60. Les Platans above Carrefour Canon near Ducifi MCZ 62f)76-82, A. S. Hand and J. Lazell coll. 4-viii-60. Carrefour Canon near Dueis MCZ 62!)8;}-f)2, A. S. Hand and J. Lazell coll. 4-5-viii-60. Tomheau Cheval MCZ 62!);)3-7, A. S. Kaiul and J. Lazell coll. 7-viii-60. Mountains on road 1o Jeremie MCZ 56145, AMNII 49504, W. G. Ilassler coll. 1!)35. About 8 miles from Camp Perrin AMNH 50098, W. G. Hassler. 1935. Five miles from Camp Perrin on J (route Road AMNH 50127, W.(}. Ilassler, 1963 ANOLIS HENDERSONI 1935. Camp Perrin AS-X 2664, 2800-2802, 2923-25, A. Schwartz coll. 1962. 13 km N Cavaillon AS-X 3646, A. Schwartz coll. 1962. Diagnosiii. A subspecies of Anolis liendersoni differing in the greater elongation of the head in large males (more than 33 per cent snout-vent length) and in coloration. (Head brown, nape verniicuhite ligliter brown on darker. On dorsum of male a few small middorsal liglit -edged transverse bars at intervals, the widest just behind nape. Flaidv stripe extending only to mid- body, narrowing and terminating rather abruptly. A narrow bhick border above the stripe and black vermiculations below it, both disappearing abruptly along with the stripe itself. The postei-ior flanks unimtterned. Belly yellowish.) Figure 3. Anoli.s lu-ndcrsoni dolirhorrphahis suhap. iiov. Type, MCZ 64510. Lateral ami dorsal views. N. Strekalovsk.v del. Color in life. Description by W. G. Hassler of AMNH 50098 9 from mountains on Jeremie road about 8 miles from (,'amp Perrin, 2000-3000 ft.: "Dark brown, striped with lighter brown or yellowish lines. Belly and throat yellowish. Belly spotted. No green. A. S. Rand for specimens from Les Platons : "5 Uniform light brown above, grayer on head and neck. Darker brown on side of neck. A yellow stripe, black bordered above and below from below eye to midbody. Below pale yellowish. Faint dark striping on throat. Dewlap area with greenish tint. Eye brown. " ? A middorsal stripe edged with light gray. Sides brown with white, black bordei-ed stripe from below eye to mid body. Belly yellow. Throat and chin white." 10 BBEVIORA No. 186 A. Schwartz for specimens from Camp Perrin : " 9 Dorsal ground color yellow-tan with a fine pale hairline with brown suf- fusions on each side. Sides dark brown with a light lower line on sides. Upper labials cream, occipital creamy. Ventral ground color pale greenish yellow." Comments. Again the color pattern is remarkably consistent in the several populations sampled. Males from the south side of the Massif de La Hotte (Les Platons etc.) differ from those from the vicinity of Jeremie only in the weaker expression of certain features, i.e. the vermiculation of the nape and the small mid- dorsal transverse markings. The striking way, however, in which the flank stripe, the black line above it, and the vermiculation below stop abruptly at the same place is exactly repeated in all Figure 4. AnoUs hendersoni hendersoni. A Boutillier Road female at the height of one foot above tlie ground. Characteristic poses. J. Lazell, Jr. del. 1963 ANOLIS HENDERSONI 11 male specimens. The extreme long-headedness of fully adult males is again characteristic of all samples. Female coloration is very like that of typical hendersoni except that in topotypic specimens from tlie vicinity of Jeremie the lateral light stripe fades rather abruptly at midbody, while in females from south of the Massif de La Hotte the stripe tends to continue very faintly to the groin as is usual in typical hender- soni. Dolichocephalus at Tombeau Cheval (3000 ft. elevation) was captured along with Anolis monticola, A. distichus subsp., A. cyhotes subsp. and A. coelestinus in the vicinity of a great heap of jagged limestone boulders overgrown with bushes and with much leaf litter — cool and shady. A. S. Rand reports that here while the monticola were found ' ' around the cliffs on the rocks, roots, twigs, small branches and leaves close to the ground," dolichocephalus was "... up on small branches, vertical stems one to four feet up and to a less extent among the rocks. A. distichus was common on the larger trees and less frequently on big rocks. A coelestinus was seen and a few A. cyhotes on rocks, trees, and bush stems. These last were much more common along the open trail and in coffee." SPECIES OR SUBSPECIES The three members of the hendersoni complex are allopatric and in most respects extraordinarily similar. Despite Cochran's statement quoted above, I do not find the structural differences between these three taxa at all clear. Abundance of material has reduced rather than strengthened any suggestion of shape or squamation difference. The color differences, on the other hand, are very marked. What is the biological significance of such differences in anoles without a functional dewlap? There is no objective evidence. I have reduced haharucoensis to a subspecies and have described dolichocephalus at this level for the following reasons: Wliile color is very important in Anolis and while the importance of pattern and color may be expected to be still more important in forms which lack a functional dewlap, anoles which do differ strongly in color and pattern (and even in structure) may inter- grade (e.g. Lazell, 1962). Further, every sibling Anolis species of which I am aware is discovered to show at least average struc- tural differences once the sample is adequate. The apparently contradictory case of .4. alter newly described by me (Williams, 12 BREVIORA No. 186 1962b) I believe to be an instance of inadequate sample size. The material of the hendersoni complex is now quite sufficient to demonstrate average differences, were they in fact present. It is true that the largest males of dolichocephalus are slightly but distinctly longer headed than any other members of the complex, but this is quite obviously not the sort of evidence that suggests species status. HISPANIOLA SCALE OF KILOMETERS hendersoni 9 bahorucoensis ■ dolichocephalus -f- Fi^iiie "). M;i]) of tlu' (lisliibution of the races of A. licnilcrsoni. ZOOGEOORAPHY The two specimens from tlie Citadelle indicate that much is still to l)e found out about this gi-oup in Hispaniola. Rare as these forms have been in previous collections, they are abundant in some of the most recent. It is therefore impossible to reason on the basis of negative evidence that A. hendersoni is really absent in any part of Hispaniola north of the Cul de Sac — whether in Haiti or in the Dominican Republic. 1963 ANOLIS HENDERSONI 13 It is thus too early to discuss the origiu or history of the hvndcrsoni complex. It is, however, worth calling attention to the tripartite division within what I have called (1961, 1962a) the "southern island" — Ilispaniola south of the Cul de Sac. This singular pattern occurs in several other instances and will be discussed further in later papers of this series. ACKNOWLEDGMENTS I am indebted to A. S. Rand and J. D. Lazell, Jr. who first collected the new subspecies of hcndersoni described herein and to Luc and George Whiteman of Port-au-Prince, Haiti, who ob- tained most of the other material reported here. I am grateful also to Dr. Doris M. Cochran and Mr. C. M. Bogert who permitted me to examine material in their charge and to Dr. Albert Schwartz who has allowed me to study material he has very recently col- lected. A grant from the American Philosophical Society and National Foundation Grant NSF G5634 have supported part of the collecting reported here. Work on these and other Haitian anoles has continued under NSF G16066. EEFEEENCES CITED Cochran, D. 1923. A new Annlis from Haiti. Jour. Washington Acad. Sci. 13: 225- 226. 1941. The herpetology of Hispaniola. Bull. U. S. Nat Mus 177: 1-398. Noble, G. K. and W. G. Hassler 1933. Two species of frogs, five new species and a new race of lizards from the Dominican Republic. Amer. Mus. Novit. No. 652: 1-17. Lazell, J. D. Jr. 1962. The anoles of the eastern Caribbean (Sauria, Iguanidae). V. Geographic differentiation in A7wlis oculatus on Dominica. Bull. Mus. Conip. Zool. 127: 466-475. Williams, E. E. 1961. Notes on Hispaniolan herpetology. 3. The evolution and rela- tionships of the Anolis semilineatus group. Breviora, No. 136: 1-8. 1962a. Notes on Hispaniolan herpetology. 6. The giant anoles. Breviora, No. 155: 1-15. 1962b. The anoles of the eastern Caribbean (Sauria, Iguanidae). IV. The anoles of the northern Leewards, Anguilla to Montserrat: New data and a new species. Bull. Mus. Comp. Zool. 127: 453-465. BREVIORA Museiiwii of Comparative Zoology Cambridge, Mass. .July 10, 1963 Number 187 THE LABYRINTIIODONT DENTITION^ By John Newland Chase Department of Zoology, Ohio Wesleyan University Because of the pliylogenetic importance of the Labyriiitho- dontia, numerous studies have been made of the cranial anatomy of members of this oreat amphibian group. Little, however, has been written regarding their dentition. My attention was called to this subject because of the unusual nature of the dentition of a newly discovered trimerorhachoid rhachitome from the Texas Permian, which is to be described in a forthcoming paper as Neldasauriis wrujhiae? In the present paper I shall review our current knowledge of the dentition of the Labyrinthodontia as a w]u>le. The data on which this review is based, summarized in Table 1, were drawn directly from specimens in a few instances but most were obtained from the literature ; measurements re- corded are tliose quoted by authors or determined from scale drawings of original or reconstructed specimens. Fifty-one gen- era, including temnospondyls, anthracosaurs and a crossoptery- gian Avere reviewed. The arrangement of labyrinthodont genera in the table is essentially as in Romer's 1947 classification (pp. 310-319). Characteristic of labyrinthodonts generally is a dentition which includes, in addition to marginal tooth rows, palatal tusks and often, at least, a shagreen of denticles on the palate and the coronoid region of the lower jaw. The primitive condition, it would seem, was one in which a relatively small number of marginal teeth were present and the palatal dentition consisted 1 This paper is part of a tliesis submitteil to the Department of Biology of Harvard University as partial fultillnient of the reqnirements for the degree of Doctor of Philosophy, August, 1U(52. 2 This is a nomen nudum here, and will enter scientific nomenclature with the publication of my projected descrijition of this new form. 2 BREVIORA NXD. 187 of a siiiti'le stout tu.sk-j)air on each of tlie tliroe lateral ]ialatal elements. There is, however, wide variation in the number of marginal teeth, and in various instances a trend toward reduc- tion in size and concomitant increase in numbers of palatal tusks. Tn this paper I have confined attention to the dentition of the upper jaw and palate, since, apart from the occasional presence of symphysial tusks, the lower dentition in great measure mirrors the marginal dentition of the upper .iaw. Further, no attempt is made to ascertain how widely a shagreen of tiny denticles is present on the jialate, since, even if present, such a shagreen is often destroyed in preparation and consequently unrecorded. The data ])resented in the tal)ulatio]i include : (1) A formula representing the palatal dentition, recorded in terms of the numbers of tusks (or tusk pits) and small teeth on the vomer, the palatine and the ectopterygoid of one side. For example, the formula for the palatal denti- tion of Lyroccphalus is 2 — 2(3) — (13), which means there are two tusks on the vomer, two tusks and three smaller teeth on the palatine, and thirteen small teeth on the ectopterygoid. (2) Information concerning the marginal upper jaw denti- tion such as (a) the number of premaxillary and maxil- lary teeth, (b) the presence of regionally enlarged teeth forming "canine peaks" Avhenever it could be determined, indicated on the table by a plus sign. (3) Skull length taken as the distance from the tip of the snout to the end of the occipital condyle. (This, taken together Avith the figure on the number of teeth, will give an indication of relative tooth spacing.) (4) Sources, placed here to avoid repetition in the text. It is generally and reasonably assumed that labyrinthodonts are descended from rhipidistian crossopterygians. The dentition is known in only a few crosso])terygiaiis, such as E nsthenoptcron, MegalicJifJnjs and Eciostcorhachis. In Eusilioioptcron of the Upper Devonian we find: (1) a series of very nuiiu'i-ous small marginal teeth, (2) a row of numerous small teeth along the outer margins of the vomer, palatine and ectopterygoid, (3) larger tusks, few in mnnber, ])lace(l more medially on these thive elements, with some indication of tusk and ])it pairing; as far as known the vomer bears only one pair of tusks. In the following discussion we will consider marginal and palatal dentitions separately. 1963 LABYRINTHODONT DENTITION MARGINAL TKETII T1h> oldest known labyriiitliodoiit, Ichthyostcf/a from East Greenland beds near the Devonian-Carboniferons bonndary, is somewhat oft' the main line of" labyi-inthodont evolntion bnt shows a nio(h^st number of well spaced teeth — 8-9 premaxillary, 16-18 maxillary — of fairly ji'ood size, most al)ont the same length but with th(> posterioi- premaxillary teeth somewhat eidarfjed. Acan- thostega, an iehthyosteo-alian of comparable age, appears to have had more maxillary teeth — about 30. Romer (1947) provision- ally associated the Colosteidae of Linton with the ichthyostegals. The marginal dentition of the colosteid Erpciosaurus, however, does not agree with that of the ichthyostegals, for here, in con- trast, there is a reduced number of rather large premaxillary teeth and very numerous small maxillary teeth. All further well-known labyrinthodonts can be clearly divided into temnospondyls and anthracosaurs. The marginal dentition of temnospondyls will be dealt with first. The most primitive temnospondyls (apart from their peculiar ''keyhole orbit") are clearly the loxommids. such as Baphetcs and Mcgaloccplialns. In the number of premaxillary teeth both appear primitive, but the maxillary series — 2\ in Bophctes, 40 in Mcgolocrphah(s — are in contrast, the larger number being perhaps proportional to skull elongation in Mcgaloccphalus. Exemplifying the next higher stage are the edopsoids, Edops and Eugyrinus (the last was considered a trimerorhachoid by Romer, but recent studies by Carroll indicate that Eugyrin\is is more primitive and generalized than the trimerorhachoids) . Here the premaxillary teeth are primitive in number (9, ?7) and the number of maxillary teeth (24, 29) is fairly low. Among trimerorhachoids, some agree fairly well with the prim- itive edopsoid pattern (the maxillary count of 19 in Saurerpc- ton is lower) but within the group there is a notable tendency toward an increase in the number of marginal teeth. The pre- maxillary count increases from 9 in Sourerpcton to 12 in Trinic- rorlKfchis and Eohrachyops and to 15 in Neldasaurus and Dvino- mnrns. The number of maxillary teeth is somewhat increased in Eohrachyops, Dvinosaurns and Trimerorhachis and increased to the spectacularly high number of approximately 93 in Nelda- sdunis. Certain, forms which Romer grouped as the Micropholoidc^a show characters more or less intermediate between primitive edopsoids and the "typical" rhachitomes of the Lower Permian. 4 BREVIORA No. 187 Three members of the group — Archegosaurus, Chenoprosopus and Lysiptcrygium — show a "normal" rhachitomous tooth count with 8-11 premaxillary and 27-29 maxillary teeth; Micro- pholis, however, has a reduced dentition with 5 premaxillary and 16 maxillary teeth, Avhile PJatyops, in correlation with an ex- tremely elongate skull, has a total of about 65 marginal teeth in the upper jaw. The "typical" rhachitomes, the Eryopsoidea, include a wide variety of Permian labyrinthodonts and a few Carboniferous predecessors. Dental formulas are given for nine members of this group (nos. 20-28, Table 1) and show some variation within the group. Premaxillary teeth range from a maximum of 15 in the broad-snouted Zatrachys and its Carboniferous relative Acan- tJtostoma, through 13 in Eryops to a low figure of 5-8 for Acti- nndon, Cacops and trematopsids. Some eryopsoids have a reduced maxillary count with a low of 12 in Cacops; Eryops, on the other hand, has an increased count of 38. The trematosaurs, a persistently rhachitomous, early Triassic, fish-eating group, reflect accompanying snout elongation in tooth numbers. Although the premaxillary teeth retain a count of 10-12, in the long-snouted forms the number of maxillary teeth increases, reaching a figure of 50 in Trcmatosuchns. The neorhachitomes of the late Permian are presumed to have been derived from eryopsoids. Such a typical form as Rhine- suchns lias a marginal dentition almost exactly like that of its morphological ancestor Eryops; Lydekkcrina, however, has a reduced formula, with only 20 maxillary teeth. Still more ad- vanced neorhachitomes of the early Triassic, those apparently leading to the capitosaurs, are such forms as WcUngasaurus, VoUjasaurus and Benthosuchus. Here, facial elongation is accom- panied by an increase in the number of maxillary teeth to 52, 56 and 61 in these genera, respectively. A high number of maxil- lary teeth occurs in capitosaurs, reaching a peak in Mastodon- saurus, which has 23 premaxillary and about 75 maxillary teeth. Parallel in development to the capitosaurs are those forms grouped as the metoposaurs. Metoposaurus itself has a "general- ized" count of 10 premaxillary and 34 maxillary teeth, but in the American species Eupdor browni and the European Ey prior frnasi the count increases to about 60 maxillary te(>th in the former and to 1)0 in the latter. Certain short-faced Triassic forms tend to have reduced counts; the brachyopid Batrachosnchus having but 17 maxillary teeth and the plagiosaur Gerrothorax having somewhat over 26. 1963 LABYRINTIIODONT DENTITION 5 A group by group account of variations of marginal tootli count in tenniospondyls has been presented above. As can be seen, no consistent ]iattern emerges. One gains the general im- pression of a probable early temnospomlyl condition of 9 or so premaxillary teeth and a maxillary count in the 20 's, with a modest increase, on the average, in typical rhachitomes. Occa- sionally there are reductions to lower figures (as, for example, in Sclerocephalus, Actinodon, or the dissorophid Cacops). On the whole, however, variations are toward higher figures in later forms. In some cases increases in tooth count are definitely associated either with notable snout elongation, as in Platyops and the trematosaurs, or, as in capitosaurs and metoposaurs, with a combination of moderate skull elongation and absolute increase in size, the teeth failing to increase in proportion to the total size of the animal. Most exceptional of all are the trimerorhachoids, particularly Neldasaurns, where an exceedingly high tooth count may be found in a skull of modest proportions. No consistent correlation between tooth number and skull size can be demon- strated. Anthracosaurian dentition is known in only a few, mainly Carboniferous forms. Here the premaxillary teeth are low in number, related to the usually narrow snout and the fairly large size of the individual teeth. Marginal teeth are essentially of two types : either large and few in number, as in Anthracosaurus and Pteroplax, or small and more numerous, as in Pholiderpeton, Neoptcroplax and, especially, Archeria, where the exact count is indeterminate although the teeth are certainly very numerous. Marginal tooth counts in the premaxilla and maxilla of ^'('^/- mouria are low; Kotlassia shows a slight increase over the primi- tive number. Regionally enlarged teeth forming "canine peaks" occur sporadically in different labyrinthodont groups but appear to have been fairly common in primitive forms. PALATAL TEETH Known rhipidistian crossopterygians have, as mentioned above, (1) a very large number of small teeth in a row along the mar- gins of the vomer, the palatine and the ectopterygoid, (2) larger tusks more medially placed on these three bones. As is well known, in all typical labyrinthodonts the larger teeth are re- tained while the lateral row of smaller teeth is lost. But, surpris- ingly, in IchtJiyostcga there is no trace at all of the "tusk" row; 6 BREVIORA No. 187 instead it appears that the lateral row of numerous, essentially even-sized teeth has been retained. In this regard Ichthyostcga is very far removed from the condition expected in the ancestor of later labyrinthodonts. Primitive temnospondyls proper almost always have one pair of tusks on each of the three bones, the vomer, palatine and ec- topterygoid. This is true almost without exception in loxommids, edopsoids and most eryopsoids. Occasionally, a replacement pit may be absent ( ? or indeterminate) but exceptions are rare. In palatal dentition, colosteids (e.g. Erpetosanrus) are typically temnospondyl in character, thus arguing strongly against Romer's suggestion of ichthyostegal relationships for these animals. In numerous more specialized or advanced labyrinthodonts there is an increase in the number of smaller palatal teeth and a tendency toward reduction of the more prominent tusks, usually those on the ectopterygoid. The presence of this pattern in different groups indicates that this trend has occurred sev- eral times in parallel fashion. It is, for example, seen in tri- merorhachoids (except Eohrachyops), mieropholoids and trema- tosaurs. In Rhinesuchus, which appears closely related to typical eryopsoids, this trend is seen again, leading to conditions in capitosaurs where, as in metoposaurs, the increase in the number of small teeth is very prominent. In some short-faced Triassic forms the increase in smaller palatal teeth is not so pronounced — a condition no doubt correlated with the short palate in these animals. Distinctive of tlie anthracosaurs is the absence of vomerine teeth — apparently related to the narrow palatal exposure of the vomers in these forms. In the few palates known, there are generally a pair of large palatine tusks and a modest number of ectopterygoid teeth, the front ones tusk-like. In Pliolidogaster, with six tusk-like teeth on the ectopterygoid, and in Anthraco- saurus, where a tusk pair is accompanied by a pair of fairly large teeth, we encounter conditions reminiscent of the crossoptery- gians, in which the palatine had a row of large teeth. Seymouriamorphs are unusual among anthracosaurs. Scy- mouria, in contrast to other anthracosaurs, has a pair of vomerine tusks and — in parallel fashion to some temnospondyl groups — has lost the ectopterygoid teeth. Kotlassia resembles other anthra- cosaurs in lacking vomerine teeth but has a close-set row of small teeth on the palatine and ectopterygoid, decreasing gradually in ]963 LABYRINTIIODONT DENTITION 7 size from front to back. These differences between known anthra- eosanrians sngwest much variation within the group, but the extent of this variation is at pi-esent incompletf^ly known. I wish to express my appreciation to Dr. Alfred S. Konier of the Museum of Comi)arative ZooU)gy for his generous aid and many helpful suggestions during the preparation of this paper. 1 am also indebted to Professor Bryan Patterson and Dr. Ernest E. Williams of the Museum of Comparative Zoology for eon- .stinu'tive criticism of the manuserii)t. EErERENCES Atthey, T. 187(3. On Antliracosaurus lusselli (Huxley). Ann. Mag. Nat. Hist., 18(-i; :146-1G7. Branson, C. C. 1935. A labyrinthodont from the Lower Goiidwana of Kashmir and a new edestid from the Permian of the Salt Range. Mem. Conn. Acad. Arts Sci., 9:568-010. Branson, E. B. and M. G. Mehl 1929. Triassic amphibians from the Rocky Mountain region. Univ. Missouri Studie.s, 4(2) : 155-225. Ekoii.i, F. 1926. Ueber Sclerocephalus liduscri Goldfuss. Sitz.-Ber. Akad. Wiss. Miinclien, Math.-Naturw. Al)t., 1926:199-222. BrOILI, F. and J. RCHROEDER 1937. Beobachtugen an Wirbeltiereu der Karrooformalion. XXV. Ueber Micropliolis Huxley. Sitz.-Ber. Akad. Wiss. Miinclien, 1937:19-38. Bystrow, a. p. 1935. Morphologische Untersuchungen der Deckknoclien des Schiidels der Wirbeltiere. I. Mitteilung. Schadel der Stegocephalen. Acta Zool., 16:65-141. 1938. Dvitiosaurus als neoteniscli,e Form der Stegocephalen. Acta Zool., 19:209-295. 1944. Kotlassia prima Amalitzky. Bull. Geol. Soe. Amcr., 55:379-416. Bystrow, A. P. and J. A. Efremov 1940. BcntlinsucJius susliTcini Efr. — a labyrinthodont from the Eotri- assic of Sharjenga River. Trav. Inst. Pal. Acad. Sci. URSS, 10:1-152. Case, E. C. 1922. New reptiles and stegocephalians from the Upper Triassic of western Texas. Publ. Carnegie Inst. Washington, No. 321:1-84. 1935. Description of a collection of associated skeletons of Trimeror- hachis. Contrib. Mus. Pal. Univ. Michigan, 4:227-274. 8 BREVIORA No. 187 Efremov, J. A. 1933. Xeuljesi-hreil)U]ig fles Labyrinthodonten PJaiyops stuchcnbergi Trautscliold aus den oberperniiscdien Ablageiuiigen des Flusses Kitjak, eines Nebenflusscs der Wjatka. Trav. iiist. Pal. Acad. Sci. UKSS, 2:117-164. 1940. Pieliminary description of the new Permian and Triassic Tetra- poda from USSR. Trav. Inst. Pal. Acad. Sci. UESS, 10(2) :1- 140. Fraas, E. 1889. Die Labyrintliodoiiten der schwiiliisclien Trias. Palaeontograph- ica, 36:1-158. Gaudry, a. 1887. Ij'Actinodon. Nouv. Arch. Miis. Nat., (2) 10:1-32. Haughton, S. H. 191.J. On the genus Rhinesuclius Broom, with notes on the described species. Ann. S. African Mus., 12:65-77. 1925. Descriptive catalogue of the Amphibia of the Karroo System. Ann. S. African Mus., 22:227-261. Jarvik, E. 1952. On the fish-like tail in the ichthyostegid stegocephalians. Medd. oni Gronland, 114:1-90. 1954. On the visceral skeleton in Eusthenopteron with a discussion of the parasphenoid and palatoquadrate in fishes. K. Svenska Vetenskapsakad. Handl. (4)5(1) :1-104. Langston, W., Jr. 1953. Permian amphibians from New Mexico. Bull. Dept. Geol. Univ. Calif., 29(7):349-416. NiLSSON, T. 1934. Vorliiufige Mitteihing iiber einen Stegocephalenfund aus dem Rhiit Schonens. Geol. Fciren. Forh., 56:428-442. 1937. Ein Plagiosauride aus dem Rhiit Schonens. Beitrage zur Kenntnis der Organisation der Stegocephalengi-uppe Brachyopoidei. Acta Univ. Lund., 34(2):l-75. Olson, E. C. 1941. The family Trematopsidae. Jour. Geol., 49:149-176. ROMER, A. S. 1930. The Pennsylvanian tetrnpods of Linton, Ohio. Bull. Anier. Mus. Nat. Hist., 59:77-147. 1947. Review of the Labyrinthodontia. Bull. Mus. Comp. Zool., Har- vard, 99(1): 1-368. 1963. The larger embolomcrous amphibians of the American Carbon- iferous. Bull. Mus. Comp. Zool., Harvard, 128(9) :415-454. RoMER, A. S. AND R. V. Witter 1942. Edops, a primitive rhacliitomous auipliiliinn from the Texas red- beds. Jour. Geol., 50:925-960. 1963 LABYBINTHODONT DENTITION 9 Save-Soderbergh, G. 193(). On the moiplioloKy of Triassic stegocephalians from Si)itzberKen, and the interpiotation of the endocraniuni in the Labyiintho- dontia. K. Sveuska Vetenskapsakad. Handl., (3)16(1) :1-181. Sawin, H. J. 1941. The cranial anatomy of Eryops megacephalus. Bull. Mus. Comp. Zool., Harvard, 86(5) : 407-461. 1945. Amphibians from the Dockum Triassic of Howard County, Texas. Univ. Texas Publ., no. 4401:361-399. Steen, M. 1931. The British Museum collection of Amphibia from the Middle Coal Measures of Linton, Ohio. Proe. Zool. Soc. London, 1930 (1931):849-891. 1934. The amphibian fauna from the South Joggins, Nova Scotia. Proc. Zool. Soc, London, 1934:465-504. 1937. On Acanthostoma vorax Credner. Proc. Zool. Soc, London, (B) 107:491-500. 1938. On the fossil Amphibia from the Gas Coal of Nyfany and other deposits in Czechoslovakia. Proc. Zool. Soc, London, (B) 108:205-283. SUSHKIN, p. p. 1936. Notes on the pre-Jurassic tetrapods from USSR. III. Dvino- saurus Amalitzki, a perennibranchiate stegoeephalian from the Upper Permian of the North Dvina. Trav. Inst. Pal. Acad. Sci. URSS, 5:43-91. Thevenin, a. 1910. Les plus anciens quadrupedes de France. Ann. Paleont., 5:1-64. Watson, D. M. S. 1912. The larger Coal Measure Amphibia. Mem. Proc. Manchester Lit. Philos. Soc, 57(1): 1-14. 1913. On MicrophoJis .siowi Huxley, a temnospondylous amphibian from South Africa. Geol. Mag., (5) 10:340-346. 1919. The structure, evolution and origin of the Amphibia. The "Orders" Rhachitomi and Stereospondyli. Philos. Trans. Roy. Soc. London, (B) 209:1-73. 1926. Croonian Lecture. The evolution and origin of the Amphibia. Philos. Trans. Roy. Soc. London, (B) 214:189-257. 1929. The Carboniferous Amphibia of Scotland. Pal. Hungarica, 1:219-252. 1940. The origin of frogs. Trans. Roy. Soc Edinburgh, 60:195-231. 1956. The brachyopid labyrinthodonts. Brit. Mus. (Nat. Hist.), Geol. Bull., 2(8):318-391. White, T. E. 1939. Osteology of Seymouria baylorensis Broili. Bull. Mus. Comp. Zool., Harvard, 85:325-409. 10 BREVIORA No. 187 Whittard, W. F. 1928. On the structure of the palate and mandible of Arcliegonaurus declieni Goldfuss. Ann. Mag. Nat. Hist., (10) 1:255-264. WiLLISTON, S. W. 1909. New or little known Permian vertebrates: Trematopa, new genus. Jour. Geol., 17:63G-G5S. 1910. Cacops, Vctimoiipondylus: new genera of Permian vertebrates. Bull. Geol. See. Amer., 21:249-284. 1911. Restoration of Seymouria baylvrensis Broili, an American cotylosaur. Jour. Geol., 19:232-237. 1963 LA15YK1NTJ10DONT DENTITION 11 ^^ i-i v'Ss^ OlO 0000000000(0050 o "T rp^Qj^OO lOOO [--cCl^t^i-lrH^EtOOt^CO IT CI o o H -^ 5 .2 " g 5 :^ o g Ti ^ '"? T' ~ ^ o I ^ 't< o o; Ot. 00 O O l~ o M-t< iL^Cl^li— (ClCCOOfO TJ 11 P( O] GO I „ ^ ^ rt ;=l CS o o -+-) -*-• !|> r^ • rH -f^ C3 rH ^ P Ph o -t^ rH T' ^, -1 01 'M "1 CI O CI 01 ^ rH ^'^ ^. CI ' ^^ 00 ',' I , I i . 1 ■ --~s C] I ' "* I ^ -^ ■ . CI Cl 01 -1 '11 ^ CI ci '^ - ] ^^ ■ — ^-^ CI CI O CI CI T-H C] C] C] rH CI -t* C] M o C OQ !^ S „ -2 m tD '^ T, « c3^aJo'C3-a'iHO> - ^ i-H Ca CO Tji id O t^ OO' ci O' r-i Cl rO TJH lo" «3 12 BREVIORA No. 187 fO CO ri f^ K +- :o; >- (-^ :cs 1-3. L_2 o;so^- .r*iS-i-^.^^-i--. •* • cH CS fTi U

!-! o Q '15 a ~ -4-^ n ■fH 3 fcX) S „ ^Cj^lOt^ CO lOlO l^ Ocoo TtHO'*''0'— (lOCOOl H^.S S^ COrH r^ "-I r^I-l i-lTtlrHCO + + ++ ++ + + + + + Oi lO W QO ■*! to 01 lO CO O , "* CI C O 01 CI C-JO rH COi— I t— I 1— ICIOI COIOCOCOI'^COCO f^ 2 J S X Ppa p,00| in coo b- GOlOt^ -OLOItOOlOClCO ^•r^ I T— tr-t r^r^r^T— Ir— ir— I oi 1-^ 01 r-i ^-^ ^-N /-^ Ol o ^^ C-. QO OJ r^ y-^ '^^ , 1— I i-l 1— I ~w' « •■T! M "^ ^' '^^ ^'J O CI c] 01 CI 01 V ^ w w 3 c3 c ' c;^ ' ■ . ■ I . . , , , , -^ ' Ol 01 ci l-H_P,._^LO -— - 01 01 01 OlCloi ClClOlij^O'^.^^^ . , '-< '^' . . . , . , . , , CI '"' ^ "® "^ > O 01 Ol Ol CI rl 1—1 CI CI Ol 01 01 CI CI 01 01 Ol o 5 .2 13 73 5 o ^ CS rt "a o .a O p4 p 6 .s 03 a o c: s O s -A .2 ci c -4-' 9 so o 2 s a: O c: be 4— • r-t !>> ^ !-i ^ O ^K o c: C3 "^ r^ !m ^ A 1-1 d, H 00 < O <; H |1h N < hJ H H S Ph t-' 00 ci o r^ Ol CO ■*' lo" o !>■' 00 o; o rH 03 CO T-( r-i r-l CM CJ Ol OJ Ol Cl OJ Ol oq Ol CO CO CO 00 ]<)(i;} LABYRINTHODONT DENTITION 13 ^^ o CO d ^ t- o f^ •^^ ;!> 0^ 3 ■^ t- JO '^ S ri "^ 55 oj i; -d Oi Oi ^^ (M j3 W ^ ^;; -)< (31 Oi r^ ,~ ---. « -t< Oi Oi ^ C. ^ r-H p^ Oi ' ' ^^ , ,^^. ^^-^ ^^ , ^^ rr [>. o O CO CO O -H CO o CO Qi 00 t^ ■O '11 'TI CO wj 1-r 00 lO CO 1 on T— 1 T— 1 r-l CO CJ tH CO r-i 1—1 1 M + + + + + O CICO'-I lO 't*^ O t^ CDffjt^'T Ot^ t-t- '^ C] lO LO CO t^ el's OD Oi I— I CI CI tH O] -^03 CC ?— I Oi ■* ' CI OO O Ol O'J fj] 00 TfH CO CO CO CO LO 'tl i-H Oi O ^-^ '-^ i-l ^-^ i-l CO ''t* O ^v v^05^^— ''-'rO ^ CO 00 -- OJ CO CO -.;♦< 3 ^ ^ cT i?r ^2. ^ g 2^ A =^ °.' "^^ *.' "^' ? ^ -^00 t^ ^ CO^-N 00 T' .^s^cac] (M t> eq -Jl o o -4-> en 3 rt CO O & o _o IB 03 _3 "a! o CS CS 3 o s -4-' o cS CC o o O o s 4-" 'TIS •rH X p o ft c3 O s CO 03 -+-' « £< 3 ■4-' OS a; "o -*-' O S +- o g H H M O s -^ < E PLh ^ OQ oo' oi o 1— ( ©q CO 'iH \6 co' t^' oo oi O CO CO ■ localities of AmpMnhaena o. occiflentalis Cope [bottom or south], and ./. <>. toirnfiendi Stejneger Ilo]' or iioitli|. 'Die dottrd line on llic lar^c scale |dctail| map indicates the I'rovince of Jacn of the Department of Cajamaii-a, I'eni. 1963 AMPHISBAENA PERICENSIS 3 (A.M. X. II.), :\Iiss Alice (i. (\ (iraiulisoii of the I'.ritish :\Ins(Mnn (Natural History) (B.M.), Dr. Alan Lcviton of the California Academy of Sciences at San Francisco (C.A.S.), Mr. Hymen Marx and Dr. Kohert F. Injicr of the Chicago Natural Ilisloiy Museum (C.N.II.M.), Dr. Ernest E. Williams of the Museum of Comparative Zoology (IM.C.Z.). Dr. Koni-ad Klemmer of the Senckenbergische Naturforschende Gesellschaft of Frankfurt a. M., Germany (S.M.F.), Dr. George S. Myers of Stanford T'ni- versity. Museum of Natural History (S.U.), Dr. Hobart M. Smith of the University of Hlinois Museum of Natural Histor^^ (U.I.M.N.H.), Dr. Doris M. Cochran of the United States Na- tional Museum (U.S.N.M.), and Dr. Heinz Wennuth [formerly] of the Zoologisches Institut der Universitat in Berlin (Z.M.U.). Dr. Virginia Cummings figured the specimen and ^liss Charlyn Rhodes furnished technical assistance. The overall project has been supported by grants NSF G-9054 and G-21819 from the National Science Foundation. Amphisbaena pekicensis Nol)Ie, 1921 Ampliishaena pericensis Noble, 1921, p. 141. Terra typic-a : "Perieo, Peru." HOLOTYPE: M.C.Z. 14(i31. PAEATYPES: A.M.N.H. 28501-03; B.M. 1929.6.1. 85-RR 1946.8.31.83; C.A.S. .■:46]4; C.N.H.M. 16106-07, 73371; B.Z. [ ?] ; M.C.Z. 14764-6.5, 14767-68, 14770, 14772-75; S.M.F. 11826, 11887-88; U.I.M.N.H. 41494; U.S.N.:M. 75970; Z.M.U. 29659 [Perieo]. M.C.Z. 14789-90 [Bellavista]. U.R.N. :\t. 59926 [Chiiic-liipe River]. U.S.N.:\1. 60057-58 [Marafion River]. Status of the types: The status of the various s})ecimeiis col- lected by G. K. Noble is thoroughly confused, possibly because the collection does not appear to have been catalogued at the M.C.Z. in its entirety, and certainly because Noble neither gave the total number of specimens examined, nor otherwise identified the "paratypes." (See table on p. 13.) There does not seem to be any difficulty about the liolutyi)e, M.C.Z. 14631. It Avas listed both by Noble and by Barbour and Loveridge (1929, p. 215). The latter authors also listed M.C.Z. 1476-1-68, 14770, 14772-80 and 14789-90 as paratypes then remaining in the ^luseum of Comparative Zoology. A cheek of the catalog disclosed that tlie series originally ran from 14764 to 14790. Of these I have been able to identify 14764-65, 14767-80 and 14786-90 in various collections, relying always on the presence of one of the original parchment labels. Since 14766 is marked "Amaral" in the 4 BREVIORA No. 189 M.C.Z. eataloo' (hence presumably in the Departamento cle Zo- ohjgia, Sao Paulo collection = D. Z.), only 14781-85 remain in question. All of the questionable specimens have their catalog entry marked " Anomalepis aspinosus Taylor" in A. Loveridge's handwi'iting. The astonishing thing is that Dunn's first report (1923, p. 185) on the collection mentions only three specimens of Anomalepis, while Taylor's original description (1939, p. 92) refers to six including 14781-85. Taylor apparently omitted mention of two other M.C.Z. specimens of this form, and the total of eight once in the M.C.Z. then suggests (1) that Dunn never saw M.C.Z. 14781-85, and (2) by implication that Noble did not check his ''paratype" series very carefully. Neverthe- less, the remaining 17 specimens may well be considered to be paratypes. An even more complex problem is posed by seven specimens also supposedly from the Noble collections; four of these are labelled paratype, stem from Perico, and are respectively in the American Museum (A.M.N.II. 28501-03) and California Academy (C.A.S. 54614) collections. Since Noble was curator of the American ^luseum collection at the time they were cata- logued, it remains possible that these are indeed part of his mate- rial. Less probability attaches to three specimens from the ''Chinchipe River" (U.S.N.M. 59926) and the "Rio Maranon" (U.S.N.M. 60057-58). Stejneger marked the catalog ''topopara- type" and these specimens Avere received from the M.C.Z., but were never catalogued there. Yet even they may be considered to have been included as paratypes by Noble's statement of species range. ^ Diagnosis: A form of Amphishacna without major fusions of head shields ; with pairs of very large first and slightly reduced second parietals ; with the head relatively blunt and not par- ticularly set off from the neck; with a cylindrical blunt-tipped tail ; with a faint autotomy constriction at the sixth to eighth caudal annulus where autotomy takes place. The form has 198 to 218 body annuli ; generally 18 or 19, occasionally 16 or 17 caudal annuli: 12 to 16 (generally 14) dorsal and 16 to 20 (gen- erally 18) ventral segments to a midbody annulus; and 4 small 1 Two spcoinicns (T'.M.AF.Z. ."."ITCiT A it B) just (liscdvorccl in the (•ipllcctioii of tlip Univorsity of Mirlii;,'aii Mnsciiin of Zoology contirm tli;it Xoblc iiippiii'cntly ilis- trilmfcd "pariitypcs" witlioiit first {riving tliciii M.C.Z. niinilitTs. 'I'licir jar contains an M.C.Z. lalirl reading (aiiimrcntly in Nolilo's liand) ■'Aniidiisliacna sp. nov. in nis. — I'prico, N. W. Pern. — Ilarv. IN'rn. Kxp." Ttirlr couiils and nicasurcnionts fall within the range of the present des('rii>tion. 1963 AMPHISBAENA PERICENSIS round preeluacal pores. Tlu' color ol' pi-eser\'i'(l spreimcns is a luiiforni dark brown, much darker dorsally than ventrally, light- ened on the tip of tail and snout and produced by a dense pig- mentation of the segments (contrasted by lighter intersegmental sutures). PERICO " ' RIVERS BELLAVI STA 200 208 216 Body Annuli 12 14 16 Max. Dorsals Fig. 2. Ampliishaena pericensls. Diagram showing distribution of counts of body annuli and maximum dorsal counts of the three groupings collected respectively at Perico, along the "rivers" of Chinchipe and Maraiion, and at Bella vista. Oeographic variation: The entire sample comes from a rela- tively restricted area and 24 out of the 29 available specimens from the single locality, Perico. Two other specimens bear the definite indication Bellavista, while the other three are marked only with the names of the river courses traversing the area. Number of segments to the dorsal portion of the midbody annulus differs clearly between the Perico sample (12 to 14) and the Bellavista sample (15 to 16). Two of the river course specimens also range to 15. It may also be significant that the two Bellavista specimens and one of the river course individuals show a slightly higher number of body annuli than does the Perico sample (Fig. 2). Here, as in so many other instances, we are handicapped by the markedly inadequate samples. 6 BREVIORA No. 189 Description: Figure 3 shows views of the head, Figure 4 the ventral surface of the cloaca and tail, and 6, 7 and 8 are photo- graphs showing coloration and other aspects of the specimens. P'igure 5 gives a scatter diagram of tail versus snout-vent lengths. Meristie data are listed in the table. This is a small species of Amphishacna, of a dark brownish color and without pattern, but with marked dorsoventral counter shading. The lightened ventral color continues anteriorly onto tlie rostral, and caudally to the cloaca and occasionally to the distal tip of the tail. The color is clearly restricted to the seg- ments, as the intersegmental sutures and the lateral grooves are significantly lighter. The specimens thus present the impression of a checkerboard under magnification. The coh)ration appears to i-esult from two levels of pigmenta- tion and two degrees of expression. There is a general darkening of the entire segmental surface, plus an additional intensifica- tion of pigmentation on tlic i-ectangular segmental centers. The dorsal segments of some well jn-eserved specimens liarely show tliis midsegmental ])igment dichotomy ; other, possibly more faded, s])ecimens show it more clearly though never as markedly as in Amphisbatna siU'cstrii (Gans, 1962) in which the margins are quite distinct from the centers. The dorsoventral fading occurs both by a general weakening of the coloration and by a complete dro])ping out of the main jiigmentation by segments. (The absence of living siieciinens does not permit decision as to whetlKn- these ventral segments are tndy unpigmented, or whether the phenomenon liere discussed applies only to those pigments least affected by methods of preservation.) The gen- eral weakening of the pigmentation may result in disappearance or reduction to flecking of the darkening of the entire surface, as well as a reduction of pigment density and size of tlie rec- tangular segmental centers. The complete depigmentation varies in extent from a blanching of most midventral rows to that of scattered groups of ventral segments. Such areas as the tail are ventrally countershaded by the first phenomenon alone and all of their segments retain (more or less faint) dark centers. The head segmentation is characterized by lack of major fu- sions, and by pairs of very large first and slightly reduced sec- ond parietals. The head is slightly flattened dorsoventrally ])ar- ticularly on the posterior half. The rostral region projects slightly across the tip of the recessed lower jaw. The entire body shows some doi-soventral flattening. The tempornl muscles are 1963 A.MPllISBAENA PERICENSIS scarcely noticeable in juveniles, l)ut are definitely swollen in adults in which they change the shape of the head. The rostral is slightly smaller than ai'c the first siipralabials and only its dorsal tip is visible I'roni above. Pairs of niediuni- FiG. 3. AmpJiisbaena pericensis. Dorsal, lateral and voiitial views of the head of the holotype, M.C.Z. 14631, from Perico. The line equals 1 mm to scale. (V. Cummings del.) 8 BREVIORA No. 189 sized nasals, large prefrontals, smaller frontals and wide first and narrower second parietals follow in sequence along the dor- sal midline of the head. The middle of the parietal lies on the level of the angulus oris. There are three supra labials of which the second is generally' the largest. The position of the angulus oris lies at the posterior edge of this segment and is generally quite easy to determine as the intersegmental sutures of the first body annulus do not tend to align with the slit of the mouth. The suture between the first and second supralabials runs at an angle of 45° to the edge of the mouth, while that between the second and third runs directly rostracl or only very slightly an- teriorly. The ocular is quadrangular, contacting the prefrontals and frontals anterodorsally, and the second and third supralabi- als anteroventrally. Its po.sterior margins adjoin the segments of the first i)odv ainiulus. Fit,. 4. A tni>hisb(ii iKi [i< iic( iisis. Yoiitial view uf cloat-a and tail li()liitYi)f. Tlie line equals 1 iiini to sc-alo. (V. Ciinimings, del.) (if the The mental is medium sized and slightly larger than the first infralabials which are followed by the very large second infra- labials (the largest segments of tlie lower jaw), and these in turn by the narrow third infralabials. The oval to pentagonal postmental is somewhat larger than the mental and is in contact with it, with the first and second pairs of infralabials, and with a row of postgenials that clearly exclude it from contact with the iiialai-s. The malars are small and fill the spaces between the wide second and nari'ow third infralabials. The antei'ior row of ])()stgenials generally consists of two tear-droj) shajjcd segments, occasionally se]iarated by a small intermediate seg- ment (indicated in the table undei- chin segments by a figure one in parenthesis). This segment occasioiudly contacts the post- mental so that a first row of three postgenials is counted. The second postgenial I'ow contains tlwee to foui- segments. Tliei'c are no postmalai's. 1963 AMl'lllSBAENA PERICENSIS Dorsally, the two paii's of very inucli ciilarjjed segments of the first body annuliis curve anteriorly and the to|)most pair contacts the lateral edges of the frontals. A dorsal half-annuhis of one to two seo-nients on each side intercalates behind this, including the first parietals. The second parietals are the mid- o- 6 J 3 16 E ^ 14 C 0) 12 - 10 8 • om • •• 7 17 Snout - Vent 15 Length - cm Fig. 5. Amphisbaena pericensis. Combined frequency histogram to sliow distribution of size classes of snout-vent lengtli (summed in units of 1 cm, from 7.0 to 7.9, 8.0 to 8.9, etc.), and scatter diagram of tail versus snout- vent length (the hollow dots indicate two coincident points each). 10 BREVIORA No. 189 dorsal segments of the second body annulus. The posterior edge of this annulus shows no forward curvature, though several of its segments are anteriorly elongate, particularly in the tem- poral region. There are 198 to 218 body annuli from the back of the third infralabial, up to and including the pore-bearing precloacals. The fourth through sixth or eighth anterior annuli are modified. The anterior ones are shorter, and the posterior ones longer than the succeeding annuli on the trunk. There is generally no complexing of the segments in the "pectoral" region, nor are there intercalated dorsal half -annuli. There are 12 to 16 (gen- erally 14) dorsal and 16 to 20 (generally 18) ventral segments to a midbody annulus, while the normal pattern appears to be 14/18 (see comments on geographical variation). The cloacal region is characterized by 4 round precloacal pores, wliich are much less noticeable in juveniles and females. There are 6 to 8 precloacal and 11 to 14 postcloacal segments and three to five (generally four) lateral rows. The not particularly apparent autotomy annulus falls on the sixth to eighth postclo- acal annulus and autotomy takes jilace here. Specimens have 16 to 19 (generally 18 to 19) caudal annuli. The cross section of the tail changes from a dorsoventrally flattened ellipse in the immediate postcloacal region, to a circle at the autotomy an- nulus, to an enlarged vertical ellipse behind this. The enlarged distal ti]^ terminates in a blunt vei'tical keel. The lateral sulci are clearly marked, starting between the twentieth and fortieth body annuli and continuing up to the level of the cloaca. At midbody, each of them is represented by little more than an elaboration of the normal intersegmental suture, which is made more comj^lex by having the extreme cor- ners of the adjacent segments diagonally cut off. The dorsal and ventral intersegmental sutures are aligned along the middorsal and midventral region, but there is no other indication of dorsal and ventral sulci. The dorsal segments of a midlxidy annulus are approximately one and one-half times as long as wide, while the ventral ones are one and three-quarter times as wide as long. Range: Pern, Depai'tment of C'ajamarca, "arid valleys of the Chinchipe and Marafion, from Pci-ico on the north to Bellavista on the south." 1963 AMPHISBAENA PERICENSIS 11 LTTERATUEE CITED Barbotr, Thomas jukI Aktiii r Loveridge 19-9. Typit'.-il reptiles and anipliiliiaiis. Bull, .\liis. Coiiip. Zool., vol. 69, no. lU, i)p. 205-360. Burt, Charles E. and May DANHEnt Burt 1930. The Soutli American lizal•d^s in the coUectidn of tlie United States National .Miiseiiin. Free. United States Nat. .Mus., vol. 78, art. G, pp. 1-52. 1931. South American lizards in the collection of the American Mu- seum of Natural History. Bull. Amer. Mus. Nat. Hist., vol. 61, art. 7, pp. 227-395. Burt, Charles E. and Ceorge S. ^NTyers 1942. Neotropical lizards in the collection of the natural history mu- seum of Stanford University. Stanford Univ. Publ., Univ. ser., Biol. Sci., vol. 8, no. 2, pp. 1-52. Cochran, Doris Mable 19(51. Type specimens of reptiles and amphibians in the United States National Museum. Bull. United States Nat. Mus., no. 220, pp. i-xv-f 1-291. Dunn, Em mett Eeid 1923. Some snakes from northwesti'rn Peru. Proc. Biol. Soc. Wash- ington, vol. 36, pp. 185-188. Gans, Carl 1962. Eedefinition and description of the Brasilian reptiles Amphis- baena silveslrii Boulenger and A. neglecta Dunn and Piatt. Copeia, no. 1, pp. 164-170. Gans, Carl and A. Allan Alexander 1962. Studies on amphisl)aenid8 (Amphisbaenia, Reptilia). 2. On the amphisbaenids of the Antilles. Bull. Mus. Comp. Zool., vol. 128, no. 3, pp. 65-158. Marx, Hymen 1958. Catalogue of type specim.ens of reptiles and amphibians in the Chicago Natural History Museum. Fieldiana : Zool., vol. 36, no. 4, pp. 406-496. Noble, Gladwvn Kingsley 1921. Two new lizards from northwestern Peru. Ann. New York Acad. Sci., vol. 29, pp. 141-143. Taylor, Edward H. 1939. Two new species of tlie genus Anomalepis Jan, with a proposal of a new family of snakes. Proc. New England Zool. Club, vol. 17, pp. 87-96. 12 BREVIORA No. 189 Data for specimens of Amphisbaena pericensis Noble Collection annul: SEGMENTS Chin and number Body/Lat/Tail Dors/Vent Segments Cloaca Length AMNH 28501 204+5+ (6) 18 14/18 2-4 4-6-13 132+13 AMNH 28502 200+4+ (6) 16 13-4/18 2(l)-4 4-6-13 146+13.5 AMNH 28503 212+4+ (7) 18 14/18-20 2-3 4-6-12 128+12 BM 1929.6.1.85 - RR 1946.8.31.83 202+4+ (7) 18 14/18 irreg. 4-7-13 140+14 CAS 54614 214+4+ (7) 18 14/18 2-3 4-8-12 150+14 CNHM 16106 198+4+ (7) 19 14/18-19 2(l)-4 4-8-12 134+14 CNHM 16107 207+4+ (7) 18 13-4/18 2(l)-4 4-6-11 101+10 CNHM 73371 203+4+ (7) 18 12-3/18 2(l)-4 4-8-12 110+11 MCZ 14631 208+4+ (7) 19 14/20 3-4 4-8-12 137+14 MCZ 14764 212+5+ (6) 18 12/18 2-3 4-8-12 153+15 MCZ 14765 200+4+ (6)x 14/18 3-4 4-7-12 104+ (3) MCZ 14767 206+4+ (7) 18 14/18 2-4 4-8-14 77 + 8 MCZ 14768 199+4+ (6) 18 12-14/18 2-4 4-7-12 103+10 MCZ 14770 200+4+ (6) 17 14/16-8 2-4 4-8-13 122+12 MCZ 14772 206+4+ (7) 19 4 12-3/17-8 2-3 4-7-12 127+12 MOZ 14773 207+1+ (7) 19 13-4/18 2-3 4-8-15 135+14 MCZ 14774 212+4+ (7) 18 12-4/18 2-4 4-8-12 86+8 MCZ 14775 215+4+ (6)x 14/18 2-4 4-7-12 142+6 SMF 11826 215+4+ (7) 19 14/16-8 2(l)-4 4-7-13 122+13 SMF 11887 206+i+(7)18 14/18 2-3 4-8-11 92+10 SMF 11888 206+4+f— )- 13-4/18 2(2)-5 4-8-13 127+11 UIMNH 41494 210+4+ (7) 18 13-4/17-8 2-4 4-8-12 84+8 USNM 75970 207 + 4+ (7)18 12-4/18 2(l)-4 4-7-13 101+10 ZMU 29659 199+4+ (6) 18 13-4/18 2(l)-3 4-6-12 122+12 MCZ 14789 217+3+ (6) 17 16/18 2(l)-4 4-8-13 133+13 MCZ 14790 217+3+ (7) 19 15-6/18-20 2(l)-4 4-7-12 130+13 USNM 59926 201+4+ (8) 19 J 13-5/18 2-3 4-8-12 143+16 USNM 60057 218+4+ (7) X 14-5/20 2(l)-4 4-8-14 142+ (5) USNM 60058 206+4+ (7) 18 12-4/18 2-3 4-8-11 126+12 1963 AMPHISBAENA PERICENSIS 13 Present status of the ^r.C.Z. imiatyix's of A. pcriccn.sifi. Orif/inal rnimber 14764-65 14766 14767-68 14769 14770 14771 14772-75 14776 14777-79 14780 Prrsoif iiinvhrr Same "Amaral" = D.Z. ? 8ame Z.M.U. 29659 Same S.M.F. 11826 Same U.I.M.N.H. 41494 C.N.H.M. 16106-07, 73371 B.M. 1929.6. 1.85-EE 1946. 8.31.83 14781-85 = Anomalepis M.C.Z. and U.S.N. M. 14786 U.S.N.M. 75970 14787-88 S.M.F. 11887-88 14789-90 Same u BREVIORA No. 189 Fig. 6. AnipliislKK nn pcricriisis. Dors.-il, l.-itcinl and ventral views of the head of the holotype, M.C.Z. U(i31. 1963 AAll'lllSUAKNA I'ERR'KNSIS 15 Fig. 7. AmpJiisbaetia pericensis. Dorsal (left^ and ventral (liglit) views of the holotype at midbofly to show size and pigmentation of segments. Fig. 8. Amphishaena periccnsis. Ventral view of cdoaca and tail of the holotype, to show pigmentation, segment arrangement and the hardly ap- l)arent autotomj' constriction. BREVIORA Musemini of Coimparative Zoology Cambridge, IMass. September 30, 1963 Number 190 CHARACTERS AND SYNONYMIES AMONG THE GENERA OP ANTS. PART III. SOME MEMBERS OP THE TRIBE PONERINI (PONERINAE, PORMICIDAE) By William L. Brown, Jr. Department of Entomology, Cornell University INTRODUCTION The present part^ is concerned with some genera and species in the tribe Ponerini of subfamily Ponerinae. It consists mainly of the justification for some rather radical revisionary changes in the tribe at the generic level. It is felt that these findings should be presented here in order that they should not distract from the "final" reclassification results eventually to be presented in synoptic form for the Ponerinae and for all Pormicidae. This paper by no means exhausts the necessary changes that must be made in the generic classification of the Ponerini, but it deals with some of the most fundamental ones that require more than sunnnary proposal. Some synonymies proposed or suggested at the species level are to be considered a by-product of the re- search into generic limits. At various stages during the last two years, the work on ponerine reclassification has been sup- ported by grants from the Bache Pund of the National Academy of Sciences, from the Society of the Sigma Xi, and from the National Science Foundation (G-23680). This help is gratefully acknowledged. SOME OLD AND NEW GENERIC CHARACTERS OF PONERINI Tribe Ponerini, the largest in subfamily Ponerinae, has under- gone no basic classificatory changes since the appearance of 1 Part I. Breviora, Mus. Comp. Zool., no. 11 : 1-13 (19.3.3). Part II. Ibid., no. 18 : 1-8 (1953). 2 BREVIORA NO. 190 Emery's fascicle covering the subfamily in the Genera Insectorum in 1911. Emery's treatment was really a refinement of his earlier reclassification of the ponerines published in 1901. In his well- known key to the genera of ants, Wheeler (1922) followed Emery's classification of the Ponerini for the most part, but Avorked in the additional genera described up to 1922. The Emery classification employed a number of characters of worker and female in distinguishing his genera of tribe Ponerini, but the two most stressed were (1) the number of segments in the maxillary and labial palpi, and (2) the number of tibial spurs on the middle and hind legs. Unfortunately, the palpal character saw extensive use only during some of the early years of ant systematics. Emery, Mayr, and even Frederick Smith often gave the count of segments in their descriptions, but curiously, after Emery used the palpal counts as a foundation stone of his 1901 classification, these ap- pendages were almost completely ignored by most later describers of ponerine species, probably due to the fact that the smaller mouthparts are so small and difficult of access. Later authors seem to have relied more on habitus than on truly diagnostic characters when they placed new (or supposed new) species to genus. As a result, some species have been described as new over and over again in different genera and subgenera (consult the case of "Trachymesopiis" hrunoi and its s^alonyms, discussed below). The primitive number of worker-female palpal segments appears to be 4, 4 (4 maxillary, 4 labial — the formula always stated in that order) ; in the male it is 6, 4, the basic number for all ants. In many Ponerini, especially the smaller or crypto- biotic species and their derivatives, the number is much reduced, and reaches as low as 0, 1 in the worker, while less consistent reduction occurs in the male. Palpal segmentation is undoubtedly important in generic classification and in the analysis of phyletic trends, but due to its neglect in species descriptions, it remains unknown for a majority of forms. "Work is now going forward to correct this deficiency, and some results are incorporated in the present paper. The other major character in question is the number and state of development of the tibial spurs. The trait is an ambigu- ous one, as can be seen from its employment in keys. Wheeler's 1922 key to the genera of the Ponerini states it this way in the very first couplet : Middle and hind til)iae with two spurs 2. Middle and liind til)iae with a single, well-developed spur, which is always pectinate; the lateral spur is rudimentary or absent 19. ANT CHARACTERS AND SYNONYMIES 3 This couplot really expresses the difference thought to separate the ''lower" Ponerini, mostly large in size, from the "higher" members, which tend to be smaller (e.g., Ponera, Cryptopone) . Generally speaking, the genera with smaller-sized species do tend to lose the lateral spurs of the middle and hind tibiae, while genera having larger-sized species normally retain the extra spur, although usually in a more or less reduced condition. The diffi- culty, of course, lies with distinguishing between the condition "two spurs" and "lateral spur rudimentary," especially when one finds that in most species the lateral spur is considerably smaller than its mate. Furthermore, the threshold at which a "vestigial spur" becomes just another apical seta of the tibia is unspecified along a gradual morphocline of species. In short, the lateral tibial spur, as an allometric character, cannot be used to split the Ponerini into two main groups, and in fact probably cannot even be used by itself as a diagnostic character at the generic level. Other adult characters of value in classification still remain: the shape of the clypeus has been and still is an important generic character. The tarsal claws may be simple, or may have one or more teeth, or may even be pectinate, as in the case of most Leptogenys species. I consider the pectination of the claws, imperfect in some African species of the genus, to be insufficient as a tribal character in view of the several other strong char- acters shared by both adult and larval Ponerini and Leptogenys, and accordingly I am placing tribe Leptogenyini as an included synonym of Ponerini (new synonymy). Another character of some importance is the presence of heavy, conical, spine-like setae on the extensor surfaces of the middle tibiae and tarsi. These structures, which appear to promote movement through soil or rotten wood by improving traction, are found in certain cryptobiotic genera (e. g., Centromyrmex, Wadeura, Cryptopone, Promyopias) as well as in a number of fossorial wasps (many pristocerine Bethylidae, most non- parasitic Scolioidea, for instance), all of which seek their prey underground or in other confined circumstances. Emery cited the presence of such spine-like, as well as merely bristle-like, setae as characteristic of Euponera subgenus Trackymesopus, which he named (in 1911) accordingly. The heterogeneity of Trachymesopus in this regard, as well as in other respects, made it a perfect catch-all for miscellaneous species of medium- to-small Ponerini from the very beginning, and it has continued in this role right down to the present. As a matter of fact. 4 BREVIORA NO. 190 heavy, conical spine-like setae that arise over half or more of the mid-tibial extensor surface are found only in a particular group of "Trachymesopus," and this group {ochracea and allies) shares this and several other characters with the mem- bers of Cri/ptoponc. The relationship between Trachymesopus and Cryptopone has long been discussed by E. 0. Wilson and myself (see Wilson, 1958: 352), and now the discovery of the concordance between the tibial armament and other characters, particularly the basal mandibular pit, makes the solution of this problem obvious. The "Trachymcsopns" species with these char- acters are all really Cryptopone, and are transferred accordingly, as discussed under that genus below. One of the chief Cryptopone characters just mentioned is a particularly interesting one ; this is a prominent oval pit or fovea near the base of the mandible in its dorsolateral surface. This pit, or its obvious homolog, is also found in all Brachy- ponera, in the members of the "Trachymesopus" sharpi group, in Hagcnsia, in Euponera sikorae Forel (type species of Eu- ponera!), and in a few other African species formerly placed in Euponera or other genera, but is lacking in T. stigma, the type species of Trachymesopus, and its closest relatives. All known species with the basal pit are from the Old World, and chiefly from Africa, with the exception of Cryptopone gilva and (per- haps) C. guatemalensis, which apparently represent a Crypto- pone invasion of the New World that has spread through North America and reached Central America. In the other direction, Cryptopone has reached southeastern Australia and New Cale- donia, but curiously, no true Cryptopone are known to occur in Ethiopian Africa. The basal mandibular pit in ponerine species was early noted as a character by Gustav Mayr and occasionally mentioned by later authors in species descriptions, but only Arnold used it as a generic character in his 1951 review of Hagensia, a genus in which the pits are unusually distinct. In most species that have it, the pit has been overlooked completely in descriptions, par- ticularly after 1900. Many Ponerini, especially among the larger and presumably more primitive species, bear another character in the form of an oblique groove across the dorsal face of the mandible, curving outward from the inner base and usually continuing into the lateral longitudinal mandibular sulcus that runs to the apex. This channel, which may be functioning to ANT CHARACTERS AND SYNONYMIES 5 distribute some product from the mandibular glands, is prob- ably not homologous with the basal pit. It is oriented differently, and at least one species of " Bothroponera" has both the pit and the groove. In the former Trachymesopus, species without the truly spine- like mid -tibial setae (except for 2-3 of these setae at the tibial apex) are divided into three groups: the stigma and darwini groups, which have no basal mandibular pit and no anal lobe on the hind wing of the sexes, and the sharpi group, which has the pit, and the lobe in the male only, so far as known. The sharpi group is close to Brachyponera, with which it shares the pit and the lobe, but further information may dictate a separate genus for each of these groups. In what remains of T r achy m^eso pits, the stigma group (e. g., stigma, cautiis, pachynodiis, rufonigrus) has palpal segmenta- tion 3, 3, while darwini has 4, 3. Thus, it may be seen that Trachymesopus is a heterogeneous grouping even after the re- moval of the Cryptopone species ; its reclassification will have to await the study of more of the larvae and adult winged forms. Among the most promising of characters to be used in ponerine systematics are those of the larvae. The Wheelers (1952) have laid the groundwork for a comparative study of the larvae of the genera of Ponerini, but for the great majority of species of the tribe, the larvae have never been studied. The larvae of Ponerini are usuall}^ covered with peculiar medullate projections, called by the AVheelers ''tubercles." In consonance with the morphological terminology applied to other holometabolous larvae, particularly Lepidoptera, I propose that these projections be called by the more specifically descriptive term chalaza (sing.), chalazae (pL). The chalazae of Ponerini are matched by ap- parently homologous structures in tribes Thaumatomyrmicini and Odontomachini, which are close to Ponerini on the basis of adult characters as well. Some of the smaller ponerine genera bear special paired mushroom-shaped chalazae on one or more abdominal tergites; these have long been known to function as "hangers" by which the larvae are stuck to the ceiling and walls of the nest by a glutinous substance covering the head of the chalaza. The number and placement of these fungiform chalazae is important in generic taxonomy, but they must be used with care owing to the fact that they may change in number and form, or be lost altogether, as the instars metamor- phose one into another. 6 BREVIORA NO. 190 REVISIONARY OBSERVATIONS ON SOME PARTICULAR GENERA Cryptopone Emery Worker : Small in size, generally under 4 mm total length, depigmented (ferruginous or yellowish) ; eyes absent or ex- tremely reduced, body compact ; mandibles downcurved, with a few (4-6) coarse teeth set on more or less oblique masticatory borders. Dorsolateral mandibular surface with a conspicuous basal pit or fovea, circular or elliptical in outline. Middle tibiae with stout spinules covering about % or more of their extensor surfaces. Palpal segmentation 2, 2 or less. Female : Aside from well-developed compound eyes, ocelli, thorax and the other obviously female characters, fit the worker diagnosis. Hind wings without anal lobe, but with two basal cells. Male : Small, slender, dark-colored, with pygidial spine so common among Ponerini. Hind wing without anal lobe. Species: Those placed here by Wilson (1958), plus the former Trachymesopvs species crassicornis (Emery), gilva (Roger), probal)ly guatemalensis (Forel), ochracea (Mayr), rotundiceps (Emery), sauteri (Wheeler), taivanae (Forel) and its probable junior synonym takahashii (Wheeler) . Of species formerly placed in Poncra, P. typhla (Karawajew) is clearly a Cryptopone, and is probably a synonym of C. testacea (Emery) ; the Australian P. mjoehergi Forel is a junior synonym of C. rotundiceps (new synonymy). Among species to be deleted from the Cryptopone roll, C. rufofcstaceus Donisthorpe belongs in Trachymesopus as that genus is presently constituted, and is the same as the large variant of T. darivini hitherto known as T. lamarki Santschi (new synonymy). True Cryptopone has not been taken in Africa south of the Sahara, and C. angustata Santschi (type examined) and C. hartwigi Arnold should be transferred to P oner a as that genus is currently constituted. I have not checked this list of species for exhaustiveness, but it should serve to give a general idea of Cryptopone in the sense of the present work. Trachymesopus Emery As already explained above, Trachymesopus (without the Cryptopone species) contains at least three groups of species. The type species, T. stigma, is common and widespread in both the Indo-Australian region and in the warmer parts of the Ncav World. The Old World synonymy is extensive, and has been ANT CHARACTERS AND SYNONYMIES 7 dealt with by Wilson (1958:355). In the New World, the synom-my of stigma is not yet clear. T. cognata (Emery) has never been differentiated satisfactorily, and T. sticcedanea (Roger) could be either stigma or the closely related coAitus Mann. T. compressinodis Boro-meier is a synonym of T. cautus Mann (types compared ; new synonymy) . Agreeing with stigma and cautus in palpal segmentation (3, 3) and in the lack of an anal lobe on the hind wings of both sexes, are two Australian species, rufonigra (Clark), transferred from Brachyponera, and pachynodus Clark (wings unknown in the latter). Two other species that apparently are close to this group are lunaris (Emery) and ferrugineus (F. Smith), though so little is known about these forms now that they cannot be confidently assigned. The second group consists of darwini and relatives. These are medium-small species known only from the ferruginous-colored females, which are taken commonly at light throughout wide areas of the Old World tropics. These females lack an anal lobe on the hind wing, and their palpal formula is 4, 3. They lack a mandibular pit and have no spine-like setae on the outer face of the middle tibia. They are found from northern Aus- tralia, Indonesia and the Philippines westward through India to tropical Africa. They vary considerably in size, but the varia- tion may be continuous; in fact, all of these forms may well represent a single species. As mentioned above, the species de- scribed as Cryptopone rufotestacea by Donisthorpe belongs here and is a synonym of T. lainarki Santschi, which in all likelihood is only a size variant of darwini. It seems likely also that Motschulsky's Amhlyopone testacea belongs in this complex. The third group is the group of sharpi Forel, which consists of medium-small species, all castes blackish in color, with a dis- tinct elliptical mandibular pit; palpal segmentation (as seen in 2 African specimens of hrunoi only) is 4, 4. Anal lobe lacking in 2 hrunoi females from Liberia, but present (though separated by a relatively shallow cleft) in a single male from Southern Rhodesia; perhaps in this group the character is variable or sexually dimorphic. Mid-tibiae without spine-like setae on ex- tensor faces. Some of the species {hrunoi, malayanus, katangana) were de- scribed originally in Ectomomyrmex, or were placed in that genus later. Actually, these species are all very close to sharpi — in fact, malayanus is almost certainly a straight synonym of sharpi. The form of sharpi from China and Japan has the propodeum 8 BREVIORA NO. 190 wider and more opaque above ; it was described by Wheeler as subspecies pilosior, but should be considered as a good species until more is known about the variation in this complex, which, if it matches the variation found in hrunoi in Africa, may be considerable. T. pilosior occurs westward to the scarp of the Tibetan Plateau, where I collected a worker on Mt. Omei, Szechuan Province, in 1945; Euponera (Trackymesopus) cJwsonensis Teranishi, 1940, from Osaka, Japan, is its new synonym. The Oriental and African forms of this group are also very close, and deserve careful comparison. The sharpi group does not really belong to Trackymesopus, and will eventu- ally have to be moved out. But in order to avoid one additional round of name changes, I am leaving it where it is until the limits of other closely related groups, especially Bracliyponera, are clarified. Trackymesopus brunoi comb. nov. Pachycondyla (Ectoviomyrmex) hrunoi Forel, 1913, Deutsche ent. Zeitsehr., beih., p. 20."), worker. Type loc: Bulawayo, S. Ehodesia. Syutype ex- amined. Euponera (Trachymesopus) nigeriensis Santschi, 1914, Boll. Lab. Zool. Portici, 8: 316, worker. Type loc: Olokomeji, Nigeria; also from Aburi, Gold Coast. Syntype examined. New synonymy. Euponera (Trachymcsopus) hayoni Menozzi, 1932, Ann'. Mus. Civ. Stor. Nat. Genova, 56: 97-98, worker. Type loc: Kome, Sesse Archipelago, Vic- toria Nyanza, Uganda. New synonymy. Ectomomyrmex nigeriensis var. Icatangana Santschi, 1933, Bull. Ann. Soc Ent. Belg., 73: 96, Type loc: Pweto, Elisabethville, Belgian Congo. Syntype examined. New synonymy. Euponera (Mesoponcra) dentis Weber, 1942, Proc. Ent. Soc Wash., 44: 43, fig. 9, dealate female. Type loc: Lotti Forest, Imatong Mts., Sudan. Type examined. New synonymy. Euponera {Trackymesopus) lamotiei Bernard, 1953, Mem. Inst. Fr. Afrique Noire, 19 (1): 195, fig. 2g, h, i, dealate female. Type loc: Keoulenta Savannah, French Guinea. New synonymy. This species occurs widely in Africa south of the Sahara. In addition to the types mentioned above, I have seen samples from Liberia, Cameroons and the Kalahari Desert. There is some small variation among these samples in relative head v.idth and in strength and opacity of the sculpture, but this appears to be both size-linked (allometric) and continuous. ANT CHARACTERS AND SYNONYMIES 9 EcTOMOMYRMEX Mayr The taxonomy of this genus has been consideral)ly siniplitied at the species level by the partial revisions of Wilson (1958) and Yasumatsu (1962), but an even more fundamental change needed is the removal of the species hrunoi and malayanus. These, as has been demonstrated above, are members of the sharpi group of "Trachymcsopus." AVith these deletions, Ec- tomomyrmex is once more restored to the status of a strictly Indo-Pacific genus, ranging from India and northwestern China eastward to Japan and Korea, Samoa and northern Queensland. It is now also possible to define the genus. Ecfomomyt-mcx consists of medium-sized to large, usually basically black-colored forms, the workers and females of Avhich tend to have the posterior cranium somewhat prismatic, and the posterior face of the petiolar node strongly striate, or at least rugose-punctate. Worker-female palpi segmented, so far as is known, 4, 4; mandibles without basal pit, the oblique basal groove present and weak, or absent. The oblique mesepisternal suture is present and reasonably distinct in the worker. Com- pound eyes fairly to rather well developed and multifacetted in the worker. Centromyrmex Mayr Centromyrmex Mayr, 1866, Verb, zool.-bot. Ges. Wien, 16: 894. Type: Centromyrmex hohemanni Mayr, 1866, monobasic. Spalacomyrmex Emery, 1889, Ann. Mus. Civ. Stor. Nat. Gcnova, 27: 489. Type: Spalacomyrmex feae Emery, 1889, monobasic. Typliloteras Karawajew, 1925, Konowia, 4: 128. Type: Typhloteras hamu- latum Karawajew, 1925, monobasic. Glyphopone Forel, 1913, Kev. Zool. Afr. 2: 308. Type: Glyphopone he- quaerti Forel, 1913, monobasic. New synonymy. Glyphopone subgenus Leptopone Arnold, 1916, Ann. S. Afr. Mus., 14: 163. Type: GlypJwpone (Leptopone) rufigaster Arnold, 1916, monobasic. New synonymy. The type female of Glyphopone bequaerti has been compared directly with a winged female kindly sent by Dr. Arnold (Abercorn, Northern Rhodesia, 10-12-1943, Arnold leg. et det.). These two specimens are virtually identical and are surely conspecific, as indicated in the formal synonymy expressed below. The median lobe of the clypeus is hardly to be considered "bi- carinate," but the median portion of its surface is very slightly concave when viewed in the proper light. Although these are 10 brevioka no. 190 large females, with dark forebody and rufous gaster, they bear the general characters of Cenfromyrmcx, and in any ease, they are probably no larger or darker than the female of C. gigas, judging from the workers of gigas I have seen. It seems best to consider Glyphopone a synonym of Centromyrmex until the worker of G. hequaerti is found and we are able to determine whether it has the definitive characters. Centromyrmex bequaerti Forel comb. nov. Glyphopone Bequaerti Forel, 1913, Rev. Zool. Afr., 2: 308, fig. 1, alate female. Type loe. : Kibombo, Belgian Congo ; examined. Glyphopone {Leptopone) rufigaster Arnold, 1916, Ann. S. Afr. Mus., 14: 163, figs. 10, 10a, alate female. New synonymy. Pro MYOPIAS silvestrii Santschi Myopias (Promyopias) Silvestrii Santschi, 1914: 324, fig. 10, worker. Type loc. : Mamou, French Guinea ; one syntype worker examined. Promyopias asili Crawley, 1916, Entomologist, London, p. 30, fig., alate female. Type loc: " Xyasaland, " [according to label on holotype: Mlanje, Nyasaland, 15-IV-1913, S. A. Neave leg.]. Holotype in British Museum, examined. New synonymy. REFERENCES Arnold, G. 1951. The genus Hagensia Forel (Formicidae). Jour. Ent. Soc. S. Afr., 14: 53-56. Emery, C. 1901. Notes sur les sous-families des dorylines et ponerines. Ann. Soc. Ent. Belg., 45: 32-54. 1911. Formicidae subfamily Ponerinae. Genera Insectorum, fasc. 118. Teranishi, C. 1940. Teranishi Memorial Volume. Posthumous works, p. 8. Wheeler, G. C. and J. Wheeler 1952. The ant larvae of the subfamily Ponerinae. — Part II. Amer. Midi. Nat., 48 : 604-672, 6 pis. Wheeler, W. M. 1922. Keys to the genera and subgenera of ants. Bull. Amer. Mus. Nat. Hist., 45: 631-710. Wilson, E. O. 1958. Studies on the ant fauna of Melanesia. IV. The tribe Ponerini. Bull. Mus. Comp. Zool., Harvard, 119: 320-371. Yasumatsu, K. 1962. Notes on synonymies of five ants widely spread in the Orient. Mushi, 36: 93-97. BREVIORA Mnjiseitiinii of Cojmparative Zoology Cambridge, Mass. December 5, H)(5o Number 191 THREE NEW SPECIES OF MANGORA (ARANEAE, ARGIOPIDAE) FROM CENTRAL AMERICA By Arthur M. Chickering In my paper dealing with the genus M angora in Panama (1954) I recognized eight species then known to exist in that country. This conclusion involved the recognition and description of three new species together with the establishment of considerable new synonymy. During my collecting trips back to the Panana Canal Zone and parts of Panama outside of the Canal Zone in 1954 and 1957-1958 a considerable number of specimens belonging to this genus were taken but no new species were found among these. I have also had the opportunity to study all of the Mangora collections from Central America now in the Museum of Comparative Zoology at Harvard College. Some of this material has only recently been sorted out of general collections which have been awaiting atten- tion for many years. From all of this material I have been able to separate out what I believe to be representatives of three new species. These are described in this paper in accord with my usual formula. If recently established new synonymy is taken into considera- tion together with the new species recognized in this paper, the complete list of species in the genus Mangora 0. P. -Cambridge, 1889, now definitely known from Central America, may be given as follows: M. bimaculata (0. P.-Cambridge) ; M. calcarifera F. P.- Cambridge; M. Candida Chickering; M. conspicua sp. nov. ; M. distincta sp. nov.; M. mobilis (0. P.-Cambridge); M. montana Chickering; M. passiva (0. P.-Cambridge) ; M. pia Chamberlin and Ivie; M. picta 0. P.-Cambridge; M. schnierlai Chickering; M. spinula F. P.-Cambridge; M. sufflava sp. nov.; M. trilineata 0. P.- Cambridge. Banks (1898) reported M. gibberosa (Hentz) from LowTr California. I have not been able to examine these specimens but I regard it as highly unlikely that the species exists in that 2 BREVIORA No. 191 part of Central America and, for this reason, the species is not included in the list given above. Kraus (1955) reported M. gib- berosa (Hentz) from El Salvador. Again, I think there is serious question regarding the validity of this identification. I am rather strongly inclined to believe it belongs to the species which I am regarding as M. conspicua sp. nov. but, in view of the uncertainty, I am omitting it from my list. Only males are known for M. distincta sp. nov., M. picta 0. P.-Cambridge, and M. sufflava sp. nov. Only females are known for M. passiva (0. P.-Cambridge), M. schnierlai Chickering, and M. trilineata 0. P.-Cambridge. All types established in this paper together with my entire collec- tion of Araneae are deposited in the Museum of Comparative Zoology at Harvard College. Genus Mangora O. P.-Cambridge, 1889 Mangora conspicua sp. nov. (Figures 1-7) Female holotype. Total length 3.77 mm. Carapace 1.408 mm. long; .99 mm. wide opposite interval between second and third coxae where it is widest; .66 mm. tall at level of greatest width and, therefore, two-thirds as tall as wide; median fovea a narrow, longitudinal groove. Eyes. Eight in two rows as usual. LE on very slightly raised tubercles; AME protrude somewhat over clypeus. Viewed from above, anterior row strongly recurved, posterior row slightly so. Viewed from in front, posterior row procurved, anterior row re- curved, all measured by centers. Central ocular quadrangle wider in front than behind in ratio of about 20 : 17; slightly longer than wide in front. Ratio of eyes AME : ALE : PME : PLE = 8 : 5.5 : 6.5 : 6. AME separated from one another by about 1.25 times their diameter, from ALE by slightly less than that distance. PME separated from one another by aljout ten-thirteenths of their diameter, from PLE by eleven-thirteenths of their diameter. Height of clypeus equal to about the radius of AME. Chelicerae. Typical of the genus. Promargin of the fang groove apparently with three teeth ; retromargin apparently with only two. Maxillae and Lip. Typical of the genus; with details regarded as unnoteworthy. Sternum. Quite convex; longer than wide in ratio of about 13 : 11; sternal suture strongly procurved; not directly continued between fourth coxae but a sclerite is present there ; fourth coxae separated by nearly two-thirds of their width. 1963 MANGORA FROM CENTRAL AMERICA l^il / / \ f-\\\4+\4l External Anatomy of Mangora Figures 1-7, M . conspicua Figs. 1-3. Epigynum from below, in posterior view, and in profile from right side, respectively. Fig. 4. Epigynum from a paratype, turned forward to reveal the dorsal surface. Fig. 5. Left second tibia of male allotype, showing ventral spines. Fig.s. 6-7. Two views of the male palpal tibia and tarsus. Legs. 1423. Width of first patella at "knee" .217 mm., tibial index of first leg 15. Width of fourth patella at "knee" .227 mm., tibial index of fourth leg 16. 4 BREVIORA No. 191 Femora Patellae Tibiae Metatarsi Tarsi Totals (All measurements in mm.) 1. 1.474 .440 1.050 1.320 .638 4.922 2. 1.350 .440 .924 1.166 .610 4.490 3. .748 .330 .528 .638 .438 2.682 4. 1.474 .450 .968 1.122 .550 4.564 All legs with fairly robust spines. The long, branched hairs on the third tibiae appear to be arranged in two oblique rows of five and four, respectively, but short trichobothria are closely contiguous. Tarsal and palpal claws appear as usual in the genus. Abdomen. 2.795 mm. long; 1.826 mm. wide; ovoid; with at- tachment to cephalothorax only slightly anterior to middle ; other- wise as usual in the genus. Epigynum. General pattern resembles that of M. mobilis (O. P.-Cambridge) and M. passiva (O. P. -Cambridge) but the central tongue is longer than in either of these species and other differences are also visible. Features as shown in Figures 1-4. Color in alcohol. This is a very clearly marked species and all three females available are remarkably consistent in their color pattern. Carapace yellowish with a narrow, median, black stripe reaching from just behind PME through the median groove where it is extremely narrow; there is a marginal stripe beginning a little behind PLE and extending to about opposite the second coxae where it widens and continues nearly to the posterior border as a gray area. The sternum is dark brown with a light spot in the mid- dle just behind the sternal suture. The legs are basically yellowish but arc rather conspicuously spotted because each spine is sur- rounded at its base by a small black dot. The first and second fem- ora each have a ventral black dot on the membrane at the joint be- tween the trochanter and the femur, then an irregular black stripe reaches nearly to the distal end where there is a curved black band ; there is also an indistinct gray stripe on both of these femora just prolateral to the black stripe. The gray stripe is lacking on some paratypes but the fourth femur has a dark, gray, retrolateral stripe extending through the middle of the segment. The abdomen is also conspicuously colored by numerous small white flecks and dark brown or black stripes. Beginning about one-third from the base there is a narrow median, brownish stripe and on each side of this there is an irregular black stripe. White dominates the whole dor- sum except for these stripes. The lateral sides each have four nar- row, irregular, dark brown or black, oblique stripes extending to the venter. The latter part is yellowish anterior to the epigynum; 1963 M ANGORA FROM CENTRAL AMERICA 5 behind the epigynum there is a broad, median stripe with a group of white flecks on each side. There are also three small, white spots on each side of the group of spinnerets. Allotype male. Three males apparently belonging to this spe- cies are in the collection but they are all defective for one reason or another. For this reason the female has been chosen as the holotype and a male with normal palps as the allotype. The abdo- men of the allotype male was crushed and of little value. Total length of a paratypc male 2.134 mm. Carapace of allotype male 1.122 mm. long; .718 mm. wide opposite interval between second and third coxae where it is widest; about .44 mm. tall. Otherwise essentially as in female. Eyes. Essentially as in female; details regarded as unnote- worthy. Chelicerae. Difficult to see teeth along fang groove ; apparently the number is the same as in female. Maxillae, Lip, and Sternum. Each maxilla has a tooth w^hich is in apparent opposition to a chitinized tubercle on the palpal femur ; otherwise these structures are essentially as in female. Legs. 1243. Width of first patella at "knee" .162 mm., tibial index of first leg 13. Width of fourth patella at "knee" .152 mm., tibial index of fourth leg 15. Femora Patellae Tibiae Metatarsi Tarsi Totals (All measurements in mm.) 1. 1.232 .418 .880 1.056 .594 4.180 2. 1.056 .360 .682 .990 .550 3.638 3. .638 .242 .396 .506 .374 2.156 4. 1.034 .330 .704 .792 .462 3.322 Palp .220 .120 .098 .462 .900 Spines on legs are considerably more conspicuous and much longer than in females. The curled hairs on the third tibiae ap- parently are arranged in two oblique rows of four and three, re- spectively. Additional trichobothria are also present but their exact distribution has not been determined. The second tibia has specialized spines; those seen in ventral view are represented in Figure 5. The first coxa has the ventral hook and the second femur has the corresponding groove and ridge. Palp. Complicated; parts are small and generally inconspicu- ous but the clavis is clearly visible. Both patella and tibia are short; the latter is broad, angular, and very irregular in outline (Figs. 6-7). The femur has the chitinized tubercle as already stated. 6 BREVIORA No. 191 Abdomen. This part of the body is badly crushed and unsuit- able for study ; apparently it is essentially like that of the female. Color in alcohol. The color in general is similar to that of the female holotype but it is less conspicuous. The marginal stripes on the carapace of the female are merely indistinct grayish areas in the male allotype. The sternum is nearly surrounded by a dark gray margin with the greater part yellowish with small grayish areas. The legs are much as they appear in the female but the black dots at the bases of the spines are less conspicuous and the prolateral, grayish stripes on the first and second femora are absent in the male. Abdomen: there are numerous white, chalky flecks on the dorsum; there are two black dots in the middle line in the anterior half of the dorsum; the posterior half of the dorsum con- tains a dark, rectangular, fairly broad spot composed of a series of black marginal spots more or less connected by very narrow, black lines extending through a brownish area. Type locality. The holotype female is from Mexico, Nuevo Leon, El Potosi, Cerro Potosi, June 13, 1938 (H. Hoogstraal). One paratype was taken with the holotype. Another in the collection is from the same general region, Sabinas Hidalgo, 1000 ft. elevation, June 13, 1940. Three females in the collection are from Honduras, Escuela Agr. Panam., 27 km. S. of Tegucigalpa, San Antonio del Oriente. Two males accompany the females. One was mature; it had undergone severe injury to its abdomen but was selected as the allotype; the other was immature. The Honduran specimens were collected by A. and M. Carr at 3800-4000 ft. elevation, No- vember 17, 1945. Two other males in the collection are from Mexico: Nuevo Leon, Villa de Santiago, Hacienda Vista Hermosa, 2000 ft. elevation, July 19, 1940 (H. Hoogstraal). Both of these males have their tarsal bulbs so distended as to be useless for description. Mangora distincta sp. nov. (Figures 8-11) Male holotype. Total length 2.535 mm. Carapace 1.202 mm. long, .902 mm. wide opposite interval between second and third coxae where it is widest, .484 mm. tall and, therefore, about .54 as tall as wide; gently arched with tallest point nearly at level of greatest width; median fovea a rather deep, narrow, longitudinal groove. Eyes. Eight in two rows as usual; LE on slightly raised tu- bercles; AME protrude i)rominently over clypeus. Viewed from 1963 MANGORA FROM CENTRAL AMERICA above, anterior row strongly recurved, posterior row slightly so. Viewed from in front, anterior row slightly recurved, posterior row moderately procurved, all measured by centers. Central ocular quadrangle wider in front than behind in ratio of 19 : 16; slightly External Anatomy of Mangora Figures 8-11, M. distincta Figure.s 12-15, M. sufflava Fig. 8. Left second tibia to show ventral spines. Figs. 9-10. Two views of the palpal tarsus. Fig. 11. Tarsal clavis. Fig. 12. Left second tibia to show ventral spines. Figs. 13-14. Two views of the palpal tarsus. Fig. 15. Tarsal clavis enlarged. 8 BREVIORA No. 191 longer than wide in front. Ratio of eyes AME : ALE : PME : PLE = 6:5:6:5. AME separated from one another by their diameter, from ALE by slightly less than this distance. PME separated from one another by slightly more than their radius, from PLE by their diameter. LE separated only by a line. Height of clypeus about equal to the diameter of AME (surrounding pig- ment makes it difficult to obtain exact measurements of eyes and their relative positions). Chelicerae. General features as usual in the genus. Unable to observe teeth along fang groove without injury to holotype. Maxillae and Lip. Maxillae with a sharply pointed tooth near their bases ; probably used in opposition to the chitinized tubercle on the palpal femur. In general, both structures appear to be typi- cal of the genus. Sternum. Scutiform; strongly convex; somewhat depressed in anterior portion; sternal suture quite procurved; widest between second coxae where it is nearly as wide as long; continued broadly between fourth coxae which are separated by 8/11 of their width. Abdomen. Ovoid; quite typical of the genus; details regarded as unnoteworthy. Color in alcohol. Legs and mouth parts generally yellowish; each chelicera has a small, irregular, grayish area in front near the base. The carapace is yellowish with a narrow central stripe from a little behind PME and extending two-thirds of the distance to the posterior border. The sternum is yellowish with a very irregular light grayish marginal area. Abdomen: yellowish with whitish flecks dorsally and dorsolaterally ; in the posterior half of the dorsum there are three narrow brownish stripes ; in the region of the spinnerets there are several black and white dots; on each lateral side there is a narrow oblique grayish stripe followed by an area of many small white flecks and finally a rounded irregularly grayish spot; the venter is generally yellowish but just in front of the genital groove there is an oval grayish spot and behind the genital groove there is an irregularly grayish, rounded spot con- taining a group of white flecks. Type locality. The holotype is from Honduras, Escuela Agr. Panam., 27 km. S. of Tegucigalpa, San Antonio del Oriente, Nov. 17, 1945, 3800-4000 ft. elevation (A. and M. Carr). There are no paratypes and the female is unknown. 1963 M ANGORA FROM CENTRAL AMERICA 9 Mangora sufflava sp. nov. (Figures 12-15) Male holofype. Total length 2.6 mm. Carapace 1.3 mm. long, .99 mm. wide opposite interval between second and third coxae where it is widest, .638 mm. tall at about the level of its greatest width and, therefore, about half as tall as wide; median fovea a narrow, well defined, longitudinal groove; considerably overlapped by anterior end of abdomen. Other features as usual in the genus. Eyes. Eight in two rows as usual. Viewed from above, anterior row strongly recurved, posterior row moderately so. Viewed from in front, anterior row moderately recurved, posterior row moder- ately procurved, all measured by centers. Central ocular quad- rangle wider in front than behind in ratio of 7 : 6; slightly longer than wide in front. Ratio of eyes AME : ALE : PME : PLE = 7.5 : 6 : 6.5 : 6. AME separated from one another by about their diameter, from ALE by about two-thirds of their diameter. PME separated from one another by about two-thirds of their diameter, from PLE by about 1.3 times their diameter. LE separated only by a broad line. Height of clypeus equal to about four-fifths of the diameter of AME. Chelicerae, Maxillae, Lip and Sternum. All apparently typical of the genus. Legs. 1243. Width of first patella at "knee" .200 mm., tibial index of first leg 13. Width of fourth patella at "knee" .162 mm., tibial index of fourth leg 14 Femora Patellae Tibiae Metatarsi Tarsi Totals (All measurements in mm.) 1. 1.364 .506 .990 1.100 .649 4.609 2. 1.320 .484 .902 1.078 .627 4.411 3. .792 .308 .506 .528 .418 2.552 4. 1.292 .418 .770 .880 .528 3.888 Palp .286 .119 .119 .440 .964 All legs with well developed spines and hairs. Second tibia with modified spines ; those seen in ventral view shown in Figure 12. The long, slender, branched hairs on the prolateral side of third tibia ap- pear to be arranged in two oblique rows of three each. The first coxa bears the usual hook and the second femur has the corresponding prolateral, proximal groove and chitinizcd ridge. Tarsal claws ap- pear to be as usual in the genus. 10 BREVIORA No. 191 Palp. The patella has the usual long, slender, terminal, dorsal spine. Both patella and tibia are short. Important tarsal features shown in Figures 13-15. The clavis appears to be different from any other found in the collection. The tooth is probably present on the maxilla together with the chitinized tubercle on the palpal femur but neither can be clearly seen without danger of damage to the holotype. Abdomen. 1.625 mm. long; 1.170 mm. wide near middle; ovoid. At the base are two long, slender, dorsal, spine-like bristles. Other features typical of the genus. Color in alcohol. Carapace yellowish with a grayish median stripe extending from AME to middle third of steep posterior de- clivity; the center of the grayish stripe is a narrow, black line. The sternum is broadly grayish around the margin but yellowish in the center. The legs and mouth parts are yellowish with many varia- tions in the depth of color. Abdomen: this part of the body has a rather unusual color pattern; dorsolaterally there is a fairly broad, white stripe on each side extending nearly the length of this part of the body; there is a large, median, grayish yellow spot in the anterior half of the dorsum containing a pair of small white dots in the posterior half with another pair of small white dots at its posterior border ; the posterior half of the dorsum has a pair of dark gray stripes separated by a yellowish stripe; these stripes do not reach to the posterior end of the abdomen; the venter is yellowish with faint graj^ irregularities. Type locality. The holotype male is from Boquete, Chiriqui, Panama, August, 1950. This specimen was overlooked among specimens belonging to different genera at the time of completion of my previous paper on Mangora (1954) and only recently found. There are no paratypes and the female is unknown. BIBLIOGRAPHY Banks, Nathan 1898. Ararhnida from Baja California and other parts of Mexico. Proc. California Acad. Sci., ser. 3, Zoology, 1(7) : 205-309, 5 pis. 1929. Spiders from Panama. Bull. Mus. Comp. Zool., 69: 53-96, 4 pis. Bonnet, Pierre 1957. Bibliograiihia Araneorum. 2. 3me partie. Toulouse: Les Artisans de rimprimerie Douladoure. Cambridge, O. P.- and F. P.-Cambridge 1889- Arachnida-Araneida. /?* .• Biologia Centrali-Americana. 1905. Dulau & Co., London. 1963 MANGORA FROM CENTRAL AMERICA 11 ChAMBERLIN, R. V. AND WiLTON IviE 1936. New spiders from Mexico and Panama. Bull. Univ. Utah, 27: no. 5, biol. series, 3, no. 5:3-103, 17 pis. Chickkring, Arthur M. 1954. The spider genus Mangora (Argiopidae) in Panama. Bull. Mas. Comp. Zool., Ill : 195-215, 28 fig.s. Kraus, Otto 1955. Spinnen aus El Salvador (Arachnoidea, Arancae). Abhand. Senck- enberg. naturforsch. Ges., 493: 1-112, 12 pis. ROEWER, C. F. 1942. Katalog der Araneae. Vol. 1. Bremen. Simon, Eugene 1892- Histoire naturelle des araignees. Deuxieme edition. 1903. Librarie Encyclopedique de Roret, Paris. BREVIORA Miasemm of Compsiratlve Zoology Cambridge, Mass. December 5, 1963 Number 192 A DESCRIPTION OF DINOPIS LONGIPES F. P.-CAMBRIDGE, 1902 (Araneae, Dinopidae). By Arthur M. Chickering F. P.-Cambridge described this highly interesting species en- tirely by means of figures and a statement of the total lengths of the two sexes. It may be that one reason for this extreme brevity was due to the fact that fully mature individuals have extraordi- narily long and fragile legs and the body is usually badly distorted in the preservative. This results in considerable difficulty in getting specimens sufficiently intact to make description significant. It has usually been assumed that this is a rare species but this is probably because collecting activities have not been sufficiently intensive. At any rate, I now have in my collection several mature specimens of both sexes and numerous immature individuals. Because of the extreme brevity of the original description it has been thought advisable to prepare a detailed description of both sexes in order to make clear the fundamental features of the spe- cies. After a careful search through the collection, individuals of both sexes have been selected for detailed description in accord with my usual formula. The selected male is smaller than some other individuals of the same sex but its parts are quite well pre- served and its legs are much less fragile than in most of the other available specimens. It appears to have been captured soon after the last moult and before its full size had been reached. The female selected for description had probably reached her full mature size. Genus DiNOPIS Macleay, 1839 DiNOPis LONGIPES F. P.-Cambridge, 1902 Male. Total length 16.282 mm. Carapace 4.42 mm. long; about 2.73 mm. Vv-ide; extremely flat; with a shallow median depression continued posteriorly by a long, very narrow groove ; with several very short spines. BREVIORA No. 192 Eyes. Eight in three rows with arrangement typical of the genus; posterior centrals enormously enlarged and moved forward to dominate the optical area. Viewed from above, only posterior eyes seen in two rows. Viewed from in front, anterior row definitely procurved. Ratio of eyes AME : ALE : PME : PLE = 13 : 15 : 32 : 16. AME separated from one another by 18/13 of their diame- ter, from ALE by 42/13 of their diameter. PME separated from one another by about 2/3 of their diameter, from PLE by nearly 3/2 of their diameter. (All measurements are made from the inner margin of the iris and not the margin where the cornea becomes continuous with the cuticle, which is often difficult to determine.) The iris of the PME is purplish and the border of the cornea is bright red. The clypeus is lacking because of the marginal posi- tion of the first row of eyes. The AME are located on a common tubercle; each ALE is also mounted on a fairly prominent tubercle. Chelicerae. Rather small for the size of the animal ; basal seg- ment about .77 mm. long; distal half of medial surface with many long, curved, stiff, brownish bristles. Fang evenly curved, without special features. Fang groove fairly well defined; with 5 definite small teeth plus a terminal nodule along the promargin; with 10-12 minute teeth along the retromargin, some of which are no more than nodules; some irregularities have been noted on left and right sides and in different specimens. mmmh liiii External Anatomy of Dinopis longipes Fig. 1. Lip. Fig. 2. Palpal tarsus of male showing coiled embolus and central apo- physis. Fig. 3. Di.stal end of embohis; removed from a second male. Fig. 4. Epigynum from below. 1963 REDESCRIPTION OF DINOPIS LONGIPES 3 Maxillae. Nearly parallel in general; massive in proximal halves; narrowed and somewhat diverging in distal halves; with well developed scopulae. Lip. General features shown in Figure 1. Legs. 1243. Width of first patella at "knee" .550 mm., tibial index of first leg 3. Width of fourth patella at "knee" .440 mm., tibial index of fourth leg 4. Femora Patellae Tibiae Metatarsi Tarsi Totals (All measurements in mm.) 1. 13.780 1.950 14.495 20.150 5.460 55.835 2. 11.700 1.755 11.310 12.675 3.510 40.950 3. 9.750 1.280 7.410 7.280 1.625 27.345 4. 10.855 1.690 10.075 10.725 1.755 35.100 Palp 2.015 .845 .520 1.170 4.550 Spines: Numerous femoral spines are present but apparently all are very small and, perhaps, irregularly placed; no patellar spines have been observed ; tibial and metatarsal spines are numer- ous and of moderate size. Three tarsal claws are present with each upper claw provided with a single row of about six teeth ; the third claw is simple and without teeth ; there are also numerous spurious claws present. Numerous trichobothria have been observed on tibiae and metatarsi; they are not confined to "ventral surface at base of tibia" as stated by Dr. Petrunkevitch (1939). The cala- mistrum is moderately well developed on the proximal half of the fourth metatarsus of immature specimens; on the specimen here described in detail it is, apparently, much reduced. Palp. Both patella and tibia are short, simple and lacking apophyses or other special modifications; the cymbium is a deep cup-shaped structure containing the bulb which has an extra- ordinarily long, coiled embolus ; enclosed in the center of the circu- lar coils of the embolus is a somewhat distinctive structure noted by F. P.-Cambridge and again noted in this study. Figures 2 and 3 show the chief features of the palp as observed by this author. The length of the free part of the embolus seems to vary in different individuals. In one example there were apparently ten circular coils making a total length of the embolus approximating two inches. In the specimen described here in detail the free part of the embolus is much shorter; when stretched out it resembles a clock spring. The median apophysis surrounded by the coiled embolus also exhibits different shapes in different specimens. The signi- ficance of these differences is not at all clear. 4 BREVIORA No. 192 Abdomen. Long, slender; of nearly uniform width; with a pair of very small dorsal tubercles slightly behind the middle. The six spinnerets and cribellum appear typical in males of the genus. The abdomen is well supplied with plumose hairs. Color in alcohol. Carapace: the basic color is a yellowish brown; with a narrow darker median stripe and a fairly broad and still darker dorsolateral stripe on each side; white plumose hair extends between the PLE up to the enlarged PME. The sternum is yellowish, with a very irregular, broken, narrow, brown border and several irregular whitish subchitinous patches. The legs are yellowish with sparsely located small black dots ; the coxae are much spotted with black ventrally. Abdomen: light brownish dor- sally and laterally ; with a darker, median, dorsal stripe ; the lateral sides possess several alternating, light and dark longitudinal stripes; the venter has a fairly broad, brownish, variegated, median stripe and many small, whitish, subchitinous, irregular spots on each side of the darker median stripe. Female. Total length about 24.7 mm. Carapace about 8.58 mm. long; 4.225 mm. wide op]:)osite interval between second and third coxae where it is widest; much narrowed immediately in front of first coxae where it is about 2.405 mm. wide; surface with nu- merous shallow depressions. Eyes. General features as in male. Ratio of eyes AME : ALE : PME : PLE = 9.5 : 11 : 44 : 12.5. Chelicerae. Essentially as in male except for teeth along the fang groove. The specimen from which this description is chiefly taken shows seven teeth along the promargin of which the second, fifth, and seventh are very small. Along the retromargin there are five or six teeth of moderate size and many minute teeth, little more than minute tubercles; immediately behind the last tooth of moderate size there is a series of five very minute teeth some of which are on the lower part of the larger tooth itself; some of the minute teeth occur in the fang groove. The fang groove in another female specimen showed the promargin with six teeth of which the fourth and sixth were small, the others of moderate size; the re- tromargin had eight teeth in an irregular row together with numer- ous minute tubercles. There is a well defined scopula along the pro- margin of the fang groove. Maxillae. Essentially as in male. Lip. This structure is abnormal in the specimen chiefly used for this description. Another specimen has the lip essentially as shown in Figure 1. Sternum. Also essentially as in male. 1963 REDESCRIPTION OF DINOPIS LONGIPES 5 Legs. 1243. Width of first patella at "knee" .88 inm., til)ial index of first leg 4. Width of fourth patella at "knee" .902 mm., tibial index of fourth leg 6. Femora Patellae Tibiae Metatarsi Tarsi Totals (All measurements in mm.) 1. 18.525 2.925 17.680 18.850 4.160 62.140 2. 18.200 2.860 16.900 16.900 2.929 57.789 3. 13.780 2.015 10.400 10.725 1.950 38.870 4. 14.170 2.665 12.545 11.895 1.950 43.225 I have found difficulties in making leg measurements as accurate as usual because of fragility and breakage. The palpal claw has three prominent teeth ; the palpal tarsus bears many stiff bristles and numerous spines. The calamistrum is fairly well developed on the proximal third of the fourth metatarsus. All legs are richly provided with spines of two kinds; the usual type are numerous, short and, apparently, not regularly placed; the second type is a slender, highly branched, unusual kind. Trichobothria have not been observed in females, perhaps because of the multiplicity of spines. Abdomen. Badly distorted by the preservative. Apparently as in male in all essential general features except that the dorsal tu- bercles somewhat in front of the middle are moderately well de- veloped in contrast to the minute tubercles in the male ; each tu- bercle has its posterior surface covered by a coating of short, white hairs. The cribellum is well developed and undivided. Epigynum. Essentially as represented in Figure 4. Color in alcohol. Essentially as in male except that the whole coloration is darkened to a dull brown in general ; the dorsal sur- face of the abdomen bears a series of isolated, small, white dots caused by clusters of short, white hairs. Collection records. Both specimens chiefly used for this de- scription were collected on Barro Colorado Island, C. Z., July, 1954. Several other adults and numerous immature specimens are in the collection from many localities in the Canal Zone and Boquete, Chiriqui, Panama. The species seems to be fairly abundant on Barro Colorado Island. All specimens referred to in this paragraph are now deposited in the Museum of Comparative Zoology at Har- vard College. 6 BREVIORA No. 192 REFERENCES Bonnet, Pierre 1956. Bibliographia Araneorum. 2. 2me partie. Toulouse: Les Artisans de rimprimerie Douladoure. , Cambridge, F. Pickard 1897- Arachnida-Araneida. In: Biologia Centrali-Americana. Vol. II. 1905 Dulau & Co., London. Petrunkevitch, Alexander 1911. A synonymic index-catalogue of spiders of North, Central, and South America, etc. Bull. Amer. Mus. Nat. Hist., 29: 1-809. 1939. Catalogue of American Spiders. Pt. 1. Trans. Conn. Acad. Arts and Sci., 27: 51-248. Roewer, C. F. 1954. Katalog der Araneae. Vol. 2, Pt. 2. Brussels. BREVIORA Mmseiinii of Comparative Zoology Cambridge, Mass. December 5, 1963 Number 193 A MIOCENE TOAD FROM COLOMBIA, SOUTH AMERICA By Richard Estes^ and Richard Wassersug^ The fossil toad described here is part of the large vertebrate assemblage of the La Venta fauna (Fields, 1959), and was col- lected by R. W. Fields during the 1949 ITniversity of California field expedition to the upper Magdalena Valley, Huila, Colombia, South America. It is the same specimen identified by D. Savage (1951) as a leptodactylid, cited by Estes (1961) as a bufonid close to the living Bufo ah>arius and B. crucifer, and discussed briefly by Tihen (1962b, p. 14) as Bufo sp., near B. marinus. Tihen's very useful paper on bufonid osteology (1962a) now makes it possible to give a much more accurate and clear-cut assessment of the relationships of this animal. Few fossil Bufo have been recorded previously from South America (Tihen, 1962b). Both authors wish to express their gratitude to Drs. E. E. Wil- liams, R. F. Laurent, J. A. Tihen and J. M. Gallardo for their many helpful suggestions, and to Mr. Howard Hamman, who prepared the illustrations. This work was supported by National Science Foundation Grant NSF G-18905, held by the senior au- thor, and research time for the junior author was provided through the Thayer Academy Advance Studies in Science Program (also supported by the National Science Foundation). We are grateful for this support. ' Department of Biology, Boston University, and Associate in Vertebrate Paleontology, Mu- seum of Comparative Zoology, Harvard University. ^ Thayer Academy, Braintree, Massachusetts 2 BREVIORA No. 193 Class AMPHIBIA Superorder SALIENTIA Order ANURA Family BUFONIDAE BuFo MARiNus Linnaeus Referred specimen — University of California no. 41159, postor- bital portion of skull, eight articulated vertebrae, both scapulae with articulating proximal ends of the humeri, both distal ends of humeri in articulation with proximal ends of radioulnae, two frag- ments from the region surrounding the acetabulum of the pelvic girdle, distal end of right femur, proximal ends of both tibiofibulae, distal extremity of right tibiale-fibulare, and two unidentified fragments. Two bone shafts which were collected at the same time, and have the same specimen number, are probably mammalian. Horizon — University of California locality V-4517, Monkey unit, Honda group (Fields, 1959, p. 419, and fig. 2). Age — Late Miocene. Fauna — La Venta. Preservation — The skeleton was apparently complete before ex- posure and erosion disarticulated and destroyed parts of it. The matrix is a silty mixture of claystone and sandstone, and the cavi- ties have been filled either with this materialor with calcite. In some cases, calcite or matrix have remained as endocasts of parts of the bones. Breakage has apparently been the result of erosion, and has occurred at the weakest and thinnest points. The shoulder and elbow complexes are articulated in flexed posi- tions, and the vertebral column has a slight ventral curve, similar in both cases to their positions in life in a normal resting position. It is possible that the toad died and was buried in such a position. Description — The maximum width of the skull across the pos- terior arms of the squamosals is 39.5 mm. A height measurement of the skull taken from the dorsal extremities of the prootic part of the otoparietal (in Bvfo, paired processes directly dorsal to the foramen for the ninth and tenth cranial nerves) to the parasphe- noid-ptcrygoid suture is 10.0 mm. A ventral measurement along the median line from the exoccipital condyles to the jiostcrior border of the sphencthmoid is 1L7 mm. In occipital view, the skull roof appears flat, except for the strong postorbital crests and suggestions of the supraorbital crests. The foramen magnum is about 65 per cent wider tlian deep, and the foramina for ninth 1963 MIOCENE TOAD and tenth cranial nerves are prominent. Only the left columella is present and its tip was broken in handling. Its true length, as indicated in the figures, was taken from a j^hotograph made before the breakage occurred. The prootic parts of the otoparietal are strongly ossified laterally and ventrally and form the ventral bor- ders of strongly marked troughs for the columellae. The pterygoids ^P^. foa opa Figure 1. Bufo marinus, U. C. no. 41159, ventral view of skull, anterior to the top; above, outline, and below, shaded drawing; X2. On the outhne, solid lines indicate sutures, dotted lines are contours, and cross-hatching indicates breakage. Abbreviations : ecc, endocranial cast; ex, exoccipital; joa, canal for occipital artery; joe, oculomotor foramen; jom, foramen magnum; jor, foramen for maxillo-mandibular branch of trigeminal nerve; fps. frontoparietal .suture; ma, attachment for nuchal muscles; oc, exoccipital condyle; oj, large opening into braincase for optic and trochlear nerves; opa, otoparietal; pas, parasphenoid ; pt, pterygoid; sph, sphenethmoid ; sq, squamosal; st, columella auris; tr, foramen for fifth, sixth and seventh cran- ial nerves; vj, foramen for ninth and tenth cranial nerves. BREVIORA No. 193 are broken just medial to the point at which they would have given off their quadrate and palatal processes, and the descending quad- rate processes of the squamosals are broken off at approximately the same level. The cultriform process of the parasphenoid is broken off just posterior to its juncture with the sphenethmoid, but the bone shows nothing unusual. The squamosal-prootic and prootic-parasphenoid sutures are difficult to discern, especially on the right side, and it appears that almost complete fusion has oc- curred. opa Figure 2. Bufo marinus, U. C. no. 41159, posterior view of skull; above, outline, and below, shaded drawing; X2. Comments and abbreviations as in Figure 1. In dorsal view, the dermal ornamentation lacks strong protu- berances other than the postorbital and orbital crests, though weak parietal crests are present, and two small bumps occur near the midline. The all-over pattern of the dermal ornament is a pitted and wrinkled one, in which the wrinkles extend more or less trans- versely across the top of the skull. The frontoparietal portions of the otoparietals appear to be fused to each other, though an anterior groove, shown in Figure 1963 MIOCENE TOAD 3, perhaps indicates part of their suture. Conjoined, these fronto- parietal areas are 21.7 mm. at their widest point. The anterior sec- tion of the frontoparietal area is broken off before its contact with the nasals. The occipital groove (Tihen, 1962a), is enclosed to form a canal. On the right side, breakage has removed the overly- ing bone and exposed the endocast of this canal. In ventral aspect, much of the brain cavity appears as an en- docranial cast. Small posterior fragments of the sphenethmoid, along with the anterior arms of the prootics and the cultriform process of the parasphenoid, completely surround an opening which was membrane covered in life and through which the optic opa Figure 3. Bufo marinus, U. C. no. 41159, dorsal view of skull, anterior to the top; above, outline, and below, shaded drawing; X2. Comments and abbreviations as in Figure 1. and trochlear nerves passed. This opening is somewhat squared- off , with notches in anterodorsal and posteroventral corners. Be- cause of greater ossification, especially on the posterior border of the sphenethmoid, the opening is relatively smaller than that of 6 BREVIORA No. 193 Ptecent Bufo. The lateral wings of the parasphenoid are broadly overlapped by rugose medial processes of the pterygoids. The fora- men for cranial nerves five, six, and seven, and the oculomotor foramen are as in Recent Bufo. The occipital grooves (here en- closed to form canals) open on the ventral surface of the orbital portions of the frontoparietals, and are continued for a short dis- tance as channels on the roof of the orbit. These paired canals, which in life transmitted the occipital arteries, diverge slightly as they pass anteriorly, but lie essentially parallel to the midline. Dorsal to the suture between the prootic and the pterygoid, im- mediately lateral to the foramen for cranial nerves five, six, and seven, and on a level with it, is a groove which passes through the prootic and emerges immediately lateral to the base of the colu- mella. The maxillo-mandibular branch of the trigeminal nerve opa for pt Figure 4. Bufo mnrinus, U. C. no. 41159, anterior view of right half of skull, outline only, showing posterior wall of orbit ; X2. Comments and abbreviations as in Figure 1. is found in this groove in Rana (Holmes, 1924, p. 296) and dissec- tion showed this to be the case in Bufo as well. The groove in Bufo terminates on the lateral borders of the prootic, leading into the supraptcrygoid fenestra (Tihen, 1962a, j). 160), but in this speci- men there has been extremely high ossification in this area and portions of the prootic, squamosal, and pterygoid have completely closed the fenestra, and enclosed the groove in a foramen. The thick and massive vertebral colunni is complete from the atlas back to about the level of the anterior third of the eighth vertebra, though all of the transverse processes are badly broken. The column was originally articulated with the skull, but was re- moved by the senior author in order to study the configuration of the exoccipital condyles and atlantal cotyles. An approximate 1963 MIOCENE TOAD overall measurement of the seven presacral vertebrae is 42.5 mm. The atlantal cotyles are separated by a shallow notch. The neural spines are flattened and capped with dermal bone that has a slightly pitted texture. None of the broken transverse processes are long enough to give positive information about their true length or orientation, though they seem to have been oriented as in Recent Biifo. The zygapophyses are robust, as are the centra. Figure 5. Biifo mannus, U. C. no. 41159, right, medial view of left scapula and humerus {h) , showing sutural contacts of scapula (sc), coracoid (c), and clavicle (cl) ; left, anterior view of distal end of left humerus and proximal end of radioulna (r), to show crista medialis (cm) ; below, ventral view of vertebral column, anterior to left; all X2. The right scapula is the better preserved. Its dorsal edge is broken parallel to, and near, the original natural border. The left scapula is broken at about the same point, but its posterior half is missing. In the glenoid fossa, the small band of cartilage that separates the scapula from the clavicle in the Recent species is completely replaced by a firm bony suture. In Bufonidae, the pars acromialis of the scapula (Proctor, 1921, p. 197) forms a strong prominence which is relatively larger in the fossil (on the right side ; it is broken on the left) than in the Recent species. The 8 BREVIORA No. 193 paraglenal cartilage (Ecker, 1889, p. 40), between the scapula and the coracoid, usually forms a noticeable portion of the floor of the glenoid cavity in modern anurans, but is absent in the fossil and replaced by a firm suture of these two bones. The large foramen in the glenoid fossa is, as a result of the heavy ossification in this region, slightly smaller than in the Recent forms. The deltoid crest of the humerus is quite prominent, and is about as high from base to peak as the circumference of the hu- merus at the same point. The shafts of both humeri are missing, distal to the crests. The radioulnar articulations are similar to those of Recent Bvfo. On the distal segment of the left humerus, the crista medialis is well developed, and has a strongly rugose muscular attachment surface. All that remains of the pelvic girdle is the ventroposterior cor- ner of the ischium and a ventral piece of the ilium from near the acetabulum. The proximal articulation surfaces and short segments of shafts of the tibiofibulae show no differences from Recent Bujo. The right tibiofibula has the longest shaft preserved. The grooves sepa- rating fused tibial and fibular components are deep but no more so than in the Recent species. The distal end of the right femur is no longer in natural articulation with the tibiofibula but still makes a perfect fit. Both proximal and distal extremities of the right tibiale-fibulare are present. The diaphyses of the bones have been broken so that a natural fit is no longer possible. Two small fragments may represent a posterior section of the ilial shaft and a section of tibiofibula, but are insufficient for posi- tive identification. Discussion — Tihen (1962a, p. 163) defines the valliceps species group of Bufo as follows: frontoparietals broad, usually pro- duced into crests; roofing bones ornamented; occipital groove en- closed to form a canal; frontoparietals and prootics fused. He then divides the valliceps group into three essentially geographical subgroups: the Mexican section in Central and North America; the South American section in South and Central America; and the Caribbean section, throughout the Neotropical Region, "par- taking to some extent of the characteristics of each of the others, besides developing its own features." This fossil is placed in the valliceps group on the basis of pres- ence of all of the above characters. Within this group, it is elimi- nated from the Mexican section by having a strong overlap of the 1963 MIOCENE TOAD 9 medial wings of the pterygoids onto the wings of the parasphenoid and complete closure of the suprapterygoid fenestra {ibid., p. 168). The strongly-overlapping pterygoids are characteristic of the South American section, and do not usually occur in the Carib- bean section, but the occlutled sui)rai)terygoid fenestra is present in the latter (ibid., p. 171 ) . This single resemblance to the members of the Caribbean section will be discussed below. Bufo chilensis of the South American section and B. retiformis of the Caribbean section were the only species not available for this study. The dermal ornamentation of the fossil is most like that of the South American section, which has crests of only mod- erate extent and development and a lined or wrinkled sculpture; while crests in members of the Caribbean section are often exten- sive, exaggerated, and the sculpture is pustular. A final factor used in allocating the fossil to the South Ameri- can section was the width of the vertebral centra, wiiich are per- ceptibly narrower in proportion to their length in the Caribbean group. As might be expected, the results are correlated with size, so that only specimens of relatively large size appear to be well separated. Many specimens of the Mexican section resemble the Caribbean forms in also having the narrow centrum. The above considerations indicate that the fossil belongs to the South American section, but it differs from these forms in one characteristic of importance to Tihen's classification. He indicates (ibid., pp. 165-166) that the suprapterygoid fenestra is not mark- edly occluded in the South American section of the imlliceps group, while it is often nearly closed by flanges of pterygoid and squmosala in members of the Caribbean section. The fenestra is completely closed in the fossil, thus indicating a possible relation- ship to the latter group in terms of the classification based on skeletons of Recent species. However, the suprapterygoid fenestra, as a taxonomic character, may be weak in some cases (as Tihen realized), owing to its qualitative nature. Tihen points out {ibid., p. 174) that B. typhonius of the Caribbean section lacks an oc- cluded suprapterygoid fenestra. One specimen of B. peltocephalus of the Caribbean section (M. C. Z. no. 23564) also has an open fenestra. Within the South American section, one specimen of B. paracnemis (M. C. Z. no. 343) has the fenestra closed on one side of the skull. It is possible that closure of the fenestra in Recent specimens may be both variable within the species as a whole, and be partly a function of age of the specimen. In addition, since there has been a trend in many groups (including anurans) toward deossification, geologic age can be a modifying factor as well. The 10 BREVIORA No. 193 latter is probably the most important with respect to this fossil, since a number of regions, e.g. shoulder girdle (see above) and prootic show a greater amount of ossification than comparable regions in any Recent specimen of either New or Old World Bufo seen by us. The crista medialis of the humerus is of some interest in this specimen, because it is a secondary sexual characteristic in some Recent frogs. The crista forms the attachment for the M. flexor carpi radialis, which aids in flexing the wrist and is important in amplexus. Ecker (1889, pp. 42-43) indicated that the crista was present in males of Rana esculenta, R. temporaria, and R. oxy- rhinus. Holmes (1924, p. 241), in discussing the same species, states that it is present in both males and females, but is more prominent in males. Inasmuch as the methods of embrace and courtship are more or less uniform throughout the Salientia (Noble, 1931, p. Ill), this characteristic may be of similar signif- icance in other anurans as well. Table I gives the results that were derived from specimens of Bufo related to the fossil. Taxon TABLE I MCZ no. sex crista medialis B. peltocepha ilus 29600 9 absent 23564 S present B. ar en arum 1263 9 absent 30650 S present B. bloinbergi 29669 9 absent J^ )t) nriYt )iR hi irribilis- 21439 9 present but small 29028 S absent — small specimen B. ictericus 321 s ? present Only a small number of the M. C. Z. collection of New World Bufo were sexed at the time of skeletonization, and of the others, only a few have the ridge to any great extent. B. marinus horribilis (M. C. Z. no. 29028) , a male which lacks the crista medialis, was the smallest individual of that species in the osteological col- lection and may not have reached maturity. The fossil is thus most probably a male, though exceptions such as the above render this uncertain. SPECIFIC ASSIGNMENT OF THE FOSSIL As indicated above, the specimen clearly belongs to the South American section. Tihen (1962a, p. 165) includes the following 1963 MIOCENE TOAD 11 species witliin this grouj): Bufo arenarum, B. blomhergi, B. chi- lensis, B. crucifer, B. ictericus, B. m. marinus, B. m. horribilis, and B. paracnemis. The fossil differs strongly from Bufo blombergi in having strong supraorbital and postorbital crests. B. arenarum has very high cranial crests, which are rugose and slightly flat- tened on their dorsalraost surfaces, and has a relatively smooth skull surface between the crests — both characters in contrast to those of the fossil. B. crucifer has a sculpture pattern much finer in texture and more deeply incised than in the fossil, has strong parietal crests, and has a relatively smooth skull surface between the crests, as in B. arenarum. The remaining taxa are very closely related, and all have been included in a single undifferentiated species, B. marinus, by some workers, or treated as subspecies or full species by others. We have followed Tihen's taxonomy (1962a) for convenience in discussion. M. C. Z. specimens of B. m. horribilis indicate that this primarily Mexican group tends to have thin, fine, almost pitted sculpture, and relatively low cranial crests, unlike the fossil. Specimens of B. ictericus contrast with the fossil in having prominent parietal crests, though this is a variable character. Both skeletons of B. paracnemis available to us resemble the fossil in sculpture and crest pattern, but on each side have a very strong anterior angulation of the posterior skull margin, in contrast to the fossil and the other species included in the South American section, though the consistency of this char- acter is not determinable at this time. Hence the strongest resem- blance between the fossil and any Recent group, on the basis of the available specimens, is with Bufo marinus marinus. The tend- ency toward obliteration of sutures, the relatively greater ossifica- tion in prootics, suprapterygoid fenestra, and shoulder girdle, as well as the general robustness of some of the bones themselves, do not, in our opinion, warrant nomenclatorial recognition at the specific level, but merely reflect the often greater ossification seen in many fossils (discussed above on p. 9) and probably indicates a stage in the evolution of Bufo marinus in the broad sense. In identifying this specimen with the Recent Bufo marinus, we real- ize, with Taylor and Smith (1945, p. 541) and J. Savage (1960, p. 235) that the concept of B. marinus as currently used is a broad one, yet further refinement in the assessment of relationships of this fossil cannot precede re-evaluation of the Recent Bufo ma- rinus complex. The beginnings of such a re-evaluation have been provided by Bertini and Cei (1962), who studied the serological relationships 12 BREVIORA No. 193 between some of the species involved. Bufo marinus, a predom- inantly Amazonian aquatic species, appears to be the most primitive type, and has given rise to xerophilic continental, and moisture-loving coastal populations in the southern half of the continent. Identification of this fossil with the populations usually referred to as B. 7n. marinus is thus consistent with the northern occurrence of the fossil within South America, and with the primitive nature of B. m. marinus Blair (1963, p. 11) has discussed evolutionary patterns in Bujo, and has suggested that the ancestral stock of the marinus group "somehow crossed into South America during its time of isolation from North America in the Tertiary." There is still insufficient fossil material to make any positive statement, but most recent work on fossil anurans suggests that by the late Cretaceous or Paleocene, family groups were well differentiated, and lines re- sembling many modern genera were already present. It is thus quite possible that the ancestral stock for the marinus group entered South America by a land connection with North America which persisted until late Paleocene time. If this is so, then it is not necessary to propose that they reached South America by some sort of "sweepstakes" method after the connection was severed. SUMMARY AND CONCLUSIONS A late Miocene toad from the upper Magdalena Valley, Huila, Colombia, South America, is referred to the Recent species Bufo marinus. Within this poorly understood group, it seems to show closest resemblance to B. m. marinus from northern South America. It differs from B. marinus only in having slightly greater ossification in the suprapterygoid fenestra, lateral parts of the prootics, and glenoid region of the shoulder girdle. This condition is insufficient evidence for recognizing a taxonomic difference from B. marinus, at least until restudy of the Recent forms is efi'ected. The broad geograi)hic and altitudinal range of the Recent species precludes ecologic interpretation. The species is aquatic, and common today along large river courses and may have been so on the Miocene floodplains as well. The fossil, in combination with the lizards described by Estes (1961), is another indication of the modernity of the hcrpetological elements of the late Mio- cene La Venta fauna, and of the greater extent of the floodplain- aquatic habitat in northern South America during the late Miocene. 1963 MIOCENE TOAD 13 REFERENCES CITED Bertini, F., and J. M. Cei 1962. Seroprotein patterns in the Bufo mnrinus complex. Hcrpetologica, 17: 231-238, 3 figs., 1 table. Blair. W. F. 1963. Evolutionary relationships of North American toads of the genus Bujo: a progress report. Evolution, 17: 1-16, 4 figs. ECKER, A. 1889. The anatomy of the frog. Oxford, Clarendon Press, translated by George Haslam, pp. i-x, 1^49, 261 figs., 2 pis. ESTES, R. 1961. Miocene lizards from Colombia, South America. Breviora, Mus. Comp. Zool., Harvard Univ., no. 123: 1-11, 5 figs. Fields, R. W. 1959. Geology of the La Venta badlands, Colombia, South America. Univ. Calif. Publ. Geol. Sci., 32: 408^44, 2 figs., 4 pis., 2 maps. Holmes, S. J. 1924. The biology of the frog. MacMillan and Co., New York, pp. v-ix, 1-370, 94 figs. Noble, G. K. 1931. The biology of the Amphibia. McGraw Hill and Co., New York, pp. 1-577, 174 figs. Proctor, J. B. 1921. On the variation of the scapula in the batrachian groups Aglossa and Arcifera. Proc. Zool. Soc. London, 1921 ; 197-214, 10 figs. S.WAGE, D. E. 1951. Report on fossil vertebrates from the upper Magdalena Valley, Colombia. Science 114: 186-187. Savage, J. M. 1960. Geographic variation in the tadpole of the toad, Bufo marinus. Copeia, 1960 (3) : 233-235, 3 figs. Taylor, E. H., and H. M. Smith 1945. Summary of the collections of amphibians made under the Walter Rathbone Bacon Traveling Scholarship. Proc. U.S. Nat. Mus., 95 : 521-613, 4 figs., 15 pis. TlHEN, J. A. 1962a. Osteological observations on New World Bujo. Amer. Midi. Naturalist, 67: 157-183, 38 figs. 1962b. A review of New World fossil bufonids. Ibid., 68: 1-50. v^ BREVIORA Museiiim of Comparative Zoology Cambridge, Mass. December 31, 1963 Number 194 A NEW SUBSPECIES OF TROPIDOPHIS GREENWAYI FROM THE CAICOS BANK By Albert Schwartz The species of small boa, TropidopJiis greenwayi, has hereto- fore been known only from Ambergris Cay.^ No additional speci- mens of T. greenwayi have been taken since the type and paratype were collected in 1936. At the time of their review of the Carib- bean snakes of the genus TropidopJiis, Schwartz and Marsh (1960) examined the two extant specimens and separated them from the Cuban and Bahaman species T. pardalis (with which they had been nomenclatorially associated) and T. canus (to which they might be expected on geographic grounds to be related). From January 11 to 22, 1961, the writer and David C. Leber collected on South Caicos and its adjacent Long Cay and visited the Ambergris Cays as well in order to secure additional speci- mens of T. greenivayi and to ascertain if this snake were more widely distributed on the Caicos Bank. We were unable to find the boa on the Ambergris Cays, but a series of fourteen individ- uals was collected for us on South Caicos and Long Cay by resi- dents of these islands. Comparison of this large series with the two specimens of topotypic T. greenwayi indicates that the South Caicos and Long Cay populations represent a different form, which may be called : Tropidophis greenwayi lanthanus, new subspecies Type: Museum of Comparative Zoology (MCZ) 69630, from 0.5 mi. north of Cockburn Harbour, South Caicos, taken 22 Janu- ary, 1961, by a native for A. Schwartz. Paratypes: MCZ 69619, 0.5 mi. east of Cockburn Harbour, South Caicos, 13 January, 1961 ; MCZ 69632, same locality, 14 lActually, "Ambergris Cay" is one of a pair of cays, both known as the Amber- gris Cays, which lie about 13 miles southwest of Cockburn Harbour on South Caicos. BREVIORA No. 194 January, 1961 ; MCZ 69620, same locality, 19 January, 1961 ; MCZ 69621, 7 mi. northeast of Cockburn Harbour, South Caicos, 20 January, 1961 ; MCZ 69622, 0.5 mi. east of Cockburn Harbour, 21 January, 1961 ; MCZ 69623, Cockburn Harbour, 21 January, 1961; MCZ 69624-25, Long Cay, off Cockburn Harbour, South Caicos, 21 January, 1961 ; MCZ 69626-28, same locality, 22 Janu- ary, 1961 ; MCZ 69629, 69631, same data as type. All specimens were collected by natives for A. Schwartz. Distribution: Known only from South Caicos and adjacent Long' Cay on the Caicos Bank. Diagnosis: A subspecies of T. grcenwayi differing from the nominate form in higher number of ventral scutes and in colora- tion and pattern. Description of type: An adult spurred male with the following measurements and counts: total length, 257 mm., tail, 30 mm.; ventral scutes 160, subcaudal scutes 28 ; supralabials 9/10 ; infra- labials 11/11; parietals in contact; preoculars 1/1, postoculars 3/2 ; dorsal scales smooth, rows 25-27-19 ; dorsal paramedian blotches 42/49 ; tail blotches four. Coloration : Head uniformly Figure 1. Dorsal uiidbody pattern of Tropidoplils grcenwayi lanthanus (MCZ 69630, type). 1963 NEW SUBSPECIES OF TROPIDOPHIS GREENWAYI 3 dark brown dorsally; neck slightly paler, quickly grading pos- teriorly into a grayish-tan middorsal zone about ten scales wide, this zone including the two paramedian rows of dark brown dor- sal blotches, each blotch faintly outlined with pale gray, the dorsal /one continuing posteriorly onto the upper surface of the tail, where it merges imperceptil)ly with the yellow coloration of the tail tip. Sides darker brown, becoming more reddish ventrally, and enclosing three rows of lateral blotches, the uppermost two rows the least conspicuous (because of the closeness of their color to that of the ground color and the reduction of pale outlining), the lowermost row outlined by gray especially along the pos- terior margins. A^entral surface reddish-tan with two rows of dark brown blotches, which may extend dorsally onto the first two scale rows, do not coalesce medially on the venter, and are out- lined posteriorly in white. Supralabials dark brown, the pos- terior three somewhat flecked and blotched with cream. Ilemi- penes extruded, weakly bifurcate and weakly s]iinose distally. Paratypes: The paratypes include three males and ten females : I cannot discern any difference in coloration, pattern or scalation between the sexes nor between the snakes from South Caicos and Long Cay, and all are discussed as one unit. All three males are spurred, as is the type. Two of the males are juveniles, each with a total length of 140 mm. Males range in total length from 140 mm. to 320 mm., whereas females vary between 250 mm. and 336 nun. Ventrals range between 156 and 165 (mean 161.2), caudals between 26 and 30 (mean 28.0). All have 1/1 preoculars. Post- oculars vary from 2/2 (two snakes), 2/3 (six, including type), and 3/3 (five), thus showing a tendency toward having three postoculars in 85 per cent of the population ; one snake has the head scales so damaged that its data are omitted. Paramedian blotches vary from 27 to 49 in the series, the largest discrepancy between right and left counts on the same snake being seven blotches, as in the type and one other individual. The average blotch number is 39.9. One snake, as noted below, lacks blotches completely. Scale rows at midbody are either 25 (eleven snakes) or 27 (three snakes). The number of rows of blotches around the body are eight or ten, with only one snake showing the reduction to eight rows of blotches. The tail/total length ratio (X 100) averages 10.5 (9.4—12.1 ; the highest ratio is shown by one of the juvenile males). All of the paratypes have the parietals in con- tact, and have the dorsal scales smooth. Upper labials vary from 9 to 10, and lower labials from 9 to 12. 4 BREVIORA No. 194 Coloration of the paratypes : All the adult paratypes except one agree very closely with the type. All have a lighter median zone containing two rows of paramedian blotches ; this middorsal zone was noted in life as matching in various individuals (color notations from Maerz and Paul, 1950) : PL MAS, PI. 15C7, PI. 13A6, PL 14A4, PL 12A6, all of which are shades of buffy tan to grayish tan. The sides are somewhat darker (PL 15A9, PL 16 A6, PL 13A10, PL 15C8, PL 15C8), then becoming lighter on the lower scale rows and venter to a more red coloration (PL 14G10, PL 15A4, for example). The blotches themselves are always dark brown. The venter varies from a rich reddish tan to an almost chocolate brown, at times relieved by rather extensive white bor- ders to the brown ventral blotches. One snake, an adult female, is unusual in that the pattern consists merely of the paler mid- dorsal zone and darker sides, without any indication of lateral or ventral blotches ; the paramedian blotch rows are represented by a rather diffuse dark brown smudging along the middle of the back. The tail tips are dark in two snakes and light (yellow) in ten ; two are indeterminate. The two juvenile paratypes are very like the adults in colora- tion and pattern ; the pattern elements are not appreciably brighter than in the adults (in contrast to juvenile and adult T. canus where the juveniles show the dorsal pattern much more distinctly than the adults). There is also no evidence of the lateral nuchal stripe which is a common feature of both juvenile and adult T. canus (Schwartz and Marsh, op. cit.: 61) and espe- cially prominent in the juveniles. In addition, the young paratypes have white, rather than brown, venters; apparently the darker pigmentation comes with increase in size. Comparisons: T. g. lanthanus requires comparison only with the typical form. It is quite distinct from the adjacent Bahaman and Ilispaniolan forms of Tropidophis. The comparison is ham- pered, however, by the paucity of specimens of the nominate form. If the type and paratype represent a fair sample of T. grccnwayi on the Ambergris Cays (and it is possible that they do not), then laniliayius is certainly very distinct from the snakes on these out- lying cays. Comparison of tlie illustration (Figure 1) of T. g. lanthanus with that of the type of T. g. greenwayi (Schwartz and ^larsh, op. cit.: fig. 7) at once demonstrates the pattern differences. In the nominate form there is no indication of the dorsal })ale zone, the entire dorsal surface has a mottled or stippled effect, 1963 NEW SUBSPECIES OF TROPIDOPHIS GREENWAYI 5 with blacks, browns, tans, and whites more or less intermixed ; the blotches are extremely obscured by the mottled and stippled mark- ings on the interspaces. Such is not the case in lanthanus where in all specimens but one (the blotchless female mentioned above) the blotches are quite distinct and there is no interspace stippling. It might be argued that the greenwaiji coloration is a peculiarity of extreme age or adulthood ; in fact the type of greenwayi is the largest specimen of the species presently available. However, in the largest specimen of lanthanus, the trend is obviously just the reverse of that in the nominate snakes, and the pattern is com- pletely obliterated except for the dorsal pale zone. Another fea- ture of greenwayi is the "salt-and-pepper" effect on the dorsal surface of the head ; such a condition does not occur in lanthanus, where the head is always uniformly dark brown. In number of ventrals, lanthanus embraces the 157-158 ventral counts known from the two greenwayi (both are males). However, the specimen of lanthanus which has the lowest count for that race (a female with 155) stands alone in the series; all other lanthanus have counts ranging from 160 to 165. I regard this single specimen as being somewhat aberrant. When large series of topotypic greenwayi finally become available, I suspect that they will have ventral counts low^er than those of lanthanus. Both specimens of greenwayi have 2/2 postoculars, w^hereas only two of thirteen lanthanus have such a count; the remainder have counts of 2/3 or 3/3, the third postocular being wedged be- tween the fifth and sixth supralabials. It is possible that a ten- dency toward 3/3 postoculars is characteristic of lanthanus. Remarks: Perhaps the most interesting feature of the new specimens of T. greenwayi here reported is their uniformity. For example, the species was partly defined (Barbour and Shreve, 1936:2) by having the parietals in contact. Schwartz and Marsh {op. cit. :57) noted that this character, as with all scale charac- ters in this assemblage of small boas, was variable ; the same com- ment applies equally to dorsal scale earination and number of dorsal scale rows. But all new specimens of T. greenwayi do indeed have parietals in contact and smooth scales. In fact, T. greeyiwayi can be in addition characterized as being a Tropidophis with smooth dorsals usuallv in 25 scale rows and tvpicallv ten rows of blotches, all features which were uncertain at the time of the review of the Caribbean Tropidophis. Likewise, additional data are now available on tail/total length ratio in T. greenwayi ; the ratio averages 10.5, which is the lowest mean of any member of 6 BREVIORA No. 194 the assemblage, most closely approached by that of T. canus canus (10.7). Even with additional material available, I am unable to guess as to the affinities and origin of T. grccnwayi. With two juveniles now at hand, the relationships of this species to T. canus seems even more remote than it did previously (Schwartz and Marsh. op. cit. : 62). This snake appears to have been long isolated from its relatives, whatever they may have been, and other than to remark that it is a member of the pardalis-maculatuf; assemblage, little can be said. I suspect that T. greemvayi will be found to occur on the other major islands of the Caicos Bank. It is surprising that it has been so long overlooked on South Caicos where the natives knew of it as soon as approached about small snakes. The name lanthanum is an allusion to the fact that the species has been overlooked on South Caicos. This boa appears to be genuinely lacking from Grand Turk and probably other islands on the Turks Bank, which is separated from the Caicos Bank by the Turks Island Passage : natives there were aware of Epicrates on some of the outer cays, but did not know of any snakes at all on Grand Turk. No snakes of any species were encountered in the one week spent by us on Grand Turk. Figure 1 was executed by Ronald F. Klinikowski. I wish to thank him for his efforts on my behalf. LITERATURE CITED Barbour, Thomas, and Benjamin Shreve 1936. New races of Tiopidophis and of Ameiva from the Balianias. Proc. New England Zool. Club, 16: 1-3. Maerz, a., and Rea Paul 1950. A dictionary of color. New York, McGraw-Hill Book Co., ])p. i-vii, 1-23, 137-208, 56 pis. SrHWAKTZ, Albert, and Robert J. Marsh 1960. A review of the paradalis-maculatus complex of the boid genus Tropidophis of the West Indies. Bull. Mus. Conip. Zool., 123(2) : r)0-84, 10 figs. BREVIORA Miisemnti of Comparative Zoology Cambridge, Mass. December 31, 1963 Number 195 CAYMAN ISLANDS TR0PID0PHI8 (REPTILIA, SERPENTES) By Richard Thomas New material from the Cayman Islands, gathered principally through the efforts of Dr. Albert Schwartz during his collecting in the West Indies, has provided a large enough series of Tropi- dopJiis from each of the three islands so that a clearer picture of the relationships can be seen. Specimens examined have been under the curatorship of Dr. Doris Cochran, United States National Museum (USNM), Dr. Ernest Williams, Mu- seum of Comparative Zoology (MCZ), Dr. Albert Schwartz, field collection (AS), and in the author's private collection (RT). The illustrations are the work of Ronald F. Klinikowski. The Cayman group consists of three islands : Grand Cayman, Ca.yman Brae, and Little Cayman. Grand Cayman is situated about 180 miles west-north-west of Jamaica and 150 miles south of Cuba ; it is about 22 miles long and eight miles wide at maxi- mum width. Cayman Brae, 89 miles east-north-east of Grand Cayman, is about 12 miles long with an average width of a mile. Little Cayman, five miles to the west of Cayman Brae, is ten miles long with a maximum width of two miles. The long axis of each island lies in a generally east-west direction. They are low-lying islands formed of calcareous rocks. Battersby (1938) described Tropidophis mcJanurus ca^iman- ensis from Grand Cayman; Grant (1940) further recorded T. m. cayDiancnsis from Cayman Brae and described Trojjidophis parkeri from Little Cayman. Schwartz and Thomas (1960) on the basis of obvious affinities with melanurus called parlicri a subspecies of T. melanurus. The new material recently acquired from the Caymans indi- cates that, despite affinities of the Cayman Tropidophis with T. melanurus, these snakes are distinct and should be known as 2 BREVIORA No. 195 Tropidophis caymanensis caymanensis and Tropidophis cayman- ensis parkeri. Tropidophis caymanensis may be distinguished from Tropidophis melanurus by the following characters: 1. Size : The average total length of caymanensis from all three islands (48 specimens) is 369.5 mm with a range of 202-564 mm. The average size of melanurus from Cuba and the Isle of Pines (77 specimens) is 607.1 mm with a range of 253-1057 mm. 2. Scale-rows : Caymanian Tropidophis have usual scale-row counts of 23-25-17 (caymanensis), 25-27-17 (parkeri), or 23- 25-17 (Brae Tropidophis). Tropidophis melanurus has a usual scale-row formula of 25-27-19. 3. Light tail tip : All specimens of Caymanian Tropidophis examined have a light tail tip ; of 16 specimens of parkeri seen in the field all but three (with incomplete tails) had yellow tail tips, and one of the three with incomplete tails showed traces of yellow pigmentation on the remaining tip. Old specimens show a slight darkening on the dorsum of the tip, but the rest of the tip is yellow. Specimens from Grand Cayman and Cayman Brae, which were not seen alive by the writer, have plain but dis- tinctly light tail tips, presumably as a result of the yellow fading in preservation; field color notes on one adult (AS16230) from Grand Cayman describe a yellow tail tip. In mclanurns the yel- low tail tip is a distinctly juvenile condition and in the adult becomes either unicolor with the ground color or black. 4. Dorsal body pattern : In caymanensis, the two rows of paramedian dorsal spots are squarish (2-4 scales long by 2-3 scales wide) and distinct from the ground color; these spots are more discrete than in melanurus, with pale edging frequently occurring on the median edges, and in addition they are distinct from the middorsal stripe. T. melanurus typically has para- median dorsal spots much less distinctly developed and they are apt to be broadly confluent at the midline and not distinct from the middorsal stripe, when the latter is present. The light edg- ing is confined to the outer edges of the spots. Tropidophis caytnavensis is, therefore, a species with closest affinities to melanurus but distinguished from that species by its dwarf size, lower number of scale rows, persistence of the yellow tail tip in adults and bolder dorsal spotting. Examination of the series of 16 specimens from Cayman Brae shows that they may be distinguished from the Grand Cayman population with which they had previously been considered identical. I take great pleasure in naming this new subspecies 1963 CAYMAN ISLANDS TROPIDOPIIIS for Dr. Albert Schwartz, who suggested this problem, in honor of his zoological work in the West Indies; it may therefore be known as : Tropidophis caymanensis schwartzi subsp. nov. Type: MCZ 69618, an adult male from The Creek, 8 mi. NE of West End, Cayman Brae, collected 25 August 1961 by a native for A. Schwartz. Paratypes: MCZ 69603-04, Cayman Brae, The Creek, 8 mi. NE of West End, 22 August 1961 ; MCZ 69605, Cayman Brae, 6 mi. NE of West End, 23 August 1961 ; MCZ 69606, Cayman Brae, The Creek, 8 mi. NE of West End, 23 August 1961 ; MCZ 69607-608, Cayman Brae, The Creek, 8 mi. NE of West End, 24 August 1961; MCZ 69609-11, Cayman Brae, West End, 25 Au- gust 1961 ; MCZ 69612-13, Cayman Brae, Spot Bay, 25 August 1961 ; MCZ 69614-17, Cayman Brae, The Creek, 8 mi. NE of West End, 25 August 1961. All specimens were collected by natives for A. Schwartz. I Fig. 1. Dorsal cephalic pattern of Tropidophis caymanensis sclnvartsi (MCZ 69618, type). Fig. 2. Dorsal cephalic pattern of Tropidophis caipnanensis cai/manensis (AS 16230). 4 BREVIORA No. 195 Diagnosis: a race of caymanensis with a normal scale-ro^\ count of 25-25-15, rather than 23-25-17 (T. c. cayma7icnsis) or 25-27-17 {T. c. porkcri), differing also from T. caymanensis in the coloration of the head and in a higher mean ventral count and from T. c. parl-cri in a somewhat lower mean ventral count. Description of type: An adult male, total length 485 mm, tail 50 mm. Body relatively stout, slightly compressed, head distinct from neck; scales weakly keeled except for five lower- most rows on each side, smooth scales becoming reduced to three lowermost rows posteriorly ; middorsal row enlarged and bi- earinate on posterior fiftli of body, ending anterior to vent ; dorsal scale-rows 25-25-17 ; ventrals 200 ; subcaudals 39 ; anal single ; spurs present ; preoculars 1/1 ; postoculars 3/3 ; temporals 3-4/3-4; supralabials 10/10, 4 and 5 entering the orbit; infra- labials 12/12. Dorsal pattern : two rows of 55 and 61 parame- dian dark spots, 2 to ^/U scales long by 2 to 2V> scales wide : light edging to spots mostly lateral, but some median edging occurs also ; lighter middorsal stripe present, broadly connected to median edges of spots. Sixth and ninth scale rows with fine, light centered, dark lines; row of small spots betw'een lines on sixth and ninth scale-rows ; two rows of staggered, dark spots on lower five scale-rows ; venter irregularly stippled with dark pigment (see Figure 3) ; lowermost row of lateral spots becom- ing enlarged and migrating to venter on posterior fifth of l)()dy: about seven spots on dorsum of tail ; subcaudal surface with two rows of al)()ut seven dark spots. Dorsal ground color of body light grey changing to cream below sixth scale-row and on ven- ter; tip of tail light. (Vphalic pattern: trapezoidal, dark, dor- sal ce]ihalic figure mucli invaded liy light stippling (see Figure 1 ) ; posterior center of figure light ; anterior portion of cephalic figure separated off to form interocular bar; nipple-like, ante- rior projection of interocular bar not extending beyond frontal ; dark, oval spot on internasals and prefrontals narrowly joined on each side to preocular portion of lateral head stripe, which contiiuics posteriorly onto neck to level of ninth ventral after which it is broken u]i into a lateral row of spots. VaridHini : The paratypes include seven males and nine fe- males; the largest male is 435 mm (the type); the lai-gest fe- male is 358 mm. The males of tliis series average larger than the females, having an av(>rage total lengt'i of 362.8 nun (202-425) : females avei-a type, average 195.6 (191-202) ; females 200 (195-205) : sub- caudals in males average 36.6 (34-39), in females 36.2 (31-39). 1963 CAYAIAN ISLANDS TROPIOOPITIS 5 All specimens have 1/1 preoculars and 3/3 postoeiilars ; all have the st-ale-row formula of 25-25-17 except for two with counts of 25-25-15. The model supralabial count is 10/10 with other counts of 12/10 (1) and 9/10 (2). Tails of males average 10.6 per cent of total length with a range of 9.1-11.8 per cent ; tails of females average 11.2 per cent of total length (10.8-12.07f ). Pattern and color are fairly uniform in the paratype series. In all of the specimens the dorsal cephalic pattern is much in- vaded and broken up by solid, light areas and by heavy stip- pling- of light pigment ; in all examples but three the interoeular ])igniented area is separated from the cephalic figure to form an interoeular bar; in the three exceptional specimens the inter- oeular bar is but narrowly joined to the dorsal cephalic figure by a median strip of dark pigment made hazy by light stippling. Typically there are several (usually 4) dark spots on each side of the lower labials. The dorsal pattern in the paratype series is much the same as in the type ; the average number of dorsal spots is 60.1 in males and 58.9 in females. Ventral coloration in the paratype series varies from a uniformly clear cream to a rather heavy, irregular, dark mottling over nearly the entire undersurface ; the typical ventral coloration for the series is like that described for the type. All specimens, except one with an incomplete tail, have light tail tips. Comparisons: A recently acquired series of 16 T. c. parkeri agree with Grant's (1940) original diagnosis of it as a 27 scale- row form with a ventral count of over 200; the ventrals of the nine males average 201.0 (199-204), those of the females 205.1 (201-212) ; the normal scale-row count for the series is 25-27-17 with variants of 25-27-16 (1), 25-27-18 (1), and 25-27-19 (1). Sell wart zi may be distinguished from parkeri by the possession of a normal scale-row count of 25-25-15. Chromatically there does not seem to be much difference in the populations; the fresh parkeri material has a richer, deeper coloration, whereas schivartzi is much lighter. The difference, however, is likely due to fading, since field color notes on one specimen of schwartzi ("dorsal ground color orange tan; dorsal blotches brown, not outlined posteriorly with orange") agree with the typical parkeri coloration. Parkeri tail tips are yellow; those of scJiwartzi are definitely light and probably faded from the yellow condition. The cephalic figure in parkeri tends to be broken up and in- vaded with lighter pigmentation much as in sch^rarfzi, though the figure is frequently broken into two or three disjunct por- tions in parkeri as o]:)posed to simple invasion by light pigment 6 BREVIORA No. 195 in schwartzi to form an erose margin and light center. The aver- age number of paramedian, dorsal spots in parkeri is 54.7 for males and 55.9 for females, slightly lower than the averages for schwartzi. In parkeri the females average larger than males; 11 females average 400.3 mm (230-564) ; nine males average 373.3 mm (306-444) ; the tails of male parkeri average 10.6 per cent Fig. 3. Dorsal midbody i)attern of Tropidophis caymanensis schwartzi (MCZ 69618, type). of the total length, those of females 9.4 per cent. The slight sex- ual dimorphism evident in average total length and average tail/total length ratios for schwartzi is present but reversed in parkeri. The ten specimens of caymanensis which were examined have a modal scale-row count of 23-25-17 with variants of 23-25-19 (1) and 25-25-19 (1). These agree closely with Battersby's (1938) type and paratype which had scale-row counts of 23- 25-17. Schwartzi is distinguishable from caytuancnsis by the possession of 25-25-17 scale-rows. Grant (1940) thought that the type of caymanensis Avas abnormal in having only 191 ven- trals ; he also mentioned a male with 183 ventrals. In his analysis, 1963 CAYMAN ISLANDS TROPIDOPIIIS 7 he apparently omitted both of these specimens from considera- tion and speculated that they might be members of a localized population along the north coast characterized by a reduced number of ventrals. The material at hand shows that similar low ventral counts are not restricted to the north coast; counts of 183, 187 and 191 are found on specimens from Georgetown in the southwest. Grant also mentioned five females and five males in his own collection which, along with the female para- type, average 205 and 193 ventrals, respectively. However, he only listed seven specimens in his own collection from Grand Cayman ; just what the provenance of the other three specimens is cannot be determined. Counts from the new material plus the counts on the type, paratype, and the individual with 183 ven- trals noted by Grant (Br. Mus. No. 1939.2.3.78) yield an av- erage of 192.6 (183-200) for seven females, 191.7 (183-195) for six males. Therefore, caymanensis is seen to average lower in ventrals than parkeri and schwartzi, there being almost no overlap between caymanensis (183-200) and parkeri (199-212). There are apparently no significant differences in body pattern and color between caymanensis and schwartzi. However, schwartzi can be distinguished from caymanensis on the basis of the cephalic figure. In caymanensis the cephalic figure is much more solidly pigmented with black and uniform in outline ; the edges are not highly indented and eroded by invasion of light pigment as in schwartzi, and the light areas within the figure are more restricted and well defined, not exten- sive and hazy in outline by virtue of light stippling (see Figure 2). Seven female caymanensis average 392.0 mm total length (222-488) and five males average 368.8 mm (192-538) ; tails of males average 12.2 per cent of the total length, those of females average 10.8 per cent of the total length. Once again, apparent sexual dimorphism is the reverse of that in schivartzi. Remarks: The author was able to do some collecting on Little Cayman in December of 1962. Some impressions were gathered concerning Tropidophis c. parkeri and are probably applicable to the species generally. A series of sixteen parkeri was col- lected during the visit to Little Cayman. These snakes gave the impression of being more secretive than T. melanurns, which is most frequently found prowling abroad at night. Parkeri was never collected in the open ; the majority were found beneath fragments of limestone which litter the island. Two were taken within rocks themselves, being secreted in eroded chambers ; one was taken among the leaves of a dead Agave, and another was 8 BREVIORA No. 195 found three feet off the gTOuiid in the leaf base of an epiphyte. One specimen was found approximately eight feet above ground in the roof of an outhouse, where it was partly concealed beneath corrugation in the tin roof. One specimen was found beneath rocks and litter under a sea-grape tree; it was this "beach situ- ation" wherein the natives claimed the snake to be most often found. A large female specimen found resting in a tin can dis- gorged a large Hula septe7ifrwnaUs shortly after capture. Many specimens autohemorrhaged freely upon capture or at subsequent disturbance; the "shutter" effect noted by Hecht, Walters and Ramm (1955) for Bimini Tropidophis was also observed; this effect is produced by the sudden and transient appearance of blood under the brille during autohemorrhage. Most of these snakes coiled in a tight, defensive ball and would so remain for long periods with their heads completely hidden. This habit is not in keeping with melanurus, which might be said to be less timid; experience with melanurus in the field indicates also that Contparafivc material: Tropidophis caymanensis caymanensis : USNM 108042-43, Grand Cayman; MCZ 44862-4, Grand Cay- man; AS 16209-11, 16230, Grand Cayman, Georgetown; AS X4876, Grand Cayman, Georgetown. Tropidophis c. parheri: rSNM 108009, Little Cayman; MCZ 44865 (type). Little Cay- man; MCZ 44866-8 (paratypes). Little Cayman; AS X489b- X4904, Little Cayman, western end; RT 36, Little Cayman, western end. Tropidophis melanurus: 79 specimens, Cuba and the Isles of Pines. LITERATURE CITED Anonymous 1961. Cayman Islands: Report for the rears 1959 and 196U. London: Her Majesty's Stationery Office, pp. 1-46: 10 figs., 1 map. Battersby, J. C. 1938. Some snakes of the genus Tropidophis. Ann. ilag. Nat. Hist., (11) 1:557-560. Grant, Chapman 1940. The herpetology of the Cayman Islands. Bull. Inst. Jamaica, Sei. Scr., 2:1-56, 13 figs. Hecht, Max K., Vladimir Walters and Gordon Ramm 1955. Observations on the natural history of the Bahaman pigmy boa, Tropidophis pardalis, with notes on autolieinonliage. Copeia, Xo. 3:249-251. Schwartz, Albert and Richard Thomas 1960. Four new snakes (Tropidophis, Dromiciis, Alsopliis) from the Isla de Pinos and Cul)a. Herpetologiia, 16('2) :73-90. ^>euo BREVIORA MnaseTLaim of Compsirative Zoology Cambridge, Mass. December 31, 1963 Number 196 A NEW GENUS AND SPECIES OF BATHYPELAGIC OPHIDIOID FISH FROM THE WESTERN NORTH ATLANTIC By Daniel M. Cohen U. S. Fish and Wildlife Service Washington, D. C. Through the kindness of Dr. Richard H. Backus of the Woods Hole Oceanographie Institution and Dr. Giles W. Mead of the Museum of Comparative Zoology I have had the opportunity to study the three fishes reported upon below. Although they were taken off the southern edge of Georges Bank, hardly a zoologically unknown area, they represent the first of their kind ever captured. And, indeed, they are so different from any other known ophidioid that still another name must needs be added to the already lengthy roster of genera in this group. ThaLASSOBATHIA gen. nov. Diagnosis. Chin barbel absent. Vertical fins continuoiLS; ventral fins each with two rays, originating close to the level of the posterior margin of the preopercle and a short distance behind the symphysis of the cleithra ; the ventral fins covered with thick skin, short, extending only to the level of the origin of the pec- toral fin; pectoral fin rounded, without separate elongated rays. Branchiostegal membranes separate ; gill rakers reduced in num- ber and size ; branchiostegal rays seven. Preopercular spine pres- ent beneath skin. Anterior nostril lacking a tube, placed high on snout. Teeth present on premaxillary, vomer, palatine and dentary, lacking on basibranchials. Maxillary not expanded pos- teriorly, not sheathed. Premaxillary not protractile, bound to the snout anteriorly by a broad frenum. Head compressed, its height greater than its width. Eyes well developed. Body compressed, relatively short and stubby. Two lateral lines. 2 BREVIORA No. 196 Peritoneum dark. Swimbladder reduced, its cavity completely occluded ; bladder not bolind to vertebrae. Neural spines on abdominal centra not depressed ; their tips not truncate. Type species. Thalassohatlua pclagica sp. nov. The name Thalassohathia is derived from the Greek GaXaaaa, the sea ; ^aO'jq, deep. The gender is feminine. Relationships. The main problem encountered in establishing this genus has been to determine its relationships rather than to distinguish it from any presently known genera. The curiously reduced ventral fins and the general shape and consistency of the body might suggest relationship with the liparid fishes ; however, there is not the slightest trace of either a suborbital stay or a sucking disc. In addition, the branchiostegal membranes are sep- arate from each other and from the isthmus, and the gill opening extends far up on the side of the body in typical ophidioid fashion. The last character also separates ThalassohafJiia from the Zoarcidae (not including Lycodapus) . The non-protractile premaxillaries and attached broad frenum suggest the condition of the upper jaw found in salariine blennnies ; however, very different dentition, branchiostegal membranes, ventral fins, caudal fin connections and other characters separate Thalassohathia from the Salariinae. In addition, the absence of a one-to-one correspondence between dorsal and anal fin rays and vertebral centra separates Thalassohathia from any of the blennioids. Among the ophidioids the closest known relatives of Thalasso- hathia appear to be the Bythites — Oligopus — Cataetyx group of genera recently discussed by Cohen (in press). Some reasons for deriving Thalassohathia from this group are the similarly compressed and foreshortened body, the typical bythitine denti- tion, and especially the double lateral line system, which greatly resembles that found in some of the species of Oligopus (Cohen, in press). Although at first glance Thalassohatlna does not appear to be a bathypelagic fish, closer examination reveals that many of its characteristics are adaptations to a midwater life. The skeleton is so poorly ossified that X-ray photographs are very indistinct. Scales are completely lacking. The swimbladder is a small, tissue- filled organ similar to those which have been found in many other species of bathypelagic fishes. The above characters and their significance in floating fishes have been treated by Denton and Marshall (1958) and Marshall (1!)()0). Other structures wbich may help to keep Thalassohathia aloft are its subdermal fatty 1963 NEW OPTTIDIOID FISH 3 layer and enlarged liver. Among the most characteristic features of ophidioid fishes are the naked ventral fin rays which probably serve as tactile organs in benthic species (Herald, 1953, and serve as tactile organs in benthic species (Herald, 1953 and Briggs and Caldwell, 1955, have confirmed this for two species of benthic Ophidiidae). These structures would of course be useless in midwater, and Thalassobathia has reduced ventral fin rays which are covered with thick skin. Because of its midwater habitat, Thalassohathia should also be compared with the ophidioid subfamily Aphyoninae (Nybelin, 1957) which comprises a small group of bathypelagic species. Aphyonines are small, pale fishes with thin, scaleless skin, weak musculature, reduced ossification and no swimbladders. Although they are livebearing, they have a copulatory apparatus which is very different from that found in benthic livebearing ophidioids. They also lack well developed eyes and lateralis systems. I believe that neoteny has played an important part in the evolution of the aphyonines. Tlialassuhathia , with its large eyes, well devel- oped lateralis system and bythitiiie dentition, seems, on the other hand, to be one more offshoot of the highly adaptive Bythites group of genera, and although it shares a common environment with the Aphyoninae, the two are not at all closely related. Specimens examined in preparing the foregoing section are listed in Cohen (in press). In addition I have examined the aphyonine material described by Nybelin (1957), specimens of Aphyonus in the Museum of Comparative Zoology and in the British Museum (Natural History), and specimens of Bara- thronus in the U. S. National Museum and the Chicago Natural History Museum. I thank the fish curators in Cambridge, Chi- cago, Goteborg, London and Washington for their kindness. Thalassobathia pelagica sp. nov. Figures 1, 2 Holotype. MCZ 42161, 221 mm standard length, collected by Richard H. Backus on board the "Capt. Bill III," haul RHB 913; October 13, 1962; 39° 26' N., 71° 00' W. to 39° 32' N., 71° 00' W.; depth to gear 390 fms. (approx. 713 m), depth to bottom 1400-1350 fms. (approx. 2561-2469 m) ; time 1020 to 1505 hrs. EDT ; 64 foot open midwater trawl. Paratypes. MCZ 42162 (2) ; data as for the holotype. Description. Body compressed and foreshortened, greatest depth between three and four in total length. Head relatively BREVIORA No. 196 S3 O tB .9 O t-i be a> a S a o o 'Si Si. e e o =0 =0 e e 2 1963 NEW OPIIIDIOID FISH 5 short, between five and six in total lengtli. Snout rounded, nioutli slightly inferior. Interoi-bital area flat or slightly convex. Out- lines of head converging anteriorly when viewed from above. Eye prominent, horizontal diameter of cornea about three and one-half in head. Posterior nostril an elongate slit immediately in front of ventral half of eye, anterior nostril small and round, closer to tip of snout than to orbit. Body quite flexible due to weak ossification, feeling much the same as icosteid fishes. All fins covered M'ith thick skin. The dorsal fin appears to originate a short distance behind the level of the posterior margin of the opercle ; it is preceded by rayless pterygiophores and ptery- giophores bearing very reduced rays. The exact number of these elements is in doubt as their weak ossification makes them difficult to count on an X-ray photograph. Pectoral fins inserted ver- tically, broadly rounded; in the holotype and smallest paratype the pectoral fin extends about two-thirds of the way from its origin to the vent, in the other paratype the pectoral fin extends only about one-half of the pectoral origin to vent distance. The ventral fins much reduced and covered with thick black skin (Fig. 2C), the longer, inner ray extending only slightly beyond the level of the base of the pectoral fin. One pectoral fin was cleared and stained to verify the number of rays. Teeth short, needle-like, extending on the dentary in a single line from the angle of the gape to the symphysis, where a broader patch of teeth is present ; premaxillary teeth arranged similarly to those on the dentary ; vomerine teeth in a roughh- circular patch; palatine teeth in an elongate patch, about three or four times longer than wide. Suprabranchial tooth patches present; however, I have been unable to find any basi-branchial teeth. Tongue massive, lacking teeth and lacking an anterior prow-like extension. The first gill arch with rakers reduced to a few fleshy flaps and protuberances laterally (Fig. 2A), medially (Fig. 2B) with seven or eight small rakers. A prominent pseudobranch located lateral to the upper arm of the first arch. Gill filaments of moderate length, about equal to diameter of lens. Color brown-grey over the body, the grey due mainly to a mucous coat ; the fins and head darker. Prominent pores on the head. The lateral canal w4th a single pore near the upper angle of the opercle ; supraorbital canal with two or three pores, one above the upper nostril, another below and slightly ahead of the upper nostril, a median pore in the interorbital region of the largest paratype ; infraorbital canal BREVIORA No. 196 Fig. 2. Thalassobathia pelagica. A, First gill arch from left side of smallest paratype, lateral view. B, First gill arch from left side of smallest paratype, medial view. C, Ventral fins of largest paratype. Drawn by Mildred H. Carrington. i 1963 NEW OPHIDIOID FISH 7 with nine pores, the anteriormost below the anterior nostril and immediately behind one of the supraorbital pores, six in an irregular row in the loose skin above the upper jaw, and two behind the eye ; preoperculo-mandibular canal with five or six pores located along the ramus of the lower jaw; in the smallest paratype the pair of pores located closest to the midline of the lower jaw have joined to form a single pore. Lateral line with an upper and lower section, the upper line originating above the opercle and arching upward and back to the level of the origin of the dorsal fin, where it extends pos- teriorly about midway between the dorsal profile and the midline ; the dorsal line disappears at about two-thirds of the distance between the posterior margin of the opercle and the base of the tail; the ventral line originates slightly behind the level of the vent and extends posteriorly to the tail in the midline of the body. Each of the two lateral lines along the body is marked by about 30, small, dark, papillae ; similar structures are sparsely dis- tributed over the body and head. Each specimen has a broad, fleshy hood over the genital area. The largest paratype has a pair of large ovaries which contain small eggs at the bases of numerous villi. The smallest paratype contains small paired ovaries with developing eggs. The swimbladder of the smallest paratype is a white, bean- shaped organ about 14 mm long. The lumen appears to be com- pletely occluded by white tissue.^ The liver is prominent in both paratypes, and the thick skin has a subdermal fatty layer. The brain case is cartilaginous. The olfactory nerves extend anteriorly in the orbit and are separated from each other by the cartilaginous interorbital. All three specimens bear numerous round marks left by squid tentacles. The smallest paratype has the top of its head bitten off. LITEEATURE CITED Briggs, John C. and David K. Caldwell 1955. The characteristics and distribution of the spotted cusk eel OtopJiidium omostigmum (Jordan and Gilbert). Quart. Journ. Florida Acad. ScL, 18(4) : 285-291. 1 I am indebted to N. B. Marshall for his kindness in examining the swim- bladder. 8 BREVIORA No. 196 Cohen, Daniel M. (in press). A review of the ophidioid fish genus Oligopus with the descrip- tion of a new species from West Africa. Proc. U. S. Nat. Mus. Denton, E. J. and N. B. Marshall 1958. The buoyancy of bathypelagic fishes without a gas-filled swim- bladder. Journ. Mar. Biol. Assoc. U. K., 37: 7.')3-767. Herald, Earl S. 1953. Spotted cusk-eel, the strange fish that stands on its tail. Cali- fornia Fish and Game, 39(3): 381-384. Marshall, N. B. 1960. Swimbladder structure of deep-sea fishes in relation to their systematics and biology. Discovery Eepts., 31 : 1-122. Nybelin, Orvar 1957. Deep-sea bottom fishes. Eepts. Swedish Deep-Sea Expedition 2, zoology (20) : 250-345, pis. 1-7. Table 1 Counts and measurements in mm of Thalassobathia pelagica Ilolotype Paratypp Parafyiic Dorsal rays Anal rays Pectoral rays Vertebrae'^ Caudal rays Standard length Head length Snout Orbit Interorbital Upper jaw length Preanal Greatest body depth Body depth at vent Pectoral fin length 1 Not including hypural. 79 72 74 65 58 61 23 24 27 49 48 48 10 10 10 221 173 200 44.0 13.5 42.0 10.0 — 7.4 11.9 — 11.0 16.0 — 13.3 30.0 21.8 22.2 87.9 71.0 87.2 59.4 47.2 64.0 54.1 45.4 50.0 28.2 23.4 25.4 X^uJ BREVIORA MMsemim of Coeiparative Zoology Cambridge, jMass. Deckmbek 31, 1963 Number 197 ANOLIS WHITEMANI, NEW SPECIES FROM HISPANIOLA (SAURIA, IGUANIDAE)i By Ernest E. Williams Miss Cochran in 1941 commented on the apparent extreme variability of Anolis cyhotcs cyhotes, mentioning among other points that "some of the specimens have heavy keeling on the ventrals ; in others from precisely the same locality, taken at the same time, the ventrals are perfectly smooth." Inspection of material referred to cyhotes in the Museum of Comparative Zool- ogy, the United States National Museum, and in the unreported collections of the American Museum of Natural History revealed that this phenomenon of sharp dimorphism in regard to keeling was curiously localized and that many of the localities were in the Cul de Sac Plain in Haiti and in its continuation in the Dominican Republic. A closer look at the specimens from this area showed further that there were indeed two kinds of cyhotes- like anoles represented, and that they were in fact reported as taken at exactly the same localities by the same collector on the same dates. There seemed, however, to be not a difference in a single character but in several independent characters. The suspicion thus arose that species difference and not intraspecies dimorphism was involved. The suspicion was sharpened when P. S. Humphrey, collecting for Yale University and the Univer- sity of Florida in 1959, lirought back only two cybotcsAike anoles from the Cul de Sac region, one of each type. The fresh material permitted also an increased confidence in a color difference which had seemed to exist in the specimens long in collections. When, therefore, in August 1959, E. E. Williams and A. S. Rand planned a visit to Haiti, one of the objectives was to obtain and examine alive a series of the keeled cyhotcs-like anole from the Cul de Sac. 1 Notes on Hispaniolaii herpetology no. 9. 2 BREVIORA No. 197 This objective was achieved, and the keeled form in the Cul de Sac Plain proves indeed to be a new species, which we name in honor of M. Luc Whiteman of Port-au-Prince who has assisted in the recent Haitian explorations. Since 1959, more material of A. whitcniaiii has been obtained by A. S. Rand and J. D. Lazell, Jr. at Mole St. Nicolas on the northwest peninsula of Haiti — an arid area very similar to the arid portions of the Cul de Sac Plain in which ivhitemani was first found. In addition to these (MCZ) specimens and the ma- terial in the United States National Museum (T'SNM) and the American Museum of Natural History (AMNH), further speci- mens have been found in the collections of the Hamburg Museum. The latter, from Monte Cristi in the Dominican Republic, extend the range of ivhitemani considerably to the east but still along a dry coast. Anolis whitemani new species ^ Type: An adult male, MCZ 60055, road to Eaux Gaillees, HaitL E. Williams and A. S. Rand collectors, 13 August, 1959. Paratypcs: Haiti. MCZ 60056-9, same locality as the type, E. Williams and A. S. Rand coll. MCZ 62844, same locality as type, A. S. Rand and J. Lazell coll. YPM 3193, Eaiix Gaillees, P. S. Humphrey and S. Van Vleck coll. AMNH 70588, Eaux Gaillees, A. Curtiss coll. USNM 117217-8, Trou Caiman, A. Curtiss coll. USNM 54189, Thomazeau, A. Curtiss coll. MCZ 62827-43, M<>lr St. Nicolas, A. S. Rand and J. Lazell coll. Dominican Republic. AMNH 50147, 50200, MCZ 61843, Las Baitoas near Duvereje, W. G. Hassler coll. Hamburg Museum 5198 (3), Monte Cristi. Diagnosis: Very close to A. cybotes but differing in squama- tion (dorsal scales larger, the middorsals hardly larger than adjacent scales, rather than abruptly larger; the ventrals smaller, hardly larger than middorsals, narrow and keeled, rather than wide, smooth and cycloid), and in color (body color very pale, nearly grey instead of distinctly brown, dewlap unmarked Avhite, rather than grey with yellow streaks or yellowish or pinkish with grey streaks). Description. Head: Head massive, snout to posterior border of eye about as long as tibia. Head scales mostly smooth. Five scales across head between second canthals. A shallow frontal 1 Named for M. Liic Whiteniaii, who has helpetl so iiiiuh iu ainassiug the recent collections from Haiti. 1963 ANOLIS WHITEMANI 3 depression. Naris in front of canthal ridge. Anterior nasal scale in contact with rostral. Supraorbital semicircles narrowly in contact, partly in con- tact with supraocular disks. Supraocular disks consisting of about 8 enlarged keeled scales separated by 4-5 rows of granules from the scales of the supraciliary rows. One to two elongate supraciliaries continued backward by a double row of moder- ately enlarged scales. Canthus distinct, canthal scales 4, the second largest, decreasing graduallj' forward. Loreal rows 6-7, the lower row the largest. Supratemporal scales somewhat en- larged, flat, grading into the enlarged scales surrounding the interparietal. Interparietal larger than ear, separated from supraorbital semicircles by 1-2 scales. Suboculars separated from supralabials by one row of scales, anteiiorly separated from the canthal ridge by one scale, pos- teriorly continued behind eye for a short distance. Six supra- labials to center of eye. Mentals somewhat broader than long, in contact posteriorly with 4 throat scales. Infralabials narrow, in contact with 3 large, wide, nearly lunate sublabials. Throat scales small, swol- len, not keeled, only the anterior ones elongate. Trunk: Middorsal scales keeled, hexagonal, not much larger than adjacent scales which grade very gradually into granules on flanks. Ventrals not much larger than middorsals, relatively narrow, distinctly keeled, the keels in lines. Postanal scales small, poorly differentiated. Gular fan: Moderate, scales cycloid, keeled, larger than ventrals. Limbs and digits: Head and foot scales multicarinate. About 17 lamellae under phalanges 2 and 3 of fourth toe. Largest scales of limbs unicarinate, those of arm as large as, of thigh larger than ventrals. Tail: Compressed. Each verticil surmounted by 4 keeled scales, ventrally 3 pairs of somewhat larger keeled scales per verticil. Color in life: Four topotypes taken on road to Eaux Gaillees: 1. Type S . Dewlap white. Ground color grey with indistinct transverse bands. Scattered brown spots on flanks and on sides of belly. 2. 6 . Dewlap pure white. Dorsum grey and cream. Pattern very obscure. Venter pure white. 3. 9 . Dorsum pale grey. Belly pure white. 4. c^ . Juvenile. Dewlap pure white. Dorsum cream and grey. BREVIORA No. 197 Notes by W. G. Hassler 1935 (Las Baitoas paratypes) : " Anolisf-like cybotes: in grass on mountain side above road and forest. These ancles very grey or grey-brown. Pattern faint. Sides slight!}' yellowish grey. Belly and throat white. S similar to 9 . Faint yellowish brown-grey longitudinal bars like ordi- nary cybotes but much lighter, less red and all very faint. Throat (very faint vermiculations), belly and fan WHITE." SPECIES STATUS AND CHARACTER DIFFERENCES The squamation and color differences between whitemani and cybotes in the Cul de Sac Plain are striking but hardly more striking than those between cybotes and " Audantia" armouri Cochran from the Massif de La Selle. New evidence, however, shows that the latter taxon, formerly considered a distinct genus, intergrades with cybotes at intermediate elevations in the foot- hills of the La Selle range, e.g. at Furcy, 10 miles S. of Port-au- Prince. Some of the s(iuamation differences between whitemani and cybotes are in fact rather similar to those between armouri and cybotes — although they are greater (Table 1, and Figs. 2-4). Thus the character differences are not in themselves great enough to establish species status for whitemani. X Anolis whitemani n. sp. in central and western Hispaniola Fig. 1. Localities for Anolu whitemani n. s]). 1963 ANOLIS WliriEAIANI Table 1. Differences separating' cybotcs-Viko anoles in the Port-au-Prince VEXTRALS TEMPORALS MIDDORSALS BODY PATTERN DEWLAP COLOR area. icliit( iiiaiii keeled, ii;irio\v. cybotes armouri (Port-au-Prince and vicinity) smooth, cycloid, smooth, cycloid, wide. very wide. granular, but finely granular, granular, but larger than ilank even smaller than larger than flank scales. flank scales. scales. enlarged but grading into flank scales. abruptly larger than flank scales. enlarged but grading into flank scales. BODY COLOR pale tan. patternless or with pale grey transverse mark- ings. white brown to reddish, olive. patternless as often with bold adults or with rhombs on flanks greenish lines on = persistence in flanks. adults of a juvenile pattern. pale yellow or rarely pale pink with white or greyish streaks or greyish with yel- lowish streaks. pinkish or yel- lowish Avith green- ish smudges centralh'.l SPECIES STATUS AND DISTRIBUTION In the final analysis the jndgment that whiteniani is a full species rather than a morph of cybotes, as Cochran (1941) be- lieved, has been based on observation of the habits and habitats of the two species in the Cul de Sac region. The topotypic series of A. whiicmani was taken in open dry scrub along the road to Eaux Gaillees in the full sun of late morning. The same day quite typical A. cybotes were obtained in the grove of trees at the spring in the village of Eaux Gaillees and in a similar grove of trees at Manneville further east on the same road, again at a spring. 1 Color notes from Information provided by Dr. Albert Schwartz. 6 BREVIORA No. 197 In the Cul de Sac Plain such groves of quite large trees around springs are quite literally oases in a thorn scrub desert. The fauna, like the flora of these oases, is characteristically that of rather moist areas. The faunas of these oases are obviously disjunct, in effect island populations derived from the faunas of the moist forests of the mountain foothills to the south and to the north. The con- trasts between the fauna within an oasis and that outside it are sharp and very characteristic. Within each oasis are animals that could be seen also in a pension garden in Port-au-Prince itself, e.g. Anolis chlorocyanus, A. cyhotes, A. distichus, Celestus costatus. Outside, in the dry scrub, are Anolis ichitemani, Celestus curtissi, Leiocephahts scmilineatus. A. cyhotes and A. whitemani thus live in the same area but are sharply segregated by ecological preference and sharply dis- tinguished by color and morphology. It is very important that the morphological differences are greater between cyhotes and adjacent whitemani than between cyhotes and more distant armouri; this is a demonstration that these two populations are behaving as good species. The discovery of A. whitemani on the dry northwest coast at Mole St. Nicolas is a very valuable addition to the data on its distribution. When its occurrence seemed limited to the Cul de Sac Plain, whitemani was something of an anomaly, since this area has only relatively recently emerged from under the sea. However, with the realization that it is probably to be found all along the dry northwest coast of Haiti and is known as far east as the similar dry coast at Monte Cristi on the north coast of the Dominican Republic, its recent extension into similar arid habi- tats in the Cul de Sac Plain becomes very plausible. It is prob- ably an autochthon of the dry west and northwest coast of tlie "northern island" — the portion of Ilispaniola north of the former seaway through the Cul de Sac. (For a discussion of the importance of the northern and southern islands in His])ain()la zoogeography, see Williams, 1961.) It is noteworthy that Humphrey obtained large series of the species cyhotes on Gonave Id. (an island mostly dry but again with oases around the springs) but no ivhifemani. The latter may thus really be absent there. This requires further confirma- tion, but the fauna of Gonave has been derived by a limited ami erratic sampling of the faunas of the neighboring mainland. The absence of any individual species is therefore not in its(^lf sur- prising and need not imply oversight in collecting. 1963 ANOIJS WniTEMANI ACKNOWLEDGMENTS T am indebted to Dr. Albert Schwartz, Dr. Philip Humphrey, Dr. Doris Cochran, Mr. C. M. Bogert and Dr. W. Ladiges for the privilege of examining specimens. National Science Founda- tion Grant G 5634 and a grant from the American Philosophical Society provided funds for expeditions to Haiti. National Sci- ence Foundation Grant G 16066 continues to support research on Hispaniolan and other anoles. REFERENCES CITED CochrjVN, D. 1941. The heipetology of Hispaiiiola. Bull. U. S. Nat. Mus., 177: i-vii, 1-398. Williams, E. 1961. Notes on HLspiuiiolau heipetology. 3. The evolution and relation- ships of the Anolis semilineatus group. Breviora, no. 136: 1-8. Fig. 2. Anolis whitemani n. sp. Paratype. MCZ 60056. Top: Side view of head. Middle : Dorsum at midbody. Bottom : Venter. 8 BREVIORA No. 197 Fig. 3. Anolis cybotes cybotes. MCZ 59883. Top: Side view of head. Mid- dle: DcTSum at niidbody. Bottom: Veuter. Fig. 4. Anolis eyhotcs annouri. MCT. ClO.'il. Top: Wide view of head. Mid- dle: Dorsum at niidliodv. Hotloiu: Venter. BREVIORA MmseiMii of Comparative Zoology Cajjbridge, :\Iass. March 10, 19()4 Number 19S AMPHISBAENA SCHMIDTI, A THIRD SPECIES OF THE GENUS FROM PUERTO RICO^ (AMPHISBAENIA : REPTILIA). By Carl Gans Departineiit of Biology, State University of New York at Buffalo, Buffalo 14, New York The island of Puerto Rieo is iiiiiciue auiouy- the islands of the Greater Antilles in possessing two almost eertainly sympatrie species of Amph'shaena. One of them, A. caeca, ranges over the entire island, while the second, A. hakeri, is restricted to the central portion of the northwestern corner of Puerto Rico (Fig. 1). .1. caeca has been shown to vary more widely in the western than in the eastern portion of its range (Grant, 1932; Gans and Alexander, 1962). A recent study of the systematics and variations of the sev- eral Antilles species (Gans and Alexander, 1962; here followed for terminology) disclosed three individuals from Puerto Rico that clearly diifered from both described forms in five character- istics, i.e. in more ways than the former differed from each other. The specimens were left incerfae sedis for three reasons : two specimens came from a rather old collection ; the third specimen had been collected on the diametrically opposite end of the island; and only these three out of more than 200 Puerto Rican siKM'hnens showed this character pattern. Dr. 11. lleatwole has now made available three additional speci- mens that stem from two coastal localities adjacent to that of the first two specimens. They make it desirable to call attention to the probable existence of yet a third species of Anrphishaena on the island, a situation that may have some interesting zooge- ographical implications. 1 Notes on amphisbaenids No. 11. BREVIORA No. 11J8 jk h- V o-^ -^ s 0 i I 0 < 0 " ' UJ - ^ ^ ^ it > \ 0 2 0^ ^ «^= 1 N^ • ® 0 n D f Q -°o — V 0. ^ ° 0 V m ** jO 3 W ^ 0 ^^ L # 7 " ^ *" 0 % m 0 ■= c o "= 0 \ 0 1 -o 1 0 H o ^ -' -; 0 0 0 ' \ ^ 9^ c 0 t 1 ^o J OJ 0 > ■3 I / ° ' 0 ''° >^ 1 -> ^ C Y / o \ 0 0 ^ 0 u \ f u 0 >- o : < 0 E 1? ■ 3 7- s. 0 D *" # 1 tf> O a> ^* 0 (^ 1 * 0 u 1 '^ 0 1 ■ ° ^, ■ O lO ^^ 1 ^ m 1 ■o „ m V "° 4, f J J D — 0 1 ■ 0 0 N 1 f "O 05 0 D \ ■ ** 0 % ■ ^ c 1 ■ ^ 0 D t r ■0 o 3 3 3 J |0 0 II / o i) C ^^L i> c 0 ^^^ ■ •' 0 i- 0 ^A 11 0 0 ' ^\ ^ ° z a 0 0 c % \ 0 1 CO ._ 0 0 cA. 0 o 0 0 / ■o 1 "o LX 0 i K " 2 o - O =M pl4 >-, a< (1) 1— ^ 0 OJ • i-< CI +— ■^ 0) bi i^ 'r^ r Oh ^ r— w s w 0 'A O K ^ -; O .Si "7^ o a> _ "Z o 5 o ,„ o i' o 19G4 AMl'lllSHAEXA SCIIMIDTI 3 1 take o-reat plPHsurc in iiaiiiiim this t'onii for the late Karl Patterson Seliuiidt. Tlic spcciniciis IicIohl;' in tlic collections of the Carneo'ie Mnscuni (CM), tlic Tniversitetets Zoolop^iske Mn- senni of Kobenhavn (KM), and the TTniversity of Puerto Rico at Kio Piedras (PIMvliP). One of the individnals from Kio Piedras has been sent on cxchano-e to the ]\Insenni of Comparative Zoology (MCZ). T am ui-atefnl to tlie curators of the several institutions, ])articularly to Dr. F. W. I>raestru]) who went to considerable tronble to check the original catalogs, and to Dr. IT. Ileatwole and students who made an effort to collect addi- tional specimens. A. A. Alexander and E. E. Williams contrib- uted by discussing this situation. Dr. Virginia Cunnnings ])re- pared the drawings and Miss Charlyn Rhodes furnished technical assistance. These studies are supported by Grant G-21819 from the National Science Foundation. KEY TO PUERTO RICAN AMPHISBAENA 1. Body aiinuli more than 205, caudal aiinuli fewer than 18, tail markedly shorter (Fig. -4), generally no enlarged parietats 2 Body annuli 205 or fewer (198-202), caudal annuli 18 or more (20-22), tail markedly longer (Fig. 4), parietals very large sclimidti sp. nov. 2. Body annuli 218-230, tail slightly longer (Fig. 4), internasal suture con- siderably shorter caeca Body annuli 239-255, tail slightly shorter (Fig. 4), internasal suture con- siderably longer taTceri Amphisbaena sciimidti s]). nov. Diagnosis: A form of Ai)tpJiish(i( iia without fusion of head segments ; with markedly enlarged parietals ; having 198 to 202 body annuli along the ventral line; 20 to 22 caudal annuli; 14 dorsal and 1(3 to 17 ventral segments to a midbody annulns; two rows of postgenial and one row of postmalar chin shields and 4 precloacal pores. The tail is cylindrical and its end rounded. The autotomy constriction is noticeable at the seventh to eighth postcloacal annulus and autotomy takes place here. Holotype: MCZ 73115, a female collected by M. J. Yelez, Jr., at Orilla (Cunta) Carrcamo, Isabella, Puerto Rico, on 21 Febru- ary 1960. Parafypes: UPRRP 12!)(), a male collected with the holotype; CM 36277 from Salinas, Puerto Rico ; KM R-4414 and R-4416 collected by Dr. Meinert at Aguadilla, Puerto Rico, in January 4 BREVIORA No. 198 1892; and UPRRP 2502 colloeted by II. IIeat\v()](> along high- way P.R. 681, 2 miles east of Cueva del Indio, Puerto Rieo, on 14 January 1968. DiscKssiou: The demonstrated differenees indicate that the new form is very distinct morphologically. They do not show whether it is a separate species, a (coastal?) race of A. caeca, or of A. hokcri. A case may be made for each of these. I have de- cided to treat the form as a full species for the following reasons : (1) The degree of morphological difference is greater than that fouiul between consiiecific populations among related species. It exceeds that between the two presently recognized species on the island. (2) The sample from Salinas contains five specimens of A. caeca and one of A. schuiidti. The coastal record of A. caeca from Arecibo lies midway between two records of A. fichinidti. Both of these cases suggest a limited degree of symjiatry. (3) Specimens of A. schniidfi from opposite ends of the island in- dicate almost no geographic variation, in contrast to the situation of A. caeca. (4) Designation of the form as a full species does not prejudge the open question of its affinities. Description: Meristic characters are listed in the table. Fig- ure 2 shows the head scalation. Figure 3 the segmentation of cloaca and tail, and Figures 5 to 7 are photogi-aphs showing details of color i)attern and midbody segmentation. Figure 4 shows the body proportions. Specimens are a dai'k bi-owii with vei-y slight ventral counter- shading. None of the sjiecimens show a droi)ping out of pigment on the midventral area. The smaller specimens have each seg- ment more or less uniformly ])igmented, but the two largest show a considerable additional darkening of the rectangular segmental centers. The intersegmental raphes are always lightened. Dorsal surfaces of head and tail are a uniform dark violet brown in freshly preserved specimens. The head scalation is characterized by lack of major fusions. An azygous rostral invisible in dorsal view is followed by a pair of contacting nasals, very large elongate prefrontals, frontals, and wide and long parietals. The latter may be divided (trans- versely) or not. There are three supra- and three infralabials, but the angulus oris lies very slightly anterior to the posterior edge of the third paii'. The second of each series is by far the largest ; indeed, the second infralabials are larger than any scales but the prefrontals. A small segment lies innnediately in line with the slit of the mouth ajid contacts the posterior half 1964 AMl'lllSBAEXA SCliAllUTi Fig. 2. AmphishaoKi .scliniidfi. Dorsal, lateral ami venti-al views of the head of the holotype, MCZ 73115 from Isal)ella, Puerto Rico. The line equals 1 mm to scale. (V. Cummings, del.) 6 BREVIORA No. 198 of the enlarged lateral postinalar segment. The sutures between the supralabials run at angles of 45° to the slit of the mouth. The oeular is (luadrangular. The mental is T-shaped and nuich lai'gcr tlian the tiny first infralabials. The third infralahials are (piite narrow. The post- mental is almost as large as a second infralabial. The triangular tips of the two segments of the first ])Ostgenial row embrace it posteriorly. There are three second postgenials. The anterior tip of the median one contacts the postnu^ntal in some specimens. The malars are relatively small and contact the second and third infralabials, but are clearly excluded from contact with the postmental. The segments of the ])ostmalar row ai-e relatively long, the lateral ones are almost wedge-shaped. Only the anterior half of the lateral ones contacts the third infralabial ; the pos- terior portion reaches the small segment back of the angulus oris. The postmalar row is thus counted as the first body annulus Fig. 3. Aniphisbacna sclnnuUi. \'('iiti;il view of cloacM ;iihI tail of the holotype, MCZ ZHllf) from Isabella, I'licito Rico. The line e(|iia Is 1 nini to scale. (V. Cuniniings, del.) 1UG4 AMl'JllSUAENA SCIIMIDTI as well. Tlu' lateraliiiost segments of the row are of the same width as the mahirs, and give the impression tliat the hitter have been split. Dorsally, the large segments of the first body annnhis curve to contact the sides of the frontals. The dorsal poi'tion of the second annulus consists of elongate segments that increase in E E c 0) 22 18 14 10 ♦ ♦ o o o OD O O 00 • 4 ♦ s chmid ti o caeca • baker i 8 16 24 Snout - Vent Length - cm Fig. 4. Amphisbaena. Scatter diagram showing plot of tail length versus snout-vent length for all specimens of A. schmidti and A. haktri, as well as specimens of A. caeca from western Puerto Rico. Inclusion of eastern Puerto Rico material of A. caeca would mask the difference between A. caeca and A. bal-eri. hut not affect that between the former and A. scJimidti. length toward the middorsal line. The posterior edge of the second annulus shows no forward curvature. There is no dor- sal intercalated half-annulus, though the parietals are split into two pairs in four specimens and on one side of another. The head is pointed, depressed, of horizontally oval cross- section. The lower jaw is but slightly shorter than the upper. 8 BREVIORA No. 198 Fig. 5. Ainp}ii.sb(t( H(i .scliiuulti . Dorsal, lateral and ventral views of the head of UCZ 73115, to show uniform coloration and the effect of the bulg- ing temporal musculature. 11)64 A.MI'HISBAENA SCHMIDTI 9 The bulji'c of the t('iii|)()i-;il itinsculature is very noticeable, pro- ducing a concave folding of tlie skin along the middorsal raphe and a clear distinction of the head from the narrower neck. There are lf»S to 20'2 body anmili from the back of the third infralabial, np to and inehiding the pore-bearing precloacals. The second throngh fifth annuli are ventrally nuu'h narroAver than the succeeding ones and their segments do not line nj) with those following. There is no pattern irregularity in the "pec- toral" region, nor are there intercalated dorsal half-annuli. There are ]4 dorsal and 16 to 17 ventral segments to a midbody annuhis. Fig. 6. Antplii.shaciia sclnuidfi. Dorsal and ventral views at midbody of MCZ 73115 to show segment proportions and coloration. The cloacal region is characterized by 4 round precloacal pores which are strongly expressed in males as well as females. There are 6 precloacal and 9 to 12 post cloacal segments and 3 (once, unilaterally, 4) lateral rows. The extreme lateral postcloacal segments lie almost directly laterad of the extreme lateral seg- ments of the curved precloacal shield, so that one is tempted to count them as precloacals. The autotomy annulus falls on the seventh to eighth postcloacal annulus and autotomy takes place 10 BREVIORA No. 198 Fig. 7. Amphisbaena schmidti. Ventral view of cloaca ami tail of MCZ 7.3115 to show pore size (in a female) and coloration. here. Specimens have 20 to 22 caudal anniili. The cross-section of the tail is circular throughout and the distal tip is capped by a hemispherical portion. The lateral sulci are distinctly marked from back of the first quarter of the trunk length to the level of the cloaca. At their widest they are narrower than the width of one of the l)order- ing segments. The dorsal and ventral sulcus and the lateral sulcus in the anterior quarter are indicated only b}- the align- ment of intersegmental sutures. The dorsal segments of a midbody annulus are approximately one and one-quarter times as long as wide, while the ventral ones are one and one-half times as wide as long. Habits: Examination of the KM catalogs (by Braestrujv) indi- cates that Dr. Meinert (an entomologist) excavated some termite nests on the days when the two paratypes were taken, and may have found the specimens at that time. The Cueva del Indio specimen was collected under a fallen ])alm leaf on liunuis. Range: Coastal Puerto Rico. LITEEATURE CITED Gaks, Carl and A. Allan Alexander 19Hi2. Studies on amphisbaenids (Amphisbaenia, Reptilia). -. On the aniphisliaenids of the Antilles. Bull. ^ilus. ('(iiii]i. /ool., vol. ll!8, no. 3, pp. 65-1.58. (iRANT, Chapman lit.'!!'. A red(>scrii)tion of Antpltisbdcrid caeca with a discussion of its relationship to A. bakcri. .lour. Dcjit. Aurii-. rucrto l^ico. vol. 1(3, no. '1, i)p. 153-155. I!)(i4 AMPIIISBAENA SCHMIDTI 11 'S a a B e rO 00 •f!^ ^ ss, s K . '^ H-t: «M ?^ J o B > >— ( "^ X -''^£ S ^ O) ft KD i^ O «(H »— ( .-H -^ t— I -<-' C3 ^■^ 72 X CI i-H 1-^ CI CI + + + + + + CH CO to ro to CI — CI CI •^ K P^ M Cl CH- Ci CI I— ( Cl rH T-H Cl I— 1 'O •-C "' •-0 o "O -^ -t -t< Ttl ^ '*< t^ t^ t^ t^ ■^' ■c re :ic c^c '^ oc -H Cl Cl Cl Cl Cl CO 'O <» 'X) rH l-H ^ •+ ^ -H [^ l^ l^ + + + + + ^'^ t't rc CO CO + + + + + Cl O Ci Cl o O O 'l^ o o eg Cl ^^ Cl Cl ^ «o Cl o Cl Ci o IT 'Tl LO T— ' r— 1 Cl t- T— ' l~ ro r\ ^ C 1 i^ l-H -H rv- ,*%- — H c^c N S 22 0^ ^-4 BREVIORA Mmseiiiinii of Coeipsirative Zoology Cambridge, Mass. March 10, 19G4 Number 199 A NEW SUBSPECIES OF VARANL\S EXANTIIEiMATICU^ (SAURIA, VARANIDAE) Bv R. F. Laurent Dr. Williams asked me recently to study a small series of voimq: Savanna Monitors collected bv C. J. P. lonides in south- eastern Tanganyika, where many novelties have already been discovered. He had been impressed by their striking color pat- tern which is quite ditit'erent from the pattern exhil)it('d ])y other juveniles of the same species. In this, he was quite right. However, the morphological fea- tures selected by Mertens (1942) as characteristic of the recog- nized species and races of the genus Varanus did not suggest that the six critical juvenile specimens belonged to an undescribed form. In squamation tbey agreed perfectly "well with Voranus {Empayusia) crantluinaticus microstictKs Boettger. Neverthe- less, I believe that it is a serious mistake to ignore coloration when this character has the same degree of constancy as is generally considered significant for morphological characters. The difficulty here .seems to be that the diagnostic juvenile pat- terns gradually fade during the life of the monitors so that the adults are far more similar, uniform and dull, than are the juveniles. As a result, the number of specimens available for comparison is greatly reduced, the adults not being distinguish- able. When I compared the lonides specimens with other juveniles of Varanus e.ranthcDiaticus, I came to the conclusion that the sul)species recognized by Mertens are themselves geographically varialjle as far as their juvenile color pattern is concerned. The six juveniles from southeastern Tanganyika are alike, but obviously different from two specimens from Kenya ; like- wise, among the so-called southern alhigularis, two juveniles from Zulidand have a striped jiattern quite different from the 2 BREVIORA No. 19!) ocellated pattern of five other juveniles from Transvaal, Becliu- analand and Southwest Africa. The photographs of specimens here discussed replace advan- tageously- any clumsy and lengthy description. Feeling that six specimens disclosing the same color pattern are enough to infer that the population to which they belong has on the whole peculiarities patently diiferent from those of other populations, I regard them as a new subspecies of Vorantis e.rantJicmaticus: Kilwa Fig. 1. Map sliowing the known localities for Varan us exanihcmatictis ionidcsi n. subsp. Varanus exanthematicus ionidesi subsp. n. Holotype: A half-grown individual (MCZ 544:9-1), Kilwa, Tan- ganyika, 1956, coll. lonides. Paratypcs: 1 adult, 2 subadults and 1 juvenile (MCZ 52536- 39), Liwale, Tanganyika, 1953; 1 juv. (MCZ 54495), Masasi, Tanganyika, 1956; 1 juv. (MCZ 55470), Liwale, Tanganyika, 1957; 1 juv. (MCZ 56176), Kilwa, Tanganyika, 1958; 1 juv. (MCZ 56981), Liwale, Tanganyika, 1958. All collected by C. J. P. lonides. Diagnosis. A subspecies of Varanus exanthematicus similar to microstictus Boettger ( ty])e locality, Ethiopia) but differing from it by a very distinctive juvc^iile color pattern som(>what sinular to that of albiyularis. VJiJ-it -\i:\v srusi'KciKs ok VxVKANus 3 The u'cucral cdloi" is li^litcf jiiid with less coiitrjist of li,L;'li1 and (lark colors than in two yonn*:- niicrosi id us from Kenya. Two l)la<-k lines start i'roni the jjostei-ior border of the eye and a])i)roaeh each other on tiie neck, stoi)])in8 in ionidcsi. The throat is blackish, as is usual, accoi'ding to ^lertens. for the other races.^ MORPHOLOGICAL CHARACTERS As stated previously, the morphological characters are not different from microstidus. Table 1 gives scale counts as com- pared with Mertens' descriptions of microstidus, alhigularis and (Difjolensis plus some specimens of microstidus (Kenya) and of alhigularis (southwest Africa and Transvaal) and angohnsis (Lobito). Certain measurements have been studied on scatter diagrams. Xo significant differences have been observed between ionidcsi and microstictus, but it seems that aJbigularis and (ixgoJcnsis have the head broader and the snout longer than liotii micro- 1 This is not the case for tlic two mii-rostii-tiis juveniles from Kenya. BREVIORA No. 199 Subspecies ionidesi microst ictus (Mertens) mierostictus (Kenya) albigidaris (Mertens) albipiihiri.s (Zululand ) albigularis (Transvaal, SW Africa) aiuiolensis (Mertens) (iiifioJensis (Loliito, Angola ) TABLE I Scales Ventral scales Scales lietween around (from gular fold the commissures N the body to the groin) of tlie moutli 5 129-150 1-i 122-152 o 131-134 19 137-1(37 O 139-147 5 137-149 3 110-138 120-141 88-99 88-94 87-90 8.1-110 98-99 88-107 77-92 94-98(!) 53-63 51-56 50-52 57-64 54-57 53-60 53-55 55-56 These data on angoleiisis suggest tlint this alleged subspecies might lie clinally merged with albigularis. stictus and ionidesi; the differences are minor and would have to be checked on larp'e series. ]\Iore interesting- perhajis is the fact that the two Zululand specimens, different from all others by their transversely banded pattern, are also ditt'erent in tliat their tail is shorter than the snout-vent length. This strongly suggests that another neglected race exists in, South Africa, but it would be imprudent to describe this on two specimens Avith relatively inexact locality. The existence of this race will have to be confirmed by studies on larger samples of Vfiraims (.raiiflK- '})}afir)(s; from South Africa. Specimens utilized for comparison with Varanus exanthematicus ionidesi VarauHS exanthonaticus micfost ictus Boettger, MCZ 41112, Sokoki Forest, Kenya (coll. H. J. A. Turner) ; MCZ 59154, Ka- koneni on Sabaki River, Kenya (coll. M. J. Coe). Varanus exanihematicus alhigularis (Daudin), MCZ 45417-18, Zululand; MCZ 39906, Waterberg, Southwest Africa (coll. W. Hoesch) ; MCZ 41888, Saltpan, Pretoria Distr., Transvaal (coll. A. Roberts) ; AMNH 57600, Otjuiarongo, Waterberg, Southwest Africa; CNHM 17141, Gaberones, Beehuanaland (coll. H. Lang) ; CNHM 75712, Hiittenhain, Southwest Africa (coll. W. Hoesch). Varanus exanthematicus angolensis Schmidt, AMNH 47721-24, Lobito, Angola. 1964 NEW SUBSPECIES OF VARANUS 5 Acknowledgmc )tf . Tliis ^v()l■k has been performed thanks to a grant o-enerously allowed by Ihe National Science Foundation (BG-18987). I take i)leasure in thanking Mr. C. M. Bogert and Dr. R. F. Ing(M' for tlie loan of sj)eeiinens belonging to the col- lections of tlie American ^luseum of Natural History (AMNH) and the Chicago Natural History Museum (CNHM), respectively. EEFEEENCE Mertens, R. 1942. Die fnmilio der Varane (Varanidae). Dritter Teil : Taxonomie. Abh. Senekenb. Naturf. Ges., 466: 235-391. BREVIORA No. 199 .:r'?tx^-&re^ Fig. 2. Faranus exanthematicus micro.st ictus juveniles from Kenya. 1964 NEW SI'BSIMMIKS Ol' \AKAXUS I Fig. 3. Varanus exanthematicus ionidesi n. subsp. Holotype sunoimded by paratypes. BREVIORA No. 199 ^i^K^ Fig. 4. Varaniis exanthematicus albigularis juveniles, showing the typical pattern. 1964 NEW SUBSPECIES OF VABANUS Fig. 5. Varanus exanthematicus albigularis juveniles from Zululand, show- ing the banded pattern. w BREVIORA MiiseTjinti of Coimparative Zoology Cambridge. Mass. April 10, 1!)()4 Number 200 AN ANGUID LIZARD FliOM THE LEEWARD ISLANDS By Garth Underwood Department of Zoology, University of the West Indies, Trinidad. Through the good offices of Mr. L. Kasasian of the University of the West Indies, I received from Mr. J. Phillips, Director of Agriculture in Montserrat, a lizard which I was amazed to see was an anguid of the genus Diploglossus; the family had not previously been known, living or fossil, from the Lesser Antilles. Scarcely less amazing was the observation that this lizard resem- bles the Central American D. monotropis in respect of several major differentiating characters, and differs conspicuously from the known Greater Antillean species. The specimen was collected in Montserrat by Mr. J. Kingsley Howes who tells me that he has only once before seen this lizard, and that twenty -five years ago. On the grounds that I am therefore unlikely soon to see another I describe for the second time (Underwood, 1959a) a new species of Diploglossus on the basis of a single specimen. Once again my description contains tedious detail, some of which may be eliminated when more specimens are examined. As explained in the earlier paper, I do not recog- nize the genus Celestus as distinct from Diploglossus. Li the condition of the claws, prefontals, nasals, and numerous scale organs on the back, this species resembles only D. ynonofropis. I therefore borrowed two specimens of monotropis and prepared the description with these before me ; the specimens are MCZ 37139 Yavisa, Panama, and MCZ 15353 Limon, Costa Rica. The name of the island of Montserrat is Spanish, from the Latin moiis serratus. As this is quite the most remarkable reptile yet reported from Montserrat, and as no other reptile bears the name of this island, I call it Diploglossus montisserrati. BREVIORA No. 200 Family ANGUIDAE Cope DiPLOGLOSsus MONTissERRATi iiew species Type and only known specimen: adult $ MCZ 76924, and from this specimen : left ramus of lower jaw; slide 1, prefontals, frontal (damaged), frontoj^arietals, interparietal (damaged), parietals and occipital scales ; slide 2, series of 10 scales back from occipital ; slides 3a, b and c, 12 longitudinal rows of 10 dorsal scales ; slide 4, transverse row of scales from middorsal to midflank ; slide 5, three rows of 10 scales forward from groin ; slide 6, transverse row of tail scales at level of ankle from middorsal to midventral ; slide 7, series of 20 scales back from mental ; slide 8, 25 midbelly scales. Collected by J. Kingsley Howes at Woodlands Spring, elevation about 600 feet, Montserrat, West Indies, 16° 45i// N, 62° 13' W. Diagiiosis: A large strong-limbed DipJoglossus with sheathed claws ; three prefrontal scales ; contact between nasal and rostral scales ; numerous scale organs on dorsal scales but none on belly scales ; very numerous scale organs on dorsal tail scales ; more than 80 scales from occipital to base of tail ; back plain brown, no crossbands. Fi^. 1. Dorsal view, head of Diplofjlossus montis.serrati type. Area of posterior frontal aud frontoparietals may be a little inaccurate owing to damage. 1964 AN ANGUIl) FROM MONTSERRAT 3 Description: As this species rather closely resembles mono- tropis, I note in square brackets the features in which monotropis differs. Upon receipt of the tir.st draft of my manuscript Mr. Benjamin Shreve kindly examined an additional seven specimens of monoivopis: in the Museum of Comparative Zoolo. montisserrati in Montserrat? Whatever may have been tlie manner of its arrival, it is now a relict, far removed geographically from relatives. I am inclined to ]i*64 AN ANGUID F1{()IM IMONTSKRIJAT 9 regard it as a survivor of the pluvial period, Avliicli lias persisted in a favourable habit()4 Number 201 FOOD HABITS AND YOUNG STAGES OF NORTH ATLANTIC ALEPISAURUS (PISCES, INIOMI).^ By Richard L. Haedrich Mui^eum of Comparative Zoology, Harvard University INTRODUCTION Other than a few general statements (Clemens and Wilby, 196U157; Gibbs and Wilimovsky, in press), there is very little information concerning food habits of the lancetfishes, genus Alepisaurus. During a recent exi)loratory fishing cruise (63-4), the U.S. Fish and Wildlife R/V DELAWARE obtained stomach contents from 36 Alepisaurus ferox and four .4.. brevirostris, taken by longline at 16 stations from the offing of New England to the Azores. These collections allowed an investigation of the food of Alepisaurus across a large portion of the North Atlantic over a relatively short period of time. In addition to the longline stations, four hauls were made with a modified Scharfe midwater trawl (Scharfe, 1960). Since the hauls were made in approximately the same levels as those in which the longlines fished, examination of this material provided some information concerning the selectivity of the lancetfish in its feeding. The longlines fished from the surface to about 80 fathoms. At most longline stations, sets were made at about 0600 hours and were hauled before 1500. Each midwater haul lasted less than an hour and reached a depth of 10 to 20 fathoms. All hauls were made at night, when vertically migrating animals would be expected to reach the upper limits of their depth ranges. The positions of the longline stations at which Alepisaurus were taken and the mid- water trawl stations of DELAWARE 63-4 are shown in Figure 1. Only five Alepisaurus smaller than 200 mm SL (standard length) were available to Gibbs (1960) for comparative purposes, and these could not be specifically identified. In the DELAWARE 1 Contriliutioii Xo. 14-12 linin i\w Woods Hole Oa'anognipliic Institution. BREVIORA NO. 201 material, AlepisaurKs was found to prey heavily on its own kind, and 42 young lancetfishes ranging from 50 to 585 mm SL were ob- tained from stomachs. Series of both species made specific identifi- cation of such small specimens possible. X ." ie«X Figure 1. Positions of longline stations at wliicli Alciiisauni.^ were taken ( • ) and midwater trawl stations ( X ) of DELAWARE 63-4. ACKNOWLEDGMENTS Peter C. Wilson and personnel of the Bureau of Commercial Fisheries, Gloucester, made possible the participation in DELA- WARE 63-4 by representatives of the Museum of Comparative Zoology, Cambridge, and the Woods Hole Oceanographic Institu- tion. Jonathan L. Treible, representative of the Museum of Com- parative Zoology for the duration of the cruise, and Martin R. Bartlett, Woods Hole Oceanographic Institution, made the collec- tions and provided information concerning depths fished by long- lines and midwater trawls. Malcolm L. Clarke, National Institute of Oceanography, England, identified tentatively the cephalopods. Daniel M. Cohen, LTnited States National Museum, Washington, D.C., commented on the nomenclature of Lepidion. Richard H. Backus, Woods Hole Oceanographic Institution, and Giles W. Mead, Museum of Comparative Zoology, have read and criticized the manuscript. My work was supported by a Woods Hole Oceano- graphic Institution Summer Fellowship. Half the ship's costs for this cruise were paid for by the National Geographic Society. The material is housed in the Museum of Comparative Zoology. 1964 FOOD AND YOUNG OF ALEPISAURUS 3 IDENTIFICATION OF YOUNG STAGES Two series of Alepisaurus were present in the stomach contents. The largest fish in each scries was identifiable by characters pro- posed by Gibbs (1960), notably the shape of the head, the body proportions, and the melanophore structure. The largest A. bre- virostris was 585 mm SL; the largest A. ferox was 267 mm SL. As Gibbs (1960) observed, marked changes in proportions occur with growth, and the morphometric diff'erences which separate large specimens cannot be used to identify small lancetfishes. However, a distinctive melanophore pattern and the relative posi- tion of dorsal and pectoral fins do distinguish the young fishes. Examination of one series, the largest specimens of which were clearly Alepisaurus brevirostris, showed melanophores along the midventral surface of the body at all growth stages seen (38 speci- mens, 49-585 mm SL). These melanophores were absent in the other series (5 specimens, 40-267 mm SL), the largest specimen of which was clearly A. jerox. Since A. ferox greater than 500 mm SL do have a fine peppering of melanophores on the belly, these spots must appear at some stage in growth between about 270 and 500 mm. The origin of the dorsal fin was in advance of the origin of the pectoral fin in all A. brevirostris, whereas the dorsal origin was over or behind the pectoral origin in all A. ferox. Alepisaurus from 260 mm SL down to at least 40 mm SL can be separated as follows: I. Melanophores present on belly, dorsal origin in advance of pectoral origin, snout profile convex AJrpisavrus brevirostris Gibbs II. Melanophores absent from belly, dorsal origin over or be- hind pectoral origin, snout profile generally straight (cf. Fig. 2) Alepisaurus ferox Lowe Anal and pectoral fin-ray counts were taken from the young fishes. Many specimens were damaged, hence dorsal counts were impractical. The meristic data (Table 1 ) show a similar modality to that found by Gibbs (1960). TABLE 1 Fin-ray count frequencies in young Alepisaurus Anal rays Pectoral rays 13 14 15 16 17 18 12 13 14 15 16 A. brevirostris 2 13 16 6 2 18 17 1 A. ferox 13 1 112 1 BREVIORA XO. 201 ij V^ "^ % g '3) o "C f^ o o "3 a V2 -S ^ o "oj -d g -4-^ CO "S c3 o 'Sn > -t^T C CD a o ^^ _Q & OJ ci O S ^ o 13 OJ ^ fl -Q f^ OS ^ K'. X' o •— ' X c O cc ■to ■ OJ Z cS "*^ -O c: g o ^' ^3j -J::; o o bC "3 ^ ,-w ^ o, Ch o o X "t^ 0 £ c5 M 0| X Ij o ^ 0 r , O O o ^ o ' — o ;H — H c b£ :-• - 1 1 _CJ O ^ yd o — c^ '^ /- o -£ o tj ^ c ^ H c -C X bi ^ r^ — ^i' c; i; ::^ O ^ b(. — . ^ O -^ rt ._ o -b c^ o o CJ o ^ tx _^ E ■*" fl r^ c 1964 FOOD AND YOrXO OF ALEPISAURUS 5 Reported here for the first time is ti 40 nuu lancetfish, MCZ 42377, taken l)y Isaacs-Kidd niidwater trawl during a shakedown cruise of R/V ANTON BRUUN; 39°10'N, 71°52'W; 1015-1420 hr.; 9 Jan. 1963. The snout of this small fish is curved in profile, hut there are no melanophores along the belly and the origin of the dorsal is well behind that of the pectoral. There are 16 rays in the anal fin, and 14 rays in the pectoral. I identify this specimen as Alepisaurus ferox, and include it in the above analysis. The 56 mm Alepisaurus figured by Maul (1946: 155, fig. 20) is refera- ble to A. brevirostris on the l)asis of head shape and fin positions. The three specimens of Alepisaurus brevirostris from 23^-'30'W (Table 2, Sta. 9, Coll. 2) constitute an eastward range extension for the species of some 1480 miles. Gibbs and Wilimovsky (in l)ress) report no specimens east of 57° W. FOOD OF ALE PI SAVE US (Table 2) Morning longline sets produced far more lancetfish than night sets. Twenty per cent of the longline stations were made at night, but only one of 45 Alepisaurus taken was caught on a set which fished primarily during the dark hours. This may indicate that Alepisaurus is a day-time feeder, but might also suggest that the bait is not readily seen by the lancetfish at night. In the dark, luminescent animals, such as the hatchetfish Sternoptyx, may be more easily detected. No difference, however, was noted between the stomach contents of the night-caught Alepisaiirus (Sta. 24) and those of fish from morning sets. Most animals from Alepisaurus stomachs were in excellent con- dition. Few had been digested to any extent. Rofen (in jiress) sug- gests that digestion proceeds almost entirely in the intestine of Alepisaurus, and that the stomach serves only for storage. This may be a mechanism allowing the lancetfish to dine at ojiportunity and to digest at leisure. TABLE 2 Alepisaurus stomach contents, DELAWARE 63-4, 27 April to 7 June 1963. Each collection represents one jar, usually the stomach contents from one fish. In some cases the stomach contents from two or more fish were put in one jar; these appear as single collections with the number of stomachs in- dicated. The fork length (FL) of the fish from which the stomach contents were obtained is given when known. Ximiber of specimens and range of standard lengths of fishes in riglilhand column. 6 BREVIORA NO. 201 Station 1. 40°09'N, 49°54'W. 0645-1510 hr. 27 Apr. 1963 Collection 1, from A. ferox: Chiasmodon niger Diretmus argenteus JJrophycis sp. Alepisaurus breviroslris Paralepis coregonoides Carinaria lamarcki Pterotrachea sji. Phrosina sp. polychaete worm (Pisces, Percoidea) (Pisces, Berycomorphi) (Pisces, Anacanthini) (Pisces, Iniomi) (Pisces, Iniomi) (Mollusca, Heteropoda) (Mollusca, Heteropoda) (Crustacea, Amphipoda) 15 1 2, 19 & 38 mm 1, 14 mm 1,21mm 4, 50-75 mm 2, 41 & 45 mm 7 10 Collection 2, from A. ferox: Schedophilus medusophagus (Pisces, Stromateoidea) 1, 120 mm Collection 3, from A. ferox: Carinaria lamarcki Pterotrachea sp. Phronima sp. Phrosina sp. Collection 4, from A. ferox: Paralepis coregonoides Pyrosoma sp. Carinaria lamarcki Pterotrachea sp. Phronima sp. Phrosina sp. Collection 5, from A. ferox: 2 small pieces of scjuid Carinaria lamarcki Pterotrachea sp. Brachyscelus sp. Phrosina sp. Collection 6, from A. ferox: Alepisaurus breviroslris Notolcpis rissoi Paralepis coregonoides Argonauta sp. Carinaria lamarcki Pterotrachea sp. Phrosina sp. polychaete worm (Mollusca, Heteropoda) 3 (Mollusca, Heteropoda) 39 (Crustacea, Amphipoda) 2 (Crustacea, Amphipoda) 1 (Pisces, Iniomi) 2, 48 & 49 mm (Tunicata, Thaliacea) 1 (Mollusca, Heteropoda) 21 (Mollusca, Heteropoda) 65 (Crustacea, Amphipoda) 1 (Crustacea, Amphipoda) 7 (Mollusca, Cephalopoda) (Mollusca, Heteropoda) 2 (Mollusca, Heteropoda) 31 (Crustacea, Amphijioda) 1 (Crustacea, Amphipoda) 16 (Pisces, Iniomi) 1,365 mm (Pisces, Iniomi) 1, 89 mm (Pisces, Iniomi) 1, 62 mm (Mollusca, Cephalopoda) 1 (Mollusca. Heteropoda) 5 (Mollusca, Heteropoda) 6 (Crustacea, Amphipoda) 5 1904 FOOD AND YOUNG OF ALEPISAUHUS Collection 7, fioiu A. jerox: Alepisaurus brevirostris Paralepis sp. Octopodoteiithid? Carin a ria la m a rcki Pterolrachea sp. Cavolinia sp. Brachyscelus sp. Phronima sp. Phrosina sp. nemertean polychaete worms (Pisces, Iniomi) 1, 585 mm (Pisces, Iniomi) 1,50 mm (Mollusca, Cephalopoda) 1 (MoUusca, Heteropoda) 9 (Mollusca, Heteroputla) 6 (Mollusca, Pteropoda) 2 (Crustacea, Amphipoda) 20 (Crustacea, Amphipoda) 2 (Crustacea, Amphipoda) 106 1 X 18 Station 2. 39°20'N, 44°50'W Collection 1, from A. jerox : Brama brama Alepisaurus brevirost ris Paralepis coregonoides Paralepis sp. Stemoptyx diaphana fish larvae doliolid test octopod Pterolrachea sp. brachyuran Palmurus larva Brachyscelus sp. Phronima sp. Phrosina sp. Collection 2, from A. jerox : Lepidion lepidion Alepisaurus brevirost ris Notolepis rissoi Paralepis coregonoides Chiroteuthis sp. Taoniinae Carinaria lamarcki Pterolrachea sp. brachyuran small euphausids Brachyscelus sp. Hyperoche sp. Phronima sp. Phrosina sp. 0650-1430 hr. 29 Aiir. 1963 (Pisces, Percoidea) 2, 34 & 101 mm (Pisces, Iniomi) 1,223 mm (Pisces, Iniomi) 1, 164 mm (Pisces, Iniomi) 36, 23-60 mm (Pisces, Stomiatoidea) 1, 34 mm 13 1 (Tunicata, Thaliacea) (Mollusca, Cephalopoda) 1 (Mollusca, Heteropoda) 2 (Crustacea, Decapoda) 1 (Crustacea, Decapoda) 1 (Crustacea, Amphipoda) 34 (Crustacea, Amphipoda) 2 (Crustacea, Amphipoda) 9 (Pisces, Anacanthini) 2, 25 & 45 mm (Pisces, Iniomi) 3, 60-295 mm (Pisces, Iniomi) 1, 105 mm (Pisces, Iniomi) 5, 30-108 mm (Mollusca, Cephalopoda) 1 head (Mollusca, Cephalopoda) 147 (Mollusca, Heteropoda) 5 (Mollusca, Heteropoda) 1 (Crustacea, Decapoda) 1 (Crustacea, Euphausiacea) 4 (Crustacea, Amphipoda) 1 (Crustacea, Amphipoda) 1 ( Crustacea, Amphipoda) 3 (Crustacea, Amphipoda) 10 8 BREVIORA NO. 201 Station 4. 40°17'N, 36°07'W. 0630-1435 hr. 1 May 1963 Collection 1, from A.jerox : Carinaria laniarcki Pterotrachea sp. Phrosina sp. Eunjdice sp. Collection 2. from A. brci Alepisaunis bre virost ris (loliolid test, containinp; a Phronima sp. Carinaria laniarcki Cavolinia sp. Brachyscelus sp. Phronima sp. Phrosina sp. ])()l>-chaete worms Collection 3, from A. Iircviroslriti: Carinaria laniarcki Pterotrachea sp. Cavolinia sp. Brachyscelus sp. Phronima sp. Phrosina sp. Platyscelus sp. polychaete worm (Mollusca, Heteropo(ia) 7 (Molhisca, Heteropo(la) 1 (Crustacea, Ami)hi])0(la) 2 (Crustacea, Isoiiotla) 3 'irostris : (Pisces, Iniomi) 1,64 mm (Tunicata, Thaliaeea) 1 (Mollusca, Heteropoda) 15 (Mollusca, Pteropoda) 4 (Crustacea, Amphi]ioda) 17 (Crustacea, Amphijioda) 3 (Crustacea, Ami)hipoda) 50 2 (Mollusca, Heteropoda) 17 (Mollusca, Heteropoda) 1 (Mollusca, Pteropoda) 6 (Crustacea, Amjihipoda) 44 (Crustacea, Amphipoda) 2 (Crustacea, Amphipoda) 125 (Crustacea, Amphipoda) 1 1 Collection 4, from .4. hrcvirostris: Anolopterus pharno Alepisaurus brevirost ris Carinaria laniarcki Pterotrachea sp. CdvoUuia sp. Brachyscelus sp. Phrosina sp. polychaete worm (Pisces, Iniomi) 1,277 mm (Pisces, Iniomi) 2, 48 & 56 mm (Mollusca, Heteropoda) ca .60 (Mollusca, Heteropoda) 3 (Mollusca, Ptero])0(la) 7 (Crustacea, Amphipoda) 69 (Crustacea, Amphipoda) 256 1 Station 6. 41 20'X, 2!) 20'\V. 1805-1050 hr. 3-4 May 1963 Collection 1, from A. jerox 1365 mm FL: Paralepis coregonoides Carinaria lamarcki Pterotrachea sp. Phronima sj). Pln'osiiKi sp. iV isces. Iniomi) ( M ollusca , Hct eropoda ) ( Mollusca, Het(>rop()da) ( Crust ac(>a, A ii 1 1 >! i i | loda ) ( Crust acea , A m ph i | loda ) 1, 123 mm 23 12 1 (I 19(U FOOD AND YOrXO OF ALEPISArRUS Station 9. 36°21'N, 23°30"W. 0635-1530 hr. 8 May 1963 Collection 1, from .1. jvrox: Urophycis sp. fish lar\a Carinana lammrld Pterotrachea sp. Brachysceltis sp. Phrosina sp. Collection 2, from A. fcrox: Lophius piscatorius Cubiceps gracilis Schedophilus medusophagits Alepisaurus brevirostris Stemoptyx diaphana fish larvae doliolid tests, one containing a, Phro7iimo unident. squids, 4 spp. Carinaria lamarcki Pterotrachea sp. Phronima sp. Phrosina sp. (Pisces, Anacantliini) (MoUusca, Heteropoda) (Mollusca, Heteropoda) (Crustacea, Amphipoda) (Crustacea, Amphijioda) (Pisces, Pediculati) (Pisces, Stromateoidea) (Pisces, Stromateoidea) (Pisces, Iniomi) (Pisces, Stomiatoidea) 1, 9 mm 1, 14 mm 1 29 1 9 2, 40 & 48 mm 1, 65 mm 1, en. 60 mm 3, 69-178 mm 2, 24 & 27 mm 8 5 (Mollusca, Cephalopoda) 4 (Mollusca, Heteropoda) 9 (Mollusca, Heteropoda) 220 (Crustacea, Amphipoda) 5 (Crustacea, Amphipoda) 5 Station 14. 37°25'X, 29 lO'W. 0610-1335 hr. 16 May 1963 Collection 1, from A. jernx 1585 mm FL: Schedophilus medusophagits Paralepis coregonoides Carinaria lamarcki Pterotrachea sp. (Pisces, Stromateoidea) (Pisces, Iniomi) (Mollusca, Heteropoda) (Mollusca, Heteropoda) 3, 33-40 mm 1,129mm 11 6 Station 15. 36 55'N, 32°32'W. 0400-1110 la-. 17 May 1963 Collection 1, from A. jerux 1437 mm FL : Lophius piscatorius Schedophilus medusoplwgiis Alepisaurus brevirostris Stemoptyx diaphana doliolid test Argonauta sp. Phronima sp. Phrosina sp. (Pisces, Pediculati) (Pisces, Stromateoidea) (Pisces, Iniomi) (Pisces, Stomiatoidea) (Tunicata, Thaliacea) (Molkhsca, Cephalopoda) (Crustacea, Amphipoda) (Crustacea, Amphipoda) 2, 62 & 80 mm 1, 35 mm 1,98 mm 28-30 mm 10 BREVIORA NO. 201 Collection 2, from ^4. jerox 1342 mm FL: Paralepis atlantica Stemoptyx diaphana doliolid test Taoniinae Phronima sp. Phrosina sp. Platyscelus sp. nemerteans polychaete worm (Pisces, Iniomi) (Pisces, Stomiatoidea) (Tiinicata, Thaliacea) (Mollusca, Cephalopoda) (Crustacea, Amphipoda) (Crustacea, Amphipoda) (Crustacea, Amphipoda) 1, 98 mm 1,28 mm 1 1 1 24 2 18 1 Station 16. 33°20'N, 39°50'W. 0410-1110 lir. 19 May 1963 Collection 1, from A. jerox 1265 mm FL: Cubiceps gracilis Alepisaurus brevirostris Macro paralepis affine Stemoptyx diaphana fish larvae doliolid tests salps Argonaata sp. Japetella sp. unident. scjuids Carinaria lamarcki Pterotrachea sp. Cavolinia sp. Limacina sp. Braclii/scchis sp. Phruiiiina sp. PJiromia sp. Platyscelus sp. ]iolychacte worms (Pisces, Stromateoidea) (Pisces, Iniomi) (Pisces, Iniomi) (Pisces, Stomiatoidea) (Tunicata, Thaliacea) (Tunicata, Thaliacea) (Mollusca, Cephalopoda) (Mollusca, Cephalopoda) (Mollusca, Cephalopoda) (Mollusca, Heteropoda) (Mollusca, Heteropoda) (Mollusca, Pteropoda) (Mollusca, Pteropoda) ( Crustacea, Amphipoda ) (Crustacea, Ami)liipuda) (Crustacea, Amphipoda) (Crustacea, Amphipoda) Collection 2, from A. jerox 1292 mm FL: Alepisaurus brevirostris (Pisces, Iniomi) Alepisaurus jerox Carinaria lamarcki Phrosina sp. Lanceola sp. (Pisces, Iniomi) (Mollusca, Heteropoda) (Crustacea, Amphipoda) (Crustacea, Isopoda) Collection 3, from A. jerox 1325 mm FL: Alepisaurus brevirostris (Pisces, Iniomi) Paralepis atlantica (Pisces, Iniomi) Stemoptyx diaphana (Pisces, Stomiatoidea) nemerteans 2, 76 & 83 mm 7, 71-439 mm 2, 64 & 114 mm 21, 8-31 mm 17 29 146 2 7 2 2 remnants 15 3 36 32 14 42 5 2, 113 & 124 mm 1, 267 mm 1 1 1 1,486 mm 1, 84 mm 1, 22 mm 15 1964 FOOD AND YOUNG OF ALEPISAURUS 11 Collection 4, from A.jerox : Alepisaunis brevirostris (Pisces, Iniomi) 1, 437 mm Station 18. 31°00'N, 47°05'W. 0555-1320 hr. 21 May 1963 Collection 1, from A. jerox Diplospinus multistriatus Cubiceps gracilis Alepisaunis brevirust ris Alepisaurus jerox Paralepis atlantica Paralepis elongata Sternoptyx diapliaiia fish larvae Pyrosoma sp. Abraliopsis sp. 1166 mm FL: (Pisces, Scombroidea) (Pisces, Stromateoidea) (Pisces, Iniomi) (Pisces, Iniomi) (Pisces, Iniomi) (Pisces, Iniomi) (Pisces, Stomiatoidea) (Tunicata, Thaliacea) 4, ca. 30 mm 1, 84 mm 1, 154 mm 1, 91 mm 1,96 mm 5, 101-162 mm 9, 18-30 mm 35 1 (MoUiisca, Cephalopoda) 2 Station 21. 32^00'N, 56°10'W. 0605-1320 hr. 24 May 1963 Collection 1, from A. ferox Brama brama Alepisaurus brevirostris Paralepis sp. Sternoptyx diaphana Onychoteuthis sp. Carinaria lamarcki Cavolinia sp. Phrosina sp. 1068 mm FL: (Pisces, Percoidea) (Pisces, Iniomi) (Pisces, Iniomi) (Pisces, Stomiatoidea) (Mollusea, Cephalopoda) (Mollusca, Heteropoda) (Mollusea, Pteropoda) (Crustacea, Amphipoda) Collection 2, from 3 A. ferox Bothus sp. Diplospinus multistriatus Alepisaurus brevirostris Alepisaurus ferox Paralepis coregonoides Lobianchia dofleini Sternoptyx diapliana fish larvae doliolid tests salps Argonauta sp. Japetella sp. Carinaria lamarcki brachyuran Brachyscelus sp. Phronima sp. Phrosina sp. 941-1043 mm FL: (Pisces, Heterosomata) (Pisces, Scombroidea) (Pisces, Iniomi) (Pisces, Iniomi) (Pisces, Iniomi) (Pisces, Iniomi) (Pisces, Stomiatoidea) (Tunicata, Thaliacea) (Tunicata, Thaliacea) (Mollusca, Cephalopoda) (Mollusca, Cephalopoda) (Mollusca, Heteropoda) (Crustacea, Decapoda) (Crustacea, Amphipoda) (Crustacea, Amphipoda) (Crustacea, Amphipoda) 2, 37 & 55 mm 1,463 mm 1, 75 mm 11, 21-32 mm 1 1 1 1 1, 19 mm 7, 20-163 mm 1, 81 mm 2, 82 & 131 mm 7, 68-99 mm 1,31 mm 5. 20-27 mm 7 4 2 4 2 3 1 1 4 3 12 BREVIORA NO. 201 Lanceola sp. nemertean polyehaete worms (Crustacea, Isopoda) 1 1 15 Station 23. 32°05'N, 59 lO'W. 0605-1340 lir. 25 May 1963 Colloctioii 1, from 4 .4. ferox 772-1276 mm FL: Diplospiiius DiullLst rial Its Anoplogaster cornuta Re go Ice us glesne leptocephalus larvae Alepisaurus brevirostris Pamlepis atlantica Paralepis coregonoicles Pamlepis eloiigoln Steriioptyx diaphana bits of Pyrosoma test unident. squids C armaria lamarcki Pterotrachea sp. Brachyscelus sp. Phroiiima sp. Phrosina sp. Platyscelus .sp. polyehaete worms (Pisces, Scombroidea) 14, 27-131 mm (Pisces, Berycomorphi) 1, 11 mm (Pisces, Allotriognathi) 1, 160 mm (Pisces, Apodes) 5, ca. 60 mm (Pisces, Iniomi) 5, 82-383 mm (Pisces, Iniomi) 2, 76 & 100 mm (Pisces, Iniomi) 5, 87-ca. 110 mm (Pi.sces, Iniomi) 2, 135 & 153 mm (Pisces, Stomiatoidea) 12, 15-27 mm (Tunicata, Thaliacea) (Mollusca, Cephalopoda) 2 (Mollusca, Heteropoda) 2 (Mollu.sca, Heteroi)oda) 1 (Crustacea, Amphipoda) 3 (Crustacea, Amphipoda) 1 (Crustacea, Amphipoda) 26 (Crustacea, Ami:)hipotla) 1 15 I I Station 24. 32°45'N, 64°36'W. 1855-0825 hr. 26-27 May 1963 Collection 1. from ,4. jerox 1146 mm FL: Anoplogaster cornuta Paralepis elongata Sternopt yx diaphana Octopodoteuthis sp. Pholidioteuthis sp. polyehaete worm (Pisces, Berycomoriilii) (Pisces, Iniomi) (Pisces, Stomiatoidea) (Mollusca, Cephalopoda) (Mollusca, Cephalopoda) 1. 45 mm 2, 127 & 138 mm 2, 27 & 30 mm 1 1 1 Station 29. 32'50'N, 68'15'\V. 0610-1340 hr. 1 June 1903 Collection 1, from .4. ferox 1028 mm FL: Diplospin us ni utt isl rial u.^ St e rn ojit i/x dia pli a n a doliolid test Phronima sp. Phrosina sp. polyehaete worms (Pisces, Scombroidea) (Pisces, Stomiatoidea) (Tunicata, Thaliacea) (Crustacea, Amphipoda) (Crustacea, Anipliipoda) 1, 116 mm 2, 20 ct 22 mm 1 1 1 3 1 i)iU FOOD AM) YOUNG OF ALKIMSAUHUS 13 Station 31. 34°45'N. 7ri5'\V. Coll(H'tion 1, rioni 5 .1. fcro-i Dij)l<)sj)iitiis Diiillisl ridtiis Brania bmma Anoplogaster coniula Alepisaiinif> brcvirosl ris Macro pamle pis ajjiiie Famlepis allantica Pa ra lepis coregonoidcs Pamlepis elongala Argyropelecus aculealus Slernoplyx diaphana 0610-1515 hr. 2 June 19G3 :• 1039-1254 mm FL: ( Pisces, iScoml)ioi(lca) 1, 131 mm (Pisces, Percoiclea) 2, 40 & 95 mm (Pisces, Berycomorplii) 1, 75 mm (Pisces, Iniomi) 2, 124 & 131 mm (Pisces, Iniomi) 1, 150 mm (Pisces, luiomi) 5, 93-lli) mm (Pisces, Iniomi) 14, 79-120 mm (Pisces, Iniomi) 2, 61 & 125 mm (Pisces, Stomiatoidea) 1, 32 mm (Pisces, Stomiatoidea) 6, 14-29 mm Station 33. 36°27'N, 67°50'W. 0610-1505 hr. 3 June 1963 Collection 1, from A. jcrox 811 mm FL : D iplospin us tn ultist rial us Pa m le pis coregonoidcs Slernoptyx diaphana Pterotrachea sp. Phrosina sp. polychaete worms (Pisces, Scombroidca) (Pisces, Iniomi) (Pisces, Stomiatoidea) (Mollusca, Heteropoda) (Crustacea, Amphipoda) 5, 61-194 mm 3, 67-106 mm 3, 20-29 mm 2 2 2 Station 37. 39°23'N, 68°33'W. 0620-1445 hr. 7 June 1963 Collection 1, from A. brcvirostris 778 mm FL: Tetraodontidae Helicolenus dactyloplerii): Lepidion lepidion Syngnathus pelagicus Macroparalepis afjine Ommastrephidae Taoniinae Brachyscelus sp. Phronima sp. Phrosina sp. (Pisces, Plectognathi) (Pisces, Scleroparei) (Pisces, Anacanthini) (Pisces, Solenichthyes) (Pisces, Iniomi) (Mollusca, Cephalopoda) (Mollusca, Cephalopoda) ( Crustacea, Amphipoda) (Crustacea, Amphipoda) (Crustacea, Amphipoda) 1, 16 mm 2, 24 & 25 mm 2, 42 & 42 mm 1,113 mm 1, 138 mm 2 6 3 1 2 With regard to the size of its prey, Alepisaurus practices no discrimination. Adapted for swallowing large prey (Marshall, 1955:330), it will also readily devour small creatures. A 1265 mm A. ferox (Sta. 16, Coll. 1 ) had eaten seven .4. brevirostris ranging from 71 to 439 mm in standard length; another A. ferox (Sta. 1, Coll. 7) contained a 585 mm A. brevirostris in addition to 106 hyperiid amphipods, each about 20 mm long. 14 BREVIORA NO. 201 Lancetfish stomachs most commonly contained hyperiid amphi- pods, heteropods, paralepidids, Sternoptyx, and other Alepisaurus. At least one of these groups was found at every station, and three or more occurred simultaneously in 66 per cent of the collections. Of the 32 collections, amphipods occurred in 78 per cent, hetero- pods in 72 per cent, paralepidids in 59 per cent, Alepisaurus in 56 per cent, and Sternoptyx in 44 per cent. Since these animals do occur so regularly in the stomachs, they must habitually freciuent the same depths as does Alepisaurus. Stomach contents reflected local abundances. Heteropods oc- curred in 17 out of 19 (89%) of the stomachs from stations 1 through 15, and they were also a major constituent of the mid- water hauls. Heteropods were found in only 6 out of 13 (46%) of the stomachs from stations 16 through 37. In contrast, Sternop- tyx had been eaten by 77 per cent of the fish from stations 16 through 37, but by only 44 per cent of the fish from stations 1 through 15. On the southerly leg of the cruise, fishes predominated both in stomach contents and in the midwater haul. Myctophid fishes were notably absent from most stomach con- tents. Only one was found, a 31 mm Lobianchia dofleini (Sta. 21, Coll. 2). None occurred in the fish from station 9, although a one- hour midwater trawl at station 12 took 101 myctophids of a dozen species. The commonness of these small fishes in midwater hauls makes their absence from Alepisaurus stomachs particularly ])uzzling. Myctophids are noted vertical migrators (Marshall, 1960). If Alepisaurus does not migrate extensively, and if its prin- cipal depth differs from the day and night levels of myctophids, the duration of the lancctfish's encounter with myctophids would be brief during any diurnal cycle, and consumption of them there- fore low. Completely absent from the stomach contents were the charac- teristic middle-depth fishes Gonostoma, Stomias, and Chauliodus. These fishes are distributed mainly between 200 and 1500 meters (Haffner, 1952; Marshall, 1960). Their absence from Alepisaurus stomachs, and the presence of primarily epipelagic fishes such as Cubiceps, Brama, and Paralepis elongata strengthens the belief (Gibbs and Wilimovsky, in press) that the lancetfish frequents the upper layers. Little difference was noted between food of Alepisaurus ferox and A. brevirostris. A. brevirostris may favor invertebrates, par- ticularly the hyperiid amphipod Phrosina. All A. brevirostris stomachs contained these amphipods, but only 75 per cent of those from A. ferox did. There were 256 amphipods in an .4. brevirostris 19G4 FOOD AND YOUNG OF ALEPISAURUS 15 from Station 4 (Coll. 4), and only two in the A. ferox from the same station (Coll. 1). The maximum number of Phrosina from a single A. ferox was 106, and the average number per fish was 9.3. The maximum number of Phrosina from a single A. brevirostris was 256; 108.3 was the average. Fishes were more common in A. ferox stomachs. Sternoptijx and paralepidids occurred in 50 and 64 per cent, respectively, of A. ferox stomachs, but were found in 0 and 25 per cent, respectively, of the A. brevirostris stomachs. Since stomach contents were obtained from only four A. brevirostris, and since three came from one station, these comparisons are inconclusive. Rarely encountered fishes found in the stomach contents were Anotoptei'us pharao (Sta. 4, Coll. 4), Regalecus glesne (Sta. 23, Coll. 1), and four pelagic young of Lophius piscatorius (Sta. 9, Coll. 2; Sta. 15, Coll. 1). LITERATURE CITED Clemens, W. A. and G. V. Wilby 1961. Fishes of the Pacific Coast of Canada. (Second edition). Fish. Res. Bd. Canada. Bull. No. 68, 443 pp. GiBBS, Robert H., Jr. 1960. Alepisaurus brevirostris, a new species of lancetfish from the Western North Atlantic. Breviora, No. 123, 14 pp. GiBBS, Robert H., Jr. and Norman J. Wilimovsky In press. Family Alepisauridae, in Fishes of the Western North Atlantic. Sears Found. Mar. Res., Mem. 1, part 5. Haffner, Rudolph E. 1952. Zoogeography of the bathjqjelagic fish, Chauliodtis. Syst. ZooL, 1: 113-144. Marshall, N. B. 1955. Alepisauroid fishes. Discovery Reports, 27 : 303-336. 1960. Swimbladder structure of deep-sea fishes in relation to their sys- tematica and biology. Discovery Reports, 31 : 1-122. Maul, G. E. 1946. Monografia dos peixes do Museu Municipal do Funchal. Ordem Iniomi. Bol. Mus. Municipal Funchal, 2(2) : 154-156. Rofen, Robert R. In press. Family Omosudidae, in Fishes of the Western North Atlantic. Sears Found. Mar. Res., Mem. 1, part 5. Scharfe, J. 1960. A new method for one-boat trawling in midwater and on the bottom. General Fisheries Council for the Mediterranean, Studies and Reviews, ser. B, 21 pp. u^ v\^\^ BREVIORA Museeim of Coainparative Zoology Cambridge, Mass. April 10, 1964 Number 202 THE BLIND SNAKES (TYPHLOPS) OF HAITI AVITII DESCRIPTIONS OF THREE NEW SPECIES^ By Neil D. Richmond Carnegie Museum, Pittsburgh, Pa. When the Herpetology of Hispaiiiola (Cochran, 1941) was published onlj^ 24 specimens of Typhlops were available from the entire island. Recent collecting (1950-1962) in Haiti has produced 93 specimens of blind snakes from that country alone, most of them now in the Museum of Comparative Zoology. Part of the field work was made possible by the following grants to the Museum of Comparative Zoology : National Science Foundation Grants NSF 16066, NSF 5634, and one from the American Philosophical Society. The six specimens from Gonave Island were obtained by Philip S. Humphrey collecting for the Yale Peabody Museum and the University of Florida. A descrip- tion of the various expeditions is given in Williams ct al., 1963. 1 wish to express my appreciation to Ernest E. Williams who was responsible for getting this material together and making it possible for me to study it. Also I wish to thank Doris M. Cochran for permission to study the type of Typldops sulcatus. Miss A. G. C. Grandison supplied counts and descriptive notes on four specimens in the British Museum. Nicholas Strekalov- sky made the drawings of the new species herein described. The following abbreviations are used to designate the collec- tions where these specimens are deposited : AMNH, American Museum of Natural History; CM, Carnegie Museum; MCZ, Museum of Comparative Zoology; UF, University of Florida Museum; USNM, United States National Museum; YPM, Yale Peabody Museum. 1 Notes on Hispanlolan herpetology no. 10. 2 BREVIORA No. 202 Examination of this material discloses that in addition to the two species previously known from Haiti, lumhricalis and piisil- lus, there are three forms new to science. All of the five species represented in this collection liave the following characters in common : relatively narrow strap-like rostrals ; completely di- vided nasals; preocular contacting the third upper labial only; scale rows either 20-20-20 or 20-20-18 ; at least one pair of enlarged parietal scales ; four upper labials ; three lower labials ; and clearly visible eyes. "With this great similarity they also show distinctive differences in the number of middorsal scale rows, number of preoculars, shape and size of ocular, shape of head, and color. In number of middorsal scales these snakes may be divided into two distinct groups : those with a relatively low number (240-330), two species, and those with over 385, three species. The three new forms are all in the group with the high number of middorsal scales. Two of these are further distinguished by having very small pointed heads. Typhlops capitulatus^ sp. nov. Figure 1 Holotype: MCZ 62636. From Manneville, Haiti, at the north- west end of Lake Saumatre. Collected by A. S. Rand and J. Lazell, 10 August 1960. Paratype: MCZ 69006. From the same locality as the type. Collected by E. E. Williams and A. S. Rand, 13-14 August 1959. Fig. 1. Typhlops capiluhiius sp. iiov. Type MCZ 62636. From Manne- ville, Haiti. Dorsal and lateral views. iFroni tlip T^atin "having a small lipad." 1964 Tvi'jiLoi's OF HAITI 3 Diagnosis: A slender small-headed species of Typhlops char- acterized by the following combination of characters : high num- ber of dorsal scales, 385-400; scale rows 20-20-20; dorsal color extending on to venter, a white anal spot ; preocnlar almost triangular ; head small and pointed ; ocular almost as wide as high. Description: Sides of head tapering from about the level of the seventh middorsal scale. Rostral narrow, its width one-third that of the head, not extending to a line connecting the eyes ; nasals completely divided by a suture that extends from the .second upper labial to the rostral ; preocular almost triangular, approximately as high as wide and in contact with the third upper labial, its anterior edge curved but not elongated. Ocular broad, but little higher than wide, its posterior edge strongly convex (Fig. 1). Tavo pairs of enlarged parietal scales; two postoculars on each side. Eye small, scale rows around body 20-20-20, dorsal scales from rostral to caudal spine about 400. Color: Light brown above, somewhat paler below although pig- mented, venter Avith scattered white scales, anal region and underside of tail white, dorsal color fades gradually on head to entirely unpigmented snout. Size: Total length 205 mm, head width at level of eyes 2.7 mm, diameter at midbody 4 mm, tail slightly longer than wide. Variatio7i: MCZ 69006 is a very small specimen, total length 93 mm, and is pale brown above and below although the charac- teristic pattern of pigmentation can be seen with magnification. This specimen has about 385 dorsal scales from rostral to spine. In detail of head scutellation it agrees with the type. Remarks: This species is most closely related to the species on Gonave Island, described below, from which it may be dis- tinguished by the shape of the preocular. Typhlops gonavensis sp. no v. Figure 2 Holotype: YPM 3003. From Point a Raquette, on the south shore of Gonave Island, Haiti. Collected by Philip S. Humph- rey and Sarita Van Vleck, 9 April 1959. Paratypes: YPM 3004, and UF A943. With the same data as the holotype. Diagnosis: A slender small-headed species of Typhlops char- acterized by the following combination of characters : high num- ber of dorsal scales, 409-423 ; scale rows 20-20-20 ; dorsal color 4 BREVIORA No. 202 extending on to venter, a white anal spot ; may be distinguished from capitulatus which it most closely resembles by the dis- tinctive shape of the preocnlar which is wider than high and narrowest anteriorly. Description: Sides of head tapering from about the level of the seventh middorsal scale. Rostral narrow, its width one-third that of the head, not extending to the level of the eyes ; nasals completely divided by a suture that extends from the second upper labial to the rostral ; preocular wider than high, narrowest anteriorly (Fig. 2), in contact with the third upper labial; ocu- lar small but little higher than wide, its posterior edge strongly convex ; two pairs of enlarged parietals ; two postoculars on each side ; eye small ; scale rows around body, 20-20-20 ; middorsal scales from rostral to spine about 418. Color in preservative: Dark purplish brown above, somewhat paler below but heavily pigmented with an irregular white area in anal region. Dorsal color fades gradually on head to unpig- mented snout. Color in life (from notes by Sarita Van Vleck) : Reddish brown inferiorly, grading to burnt umber on tail, venter purplish gray anteriorly becoming pale gray-brown posteriorly. Size : Total length 189 mm, tail 4.2 mm ; head width at level of eyes 2.7 mm ; diameter at midbody 4 mm ; tail slightly longer than wide. Remarks: This species is most closely related to capitulatus on the mainland of Hispaniola and apparently has differentiated from that species following isolation on Gonave Island. It is easily recognized by the peculiar shape of the preocular. These two species have the smallest and the most pointed heads of any of the known Antillean Typhlops. Fig. 2. Typhlops gonavensis sp. iiov. Tyjje YPM 3003. From Point a Raquette, Gonave Island. Dorsal and lateral views. 1964 TYPllLOPS OP HAITI 0 Typhlops haitiensis sp. IIOV. Figure 3 Holotype: MCZ 62635. From Manneville, Haiti. Collected by A. S. Band and J. Lazell, 10 August 1960. ParatijiJes (14) : MCZ 69007-12, from Manneville, Haiti. Col- lected by E. E. Williams and A. S. Rand, 13-14 August 1959. USNM il7273-74, 117276, from Trou Caiman, Haiti. Collected by A. Curtiss, 18 February 1943. CM 38804-8, from Manneville, Haiti. Collected by George Whiteman, 1963. Manneville is near the northwest end of Lake Saumatre, on the north edge of the Cul de Sac Plain; Trou Caiman is a small freshwater lake just west of Manneville. Fig. 3. Typlilnpn haitiensifi sp. nov. Type MCZ 62635. From Manne- ville, Haiti. Dorsal and lateral views. Diagnosis: A species of Typhlops characterized by a high number of dorsal scales, 400-452; scale rows 20-20-20; dorsal color extending on to venter, white anal spot; preocular higher than long, in contact with the third upper labial ; ocular narrow, more than twice as high as wide ; head broadly rounded as seen from above. Description: Snout, as viewed from above, truncate and broadly rounded, not tapering; rostral width about one-fourth the width of the head, rostral extending to the level of the eyes ; nostrils lateral ; nasals completely separated by a suture extend- ing from the second upper labial to the rostral. The preocular higher than wide in contact with the third upper labial only (Fig. 3). One pair of enlarged .strap-like parietals each as wide as two body scales ; a single postocular ; eye large and conspicu- ous. Scale rows around body 20-20-20, dorsal scales from rostral to tail spine about 435. 6 BREVIORA No. 202 Color: The head scales have dark brown centers and relatively wide light margins giving the top of the head a striped appear- ance ; dorsal brown color extends on to the venter where it is interrupted by irregular light areas along the midventral line. The dorsal scales are dark brown with light margins. The light margins are aligned giving the effect of fine light longitudinal lines. Size: Total length 240 mm, tail 5 mm, width of head at level of eyes 3.5 mm, diameter at midbody 5 mm, tail slightly longer than wide. Variation: The 14 paratypes agree in details of head scutella- tion, and in having 20-20-20 scale rows. The number of dorsal scales in this series varies from 400 to 452. The number of enlarged parietal scales also varies. All have one pair of greatly enlarged strap-like scales ; some have a second pair equally enlarged, while others show a graded change from this condi- tion to where the second pair is not distinguishable from the body scales. All have a single postocular. The color varies from dark to light brown in the large specimens. The three smallest specimens are very pale above and below, but under magnifica- tion it can be seen that the dorsal pigment extends on to the venter in the same pattern observed in the larger specimens. The extent of the light areas on the venter varies considerably ; the three specimens from Trou Caiman have little or no brown pigment on the venter, while MCZ 69012 has the venter almost entirely pigmented. Typhlops pusillus Barbour The 27 specimens examined from the mainland of Hispaniola agree with those reported by Cochran (1941) except that in this series all have two postoculars between the fourth upper labial and the parietal. The greatest variation is shown by the three specimens from Gonave Island. These agree with the mainland forms in pos- sessing divided preoculars, two postoculars and in other details of head seutellation. They differ from the mainland specimens in having a lower number of dorsal scales (258-268 compared with 285-319 for the specimens from Bombardopolis on the northwest peninsula). They also differ in color as they are brown above and pale below and in this respect resemble lum- hricalis. l!)64 TYPIlLOrS OF HAITI 7 Typhlops lumbricalis (Linnaeus) This species has the most extensive range of any of the Antil- lean Typhlops. It has been found on Hispaniola, Cuba, several islands in the Bahamas, and in British Guiana, South America. Throughout this range, it is remarkably consistent in characters : scale rows 20-20-18, rarely 20-18-18 ; usually one pair of enlarged parietals and usually two postoculars. The middorsal scales range from 240-320 (Legler, 1959). The established lack of variation in the number of scale rows in various parts of its extensive range makes it very surprising to find a population of this species on the southwest peninsula of Haiti that is highly variable. There are 41 specimens of lumhricalis at hand from the south- west peninsula, representing several localities from Miragoane to Jeremie, and one specimen from Grand Cayemite Island. In number of rows of scales around the body they vary as follows: 22-20-20 (2), 20-20-20 (27), 20-20-18 (12). 'The num- ber of enlarged parietals also varies: IL-IR (14), 1L-2R (11), 2L-1II (5), 2L-2R (10). All but one have two postoculars on each side. The exception has the upper postocular on each side fused to the parietal. The number of middorsal scales ranges from 273-324, average 801. Although this is within the known range of variation for lumlyricalis, it is higher than for most specimens of that species from other parts of Hispaniola. The shape and arrangement of the head scales is the same as in other populations. The color pattern is also the same. Although the number of rows of scales is both high and variable, this population can not be distinguished from lumhricalis by any character or combination of characters, and it appears to be one variable population rather than a mixture of two or more forms. In addition to the specimens from the southwest peninsula, there are five specimens from Port-au-Prince and Manneville. Tliese have the scales in 20-20-18 rows, middorsal scales ranging from 261 to 290, and are typical liunhricalis. Grant (1956) discusses four specimens of lumhricalis, three from Eaux Gaillees and one from Port-au-Prince. The number of middorsal scales ranges from 257 to 275 and indicates that these are lumhricalis as currently defined. There are also four specimens of lumhricalis in the British Museum from Haiti, two from Pont Beudet and two from Hinche. Miss A. G. Grandi- son supplied the counts for these and the middorsal scales 8 BREVIORA No. 202 range from 260-294 — well within the range of variation for lumhricalis. The average number of dorsal scales for all 13 specimens from Haiti outside the southwest peninsula is 279 with a range of 257-294, ^&^ DISCUSSION The five species now known from Haiti are similar in having a maximum of 20 scale rows, preocular in contact with the third upper labial only, and a relatively narrow rostral. Yet with this apparent similarity the five forms represent three distinct phyletic lines of Typhlops. 1. The group including lumhricalis and pusilhis is dis- tinguished by having a low number of middorsal scales, 240-330. These two forms on Hispaniola are readily separated from each other by the presence of a single preocular in lunihricalis and by a divided preocular in pusillus. 2. A slender, small-headed group with a high number of middorsal scales, 385-423 ; this includes two species, gon- avensis on Gonave Island, and capitulatus on the main- land. These are the most distinctive Typlilops in the area as no other species known from, the Antilles has such a small head. The two forms are obviously closely related and are probably insular populations of what was originally one form. The two may be distinguished from each other by the shape of the preocular. 3. The third group is represented by one species, haitiensis. Like the small-headed forms it has a very high number of middorsal scales (400-454) but unlike the species of that group it has a broadly rounded truncate snout, as well as a distinctively shaped preocular and ocular. Of the five species known from Haiti it is the only one that typically has a single postocular. The others have two postoculars with one occurring as an uncommon variant. These three different groups do not seem to be closely re- lated to each other. The distributional data, although limited, seem to support the concept that only unrelated forms can occur sympatrically, and even then probably l)y occupying different ecological niches. For example, Iwnibricalis and pusil- lus appear very similar in body proportions and scutellation, and differ in the number of preoculars. A divided preocular is also known as a rare variant in lumhricalis (Legler, 1959). 1!)G4 TYPIU.OI'S OF HAITI 9 Probably pusiUus was derived from lumhricalis at sometime when Hispaiiiolfi was a series of islands. Todaj^ pusillus occurs on the northwest peninsula and ranges across the northern part of the island at least as far as the southern shore of Bahia de Samana. Cochran (1941) reported one specimen from Sanchez and I have examined four from Samana (AMNH 50353-56). The range of lumhricalis includes all of the southwest penin- sula and extends eastward along the southern coast where it is known from Barahona (Noble and Hassler, 1933)' and San Pedro de Macoris. On the northern coast it is known as far west- ward as Puerto Plata. The range of the two species overlap on the northeast coast. Whether they are ecologically separated in this area of overlap can not be determined from the available data. At Manneville, the one locality where three species have been found, each species represents a different group. In view of the large number of Typhlops in collections from the northwest peninsula and all of them pusillus, it would appear that lumhricalis does not occur there. The opposite is true of the southwest peninsula where extensive collecting has revealed only lumhricalis. Typhlops sulcatus Cope The finding on Haiti of species with a high number of dorsal scales raised the question of w^hat relationship these forms might have with sulcatus on nearby Navassa Island, the only other Typhlops known from the Antilles wdth both 20 scale rows and a high (397) number of dorsal scales. As no other specimens of sulcatus have been collected since Cope (1868) described it on the basis of one specimen, the type was examined. As the type appears today, it is a specimen that was pre- served just prior to shedding ; in fact, the old head shields have been shed (probably after preservation), and the new head shields are represented by soft raised areas outlined by deep sulci that mark the position of the original sutures. The exten- sions of the nasal sutures from the nostril to rostral are little more than faint lines but sufficient to indicate that this species does have completely divided nasals. Since Cope (1868) also iThe specimen with 385 micUlorsal scales from Alta Vela Island reported as Uunbricalis by Noble and Hassler (iy;j3) is not that species as currently defined. A preliminary examination discloses this specimen as more closely related to sulcatus than it is to the mainland forms. 10 BREVIORA No. 202 described this specimen as having completely divided nasals he may have seen it with the nasal plates in place Avhen the sntures might have been more distinct. Other characters of this specimen are : very long parietals extending down the posterior edge of the oculars to well below the level of the eyes, and a single postocular; both of these characters are well shown by Cochran (1941, p. 310, fig. 88). As stated by Cope the body is more slender than that of lumhricaUs ; it is also distinguished from that species by the high number of middorsal scales. From the species on Haiti with middorsal scales over 385, it is distin- guished by the shape and size of the ocular and preocular, and by the very long parietals. This specimen has the snout tapered in front of the eyes very much like monensis but not at all like the narrow-headed forms on Haiti which have the entire head tapered from well back of the eyes. Until such time as more specimens are available it seems best to recognize sulcatus as a distinct species. The other characters for this species mentioned in the litera- ture, trilobate snout and incomplete nasal sutures, are conditions that are associated with shedding (Richmond, 1961). The ex- tension of the nasal suture from nostril to rostral is very faint on most specimens of small Typhlops and difficult to see at best. As the nasal plate thickens, prior to shedding,, the nasal suture from labial to nostril becomes more conspicuous while the suture from nostril to rostral becomes fainter and may actually disap- pear as a suture although its presence can usually be detected as a fine dark line crossing an otherwise smooth area of the nasal plate. Also associated with shedding are the deep sulci repre- senting the sutures between the other head scales. Since sulcatus is known from only one specimen, and that one shedding, it is little wonder that it has been variously described as having complete nasal sutures (Cope, 1868), incomplete nasal sutures (Cochran, 1924, 1941), and complete on one side and incomplete on the other (Legler, 1959). TYPHLOPS Key to the forms occurring in Haiti A. Preocular divided pusillus AA. Preocular single B B. Ocular narrow and liigli, over twic-e as liigli as wide, posterior edge of ocular almost straight, preocular higher than wide (400- 460 dorsal scales) (Fig. 3) haitiensis i li)64 TYPIILOl'S OF HAITI 11 BB. Ocular wide, but little higher than wide, posterior edge strongly convex, preocular as wide or wider than high C C. Venter unpigmented, same color as underside of tail, less than 330 dorsal scales, head rounded lumhricalis CC. Dorsal dark color extending on to venter, underside of tail and anal area white, 380-430 dorsal scales, head tapering . D 1). Preocular longer than high, anterior angle acute (Fig. 2) gonavensis DD. Preocular almost triangular, anterior angle rounded (Fig. 1 ) capitulatus EEFERENCES CITED Cochran, D. M. 1924. Typldops lumhricalis and related forms. Jour. Washington Acad. Sci., 14(8): 174-177. 1941. The herpetology of Hispaniola. Bull. U. S. Nat. Mus., 117: VII + 398 pp. Cope, E. D. 1868. Additional descriptions of neotropical Eeptilia and Batrachia not previously known. Proc. Acad. Nat. Sci. Philadelphia, 1868: 128. Grant, C. 1956. Eeport on a collection of Hispaniolan reptiles. Herpetologica, 12: 85-90. Legler, J. M. 1959. A new blind snake (genus Typhlops) from Cuba. Herpetologica, 15: 105-112. Noble, G. K. and W. G. Hassler 1933. Two new species of frogs, five new species and a new race of lizards from the Dominican Eepublic. Amer. Mus. Novit., no. 652: 1-17. (Eeceived Sept. 26, 1963) ElCHMOND, N. D. 1961. The status of Typhlops silus Legler. Copeia, 1961 (2) : 221-222. Williams, E. E., B. Shreve, P. S. Humphrey 1963. The herpetology of the Port-au-Prince region and Gonave Island, Haiti. Parts I-II. Bull. Mus. Comp. Zool., 129: 291-342. (Eeceived Sept. 26, 1963.) 12 BREVIORA No. 202 0-, b j ^Y u> .i5 ." ?- UJ 3 CO o, — >- UJ o 5 W O U) 1 s £ S - o fO 2 '•/ 1 i^ 0 0 0^3 UJ _J O CTJZ ^ Q. o- o O • < 4 ■ o- \ •< 2 "^ • i-t 1—1 ^_^ CS 2. rt C o 2 ^ IS ft be c« •S S ^ w 7J o en o snsiip<'tc(l on tlio basis of sizp and ^'oncral Htrufturc that the spf'cinicii is an upper clu'cl^todtli (as yet iinkiiown) of (Jncniisid. However, tlie walls of eacii rei'ntrant folil in tli(! upper elieckteetli of Qiirininia would un- doubtedly be tif;lill.v appressecl as they are in the lower cheekleelh. Thoui;!! the folds in .M("Z 707."i are narrow, tlieir walls are not appressed ( l''i>;. 1). "^ (ipOCtOV' narrow. 1964 NEW RODKXT FROM IIISrANIOLA S MM Figure 1. Occlusal views of (A) MCZ 7675, left DP^ of Plagiodoniia araeum, the holotype, aiul (B) MCZ 35314, right DP^ (reversed) of P. hylaeum. Fine stippling represents cenientuni ; conrse stippling, dentine. Drawn bv 8ue Hirschfeld under XSF GB 178. Description : Anterior margin of tootli strongly convex as in DP^ of Plagiodontia, not flattened as in M^- IVP; incipient anter- olabial fold absent, as in all Plagiodontia, excepting P. aedium (Johnson, 1948) ; reentrant folds relatively long and narrow, parallel- sided ; lingual reentrant with slight posterad flexure at internal extremity, not observed in other Plagiodontia; labial reentrant with concomitant flexure near its mouth ; posterolabial concavity in enamel wall shallow in comparison to other species ; longitudinal axis of tooth curved (convex anterolingnally) as in upper cheekteeth of other Plagiodontia. Measurements : All measurements of DP^ are given in mm in the following table. Anterolingual-posterolabial diameter of occlusal surface in MCZ 7675 is estimated owing to damage along posterolabial wall of crown. BREVIORA No. 203 P. araeitm P. ipnaeum P. hylaeum MCZ 7675 USNM 254376, MCZ 35314, right DP^ right DP4 Maximum diameter of occlusal surface (A) Anterolingual-posterolabial diameter of occlusal surface (B) Eatio of B/A Maximum diameter of crown perpendicular to longitudinal axis (C) Minimum diameter of crown perpendicular to longitudinal axis (D) Eatio of D/C Height of crown perpendicular to occlusal plane 11.0 5.2 .47 10.7 5.0 .47 15.0 8.2 6.7 .82 7.6 5.6 .74 19.5 6.8 5.0 .74 6.8 4.1 .60 12.4 EEFEEENCES Hooijor, D. A. and C. E. Eay [in press.] A metapodial of Acratocnus (Edentata: Megalonychidae) froni a cave in Hispaniola. Proc. Biol. Soe. Washington, 77, 1 fig. Johnson, D. H. 1948. A rediscovered Haitian rodent, Plagiodontia acdium, with a synopsis of related species. Proc. Biol. Soc. Washington, 61: 09-76. Patterson, B. 1962. An extinct solenodontid insectivore from Hispaniola. Breviora, Mus. Comp. Zool., no. 165: 1-11, 4 figs. BREVIORA Mmseiim of Comparative Zoology Cambridge, Mass. May 15, 1964 Number 204 THE STATUS OF PSEUDOGEKKO SHEBAE AND OBSERVATIONS ON THE GECKOS OF THE SOLOMON ISLANDS By Walter C. Brown^ INTRODUCTION Brown and Tanner (1949) referred a unique specimen (Brig- ham Young University [BYU], No. 7002) of a previously un- described geckonid lizard from Guadalcanal in the Solomon Islands to the new species shebae in the genus Pseudogekko Taylor (1922), thus establishing the second species known for the genus and extending the range to include a second peri- pheral group of islands, analogous in position to the Philippines. At that time we had not had the opportunity of examining any material of the type species of the genus, the Philippine species Pseudogekko compressicorpus. New material has now provided the opportunity to reassess the relationships of P. shehae and to redefine its differences from the other small geckos of the Solomon Islands. Dr. Ernest Williams, Museum of Comparative Zoology, recently called my attention to the difficulty of identifying certain specimens, which, on the basis of descriptions in the literature, were apparently referable either to Lepidodactylus guppyi or Pseudogekko shehae. The series of specimens in question (Museum of Comparative Zoology, Nos. 64152, 65862, 67122, 67124, 74517-19, and Stanford Uni- versity, No. 23720) were collected by Mr. Fred Parker on Bougainville Island, Solomon Islands, during 1961-62. These have provided the point of departure for the present paper. This study is part of the author's investigations on the herpetofaunas of the Islands of the Pacific area, supported by 1 Division of Systematic Biology, Stanford University and Menlo College, Menlo Park, California. 2 BREVIORA No. 204 a grant from the National Science Foundation. Illustrations were prepared by Mr. Walter Zawojski, Stanford Research Institute. THE RELATIONSHIPS OF THE GENUS PSEVDOGEKKO Taylor (1922, p. 103), in erecting the genus Pseudogekko, suggested that it might have its closest affinities with Thecadac- tylus (I assume T. australis = Pseudothecadactylus austraUs: Brongersma, 1936, p. 136). There is reason, however, to believe that the closeness of this relationship is doubtful. Recently, five specimens of Pseudogekko compressicorpus became available : Stanford University Nos. 23548-49, from Zamboanga, Mindanao Island, and 23654-55 from Bohol Island, and Museum of Com- parative Zoology No. 44130, the latter collected by Taylor at Saub, Mindanao, and never previously reported. An examination of the foot structure of these specimens indicates probable close affinities with three other Oriental-Pacific genera (Lepidodac- tylus, Gekko and Luperosaurus) , which also belong to the sub- family Gekkoninae as defined by Underwood (1954). The six Oriental-Pacific genera, Gekko, Hcmiphyllodactylus, Lepidodactylus, Luperosaurus, Pseudogekko and Pseudothecadac- tylus^, all belong to that group with moderately to strongly dilated digits, with the distal joint relatively short, compressed, and arising from the tip or near the tip of the dilated part. If digital structure alone is considered, these six genera fall into four sections. Hemiphyllodactylus is rather sharply dis- tinguished by the greatly reduced first digit and the fact that the distal compressed phalanx is not attached all the way to the tip of the dilated portion, as pointed out by Stejneger (1899) and Smith (1933). Pseudothecadactylus forms a second section distinguished by the double series of lamellae which are widely separated distally. As they are presently understood, Gekko and Luperosaurus fall into a third section which may be distinguished from Lepidodactylus and Pseudogekko on the basis of the lamellae being entire throughout the length of the digit. In general the species of the genus Gekko are larger than are the species in the other genera, and the species of Luperosaurus exhibit more extensive webbing. However, a critical study of 1 Pseudothecadactylus confined to the islands of Torres Straits and Australia is not properly Oriental-Pacific but is here considered because of Taylor's belief in its relationship to Pseudogekko. 1964 SOLOMON ISLANDS GECKOS Figure 1. PseudogelcTco compressicorpus. all known species in these last four genera and consideration of other characters than digital structure will be necessary before the generic limits and relationships can be clearly understood. This is borne out by the difficulties which have at times arisen in assigning certain species to one or the other of these genera. Thus Boulenger (1885a, p. 162) included in Lepidodactylus three species later placed in Hemiphyllodactylus by Stejneger (1899, pp. 788, 799) and Smith (1933, p. 15). Again, Boulenger (1885b, p. 473), in describing Gekko pumilus, noted that the species was very like a Lepidodactylus in many characters, being placed in Gekko on the basis of the undivided lamellae. Exam- ination of one specimen of pumilus (MCZ 69216) suggests that BREVIORA No. 204 it may be more closely related to Liiperosaurus with reduced webbing and skin folds. Taylor (1915, p. 96), on first describing compressicorpus, placed it in the genus Luperosaurus but later (1922) separated this species from Luperosaurus and erected Figure 2. PseudogeTcTco shebae. the genus Pseudogekko for it. Again, an examination of the paratype of Luperosaurus rnacgrcgori Stejneger (Stanford Uni- versity No. 6263, a hatchling measuring 23.5 mm from snout to vent) reveals that the sub-terminal lamellae are divided in the midline, and hence that this specimen should be referred to Lepidodactylus species as that genus is presently understood. (The assignment of macgregori to the genus Luperosaurus is thus placed in doubt and the type .should be re-examined in this light.) 1964 SOLOMON ISLANDS GECKOS DISTINGUISHING CHARACTERS OF PSEUDOGEKKO SHEBAE AND LEPIDODACTYLUS GUPPYI Pseudogekko would appear to be distinguished from Lepi- dodactylus primarily on the basis of the more slender habitus and the more narrowly but uniformly dilated digits. Both P. shehae and P. compressicorpus are more slender in body and exhibit less broadly dilated digits (Fig. 1) than most of the several species of Lepidodactylns which I have had the oppor- tunity of examining. If the ratio ' ' breadth of head : snout-vent length" is used as a measure of habitus, the range for Pseudo- gekko shehae and Pseudogekko compressicorpus, based on the few adult specimen,s available, is from about 14-16 per cent ; for five species of Lepidodactylns {christiana, lugiibris, aureo- lineatus, planicaudus and giippyi) the range is 18-21 per cent. Figure 3. Lepidodactylns guppyi. 6 BREVIORA No. 204 (Lepidodactylus hrevipes from the Philippines is an exception with respect to both of these characters and is closer to Pseudo- gekko. A careful study of this species will probably show that it should be placed with P. shehae and compressicorpus.) The differences in foot structure of Lepidodactylus giippyi and Pseudogekko compressicorpus are illustrated in Figures 1 and 3. The condition of the terminal lamella — divided or entire — is not a generic character, since it is divided in Pseudogekko compressicorpus and Lepidodactylus lugubris and entire in Pseudogekko shehae and Lepidodactylus guppyi. The number of preanal and femoral pores in males and the size of the head scales, as illustrated by the number of scales between the eyes in the mid-orbital plane, will help to dis- tinguish Pseudogekko shehae from Lepidodactylus guppyi (Table 1). In addition to L. guppyi and P. shehae, members of the lugu- hris-woodfordi species complex occur also in the Solomon Islands. Individuals of this species or group of species, however, are readily distinguished from Lepidodactylus guppyi and Pseudo- gekko shehae on the basis of the divided terminal lamella on all toes but the first. This scale is undivided on all toes on specimens of L. guppyi and P. shehae. SUMMARY The Oriental-Pacific geckonid lizards of the genera Gekko, Lepidodactylus, Luperosaurus and Pseudogekko represent cat- egories which probably include closely related groups of species but, as they are presently understood, are not sharply and clearly delimited from each other. Two additional genera, Hemi- phyllodactylus and Pseudothecadactylus, although they probably represent lines of evolution distinct for a longer period of time, have by some authors been regarded as very closely related to these four genera. Superficial resemblances between species and the lack of sharp lines of demarcation between the genera not infrequently have made difficult the proper generic assignment of some of the species and even the determination of the species to which isolated individuals belong. A case in point is the identification of specimens of Pseudogekko shehae and Lepidodactylus guppyi, both known from the Solomon Islands. Their distinguishing characteristics and present generic assignment have been briefly discussed in the present paper. 1964 SOLOMON ISLANDS GECKOS 7 LITERATUEE CITED BOULENGER, GEORGE A. 1885a. Catalogue of the lizards in the British Museum. London (Taylor and Francis), vol. 1, xii+436 pp., 32 pis. 1885b. Descriptions of three new species of geckos. Ann. Mag. Nat. Hist., (5) 16: 473-475. Brongersma, Leo D. 1934. Contributions to Indo-Australian herpetology. Zool. Meded., 17: 161-251, 2 pis. 1936. Herpetological note XIII. Zool. Meded., 19: 136. Brown, Walter C. and Vaso M. Tanner 1949. Rediscovery of the genus FseudogeTcTco with description of a new species from the Solomon Islands. Great Basin Naturalist, 9: 41-45. Smith, Malcolm A. 1933. Remarks on some Old World geckoes. Rec. Ind. Mus., 35: 9-19. 1935. The fauna of British India. Reptilia and Amphibia, vol. 2, XIII+440 pp., 1 pi. Stejnegek, Leonard 1899. The land reptiles of the Hawaiian Islands. Proc. U.S. Nat. Mus., 21: 783-813. Taylor, Edward H. 1915. New species of Philippine lizards. Philippine Journ. Sci., 10: 89-109. 1922. The lizards of the Philippine Islands. Bur. Sci. Manila, Publ. 17: 1-269, 23 pis. Underwood, Garth 1954. On the classification and evolution of geckos. Proc. Zool. See. London, 124: 469-492. 8 BREVIORA No, 204 o ^._^ ;-< ^ -^ +-• s > c: g e -4-2 S-l • rH o o '^"^ o .rj tn a -fj r2 ai '3 s o dT 4.4-51, N=6 M ^ 5 s § ft fe -—1 ^ CO ft ^ --^ Ph CO iH i-H C] r-l i-H ^H ^ji^ oj oocococococo go fl g ^ '=*;§'^ft£ "-I i-lrH fCiH O] CD «^ rt -- ^ ft -^ s g d ^S^ CO cocoTt— ' ,,^ O ^ >, 2 o '^ ^S-^^,-; --I T-((Mco- M -S" o r-t m CM M cr> eg r-t o a> eg CO eg X ■:)< r-t o +J 1—) CM CM rH CM r-t t—* CM r-t r-t CM eg r-t eg r-t r-i eg eg eg r-t eg r-t eg r—t CM CM CM bC + + + + + + + + + + + + + + + + + + + + + + + + + + + c in 00 t^ in t^ GO in CO CO (Jl t^ o t^ CM t- eg r^ CO CT) t~ r-t o t^ o X Tt< r—t 0) o CD 00 t^ ■^ CM CO tr r^ CO eg in r-t CD i-H in ■^ I* CO o in 'J" •* r-t t^ in CD J eg CM CM r^ CM CM CM N CM r-t CM OJ N eg CM eg eg eg eg eg eg r^ eg eg eg eg CM U3 •^ rH •!-> n CM IM CO <—i rH O Cvj Cvl eg in r—t eg r-t r—t o o o o r-t o O "a" r^ 1 eg 01 c r-i r— t r^ rH 7-^ T—t ,—t r-t rH r—t rH r-t r—t r^ r-t r-t r-t r—t r-t r^ r-t r-t r-t 1 1 r-l 05 o 0) 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 o 1 1 1 rt e «5 CD (D CD 00 X 00 00 CD CD CD CD CD CD CO 00 X CD t~ CD cc in CD r^ in CD t^ 0 bC 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I 1 1 1 1 1-1 0) CM (:m bH N CM CO CM CM (N o eg f^ eg bn &H eg eg 6h o eg U-, bH eg eg CO i^ CM O m N CM t^ 1 eg ^-^ eg eg CM eg eg CM /--^ u c c t^ t^ t^ 00 t^ 1 00 r^ 1 t^ 00 CD t^ 00 r- t^ r~- t^ f~ t^ X t^ X CD t^ CD X 0) •H 0) ^— ' \— ^ > — y-N \.^ ^— ' 1 ^-^ '— ' "— ^ ^-^ *— ^ N^ ^-^ ^^ •^-^ v-^ ^-^ v-^ ^-^ ^-^ t-{ Jl E 1 1 1 1 1 r-t 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 iH o M ■* CO CO CO CO *— ^ CO CO 1 CO ■* CO 'r •^ ro CO CO CO CO CO CO -^ • 1 1 1 1 1 1 1 0 w + to 00 00 00 00 ■^ CO CO CD CD CD 00 t^ r~ CO M CD « t^ X CO X CD in CD X S r^ rH rH rH r-t r-l r-l r-t r^ r-t r—t r-t r-t r-t r-t r-t r-t r-t r-t r-l r-t r-t r-l rH r-l r-l r-t 10 O m + + + + + + + + + + + + + + + + + + + + + + + + + + + c w & •* lO Tf "S" ^ ■ 1-1 a •H CO h- 1 in r-t in r-l h J + in CO in in 1< Tf CO rf ■^ t> •V ■^ CD in in CO in ■* ■^ rr + Tf ■^ CD + in in 0 1-1 rH t-^ ,-t r-i i-t r^ rH r-t r-t r^ r-t r-t r-t r-t r-t r-t r-t r-t r-t rr r-t r-l rH TJ< r-t r-t ■H 3 ■P + + + + + + + + + + + + + + + + + + + + \ + + + \ + + rt M CO CO CO n to CO CO CO CO CO n ■^ CO TT CO ■* ro CO CO ro to CO ■^ CO n CO ei Z J + + + + + + + + + + + + + + + + + + + + + + + + + + + ■P + CO ■^ CD w CO iO CM CO 00 r-t Tl< -rf CO CO r-t (S ■* t^ CD CD (J> CO > X r^ to in OJ ;s >> CM CM CM r-i r^ r^ CM r^ r-t CM r-t r-t eq eg eg r—t eg eg r-t eg r-t r-t eg eg CO r-t eg Q <; ■a 0 CD CM CM CM CM CM CM CM CM CM CM CM eg CM eg eg CM CM eg eg CM eg CM eg eg eg eg eg •o o 0» ^• "O 1 1 1 1 O 1 1 1 1 1 1 1 1 o •0 1 1 1 1 1 o 1 c u CO o ■^ o Qi CM in O t^ 1 •* CO •H x> • in in in in O) r^ ■H B 1—1 to (D «3 CO CM CM a^ CD CO CD o o in 00 CO CD •* eg to ro ■^ r-l n t- X t^ O 3 «3 o O o CO ^ CO CO CO CD CD CM ■* CM CO in t^ in CD ro X CD • 'i' CD r-l •'>ha|)C'd cont'ix'tiun of den^e irontstonc encloses the pyritized skeleton and promises formidable difficulties to the pre- parator. Fractures in the frontal plane at various levels, however, reveal aistopod characteristics: an elongate, apparently limbless body armored with pebble-like dorsal osteodcrms and needle-like gastralia; vertebrae with hourglass-shaped centra and long, low neural s})ines; and sharp, straight-shafted ribs. The bulbous, well- ossified calvarium recalls that of Phlegethontia. Although the proper classification of this genus must await preparation and de- tailed anatomical study, we may say that canny Thom Stock hit close to the mark in labeling it "Ophiderpeto7i." From the remarkable specialization already achieved by early Mississippian time we must infer that the history of the Aistopoda carries well back into the Devonian. Here the total absence of limbs and girdles in all known aistopods provokes speculation. Did the group originate from a tetrapod ancestor, or directly from some crossopterygian fish through the suppression of the fins? We have no proof either way. But the many features which aistopods share with the limbed leposj-jondyls, most obviously their basic skull structure and the division of the body into neck, trunk and tail, make it extremely unlikely that the forefather of the group was anything but a tetrapod. Affinities of the Aistopoda, as mentioned earlier, have been much debated. In surveying the literature we will do well to restrict our attention to those commentators who have had first-hand knowl- edge of the specimens. J. T. Gregory in 1948 assigned Phlegethontia to the "Dolicho- somidae" and classed this family with the Ophiderpetontidae as the suborder Aistopoda of the order Microsauria. Influenced by the reptilian features of Phlegethontia and by the belief that Cephalerpeton is a microsaur (which it is not) as well as a caji- torhinoid reptile (which it is ) , he placed the orders Microsauria and Captorhinomorpha together in the infraclass Captorhina of the subclass Eureptilia. In 1950, however, he conceded that the order Aistopoda should probal)ly be assigned to the amphibian subclass Lepospondyli. Gregory's 1948 paper convincingly demon- strated that the similarities of the Aistopoda to the Lysorophidae, Apoda and Amphisbacnidae are the result of convergence in ani- mals of similar ecotype, and that these groups are only remotely related. The Turnbulls in 1955 pointed out that Lydekker's family name Dolichosomatidae is objectively synonymous with Cope's earlier name Phlegethontiidae ; they assigned Dolichosoma to the latter 10 BREVIORA No. 206 family. They also noted without comment that both Lydekker and Schwarz had included the Alolgophidae in the Aistopoda." Stress- ing that the postcranial skeleton has more fundamental signifi- cance than the skull, they applied the same classification— class Amphibia, subclass Lepospondyli, order Aistopoda — used in this paper. Gregory, Peabody and Price (1956) carefully compared the vertebral structure of the various lepospondyl groups and con- cluded that the "Aistopoda are a highly specialized limbless group of great antiquity [which] may ]irovisionally be regarded as an early offshoot of the nectridian stock." Basing his observations in part on my preparations and the un- described Alississippian aistopod, Williams (1959) emphasized the antiquity of the tyj^ical lepospondylous centrum which he homol- ogized (quite rightly, in my opinion) with the pleurocentrum of other ancient tetrapods and the centrum of the living amphibians. Rejecting the theory that leposj^ondyls had an independent de- rivation from the crossopterygian fishes, he observed that "if then the lepospondyls are to be derived from the ichthyostegalians, rhachitomes, or anthracosaurs, the pleurocentral centrum speaks for anthracosaur affinities." Without questioning the morphological theory on which Wil- liams' hypothesis of anthracosaur affinities is based, I would like to point out some uncomfortable facts. The oldest true reptile which is adequately known, Cephalerpeton, is of mid-Pennsyl- vanian age and has composite vertebrae in which the relative sizes of pleurocentrum and hypocentrum ( = intercentrum) are the same as in Permo-Carboniferous mammal-like rejitiles and the living Sphenodon. In the same swamj) witli Cephalerpeton lived Di- plovertebron [Eusauropleura] , a form in which the hypocentrum is much larger in proportion to the ])leurocentrum; this genus may reasonably be taken as a morphological ancestor of the Seymouri- amorj^ha from which the true reptiles are believed to be derived. Diplovertebron, however, is so similar to the embolomerous labyrinthodonts in its skull and appendicular skeleton that it must be classed as an amphibian — an anthracosaur. Amphibians with similar vertebral proportions coexisted with true embolom- eres at least as far back as late ^Nlississippian time, so the com- mon ancestry of the diplovertebrontids and embolomeres can hardly be dated later than the middle Mississippian. That the ancestors of the reptiles had a relatively large hypocentrum, and that reduction of the hypocentrum characterized the transition from amphibians to rei^tiles, can hardly be doubted. Derivation of 1964 a'istopod amphibians 11 this pi'olo-ii'ptiliaii verti-hral structure from tlu' ichthyostegid pattern of late Devonian time is (juite undocumented but presents no great theoretical difficulties; the ichthyostegalian vertebra in turn is readily dei-ivahle from that of crossopterygian fishes such as Eusthcnopteron. In striking contrast to this plausible sequence of morphological stages stands the oldest known leposi)ondyl, Stock's early Missis- sippian ai'stopod. Its vertebrae are holospondylous without any trace of an ossified hypocentrum or of laterally paired pleuro- centra like those of the ichthyostegalians and crossopterygians. In vertel)ral form this eo-lepos))ondyl apjicars to be much further removed from the anthracosaur-reptile lineage than that lineage is from the rhachitome-stereospondyl line. Indeed, the evidence now available suggests rather that the dominance of the ])leuro- centrum evolved in convergent fashion in the lepospondyls and anthracosaurs. Thus in my opinion the evidence for leposjwndyl- anthracosaur affinities is too tenuous, and the i)oint of supposed common origin too remote in time, to justify speculation. Relationships of the Aistopoda to other lepospondylous amphib- ians are almost equally difficult to determine. The urodeles and anurans (Lissamphibia) need not be considered here, for their ]X'rsistent four-leggedness and their relatively recent pedigree testify that any connection between Lissamphibia and Aistopoda must be indirect, by way of one of the Paleozoic lepospondyl groups. The apodans (Gymnophiona), though similarly devoid of limbs and girdles, show features of the scales and vertebrae which ally them rather with the microsaurs. Among Paleozoic Icjiospondyls the Nectridea (let us use the correct spelling if only for the sake of novelty) display similarities in vertebral structure which have led Gregory, Peabody and Price (1956) to consider them as distantly related to the aistopods. Tills liypothesis is the most plausible one yet advanced. But it should be emphasized that all known nectrideans had inter- vertebral spinal nerves and well-developed (though sometimes diminutive) appendicular skeletons, including the sculptured clavicles and interclavicle which characterize Paleozoic amphib- ians in general. Aistopods had intravertebral spinal nerves and lacked limbs and girdles, apparently since early iSIississippian time. Thus if these two groups had a common ancestor it must be sought in Devonian rocks — rocks in which amphibians of any kind are extremely rare. Until pertinent fossils can be found and de- scribed, therefore, the affinities of the Aistopoda must remain in doubt. 12 BREVIORA No. 206 The foregoing discussion will have served its purpose if it stimu- lates further research by clarifying the nature of this peculiar group of amphibians and pointing up some of the many unan- swered questions concerning them. Perhaps it is too much to hope that anything can remedy the common but unfortunate tendency to omit the diaeresis and pronounce the name as "ACE-topod" or ''ICE-topod." Like the proverbial politician, the ai'stopods should be happy to be called by any name so long as they are recognized in print. OPHIDEEPETOX FROU CANNELTON, PENNSYLVANIA Seven associated vertebrae (PU 17293) whicli recently came to light in the Princeton paleobotanical collections provide a new horizon and locality record for the genus Ophiderpeton. This speci- men, collected by I. F. ]\[ansfield in 1877, comes from the richly fossiliferous roofing shale of the L'pper Kittanning Coal, a member of the Allegheny Group and late Westphalian D (^Middle Pennsyl- vanian ) in age. The sotu'ce locality is ]\Iansfiekrs mine near Can- nelton in Darlington Township, Beaver County, Pennsylvania. This is the first lepospondylous amphibian to be recorded from Cannelton. ]\Iuch of the bone has flaked away and little remains but the neural arches and spines. Absence of the diagnostic ribs and gas- tralia makes generic identification difficult, as the neural arches of Ophiderpeton and Phlegethontia appear very similar when crushed. Dimensions of the vertebrae fall within the range of the pene-contemporaneous 0. amphiumvnum, but they are also com- patible with the exceptionally large Phlegethontia specimen named P. serpens by Cope. The most diagnostic feature available seems to be the medial notch in the posterior border of the neural arch: this is deep in the Cannelton aistopod and Ophiderpeton amphiuminum but shallow in PJdegethontia. As Ophiderpeton is common in the Linton fauna while only one Phlegethontia speci- men of e(iuivalcnt size is known, statistical probal)ilitics also favor assignment of the Cannelton ai'stoj^od to Ophiderpeton. Although the difference in age between the Linton and Cannelton deposits makes a specific difference possible, in default of morphological evidence the specimen may best be designated Ophiderpeton cf. nmphivmivinn fCope). 1964 AisToi'on ami'hibians 13 SUMMARY The order A'lstopoda is a natural gi'oiip of snake-like lepo- spondylous amphibians which inhabited the Paleozoic coal- swamps of Europe and North America. Two families, each es- sentially monotyi)ic, are distinguished: the Ophiderpctontidac of the early middle to late Pennsylvanian and the more highly spe- cialized Phlegethontiidac of the early Pennsylvanian to early Permian. The oldest known leposi)ondyl, an undescribed aistopod from the early Mississii)pian of Scotland, is of uncertain family position. Pending detailed comparison of the skulls, ordinal characteris- tics may be found in the postcranial skeleton : ( 1 ) body long and snake-like, reaching a vertebral count of about 230; (2) holo- spondylous, amphicoelous vertebrae with low, blade-like neural spines, transverse jirocesses formed by parapojihyses, hypapophy- seal flanges on the caudals, and foramina for the spinal nerves; (3) relatively straight, unicipital, tetraradiate ribs; (4) sickle- shaped hyoids; (5) limbs and girdles entirely lacking; (6) ventral armor of needle-shaped gastralia arranged en chevron, and (ex- cept in phlegethontiids) dorsal armor of pebble-shaped osteo- derms. The remarkable specialization already achieved by the early Mississippian implies an origin well back in Devonian time; a tetrapod ancestry rather than direct derivation from the crossop- terygian fishes is indicated. Relationships of the order are obscure. A proposed connection with the anthracosaur-reptile lineage is controverted by the available evidence, and among lepospondyl groups only the Nectridea show even distant affinities with the Aistopoda. NOTES 'The man who actually collefted the superb skeletons of Linton amphib- ians which the U. S. National Museum acciuired from R. D. Lacoe, and whicli Moodie (1909B) attempted to describe, has not previously been identified. Circumstantial evidence points to Thomas Stock of Edinburgh whose great collection of Scottish Carboniferous fishes and amphibians was purchased by the Museum of Comparative Zoology in 1883. In the MCZ archives is a letter from Stock to Alexander Agassiz, sent from Wellsville, Ohio (the town just north of Linton) on April 30, 1888, which states in part: "I am staying about here . . . collecting the Linton fishes and labyrinthodonts. ... I have been very .successful with the rare group of Carboniferous laby- rinthodonts and have already several nice ones that I should think would 14 BREVIORA No. 206 be a valuable addition to your collection." Unfortunately for Harvard, Agassiz declined to buy Stock's collection. Evidently it was sold to Lacoe, for se\-eral of the Lacoe specimens from Linton bear stickers inscribed "Stock #— ." " Current stratigraphic practice, followed in this paper, places the Mazon Creek fauna close to the Westphalian C-D boundary or in earliest West- phalian D and assigns the Cannelton fauna to late Westphalian D; the Linton fauna is very late Westplialian D. However, recent paleobotanical work by Cridland, Morris and Baxter (1963, p. 63) implies that the Francis Creek Shale (Mazon Creek fauna) is only slightly older than the Upper Freejiort Coal (Linton fauna) but considerably younger than the Kittanning Coal Group (Cannelton fauna). ' Names of the Jarrow amphibians are usually cited as of Huxley, 1867. However, Ophiderpeton and four otiier genera were first validly published in January of 1866 by Etheridge, who credited their authorship to Huxley. Generic descrijjtions were pro\'ided although no species were named. As a genus coelebs (i.e. without named species) is valid under the International Rules of Zoological Nomenclature, the genus Ophiderpeton properly dates from January, 1866; its type sjiecies was first described in April of that year by Huxley in a paper by Wright and Huxley. Duiichusoma emersotti was not described until 1867. ^Preparation of the type specimen (AMNH 6900) of the type species Thyrsidium jasciculare Cope confirms Romer's (1930D) identification of it as Ophiderpeton amphiumirinm. I mention this fact in order to resolve Mile. Dechaseaux's doubts (1955, p. 278) about tlie identity of Thyrsidium. ' Steenimurus was proposed by Oskar Kuhn (1938A, p. 51) in a footnote to Ophiderpeton gmnulosuni Fritsch : "OpJiiderpeton Fritsch non Huxley (teste Steen) = Steeidsaiirus nov. nom." Steen's publications do not suggest any such generic distinction, nor do the specimens document it. "The specimen figured by Scliwarz (1908B, figs. 1-4) under the name Ophiderpeton vicinum is apparently from Nyrany. Because of the difference in age I prefer to assign the late Westphalian material from Nyrany to O. grduiilosiim and restrict the name O. vicinum to late Stejthanian material from Kouno\'a. ' A si)ecimen (L^SXM 4313) which was assigned to PIdcgethontin mazonen- sis by Gregory (1950A, p. 867) proves on preparation to be Cocytinus, the first record of a lysorophid microsaur in the Mazon Creek fauna. '* Dy Fritsch (1883A, p. 125). So far as one can 1964 Al.sToi'UU AMrillBlANS 15 judge from Geinitz and Deichmiilh^r's lithograph (1SS2A, \A. 9) I he centra are hourghiss-shaped only externally: they lack the funnel-like amphicocly which characterizes the Aistoi)oda. Until preparation and rt^study of lliis .specimen reveal its affinities it remains inccrlae sedk. The iiuprinf of a small skeleton from Ihc late Stephanian or early Auluiiian of La Machine, Xicvre, France, has been tentatively identified as an aistopod by Thevenin (1910A, p. 39). Its preservation is so ])oor that little more than the gross niorjjhology can be made out. The skeleton looks more lysoroi)hid than aistopod to me, Ijut this suggestion must be tempered by the fact that lysorophids are known at present only from North America. Until further material is a\ailable tlie affinities of Thevenin's animal must be left snh j II dice. "The element in Phlegethontia longissimn ("Dolichosoma scKtifennn") which Fritsch (1901A, p. 88) called a grill-bar and Steen (1938, p. 226) called a rib is readily identifiable with the hyoid in P. linearis and P. nidzotiensis. Structures identified by Fritsch as gill-rays and shown forming external gill-tufts in his well-known reconstruction are actually gastralia, as Steen has pointed out. "Preparation of type specimens from Linton makes it abundantly clear that the Pennsylvanian genera Molgophis, Coeytiniis and Lysorophus and the Permian species commonly assigned to Lysorophus form a closely-knit family of microsaurian affinities. Although the family name Molgophidae Cojie 1875 has priority, I prefer to retain, at least for the present, the more familiar name Lysorojihidae Williston 1908. Whether Lysorophus Cope 1877 is genericallj' distinct from Molgophis Cojie 1868 remains to be determined. REFERENCES References which are cited by date and letter (e.g. Cope, 1868J) will be found in the standard bibliographies of Hay, Camp and Romer wliich are listed below. B.AIRD, D. 1955. Latex micro-molding and latex-plaster molding mixture. Science, vol. 122, no. 3161, p. 202. 1963. Development of the Westphalian (early Pennsylvania) coal- swami) chronofauna. Geol. Soc. America Special Paper 73, pp. 107-108. C.\MP, C. L., et al. 1940-1961. Bibliography of fossil vertebrates, 1928-1933. Geol. Soc. America Special Paper 27, 1940; 1934-1938, Special Paper 42, 1942; 1939-1943, Geol. Soc. America Mem. 37, 1949; 1944-1948, Mem. 57, 1953; 1949-1953, Mem. 84, 1961. Cridl.and, a. a., J. E. Morris, and R. W. B.axter 1963. The Pennsylvanian plants of Kansas and their stratigraphic sig- nificance. Palaeontographica, vol. 112, Abt. B, Lief. 1-3, pp. 58-92, pis. 17-24, 4 text-figs. 16 BREVIORA No. 206 Dechaseaux, C. 1955. Lepospondyli. In PIVETEAU, J., Ed., Traite de Paleontologie, vol. 5, pp. 275-305, 32 text-figs. Paris: Masson et Cie. Gregory, J. T., F. E. Peabody, and L. I. Price 1956. Revision of the Gymnarthridae, American Permian microsaurs. Peabody Mus. Nat. Hist., Yale Univ., Bull. 10, ix + 77 pp., 1 pi., 33 text-figs. Hay, O. p. 1902. Bibliography and catalogue of the fo.ssil Vertebrata of North America. U. S. Geol. Survey Bull. 179, 868 pp. 1929-1930. Second bibliographj' and catalogue of the fossil Vertebrata of North America. Carnegie Inst. Washington Publ. 390; vol. 1, viii + 916 pp. ; vol. 2, xiv + 1074 pp. M.\TTHEW, W. D. and C. De Pavla Couto 1959. The Cuban edentates. Amer. Mus. Nat. Hist. Bull.; vol. 117, art. 1, pp. 1-56, 42 pis., 5 text-figs. McGiNNis, H. J. 1964 MS. The skeletal moriihology of Phlcgcthontia, a Permocarbonif- erous aistopod. Unpublished MA thesis, Univ. California, 154 pj)., 34 text-figs. RoMER, A. S., et al. 1962. Bibliography of fossil vertebrates exclusive of North America, 1509-1927. Geol. Soc. America Mem. 87, 2 vols., Ixxxix -^ 1544 pp. TuRNBULL, W. D. and P. F. Turnbull 1955. A recently discovered Phlegethontia from Illinois. Fieldiana: Zoology, vol. 37, pp. 523-539, pis. 6-9, text-fig. 114. Williams, E. E. 1959. Gadow's arcualia and the development of tetrapod vertebrae. Quart. Rev. Biol., vol. 34, no. 1, pp. 1-32, 22 text-figs. 1964 AISTOPOD AMPHIBIANS 17 (% HT _5 >< a) 3 ^ J> Sh ^ s o -a C3 ;:r Cl, ^ bi M ° J-< &■ ^ s > ^ cc C 5r ft o -f -^ s BREVIORA Miuiseiuioii of Compsirsitive Zoology Cambridge, Mass. July 28, 1964 Number 207 NOTES ON THE HORSESHOE BATS HIPPOSIDEROS CAFFER, RUBER AND BEATUS By Barbara Lawrence INTRODUCTION Attempts to identify a small series of Hipposideros, collected in the Belgian Congo by Alvin Novick in 1956, have made necessary a re-examination of the distribution and specific char- acters of caffer, ruber and beatus. This has brought to light some new characters of the nasal swellings and of the nose leaves, and emphasizes the importance of the nasal region in general for showing specific differences. Hipposideros ruber centralis Andersen Study of a series of six males and five females from Kivu Forest, Ituri Province, has led to a re-examination of the caffer- ruber group in the Museum of Comparative Zoology collections in the light of Verschuren's and Aellen's recent discussions of these small hipposiderids. The group is admittedly a confusing one. It is the more so because as competent a worker as Andersen (1906, pp. 281-282) interpreted his findings as evidence of a complicated distribution pattern which brought together in a single area different subspecies of the species caffer. Subsequently Hollister (1918, pp. 85-87) pointed out that two clearly distinct species, ruber and caffer, occur together in East Africa. Our mate- rial amply confirms this; we even have specimens of ruber and caffer taken in the same basement. This situation was apparently not clear to Aellen (1952, p. 76) when he followed Hollister in synonymizing centralis with ruber, although he failed to give ruber full specific status, considering it instead a subspecies of caffer. In addition he extended the range of what he called caffer ruber clear across the Congo to the Cameroons, whence 2 BREVIORA No. 207 he recognized three other races of caffer. Such a distribution pattern evidently did not satisfy him and subsequently (Perret and Aellen, 1956:435; Aellen, 1957:198) he stated that the whole caffer complex was in need of revision, though he did not further discuss the relation of ruber to centralis. Verschuren (1957, pp. 346-374), although he deals with a limited area, seems to have understood the situation better and it is significant that much of his work was based on field studies of the live bats. Aside from H. beatus which is easily distinct, he recognizes two species: caffer centralis, a forest form which readily adapts to life in houses, and nanus which is more char- acteristically a savannah form. According to him, the latter were far less numerous than the fonner and the two were only found together in one cave in heavily wooded savannah country. Verschuren's distinction between nanus and centralis is good and on a subspecific level our material agrees well with his. On a specific level, the evidence at hand for considering centralis a race of caffer rather than of ruber is not convincing. Length of tooth row, width across the molars at M^, and the relation of zygomatic to mastoid width, while useful as key characters, are not good indicators of specific relationships. Careful examina- tion of series of the caffer-ruber and beatus groups from a num- ber of localities in a belt across central Africa suggests very strongly that modifications of the nasal structures both internal and external are the best index to specific relationships. This is perhaps not surprising. Since hipposiderids make their sounds through their noses, small variations in the relative size of the diff"erent parts of the nasal swellings and the septa which in- completely divide them should be important. Before final decision can be made on the classification of this group, the functional significance of these morphological differences needs to be much better understood. As far as the material studied is concerned, attempts to separate ruber and caffer show that, whereas most of the differ- ences between the two species are of size not proportion, there is a real difference of kind in the compartments of the nasal swellings. In all forms, in dorsal view (strongly illuminated from below), three pairs of compartments are visible: 1. anterior, more or less dorsal and lateral to the external nares; 2. medial, be- tween one and the olfactory part of the brain; and 3. postero- lateral to the others, and the largest. Variation in size of these is a good diagnostic character. When caffer is viewed from above, 1964 NOTES ON HIPPOSIDEROS 3 the relatively very small size of compartment two and the some- what larger size of three are sufficient to distinguish it from ruber. No doubt sections and other more detailed studies would show further differences in this region. Present knowledge of the small hipposiderids, excluding beatns, of east and central Africa can then be summarized as follows. In the savannah country of eastern Africa two species, ruber and caffer, occur together, of similar proportions externally and cranially, but conspicuously different in size. They may be fur- ther distinguished by the jDroportions of the nasal compartments. Near topotypes of niber from Kilosa in southern Tanganyika and s])ecimens from Mtimbuka south of Lake Nyasa resemble each other closely, while to the north a series from Mt. Elgon is somewhat intermediate towards a more western race. This western race has the skull as long as, but slightly narrower than, typical ruber and small nasals which give it the appear- ance of a large caffer. The proportions of the nasal compartments, however, are of the ruber type. The specimens at hand from the Ituri Forest, two from Rutshuru and one from Avakubi all be- long to this western form, as does probably a specimen from Ulkerewe Island. Whether the name centralis is available for this western race or whether Hollister is correct in synonymizing centralis with ruber, cannot be determined until the type series is re-examined. Since the type locality of centralis, Entebbe, Uganda, lies west of Mt. Elgon whence we have intermediate specimens, it seems advisable for the time being to continue to use centralis for the Congo race. As for the specific status of centralis, the structure of the nasal compartments and apparent intergrading with ruber are evidence of a closer relationship with this form than with caffer. This is in accord with Andersen's (1906, p. 281) subdivisions of caffer, except that his group 2, "H. c. centralis" ( = ruber) , is raised to specific rank and two subspecies of this, ruber ruber and ruber centralis, are recognized. There also occurs in central Africa a bat with ca^er-like nasal compartments but of nearly the same size as centralis. We have one specimen from Beni of this apparent race of caffer which we take to be nanus and probably the same as the series so identified by Verschuren. The sympatric occurrence of rep- resentatives of two rather similar species of small Hipposideros recognized by Hollister in East Africa seems to be typical of the Congo region as well. 4 BREVIORA No. 207 Verschuren's habitat notes for caffer centralis ( = ruber cen- tralis) and nanus {— caffer nanus) suggest that caffer may be typically a savannah form which has spread west where suitable country is to be found, while ruber, a forest form, has moved into dry country fairly extensively, possibly because of its adaptability to roosting in buildings. HiPPOSIDEROS BEATUS MAXIMUS VeRSCHUREN Two males, collected at Mabali, agree with Verschuren's (1957, p. 364) description in being larger than beatus from the Cameroons, and extend the range of b. maximus considerably to the southwest. Until Verschuren published his account of the bats of Garamba, it had generally been accepted that nanus was a small relative of beatus and represented this species in the northeastern Belgian Congo. Actually, as Verschuren has said (1957, pp. 370-72), the eastern relative of beatus is a larger not a smaller race and nanus is quite a different animal. As stated above, it is very likely a forai of caffer. With the discovery of this larger race, Andersen's (1906, p. 275) characters of tooth row and zygomatic arch no longer hold for the species. One of the specimens of beatus maximus from Mabali is almost identical in zygomatic-mastoid proportions with a similar-sized guineensis, a member of the caffer-ruber group, from Metet, while length of upper cheek teeth and width across the tooth rows taken outside IVP are exactly the same in both. As for the supposedly shorter tibia of beatus, the typical race is smaller than guineensis and so has a conspicuously shorter tibia. In the larger b. maximus, the tibia is only slightly shorter and the ratio of length of tibia to length of forearm while usually less than in guineensis occasionally shows no difference. A search for better diagnostic characters to separate the caffer-ruber races from beatus again has shown the importance of the nasal region. Cranially, these differences are hard to mea- sure though they are apparent enough to the eye. In beatus the rostrum bulks larger in proportion to the cranium with the width across the nasal swellings relatively greater in proportion to their total length. In spite of this, the inflation of the nasal region is less than in the caffer-ruber group. In particular, compartment three is relatively small as seen both in dorsal and in side view. Externally, width of nasals is usually reflected in width of nose leaf. While this character is often of subspecific value, specific relationships are better shown by differences in details of structure of nose leaves than by absolute size. Here again 1964 NOTES ON HIPPOSIDEROS 5 we find beatus sharply distinct from the ruber-caffer group. In beatus, the center of the horseshoe is shallower, and associated with this general difference are a number of particular char- acteristics. The septum slopes rather evenly inward to a point between the nostrils, whereas in caffer and ruber it has a very marked lumpy projection midway of its length, internal to which it makes a sharp angle where it meets the face between the nostrils. The nostrils themselves are less deep set and the pits ventral to them are shallower in beatus than in caffer and ruber. Further, in beatus, the two lateral accessoiy leaflets are of nearly equal length with a small warty growth dorsal to the end of the inner one, while in caffer and ruber the inner accessory leaflet is longer than the outer and the wart lies posterior rather than dorsal to its end. Verschuren's figures (1957, figs. 137, 141, and 146) show well the differences in septum between the two groups, while the pale rim around the nostrils seems to be associated with their shallowness in beatus. The difference which he figures in surface of the sella is harder to see in alcoholic specimens and the development of the medial thickening is not a character which separates all the caffer-ruber group from beatus. Other external characters which help to confirm the distinct position of beatus are the average more distal attachment of the wing membrane and less funnel-shaped ears. LITERATURE CITED Aellen, V. 1952. Contribution a I'etude des chiropteres du Cameroun. Mem. Soc. Neuchateloise Sci. Nat., 8, fasc. 1:1-121, figs. 1-26. 1957. Les chiropteres africains du Musee Zoologique de Strasbourg. Rev. Suisse Zool., 64, fasc. 1, no. 6:189-214. Andersen, K. 1906. On Hipposideros caffer, Sund. and its closest allies; with some notes on H. fuliginosus, Temm. Ann. Mag. Nat. Hist., (7) 17 : 269-282. HOLLISTER, N. 1918. East African mammals in the United States National Museum. Part I. Insectivora, Chiroptera and Carnivora. Bull. U. S. Nat. Mus., 99:1-194, pis. 1-55, figs. 1-3. Pereet, J. L. and V. Aellen 1956. Mammiferes du Cameroun de la collection J. L. Perret. Rev. Suisse Zool., 63, no. 26:365-450. Verschuren, J. 1957. Ecologie, biologie et systematique des cheiropteres. Institut des Pares Nationaux du Congo Beige, Exploration du Pare National de la Garamba, Mission H. de Saeger, fasc. 7:1-473, 1 pi., 1 map, figs. 1-178. I cO BREVIORA Miasenam of Compsirative Zoology Cambridge, Mass. November 12, 19G4 Number 208 THREE NEW SPECIES OF FROGS (LEPTODACTYLIDAE, ELEUTHERODACTYLUS) FROM HISPANIOLA By Albert Schwartz 10,000 S.W. 84th Street, Miami, Florida Among specimens collected during the summer of 1962 in southwestern Haiti are representatives of two undescribed species of Eleutherodactylus, both of which belong to the ricordi group. A third new species from the Republica Dominicana was taken during the summer of 1963 ; the affinities of this third species are distinctly with species from Haiti, and it is ap- propriate to describe the three new forms in a single paper. In Hispaniola, I have had the capable assistance of Miss Patricia A. Heinlein, and Messrs. Ronald F. Klinikowski, Dennis R. Paulson, David C. Leber and Richard Thomas. For significant comparative material I wish to thank Dr. Doris M. Cochran, United States National Museum (USNM), and Dr. Ernest E. Williams, Museum of Comparative Zoology (MCZ), whose as- sistance in these studies is very deeply appreciated. The il- lustrations are the work of Mr. Leber, and I wish to thank him and the above mentioned persons for their assistance in the field. The types have been given to the Museum of Comparative Zoology. Paratypes of the new forms have been deposited in the American Museum of Natural History (AMNH) and the Museum of Natural History, University of Kansas (KU) ; material in my own collection is designated as the Albert Schwartz Field Series (ASFS). In July 1962, during a stay at Camp Perrin in Haiti, Mr. Leber took an especially interesting and rather large frog from a nearby cave during a diurnal visit. Nocturnal visits to the cave resulted in the taking of no more specimens of this frog, so that 2 BREVIORA No. 208 it is represented only by one individual. Although I am reluc- tant to describe a new species of Eleutherodnctylus on such limited material, the species is so distinctive that I have no hesitancy in naming it as Eleutherodactylus counouspeus new species Figure 1 ^o " Holotype: MCZ 43199, an adult male, from Grotte de Counou Bois, 1 mile (1.6 km) southwest of Camp Perrin, Dept. du Sud, Haiti, taken 30 July 1962 by David C. Leber. Original number X3266. Diagnosis: An Eleutherodactylus of the ricordi group charac- terized by a combination of large size (snout-vent length 48 mm), large digital discs, a greenish yellow dorsal ground color with black markings consisting primarily of a dark interocular bar, a dark postocular V, and a scapular cross, a heavily mottled throat, and lacking inguinal glands. Description of type: An adult male, with the following meas- urements (in millimeters) : snout-vent length, 48.0; head length, 18.5; head width, 17.2; diameter of tympanum, 3.1; diameter of eye, 7.1 ; naris to anterior corner of eye, 6.4 ; femur, 21.3 ; Fig. 1. Eleutherodactylus counouspeus, new species, type, MCZ 43199, snout-vent length 48.0 mm. 1964 NEW FROGS FROM HISPANIOLA 3 tibia, 22.5; fourth toe, 18.1. Head slightly narrower than dis- tance from snout to posterior border of tympanum ; snout de- cidedly truncate with nares prominent at anterior end of can- thus rostralis ; diameter of eye greater than distance from naris to anterior corner of eye ; diameter of tympanum about one-half diameter of eye, distance from tympanum to eye about one- third diameter of tympanum ; tympanum oval, the vertical diameter slightly greater than the horizontal. Interorbital dis- tance 5.5, less than diameter of eye. Digital discs present and well developed, those on fingers three and four distinctly larger than those on digits one and two, disc of finger three the largest and equal to about three-quarters size of tympanum. Fingers long and slender, unwebbed, 3-4-2-1 in order of decreasing length ; subartieular tubercles prominent, gray. Toes moderately long, with vestigial webs, 4-3-5-2-1 in order of decreasing length ; subartieular tubercles prominent, gray. Heels touch when fe- mora held at right angles to body axis. Inguinal glands absent. Dorsum smooth ; upper eyelids with low rounded tubercles. Throat slightly granular, belly smooth ; abdominal disc poorly developed but lateral margins of disc fairly conspicuous. Dorsal surfaces of forelimbs smooth, of hindlimbs granular, especially on dorsal surfaces of crura. Posterior face of thighs with many small, juxtaposed rounded granules. Tongue small, entire, free behind, its greatest width equal to about one-half that of floor of mouth. Vomerine teeth in two short, almost straight series, beginning well within the inner margins of the choanae and separated from the choanae by a distance equal to about three times the diameter of a choana, the two series separated from each other by a distance equal to the diameter of a choana. Coloration of type: Dorsal ground color in life greenish yellow (slightly brighter than PI. 21L1 of Maerz and Paul, 1950), with dorsal black markings as follows: a black interocular bar, a black median postocular V, a black scapular cross with a rounded black spot anterior and lateral to the side "arms" of the cross, a rather linear blotch on the sides posterior to the forearm in- sertion, and the balance of the dorsum and sides with lighter and irregular obfuscations of dark (Fig. 1). Fore- and hind- limbs diffusely spotted with dark with no apparent banding or barring; concealed surfaces pale purplish gray overlaid with brown suffusions. Throat yellowish green, heavily mottled with brownish; belly pinkish with admixture of yellow-green, and suffused with brown stippling ; underside of crura rather heavily 4 BREVIORA No. 208 marbled with brown. Lores yellow-green with an indistinct darker canthal line. Iris bronzy in life, with a reddish pupillary ring. Comparisons: No other Hispaniolan members of the ricordi group reach the size of connouspeus except E. schmidti Noble and its races rucillensis Cochran and limhensis Lynn, and pos- sibly E. femurlevis Cochran. The webbed feet in E. schmidti, along with the entirely different color pattern and the reddish- orange venter will distinguish it from counouspeus. E. schmidti is known only from the Cordillera Central and the Cordillera Septentrional in the Republica Dominicana and from the area about Limbe in Haiti and is thus not sympatric with counou- spejis. I have not examined the type and apparently only speci- men of femurlevis, which has a snout-vent length of 42 mm (Cochran, 1941:62); however, the detailed description of the pattern as well as the photograph of the specimen {ihid., pi. IOC) indicate that there is little similarity between it and counouspeus. E. femurlevis has a pale, wide, interoeular bar banded narrowly anteriorly with brown, a brown loreal spot, pale gray crossbands on the legs, immaculate venter — none of which corresponds to the pattern of counouspeus. The photo- graph of the type does not show a conspicuously large-eyed frog, and the legs appear shorter and the discs much less en- larged. It has seemed most pertinent to compare these two species not only because of some similarity in size but also because they both occur in the same general region on the southwestern Haitian peninsula, femurlevis having been de- scribed from Desbarriere, on the north and east foothills of the Massif de la Hotte. Observations: The type of E. counouspeus was taken by Mr. Leber as it actively hopped about on rocks about fifty feet back from the entrance of the Grotte de Counou Bois in the morning. No other specimens were secured on a later visit and no voice was heard which might be associated with this species ; E. pictissimus was taken fairly abundantly in the same cave. In general habitus, E. counouspeus resembles E. zeus from Cuba ; both are rather long-legged, large-eyed species with large digital discs. The pattern of the two is vaguely similar although they are not at all comparable in coloration. Since both are petricolous and/or cavernicolous forms, I feel that these simi- larities are due to convergence and do not show relationships. In Antillean Eleutherodactylus there is often a remarkable 1964 NEW PROGS FROM HISPANIOLA 5 similarity in pattern and coloration between unrelated forms which are petricolous (Schwartz, 1960:41-42). A second new species of the ricordi group from the western portion of the Tiburon Peninsula is named in honor of Dennis R. Paulson who helped collect the type series and whose interest and diligence in field work in the West Indies deserves special mention : Eleutherodactylus paulsoni new species Figure 2 Holotype: MCZ 43200, a gravid female, from 4.5 miles (7.2 km) northwest of Les Cayes, Dept. du Sud, Haiti, one of a series taken 7 August 1962 by David C. Leber and Dennis R. Paulson. Original number X3796. Parahjpes: ASFS X3797-98, AMNH 71987-89, same data as type; ASFS X2799, Camp Perrin, Dept. du Sud, Haiti, 23 July 1962, collected by a native ; MCZ 43201, Grotte de Counou Bois, 1 mi. (1.6 km) SW Camp Perrin, Dept. du Sud, 26 July 1962, Lucien Rigaud; KU 79811, 4 km (6.4 km) NW Les Cayes, Dept. du Sud, Haiti, 7 August 1962, D. C. Leber; MCZ 33825, Place Negre, near Jeremie, Dept. du Sud, Haiti, Luc and George \¥hiteman, 11 December 1960. Diagnosis: A species of Eleutherodactylus of the ricordi group characterized by a combination of rather small size (females snout- vent length to 26 mm), very small digital discs, a dorsal Fig. 2. Eleutherodactylus paulsoni, new species, type, MCZ 43200, snout- vent length 25.4 mm. 6 BREVIORA No. 208 pattern of dark brown marbling with a pair of butfy dorso- lateral stripes, scattered dorsal pinkish warts, without red or orange on the hindlimbs, and with inguinal glands. Description of type: A gravid female with the following measurements (in millimeters): snout-vent length, 25.4; head length, 10.5 ; head width, 10.0 ; diameter of tympanum, 2.1 ; dia- meter of eye, 3.6 ; naris to anterior corner of eye, 3.0 ; femur, 11.9; tibia, 13.5; fourth toe, 11.5. Head slightly narrower than distance from snout to posterior border of tympanum ; snout truncate with nares prominent at anterior end of canthus ros- tralis; diameter of eye slightly longer than distance from naris to anterior corner of eye; diameter of tympanum greater than one-half diameter of eye, distance from tympanum to eye equal to about one-half diameter of tympanum. Fingers rather long and slender, unwebbed, 3-4-2-1 in order of decreasing length; subarticular tubercles prominent, gray. Toes moderately long, with vestigial webs, 4-3-5-2-1 in order of decreasing length ; sub- articular tubercles prominent, gray. Heels overlap when fe- mora held at right angles to body axis. Inguinal glands present. Dorsal surface, including eyelids, snout, and upper surfaces of all limbs heavily rugose ; a raised median dorsal line. Throat and belly smooth ; abdominal discs poorly developed. Posterior and ventral surfaces of thighs with many large, conspicuous juxtaposed granules. Tongue rather small, entire, free behind, its greatest width equal to about one-half that of floor of mouth. Vomerine teeth in two rather long, slightly bowed, series, be- ginning at outer margin of choanae and separated from the choanae by a distance equal to one-half the diameter of a cho- ana, the two series separated from each other by the same distance. Coloration of type: Dorsal ground color in life brown, with a rather irregularly marbled darker brown pattern, a dark brown interocular bar preceded by a paler tan snout, much marbled with dark brown. Scattered conspicuous warts from the level of the eyes onto the lower back, pinkish in life and strongly contrasting with the dark ground color. Dorsolateral stripes indistinct and buffy, almost completely obscured in the pre- served specimen by dark pigment. Dorsal ground color dis- tinctly more reddish posteriorly than anteriorly. Antebrachia with about three obscure crossbars, brachium marbled dark ; hindlimbs very obscurely crossbarred and dotted w^ith paler, their concealed surfaces dark brown, with some lighter stippling 1964 NEW FROGS FROM IIISPANIOLA 7 medially. Belly immaculate opalescent whitish ; throat with large to small discrete dots and stipples; underside of ante- brachia clouded with brown ; underside of thighs and crura rather heavily marbled and clouded with brown. Variation: Five adult females (including the type) have the following measurements: snout-vent length, 24.0 (21.5-25.8); head length, 9.6 (8.3-10.6); head width, 9.1 (8.1-10.1); tym- panum, 2.0 (1.9-2.1) ; eye, 3.5 (3.2-3.8) ; naris to eye, 2.8 (2.4- 3.2); femur, 11.3 (10.0-13.6); tibia, 12.3 (11.3-13.9); fourth toe, 10.7 (9.8-12.2). A single adult male paratype measures: snout-vent length, 22.1 ; head length, 8.8 ; head width, 7.8 ; tympanum, 1.9; eye, 3.3; naris to eye, 2.3; femur 10.4; tibia, 11.4; fourth toe, 10.7. Females reach a larger size than males; interestingly, the largest female (MCZ 33825) is from the north side of the Massif de la Hotte. The condition of this specimen is such that it is not readily compared with specimens from the south side of the Massif, but superficially it seems identical except for its slightly larger size. It is, of course, possible that such northern frogs may be found to be racially distinct from southern frogs. In color and pattern, the series shows variation principally in the distinctness of the dorsolateral stripes. When present, they are boldly outlined in dark brown and extend from the posterior portion of the upper eyelid to the groin. The brown dorsum with a more reddish (chestnut) tinge posteriorly was noted for all fresh material ; the pinkish warts are likewise a characteristic feature. The dorsolateral stripes vary from buffy to reddish; the interocular bar can be either tan or light buffy. The belly varied from white to translucent gray. The degree of throat spotting varies considerably, with the type showing the mean condition ; in some specimens the dots form almost a complex dark brown reticulum and in others they are larger and more uniform than in the type, but nonetheless discrete. All specimens, even the smallest juvenile (snout-vent 11.5), have the throat marked with dark spots. Comparisons: Shreve and AVilliams (1963:331) referred to a single specimen (MCZ 33825) under their discussion of E. furcyensis saying that "a single specimen from the vicinity of Jeremie . . . somewhat resembles this species and may be the La Hotte region representative." The specimen referred to is here designated as a paratype of E. paulsoni. Their comment, however, points out the comparison which it is most necessary 8 BREVIORA No. 208 to make — that between paidsoni and furcyensis: In coloration and pattern these two species are quite different. E. furcy- ensis has prominently banded crura and thighs, and lacks dorso- lateral lines. The anterior and posterior faces of the thighs are red (PI. 4D11) in furcyensis (these same areas are dark in paulsoni) and there is additionally a faint orange wash in the groin. Female furcyensis have the throat slightly purplish, flecked with black, whereas male furcyensis have the belly pale yellow; neither of these features occurs in paulsoni. Furcyensis has a dark brown dorsum with a pair of pale orange-buff re- versed parentheses, not a pair of dorsolateral lines. The two species do resemble each other in the throat spotting, which is a variable but always present feature in furcyensis, although no specimens at hand have the throat with a dark reticulum. Structurally, the two species differ in that paulsoni has in- guinal glands and furcyensis lacks them. In comparably sized specimens, the discs of furcyeyisis are larger than those of paulsoni. The differences in coloration and pattern are so strik- ing as to preclude the possibility that these two forms are related on a subspecific level. In size, furcyensis exceeds paadsoni in both sexes (largest female furcyensis snout-vent, 37.0; largest male, 28.0), although the small series of paulsoni makes comparison, difficult. At least fourteen adult female furcyensis have higher values in all measurements taken. The remaining ricordi group members in Hispaniola are : glandulifer, darlingtoni, weinlandi, rufifenioralis, schmidti, fe- murlevis, pictissimus, leoncci and counouspeus (list j^rincipally from Shreve and Williams but with some interpolations of my own). Of these species, schmidti and counouspeus are known to lack glands. These two forms are likewise much larger than paidsoni, are differently patterned, and schmidti has webbed feet. Of the forms with glands, paulsoni is exceeded in size by glandulifer, darlingtoni, weinlandi, pictissimus, and leoncei. None of these larger species have patterns which are in any way comparable to that of paulsoni, although the dorsum of weinlandi with its anteriorly and posteriorly differing dorsal ground colors is vaguely similar. E. darlingtoni has large digital discs, whereas leoncei, weinlandi and pictissimus have small digital discs. I have not examined specimens of either femur- levis or rufifenioralis; the latter is however much smaller than paulsoni (largest recorded snout-vent length 18) and is much 1964 NEW FROGS FROM HISPANIOLA 9 differently patterned. E. femurlevis, on the other hand, is much larger (snout -vent 42) and again with a different pattern. Observations : The type and paratopotypes were taken in leaf litter under large trees along the bank of an intermittent river (Riviere de la Grande Ravine du Sud) ; the adjoining area was thorn scrub. Another was taken at the cave entrance of the Grotte de Counou Bois, and a single individual from 4 km NW Les Cayes was taken on the ground among dead leaves in grass about 2 m high. At the type locality of paulsoni, E. pictissimus was the commonest frog encountered. Shreve and Williams (1963:338-9) discussed the Hispaniolan and Cuban components of the dimidiatus and varleyi groups. To the former they assigned the Cuban species alhipes, dimi- diatus, and emiliae and the Hispaniolan jugans, and to the latter the Cuban varleyi and cuhanus (which has customarily been placed in the dimidiatus group) as well as E. phyzelus Schwartz which I consider a synonym of E. varleyi Dunn. I concur in these assignments. Shreve and Williams also consider E. ventri- lineatus a member of the diynidiatus group (although it is pe- ripheral) in spite of its short vomerine series {dimidiatus group members are characterized by long vomerine series). E. ventri- lineatus is so obviously related to E. jugans that, except for the vomerine series, one might be inclined to regard them as con- specific. Thus, on Hispaniola, jugans and ventrilineatus are the only known members of the dimidiatus group. Both are restricted to higher elevations in the mountains of the south island, jugans in the Massif de La Selle, ventrilineatus in the Massif de la Hotte. A small series of a third Hispaniolan species belonging to this assemblage was taken by Richard Thomas in the Sierra de Neiba in the Republica Dominicana; since this form is the first recorded from the north island (and has thus "crossed" the Cul de Sac-Valle de Neiba plain), I propose that it be named, from the Greek word for "transgressor," Eleutherodactylus parabates new species Figures 3, 4 Holotype: MCZ 43202, a gravid female, from 20 km south- west of Hondo Valle, 5950 feet (1800 m), Independencia Pro- vince, Republica Dominicana, taken 11 August 1963 by Richard Thomas. Original number V366. 10 BREVIORA No. 208 Parahjpes: ASPS V365, V367-70, same data as type. Diagnosis: An Eleutherodactylus, related to E. ventrilineatus Shreve, with short vomerine series but differing from both ventrilineatus and jugans in smaller size, in the presence of digital discs, in ventral pattern, and in relative size of tym- panum. Description of type: A gravid female with the following measurements (in millimeters): snout-vent length, 24.1; head length, 8.6 ; head width, 8.8 ; diameter of tympanum, 1.7 ; dia- meter of eye, 2.7 ; naris to anterior corner of eye, 2.4 ; femur, 8.6; tibia, 9.8; fourth toe, 9.3; tympanum/tibia ratio 16.8. Head slightly broader than distance from snout to posterior border of tympanum ; snout rather pointed with nares conspicuous at anterior end of canthus rostralis; diameter of eye greater than distance from naris to anterior corner of eye ; diameter of tympanum greater than one-half diameter of eye, distance from tympanum to eye equal to about three-quarters diameter of tympanum. Interorbital distance 3.2 mm, greater than diameter of eye. Digital discs present and moderately well developed, those on fingers three and four larger than those on digits one and two, disc on finger three the largest and equal to about one-half size of tympanum. Fingers rather short, unwebbed, 3-4-2-1 in order of decreasing length ; subarticular tubercles Fig. 3. Eleiitlierodactylus pardbates, new species, type, MCZ 43202, snout- vent length 24.1 mm. 1964 NEW PROGS FROM HISPANIOLA 11 fairly prominent, gray. Toes moderately long, nnwebbed, 4-3-5-2-1 in order of decreasing length ; subarticular tubercles prominent, gray. Heels overlap slightly when femora are held at right angles to body axis. Inguinal glands absent. Dorsum smooth. Throat and belly smooth ; abdominal disc not prominent with only the pectoral fold moderately conspicuous. Dorsal surface of forelimbs smooth, of thighs smooth, but dorsal sur- face of crura studded with a few scattered low tubercles. Pos- terior face of thighs with many low rounded juxtaposed gran- ules. Tongue rather large, ovate, entire, free behind, its great- est width equal to slightly more than half of floor of mouth. Vomerine teeth in two short transverse patches, beginning at the midlevel of the choanae and separated from the choanae by a distance equal to about two and one-half times the diameter of a choana, the two series separated from each other by a distance equal to twice the diameter of a choana. Coloration of type: Dorsal ground color brown in a longitudi- nal band from snout to groin, enclosing a median tan dorsal line from snout to above vent, where it divides, each branch proceeding across the posterior of the thigh to behind knee; dorsal band separated from darker brown sides by a faint tan dorsolateral stripe which begins at the eye and proceeds to near the groin (Fig. 3). Sides darker brown. A tan canthal rig. 4. Eleutherodactylus paraiates, new species, ventral view, type, MCZ 43202. 12 BREVIORA No. 208 line above a dark brown stripe which continues posteriorly as a supratympanic stripe to above the forelinib insertion; lores and cheek brownish, lips mottled. Dorsal surface of forelimbs brown, brachia slightly paler, antebrachia with remnants of two or three crossbars. Dorsal surface of hindlimbs dark brown, as are con- cealed surfaces, with a vague remnant of one darker crossbar on the crura ; a darker brown postanal triangle present. Throat and belly yellowish tan, heavily mottled with brown ; the mottling consists of dots of varying intensities, with the large dots gen- erally darker, the smaller lighter, giving an irregularly mottled appearance (Fig. 4). Undersurface of all limbs mottled with brown, heaviest on crura and pes. Variation: Four gravid females (including the type) have the following measurements (means followed by ranges) and ratios: snout-vent length, 23.8 (22.3-24.3); head length, 8.6 (8.0-8.9) ; head width, 8.9 (8.5-9.4) ; tympanum, 1.6 (1.5-1.7) ; eye, 2.7 (2.6-2.8) ; naris to eye, 2.1 (1.9-2.4) ; femur, 8.6 (8.3- 9.0) ; tibia, 9.6 (9.3-9.8) ; fourth toe, 9.2 (9.0-9.3) ; tympanum/ tibia ratio, 16.5 (15.6-17.3) ; tibia/snout-vent length, 40.3 (38.3- 43.0). A single adult male measures: snout-vent length, 17.6; head length, 6.2; head width, 6.7; tympanum, 1.0; eye, 2.0; naris to eye, 1.5; femur, 6.7; tibia, 7.3; fourth toe, 7.2; tym- panum/tibia ratio, 13.7 ; tibia/snout-vent length, 41.5. In life the dorsal ground color of the entire series was like that of the type ; however, only one other paratype has the median dorsal line. In this specimen the dorsal band is some- what lighter than the type and there is conspicuous deposition of dark pigment adjacent to the median line and laterally along the region of the dorsolateral line, so that the dorsal band is rather conspicuously cut off from the sides, and the median line is also set off from the dorsal band. The four specimens which lack the median line also seem to lack a dorsal band. In these there is a dark interocular bar setting off a paler (more grayish) snout, and the sides are vaguely barred posteriorly. These specimens also have a single dark antebrachial bar and a single crural bar with a faint indication of a thigh crossbar as well. Even the smallest juvenile (snout- vent 11.0) shows the dark ventral mottling of the type, which is also a standard fea- ture of the balance of the series. One female paratype addi- tionally has a pair of tan scapular spots, each accompanied by a more laterally placed darker brown spot. 1964 NEW FROGS FROM HISPANIOLA 13 Comparisons: Only two known Ilispaniolan species of Eleu- therodactylus are comparable to E. parahates in squatty habitus — E. jugans and E. ventrilineatus. From both of these, para- hafcs differs in possessing- digital discs. From jugans, par abates differs as well in having a short vomerine series. I have not seen either jugans or venfrilincafus in life, and preserved specimens are all rather old and discolored. However, specimens of ventri- lineatus lack a middorsal line, which occurs in parahates. Jugans has a double crossbar on the crura and about three crossbars on the thigh, whereas ventrilineatus has a single bar as does parahates. The ventral coloration of the three species is very distinct ; the differences between jugans and ventrilineatus are well shown by Cochran (1941:34 and 35), the former having the belly heavily and irregularly pigmented, and ventrilineatus having the belly uniformly pigmented with a median ventral clear line. Thus, parahates has the least ventral pigmentation of the three. Of the three species, jugans reaches the largest size (females to snout-vent length of 33.3), with ventrilineatus smaller (fe- males to 30.6) and parahates smallest (females to 24.3). The single male of ventrilineatus (snout-vent length 24.9) is larger than the largest male jugans (22.5) or parahates (17.6). Para- hates and ventrilineatus may be separated by the tympanum/ tibia ratio; in the former, this ratio ranges (in females) between 15.6 and 17.3, in the latter (females) between 20.0 and 21.9; the single male parahates has a ratio of 13.7, the single male ventrilineatus 19.6. Tibia/snout-vent length ratio likewise is diagnostic between these two species, varying in parahates be- tween 35.6 and 37.6 and in ventrilineatus between 38.3 and 43.0 (all females). Ohservations and comments: The type series of E. parahates was collected under rocks and wood adjacent to a road through dense rain forest in the Sierra de Neiba (southern range). The forest in this area was composed of hardwoods and some ehano vercle (Magnolia doming uensis) and in the vicinity were low areas which supported small marshes with cat-tails, a rather unusual feature for such a high elevation. The members of the dimidiatus group, as presently understood are: dimidiatus Cope (with its race amelasma Schwartz) emiliae Dunn intermedius Barbour and Shreve 14 BREVIORA No. 208 alhipes Barbour and Shreve jugans Cochran ventrilineatus Shreve parabatcs Schwartz Of these, four are Cuban and three are Hispaniolan ; two of the Hispaniolan species occur on the south island, and the other just across the Cul de Sac-Valle de Neiba plain, in the Sierra de Neiba. Ve^itrilineatus and parabaies are slightly aberrant in that they both possess short vomerine series. Para- hates also has digital discs; as originally diagnosed (Dunn, 1926 :210) this group was composed of members with very feebly developed discs. Aside from the color features which Dunn considered as diagnostic of the group, the structural fea- tures regarded as pertinent are : 1 ) discs present and small, or absent, 2) smooth belly, 3) long vomerine series (or short in two species), 4) no external vocal sac, 5) squatty habitus, and 6) inguinal glands absent. The last character may well not be valid ; I have examined only dimidiatus, jugmis, ventri- lineatus and parahates for this feature. Of the included species, E. dimidiatus, as Shreve and Wil- liams (1963:338) pointed out, is somewhat different in that it lacks the stocky build of the remaining forms. It too is the only known species which is not restricted to high elevations, although it does occur at high elevations in the Sierra de Trini- dad and Sierra de Grand Piedra in Cuba. As far as known, emiliae occurs only in the Sierra de Trinidad, and intermedius and alhipes in the Sierra Maestra. There is a possibility that E. unicolor Stejneger is the single Puerto Rican representative of the dimidiatus group. Relying on Stejneger 's (1904:597-598) description of the type, I feel that unicolor agrees with my concept of the group in all features except that the belly is granular rather than smooth. Additionally, it is a high mountain form, and thus agrees with most dimidiatus group species. No Jamaican species are obvi- ously associated with the dimidiatus group. I have had field experience only with dimidiatus. At least this species is vocal; the call is an insect-like twittering, males calling from the ground in mesic forested situations. No calls have been reported for any other species. Specimens examined: Eleutherodactylus ventrilineatus: MCZ 19857-61, 19863 (type and paratypes), Mt. La Hotte (=Pic Macaya), Dept. du Sud, Haiti; Eleutherodactylus jugans: MCZ 1964 NEW FROGS FROM HISPANIOLA 15 21594-95, MCZ 19852-56, plus four unnumbered specimens (type and paratypes), near La Visite, La Selle Range, Dept. de 1 'Quest, Haiti; USNM 95423-27 (paratypes), La Visite, Dept. de I'Ouest, Haiti. LITERATURE CITED Cochran, Doris M. 1941. The herpetology of Hispaiiiola. Bull. U.S. Nat. Mus, 177:i-vii, 1-398, 120 figs., 12 pis. Dunn, E. R. 1926. Additional frogs from Cuba. Occ. Papers Boston Soc. Nat. Hist., 5:209-215. Maekz, a., and M. Rea Paul 1950. A dictionary of color. New York, McGraw-Hill Book Co., pp. i-vii, 1-23, 138-208, 56 pis. Schwartz, Albert 1960. Nine new Cuban frogs of the genus EleutJierodactylus. Reading Public Museum and Art Gallery, Sci. Publ., 11:1-50, 6 figs. Shreve, Benjamin, and Ernest E. Williams 1963. The herpetology of the Port-au-Prince region and Gonave Island, Haiti. Pt. 11. The frogs. Bull. Mus. Comp Zool., 129(5) :302- 342, 5 pis. Stejneger, Leonhard 1904. The herpetology of Porto Rico. Rept. U.S. Nat. Mus., 1902:549- 724, 1 pi., 196 figs. BREVIORA MiLiseiui.-nTi of Connparative Zoology Caimbridge, ]\Iass. .\()\ f.mhkr 12, 19()4 Number 209 A XEAV SKATE. RAJA CERVIGONI, FROM VENEZUELA AND THE GUI ANAS By TTexkv T>. Bigelow and William C. SchroederI Raja cervigoni sj). iiov. Holofype: An immature male, 357 mm in total length, from 10 miles (16 km) northeast of Carupano, in 20-30 fathoms (37-55 m), Museo Historia Natural La Salle, Venezuela, No. 873. Paratjjpes: A male of 206 mm and a female of 229 mm from Punta Arava, Estado Sucre, in about 20 fathoms (37 m), from off the eastern part of Venezuela, and a male of 235 mm from off tlic Guianas, 07° 25' N, 54° 35' W, in 75-80 fathoms (137- U5 m), "Oregon" station 2289. Distinctive characters. Among rajids from the western Atlan- tic that have a pair of ocelli on the disc, cervigoni most ckjsely resembles Raja cyclophora Regan 1903, but it differs in lacking dark mucous pores on the under surface, in having orbital and nuchal thorns, and 3 rows of thorns on the tail, whereas cyclo- phora has prominent blackish streaks below marking mucous pores, lacks orbital and nuchal thorns, and has a single row of thorns on the tail. In cervigoni the more sharply rounded outer corners of the disc, which is also relatively wider (averaging 73.2 per cent of the total length of the specimen on 4 indi- viduals), distinguishes it from texana Chandler 1921 (average disc width 63.8 per cent on 23 specimens) and also from ackleyi Garman 1881 (average 59.4 per cent on 3 specimens) ; and its fewer thorns (16-26) in the midrow on the tail, between the axils of the pectorals and the first dorsal fin, tcgether with the axis of greatest width further rearward (73-77 per cent, meas- ured between the tip of the snout and axils of pectorals), set 1 Contrilnition No. 1522 from rhn Woods Hole Oceanograiihic Institution. 2 BREVIOKA No. 209 it apart from hahanu usis Hv^ehnv and Schroeder- on which the thorns number 34-47 (4 specimens) and the greatest disc width is 64-69 per cent rearward. Description of Jioloti/jx'. Proportional dimensions in per cent of total length. Disc. — Extreme breadth 72.8 ; length 52.1. Snout length. — In front of oi-bits 12. 6 ; in front of mouth 14.6. Orbits. — Horizontal diameter 4.4; distance between 4.2. Spiracles. — Length 2.8 ; distance between 6."). Mouth. — Breadth 8.4. Exposed nostrils. — Distance between inner ends 8.2. Gill openings. — Length 1st 2.0; 3rd 2.0; 5th 1.5; distance be- tween inner ends, 1st 15.6 ; 5th 8.0. First dorsal fin. — Height 3.1 ; length of base 5.3. Second dorsal fin. — Height 3.1 ; length of base 4.8. Pelvics. — Anterior margin 13.2. Distance. — From tip of snout to center of cloaca 47.6 ; from center of cloaca to 1st dorsal 33.0 ; to tip of tail 52.4 ; from rear end of 2nd dorsal to tip of tail 3.7. Interspace. — 1st and 2nd dorsals 5.6. Disc 1.4 times as broad as long; maximum angle is front of spiracles 102°; snout pointed; anterior margins of disc nearly straight, outer corners shari)ly rounded; posterior and inner margins moderately convex. Axis of greatest breadth 73 per cent of distance back from tij) of snout to axils of pectorals. Tail rath]k G57-658. Garman, Samuel 1881. Eeports on tlic results of ull. .Mus. Comp. Zool., vol. 8, no. 11, pp. 231-2.37. Regan, C. T. 1903. On a collection of fishes made li\' Dr. (ioeldi at Eio .Janeiro. Pioc. Zool. Soc. Lomlon, jiai-t 2, pj). n9-68, 2 pis. 1964 NEW SPECIES OF SKATE 5 Haja cenigoni sp. nov., holotype, immature male, 357 mm long. i^MJ BREVIORA Mmseiuiini of Compsirsitive Zoology Cambridge, Mass. November 12, 1964 Number 210 THE ANTS OF THE FLORIDA KEYS By Edward 0. Wilson Biological Laboratories, Harvard University The Florida keys should be of more than usual interest to zoogeographers on several grounds. They are an extensive sub- tropical archipelago adjacent to the United States mainland, easily reached by a main highway that runs for almost their entire length. They are in the hurricane belt, providing an opportunity to study the effects of severe periodic storms on population dynamics and dispersal. Finally, they are separated by the Florida current of the Gulf Stream — a narrow but formidable faunistic barrier — from the topographically similar Bimini Islands, with which they can profitably be compared. These advantages induced me to make a special collecting trip in June, 1958, to conduct an initial faunal survey. Time permitted only several islands to be investigated thoroughly. I selected Key Largo, Plantation Key, Big Pine Key, and Key West in order to insure the maximum geographic spread and diversity of habitats (see Davis, 1943, and Duellman and Schwarz, 1958). The ant fauna of the Bimini Islands had al- ready been surveyed by M. R. Smith (1954), while other studies of Floridian ants were available in the publications of Wheeler (1932) and Van Pelt (1956, 1958). The project was supported in part by a grant from the Na- tional Science Foundation. I am grateful to my wife Irene for her assistance in the collection of specimens and recording of ecological information, ZOOGEOGRAPHIC ANALYSIS In Table 1 the Keys fauna is classified according to nest site and probable origin. From a consideration of this partition plus 2 BREVIORA No. 210 other data given in the Systematic List later, the following generalizations are permissible. 1. Elements introduced by human commerce are numerically more prominent in both species and individuals than on the Florida mainland. Eight of the 30 species recorded, or 26.7 per cent, are in this category. In the AVelaka Reserve in central Florida, according to published data by Van Pelt (1958), 8 of 75 species, or only 10.7 per cent are in this category. The prev- alence of introduced "tramp" species is shared with most other small islands in the tropics, including Bimini, hence is a distinctly "insular" trait. (Of 12 introduced species occurring jointly in the Keys and Welaka Reserve, onlj^ four are held in common : Cardiocondyla emeryi, Tetramorium guineense, Mono- morium floricola, and Paratrechina longicornis.) 2. The Antillean elements are mostly arboricolous, whereas the Floridian elements are mostly terricolous. 3. Most of the arboricolous (hence Antillean) elements are very well adapted to life in the mangroA'e swamps, maintaining unusually dense populations there. The same species also oc- cur, for the most part, in other habitats in the interiors of the Keys. 4. The arboricolous elements make up a disproportionate part of the Keys fauna. Twelve of the 30 species, or 40 per cent, are in this ecological category ; on the other hand, only 12 of 75, or 16 per cent, are arboricolous in the Welaka Reserve of Central Florida. Judging from data by Smith (1954), the Bimini fauna resembles that of the Florida Keys in this respect ; of 23 species for which there is ecological information, 8 or 34.8 per cent are arboricolous. Generalizations (2), (3), and (4) are probably interrelated. The mangrove swamps provide an excellent portal into such small islands as the Florida Keys. They form an extensive, per- sistent habitat around most of the margins of the islands. Dur- ing severe storms, such as the famous hurricanes of 1935 and 1960 (Craighead and Gilbert, 1962), the mangroves were badly damaged but not exterminated. In 1960 many tree branches were torn loose and undoubtedly transported long distances by sustained winds of 140 mph over a period of nearly 36 hours. Such debris, unless submerged in water or literally torn to shreds, should provide vehicles of transport for intact colonies 1964 ANTS OF THE FLORIDA KEYS 3 of the arboricolous ant species found in the Keys, although the point has not been confirmed. The rapid regrowth of the man- grove swamps would provide the means for the reproduction of these propagules. The terricolous ant species, on the other hand, are not so well favored. Subterranean and log-dwelling colonies are not likely to be transported intact by the high winds. At the same time they are much more susceptible to mortality by drowning. In the hurricanes of 1935 and 1960, tides of 11 to 18 feet were reported. The effects on the ground ant fauna were not studied but must have been extensive. In sum, it is reasonable to postulate a higher immigration rate and lower extinction rate for arboricolous species in comparison with terricolous species in the Florida Keys. Whether or not the fauna is in equilibrium (see MacArthur and Wilson, 1963), it follows that it should consist disproportionately of arboricolous species. Hence generalization (4) of the present study seems to have a reasonable explanation in one aspect of the physical environment. But further field studies before and after tropical storms are clearly needed to illuminate this interesting situation. The faunas of the Florida Keys and Bimini Islands show some striking differences. There are, to begin with, the set of in- ferred ecological vicars listed in Table 2. In at least one case, Solenopsis geminata versus Pheidole megacephala, the two species are known to compete directly and replace each other on very small islands in the West Indies and Pacific. Whether this is also the case for the other pairs might be determined by ex- perimental introductions. According to the hypothesis, such introductions of one vicar into the territory of the other should result in few successful colonizations. Certain species appear to have no ecological equivalents. For the Bimini group this is true of the Antillean and endemic elements Macromischa pastinifera, Smithistruma nigrescens, and Brachymyrmex ohscurior. For the Florida Keys it is evidently true of the species Aphaenogaster miamiana and Xenomyrmex floridanus. Again, experimental introductions, conducted for comparative purposes with exchanges of the hypothesized ecological equivalents listed above, might prove instructive. According to the hypothesis, this second class of introductions should prove relatively successful. 4 BREVIORA No. 210 SYSTEIVIATIC LIST Platythyrea punctata (Fr. Smith) Center of Key Largo, winged queen at light, June 14. Ranges from northern South America to southern Texas, also through- out the West Indies to the Bahamas and southern Florida. Odontomachus ruginodis AVheeler Key Largo, Big Pine Key. According to W. L. Brown {in litt.) ruginodis is the correct name for the Floridian species, whereas insularis is correctly applied to the species referred to by Smith (1945) as insularis var. pallens Wheeler. 0. ruginodis ranges from the United States through the West Indies to South America as far south as Paraguay. Within the United States it is found throughout Florida and reaches extreme southern Georgia and southeastern Alabama. An isolated population, probably introduced, occurs within the city limits of Mobile, Alabama. This large ponerine, distinguished by its trap-like mandibles, was very abundant in undisturbed tropical hammocks in the center of Key Largo. Colonies were found nesting in cavities in the rather thick leaf litter. On Big Pine Key two colonies were found nesting beneath logs in open pine-palmetto woods. PSEUDOMYRMEX ELONGATUS Mayr Key Largo, Plantation Key, Big Pine Key, Key West (winged males in nest June 21). Ranges from Brazil north through the West Indies to the Bahamas and southern Florida. This typically tropical species was abundant on trees in open habitats, including mangroA^e swamps, lawns, and disturbed hammocks on Plantation Key. On Key Largo it did not pene- trate the denser hammocks in the center of the island. Colonies were found nesting in hard dead twigs attached to living trees. PsEUDOMYRMEX PALLiDUS Fr. Smith Big Pine Key. A species adapted primarily to savannas, P. pallidus ranges around the Caribbean from the West Indies through the coastal plain of the Gulf States into Mexico. 1964 ANTS OF THE FLORIDA KEYS 5 On Big Pine Key, F. pallidus occurred in both the mangrove swamps and on grass in the open pine-palmetto woods of the interior. Colonies were found nesting in dead hollow twigs at- tached to live mangrove trees. Cardiocondyla emeryi Forel Plantation Key, Key West. A pantropical "tramp" species that originated in Africa and has been introduced in the New World, including the warmest parts of the Gulf States, by human commerce. Workers were found foraging on hot, bare ground in open habitats during the day. The species is generally limited to the most disturbed habitats. Crematog ASTER ASHMEADi Mayr Key Largo (a series in Museum of Comparative Zoology col- lected in 1904), Plantation Key, Big Pine Key, Key West. Limited to the United States, from the Florida Keys north to North Carolina and southern Tennessee and w^est to Texas. The workers occur in trees in mangrove swamps, lawns, and disturbed hardwood forest. None could be found in the deep hammocks of Key Largo. Nests were found in dead, dry branches of living trees and (in one instance) under loose bark on the trunk of a small tree. Pheidole floridana Emery Big Pine Key (males in nest June 19-20). The above specimens compare well with floridana syntypes and are closer to other Florida material placed with this species than to the closely related (and possibly conspecific) flavens Roger of Cuba and the Bahamas. Both floridana and flavens are highly variable but can be distinguished by apparently consistent sculp- tural characters in the soldier. Floridana is known only from Florida, and the Florida mainland is therefore the inferred source of the Keys population. This small yellow species was abundant in open pine-palmetto woods, nesting in the soil beneath coral rock and rotting logs. Each colony appeared to contain between 100 and 200 workers. 6 BREVIOBA No. 210 Aphaenogaster miamiana Wheeler Key Largo (North). This native species, a typically Nearctic element, ranges over all or most of Florida, just reaching extreme southeastern Alabama. A single colony, containing between 30 and 50 workers, was found nesting in a rotting log at the edge of a clearing in a dense tropical hammock. Tetramorium guineense (Fabricius) Key Largo (Center). This species originated in Africa and has been spread by human commerce throughout the tropics of both hemispheres. It is a characteristic element of disturbed habitats everywhere, even in the smallest, most remote oceanic islands. Workers were found foraging at sunset on the trunk of a guava tree and fronds of a small royal palm in a lawn in the center of Key Largo. Tetramorium simillimum (Fr. Smith) Key West. Like T. guineense, this little species evidently originated in Africa and has been spread by commerce throughout the tropics. Whereas guineense is predominantly arboreal, simillimum is mostly or entirely terrestrial. Paracryptocerus (Cyathomyrmex) varians (Fr. Smith) Key Largo (North), Plantation Key (winged queens June 14), Big Pine Key, Key West. According to W. W. Kempf (1958), varians occurs in southern Florida, south of Miami, in the Bahamas (Andros, New Provi- dence, Bimini), and in Cuba and Jamaica. Previously published records of the species from Honduras, Trinidad, and northern South America were based on misdetermined specimens of P. pallens (Klug). The species is abundant in the Florida Keys in a variety of major habitats. The workers are exclusively arboreal and noc- turnal. Colonies were found nesting in hard, dead branches, one to two inches in diameter, attached to living trees of several species, including gumbo limbo, cocoplum, and mangrove. A 1964 ANTS OF THE FLORIDA KEYS 7 dealate queen, evidently in the act of nest-founding, Mas found in a dead branch on Plantation Key June 15. She was blocking the opening of a small cavity with her saucer-shaped head. Strumigenys GUNDLACHi (Rogcr) Key Largo. This characteristic Neotropical species also occurs in tropical Mexico, Central America, Trinidad, and the Greater Antilles. Brown (1959) records it from the Everglades National Park in Florida. Two specimens, a worker and a dealate queen, were collected in leaf litter in a dense, relatively undisturbed hammock in northern Key Largo. Xenomyrmex floridanus Wheeler Key Largo (North), Plantation Key (winged males June 14, winged males and queens June 16), Key West (males June 21). According to the recent revision of the genus by Creighton (1957), X. floridanus occurs in southern and central Florida, the Bahamas, Cuba, and Mexico. Colonies were found in abundance along the edge of forest and in mangrove swamps everywhere I collected. They nested exclusively in dead branches of trees. AVorkers were found foraging singly and in files during the day. They are apparently exclusively arboreal. MoNOMORiUM destructor (Jcrdou) Key Largo, Plantation Key. This aggressive little species originated in the Old World tropics, possibly Asia, and has been spread by human commerce throughout the tropics. Workers were abundant in lawns; a single colony was found nesting in the soil of a grassy roadstrip in North Key Largo. MoNOMORiUM FLORicoLA (Jcrdon) Key Largo (North), Plantation Key, Key West. M. fioricola is a pantropical tramp species that originated somewhere in the Old World tropics. Several colonies were located in dead branches of standing trees at the edge of hammocks. 8 BREVIORA No. 210 MoNOMORiUM PHARAONis (Linne) Plantation Key. This remarkable species is perhaps the ant most intimately associated with man. Originating in Africa, it has been spread by commerce throughout the world. In the tropics it nests out- of-doors in disturbed habitats, while in temperate zones it is abundant in greenhouses and dwellings. On Plantation Key it was discovered in both situations. Colonies were abundant in dead tree branches in disturbed native w^oods, and workers were foraging in the walls of a restaurant far from any native woods. SoLENOPSis (S.) GEMiNATA (Fabricius) Big Pine Key. This species, commonly referred to as the native fire ant, is found throughout the New World tropics, ranging northward well into the Gulf States. A reddish color phase, to which the Big Pine Key series belongs, occurs through much of this range and in addition has been carried by human commerce to many parts of the Old World tropics. It is especially successful in open habitats. On Big Pine Key workers were found foraging over the crushed coral surface of a parking lot. SoLENOPSis (Euophthalma) globularia (Fr. Smith) Plantation Key, Big Pine Key (winged queens in nest June 19-20) . aS^. globularia ranges from the extreme south of Alabama and Mississippi through Florida and the West Indies to Mexico, Central America, Brazil, and the Galapagoes (Creighton, 1930). Through much of this range it is limited to the coast. In colora- tion and propodeal sculpturing the Keys samples are closer to series from Florida (subsp. littoralis Creighton) than to series from Cuba, Haiti, and Puerto Rico. Hence the Keys population can be inferred to have originated from Florida. Near the center of Big Pine Key a single colony was found nesting in soil under a piece of coralline rock in pine-palmetto woods. SOLENOPSIS (Diplorhoptrum) longiceps M. R. Smith Key Largo (North). Workers collected at the above locality correspond well with 1964 ANTS OF THE FLORIDA KEYS 9 loncficeps paratypes in body proportions, pilosity, and color but are somewhat smaller in size. Longiceps ranges from Florida to Texas north to Tennessee (Creighton, 1950). A single colony was discovered nesting in a small cavity in firm leaf litter on the floor of an undisturbed tropical hammock. SoLENOPSis (Diplorhoptrum) picta Emery Plantation Key (winged queens and males in nest June 15, 1958) ; Lower Matecumbe Key (collection by W. M. Wheeler, 1930). The species is known from extreme southern Alabama through Florida to the West Indies and Central America. The Plantation Key colony was found nesting in a dead stem in disturbed native woods. 8. picta is characteristically an ar- boreal species. Cyphomyrmex minutus Mayr Key Largo (North) (winged queens and males in nest June 16) ; Plantation Key. This common, primitive little fungus-grower ranges from ex- treme southern Alabama through Florida to the Bahamas, Greater Antilles, tropical Mexico, Trinidad, and South America as far south as Manaos. On Key Largo a colony was discovered in a small, crumbling log in a relatively undisturbed tropical hammock. On Plantation Key one colony was nesting in the soil beneath a piece of coral- line rock. Another was in rotting wood in a tree hole about three feet from the ground. Both were in disturbed, open na- tive woods. Tapinoma littorale Wheeler Key Largo (North), Plantation Key (males in nest June 15), Big Pine Key (males June 19-20), Windley Key (queens and males at light June 17). The species is distributed from extreme southern Florida to the Bahamas, Cuba, and Puerto Rico. This was one of the most abundant ant species in leaf litter on the floor of undisturbed tropical hammocks on Key Largo. Colonies were found nesting in small pieces of rotting wood buried in the litter. Colonies were also abundant in dead stems attached to living bushes and trees, especially in disturbed forest 10 BREVIORA No. 210 on Plantation Key and the mangrove swamp around Big Pine Key. Each colony appeared to consist of 100 to 300 workers and, in at least one case, up to several nest queens. CONOMYRMAPYRAMICA (Roger) Big Pine Key, Key West. This distinctive, terrestrial dolichoderine ranges from New Jersey to Florida west to Arizona and southward through the West Indies to Mexico, Central America, western South America and the Galapagoes. Colonies were found in open soil in elevated spots in the mangrove swamps. Paratrechina (P.) LONGicoRNis (Latreillc) Key Largo, Plantation Key. This species ranks with Tetramorium guineense as one of the most abundant and ubiquitous of all the pantropical "tramp" species of ants. It is characteristic of open, dry, highly dis- turbed habitats, from farmland to the centers of the largest cities. It originated somewhere in the Old World tropics. On the Keys listed above, P. longicornis was common in com- pletely open situations, especially around human dwellings. It was not found in the native woods. During a field trip to eastern Cuba in 1953, I noticed a similar distribution. P. longicornis workers abounded through sugar cane fields and along roads up to the very edge of the native forests that clung to limestone outcrop])ings ; inside the forests, over a distance of only a few feet, they were replaced by native Cuban ant species. Paratrechina (Nylanderia) bourbonica Forel Key West. This is another prominent pantropical "tramp" species of Old World origin. The Keys specimens have been compared with a syntype worker in the collection of the Museum of Com- parative Zoology. Workers were commonly seen foraging during the day at several locations on the streets of residential sections of Key West. Paratrechina (Nylanderia) parvula (Mayr) Key Largo (North). The workers of species in the parvula complex are too similar 1964 ANTS OP THE FLORIDA KEYS 11 to make tlie present identification more than tentative. P. par- vula, according to Creigliton (1950), ranges from soutliern New York west to Iowa and Texas and south to Florida. A single colony was found nesting in a small piece of rotting wood buried in leaf litter on the floor of a relatively undisturbed tropical hainmoek. Camponotus abdominalis ploridanus (Buckley) Key Largo, Plantation Key, Big Pine Key. C. floridanus is such a well-defined form with reference to the remainder of the abdominalis complex that it may well be a distinct biological species. I have used the trinomen in this case as an indication of the uncertainty of its biological status. Floridanus ranges from the east shore of Mobile Bay, Alabama, west through southern Georgia and south to the Florida Keys. The remainder of the abdominalis forms, which may or may not constitute a single species, range from southern Texas to Ecua- dor, the Amazon basin, and thence north again to the Lesser Antilles. The species is evidently absent from the Greater Antilles. Floridanus is one of the most abundant and adaptable ant species in the Keys. It occurs in undisturbed and disturbed native hammocks and on the lawns around human dwellings but is evidently absent from mangrove swamps. Nests are formed in leaf litter, under rotting logs, and in dead branches of living trees. Workers forage during the day, mostly on vegetation. Camponotus planatus Roger Plantation Key (winged queens and males in nest June 14-18), Key West. The species is found in extreme southern Florida and Texas, Cuba, Mexico, and Central America. C. planatus is a very abundant species in tropical hammocks but rare or absent elsewhere. On Big Pine Key one colony was found in a small, isolated hammock, but not a single specimen was collected in the extensive pine-palmetto woods that dominate the unsettled part of the island. Unlike floridanus, planatus is absent from the vicinity of human dwellings. Colonies were found nesting in dead tree branches, both attached to living trees and lying loose on the ground. The alert, swift-running workers forage during the day. 12 BREVIORA No. 210 Camponotus tortuganus Emery Plantation Key (winged queen June 14), Big Pine Key. As Creigliton (1950) points out, the status of this form, originally described as a subspecies of the widespread tropical species C. maculatus, will remain unsettled until a careful revi- sion can be made of the difficult species group to which it be- longs. Tortuganus was originally described from the Dry Tor- tugas. As currently delimited it occurs on the Keys and in the Florida mainland as far north as Lake Worth. A single colony was found beneath a rotting log in pine- palmetto woods on Big Pine Key. Camponotus (Colobopsis) sp. Big Pine Key. A single colony belonging to the subgenus Colohopsis was found nesting in a dead stem in a mangrove swamp on the south shore of Big Pine Key. It is similar to C. impressus (Roger) but differs sufficiently in the shape of the propodeum of the minor worker to warrant distinguishing it as a different, and possibly undescribed species. Additional materials considered in the context of a generic revision are needed to clarify the matter. SUMMARY Thirty ant species are listed from the Florida Keys. The disproportionate fraction of introduced species is cited as a typically insular feature of the fauna. Also, an exceptional prevalence of arboricolous (and Antillean) species is noted and is hypothesized to be the outcome of the periodic severe tropical storms that alter the Keys environment. Finally, comparisons are made with the nearby Bimini Islands fauna. 1964 ANTS OF THE FLORIDA KEYS 13 Table I Ecology and Zoogeographic Origin of the Keys Species Terricolous FbORiDiAN: Aphaenogaster miamiana, Phcidole floridana, Solenopsis glohidaria, S. longiceps, Paratredhina parvula. Antillban: Odontomachus ruginodis, Strumigenys gundlachi, Cyphomyr- mex minutus. Natives of uncertain origin : Solenopsis geminata, Conomyrma pyra- mica, Camponotus tortuganus. Introduced by man: Cordiocondyla emeryi, Tetramorium simillimum, Paratrechina longicornis, P. hourhonica. Arboricolous Floridian : Crematogaster ashmeadi. Antillean: Platythyrea punctata, Pseudomyrmex elongatus, Xenomyr- mex floridanns, Solenopsis picta, Paracryptocerus varians, Tapinoma lit- torale. Natives of uncertain origin: Pseudomyrmex pallidus, Camponotus (Colobopsis) sp. Introduced by man: Tetramorium guineense, Monomorium floricola, M. pharaonis. Both Terricolous and Arboricolous Floridian : Camponotus abdominalis floridanus. Antillean : Camponotus planatus. Introduced by man : Monomorium destructor. Table II Ecological Equivalents (with Zoogeographic Origins) Florida Keys Bimini Islands Odontomachus ruginodis Odontomachus insularis Pheidole floridana (Floridian) Pheidole flavens (Antillean) Solenopsis geminata (Neotropical) Pheidole megacephala (introduced) Crematogaster ashmeadi (Floridian) Crematogaster steinheili (Antillean) Cyphomyrmex minutus (Antillean) Trachymyrmex jamaicensis (Antillean) Camponotus bermudezi (Antillean) ^ (Camponotus floridanus (Antillean) C. lucayanus (endemic) U C. tortuganus (Antillean?) C. ramulorum (Antillean) J [c. planatus (Antillean) C. (Colobopsis) (mlmicola (endemic?) C. (Colobopsis) sp. (endemic?) 14 BREVIORA No. 210 EEFEEENCES Brown, W. L. 1959. The Neotropical species of the ant genus Strumigenys Fr. Smith: group of gundlachi (Roger). Psyche, 66: 37-52. Craighead, F. C. and V. C. Gilbert 1962. The effects of hurricane Donna on the vegetation of southern Florida. Quart. J. Florida Acad. Sci., 25 : 1-28. CREIGHTOlSr, W. S. 1930. The New World species of the genus Solenopsis (Hymenop. Formieidae). Proc. Amer. Acad. Arts Sci., 66: 39-151. 1950. The ants of North America. Bull. Mus. Comp. Zool., 104 : 1-585. 1957. A studj' of the genus Xenomyrmex (Hymenoptera, Formieidae). Amer. Mus. Novit., no. 1843, 14 pp. Davis, J. H. 1943. The natural features of southern Florida, especially the vegeta- tion, and the Everglades. Bull. Florida Geol. Sur., (Dept. Conservation), 25: 1-311. DuELLMAN, W. E. and A. Schwartz 1958. Amphibians and reptiles of southern Florida. Bull. Florida State Mus., Biol. Sci., 3 : 181-324. Kempf, W. W. 1958. New studies on the ant tribe Cephalotini (Hym. Formieidae). Studia Ent., 1 : 1-68. MacArthur, R. H. and E. O. Wilson 1963. An equilibrium theory of insular zoogeography. Evolution, 17: 373-387. Smith, M. R. 1954. Ants of the Bimini Island Group, Bahamas, British West Indies (Hjnnenoptera, Formieidae). Amer. Mus. Novit., no. 1671, 16 pp. Van Pelt, A. F. 1956. The ecology of the ants of the Welaka Reserve, Florida (Hy- menoptera: Formieidae). Amer. Midi. Nat., 56: 358-387. 1958. The ecology of the ants of the Welaka Reserve, Florida (Hy- menoptera: Formieidae). Part II. Annotated list. Amer. Midi. Nat.,59: 1-57. Wheeler, W. M. 1932. A list of the ants of Florida with descriptions of new forms. J. New York Ent. Soc, 40: 1-17. BREVIORA Miasenainii of Comparsitive Zoology Cambridge, Mass. December 21, 1964 Number 211 A NEW SPECIES OF THE SNAKE LEPTOTYPHLOPS FROM COLOMBIA By Benjamin Shreve In a shipment of snakes sent to me for identificaton by Hermano Niceforo Maria of the Instituto de La Salle of Bogota, Colombia, one appears to represent an undescribed species. An- other specimen, in the Museum of Comparative Zoology collec- tion, is apparently of the same species. Thanks are due the following individuals for lending compara- tive material from the collections in their charge : Mr. Charles M. Bogert of the American Museum of Natural Histoiy (AMNH), Dr. Doris M. Cochran of the United States National Museum (USNM), Miss Alice G. C. Grandison of the British Museum (Natural History) (BMNH), and Dr. William Duell- man of the University of Kansas Museum of Natural History (UKMNH). The new snake will be given a name suggested by the exceed- ingly low dorsal scale count. Leptotyphlops brevissima sp. nov. Holotype. Museo del Instituto de La Salle No. 1311, from Florencia, Caqueta, Colombia, collected by Niceforo Maria Feb- ruary 10, 1951. Paratype. MCZ No. 38950 (taken from the stomach of Micniriis tnipartitus (Dumeril, Dumeril and Bibron), MCZ No. 21988, from Sonson, Antioquia, Colombia, collected by Niceforo Maria in 1925. This specimen has about a third of the head miss- ing on the right side. Diagnosis. Apparently most closely related to Leptotyphlops macrolepis (Peters) and L. anthracina Bailey, from which it differs in a lower dorsal and subcaudal count (Table 1), and in 2 BREVIORA No. 211 coloration (brown rather than black as in anfhracina and with- out the prominent white borders of the ventral scales of macrolepis) . Also related to L. dugandi Dunn to which it is nearest in longitudinal dorsal count (but without overlap, Table 1). L. dugandi, however, has still fewer subeaudals and a very different coloration : a whitish unmarked venter, a striped dorsal pattern, and a whitish fore part of the head. Description. (Paratype variation in parentheses.) Snout rounded ; supraocular present, small ; rostral extending nearly to level of eyes ; nasal completely divided into two ; ocular border- ing the lip ; three labials, two in front of, and one behind the ocular ; second labial not reaching the eye ; six lower labials ; 14 scales around body at midbody ; dorsal scales from rostral to tail spine 152 (164) ; subeaudals is (14). Coloration in alcohol. Above, dark brown ; below, light brown ; scales, both above and below, narrowly bordered with whitish. The coloration resembles quite closely that of L. goudotii (Dumeril and Bibron). Measurements (in millimeters) head and body tail holotype 63 5 paratype 127 11 Remarks. The paratype was formerly identified as L. tnacro- lepis by the late Karl P. Schmidt, who found it during his work on the coral snakes. It may be that other specimens of hrevissima have been similarly misidentified ; longitudinal counts have been made all too infrequently. The dorsal coloration of the two species is quite similar. The ventral coloration of macrolepis gives the impression of a whitish ground color covered with brown spots, as each ventral scale has a brown spot in its center with much of the ground color visible around it. In contrast, the ventral coloration of hrevissima appears to be a uniform light brown, but on close examination it is seen that each scale has a very narrow light border. The coloration of anthracina appears to be even more similar to that of the new form. The pattern of spotting on each scale seems identical, but the brown coloration of hrevissima is repre- sented by black in anfhracina (or gray in some preserved ani- mals). 1964 NEW SNAKE FROM COLOMBIA 3 In order to shed more liglit on the relationships, material re- ferred to macrolepis was borrowed from various institutions. The following macrolepis were examined : V enezuela. Caracas : rSNM 62206, 107891. Chama, 200 meters: BMNH 1905.5.31.63. El Valle, Federal District: AMNH 59406. Merida, 1600 meters: BMNH 1903.4.28.12, 1904.6.30.2-5. Pauji, Acosta District : MCZ 48752-3, 48918. Puerto Cabello : BMNH 1913.9.10.2. No definite locality: BMNH R.R.1964.33. Colombia . In swampy country of the Choco: MCZ 39705. New Grenada, Amazonas: AMNH 17538 (in very poor condition). Brazil. Maues, Amazonas: AMNH 91642. British Guiana. Rupununi District, north of Acarahy Mts., west of New River : UKMNH 69821. E. R. Dunn (1944, p. 52) lists specimens of macrolepis from the following localities of Colombia : Norte de Santander : Ocana (1200 meters), two; Santander: Barichara (1336 meters), three; San Gil (1095 meters), four; Tolima: Mariquita (535 meters), one; Guamo (402 meters), one; Chaparral (800 meters), two. It can be seen from this list and the material mentioned pre- viously that there is no absolute evidence of sj'mpatry between the new form and what is regarded as macrolepis. In addition, there is no positive evidence that all of Dunn's material was properly identified ; he only reports the longitudinal dorsal scale count of one example and this is, of course, a critical point in separating the species. Unfortunately, all the Colombian ma- terial examined by Dunn was consumed by fire some years ago at the Museo del Instituto de La Salle. The dorsal and subcaudal counts of the specimens of macro- lepis examined, and those taken from the literature, are 202-246 and 16-26, respectively. Those taken from the literature are those of J. A. Roze (1952, p. 153), E. R. Dunn (1944, p. 52) and J. R. Bailey (1946, pp. 3, 4). It may be that with such a wide range of extremes in these counts, more than one form is in- volved. At all events, the dorsal scale counts are considerably higher than those of hrevissima (152-164). Among the specimens received on loan as macrolepis was AMNH 35953 from Balzapamba, Ecuador. This has 183 dorsal scales from rostral to tail spine, 16 subcaudals, the usual 14 scales around midbody, and a head and body measurement of 113 mm, tail 12 mm. In these characters and in coloration the specimen matches anthracina, the only member of this complex known from Ecuador ; as anthracina, it represents the third known locality for this species. BREVIORA No. 211 L. anthracina with scale counts (J. R. Bailey, 1946, p. 3) of 182-189 dorsals and 15-19 subcaudals is closely related indeed to hrevissima. It may be that anthracina and hrevissima are only subspecifieally distinct. If so, it may be that the very low dorsal counts of hrevissima may represent "character displacement" in northern anthracina populations, sympatric with macrolepis. However, the difference in coloration between anthracina and hrevissima or macrolepis is not readily explainable on this hy- pothesis. It would be expected that anthracina, outside the range of macrolepis, would be as similar in coloration as it is in scales to macrolepis, instead of exhibiting a coloration which is more different from that of rnacrolepis than is that of hrevissima. KEFEEENCES CITED Bailey, Joseph K. 1946. LeptotypMops anthracinus, a new blind snake from eastern Ecuador. Occ. Pap. Mus. Zool. Univ. Mich., No. 492 : 1-5. Dunn, Emmett Eeid 1944. A review of the Colombian snakes of the families Typhlopidae and Leptotyphlopidae. Caldasia, 3 (11): 47-55. EozE, Janis a. 1952. Contribucion al conocimiento de los ofidios de las familias Typhlopidae y Leptotyphlopidae en Venezuela. Mem. Soc. Cien. Nat. La Salle, 12 (32) : 143-158. head color: TABLE 1 hrevissima anthracina macrolepis dugandi dorsal longitudinal scale count: 152,164 182-189 202-246 172-184 subcaudal scale count : 13,14 15-19 16-26 9-12 dorsal color: dark brown lilack brown striped ventral color: light brown Idack spotted (scales witli wide white borders) whitish dark brown black brown whitish BREVIORA MiaseMm of Comparative Zoology Cambridge, Mass. Dkckmbkk ."^0, 19G4 Xumbek 212 A NEW SPECIES OP^ FRESHWATER GASTROPOD MOLLUSC OF THE GENUS SAULEA FROM THE MIOCENE OF KENYA Bv Thomas Pain^ and Dawn Beatty The genus Saulca Gray 1867 (family AmpuUariidae) has a thill, wholly corneous operculum, as found in the American genus Pomacca, and in this respect differs widely from the other African x\mpullariidae, all of which have a calcareous oper- culum. The corneous opercula are unfortunately not known in a fossil state. The only recent species of the genus Saiiha is *S'. vitrea {Born)=HeIi.r vitrea Born (1780, p. 383, pi. 15, figs. 15-16), originally described from an unknown locality, but now known to occur in the West African coastal regions of Liberia and Sierra Leone. The anatomy is at present unknown. The possibility of the occurrence of Saiilea in East Africa was first suggested by V. E. Fuchs (1936, p. 100) who con- sidered a poorly preserved fossil from the Kaiso (Pleistocene) beds of Lake Albert as possibly referable to this genus. Adam (1957, p. 48), however, pointed out that this strongly carinated shell was more probably a Vivipa)-us, perhaps a form of his own species V. coxi from the same deposit, and he later (1959, p. 15), correctly in our o]5iiiion, listed it in the synonymy of that species. In 1962, Dr. B. A'erdcourt very kindly sent to the senior author for examination a fossil shell, in the form of an internal cast in limestone, from Miocene deposits at Kirimon, east of Lake Bariiigo, Kenya, which, in spite of its poor state of preser- vation, he had no hesitation in referring to the family Ampul- lariidae (Pain, in Verdcourt, 1963, p. 7, tig. 10). This species is dextral in form, with a short, conical spire and somewhat barrel-shaped body whorl, features typical of the genus Saulea. 1 47 Keyuolcls House, Millbank, Liunldu SW 1. 2 BREVIORA No. 212 It was not, however, sufficiently well preserved to make a description of it possible. Recently we have, through the kindness of Prof. Bryan Pat- terson of the Museum of Comparative Zoology, Harvard Uni- versity, received for study two further examples of this fossil collected at Kirimon by the Harvard University 1963 Kenya Expedition. These are both in a far better state of preserva- tion than the original specimen, and have convinced us that all represent a so far unknown fossil species of Saulea, which is described and figured herein. It is to be hoped that future col- lecting at Kirimon will produce further and perhaps even better preserved specimens of this unique fossil, which has established beyond reasonable doubt the existence of this genus in East Africa in Miocene times, and its probably far wider distribution during the Tertiary period. Saulea Gray Sduica Gia.v, 1867, ji. lOllU. Type species l).v moiiutypy: Hrli.r ritrrn Born. Saulea lithoides n. sp. Holotijpc. MCZ 28018, internal cast. Pamtypc. MCZ 28017, internal cast. Horizon and locality. Late( ?) Miocene, at Kirimon. approxi- mately 42 miles east of Lake Baringo and 38 miles SSE of Mara- lal, Kenya, East Africa. Diagnosis and description. An internal cast in limestone; 51/4 whorls are preserved, the apex being decollate ; the spire short, somewhat convex, the body-whorl oval or barrel-shaped in outline and a little inflated at tbe periphery, % the total length ; umbilicus very narrow, columella erect, curving slightly toward the base. Aperture of moderate width, vertical, oblong. The outer lip and base of tlic columella are missing in all the ma- terial examined. Measurements fin millimetres: approx. only) Greatest -ciiKttl width u 32 MCZ 28017 Paratyi)e 39 27 MCZ 28018 Ilolotype 34 :{() Coryndmi .Museum Topotype 1!)(U NKW MIOCENE GASTROPOD FROM KENYA 3 The authors" tlianks arc tluc to l*i'ot'. Bryan Patterson for the loan of material for study, to Mr. T. E. Crowley for much helpful adviee and to Mr. 1\. F. Cumberland for i)lioto»>rai)hindo)ii mi. S Revnoldsville, Seminole Co. 4' BREVIORA Xo. 21;^ OCHLOCKONEE KIVER SYSTEM Ochlockonee River Drainage. — Florida : Little River. 7 mi. SW Havana, Gadsden Co. Oehloi-konee River, 8 mi. \V Talla- hassee ; SE shore. Lake Talqiiin ; both Leon Co. Ochlockonee River, 71 o mi. E Hosford. Liberty Co. WAKT7.LA RIVER SYSTEM Florida: Wakidla River, In mi. S Tallahassee, \Vakulla Co. ST. MARKS RIVER SYSTEM Florida: St. Marks River, Natural Bi-idfre Sink, Leon Co. ST^WAXXEE RIVER SYSTEM Suwannee River Drainage. — Florida: Suwannee River, Old- town, Dixie Co. Suwannee River, Fannin Springs, (lilchrist Co. WITHLACOOCHEE RIVER SYSTEM Withlacoochee River Drainage. — Florida: Withlaeoochee River, 1 mi. N\V Lacooclice, Pasco Co. Little Withlacoochee River, Rerdell, Hernando Co. X shore, Lee's Lake, Panasott'kee ; Lake, 6 mi. XXW Panasolfkee ; both Sumter Co. SW shore. Lake Tsala Apopka, Floral City, Citrus Co. XW shore. Lake Tsala Apopka, Hernando Co. HILLSBOROT^GH RIVER SYSTEM Florida: Blackwater Creek. 8 mi. N Plant City; Hillsborough River. 4 mi. XE Tem])h' Ten-ace; IxdJi Hillsboroug'h Co. I l!)(ir) Xi:\V SI'KCIKS OF AXonoNTA I'lMtC 1 The known distiiliution of Annthinta pri/c/yae, n. sp. in tlu' (uilf dijiinage of Florida is represented by a (dosed circle. The known distribution of Anodonta imbecilis Say in the southeastern United States is represented liy a partially filled circle. The known distribution of Anodonta coiipeiiana Lea is represented by an open circle. o 1-^ c; BREVIORA No. 218 d -S -^ ci ^ ^ i:^ S S S 3 M ^ -5 £ S ■-H fe & e - M ^ lO "^ '^ £ S « -J^ (^J CM - ■ - JS -= 'x CI CO 00 O oj ,-, ^ N CO CI - -j:: -c O ^ ^ ^_ S = ^ ^. i ."I 6 S' ^ ^ '^ -2 ^ ^ B r. O ±^ ^ w ^ ^ ^S^^ .^ !2 S k) ^ ^ CO ^ ,-4 "* '^ '^ S "So £ o r:! ^ ^ ^ _ S t^ O S ."^ i^ ~ »; ^ ^ . f^ o -3 "-- . • g CO CO rH ^ ^ J l^ '-'^ K S ^- ^- e ^ 5' poi «' .5 c 1 - ' ^ M -^ — M fat _fcJ5 -^ _faij 5 &i S Mo U 'i 9, M ^ Ss liJ" S" St Sc.s S.^ K 196.3 XKW SI'KCIKS OK AXOnoXTA BREVIORA Miiseiim of Comtiparative Zoology Cambridge, Mass. February 15, 1965 Number 214 A REVISED CLASSIFICATION OF THE DENDROCHIROTE HOLOTHURIANS By David L. Pawson U.S. National Museum and H. Barraclough Fell Museum of Comparative Zoology Recent studies on some unusual dendrochirote liolothurians and examination of available data on fossil liolothurians have led us to conclude that the hitherto accepted classification of the dendrochirote holothurians conceals some important evolutionary trends. In this article a new classification is proposed in which tentacle numbers are abandoned as criteria for diagnosing major taxa; instead, importance is attached to the shape of the cal- careous ring, the shape of the tentacles, and the calcareous de- posits of the body wall. The reasons for such changes in classifica- tion are given in detail elsewhere (Fell and Pawson, in press). NEW TAXA Certain genera which were hitherto grouped in the order Dendrochirotida are known to lack the richly branched tentacles which are so characteristic of other genera within that order. These genera with essentially unbranched tentacles also share other important features, which suggest a natural grouping of three families. We therefore propose a new order to accommodate these families. Order DACTYLOCHIROTIDA nov. Diagnosis: Tentacles 8-30 in number, not branched but digit- iform or digitate, the digits sometimes bifurcate. Body enclosed 2 BEEVIORA No. 214 in a test comprising imbricate plates. Calcareous ring simple, lacking complex posterior processes. Included families: Ypsilothuriidae Heding, 1942; Rhopalod- inidae Perrier, 1902 ; and a third family, diagnosed here as new : Family VANEYELLIDAE nov. Diagnosis: Anus and mouth at opposite ends of the body, which is U-shaped or fusiform. Plates in the body wall with or without small spires. Tentacles 10-20 in number. Included extant genera : Vaneyella Heding and Panning, 1954 (type genus by tautonymy) ; Mitsukuriella Heding and Panning, 1954. The order Dendrochirotida as now restricted may be diagnosed as follows : Order DENDROCHIROTIDA Grube, 1840 (restricted herein) Diagnosis: Tentacles 10-30 in number, richly branched. Cal- careous ring simple or with complex posterior processes. Test sometimes well developed, or reduced to non-contiguous calcare- ous spicules. Morphological and paleontological evidence (Fell and Pawson, in press) suggests that this order includes some very primitive extant holothurian genera. These genera seem best assigned to new families ; one new genus and two new families are diagnosed here : Family PLACOTHURIIDAE nov. Diagnosis: Body completely enclosed in a test of imbricate plates without spires. Sole lacking. Calcareous ring long, slender, with long posterior processes. Included extant genus : Placothitria n.g. Placothuria n.g. Diagnosis: Tentacles 10, richly branched. Body U-shaped. Cal- careous ring long, with posterior processes composed of a mosaic of small pieces. Type-species: Stolus Juittoni (Dendy) (illustrated in Pawson, 1963, pi. 7). 1965 NEW HOLOTHURIAN CLASSIFICATION 3 Etymology: The generic name is derived from the Greek plax (a plate) and Holothuria, and refers to the test formed by over- lapping plates. Gender, feminine. A second new family within the order Dendrochirotida is pro- posed : Family PARACUCUMIDAE nov. Diagnosis: Body completely invested in a test of imbricate scales with spires. Sole lacking. Calcareous ring simple, lacking posterior processes. Included extant genus : Paracuciimis Mortensen, 1925. A REVISED CLASSIFICATION The relationships of the orders Dendrochirotida and Dactylo- chirotida are discussed elsewhere, in the broad context of the inferred evolution of the Holothuroidea (Fell and Pawson, in press). From the new data the following revised classification of holothurians emerges : Subclass DENDROCHIROTACEA Grube, 1840 (Nomen translatum herein, ex Dendrochirotae Grube, 1840) Introvert (with retractor muscles) always present. Tubefeet and respiratory trees usually present. Madreporite free in the body cavity. Mesentery of the posterior loop of the intestine in the right or left ventral interradius. Free tentacle ampullae lack- ing. Gonad in two tufts, one tuft to each side of the dorsal mesen- tery. Order DENDROCHIROTIDA Grube, 1840 (restricted herein) Tentacles richly branched, 10-30 in number. Key to Included Families 1 (6) Body partly or completely invested by plates. 2 (5) Body enclosed by a test comprising conspicuous imbricate plates; sole lacking, 3 (4) Calcareous ring complex, with long paired posterior processes. Placothuriidae nov. (Included ex- tant genus: Placoihuria nov.) 4 BREVIOBA No. 214 4 (3) Calcareous ring simple, lacking posterior processes Paracucumidae nov. (Included ex- tant genus: Paracucumis Mortensen, 1925) 5 (2) Body invested dor sally by conspicuous plates; sole present Psolidae Perrier, 1902 (Included extant genera: Psolus Oken, 1815; StoUnus Selenka, 1868; Psolidium Ludwig, 1886; Lepidopsolus Bronn, 1860; Thyonepsolus Clark, 1901) 6 (1) Body more or less naked, not enclosed by a test; calcareous de- posits small, inconspicuous. 7 (10) Calcareous ring complex, with paired or unpaired posterior proc- esses. 8 (9) Processes composed of a mosaic of small pieces Phyllophoridae Ostergren, 1907 (emend, herein). (Included subfamilies: Phyllophorinae Ostergren, 1907; Sem- periellinae Heding and Panning, 1954; Thyoninae Panning, 1949). 9 (8) Processes entire Sclerodactylidae Panning, 1949 (Nomen translatum herein, ex Sclerodactylinae Panning, 1949). (Included subfamilies : Sclerodactylinae Panning, 1949 ; Cladolabinae Heding and Panning, 1954). 10 (7) Calcareous ring simple, lacking posterior processes Cucumariidae Ludwig, 1894 (emend. herein). (Included subfamilies: Cucumariinae Ludwig, 1894; Col- ochirinae Panning, 1949 [partim?] ; Thyonidiinae Heding and Panning, 1954). Order DACTYLOCHIROTIDA nov. Tentacles digitiforiii or digitate, the digits sometimes bifurcate, 8-30 in number. Body enclosed by a test comprising imbricate plates. Calcareous ring simple, lacking complex posterior pro- cesses. Key to Included Families 1 (4) Anus and mouth at opposite ends of the body, which is fusiform or U-shaped. 2 (3) Plates with a prominent spire. Tentacles 8-10 in number Ypsilothuriidae Heding, 1942 (Included extant genera: Tpsilotlmria Perrier, 1886; Echinocucumis Sars, 1895; Ypsilocucumis Panning, 1949) 3 (2) Plates with small spires, or none. Tentacles 10-20 in number. Vaneyellidae nov. (Included extant genera: Vaneyclla Heding and Panning, 1954; MitsuJcuriella Hed- ing and Panning, 1954) 4 (1) Anus and mouth opening close together, body flask-shaped Rhopalodinidae Perrier, 1902 (Included 1965 NEW HOLOTHURIAN CLASSIFICATION 5 extant genera: Bhopalodina Gray, 1853; Rhopalodinopsis Heding, 1937). The above arrangement of taxa is believed to reflect phylo- genetic relationships more satisfactorily than that currently em- ployed. Of the fossil holothurian families defined by Frizzell and Exline (1955), the Calclamnidae (including genus Thuroholia) and some members of the Stichopitidae and Priscopedatidae should be included within this subclass. OTHER HOLOTHURIAN GROUPS In view of the above proposed revision of the Dendrochiro- tida, it is desirable to consider briefly the arrangement of the remaining holothurian orders, the Aspidochirotida, Elasipodida, Apodida and Molpadida. Aspidochirotida and Elasipodida Both the aspidochirotes and elasipods have tentacles which terminate in an approximately circular disc. The body is usually bilaterally symmetrical in both groups, with the dorsal tubefeet modified into papillae or warts (Aspidochirotida) or into elongate sensory processes (Elasipodida). The two groups are distinguish- able on the basis of some anatomical features, but both may be conveniently placed into a single subclass, as follows : Subclass ASPIDOCHIROTACEA Grube, 1840 (Nomen translatum herein, ex Aspidochirotae Grube, 1840) Diagnosis: Tubefeet present, tentacles shield-shaped, 10-30 in number. No introvert, hence retractor muscles lacking. Body with conspicuous bilateral symmetry. Key to Included Orders 1 (2) Respiratory trees present. Mesentery of posterior loop of intestine attached to right ventral interradius Aspidochirotida Grube, 1840 2 (1) Respiratory trees lacking. Mesentery of posterior loop of intestine attached to right dorsal interradius Elasipodida Theel, 1882 6 BREVIORA No. 214 Theel (1882) has discussed the possible antiquity of the elasi- pod holothurians and concluded that they are certainly not rep- resentative of an ancestral holothurian stock; rather they are secondarily adapted to deep-sea life. In elasipods, and in some aspidochirotes, the madreporite opens to the exterior, and does not hang free in the body cavity. This can be regarded, not as a primitive feature, but as a logical compensatory consequence of the absence of respiratory trees. The extremely fragile calcareous ring in elasipods is apparently secondarily reduced. Within this subclass should be placed some representatives of the fossil families Priscopedatidae and Theelidae as defined by Frizzell and Exline (1955). APOOroA AND MOLPADroA The important character shared by adopids and molpadids is the almost complete absence of tubefeet. Also both groups have simple digitate or pinnate tentacles. It is possible that the apodids and molpadids bear no close relation to each other, and the characters they have in common may have arisen through parallel evolution and convergence. Subclass APODACEA Brandt, 1835 (Nomen translatum herein, ex Apodes Brandt, 1835) Diagnosis: Tentacles simple, digitate or pinnate. Tubefeet markedly reduced, or, more usually, lacking altogether. No intro- vert, hence retractor muscles lacking. Deposits may include anchors and anchor plates. Key to Included Orders 1 (2) Body cylindrical. Eespiratory trees and anal papillae lacking. Deposits often include wheels Apodida Brandt, 1835 2 (1) Body fusiform, often with tapering caudal portion. Eespiratory trees and anal papillae present. Wheels lacking Molpadida Haeckel, 1896 It is remarkable that some members of both Apodida and Molpadida have anchors and anchor plates in the body wall. The anchors and their plates differ morphologically in the two orders, but presumably have the same functional significance as accessory locomotor organs, since the anchors project through the body wall. 1965 NEW IIOLOTIIURIAN CLASSIFICATION 7 The fossil family Achistridae and some members of the fossil families Stichopitidae, Theelidae, Synaptitidae and Calcanoridae as defined by Frizzell and Exline (1955) may be included in this subclass. LITERATURE CITED Fell, H. B. 1965. The early evolution of the Echinozoa. Breviora (in press). Fell, H. B. and D. L. Pawson 1965. Chapters on the Echinozoa; and evolution and phylogeny of the Holothuroidea. In Treatise on Invertebrate Paleontology, U, Echinozoa and Asterozoa. (In press.) Frjzzbll, D. L. and H. Exline 1955. Monograph of fossil holothurian sclerites. Bull. Univ. Missouri Mines Met., 89:1-204, 11 pis., 21 text figs. Pawson, D. L. 1963. The holothurian fauna of Cook Strait, New Zealand. Zool. Pubis. Victoria Univ. Wellington, 36:1-38, pis. 1-7. Theel, Hj. 1882. Report on the Holothurioidea. I. Challenger Sci. Results: Zool- ogy, 4(13):1-176, pis. 1-46. BREVIORA Museimi of Comniparsitive Zoology Cambridge, Mass. February 15, 1965 Number 215 TWO NEW SUBSPECIES OF AMPHISBAENA (AMPHIS- BAENIA, REPTILIA) FROM THE BARAHONA PENIN- SULA OF HISPANIOLA By RicHAED Thomas^ As a result of collecting sponsored and led by Dr. Albert Schwartz during the summers of 1963 and 1964 in which I was fortunate to be able to participate, 33 specimens of Amphishaena were collected in the low areas on and near the Barahona Penin- sula of the Dominican Republic. The general affinities of these specimens are with Amphishacna innocens Weinland. Compari- son of the Barahona Peninsula specimens with representatives of the forms assigned to the species innocens shows them to be distinct in themselves, but to resemble most closely A. i. gona- vensis Gans and Alexander. These authors, in their recent study of West Indian amphisbaenids (1962), examined the available Hispaniolan specimens and reviewed the forms. The combina- tion AmpJiisbaena innocens caudalis Cochran was first used by them, and the new name Amphishaena i7inoce7is gonavensis was proposed for the population on Gonave Island. Fourteen addi- tional specimens of gonavensis, recently acquired by Dr. Ernest Williams, allow a better comparison of the Gonave and Bara- hona populations than was possible previously. Although the differences between the two populations are perhaps on a level of specific separation, their relationship with one another is ob- vious. I feel that relationships are best expressed by regarding these two as conspecific and distinct from A. innocens. These related amphisbaenids of Gonave Island and the Barahona Peninsula should therefore be known as Amphishacna gonaven- sis Gans and Alexander. In lacking major fusion of the head scales, Amphishaena gonavensis is separable from all other West Indian Amphis- haena except innocens Weinland, caeca Cuvier, hakeri Stejneger 1 10,000 S. W. 84th street, Miami, Florida .33143 BREVIORA No. 215 and fcnestrata Cope. From fcnestrata and bakeri it is imme- diately separable on the basis of lower number of body annuli, 199-225 for gonavensis versus 236-249 for fcnestrata and 239- 254 for hakeri; from fcnestrata it is additionally different in lacking the backward penetration of the rostral between the D Fig. 1. Ventral views of the heads of four forms of Amphishaena. The malar scales are stippled for points of reference. A, A. g. gonavensis (MCZ 80296), note three rows of postgenials and anterior penetration of first postgenials between genial and second infralabial; B, A. g. leheri (ASFS V2596), note two rows of postgenials and lack of anterior penetration of first postgenials (position of first two postgenials in hyporissor is inter- mediate between A and B) ; C, A. innocens (ASFS X3114) ; D, A. caeca (ASFS X7382). nasals. From A. caeca, A. gonavensis may be distinguished in the possession of two scales in the first and second rows of postgenials (versus three for caeca). Gonavensis also differs from caeca in having enlarged precloacal scales forming a dis- tinct convex flap as opposed to the relatively undifferentiated precloacals of caeca, and in heraipenial structure (see discus- sion). 1965 HISPANIOLAN AMPHISBAENA 3 Employing the concept of accessory dorsal half -annul! in the nuchal region, as expressed by Gans and Alexander (1962: 79- 80), the first two body annuli in innocens, gonavensis and caeca typically correspond to three dorsal half -annuli (Fig. 2). Gona- vensis (and innocens, typically) has the first complete annulus including (continuous with) the anteriormost of the first three dorsal half -annuli (temporals, postoculars, frontals) (Fig. 2, B-D). Exceptionally, the condition depicted in Figure 2 A occurs in innocens. Examination of 49 specimens of A. caeca shows the anteriormost half-annulus to appear split off from, in- stead of continuous with, the first complete body annulus (i.e. first ventral half-annulus) (Fig. 2, E-F) . A. gonavensis differs from its neighbor species A. innocens in the possession of a malar scale ^ (related to this are differences in shape and arrangement of the ventral head scales. Fig. 1), and in the condition of the first pair of parietal scales which nor- mally are in contact apically (thus only narrowly) with one another at the midline or may even be slightly separated from contact by the second pair of parietals (Fig. 2, C-D). In inno- cens by contrast there are typically two or four large squarish parietals (there may be other small supernumerary scales pres- ent, but these do not alter the basic appearance) formed by the median scales of the second and third dorsal half -annuli (Fig. 2, A, B). Gonavensis is additionally distinguished from innocens in the shape of the tail which is rounded rather than tapered (Fig. 3, A, B), in the precloacal scales which form a more prominently convex flap, and in hemipenial structure (see be- low). A. gonavensis is broader and shorter headed, has a less sharply pointed snout, and is distinctly smaller than inno- cens. The largest of 55 specimens of gonavensis measures 242 mm ; of 73 innocens 17 measure more than 242 mm, the 1 The malar scale as defined by Gans and Alexander is a major and obvious diflference between those Amphisbaena which possess it and those which do not. With respect to the species discussed here, the malar, possessed by A. gonavensis, seems at least in part homologous to the two end scales of the second row of postgenials of A. innocens. The advisability of denoting this scale with a sepa- rate name seems somewhat dubious inasmuch as it implies an all-or-nothing diflference where there is really one of degree. Further, the designation of the row of scales behind the malar (the row just in front of the first body annulus) as postmalars in contrast to this row being merely the third row of postgenials in forms having no malar seems also inadvisable and misleading. The term "malar," however, is a very convenient one in description, and its usage is re- tained with the above comments duly noted. The postgenials are here used to apply to all rows of small scales hetween the genial and the first body annulus, including the "postmalar row" used by Gans and Alexander. BREVIORA No. 215 largest being 279 mm. Gans and Alexander noted that a single specimen from Petite Gonave (MCZ 25549) was at that end of the range of values (Gonave population) of several characters which was closest to the Gul-de-Sac sample of innocens. On the basis of the new material, this specimen falls within the range of the Gonave population in those characters for which the trend was most noticeable (body annuli, dorsal segments, postcloacals). Numerical data used in above comparisons and in ' ' Comparisons Fig. 2. Diagrammatic illustrations of Amphisbaena lieads (dorsal view, half -heads) showing sealation differences. First two body annuli are shown "peeled out" from heads. A, atypical A. innocens (Camp Perrin) ; B, typical A. innocens (Camp Perrin) ; C and D, alternative common condi- tions of A. gonavensis ; E and F, A. caeca. Stippled scales are second dorsal half -annuli (complete annulus in F). In A-D the second half -annuli are accessory • in E and F the first are "accessory." 1965 HISPANIOLAN AMPHISBAENA 5 and Discussion" below are in i)art obtained from Gans and Alexander's paper and in part from personal observation (see ' ' Specimens Examined ") . Amphisbaena gonavensis hyporissor new subspecies Holotype: MCZ 77149, an adult male, collected 13.1 mi. (20.9 km) SW of Enriquillo, Pedernales Province, Republica Domini- cana, 30 July 1963, by David C. Leber and Richard Thomas. Paratypes: KU 79824-25, same locality as type, 22 July 1963, Richard Thomas ; ASFS X9974-76, AMNH 92792-95, RT 754-55, UIMNH 55600-02, same data as type; MCZ 77150, Republica Dominieana, Pedernales Province, 5 mi. (8 km) NE Oviedo, 30 July 1963, Richard Thomas. Diagnosis: A subspecies of Amphishaena gonavensis charac- terized by lack of fusions of head scales, a high number of caudal annuli, more than six precloacals ; an occasional caudal autotomy; and a mottled but faded coloration. Range: Presently known from the southeastern portion of the Barahona Peninsula of Hispaniola. Description of type (Fig. 4) : (Methods of counting and ter- minology follow Gans and Alexander, except for differences in terminology already noted). Head scales not fused; prefrontals as broad or slightly broader anteriorly than posteriorly. First two body annuli correspond to three dorsal half-annuli. First pair of parietals border anteromedially on frontals and are oc- cluded from apical contact by an anterior extension of the left second parietal. Genial .81 times as broad as long. Three rows of postgenials present; first row with two enlarged scales (an abnormal, minute middle scale is present), apex of each scale projecting slightly beyond malar-second infralabial suture be- tween genial and second infralabial ; second row with three scales not including the two large triangular malars at each end ; third row with four scales not including the two enlarged, ter- minal " postmalars. " Body annuli 213, four laterals on each side, and 20 caudal annuli. Sixteen dorsal segments and 24 ventral segments to an annulus counted at midbody. Four cloacal pores ; 8 precloacal scales and 13 postcloacals, including 2 median, rectangular postcloacals. Snout-vent length 200 mm ; tail 20 mm. Overall coloration tan, becoming darker and slightly mottled dorsally due to increasing amount of central dark pig- mentation on more dorsal segments ; four ventralmost segment rows without dark pigment spots. BRE\nORA No. 215 Variation: The paratypes are similar to the type in the con- figuration of the first pair of parietals. Two have the first parietals in broad contact and squarish ; in six they do not meet at the midline, the second pair of parietals contacting the frontals (Fig. 2C) ; in the balance of the paratypes the first pair of parietals are in apical or only narrow contact. Consid- erable asymmetry occurs in the degree of contact of the first pair of parietals with the midline. In all specimens of hyporis- sor the two scales of the first row of postgenials project not at all or only slightly beyond the malar-second infralabial suture and thus between the genial and second infralabial (Fig. 1). All paratypes have three rows of postgenials, the modal count being 2 + 3 + 4 but with variants of four and five in the second row and five in the third row. Body annuli vary from 199 to 221; laterals two to five (mode four), most permutations of these counts occurring except 2/2, 5/5, 4/5, and 5/2 ; caudal Fig. 3. Ventral views of the caudal aud cloacal region of four forms. A, A. g. gonavensis (MCZ 80291); B, A. g. leberi (ASFS V2596, except for median postcloacals, is also typical of hyporissor) ; C, A. innocens (ASFS X3111); D, A. caecu (ASFS X940). Median postcloacals of A and B are stippled ; those of A are typical of gonavensis and hyporissor ; those of B are diagnostic of leberi. 1965 HISPANIOLAN AMPHISBAENA 7 aiinuli vary from 1!) to 21 (mode 19). Dorsal segments of a midbody ammlus are 16 (mode), 17 or 18; ventral segments are 22 (mode), 23 or 24; totalled midbody segments 38-42. Pre- eloacal scales range from eight to eleven; postcloacals from 12 to 14; totalled cloacals 20-24. There are two enlarged, median postcloacals; these are typically straight-sided (Fig. 3A). Clo- acal pores are four in every specimen. The smallest specimen measures 106 mm total length (tail 10 mm), the largest 231 mm (tail 20 mm). Apparent caudal autotomy constrictions are evident between the fifth and sixth annuli of the tails of some of the juvenile specimens; adults do not possess obvious au- totomy constrictions. One medium size specimen possesses a broken tail, presumably autotomized, missing beyond the fifth annulus. The coloration is pale tan, darker dorsally and with some scattered dark mottling ; some specimens were noted as be- ing purplish dorsally in life. Specimens collected in the vicinity of the town of Pedernales in the northwestern section of the Barahona Peninsula represent a subspecies distinct from the other two known populations of this species. This form is named in honor of Mr. David C. Leber, for his enthusiastic participation in collecting in the Dominican Republic. Amphisbaena gonavensis leberi new subspecies Holotype: MCZ 77218, an adult male, collected 5 km N of Pedernales, Pedernales Province, Republica Dominicana, 25 June 1964 by Richard Thomas. Paratypes: MCZ 77219, same data as type; ASFS V2595-96, 8 km N of Pedernales, 26 June 1964, Richard Thomas; KU 79855-56, UIMNH 55627-28, ASFS V2676-78, AMNH 92827-28, Pedernales, 29 June 1964, Richard Thomas; RT 985, same lo- cality as previous series, 3 July 1964, Richard Thomas. All specimens from Pedernales Province, Republica Dominicana. Diagnosis: A subspecies of Amphishaena gonavensis most closely related to A. g. hyporissor in the possession of a high number of caudal annuli, but differing from that form in con- tact of first pair of parietals with one another, in having typi- cally two rows of postgenials, and in the configuration of the median postcloacals. Distribution: Known presently from the low elevations of the northwestern portion of the Barahona peninsular region of Hispaniola. 8 BREVIORA No. 215 Description of type: Head scalation much like that described for the type of hyporissor. First pair of parietals in broad con- tact across the midline giving a squarish appearance to the parietals. Two rows of postgenials present, two scales in the first row and five in the second. Anterior penetration of first two postgenials between the genial and second supralabial very slight. Body annuli 212 ; four laterals on each side ; 16 caudal annuli. Sixteen dorsal segments and 22 ventral segments to an annulus counted at midbody. Four cloacal pores, nine pre- cloacal scales and 14 postcloacal scales. Enlarged median post- cloacal scales four. On each side of the midline, two median postcloacals separated by a transverse suture which curves pos- teriorly and laterally giving a rounded outline to the posterior two of these four scales (Fig. 3B). Total length 223 mm, tail 19 mm. Coloration generally like that of the type of hyporissor, but the tan is more uniform with less dark mottling. Variation: Head scalation of the paratypes is much like that of the type. The first pair of parietals are in broad contact medially in all but one specimen in which the contact is slightly more than apical. First two postgenials typically have none or only slight anterior penetration between the genial and the second inf ralabial ; the characteristic condition is that illustrated in Figure 1 B. Body annuli vary from 207 to 220 ; laterals two to five with the same variation noted for hyporissor; caudal annuli range from 16 to 19 (mode 19). Many specimens have the tips of their tails scarred ; were this not so, higher caudal counts would probably be obtained. Dorsal segments of a mid- body annulus are 15 (two specimens) or 16; ventral segments 22-25 (mode 24) ; totalled midbody segments 38-41. Precloacal scales range from 9 to 11; postcloacals 11-14; total cloaeals 20-24. Median postcloacals are essentially the same as described for the type in all except one specimen in which they are very abnormal and frgamented and another in which the posterior scales reach the cloacal border between the two anterior scales which are reduced. Apparent caudal autotomy is found in one of the paratypes of leheri (UIMNH 55628) and in another speci- men from Pedernales which is currently being kept alive. In both the tail is missing beyond the fifth annulus. Coloration is generally darker and more uniform than that of hyporissor. Hemipenes are everted or partially everted in four specimens; the organs are attenuated and naked ; apparently they are only slightly bilobcd (the most completely everted organs are not 1965 HISPANIOLAN AMPHISBAENA 9 bilobed, but terminal bifurcation of the dissected retractor penis indicates that they may be slightly bilobed when completely everted). Comparisons and discussion: To facilitate comparisons of the three races of A. gonavensis the following table showing scale count data is presented. Table 1. Scale count ranges for the three races of A. go7iavensis Body annul! Cauflal annul! Total niidbody sosments Pre- cloacals Post- cloacals Total cloacals gonavensis 207-225 10-12 36-41 6-7 11-14 16-20 hyporissor 199-221 19-21 38-42 8-11 12-14 22-24 leberi 207-220 16-19 38-41 9-11 11-14 20-24 As shown by the table, gonavensis differs strikingly from the Barahona populations in the low number of caudal annuli. It further differs in scalation in the low number of precloacals (the condition of seven precloacals occurs in only one specimen), and in total cloacals. The specimens of gonavensis examined by me (18) and the type (not seen by me but illustrated by Gans and Alexander) are rather uniformly characterized by the first pair of parietals meeting in apical contact at the midline ; in but three specimens was there a short suture between the parietals. As described above, hyporissor is generally, but much more variably, characterized by narrow contact of the first pair of parietals, while in leheri the parietals with one exception are in comparatively broad contact with one another. Gonavensis and hyporissor uniformly have three rows of postgenials, while leheri is typically missing a segment in the gular region and consequently has only two rows. Of the three exceptions noted, one is abnormal in having the "missing" segment partly inter- calated between the second postgenial row and the first body annulus, a postgenial count of 2 + 4 + 2 being the result. The two scales of the first row of postgenials penetrate very far for- ward between the genial and the second infralabial in gonaven- sis (with 35 to 69 per cent of the length of the scale lying an- terior to the malar-second infralabial suture) ; in hyporissor the penetration is less (10-30 per cent, four scales have no penetra- tion) ; and in leheri even less (6-18 per cent, 11 scales have no penetration). The corresponding differences in the configura- tion of these scales can be seen in Figure 1 A, B. In the con- dition of the median postcloacals, gonavensis and hyporissor 10 BREVIORA No. 215 agree in having but two roughly rectangular, undivided scales (small marginal scales which are normally folded inside the cloaca are not considered). The condition which characterizes leheri and the one exception have already been noted. One speci- men (ASFS V2507) from approximately 10 km NW of Oviedo B Fig. 4. A and B, dorsal and ventral views of the head of the type of A. g. hyporissor (MCZ 77149) ; C and D, dorsal and ventral views of a specimen of A. innocens from Camp Perrin (ASFS X3123). on the road to Pedernales has the leheri configuration of the median postcloacals ; it also agrees more with leheri in the con- dition of the parietals, while in having three rows of postgenials it agrees more with hyporissor. More specimens are necessary from intermediate regions before the status of this specimen can be determined. The presumed caudal autotomy noted for some specimens of hyporissor and leheri has not been observed in gonavensis, nor does it occur in A. innocens (both have relatively short tails). 1965 HISPANIOLAN AMPHISBAENA 11 Gans and Alexander noted that caudal autotomy is variable in some forms, e. g. A. caeca. If the hemipenial structure noted for leheri holds true for the species as a whole, this will serve as an additional distinction from A. innocens and A. caeca. In contrast to the simple, naked structure of leheri, the hemipenis of innocens is strongly bilobed, very heavy and fleshy; the sulcus spermaticus bifurcates at the fork of the organ and each branch proceeds to the non-sulcate side and thence to the apices of the lobes which are flattened and disk-like. Proceeding from each branch of the sulcus sper- maticus over each lobe and onto the distal sulcate surface of the organ are very fine but regular flounces. The non-sulcate sur- face is naked (from ASPS X3112). The hemipenes of caeca are of the same general appearance as those of innocens, but the sulcus spermaticus is much more prominent (thick edged) and forks slightly before the bifurcation of the organ itself ; the apices are rounded not flattened, and there is no evidence of flouncing (from ASFS X937, X4111). The distribution of the races of Aniphishaena gonavensis appears zoogeographically a bit strange at first glance (see Fig. 5). The possibility exists that it is in reality not so strange. The Cul-de-Sac plain, a channel (in places below sea level) be- tween the north and south ''islands" of Hispaniola (Williams, 1961) bordered on both sides by mountainous regions, debouches to the west at the angle between the Tiburon Peninsula and the main part of the island and to the east at the northeastern corner of the Barahona Peninsula just east of the Sierra de Baoruco. A. gonavensis may occur (or have occurred in the recent geologic past) up the eastern coastal fringe of the Bara- hona (presently almost nonexistent in places) and into the Cul-de-Sac plain. This being so, the island of Gonave would not be an unlogical extension of its range. The close affinities of hyporissor with gonavensis seem to support such a distribution. It is interesting to note that at least two other Hispaniolan lizards have a similar distribution. Anolis hrevirostris Bocourt occurs along the eastern Barahona into the Cul-de-Sac, east to the Golfe de Gonave and on Gonave. Diploglossus curtissi Grant, though of wider distribution to the east and north (Schwartz, in press), also occurs along the eastern Barahona coast and is channelled to the west through the Cul-de-Sac (lo- cality records not continuous) and occurs on Gonave. 12 BREVIORA No. 215 It is interesting to note that despite the geographic continuity of the Sierra de Baoruco with the mountain mass of the Tiburon Peninsula, Amphishaena innocens is not known from the Bao- ruco. On the contrary it is A. manni^ which has been collected there. Gans and Alexander record four specimens of manni from near Paraiso at 1800 feet (about 600 m) (Fig. 5). In the summers of 1963 and 1964 we obtained four additional speci- mens (ASFS X9907-09, V2911) from the eastern end of the Sierra de Baoruco (see Fig. 5) at an elevation of 2600 feet. (790m). Although further collecting may prove otherwise, it ap- pears that manni is geographically the nearest Amphishaena to hyporissor, possibly even interposed between it and innocens. Com- parable geographic relationships are seen in the species men- tioned above as having distributions similar to the one predicated for A. gonavensis. Diploglossus ciirtissi and Anolis hrevirostris are "replaced" in the highlands of the Sierra de Baoruco by other related species (Diploglossus cost at us Cope and Anolis distichus Cope). The type and most of the paratypes of A. g. hyporissor were collected in one of those occasional but not altogether rare situa- tions where fossorial creatures such as Amphishaena and Typh- lops are found concentrated near the surface. Such localities are generally but not invariably characterized by a rather friable soil, a scattering of moderate to large sized trees, a cover- ing of plant litter, and an abundance of rocks. This particular locality is a region of xeric woods inland from the western mangrove margin of the Laguna de Oviedo. The locality gener- ally satisfied the above conditions ; an abundance of limestone rocks was present. Aside from the amphisbaenids, two species of Typhlops were found in this region, Typhlops cf. sulcatus and another species (Thomas, MS). The localities to the north of Pedernales where A. g. leheri was collected were similar to the locality just described, at least in degree of aridity. The locality at Pedernales, however, was somewhat more xeric; the tree cover was primarily Acacia with an undergrowth of scat- tered clumps of Opuntia growing among outcroppings of lime- stone rock; the soil was sandy. Typhlops haitiensis, Typhlops 1 A further note of comparison between gotiarensix and manni may be in order. A. manni, though distinct from {/onavcngis in having the nasals fused with the rostral, differs further in having the condition of the half-aniiuli in the nuchal region as described for caeca ( ride xiipra) . 6-9 (versus 4) cloacal pori's, a com- paratively stout tail with a very prominent autotomy constriction, and a deeply bilobed heniipenis. In meristic characters there are no striking differences. 1965 HISPANIOLAN AMPHISBAENA 13 sp. and Leptotyp/ilops sp. (Thomas, MS) were found in the same macrohabitat with A. g. leberi in the Pedernales region. In both of the localities where the two largest series were ob- tained, numbers of shed skins of these amphisbacnids were seen while collecting. Little can be said about the habits of these lizards from our encounter with them. They w^ere mostly col- lected under rocks; their passageways were frequently evident. One grasped and held onto a rootlet with its mouth in a pos- sible attempt to resist capture. None were found above ground in the open, and their abundance did not seem to be correlated with any weather phenomena. Specimens examined Amphisbaena gonavensis Jiyporissor : As listed for type and paratypes. Amphishaena gonavensis leberi: As listed for type and paratypes. Amphisbaena gonavensis hyporissor x leberi: Repiibliea Dominicana; Pedernales Province, ca. 10 km NW Oviedo, ASFS V2507. Amphisbaena gonavensis gonavensis: Haiti: Gonave Island: Pointe-a- Eaquette, PM 3385 (allotype), PM 3386, 3388 (paratypes), MCZ 80289; Ti Palmiste, 6 km from Pointe-a-Eaquette, MCZ 80290-302. Amphisbaena innocens innocens: Haiti: Department de 1 'Quest, Manne- ville, MCZ 62511, MCZ 8748; Thomazeau, MCZ 37595-97; Purey, MCZ 51417, ASFS X3862; Department du Sud, Camp Perriu, ASFS X3109-34, X3240-41, DEP 2403. Amphisbaena innocens caudalis: Haiti, Grande Cayemite Island, MCZ 25550 (type), MCZ 25551 (paratype). Amphisbaena caeca: Puerto Eico: Isla Verde, ASFS X937-43, X4104-25, X7381-98; 2.2 mi. SW Sabana, ASFS X7433-34. ACKNOWLEDGMENTS I wish to express my appreciation to Dr. Albert Schwartz, who allowed me to study this material gained as a result of his West Indian collecting; to Dr. Ernest E. Williams, Museum of Comparative Zoology at Harvard ; Dr. Charles A. Reed, Yale Peabody Museum (PM) ; and Mr. Dennis R. Paulson (private collection, DRP) for loan of specimens in their care. I also wish to thank Mr. Ronald E. Klinikowski for executing some of the illustrations for this paper. Types and paratypes designated herein now reside in the following collections: Museum of Comparative Zoology at Har- vard (MCZ), American Museum of Natural History (AMNH), 14 BREVIORA No. 215 University of Kansas Museum of Natural History (KU), Uni- versity of Illinois Museum of Natural History (UIMNH), Al- bert Schwartz Field Series (ASFS), Richard Thomas private collection (RT). LITEEATURE CITED Gans, Carl and A. Allan Alexander 1962. Studies on amphisbaeiiids (Amphisbaenia, Eeptilia). On the amphisbaenids of the Antilles. Bull. Mus. Comp. ZooL, Vol. 128. no. 1, pp. 65-158. Schwartz, Albert (In press.) Biploglossus cosiatus Cope (Sauria, Anguidae) and its relatives in Hispaniola. Reading Public Museum and Art Gallery, Sci. Pub. Williams, Ernest E. 1961. The evolution and relationships of the Anolis seinilineatus group. Breviora, Mus. Comp. Zool., No. 136, pp. 1-7. -18* O 20 40 ■ ■ ■ I I km 74* I Fig. 5. Map of southwestern Hispaniola showing Gonave Island, the Tiburon Peninsula and the Barahona Peninsula with localities for the spe- cies of AfnpMshaena (some localities are from Gans and Alexander, 1962). Hollow circles, A. innocens; circles with crosses, A. gonavensis ; solid circles, A. Ttmnni. BREVIORA Meseum of Comparsitive Zoology Cambridge, Mass. February 15, 1965 Number 21() THE GEOGRAPHICAL VARIATION OF THE FROG HY- PEROLIUS MAEM0EATU8 (FAMILY HYPEROLIIDAE) IX RHODESIA, NYASALAND AND TANGANYIKA By R. F. Laurent Mr. Yesey-Fitzgerald has generously provided the Museum of Comparative Zoology with several very interesting series of tree frogs collected in various localities of Rhodesia, Nyasaland and Tanganyika. Some of them belong to the marbled tree frog Hyperolius marvvoratus Rapp, the geographical variation of which is so striking and has already required the recognition of many subspecies. This new material permits a summary of our knowledge of this species in eastern Africa plus some significant additions, including one very distinct new subspecies. Thus far, the marmoratus subspecies recognized in the present area are as follows : orgentovittis Ahl : Shores of Lake Tanganyika. rhodoscelis (Boulenger) : Luapula River drainage and Lake Rukwa, Tanganyika. This is apparently a disjunct range and suggests that Lake Rukwa might have been part of the drainage of the Chambeshi River, which is itself loosely connected with the Luapula drainage through a swampy area south of Lake Bangweolo. lestagei Laurent (a doubtful form) : Lake Bangweolo. mclanolcucus Laurent: Primarily the Lufira basin in the Congo, but samples from Sakania (near the Rhodesian border) show conclusively that its range exceeds the limits of this drainage in a southward direction. nlhorufus Laurent: Recently described (Laurent, 1964) from the Moxico Province, Angola, but also present in the Lua- laba District (southwestern Katanga). This race is likely to be found in the extreme northwestern part of Rhodesia. 2 BREVIORA No. 216 aposematicus Laurent : Only known to me from Lealui. Prob- ably restricted to Barotseland. Possibly intergrading with alhornfus to the north. rlwdesianus Laurent: Described from Matetsi, Southern Rho- desia and presumably ranging into the Wankie region. swynnertoni FitzSimons : Described from Chirinda, eastern Southern Rhodesia. tneniatus Peters: Lowlands in Mozambique southward to northern Natal and eastern Transvaal. aJhofasciatus Hoffman : Southern Nyasaland. nyassae Ahl : Described from Langenburg. The precise loca- tion of this locality appears controversial. Loveridge has stated (1957, p. 329) that it is Manda, Lake Nyasa. Manda is on the eastern shore at 10°28' S. In the Gazetteer No. 1 (British East Africa) published by the United States Board on Geographic Names, no Langenburg is found, but there is a New Langenburg, which is Tukuyu. Tukuyu is near Rungwe where many other specimens of nyassae and filllehorni (which is a synonym) have been collected. This, therefore, sounds more probable as a type locality than Manda which is rather far from the Rungwe Mountains. Lastly, according to Stieler's Atlas (1905), Langenburg is on the eastern shore of Lake Nyasa, but almost at its north- ern tip. It would require fresh material from the Rungwe region and the northern part of Lake Nyasa to resolve the point. Vesey-Fitzgerald's material includes good series from Lusaka, Seremje, Mazabuka and Lake Chilwa, in addition to series from Lake Tanganyika and Lake Rukwa, which merely confirm that argcntovittis Ahl and rhodoscclis (Boulenger) are respectively present there. Hypeholius marmoratus melanoleucus Laurent Hyperolius melanoleucus Laurent, 1941, Kev. Zool. Bot. Afr., 34: 157, pi. VIII, figs. D, E, r. — Lukaf u, Upper Katanga, Congo. New material. 3 $6,1 9 (MCZ 37362-65), Lusaka, North Rhodesia, 7-9-III-1962 ; 1 9 (MCZ 38824), River Mwambeshi, north of Lusaka, North Rhodesia, 14-III-1963 ; 2 9 9 (MCZ 38807-08), Mkushi District, North Rhodesia, 3-IV-1963. 1965 HYPEROLIUS MARMORATUS 3 Color pattern. Two males (31-34 mm) have still the juvenile dull color pattern. The third male which is small (26 mm) and the female (34 mm) have the vivid " melanoleucus" markings: black with white bands and a few white spots. The pattern com- prises essentially one mediodorsal and two laterodorsal main elements with some irregularities (branches, lumbar inflexions and isolated spots) ; in the female, the center or the axis of these white spots and bands shows red spots or lines. In the male with the adult pattern, these red markings exist only on the sides. The belly also shows red spots. In live specimens (I saw them very often when I was in Katanga) these red ventral markings stand out on the paler pink general coloration. The female from Mwambeshi River (30 mm) has a somewhat disrupted pattern with red only on the sides, the belly and the throat. The two females from the Mkushi District have the usual pattern, with the remarkable feature that the red dorsal lines are very broad. Range. The older record of this subspecies at Sakania was already a proof that its range was not limited to the upper Lufira drainage. The specimens from North Ehodesia suggest that this range actually overlaps the Zambesi basin, at least in the Kafue drainage. It must be remembered, however, that two specimens with an adult pattern are far from an adequate sample and that a situation similar to that which I shall describe for the Serenje population is surely not excluded. HyPEROLIUS MARMORATUS NYASSAE Ahl Eyperolius nyassae Ahl, 1931, Daa Tierreich, 55: 339, fig. 213. "Langen- burg, " northeastern shore of Lake Nyasa, Tanganyika. Eyperolius fiilleborni Ahl, 1931, Das Tierreieh, 55: 349, fig. 224. "Langen- burg, ' ' northeastern shore of Lake Nyasa, Tanganyika. New material. 13 S S,2 $ $ (MCZ 38809-38823), Serenje, Chikoli River, North Rhodesia, 10-III-1963. Color pattern. Three males have the juvenile pattern. One has a " melanoleucus" adult pattern. The nine others have irregular black spots, a few large ones or many small ones, with a general predominance of the light ground color. Red lines in the middle of the light network are generally distinct, but they are less conspicuous when the black spots are small and numerous. Three of the males have still a white mediodorsal band free of black spots (this is a remnant of a " melanoleucus''' 4 BREVIORA No. 216 pattern) as well as the red markings. One of the males has red spots on the throat. One of the females has a moderately dis- rupted " melanoleucus" pattern, with heavy red markings; in the other, the pattern is strongly disrupted and, like most males, still with clear red lines. Size. Males with juvenile pattern : 31, 32, 35 mm. Males with adult pattern: 31 (2 specimens), 32 (1), 33 (4), 34 (1), 36 (2) mm. Females: 35, 36 mm. Discussion and range. The pattern displayed by the majority of the specimens is quite different from the "melanoleucus" pat- tern. On the other hand, it appears identical with that of filllehorni Ahl over which the name nyassae Ahl, based on specimens with a juvenile pattern, has page priority. This judgment is confirmed by the comparison with two fiillehorni paratypes. I still have misgivings about the identification of the new material from the Chikoli River because nyassae and fiillelorni are based on populations coming from the northern part of the Lake Nyasa basin, while the Chikoli River seems to belong to the Luangwa drainage. However, there are already several examples in Hyperolius marmoratus of ranges which overlap two adjacent drainages while, on the other hand, ranges can also be separated by dividing crests or highlands as well as by escarpments and falls. Therefore, we must now admit that the range of the nyassae subspecies should be extended to in- clude the Luangwa basin. Hyperolius marmoratus albofasciatus Hoffman Hyperolius alhofasciatus Hoffman, 1944, Soolog. Navors. Nas. Mus. Bloem- fontein: 178, fig. 8. Limbe, Nyasaland. Hyperolius marmoratus alhofasciatus, Loveridge, 1953, Bull. Mus. Comp. Zool., 110: 350. Limbe, Mtimbuka, Euo River, Nyasaland. New material. 3 $ $,1 $,6 juv. (MCZ 46186-95). Lake Chilwa (= Lake Shirwa), Nyasa, 3-X-1943. Size. Males : 20.9, 22.7, and 23.9 mm. Female : 22.4 mm. Color pattern. The four adults have the typical albofasciatiis pattern ; in the smallest male, however, the dark coloration is brown rather than black, reminiscent of the juvenile color. The juveniles show a strong tendency towards a longitudinal pattern similar to the ''hayoni" pattern; this is not surprising, since this pattern is established in taeniatns, a Mozambican subspecies, which occurs not far from southern Nyasa. The shift to a longi- tudinal pattern occurs by the elongation in a backward direction 1965 HYPEROLIUS MARMORATUS 5 of the hour-glass spots which are located between the eyes and in the scapular region in the juvenile " undidatus" pattern. Similarly, the lumbar undulations of the undulatus pattern are also elongated, but in a forward direction. Discussion. Loveridge (1953) rightly objected to my synon- ymizing melanolcucus with alhofasciatus (Laurent, 1947b). Indeed, the red elements characteristic of the "melanoleucus" pattern are not apparent in alhofasciatus. Another difference is that the sides are predominantly white in alhofasciatus. This is similar to the situation in the southern subspecies {marmor- atus and verrucosus) where the dark dorsal pattern is generally restricted to the back, showing no encroachment on the flanks. In melanolcucus, there is still a large black elongated spot on the sides or several smaller black markings. In addition, the juvenile pattern does not show any kind of elongation in melano- lcucus. The size of the specimens from Lake Chilwa is unusually small for the species. This striking peculiarity is not shared by the series collected at Limbe (type locality of alhofasciatus), Mtim- buka and the Ruo River (Loveridge, 1953). Apparently the populations from Lake Chilwa, which seems completely isolated (i.e. without any efferent river), are dwarfed. An additional problem derives from the fact that the type of alhofasciatus is an adult female of only 24 mm, which does not match the size of the frogs collected at Limbe by Loveridge, but corresponds to our Lake Chilwa series. Consequently, I wonder if the type was not actually collected on the shores of Lake Chilwa. HyPEROLIUS MARMORATUS PYRRHODICTYON^ SUbsp. U. Holotype. 1 9 (MCZ 46182), Mazabuka, in the water-grass fringe of the Kafue River, 31-1-1963, North Rhodesia, Vesey- Fitzgerald coll. Paratypes. 2 $ $, 1 juvenile, same data (MCZ 46183-85). Diagnosis. A race of Hyperolius marmoratus, characterized by the absence of any spots or marbling on the back, and the presence of a dark red network on the belly and throat. Color pattern. The absence of any definite dorsal pattern is not quite unique in the marmoratus group : it is characteristic of some well differentiated subspecies such as rhodoscelis (Bou- lenger) of the Luapula drainage (Congo and North Rhodesia) 1 From the Greek for "red network." 6 BREVIORA No. 216 and Jcarissimhiensis Ahl of the region of the Karisimbi volcano. These two races differ in their juvenile patterns and their color- ation in life : belly vermilion red with bluish flanks and a white laterodorsal streak in rJiodoscelis, belly purple red with back almost entirely dark blue in karissimhiensis. I don't know the colors in life of pyrrhodictyon, but a red network on an otherwise white belly is a quite unknown feature in any other member of the genus Hyperolius. On close examination under the binocular, it appears that some black spots may be present on the flanks, around the vent, and on the limbs along the boundary between the dorsal gray or brown (in alcohol) and the ventral red which becomes a solid color (not reduced to a network) on all the parts of the limbs that are hidden in the normal resting position of the tree frog. Therefore, as generally in the genus, the thigh is almost entirely vividly colored (red in this case) except for a dorsal stripe which is gray or brown. In the smallest of the two female paratypes, the black spots are also distinct on the throat between the meshes of the red network. The juvenile is almost uniformly greyish. Discussion. Since the juvenile, as sometimes happens in the marmoratus group, has no distinct pattern, specific identifica- tion may seem questionable. However, the morphology appears to be that of H. marmoratus which is different enough from that of other species to be relied upon at least tentatively. Size. The holotype measures 32 mm, the two adult paratypes 30 and 26 mm respectively. Range. Since there are only the four specimens from Maza- buka, we do not know the range of this form. Mazabuka is not far from the right bank of the Kafue River. We already know that the left part of the Kafue drainage is inhabited by melano- leucus, that aposcmaticus lives in the Upper Zambezi region (Barotseland), and rhodesianus exists to the south of the Zam- bezi in western Southern Rhodesia. Poynton {in litt.) believes that aposematicus cannot belong to the same species as rho- desianus since they do not form hybrid populations where they meet; if he is correct, we can expect similar discoveries in the future and the species Hyperolius marmoratus would then have to be split into several species. However, if the dividing line is the Zambezi itself or the Victoria Falls, the lack of a hybrid belt has no meaning at all, reflecting as it does mere physical in- ability to meet. 1965 HYPEROLIUS MARMORATUS 7 The real relations of pyrrJiodictyon with the surrounding populations attributed to //. marmoratus will thus remain un- certain until more material is collected. Acknowledgment. This work has been supported by National Science Foundation Grant NSF G-1342. BIBLIOGRAPHY Laurent, R. F. 1941. Contribution i la systematique du genre Hyperolius Rapp (Batraciens). Rev. Zool. Bot. Afr., 34: 149-167. 1943. Les Hyperolius (Batraciens) du Musee du Congo Beige. Ann. Mus. Congo, Zool., (1) 4: 61-140. 1947a. Two new forms of the genus Hyperolius. Ann. Mag. Nat. Hist., (11) 14:294-296. 1947b. On some misuses of Hyperolius names. Ann. Mag. Nat. Hist., (11) 14: 288-294. 1957a. Catalogue des rainettes africaines (genres Afrixalus et Hypero- lius) de la collection du Museum National d'Histoire Naturelle de Paris. Ann. Soc. Roy. Zool. Belg., 82: 23-50, figs. 1-2. 1957b. Apergu des formes actuellement reconnaissables dans la super- espece Hyperolius marmoratus. Ann. Soc. Roy. Zool. Belg., 82: 379-397. 1964. Reptiles et amphibiens de 1 'Angola (troisieme contribution). Publ. Cult. Museu do Dundo, No. 67 : 1-165. LOVERIDGE, A. 1953. Zoological results of a fifth expedition to East Africa. IV. Amphibians from Nyasaland and Tete. Bull. Mus. Comp. Zool., 110: 325-406. 1957. Check list of the reptiles and amphibians of East Africa (Uganda; Kenya; Tanganyika; Zanzibar). Bull. Mus. Comp. Zool., 117: 151-362 -f i-xxvi. 8 BREVIORA No. 216 Fig. 1. Ventral aspect of Hyperolius marmoratus pyrrJiodictyon n.sp. Type, MCZ 46182, \ 1965 IIYrEROLIUS MARMORATUS A ALBORUFUS A APOSEMATICUS e ARGENTOVITTIS • RHODOSCELIS N NYASSAE A ALBOFASCIATUS TAENIATUS PYRRHODICTYON RHODESIANUS SWYNNERTONI LESTAGEI MELANOLEUCUS Fig. 2. Map of the localities for the subspecies of Ryperolius marmoratus in East Africa. BREVIORA Miiseiim of Comparative Zoology Cambridge, Mass. March 1, 1965 Number 217 THE AUDITORY REGION OF THE BORHYAENID MARSUPIAL CLAD0SICTI8 By Bryan Patterson Since the appearance of Sinclair's memoir on the marsupials of the Santa Cruz formation of Patagonia (1906), it has been known that certain members of the Borhyaenidae, such as Borhyaena and Prothylacynus, were peculiar among marsupials in lacking a tympanic process, or bulla, of the alisphenoid. So far, indeed, had reduction of this bone been carried in Borhyaena that all trace of the foramen ovale had disappeared, the mandibular nerve evidently having passed out between ali- sphenoid and tympanic without so much as a notch to mark its passage. Sinclair was, however, able to show that lack of a tympanic process of the alisphenoid was not characteristic of the family as a whole, this structure being present in at least two genera, Cladosictis and " Aniphiproviverra" (=TJiylacocUc- tis). His observations have been confirmed by others (e.g. Wood, 1924). Recently, the surprising statement has been made by McDowell (1958, p. 173) that "the borhyaenines show no clear relationship to other marsupials (they have no alisphenoid bulla, otherwise characteristic of marsupials, for example)." Even if this claim were correct, these animals would not be unique in this respect within their order. As has long been known (e.g. van Kampen, 1905, pp. 406-407), Phascolomis lacks an alisphe- noid bulla, although it is remarkable in possessing a process from the squamosal that fulfills the same function. In the course of a recent visit to Princeton University, I was able once more to examine the borhyaenid material described by Sinclair and reassure myself that his descriptions were correct. In the course of this examination I came across an undescribed 2 BREVIORA No. 217 specimen that adds to knowledge of the borhyaenid auditory region. This is Princeton University no. 15705, consisting of various fragments, among which is an incomplete cranium. It had been identified, presumably by Sinclair himself, as Clado- sictis histratus Ameghino, a determination which there is no reason to doubt. From the fact that Sinclair made no mention of the specimen in his memoir, I suspect that it was not prepared prior to publication. Professor Glenn L. Jepsen very kindly loaned me the cranium for description. The accompanying fig- ures are the work of Mrs. Dorothy Marsh. DESCRIPTION The cranium is broken off anteriorly in the vicinity of the postorbital constriction and has been considerably damaged by weathering, which has removed the condyles, the bases of the zygomatic arches together with the glenoid cavities, and much of the bone of the right side. Compensating for these losses is the almost complete lack of crushing, something that cannot be said for the specimens of this species that Sinclair described and figured. As in other borhyaenines, and unlike the crushed specimen figured by Sinclair (pi. 55, fig. 1), the basisphenoid and basi- occipital progressively increase in width posteriorly. As he noted, neither bone has a distinct median keel and both are relatively flat transversely; at the suture they form a fairly prominent transverse ridge. No carotid foramen is to be seen in the basisphenoid ; in fact the only foramen visible in this bone is a minute one immediately antero-internal to the foramen ovale. Between basioccipital and periotic there is a conspicuous foramen antero-internal to and almost as large as the foramen lacerum posterius. This is the posterior carotid foramen of Gregory (1910, p. 233) ; a wide, deep groove in the postero- lateral portion of the basioccipital leads forward to it. The significance of this foramen is discussed below. The pars petrosa of the periotic, as is clearly seen on the left side, is rather sharply pointed anteriorly and extends far forward. Except for the opening of the foramen just described, the medial border of the pars petrosa abuts very closely against the basioccipital and basisphenoid. There is a long and low, slightly curved and narrow auditory bulla that does not project ventrally below the basicranium. The 1965 AUDITOEY REGION OF CLADOSICTIS anterior two-thirds of this structure is formed by the alisphe- noid. Antero-externally, this portion is suturally united with the squamosal (part of this area is broken away in the specimen). Medially, there is a long, fairly wide and deep depression be- tween the basicranium and the alisphenoid, which is roofed by the pars petrosa. The Eustachian tube opened into the posterior portion of this depression and passed forward within it. On the 1 u /^^ r.er-- •fen. r.'' Fig. 1. Cladosictis lustratus Ameghino. Ventral view of incomplete cranium. Princeton University no. 15705, X2. Abbreviations: al., alisphe- noid; al. b., alisphenoid bulla; Eu., opening for Eustachian tube; f.c.p., posterior carotid foramen; f.l.p., foramen lacerum posterius; /. ov., foramen ovale; fen. r., fenestra rotunda; p. mas., pars mastoidea of periotie; p. p., paroccipital process; p. pet., pars petrosa of periotie; r. e., epitympanic recess ; sq., squamosal ; ty., tympanic. broken right side of the specimen a natural section of the ali- sphenoid portion of the bulla is preserved. This is revealed as a deep, oval chamber situated almost entirely within the lateral wall of the cranium. It is separated from the rather large, circular epitympanic recess by a low ridge formed by the alis- phenoid anteriorly and the squamosal posteriorly. The remain- ing third of the bulla is formed by the periotie, which meets the BREVIORA No. 217 alisphenoid in a suture anteriorly and is clasped posteriorly by the heavy, squat and blunt paroccipital process. The portion of the periotie involved is not the pars petrosa, which, as in other borhyaenines, exhibits no trace whatever of a tympanic process, but the pars mastoidea, which, in this form at least, had enlarged downward and forward from the ventro-lateral corner of the occipital surface. — a.t.c.f: mas. f. ov. I p.pet Fig. 2. Cladosictis lustratus Ameghino. Lateral and slightly ventral view of incomplete cranium. Princeton University no. 15705, X2. a.t.c.f., aper- tura tympanici canalis facialis; /. pgl., postglenoid foramen; /. sub., sub- squamosal foramen; VII, grooves for facial nerve; other abbreviations as in Figure 1. The meatus — a meatus spurious — is large, very short, and deeper than long; it is bounded anteriorly and dorsally by the squamosal, posteriorly by the squamosal above and the pars mastoidea below (the suture between these bones is unfortunately not distinct), and ventrally by the alisphenoid in front and the pars mastoidea behind. The tympanic is present on the left side. It is free of the bulla and gives the appearance of having been situated wholly or almost wholly within it, although evidently lying close to the bony meatus. The bone has the primitive 1965 AUDITORY REGION OP CLADOSICTIS 5 horseshoe shape so frequently encountered in the order ; it is, so far as can be seen, slightly enlarged ventrally, and has a short, dorsally tapering posterior crus. Due to slight forward dis- placement, the anterior crus is concealed by the squamosal ; this displacement is evidence that, as is usual in marsupials, the tympanic was not fused to the skull. The internal surface of the pars mastoidea is grooved anteriorly for the passage of the facial nerve, and there is a corresponding slit in the ventral face of the tympanic. DISCUSSION This is the first specimen of a borhyaenine in which a fully formed auditory bulla has been preserved, but it is not unique in the family. A complete bulla, superficially similar to that of the machairodontines, is known in Thylacosmilus. In a brief description that I gave of this structure (m Riggs, 1934, p. 13), the alisphenoid was identified as a component, although the extent of its participation could not be determined, and the pars mastoidea and paroccipital were recorded as "overlapping" the bulla; uncertainty was expressed as to whether or not the tympanic and a tympanic wing of the pars petrosa also partici- pated. With Princeton University no. 15705 in hand I have recently gone over the Thylacosmilus material in the Chicago Natural History Museum, hoping for a clearer interpretation in the light of this new evidence. It now seems virtually certain that the tympanic was situated within the bulla, as in Clado- sictis, but, for the rest, cracks, absence of recognizable sutures and slight distortion prevent any assurance regarding the ele- ments contributing to the posterior portion of the bulla. I am now very dubious as to the existence of a tympanic wing of the pars petrosa and suspect that only the pars mastoidea was in- volved in the bulla, but certainty on these points can only come with additional evidence. As has long been known, the borhyaenids show very definite resemblances to two marsupial groups, the Didelphoidea and the Dasyuroidea, and there has been controversy as to whether they were definitely dasyuroid, even thylacinine, or an ofi'shoot from the didelphoid stock. The evidence has been summed up by Simpson (1941), who, in general agreement with such earlier authors as Winge, Ameghino, and Matthew, concluded that while all three groups are clearly related, and very probably had a common ancestry, the Borhyaenidae were descended from the 6 BREVIORA No. 217 Didelphoidea, the resemblances between them and the larger members of the Dasyuridae being due to parallelism. With this view I am in complete accord, and the structure of the auditory- region contributes a few additional points to its support. In the Dasyuridae there is a large depression or recess in the squamosal, posterolateral to the small recessus epitympanicus, which is covered by the pars flaccida of the tympanic membrane ; no such depression is present in the Didelphidae and Borhyae- nidae. The facial nerve in dasyurids, after leaving the middle ear cavity, runs through a bony canal formed mainly by the pars mastoidea with the squamosal contributing antero-ventrally {Thylacinus, in which there is only a groove in the pars mas- toidea, is an exception) ; no such canal occurs in didelphids and borhyaenids. Borhyaeuids, so far as known, lack a tympanic process of the pars petrosa. In the Didelphidae this process is less developed than in the Dasyuridae, in which it usually con- tributes largely toward the formation of the bulla {Thylacinus, which lacks the process, is again an exception). Two borhyaenine genera are known to lack an alisphenoid bulla. This structure is smaller in a number of didelphids than is the rule among dasyurids. Cladosictis is peculiar among marsupials in the par- ticipation of the pars mastoidea in the bulla, but it is not unique in this respect. The remarkable little Dromiciops, which, with Reig (1955), I believe to be a surviving microbiotheriine didelphid, has a large, complete bulla in which, in addition to a tympanic wing of the pars petrosa, the pars mastoidea is involved posteriorly. Sinclair (p. 410, pi. 62, fig. 7) has figured a speci- men of Microhiotherium with a fully formed bulla, to which this element may well have contributed, and, as noted above, the same may well have been true of Thylacosmilus. THE INTERNAL CAROTID ARTERY IN THE THERIA Shortly after preparing this description I had the opportunity to make a field dissection of the head of an opossum (Didelphis marsxipialis) . In the course of this, I noted a small branch leaving the internal carotid artery at the level of the foramen laeerum posterius, passing into the cranial cavity through the posterior carotid foramen and there joining the circle of Willis. The observation was subsequently confirmed in injected specimens by Miss Suzanne Kreinbrook in the Biological Laboratories, Harvard University. This condition is widespread in the Mar- supialia, to judge from the nearly universal occurrence of the 1965 AUDITORY REGION OF CLADOSICTIS 7 posterior carotid foramen. Among the Recent genera I have been able to examine, the foramen appears to be lacking only in the microbiotheriines, and in Acrohates — and possibly Phasco- larctos — among the Phalangeridae. Even in these cases it is impossible to be certain short of dissection, for the foramen may be little more than a slit between the basioccipital and the peri- otic, as in caenolestids, or, apparently, confluent with the foramen lacerum posterius, as in some macropodids. The internal carotid artery passes through the basisphenoid in pelycosaurs and therapsids (e.g. Romer and Price, 1940 ; Olson 1944), and also in non-therian mammals so far as known — Monotremata, Triconodonta (Kermack, 1963) and Multituber- culata (Simpson, 1937). The marsupials have retained this prim- itive condition, but have an alternate route in the small branch of the artery that passes through the posterior carotid foramen. Cladosictis, so far as I am aware, is the only member of the order in which this alternate route was fully followed, and in which the anterior portion of the artery and the foramen in the basi- sphenoid was lost. Cladosictis had, in fact, essentially and inde- pendently achieved the placental condition. Retention of the primitive condition in marsupials suggests that the placental arrangements were not attained until after the eutherian-metatherian dichotomy, presumably at some time or times during the Cretaceous. Just how these arrangements came about is, of course, unknown. The posterior carotid foramen alternate route may have been followed by all or by some pla- cental groups, but the possibility would appear to exist that a small, more anterior branch may have reached the cranial cavity via an opening antero-internal to the tympanic region (a fore- runner of the foramen lacerum medium), and that this even- tually became the anterior portion of the internal carotid. In either case the artery would have run forward along the medial edge of the periotic and could thus have readily become enclosed in a groove or canal between basicranial and auditory elements. The stapedial artery presumably came into existence quite early in placental history. Matthew (1909), who recorded its presence in arctocyonids, creodonts and miacids, suggested, rightly I sus- pect, that the internal carotid early became divided into medial and lateral (stapedial) divisions. With the establishment of these main branches, and given the relative ease with which blood ves- sel patterns may change, a variety of combinations became pos- sible. Independent acquisitions of similar patterns could have 8 BREVIORA No. 217 occurred and almost surely did occur. Since the supply of blood rather than the route taken by it is the important thing, it is doubtful if there is any selective advantage in one vessel pattern over another. How varied these patterns may be has recently been demonstrated by Guthrie (1963) for the rodents. Within this one order almost all possible variations are present ; in some groups both internal carotid (medial) and stapedial (lateral) branches occur ; in others one of these is absent, and in yet others both are wanting. Details of the carotid circulation appear to be of dubious value as an item of evidence for determining higher affinities within the Eutheria. EEFEEENCES Gregory, W. K. 1910. The orders of mammals. Bull. Amer. Mus. Nat. Hist., 27: 1-524. Guthrie, D. A. 1963. The carotid circulation in the Eodentia. Bull. Mus. Comp. Zool., 128: 457-481. Kampen, p. N. van 1905. Die Tympanelgegend des Saugetierschadels. Morph. Jahrb., 34: 321-722. Kermack, K. a. 1963. The cranial structure of the triconodonts. Phil. Trans. Koy. Soc. London, (B), 246: 83-103. McDowell, S. B. 1958. The Greater Antillean insectivores. Bull. Amer. Mus. Nat. Hist., 115: 115-214. Matthew, W. D. 1909. The Carnivora and Insectivora of the Bridger Basin, middle Eocene. Mem. Amer. Mus. Nat. Hist., 9: 291-567. Olson, E. C. 1944. Origin of mammals based upon cranial morphology of the therapsid suborders. Geol. Soc. Amer. Spec. Pap. no. 55 : i-vi, 1-136. Eeig, O. a. 1955. Noticia preliminar sobre la presencia de microbiotherinos vivi- entes en la fauna sud-americana. Investigaeiones Zool. Chilenas, 2: 121-130. ElGGS, E. S. 1934. A new marsupial saber-tooth from the Pliocene of Argentina and its relationships to other South American predacious mar- supials. Trans. Amer. Phil. Soc, n.s., 24: 1-31. 1965 AUDITORY REGION OF CLADOSICTIS 9 ROMER, A. S. AND L. I. Price 1940. Eeview of the Pelycosauria. Geol. Soc. Amer. Spec. Pap. no. 28: i-x, 1-538. Simpson, G. G. 1937. Skull structure of the Multituberculata. Bull. Amer. Mus. Nat. Hist., 73: 727-763. 1941. The affinities of the Borhyaenidae. Amer. Mus. Novit., no. 1118: 1-6. Sinclair, W. J. 1906. Marsupialia of the Santa Cruz Beds. Repts. Princeton Univ. Expeds. Patagonia, 4: 330-460. Wood, H. E. 1924. The position of the "sparassodonts": with notes on the rela- tionships and history of the Marsupialia. Bull. Amer. Mus. Nat. Hist., 51: 77-101. (Received 15 December, 1964.) BREVIORA Mnaseem of Cooiparsitive Zoology Cambridge, Mass. ^r.w 7, 1965 Number 21S NEW FK()(;S OF THE ({ENUS CORNUFER (RANIDAE) FROM THE SOLOMON ISLANDS By AV ALTER C. Brown ^ INTRODUCTION Large collections made by Mr. Fred Parker on Bougainville and neighboring small islands are providing very consider- able additions to onr knowledge of the fauna of this area. The present paper reports three new species and one new subspecies of the genus Cornufer discovered by Mr. Parker. Future papers in this series will describe other novelties in both frogs and rep- tiles and will record important ecological and behavioral obser- vations. PLATY3IANTIS SYNONYM 1 ZED WITH CORNUFER In my revision of the amphibians of the Solomon Islands (Brown, 1952), I followed Boulenger (1918, p. 372), Noble (1931, p. 522), and Deckert (1938, p. 148) in maintaining Platy- mantis and Cornufer as distinct genera. My separation of the two genera was based primarily on the structure of the digital pads, as emphasized by Boulenger (1918, p. 372). On the basis of this character, the Solomon Islands representatives of this grou]i of ranid frogs, which were known at that time, fitted rather readily into one or the other of the two categories. Inger (1954, p. 348), on the basis of his experience with the Philippine ranid frogs, again placed Platymautis in the synonymy of Cor- nufcv. In so doing, he pointed out the difficulty of maintaining a separation of these two genera when a majority of the species is considered and the apparent evolution of the digital pads is taken into account. ITowevei-, he did note that the species of 1 Division (if Systi'iiia tic Biolnj^.v. StJiiil'onl I'lii \ nsit y ami Mi'iilo ('(illrf;c. Mciilo Parlv, Califitrnia. BREVIORA No. 218 this group do have a number of characters in common, which separate them from Rana {sc7isu stricto). As to their relation- ships with other ranid genera, it has been noted by Noble (1931, ]). 520) and Brown (1952, p. 86) that as a group they are prob- ably more closely related to the genus Discodcles, which is some- what intermediate in position when the Bana-Discodeles-Cornu- fer-Batrahylodes series is being considered. Since 1952 I have worked exteusively with this group of ranid frogs (I have now examined 23 of the 28 species referable to the genus), and I agree with Inger that it is indeed difficult to maintain the two genera as distinct entities on the basis of our present knowledge of the digital or other known characters, or on the basis of any great difference in ecological adaptations. Therefore, in the following list I have assigned all of the species previously placed in PJatymantis to the genus Cornufcr and noted their distribution. However, a thorough study of these frogs in terms of their morphology and life histories is much needed in determining relationships within the group. Cornufer acrocliordiia new species, PI. 2 Soloiuon Islands (Bougainville) Cornufcr acuhodactylus (Brown), PI. 1 Solomon Islands (Bougainville, Choiseul ) Cornufcr bcauforii (Van Kanipen j Cornufer boulengeri Boettger Cornufcr cliccsniani (Parker) Cornufcr coriiutus Tayloi' Cornufer corrugatus (Dumeril) Cornufer dorsnli.'i Dumeril Cornufcr {lilliardi (Zweifel) Cornufcr gucntlicri Boulenger Cornufer guppyi Boulenger, PI. 1 Cornufcr hazelac (Taylor) Cornufer ingeri Brown and Alcala Cornufer macrops new species, PI. 1 Cornufer mo.szl-ou-ski (Vogt ) Cornufer mycrsi (Brown), PL 2 Waigeu Island (not seen by me) Bismark Island (not seen by me) New Guinea Philii»pine Islands l'liilipi)ine Islands l'hili])pine Islands Bismai'k and Adiiiiialty Islands (not seen by me) IMiilippine Islands Solomon Islands Philippine Islands Philippine Islands Solomon Islands (Bougainville) New (luinea (not seen by me) Solduion Islands (Bongainville ) Cornufer necheri Brown and Myers, PI. 1 Solomon Islands Cornufer p. papucnsis (Meyer) Cornufcr papuen.sin wchcri (Sclunidt), PI. 2 Cornufcr p. porkcri new species, PI. 1 Cornufcr parJ.cri huluK itsix new subspecies, Cornufer pelewensis (Peters) Xew (Uiinea, Bismarks Solomon Islands Solomon Islands ( P>ongain ville i Solomon Islands (Bukaj Palau Islands 1niiif< r polillciisis (Taylor) Pliilil)|iiii(' Islands Cornufcr punctata (Peters and Ddi-ia) New (iuinea (not seen ]>y me) Connifrr ruhififriatus (Bari)our) Koou Island ('(>r)uif( r sdlonioni.s Boiilentjer, I'l. 2 Solomon Islands Cuntiif( r suht('rr(!'raphical point of view, as shown by this list of species, Cornufcv appears to represent a relict, peripheral group of ranid frogs with the greatest number of species occurring at present in the fringing Philippine and Solomon archipelagos, and a smaller remnant in New Guinea and related islands, as well as in the Palau and P'iji Islands in the outer Pacific. NEW SPECTES The new species collected by Mr. Parker are ail relatively small and have been compared directly with the tyjx' of C. andvodartijlnx ( USNM 119769) and with a paratype of C. chccs- mani (MCZ 26501). (The relative sizes at maturity of the sev- eral species known from the Solomon Islands are given in Table 1.) CORNUFER PARKERI I'ARKERI ^ sp. and Subsp. IIOV. This diminutive frog with its rough, tuberculate skin sui)er- ficially resembles some of the small Oriental bufonids. However, its firmisternal girdle, well developed omosternum, undilated sacral diapophyses, teeth only on the upper jaw, reduced webs and united outer metatarsals place it in the ranid genus Coruu- fer. Holofjipc: MC'Z 86928, a mature female collected at Kunua area, Bougainville Island, Solomon Islands, on July 10, 1962, by Fred Parker. ' Paratypes: MCZ 86911-18, 36921-22, 38194. 41860, 41866-69, 42524-81, 48741-44, Stanford University Nos. 21778-74. AMNII 70069-71, collected from the same general locality as the holo- type, during 1962-1968. Diagnosis: A diminutive Cornufcr, largest available mature female measuring 18.5 mm, and largest male 15.9 nnn from snout 1 Niinicil f(ir Mr. Fred Parker. 4 BREVIORA No. 218 to vent; skin witli numerous warty tubercles on dorsum, lateral surfaces and limbs ; first finger shorter than the second ; tips of fino'ers blunt or slio-htly pointed, occasionally a eircummarp:inal o-roove faintly indicated; tips of toes scarcely dilated, rather pointed, a shallow groove, most prominent laterally, separating the dorsal from the ventral portions ; diameter of eye usually greater than, rarely equal to, length of snout. Description: A very small Cornufer, snout-vent length of 16 mature females 15.1 to 18.5 mm, of 8 mature males 14.0 to 15.9 mm, habitus slender; hind limbs long, the snout-vent length ranging from 60 to 72 per cent of the length of the hind liml) for 10 specimens ; head about as broad as long ; snout round- pointed, upper jaw scarcely protruding ; eye large, its diameter somewhat greater than the length of the snout (rarely equal to) ; tympanum distinct, large, its diameter about 50 to 70 per cent of the diameter of the eye, and usually almost 25 per cent of the breadth of the head ; canthus rostralis rounded ; loreal region concave, only slightly obli(iue; a moderately to faintly distinct, oblique fold dorsal and posterior to the tympanum ; forelimbs well developed, fingers relatively uniformly slender, bluntly round or slightly pointed, occasionally with a very faint groove separating a ventral ]iad laterally from the dorsal por- tion; witliout webs; subarticular and metacarpal tubercles large and well developed ; first finger shorter than the second which is about equal to the fourth (PI. 1, fig. 3) ; hind limb long; toes very slender without web, rather pointed with the ventral pad delimited by a shallow groove except at the tip (the fifth toe is more blunt and usually lacks the groove) ; subarticular and both inner and outer metatarsal tubercles well developed. Skin of dorsum and dorsolateral surfaces marked by scattered, moderate-sized, round, oval or oblong tubercles ; ventral and dis- tal posterior part of the thighs granular; the proximal posterior part of thighs marked by elongate folds. Color (in ])reservative) : dorsum and lateral surfaces from grayish brown, through brown, to brownish black, the lighter shades with irregular darker blotches; several of tlie sj)ecimens exhibiting a tan, silvery or whitish middorsal band, wider anteriorly than posteriorly, and beginning anteriorly on the posterior jiart of the head or tlie pectoi-al region ; lower limbs and (Hh^r of jaws with wide, dai-k ti-ansverse bars; venter rather hea\ih- mottled with dark brown. IDGf) NKW KKOdS KKO.M 'rUb: SOLOMON ISLANDS o CO ^_^ ;-c ii (l; is ^ o p ^ r-l V. II ^--, s ^ 23 *- ^ M ^^ cr o m c; II rw^ m (D +J «*^ O >» -4J -*- 03 s -!-:> rt ai N m S-np.lOll.X).!,)!) 'J SIIIOUIOIOX 'J j.tJ.i.ivd i.t.i.i.iixl 'J ns'UJD.imi LI >:(.n>(l 'J y.ciq.im sifmindnd 'Q i.i^^^,^,ni 'J }s:i,jfuu 'J xdo.tJDUl 'J iliddn6 -j s-ii]l\ i.}ni)0>in.n} 'j K 1^, K p^ 'A Ph ^ CO P^ ^ p^ \n CO P^ to cr, CO i--i 01 01 re P5 o o Pi; <^ O 1- CO -o OC t-l p:; ^ CO >ra K 05 pi! ;2; O CD 00 CO P5 _• 01 Ph ^ o o CO o P^ rci OC 01 01 K p? lO 00 cc -. 1- p^ lO 00 oi — Ol 01 — p^ 01 CO -H M cc .J 6 BREVIOEA No. 218 Measurements of holotype (in mm) : Snout-vent length 18.5; length of head to posterior edge of tympanum 6.8 ; breadth of head 6.9 ; diameter of eye 2.4; diameter of tympanum 1.7 ; length of snout 2.4 ; length of hind limbs 27.5 ; length of tibia 8.1. Ecological Jinfc: Parker (personal eommunieation ) states that the specimens of this small frog were found under stones and logs in lowland secondary growth areas. Comparisons: This Cornufer is much smaller at maturity than any other known species of the genus. In the warty nature of the skin it is most similar to C. acrochordus. Cornufer parkeri bukaensis subsp. no v. Holotype: MCZ 35777, a mature female, collected in lowland forest at south end of Buka Island, Solomon Islands, on January 28, 1962. l)y Fred Parker. Paratypcs: AMXII 69814-15, same locality as holotype. Diagnosis: The Buka population is distinguishable from the nominate subspecies by the much less warty skin (both dorsal and ventral surfaces) ; the larger eye as measured by the ratio of eye diameter to breadth of the head and the relatively broader head as measured by the ratio of the length of the head to its breadth (Table 2). ' Because of their obvious close affinities, and the fact that the observable morphological differences between individuals of these populations of diminutive Cornufer are based upon a very small sample of the Buka population, I prefer to regard these two populations as geographic subspecies of a polytypic species. AVere these populations overlapping in range and wer(^ there no intergradation, they would be recognized as full species. As pointed out by Mayr (1963, pp. 481-515), geographieally isolated populations such as these island populations are certainly incip- ient species wliether or not marked by pronounced morphologi- cal differences. If the isolation is relatively complete for a suf- ficiently long period of time, true reproductive^ isolates f distinct species) may arise. Color (in preservative) : Dorsum and iippci- hiteral surfaces more or less uniforml\- piifplisli bi'owu or with lighter blotches; h)wer fore limbs and to some (Icgi'cc the thighs uiMi'kcd with dai-k transvei-se bands; venter with a i-cticuiate pattern of bi-own and grayish white. 196.') NEW FROGS KKOM THE SOLOMON ISLANDS Measnrcmcnts of liolotypc (in mm) : Snoiit-vcnt lon^tli 15.9; length of head to posterior eder) iJiameter of eve Length of head Breadth of head Breadth of head R = 0.339-0.405 R = 0.966- 1.050 ('. parkeri parkeri M = 0.376 M = 1.003 X = 20 X = 20 R = 0.318-0.328 R = 0.922-0.952 C. parkeri hukaensis M = 0.323 ^r = 0.937 X = 3 x = a CORNUFER MACROPS^ Sp. nOV. Hulutype: MCZ 41864, an adult female, eolleeted at 3U00 to 4000 feet, in mountains of Aresi area, south of Kunua, Bougain- ville Island, Solomon Islands, 1963, by Fred Parker. Faratijpes: MCZ 38195-96 and 43740 in mountains of Kunua area, Bougainville Island ; Stanford University Xo. 21795, Kieta area, Bougainville Island, collected by Fred Parker, 1962. Diagnosis: A small Cornufcr, largest available mature female measuring 28.5 mm, and largest male 25.9 nnn from snout to 1 From the Greek meaning "large eye." 8 BREVIORA No. 218 vent ; second finger longer than first ; slightly dilated disks at the tips of the fingers and toes ; the ventral pad separated from the dorsal by a circummarginal groove ; eyes relatively large, diameter of eye greater than length of snont (Table 3), about 4U per cent of head breadth. Description: A moderately small Cornufer, snout-vent length 26.0 to 28.5 mm for the two females; 23.2 to 25.9 mm for the three adult males ; habitus slender, tapering from head to groin ; hind limbs long; the snout-vent length about 65 per cent of the length of the hind limb ; head about as broad as long ; snout rounded, upper jaw not protruding; eye very large, its diameter about 16 to 17 per cent of the snout-vent length, greater than the length of the snout (Table 3) ; tympanum distinct, its diam- eter slightly more than 20 per cent of the breadth of the head ; canthus rostralis rounded ; loreal region oblicpie, concave ; a rela- tively inconspicuous fold above and ]iosterior to the tympainim ; fingers relatively long, slender, without web ; finger tips slightly dilated and more or less rounded, the ventral pad separated from the dorsal portion by a shallow marginal groove, first finger much shorter than the second which is shorter than the fourth ; distal subarticular tubercles large and strongly protruding, basal and metacarpal tubercles less protruding (PI. 1, fig. 5) ; hind limb long ; toes slender, without web ; tips of toes slightly dilated, round or round-pointed, the ventral portion separated from the dorsal by a circummarginal groove ; subarticular tubercles moderately large and strongly protruding; inner metatarsal tubercle large and broadly oval, the outer small and round. Skin of dorsal and dorsolateral surfaces of body and u])per surfaces of limbs without pronnnent tubercles or folds ; skin of ventral abdominal region with faint small granules. Color (in preservative) : Dorsal and lateral surfaces blotched light and dark bi'own ; hind limbs with dark crossbars; ventral surfaces heavily fiecked with brown. Mcasnrcmoiis of holoiijpv (in mm) : Snout-vent length 26.0; length of head to posterior edge of tympanum 10.7; breadth of head 11.1; diameter of eye 4.9 ; diameter of tympanum 2.3 ; length of snout 3.9; lengtli of liind limb 44.9; length of tibia 13.6; length of third finger 4.3; diameter of third finger disk 0.8. Comparhons: Cor nu fry ))iac)'ops is distinguished from C. myersi by its snudler size at maturity ami relatively lai-ger eye, diameter of eye greater than length of snout (not less as for /y///r/'.s('). It is distinguished fi-oni ('. pdrhivi by its larger size. 1!)( ).) NEW FKUCiS FKO.M THE SULU.MON ISLANDS S 2 o ^ i o ^ 3 ^ re 3 5 -^ 2f ? ^ « &j3 K 'C o -*-' be o CO o OO i-H CO -* rJH 00 «t-l OJ I-H 1—1 1— i I-H o +j o o o p" S ^ 00 r-H ■yi re CO o c'l CO CO T— 1 1-H I-H CO rH § s <= — i-H — ' — ■ 1-H c — " -+ d =' >-'^ ■S o II 11 II II II II II II II II II II — ''m « g ^ « ^ ^ K "7^ ^ p? r-^ 'A -♦- ~ &£ o CO o t^ o ^ o CI I— I C3 ■~" O '*. Tj* ^ 'I* o +- o o o o S t^ r-c -H CO o to CO 00 -^ O CI -f CK Oi a-. 0-. 00 Ol ^ K* 't -t T— 1 cc CO o CO CO CO CO ^ o" — I-H c o rH o d ^ d d lO r5 o II II II II II II II II II II II II '^ OQ p^ % ^ w ?^ « K s ^ « <^ k'n c.> -f^ o OO to I-H o M o Ol CO =4-1 !— 1 p T-H o -*— C5 CO 5? rc c'l 5; cc a-. o I-H ci CI § EX o '=■ T— ( o" — r— 1 rH ^ 1" >— \ -1 II II II II II II II II il Hh « s ?-. « ^ P5 UH y. r^ ^ ^ QJ CI ^ CO '^ r^ o ^ +-» ■ "^ cc ^. ^ s =t-l o o o d o be CI iro CI CI t~ -t ^ ^ -^ 1^ CI CI cc CI 00 CI ^^. oc CO CI -t he re — ' 1 1 1— I o o 1-H d - lO S II II II 1 1 II II 1 1 II II h:j cS II 1 1 1 1 1 1 1 1 1 1 1 1 p: ^ v. '%- k— 1 ^ p: f^i y =0 'K ^ t— r^ ^ ^ 1^ ^ --- ^ c r— r»^ F^ o V- ^ ;. ^ i. o ■^ >*** "-- «j sj cs !=1 ^ o o o o c s: lU BREVIORA No. 218 In size, C. Diao-ops is most similar to the Papuan-Solomon species ('. acnleodactylus, C. acrochordus, n. sp., C. checsmmiae, C. (jilliardi and C. unicolor. It differs from C. aculeodactylus in its much less pointed fint^ers ; the first finger shorter than the second (not longer) and the fingers longer relative to other measurements, length of third finger to base of second subarticu- lar tubercle about 40 per cent of breadth of head as compared to 25 to 32 per cent in (\ aculeodactylus (Table 8); smaller, rounded, outer metacarpal tubercle ; and eye larger relative to length of snout (Table 3). It differs from C. (jiUiardi in that the first finger is shorter than the second (not longer) ; head narrower ; dorsal folds absent. It differs from C. acrochordus in having tlic skin much less warty; the fingers less pointed; the fingers longer ( ditf'erences for third finger length relative to breadth of head are shown in Table 3) ; and the eye larger rela- tive to length of snout (see also Table 3). C. macrops diflPers from C. checsmanae in the larger eye (diameter of eye greater than length of snout, not less than, and more than 35 per cent of breadth of head, not less than as for chcesmanae) ; and the more granular posterior venter. It differs from C. unicolor in the absence of a web at the base of the toes, tlie relatively longer hind limbs; and the relatively broader head. CORNUPER acrochordus' sp. IIOV. _^. llolutypc: MC'Z 442(i4, a mature female collected at Aresi Mountain region, south of Kunua, between 2()()()-4000 feet, Bougainville Island, Solomon Islands, on 6 September 1!)63, by Mr. Fred Parker. Paratypcs: MCZ 44256-63, 44265-66, same general area as the holotype; MCZ 41871-2 and Stanford University 21832, Aresi area south of Kunua (elevation about 30()()-4()0() feet), Bougainville Island, Solomon Islands. Diagnosis: A moderate-sized Cornufer, largest available mature female measuring 39.3 nun and largest male 27.1 nnii from snout to vent ; dorsal surfaces of limbs and body with scattered, pi'ominent. i-oundish tubercles, dorsum also with some elongate folds, venter with coarse, rounded granules; fingers short ; first finger longer than the secoiul ; fingers and toes dis- linctly |)()inted; subarticulai- and innei' metacarpal and meta- tarsal tubercles very large and sti-ongly piv)1 i-nding. 1 From the (Iicck lUr "wmiIv." 1 !)(;,■) MEW KKOCiS FKO.M THE SUEOMON ISLANDS 11 Description: A moderate-sized Cornufcr, siiout-veiit l('n<:tli about 25 to 28 mm for mature males (4 measured) ; 37.0 to 40.0 mm for mature females (6 measured). (Two females about :5() mm in len-th 60 to "to per cent of the length of the hind limb; head broader than long; snout broadly rounded; ujiper jaw not or scarcely protruding:; eye moderately large, its diameter slightly less to slightly greater than the length of the snout and about 15 to 16.5 per cent of the snout-vent length ; tympanum distinct, its diam- eter about 50 to 70 per cent of the diameter of the eye and 19 to 24 per cent of the breadth of the head ; canthus rostralis broadly rounded ; loreal region strongly oblique and only slightly con- cave ; a prominent fold above and posterior to the tympanum ; fingers slender, round-pointed to pointed, ventral pad lacking, without web or lateral fringe; first finger usually longer than second (rarely equal to) ; second finger about ecjual in length to the fourth; subarticular tubercles very large and protruding but not pointed (PI. 2, tig. 4) ; metacarpal tubercles large, the inner protruding laterally; hind limb relatively long; toes slen- der, tips of toes slightly dilated, pointed, the ventral portion separated from the dorsal by lateral grooves; subarticular tuber- cles moderate, strongly protruding, distally pointed ; inner meta- tarsal tubercle large, strongly protruding; the outer a rounded cone; solar and palmar tubercles small but prominent; skin of dorsal and lateral surfaces of head, body and limbs with numer- ous small to moderate, prominent, rounded or elongate tubercles ; dorsum also marked with relatively short folds, the longest pair forming an urn-shaped pattern between the post-orbital and the axillary levels; venter posterior to the fore limbs, and the pos- terior surface of the thighs marked by prominent, relatively large, rounded tubercles. Color (in preservative) : Dorsum variable, grayish to black, mottled usually with a broad occipital blotch ; fore and hind limbs marked by light and dark transverse bars of about e(iual width; lips with dark bars; venter with brown flecks, heavily concentrated anterior to the fore limbs ; areolated light-dark pat- tern on inner and usually lower surface of thighs. Measurements of holotype (in mm) : Snout-vent length 37.7; length of head to posterior edge of tympanum 15.1 ; breadth of head 16.5; diameter of eye 5.6; diameter of tymi)anum 3.3; length of snout 5.9 ; length of hind limbs 60.5 ; length of tibia 12 BREVIORA No. 218 18.0 ; length of third tinger to base of second subarticular tu- bercle 5.0. Eggs: A small clutch of 10 eggs, stated to be of this species by Parker, were measured. In the preserved state they measure 3 or 4 mm in diameter. They are creamy white, without any indi- cation of pigment. Comparisons: Cormifer acrochordus is intermediate in size at maturity between C. macrops and C. papuensis weheri of species known from the Solomon Islands, and closest to C. macrops (Table 1). AVhen compared with extraterritorial species it is of about the same size as C. dorsalis from the Philippines and slightly smaller than C. gilliardi from New Britain. The sharply pointed and relatively short fingers distinguish C. acrochordus from known species of the genus other than C. aculcodactylus and possibly C. giUiardi. It differs from C. aculeodactyhis in its larger size (Table 1), much more warty skin, color pattern, larger eye relative to length of snout, and the broader head relative to snout-vent length (Table 3). It dif- fers from C. gilliardi in its smaller size, more pointed fingers, more warty skin, broader head, and larger eye (Table 3) ; the differences from C. macrops have been discussed in the section on that species (p. 10). AKTIFICIAL KEY TO SPE(^IES OF COBNUFEK KNOWN FEOM THE SOLOMON ISLANDS 1. Tips of fingers liroadl.v dilated, hreadth of disk of tliird finger more than 30 per cent of the length of the third finger as measured to the base of the second subarticular tubercle 2 Tips of fingers not or scarcely dilated, breadth of disk of third finger, if dilated, less than 20 per cent of the length of the third finger as measured to the base of the second subarticular tubercle ?> '2. Head relatively narrow, its breadtli usually less than 40 i)er cent snout- vent h'Ugth ; loreal region slightly or moderately oblique; eye large, its diameter nearly equal to length of snout nrcl'cri Head relatively broad, its breadth usually greatei- than 40 per cent snout-vent length; loreal region strongly obli(|ue; eye moderate, its diametei' ciiual to oi' slightly greater than the distance fi'oiu eye to nostril .'/"/'/'.'/' 3. First finger distinctly shorter than the second 4 First fiugei' loiigci- than (occasionally about ('(lual to) the second ;") 4. Snout-vent length of adults L'l) to .''II unn ; tips of fingers aiul toes idund, witli moderately dilated disks; fourtli finger longer than second ... mormps lill),") XKW FK()(iS FKOiSl THE SOLOMO.N ISLANDS I'S SiKiiii-vt'iit li'iijitli (if mliilts less tli;iii l^n nun; tijis of finjicrs .-iihI toos lilunt or slifjlitly i)oiiite(l, scarcely dilated; fonrtli fiiit^ei' usually shorter than oi- ahout eriual to second parlceri o. Tijjs of tintieis sliaiply jiointed; fourth finger usually shorter than the sc-iind when ad]iressed (i Ti])s of fingers lilunt or rounded; fourtli finger usually kinger than the second when adjiressed 7 0. Skill with nunu'rous itroiuinent warts and dorsal folds; solar area with numerous tubereles acltrucliordus Skin relatively smooth; solar area without tubercles acideodactylufi 7. Tills of fingers lilunt ly swolk'u, la(d\ing a niai'ginal groove (kdiiniting a veiiTial ]iad ; length of tiliia usually less than .111 jier cent of snout-vent length solonionis Tips of fingers with slightly dilated disks, a nuirginal groove delimiting a, ventral pad; length of tibia usually greater than 50 per cent of snout-vent length 8 8. Solar area with prominent tubercles; weli at base of toes not reaching ])roximal edge of suliarticular tuliercle on inner margin of second toe; dorsum, especially of adults, with numerous narrow folds l)apuensis weh,eri Solar area lacking prominent tuliercles; web at base of toes reaching midpoint of subartieular tul)ercle on inner margin of second toe; dorsum lacking numerous narrow folds myersi ACKNOWLEDGMENTS I wish to thank Dr. Alan Leviton, California Academy of Sciences, Dr. Doris Cochran, United States National Museum (USNM), Dr. Richard Zweifel, American Museum of Natural History (AMNII), Dr. Alice Grandison, British Museum (Natu- ral Hi.story), Drs. L. D. Brongersma and M. Boseman, Leiden Museum, for the opportunity of examining pertinent materials in the collections of their institutions; and Dr. Ernest Williams, Museum of Comparative Zoology, for suggesting that I describe these interesting frogs. The study of this genus of frogs is part of the author's pro- gram concerned with herpetofauna of the Pacific Islands. This program is sponsored by the National Science Foundation grant no. GB-409. Drawings are by Mr. AYalter Zawojski, Stanford University. 14 BREVIORA No. 218 LITERATURE CITED BOULENGER, G. A. 1918. Reinark^i uii tlii' liatracliiaii s'-'n^^'i':' Coniiifer Tsc-hudi, Plattj- mantis Giinther, Simomantis g. ii., and Staurois Cope. Ann. :\raR. Nat. Hist., (it) 1: 372-375. Brown, Walter C. 1952. Aiii])hil)ians of the Solomon I.slands. Bull. ^lus. Conij). Znol., 107: l-()4, pis. 1-8. Deckert, Kurt 1938. Beitrage ziir Osteologie und Systeniatik lanider Fruschluiche. Sitz.-Ber. Ges. Naturforscdi. Fr., Berlin, Jahrg. 1938; 127- 184. Inger, Robert F. 1954. Philippine zoological e.xpedition 194(5-1947. Systematies and zoogeogra))liy of Pliili])])ine Anijiliiliia. Fieldiana : Zool., 33: 183-531. Mayr, Ernst 19()3. Animal species and evolution. Harvard University Press, Cam- bridge, Massaehu-setts, xiv + 797 pp. Noble, G. K. 1931. Biology of the Aniiihihia. Mcdiaw-Ilill Hook Co., New York, xiii + 577 i)p. ZwEiPEL, Richard (J. 19fin. Results of the 1958-1959 Gillian! New Britain expedition. 3. Notes on the frogs of New Britain. Amer. Mus. Novit., No. 2023; 1-27. (Received 17 Decemlier, 19(i4.) 196.') XKW FROGS FROM THE SOLOMON ISLANDS 15 ■^•*fef FIGURE I FIGURE 2 FIGURE 3 FIGURE 4 FIGURE 5 PLATE 1 Fig. 1. Cornufer (/uppyi, inferior view of hand. Fig. 2. Cornufer necT:eri, inferior view of hand. Fig. 3. Cornufer p. parkeri n. sp., inferior view of hand. Fig. i. Cornufer aculeodactyJus, inferior view of hand. Fig. 5. Cornufer maerop-s n. si)., inferior view of hand. 16 BREVIORA Xo. 218 FIGURE I FIGURE 2 FIGURE 3 FIGURE 4 I 'LATH -2 Fiii. 1. Coniitfcr mi/crsi, inferior view of li.-iiid. Fig. 2. Cornufcr papurn^is inhrri, iiift'rioi' \ic\v of IkiikI. Fig. 3. Cnrnufcr .solo)iioni.s, infcMior view of IiiiikI. Fig. 4. Cornufcr acrocliordus n. sp., inferior view of hand. BREVIORA Meseenra of Coeipsirative Zoology Cambridge, Mass. May 7, 191)5 Number 219 THE EARLY EVOLUTION OF THE ECHINOZOA By H. Barraclough Fell Museum of Oomiinrntive Zoology INTRODUCTION The phylum Echinodermata is customarily considered to embrace two contrasted siibphyla. One of these, the Pelma- tozoa, comprises forms which are attached to the substrate for part or the whole of the life history, and which have a U-shaped gut, with the mouth and anus both directed upwards. The other, the Eleutherozoa, comprises free-living forms in which the mouth is directed downwards, and the anus (if present) is normally placed on the upper surface. The best known members of the Eleutherozoa are the sea urchins, the sea cucumbers, the star- fishes and the brittlestars. However, these included forms differ so widely that is has been a difficult task to elucidate their interrelationship and the probable nature of their presumed common ancestor. Further, there are strong grounds for sus- pecting that the subphylum Pelmatozoa includes some forms which are really related more closely to certain Eleutherozoa than they are to other members of the Pelmatozoa. These grounds are here set out, and it is proposed to abandon the Pelmatozoa as a formal classificatory division, and to adopt patterns of body symmetry as the main criteria for defining subphyla, in- stead of using habit and attitude. Recent morphological and paleontological studies have led to the conclusion (Fell, 1962, 1963a, 1963b) that the star-shaped members are interrelated and comprise a single grouping which may be regarded as a subphylum, and for which the name Asterozoa is already available. Similarly, other evidence implied that the globoid members were probably interrelated, and these have been associated as another subphylum, Echinozoa. Conse- quently, the so-called Eleutherozoa proved to be a polyphyletic 2 BREVIORA No. 219 assemblage, and the name should therefore be abandoned. The present contribution is aimed principally at clarifying the in- ferred interrelationships between the various classes of Echino- zoa, and adapting the current classification to reflect these rela- tionships. The discovery of the hitherto unknown class Helicoplacoidea (Durham and Caster, 1963) has shown that primitive, free-living echinoderms, with characters intermediate between those of Echi- noidea, Holothuroidea and Edrioasteroidea, had already dif- ferentiated in early Cambrian times. The morphological char- acters of the Helicoplacoidea suggest a relationship to some common ancestral stock from which arose, on the one hand, the eleutherozoan Echinoidea and Holothuroidea, and on the other hand, the pelmatozoan Edrioasteroidea. Hitherto, the pelmato- zoan echinoderms have generally been thought to represent a single natural assemblage, the subphylum Pelmatozoa, but this concept now becomes suspect. Further grounds for doubting the validity of subphyla based on eleutherozoan or pelmatozoan habit are provided by recent work on Paleozoic sea-stars, from which it has become evident that the subphylum Eleutherozoa, erected to comprise the free-living echinoderms, consists actually of two entirely separate stocks, the Echinozoa and Asterozoa (Fell, 1962, 1963a). The Echinozoa represent an ancient, pre-crinoid stock, of which the modern representatives are the Echinoidea and Holothuroidea ; whereas the Asterozoa are of relatively late origin, derived from a pinnulate pelmatozoan stock, provisionally identified with Crinoidea. Some analogous results emerge from recent work by Ubaghs (1961), from which it is apparent that the lower Paleozoic Homalozoa comprise a stock of asymmetrical, or bilaterally symmetrical, echinoderms, some members of which were free-living (i.e., eleutherozoan), while others were stalked and attached to the subtrate (i.e., pel- matozoan). Lastly, data given later in this paper imply the essentially archaic character of the dendrochirote orders of Holo- thuroidea, and point to possible relationships between these forms and the Cambrian Helicoplacoidea. Certain parallels be- tween the dendrochirote psolid holothurians, on the one hand, and the Edrioasteroidea, on the other, serve also to reinforce suspicions that the Edrioasteroidea should be classified with the echinozoan echinoderms, and not with the so-called Pelmatozoa, where they are commonly placed. Indeed, this inference is already implicit in a phylogenetic diagram published by Fell (1962). 1965 EARLY EVOLUTION OF THE ECIIINOZOA 3 PATTERNS OF SYMMETRY Four structural patterns may be contrasted in echinoderms; these are : (1) Homalozoan pattern, seen in those early Paleozoic echino- derms in which the skeletal plates are arranged either asym- metrically, or with more or less bilateral symmetry. These forms have been assigned to a separate subphylum, the Homalozoa (Whitehouse, 1941; Ubaghs, in press). (2) Echinozoan pattern, seen in the Helicoplacoidea, Holo- thuroidea, Echinoidea, Ophiocistioidea and Edrioasteroidea, all essentially globoid forms lacking arms, with meridional sym- metry. The Echinoidea and Holothuroidea were placed by Zit- tel (1895) and Jaekel (1918) in a subphylum Echinozoa, and the same name may be retained in a more extended sense, to com- prise all the classes listed here. (3) Crinozoan pattern, seen in the pelmatozoan classes Eocri- noidea, Paracrinoidea, Cystoidea, Blastoidea, Edrioblastoidea, and Crinoidea, initially globoid forms with partial meridional symmetry, but acquiring radially divergent systems of ambula- cral feeding appendages (brachioles or arms). These groups, together with the Edrioasteroidea and some dendrochirote Holo- thuroidea, exhibit a sessile habit, involving certain morphological features normally utilized in defining a subphylum Pelmatozoa. However, although the sessile holothurians have never been grouped, the so-called Pelmatozoa cannot be defined so as to in- clude the one without the other. It is evident that two categories of diagnostic criteria have been intermingled, and a more critical definition is required. (4) Asterozoan pattern, seen in the Somasteroidea, Asteroi- dea and Ophiuroidea, in which radial divergent axes of sym- metry produce arms, and the earliest morphological features of the arms correspond to those seen in pinnulate Crinoidea. These taxa fall within the subphylum Asterozoa, as defined by Zittel (1895), Jaekel (1918), and at greater length by Fell (1963a). It will be noted that whereas categories (1), (2), and (4) above appear to be natural groupings, and offer no diagnostic difficulties, some unsatisfactory features arise under category (3). These are now examined, in the light of evidence supplied by the other three groupings. 4 BREVIORA No. 219 BLEUTHEROZOAX AND PELMATAZOAN TRENDS In each of the subphyla Homalozoa, Echinozoa and Asterozoa, irrespective of the pattern of symmetry adopted, two mutually opposed evolutionary trends may be observed, fundamentally governed by the attitude which the animal adopts with respect to its habitat. These are: (a) Qeutherozoan tendencies, that is, adoption of a free- living habit, in which the animal acquires locomotor mechanisms permitting it to seek out food wherever it is to be found, by browsing on available algae, preying upon other animals, or swallowing large quantities of mud for the sake of its slight organic content. Such tendencies are invariably accompanied by the evolution of jaws, or of some special oral appendages adapted to gross (macrophagous) feeding. The anus, if de- veloped, tends to lie on a part of the body remote from the mouth. (b) Pelmatozoan tendencies, that is, adoption of a sessile habit, by which the animal becomes attached more or less per- manently to the substrate, either by the aboral surface itself or by an aboral stalk. Locomotor organs are atrophied or lost alto- gether, and the animal is then dependent upon such planktonic sources of food as the sea-currents may provide. It secures the food by some ciliary or comparable advective mechanism medi- ated by the tube-feet, the food particles being conveyed to the mouth by food grooves on the upper surface, the nutrition being therefore, of the microphagous type. The mouth and anus neces- sarily both lie on the upper surface, and the alimentary canal is consequently bent into a U-shape in the vertical plane. Al- though the modifications are here considered only in the context of echinoderms, analogous features, of course, occur in other phyla with sessile members. It may be noted here that radial symmetry is by no means a consequence of the adoption of sessile habits. On the contrary, echinoderms which already possess radial symmetry, if they adopt a sessile habit, may acquire a strongly marked bilateral symmetry, very similar to that acquired by the sessile tunicates, for example, with which the psolid holothurians were once confounded. Further, the discovery of Hclicoplacus (Durham and Caster, 1963) implies that the echino- zoan echinoderms were already free-living forms hefore radial symmetry was fully developed, and that no subsequent sessile stages supervened between the Cambrian Helicoplacoidea and 1965 EARLY EVOLUTION OF THE ECIIINOZOA 5 their presumed Ordovician successors, which include the earliest known echinoids and holothurians. It would appear that eleutherozoan and pelmatozoan ten- dencies are not directly related to the pattern of symmetry of the body in echinoderms, and that the two categories of evolu- tionary change, namely body symmetry and habit, have operated as simultaneous variables. This may be illustrated by reference to the echinozoan classes. EVOLUTION OF THE ECHINOZOA The oldest known echinozoan is the lower Cambrian Helico- placus, in which the body is fusiform in shape, with the mouth at a broad anterior end, and the anus at the tapering opposite extremity (Fig. 1). The skeleton comprises numerous quad- rangular or lozenge-shaped plates, sometimes bearing a rigid, erect spine, and disposed in counter-clockwise helical series. A single band of smaller platelets winds in a helix (sometimes bi- furcated) around the body, and evidently indicates the position occupied by a single external ambulacral water-vessel. The symmetry would appear to be bilateral, therefore, but distorted by the counter-clockwise torsion, and combined with an apparent radial symmetry displayed by the arrangement of the thecal plates. Similar torsion is observable in the earliest Echinoidea, notably Eothuria in the Ordovician; here, however, the sym- metry is overtly radial, or more correctly meridional, for there are now five ambulacra, disposed at regular intervals of 72°, forming twisted meridians. Analogous torsion is seen in the Edrioasteroidea, but has not yet been reported from Holothuroi- dea or Ophiocistioidea. The torsion was eventually lost in the echinoid line, but it persisted in the edrioasteroids until their extinction in the Carboniferous. As already noted above, eleutherozoan and pelmatozoan trends have arisen independently from time to time in the various groups of echinoderms. Within the Echinozoa, the classes Echi- noidea and Ophiocistioidea are not yet known to have produced any sessile forms. The initial echinozoan stock, to judge by the Helicoplacoidea, was itself free-living, too. Helicoplacus must have been a motile, bottom-feeding echinoderm, resembling a plated dendrochirote holothurian, as suggested below. The skele- tal plates formed a complete, robust though flexible test. The 6 BREVIORA No. 219 varying degrees of expansion and contraction reported in fos- sils by Durham and Caster (1963) imply an underlying muscula- ture, able to operate concertinawise. Thus, Helicoplacus prob- ably crept over the sea floor, like an annelid. The fossils occur in a fine clastic matrix, implying that the habitat was mud; Helicoplacus was probably a gross mud-swallower, like many aspidochirote holothurians. The primitive state of the ambula- crum suggests that the organ may have carried, at best, only rudimentary tube-feet, which could hardly have been more than respiratory organs, and probably Avere only sensory tentacles, like the dorsal tube-feet of many holothurians. The complete plating of the body-wall, and the probably rudimentary form of the tube-feet, imply the lack of an effective respiratory mech- anism on the outer surface of the body. If this is correct, then it may be inferred that rectal respiration was required, either of the pulsatory crinoid type, or by means of respiratory trees, as in holothurians. Study of the distribution of respiratory trees among holothurian orders suggests a direct relationship to the habits of these animals, and also implies that the earliest holothurians had already developed these structures. It therefore seems likely that respiratory trees of a rudimentary type were present in Helicoplacoidea. The earliest Echinoidea, such as £^o#/i?(na, possessed a multi- plated, flexible, spirally twisted body wall, similar to that of the Helicoplacoidea, and perhaps inherited from a helicoplacoid ancestry. They, however, had five well-developed ambulacra, on which the meridional water-vessels lay as external structures, though with internal ampullae for the tube-feet. Structural de- tails of the ambulacral pores show that the tube-feet were large, and probably suctorial — certainly extensile and muscu- lar. They would, therefore, serve the double function of loco- motor organs and respiratory organs, as in modern echinoids. The fossils exhibit a developed jaw mechanism, showing that the early echinoids were already capable of feeding in the manner of their extant endocyclic descendants, that is to say, by biting and grinding organisms in the substrate, and chewing algae. This type of feeding demands an eleutherozoan habit, and con- traindicates any pelmatozoan tendencies, since an animal with such feeding mechanisms would rapidly starve if it adopted a sessile manner of life. The Ophiocistioidea developed a rigid skeleton by soldering of adjacent plates of the test, in much the same manner as in the 1965 EARLY EVOLTTTION OF TITE ECfllNOZOA Figures 1-5. Archaic types of placoid echinozoans. 1, Helicoplacus lat- eral aspect, X 3. 2, Ypsilothuria lateral aspect, X 3. 3, Placothuria lateral aspect, X 2. 4, Isorophus actinal aspect, X 3. 5, Lepidopsolus actinal aspect, X 2. Fig. 1 from Durham and Caster, 1963; Figs. 2, 3, 5, drawn by D. L. Pawson, Fig. 4, from Kesling and Mintz, 1960. later echinoids. Locomotion, however, was effected by the use of the grossly enlarged and plated oral tube-feet, so characteristic of the class. The enlarged oral tube-feet on the lower surface would also subserve the function of nutrition, by sweeping up detrital 8 BREVIORA No. 219 material, and cramming it into the mouth, which was directed downwards. The anus, as in the endocyclic echinoids, lay on the upper surface, though not at the apical pole. Here, as in the endocyclic echinoids, the feeding and locomotor habits imply an eleutherozoan mode of life, and no sessile forms are known to have developed. The early Ilolothuroidea are known at present only from isolated skeletal plates. However, on the basis of recent studies by Pawson (1965) it would appear extremely probable that the Ordovician and later Paleozoic holothurians resembled the extant Ypsilothuriidae (Fig. 2). Further, when once the dendrochirote tentacle had been evolved, they would resemble the extant genus Paracucumis, or PlacotJiuria (Fig. 3). These are all heavily plated forms, with a complete test, flexible, made up of large plates with or without rigid spinous processes. The early holo- thurians would also be comparable with Helicoplacus, and with the flexible-bodied Ordovician echinoids, such as Eofhuria (the latter genus having originally been regarded as a holothurian). Suctorial tube-feet may have been lacking from the earliest holothurians, to judge by their rudimentary state in extant plated genera, though this is uncertain. If they were initially lacking, then locomotor movements would have been effected by contraction and expansion of the body wall and its flexible test. Once suctorial tube-feet had developed, locomotion on the echi- noid plan would be possible. There is morphological evidence that some kind of jaw apparatus, resembling the echinoid lan- tern, was developed early in the holothurians. Apparently it was abandoned once the dendrochirote tentacle had developed, but the skeletal elements of the presumed lantern acquired a new purpose — the attachment of the radial (and retractor) muscles — and hence the organ persisted in later holothurians as the calcareous ring on the pharynx. It is probable, especially from data given by Pawson (1965), that the dendrochirote tentacle evolved from a formerly simple oral tube-foot. Repeated dichotomy led to the complex dendritic tentacle of the Dendrochirotida. This is an efficient collecting organ for planktonic material which is conveyed to the mouth by the contractions of the tentacles, ciliary action, and the "spooning" action of two ventral tentacles. Dendrochirote holo- thurians, whether motile or not, are able to trap sufficient nour- ishment by filtering the surrounding sea water, provided there are currents replacing the surrounding water, and bringing fresh 1965 EARLY EVOLUTION OP THE ECUINOZOA 9 supplies of food particles. From such deudrochirote holo- tliurians more than one line of evolution is possible, for they have the means of adopting either eleutherozoan or pelmatozoan habits. If the locomotor system is retained, the oral tentacles can be adapted to serve as food-collecting- organs operating in various ways. In the Cucumariidae, for example, the body may be held erect, attached only by the posterior tube-feet, whilst the tentacles are spread out in a ring around the mouth, which is directed upwards, so that the habit of the animal resembles that of a sea anemone. In holothurians which adopt the horizontal attitude (lying on the ventral side where the tube-feet are re- tained), the tentacles can readily evolve from deudrochirote to aspidochirote forms, thus permitting gross mud-swallowing, and a markedly eleutherozoan habit. A further possibility is for the locomotor system to fulfill a purely adhesive role, leading to a sessile (pelmatozoan) habit. This is illustrated by the psolid dendrochirotes, in some genera of which the body is converted into a flattened, limpet-like form, adhering by a broad, flattened ventral surface, applied to a firm substrate (Fig. 5). The exposed dorsal and lateral surface is covered by a test of imbricating, robust plates. The mouth and anus lie on the upper surface, often protected by valvate plates, similar to those of cystoids or edrioasteroids. The whole body, in fact, is comparable to that of an edrioasteroid, the only dis- tinction being the lack of external ambulacral plates (Figs. 4, 5). Thus the psolids are actually closer to edrioasteroids in morphological features than to many holothurians, or even to eehinoids, and the only character by which they differ from edrioasteroids is the same as that which distinguishes them from eehinoids — namely, the fact that the water-vessels are internal, and the test consequently does not form ambulacral plates. Hence a comparison of Psolus with an edrioasteroid illustrates two important features: (1) Edrioasteroids have the same pat- tern of symmetry as Echinozoa, and have evidently arisen from an early echinozoan stock similar to that which gave rise to the deudrochirote holothurians, of which the psolids are obvious members. (2) The pelmatozoan characters of edrioasteroids and of psolid holothurians have arisen as a direct habit response to adoption of a sessile mode of life, and do not indicate any near relationship to such pelmatozoan groups as are customarily in- cluded in the subphylum Pelmatozoa. Comparative study of the internal skeleton of the pliaryngeal 10 BREVIORA No. 219 region of dendrochirote holothurians suggests to me that the original holothurians must have had external ambulacra formed by modified plates of the test and similar to those seen in the fossil edrioasteroids and illustrated by Kesling and Mintz (1960) (Fig. 7). Apparently, when once evolved, the large dendrochi- rote tentacles required a protective mechanism, by which they could be withdrawn into the body. The protection was achieved by telescoping of the anterior end (termed the introvert), capable of retraction under the action of muscles derived from the radial muscle group. The evolution of the introvert, in turn, implied the conversion of the original external ambulacral areas of the test into internal structures, surrounding the pharynx, and serving for the insertion of the retractor muscles. In primi- ive dendrochirotes the pharyngeal skeleton is still recognizable as equivalent to the ambulacral plates of an edrioasteroid, but in most surviving holothurians the mechanism is very reduced or vestigial. Stages in the reduction are illustrated in Figures 6 and 8 to 12. Inferred homologies of edrioasteroid and dendro- chirote skeletal elements are indicated in the captions to these figures. The Edrioasteroidea (Figs. 4, 7) adopted a similar habit to that of psolid Holothuroidea, but the ambulacra remained ex- ternal, instead of sinking inwards, and this permitted an alter- native method of feeding, suited to a pelmatozoan way of life. The feeding mechanism was provided by the whole complex of tube-feet. The five ambulacra grew outwards from the mouth as meridians, but only on the upper surface. Each ambulacrum Figures 6-12. Ambulacral plating systems and calcareous ring elements in holothurians (6, 8-12), and an edrioasteroid (7), considered in this paper to be homologous structures. 6, Pentadactyla (Dendroehirotida), X 3. 7, Isorophus (Edrioasteroidea), X 4. 8, Placothuria (Dendroehirotida), X 4. 9, Neothyonidium( Dendroehirotida), X 2. 10, Psolus (Dendroehirotida), X 3. 11, Euthyonidiella (Dendroehirotida), X 3. 12, Mitsukuriella (Dendro- ehirotida), X 4. Abbreviations: a.n., anterior notch of radial plate; g.d., gonoduct; ir.p., interradial plate; mad., madreporite; viad. d., stone canal; m.f., radial water-vessel; oes., esophagus; p.p.r., posterior process (considered in this paper to be the distal ambulacral plates) ; ph., pharynx; p.v., polian vesicle; r.m., retractor muscle; r.p., radial plate (here considered to be an ambulacral element). Figs. 6, 8, 9, from Pawson, 1963; 7, from Kesling and Mintz, 1960; 10, drawn by D. L. Pawson; 11, 12, redrawn from Heding and Panning, 1954. 1965 EARLY EVOLUTION OF THE ECIIINOZOA 11 an nm. 10 12 BREVIORA No. 219 carried a median groove, bordered on either side by the tube- feet. The latter must have functioned in a manner similar to the tube-feet of sea lilies, that is, they would wave about in the water, trapping small plankton and organic particles on their mucous surface, and sweeping them inwards towards the mouth, along the food-groove on the ambulacrum. Evidently no intro- vert evolved, and this implies that dendrochirote tentacles never developed. It may be inferred that the Edrioasteroidea, soon after their differentiation from the initial echinozoan stock, adopted the sessile habit but, unlike the dendrochirote holothurians, had no oral tentacles on which to rely for nutrition. The ambulacra as a whole, therefore, took on the function of nutrition, mediated by the tube feet; and the advective food grooves, and the ambu- lacral plates on which they were carried, were a natural conse- quence. In the holothurian line, the radial water vessels were early converted into internal canals, as also occurred in the post- echinocystitoid echinoids. Hence the adoption of pelmatozoan habit inevitably demanded a pre-existing dendrochirote nutritive mechanism, and never involved external ambulacra in holo- thurians. As Bassler (1935) has pointed out, the earliest (i.e. mid-Cambrian) edrioasteroids retain a fully-plated ventral sole, unlike their later derivatives, and this may be taken as evidence that the test was originally spherical in edrioasteroids. The am- bulacra in the early edrioasteroids were also more simple than in later forms, and most of the so-called pelmatozoan features, such as cover-plates, may well be later specializations, analogous to the development of cover-plates in crinoids and somasteroids. Comparison of dissections of psolids with edrioasteroids sug- gests certain inferences as to the internal anatomy of edrioaster- oids. In the absence of any evidence to the contrary, we may assume that the edrioasteroids had a gonad placed in the dorsal interradius. In psolids the gonopore lies on the introvert, just posterior to the mouth. The corresponding position on an edrio- asteroid would be that in which a pore is known to occur, but the pore has hitherto been supposed to be a hydropore. Psolids, however, respire (at least in part) by rectal respiratory trees. It seems probable that respiratory trees would be required by edrioasteroids also, and that the hydropore would have been internal, as it is in dendrochirote holothurians. Irrespective of these latter inferences, the main conclusion emerges that edrioasteroids should be treated as members of 1965 EARLY EVOLUTION OF THE ECIIINOZOA 13 the Echinozoa aud that their pelmatozoan features are no more fundamental than the same features in psolid holothurians, being purely secondary responses to the demands of sessile habit. The Echinozoa have here been selected to illustrate an argu- ment which could also be developed on the basis of evidence pro- vided by other groups of echinoderms. The hitherto puzzling features of Homalozoa, for example, some being apparently pel- matozoans, others eleutherozoans, would appear to be no more unusual than the circumstances found within the Echinozoa. Without prolonging the discussion at this stage by reference in detail to other subphjda, the inference may be drawn that the eleutherozoan and pelmatozoan characters have arisen independ- ently, and at different times, in various echinoderm groups, and it is not possible to devise a natural classification on the basis of such characters alone. A revised classification now emerges in which the Edrioasteroidea are transferred to the subphylum Echinozoa. When this transfer is made, the residual pelmatozoan classes prove to comprise a much more uniform assemblage which may be defined, not in terms of their habit, but in terms of their morphological symmetry, as follows : Subphylum Crinozoa Matsumoto, 1929 (redefined) Fundamentally globoid echinoderms with partial meridional symmetry tending to produce an aboral calyx, the ambulacra de- veloping as aboral semi-meridians later forming radially diver- gent systems of ambulacral feeding appendages w^hich take the form of brachioles or arms. Included classes : Cystoidea, Eocrinoidea, Paracrinoidea, Blas- toidea, Edrioblastoidea and Crinoidea. The following extended diagnosis of the Echinozoa also emerges : Subphylum Echinozoa Zittel, 1895 (redefined) Fundamentally globoid echinoderms which never develop arms. In earliest members mouth and anus lay at opposite ends of the body, but in some later forms these have become sec- ondarily displaced. Meridional w^ater-vessels traverse the body wall in the direction of the anus, the vessels lying originally on the surface, but sinking into its substance in later Paleozoic and all post-Paleozoic groups. Skeleton, nervous system, repro- ductive organs and muscular system tend to differentiate into meridional systems, although an underlying bilateral symmetry is always evident. 14 BREVIORA No. 219 Included classes : Helieoplacoidea, Edrioasteroidea, Ophiocisti- oidea, Holotliuroidea, Echinoidea. We can now set out a general classification of the phylum Bchinodermata, incorporating the proposals relating to the eleutherozoan groups already made in an earlier paper (Fell, 1963a), together with the results of the present investigation. These are given in Table 1. It will be noted that uniform ter- minations in -zoa are adopted for subphylum categories, whilst class names end in -oidea. Conventional Classification Classification Herein Adopted Subphylum Pelmatozoa Class Carpoidea Class Cystoidea Class Eocrinoidea Class Paracrinoidea Class Blastoidea Class Edrioblastoidea Class Crinoidea Class Edrioasteroidea Subphylum Eleutherozoa Class Ophiocistioidea Class Echinoidea Class Holothuroidea Class Asteroidea Class Ophiuroidea Subphylum Homalozoa Class Carpoidea Subphylum Crinozoa Class Cystoidea Class Eocrinoidea Class Paracrinoidea Class Blastoidea Class Edrioblastoidea Class Crinoidea Subphylum Echinozoa Class Helieoplacoidea* Class Edrioasteroidea Class Ophiocistioidea Class Echinoidea Class Holothuroidea Subphylum Asterozoa Class Stelleroidea (including the subclasses Somasteroidea, Asteroidea and Ophiuroidea) ^Helieoplacoidea unknown before 19G3. Table 1. Comparison of the conventional classification (left) of Echino- dermata with that herein adopted (right). The new arrangement of the classes in four subphyla is believed to avoid the polyphyletic categories which have been demonstrated in the conventional classification, and it at- tempts to reconcile the systematic treatment with the evidence of the fossil record. ACKNOWLEDGMENTS In preparing this paper I was assisted by having access to then unpublished data on the morphology of dendrochirote holo- thurians assembled bv Br. David L. Pawson, at the time when 1965 EARLY EVOLUTION OF THE ECHINOZOA 15 we were working together on South Pacific echinoderms; these data are presented elsewhere (Pawson, 1965). A revised classifi- cation of Holothuroidea recently proposed (Pawson and Fell, 1965) synthesizes the sj'^stematic outcome of our separate con- tributions. A further contribution, incorporating some material <^^W?i^ z < < Volchovio V \ \ OPHIOCISTIOIDEA BothriocidarisEothuria \ \ / ECHINOIDEA Thuroholia Cystaster / / HOLOTHUROIDEA / / Edriooster Walcottidiscus Helicoplacus HELICOPUkCOlDEA / (^ Stromctocystltes EDRIOASTEROIDEA Figure 13. Approximate inferred phylogeny of the Echinozoa as deduced from evidence discussed iu this paper. 16 BREVIORA No. 219 added by Professor Raj'mond C. Moore, University of Kansas, will be given in the forthcoming Treatise on Invertebrate Paleon- tology, volume U, where the material of this paper is given a formal text-book presentation (Fell and Moore, 1965) ; for assist- ance in re-lettering Figure 13, I am grateful to Dr. Moore. I have to thank Dr. Ernst Mayr for reading and criticizing the manuscript, and for suggesting the inclusion of Table 1 ; Dr. David Pawson for assistance in illustrating holothurians, and for permission to use Figures 2, 5, 6 and 8 ; Dr. Porter M. Kier for some suggestions ; and other colleagues and students who have discussed the interpretations offered here, and who have all to some extent influenced the manner of presentation here adopted. SUMMARY The morphology of fossil and extant Echinodermata implies that pelmatozoan and eleutherozoan characters must have arisen independently in the various classes at several different times since the Pre-Cambrian. Hence it is not possible to base a natu- ral classification on these two categories of contrasting characters alone ; for sessile and free-moving echinoderm assemblages each comprise two or more unrelated stocks, the similarities of which are due only to convergent evolution. At least four subphyla may be defined on the basis of four recognized patterns of growth and symmetry, mediated by the hydrocoel, evidently as innate trends initially independent of the environment. At least three of the four subphyla include pelmatozoan and eleutherozoan stocks, which are here regarded as arising as simple habit responses to the habitat. Edrioasteroidea are interpreted as members of the subphylum Echinozoa, and should be removed from their present conven- tionally accepted position among the so-called Pelmatozoa. The latter assemblage, after removal of the Edrioasteroidea, is re- defined and assigned the rank of subphylum, under the name Crinozoa. The so-called calcareous ring of holothurians is here interpreted to be the homologue of the ambulacra! plating system of edrioasteroids, telescoped within the body, following the evo- lution of the introvert. 1965 EARLY EVOLUTION OF THE ECHINOZOA 17 EEFERENCES CITED Bassler, E. S. 1935. The classification of the Edrioastcioidea. Smithson. Misc. Coll., 93 (8): 1-11. Durham, J. W. and K. E. Caster 19G3. Helicoplacoidea, a new class of echinoderms. Science, 140: 820- 8-2-2. Fell, H. B. 1962. A classification of echinoderms. Tuatara, 10 (3), 138-140. 1963a. The phvlogeny of sea-stars. Phil. Trans. Eoy. Soc. London, B, 246: 381-435, figs. 1-18. 1963b. The evolution of the echinoderms. Ann. Eept. Smithson. Inst, for 1962 : 457-490, pis. 1-3. Fell, H. B. and E. C. Moore 1965. General features and relationships (of echinozoans). Treatise on Invertebrate Paleontology, U: (in press). Jaekel, 0. 1918. Phylogenie und System der Pelmatozoen. Pal. Zeitschr., Berlin, 3: 1-128. KESLING, R. V. AND L. W. MiNTZ 1960. Internal structure in two edrioasteroid species. Univ. Michigan Contrib. Paleont., 15: 313-348, figs. 1-4, pis. 1-9. Matsumoto, H. 1929. Outline of a classification of the Echinodermata. Tohoku Imp. Univ., Sei. Eepts., ser. 2( Geol.), 13 (2): 27-33. Pawson, D. L. 1963. The holothurian fauna of Cook Strait, New Zealand. Zool. Pubis. Victoria Univ., Wellington, 36: 1-38, pis. 1-7. 1965. Phylogeny and evolution of holothuroids. Treatise on Inverte- brate Paleontology, U: (in press). Pawson, D. L. and H. B. Fell 1965. A revised classification of the dendrochirote holothurians. Brevi- ora, Mus. Comp. Zool., No. 214: 1-7. Ubaghs, G. 1961. Sur la nature de I'organe appeletige ou pedonele chez les car- poides Cornuta et Mitrata. C. E. Acad. Sci. Paris, 253: 2738- 2740, figs. 1-5. Whitehouse, F. 1941. The Cambrian faunas of north-eastern Australia. Mem. Queens- land Mus., Brisbane, 12: 1-28, figs. 1-9, pis. 1-4. Zittel, K. a. von 1895. Grundziige der Palaeontologie (Palaeozoologie). Munich, 971 pp. (Eeceivcd 18 January, 1965.) BREVIORA Mmseminti of Comnpsirative Zoology Cambridge, Mass. May 7, 1965 Number 220 A NEW SPECIES OF ELEUTHERODACTYLUS FROM GUADELOUPE, WEST INDIES By John D. Lynch Department of Zoology and Museum of Natural History University of Illinois Urbana, Illinois The numerous Greater Antillean frogs of the genus Eleu- therodactylus have, for the most part, been studied as groups and in faunal studies. Cochran (1941) studied the forms occurring on Hispaniola, and Lynn and Grant (1940) studied the Ja- maican forms. In a series of papers, Schwartz (1957, 1958a-d, 1959a, b, and 1960) has clarified the status of the forms occurring on Cuba. The frogs of this genus in the Lesser Antilles are less well understood. Five names have thus far been applied to the frogs occurring there. These are : E. martinicensis (Tschudi), E. johnstonei Barbour, E. lentus (Cope), E. aiitillensis Reinhardt and Lutken, and E. darludensis (Auffenberg). The last form was described as an extinct Hyla by Auffenberg (1958) but I have recently shown (Copeia, in press) that it is an Eleuthero- dactylus and is probably not extinct. My studies on the osteology of these frogs and the researches of Albert Schwartz, who is currently revising the Lesser Antillean Eleutherodactylus, show that in reality there are numerous forms of this genus occurring on the Lesser Antilles. James Lazell, Jr. and Patrice Barlagne collected two forms at Matouba, north of Basse Terre, La Guadeloupe, which they could distinguish by voice and habit of calling. On external features, however, they are nearly identical. But from Lazell's field notes and discussions with him it appeared that two species were involved. Inasmuch as I had had good fortune in separat- ing some of the other Eleutherodactylus on neighboring islands by the structure of their pelvic osteology, specimens of these BREVIORA No. 220 forms were macerated, and strong differences were found be- tween the ilia of the two forms. These differences are as great or greater than those between any other of the forms of this genus occurring on Antigua, Barbuda, Martinique, St. Kitts or Grenada. In view of this, as well as the minor external differ- ences, and the call difference (which must be an important isolat- ing mechanism), it is evident that there are two species occurring in the vicinity of Matouba. One of these, the larger, also occurs on Martinique and is very probably E. martinicensis. The second species is apparently cryptic (although not sibling). It is here named for M. Patrice Barlagne, who collected the majority of the specimens and aided Lazell in collecting on the Souffriere-Saus Toucher massif of La Guadeloupe. Eleutherodacttlus barlagnei sp. nov. Holotype. Adult female, MCZ 35334, collected by Patrice Bar- lagne and James Lazell, Jr., at Matouba, La Guadeloupe, ca. 700 meters elevation, on 17 August, 1961. FiGUEE 1. (Left) Eleutherodactylus martinicensis MCZ 35322; top, side of head; bottom, roof of mouth. (Right ( EleutTierodactylus harlagnei sp. nov., holotype MCZ 35334 ; top, side of head, bottom, roof of mouth. 1965 NEW GUADELOUPEAN FROG 3 Faratijpes. MCZ 35330-33 (4), same data as for holotype. MCZ 35331 is a skeleton. Diagnosis. An Eleutherodactylus apparently related to E. martinicensis and separable from that species and all other Lesser Antillean species of the genus by the following combina- tion of characters : head narrow, not wider than body ; tympanum small, hidden dorsally, a small tubercle in the area of the hidden tympanic membrane ; tympanum separated from commissure of jaws by three-quarters to more than the horizontal diameter of the tympanum ; choanae completely visible from below ; vom- erine tooth bosses triangular in outline, within the borders of and posterior to the choanae ; the voice assumed to be that of the new form (since the species could not be separated on other bases in the field) is described by Lazell as "Teeen." Description of holotype. Adult female (see Fig. 1) : head very slightly broader than long; head not broader than body; eyes small, width of eyelid less than interorbital distance; can- thus rostralis distinct, not sharp ; loreal region slightly concave, sloping sharply to lip ; nostrils closer to tip of snout than to eye, area around them swollen; length of eye greater than distance from eye to nostril ; tympanum small, about one-third diameter of eye, upper portions hidden, not distinguishable ; tubercle pres- ent in area where upper rim of tympanum should be ; distance from tympanum to commissure of mouth slightly greater than hor- izontal length of tympanum ; anterior edge of tympanum from eye about one and one-half times horizontal width of tympanum ; no supra- or post-tympanic fold present. Tongue oval, free for about one-half its length; no vocal sac or slits; choanae not concealed by rim of upper jaws, small, round, slightly smaller than area of a vomerine tooth boss ; vom- erine tooth bosses between and posterior to choanae, triangular in outline, separated by a distance about equal to width of a single boss. No axillary membrane ; no tubercles or ridges on arm ; palmar tubercles small ; supernumerary tubercles on palms very faint or lacking; subarticular tubercles large, rounded, simple; lateral fringe present on fingers ; no webbing between fingers ; order of finger length, shortest to longest, 1, 2, 4, 3 ; circular disks on fingers, somewhat like pads of hylids in lateral view, notch pres- ent ; no tarsal fold or tubercles ; inner metatarsal tubercle small, elongate; outer metatarsal tubercle faint, but large; no super- numerary tubercles on soles ; subarticular tubercles large, round, BREVIORA No. 220 simple ; lateral fringes on toes ; pads on toes like those on fingers, but smaller; faint webbing on toes except for web between toes 3 and 4 which is clearly visible for about one-half of toe 3 ; legs short, heels do not overlap when flexed legs are held at right angles to body ; heel of adpressed hind limb extends to mid-eye. Figure 2. (Top) Eight ilium of Eleutherodactylus martinicensis, MCZ 35321. (Bottom) Right ilium of Eleutherodactylus harlagnei sp. nov., para- type, MCZ 35331. 1965 NEW GUADELOUPEAN FROG 5 Skin of dorsum and sides smooth, that of belly and ventral and posterior surface of thighs granular. Color in alcohol. Dorsum nearly uniform red-brown with scattered faint reticulations of black. Bands are evident on the limbs. Those of thigh are broad, three on each side. Those of tibia are narrower and three on each side. Two bands are present on tarsus and foot and two on the forelimbs. The venter is a dusky brown with small lighter spots. The uudersurfaces of the forelimbs and the area across the chest lack the dusky brown pigment and are yellow. Measurements in mm. Snout to vent 33 ; width of head 10.2 ; length of head 9.8 ; horizontal length of tympanum 1.2 ; length of eye 3.6 ; eye to nostril 3.2 ; eyelid width 2.2 ; interorbital dis- tance 2.6; length of tibia 13.0. Variation. The four paratypes are quite similar in appearance. In all specimens the venter is grey-cream with cream spots on chin and throat. The legs are flecked with brown pigment giving the appearance of cream flecks. All specimens have a light tri- angular interorbital spot, although it is weak in the holotype. Two specimens have a dark chevron on the dorsum (MCZ 35330, 35333). None of the type-series shows the wide or narrow ver- tebral stripe, although this variation is seen in examples of E. martinicensis collected with the type-series. This species has a narrower head than does E. martinicensis. The head width/body length ratio ranges from 0.27 to 0.32 with a mean of 0.30, whereas of twenty-four E. martinicensis taken at the type locality by Barlagne and Lazell, the ratio varies from 0.32 to 0.41 with a mean of 0.33. Comparisons. While morphologically E. harlagnei is quite similar to E. martinicensis, there are several differences. E. harlagnei has a dark venter, the tympanum appears smaller and is farther from the mouth and the vomerine tooth bosses are triangular in outline, not elongate, and do not extend laterally beyond the inner borders of the choanae (in E. martinice'usis the bosses extend laterally as far as the outer borders of the choanae). E. urichi has less prominent vomerine tooth patches which are round. E. johnstonei has shorter limbs and the choanae are not completely visible when the roof of the mouth is viewed from directly below. E. harhudensis has elongate vomerine tooth patches. A further difference can be noted between these frogs. In 6 BREVIORA No. 220 connection with a study regarding the identity of Hyla harbu- densis (= Eleutherodactylus harhudensis) , I prepared skeletons of the two species found at Matouba. E. harlagnei is very dis- similar to E. martinicensis with regards to the form of the ilium. These differences are readily apparent in Figure 2. E. harlagnei has a thinner ilial shaft, smaller angle of ventral ace- tabular expansion, less elevated ilial prominence and a very short ACTUAL HEAD WIDTH Figure 3. Head width iu per cent of snout-vent length vs. actual head width in mm. Crosses are E. martinicensis, circles are E. harlagnei sp. nov. Large cross is a small E. martinicensis with a narrow head ; in other fea- tures it is typical of its form. crest beginning at the anterior edge of the ilial prominence and extending anterad about one and one-half times the length of the prominence. There is also less of a ventral acetabular ex- pansion in E. harlagnei. Acknowledgment. The collections on which this study was based were made under the auspices of National Science Founda- tion Grant NSF-G 16066. 1965 NEW GUADELOUPEAN FROG 7 LITEEATUEE CITED AUFFENBEBG, W. A. 1958. A small fossil herpetofauna from Barbuda, Leeward Islands, with the description of a new species of Hyla. Quart. J. Florida Acad. Sci., 21: 248-254. Cochran, D. M. 1941. The herpetology of Hispaniola. Bull. U. S. Nat, Mus., 177: vii + 1-398 pp. Lynch, J, D. In press. The status of the tree frog, Hyla iarhudensis Auffenberg, from Barbuda, BWI. Copeia. Lynn, W. G. and C. Grant 1940. The herpetology of Jamaica. Bull. Inst. Jamaica, Sci. Ser., 1: 1-148. Schwartz, A. 1957. A new species of Eleutherodactylus (Amphibia: Leptodactyli- dae) from Cuba. Proc. Biol. Soc. Washington, 70: 209-212. 1958a. Four new frogs of the genus Eleutherodactylus { Leptodactyli- dae) from Cuba. Anier. Mus. Novit. No. 1873: 1-20. 1958b. Another new species of Eleutherodactylus (Amphibia: Lepto- dactylidae) from Western Cuba. J. Washington Acad. Sci., April, 1958: 127-131. 1958c. Another new large Eleutherodactylus (Amphibia: Leptodactyli- dae) from Western Cuba. Proc. Biol. Soc. Washington, 71 : 37-42. 1958d. A new frog of the auriculatus group of the genus Eleuthero- dactylus from Western Cuba. Herpetologica, 14: 69-77. 1959a. A new species of frog of the Eleutherodactylus ricordi group from Central Cuba. Amer. Mus. Novit., No. 1926: 1-16. 1959b. The status of Eleutherodactylus pinarensis and a new species of the genus from Western Cuba. Herpetologica, 15(2) : 61-69. 1960. Nine new Cuban frogs of the genus Eleutherodactylus. Eeading Pub. Mus. and Art Gallery, Sci. Pubis., 11: 1-50. (Eeceived 22 January, 1965.) ■- \J BREVIORA Mmseium of Comparative Zoology Cambridge, Mass. May 7, 1965 Number 221 NEW MELANESTAN ANTS OF THE GENERA SIMOPONE AND AMBLYOPONE (HYMENOPTERA- FORMICIDAE) OF ZOOGEOGRAPHIC SIGNIFICANCE ^-^ By Robert W. Taylor Biological Laboratories, Harvard University The two ants described below are of special zoogeographic interest. Simopone gressitti sp. n. (subfamily Cerapachyinae) is the second species of its genus recorded from the Indo-Australian area, and the first east of the Philippines. Amhlyopone noona- dan sp. n. (subfamily Ponerinae) is the first apparently endemic Amhlyopone to be described from Western Melanesia. Simopone Forel includes ten described Ethiopian and Mala- gasian species: 8. grandicUeri Forel, 1891 (in Grandidier, Hist. Nat. Phys. Madagascar, 20: 141, pi. 4, fig. 8, Imerina, Madagas- car) ; ^S*. emcnji Forel, 1891 {ihid. 247, Anoside, Madagascar) ; S. conradti Emery, 1899 (Ann. Soc. ent. Belg., 43: 475, Cam- eroon; 1911, Genera Insect., 118: 16, pi. 1, fig. 7) ; 8.{f) mayri (Emery), 1900 (Bull. Soc. ent. Ital., 31: 264 (Cerapachys) ; 1911, Genera Insect., 118 : 16, Antongil Bay, Madagascar) ; S. marleyi Arnold, 1915 (Ann. S. Afr. Mus., 14 (1) : 20, Stella Bush, South Africa) ; S. grandis Santschi, 1923 (Rev. Zool. Afr., 11 (3) : 259, Kungu, Congo) ; 8. schoutedeni Santschi, 1923 (ibid. : 260, fig. 1 a-c, Kamaiembi, Congo) ; ^S*. fulvinodis Santschi, 1923 (ihid.: 262, fig. Id, Kiclaba [Kitabola], Congo) ; ^^. wilhuri Weber, 1949 (Am. Mus. Novit., 1396 : 7, figs. 6, 7, N. of Beni, Congo) ; 8. laevissima Arnold, 1954 (Ann. Mus. Congo, n.s., 4°, Zool., 1:291, figs. 1, la, Zika Forest, Uganda). An eleventh species, 8. bakeri Menozzi, 1926, was described from Singapore (Atti Soc. Nat. Mat. Modena, (6) 4: 92 (1925)). All these species are known only from the worker, except 8. mayri, which 1 Research supported by U.S. National Science Foundation Grant No. GB. lfi.34. 2 The specimens discussed liere were provided by Drs. J. L. Gressitt (Bishop Museum, Honolulu) and B0rge Peterson (Universitetets Zoologiske Museum, Copen- hagen), whose assistance is gratefully acknowledged. 2 BREVIORA No. 221 is based on a unique male and may not really belong in Simo- pone. Simopone workers are small to medium sized slender ants (length about 5.0-8.5 mm), usually dark brown or black in color, with very weak to moderately intense sculpturation and pilosity. The head is elongate-subrectangular, prismatic behind, with a transverse occipital carina. Frontal carinae horizontal, diverging posteriorly and obscuring the antennal insertions in facial view; fused anteriorly with the median part of the cly- peus, and forming with it and the frontal area a continuous planar surface, thus producing an anterior cephalic structure much as in the aberrant Indo-Australian myrmicine genus Meta- pone. Eyes very large (maximum diameter about 0.3 to 0.5 x the head width), situated at or just behind the middle of the sides of the head. Ocelli present, usually minute and closely ap- proximated. The 11-segmented antennae have flattened stubby scapes (usually only about 3 x as long as broad) which lie at rest in well developed preocular antennal scrobes, each of which is enclosed dorsally by the frontal carina and ventrally by the characteristically cerapachyine genal carina. These carinae usu- ally reach the eye posteriorly and may become continuous with a very fine postorbital carina, so that the eye is essentially en- closed within the scrobal area. Mandibles obtusely triangular, strongly arched ventrally ; masticatory border with a number of small regular teeth. The palpal formula of a single African specimen (species evidently undescribed) in the MCZ collection is maxillary 6: laMal 2, possibly 3 (inspected). The structure of the mesosoma is generally like that of Phyra- caces niayri Forel, with its dorsolateral borders broadly or nar- rowly rounded, sometimes angled but never carinate. Pronotum prismatic anteriorly, with a transverse carina between the humeri. A similar carina may separate the dorsal and declivitous faces of the propodeum and the declivitous face may be laterally margined. Sutural traces on mesosomal dorsum weak or ves- tigial, the mesometanotal suture sometimes lacking. The leg segments, especially the femora, are often inflated, the fore and hind tibiae each bear a single pectinate spur, and the pretarsal claws are toothed or pectinate.^ The posterior flange of the hind coxa may be produced dorsally as a more or less raised lamella 8 The characters of the tibial spurs and the pretaisiil claws have been selrlom mentioned in specific descriptions. All specimens which 1 have seen lack tibial spurs on the middle legs, and have a single median tooth on each tarsal claw. 1965 MELANESIAN ANTS d (another character common to many cerapachyines). Petiole longer than broad, subrectangular-trapezoidal in dorsal view, the dorsolateral margins acarinate, though sometimes angled; there is usually a transverse anterior carina, and sometimes a posterior one. Postpetiole strongly constricted behind, sub- rectangular in dorsal view, about equal in size to the petiole or larger. Pygidium flattened at its apex, with a full or reduced complement of bristle-like marginal setae, the presence of which indubitably establishes the cerapachyine affinities of this genus (see Brown, 1954). Little is known of the biology of Simopone but its general habitus strongly implies that it is arboricolous ; several of the older types were collected on vegetation and one species (;Si. marleyi — see Arnold, 1915) has been taken in hollow twigs. Specimens in the MCZ collections are either labeled as having been swept from foliage, or else carry adherent moth wing scales, an almost sure sign that they were collected by sweep- net. The genus is apparently an aberrant arboricolous offshoot from Phyracaces-like stock. The feeding biology needs study, especially since many cerapachyines, including some Phyracaces species, are apparently specialized myrmecophagous feeders (Wilson, 1958). Amhlyopone Erichson is an almost cosmopolitan genus now containing 50 described and apparently valid species, including 31 from the Indo-Australian area. The world fauna was exten- sively reviewed by Brown (1960), and one subsequently de- scribed species is known (Brown, 1962). A. noonadan sp. n. is a member of the Indo-Australian luzonica group, which includes the following species: luzo7iica (Wheeler and Chapman) 1925, Philippines {=williamsi (Wheeler and Chapman) 1925; synon- ymy by Brown, I960); silvcstrii (Wheeler) 1928, Japan; am- hlyops (Karawajew) 1935, Indo-China; and probably also celata (Mann) 1919, Solomon Islands (see Brown, 1960, for details and references). The features distinguishing noonadan from these species and others present in Melanesia are given below. Simopone gressitti Taylor, new species Type locality. NEW GUINEA (WEST) : HoUandia-Binnen, 100 m. The unique holotype worker was collected on November 1, 1958, by Dr. J. L. Gressitt for whom this species is named. Type deposition. Holotype deposited in the Bernice P. Bishop Museum, Honolulu, Hawaii. BREVIORA No. 221 «4H o p be cS C c (D -4-) a a -k^ _g Ph l-H tn CI •-s^ -f* ffi •^ o =0 CO CO Ji Si ■» •4-^ .rH O S Sj o o CO C OJ -M Oi d cS a 2: < (U T— ( m o Ph 1965 MELANESIAN ANTS 5 Description. Dimensions (in mm): Total length (TL) 6.4; head length at midline (HL) 0.98; maximum head width (IIW) 0.79; maximum pronotum width (PW) 0.70; Weber's length of mesosoma (WL) 1.4; petiolar node length at midline 0.72; maxi- mum petiolar node width 0.69 ; postpetiolar length at midline 0.70 ; maximum width of postpetiole 0.75. General form as shown in Figures 1 and 2. Mandibles small, their outer borders sinuate, convex basally ; angle between masticatory and posterior borders broadly rounded; dentition worn, almost effaced, probably orig- inally like that of S. hakeri. Head 0.81 x as broad as long ; sides almost parallel, slightly concave before eyes, slightly convex behind ; occipital border feebly concave ; occipital carina well developed, ribbed along its anterior edge, enclosing the occiput laterally as well as dorsally, its ventral traces extended for- wards for a short distance on each side, along the sides of the postgenae. Eyes large, maximum diameter 0.30 mm (0.38 x the head width), separated by a distance of 0.46 mm (1.53 x their maximum diameter) ; ocelli minute. Scapes barely reaching anterior margins of eyes, flagellar proportions as in Figure 2. Anterior clypeal border feebly convex. Frontal carinae diverg- ing posteriorly, on each side meeting the postorbital carina, which is very fine and continuous below the eye with the genal carina. Mesosoma twice as long as broad in dorsal view, its dorsolateral borders approximately right-angled but acarinate; transverse pronotal carina well developed; angle between dorsal and de- clivitous propodeal faces abruptly rounded, acarinate. Sutura- tion of mesosomal walls as shown in Figure 1 ; promesonotal and mesometanotal sutures represented by transversely ribbed traces on mesosomal dorsum, mesometanotal suture weakest. Femora and tibiae moderately inflated (Fig. 1) ; posterior edge of hind coxa raised but not lamellate ; pretarsal claws each with a single median tooth. Petiole trapezoidal in dorsal view, broader behind than in front, with a distinct ribbed transverse anterior carina; lateral borders strongly angled at about 60 degrees, but not carinate ; profile as in Figure 1. Sides of postpetiole slightly con- vergent posteriorly in dorsal view. Pygidial spines reduced to a single minute pair on each side, at the extreme apex. Mandibles smooth and shining, with a few minute piligerous punctures. Entire body strongly shining, virtually lacking sculp- turation except for scattered minute piligerous punctures, and some effaced longitudinal rugosity along the sides of the pronotal dorsum and on the metepisternum, which is in part coarsely punctate-rugose. Pilosity reduced. A few moderately long 6 BREVIORA No. 221 suberect to reclinate hairs on mandibles, clypeus and underside of head, propleurae, coxae, and undersides of petiole and gaster ; hairs most abundant on the propleurae and coxae, and on the petiolar sternite where they form a peculiar brush-like series behind the subpetiolar process. Single, slightly longer erect hairs in the following positions : at the midlength of each frontal carina, above eyes, on pronotal humeri, on the anterior half of the sides of the node (2 pairs) and the anterodorsal corners of the postpetiole. Similar, but slightly less erect hairs increasingly long and abundant towards the gastric apex, which is surrounded by very long arched hairs. Pubescence virtually absent. Color very dark brown, almost black, the following areas weakly in- fuscated with reddish brown : mandibles and front of head, an- terior parts of each gastric tergite, gastric apex and legs, es- pecially the tibiae and tarsi; antennae medium dull reddish brown. Diagnosis. According to Menozzi's description (Menozzi, 1926) S. hakeri is smaller than gressitti, with a narrower head and petiolar node. I have tentatively identified as 8. hakeri a speci- men in the J. W. Chapman collection (MCZ) from the Philip- pine Island of Negros (Horns of Negros 3,600 ft., J. W. Chap- man). This individual agrees well with Menozzi's description, and resembles gressitti in color and habitus, but has very different cephalic, ocular and petiolar proportions as follows : TL c. 5 mm ; HL 0.91 mm; HW 0.68 mm (head 0.75 x as broad as long) ; PW 0.55 mm; WL 1.2 mm; petiolar node length at midline 0.68 mm; maximum petiolar node width 0.56 mm (node 0.52 x as broad as long); postpetiolar length at midline 0.64 mm; maximum width of postpetiole 0.61 mm. The maximum diameter of the eyes is 0.30 mm (0.44 x head width) and they are separated by a distance of 0.34 mm (1.14 x their maximum diameter). Apart from these proportional differences the post-cephalic structure is similar to that of gressitti. The pilosity is similarly distributed but less abundant, and the subpetiolar "brush" is lacking. Amblyopone noonadan Taylor, new species Type locality. TERRITORY OF NEW GUINEA : New Brit- ain: Yalom, 1,000 m. May 19, 1962 (Danish Noona Dan Ex- pedition). The types were collected "in or on the ground in newly cleared secondar.v growth," no collector specified. 1965 MELANESIAN ANTS Type deposition. The holotype is deposited in the Universi- tetets Zoologiske Museum, Copenhagen, Denmark; the paratype is in the Museum of Comparative Zoology (Type No. 31148). The species is named, in apposition, for the Danish expedition vessel Noona Ban. Worker. The following description is based on the holotype and single paratype. Dimensions (in mm, holotype cited first). TL c. 5.5, 6 mm; HL (including clypeal denticles) 1.04, 1.12; HW (immediately behind genal teeth) 0.96, 1.04; maximum scape length (exclud- ing articular condyle) 0.64, 0.68; outside total length of man- dible 0.85, 0.90 ; PW 0.58, 0.65 ; WL 1.38, 1.42 ; midline length of petiolar dorsum 0.49, 0.55 ; maximum petiolar node width 0.50, 0.57 ; postpetiolar length at midline 0.36, 0.40 ; maximum width of postpetiole 0.65, 0.73. General habitus as in Figures 3 and 4. Head with occipital border feebly concave, sides feebly convex, converging posteriad ; Figs. 3 and 4: Amblyopone noonadan sp. n. Fig. 3. Lateral view of meso- soma and node. Fig. 4. Frontal view of head, right antenna omitted. Scale line 1 mm. Holotype worker. 8 BREVIORA No. 221 anterior corners with strong acute genal teeth, the inner edges of which are about as long as the maximum width of the mandib- ular shafts. Frontal lobes approximate, separated by a deep linear groove. Clypeal apron strongly convex, with eight small denticles ; the four median ones closely approximate, less deeply separated from each other than from the more lateral denticles, their apices diverging from the midline ; the innermost of the two lateral denticles on each side moderately large, triangular, separated from the median quartet by a gap equal to its width at base ; the outer tooth large and blunt, its apex jagged, form- ing two or three indistinct cusps. Mandibles slender, their ex- ternal margins feebly concave, each bearing ten acute, slightly recurved teeth. The two basal teeth simple, conical, the basal- most blunt, the second acute ; the eight apical teeth arranged in four more or less separated pairs, in typical "stigmatommine" fashion ; the dorsalmost tooth of each pair lies slightly distal to its partner ; a distinct low reclinate tooth is present on the inner edge of the strong acute mandibular apex. The mandibular apices cross when the jaws are closed, leaving a triangular gap between the mandibular and clypeal teeth. Palpal formula maxillary 4: labial 3 (paratype dissected). Scapes slender, slightly incrassate ; funiculus with 11 segments proportioned as in Figure 4. Eyes small, variable in size, maximum diameter 0.02 mm in holotype, 0.06 mm in paratype, with four and nine or ten indistinct facets, respectively. Mesosomal profile as in Figure 3. In dorsal view this tagma is widest at the pronotum and strongly narrowed at the base of the propodeum. Pronotal dorsum almost flat, with marginate lateral borders; inferior angles of pronotum broadly rounded. Promesonotal suture flexible ; mesonotum transverse ; mesometa- notal suture A^estigial, represented only by a sculptural break between the subopaque mesonotum and the moderately shining propodeum. Propodeal dorsum about as long as broad, its sides diverging posteriad; declivity feebly concave, rounding into dorsum, its lateral edges slightly raised, forming angles of a lit- tle more than 90 degrees in dorsal view. Petiole sessile, its profile as in Figure 3 ; subpetiolar process afenestrate; nodal dorsum slightly wider than long in dorsal view, the anterior border Avith a slight median emargination, the sides converging slightly anteriad. Postpetiole wider but shorter than node, and also shorter than the succeeding segment which is of about the same width. Gastric apex laterally compressed, 1965 MELANESIAN ANTS 9 sting stout. Tibial spurs vestigial on middle legs ; posterior tibiae each with a broad flat pectinate spur and a more slender simple conical one. Mandibles and frontal lobes obscurely longitudinally striate. Clypeal apron with somewhat radial longitudinal striae; lateral parts of clypeus similarly sculptured, the striae on each side radiating back from a focal point at about the level of the inner basal edge of the mandible. These lateral clypeal striae arch back over the cheeks, where those nearest the midline are almost longitudinal, reaching back to the base of the frontal carinae; the cheek striae become increasingly divergent towards the sides of the head and the most lateral ones gather apically, at the base of the genal tooth. Remainder of head coarsely and roughly punctate-rugose ; the sculptural trend faint, mainlj^ longitudinal, but transverse across a narrow posterior strip. Scapes finely shagreened. Postgenae somewhat obscurely and irregularly longitudinally striate, the striae diverging posteriad. Lateral parts of the dorsa of the pronotum and propodeum with scattered punctures, about 0.02 mm in diameter, separated by about twice this distance on pronotum and more widely spaced on propodeum ; a narrow longitudinal median strip on these scler- ites lacks punctures, the surface here is shining, with a very fine superficial scale-like surface pattern which is also present on the interpunctural areas of the lateral strips, and which has a transverse trend on the posterior propodeal dorsum. Mesonotum subopaque, coarsely and irregularly shagreened. Sides of mesosoma, except metepisternal area, subopaque, bearing somewhat effaced and polished, almost vertical fine striae, which are slightly curved (concave anteriorly) and slope posteriorly; sculptural intensity diminishing posteriad, with striae virtually absent behind the propodeal spiracle. Metepister- nal area longitudinally striate. Declivity of propodeum shin- ing, with very superficial, minutely scale-like transverse sculptur- ation. Petiolar dorsum subopaque, with scattered fine piliger- ous punctures ; gastric tergites similar, the punctures finer and more abundant. Pubescence adpressed and subadpressed, generally distributed over body except for the post-pronotal sides of the mesosoma and the sides of the petiolar node. Erect pilosity moderately abundant, especially on the dorsum of the body and towards the gastric apex, where the hairs are longest. Ground color dark 10 BREVIORA No. 221 chocolate browii, with the following areas infuscated with red- dish brown: clypeus, frontal lobes, anterior corners of head, including genal teeth, pronotal collar, area of mesosomal-petiolar junction, subpetiolar process, posterior edges of gastric tergites, and gastric apex. Mandibles, antennae, legs and sting rich golden brown. Diagnosis. A. noonadan is readily distinguished from the only other known western Melanesian Amhlyopone, A. australis Erichson, by the characters of its " stigmatommine habitus" (i.e., "double ranked" mandibular dentition and enlarged clypeal teeth, etc., — see Brown, 1960). In addition, australis is larger (minimum known HW on New Guinea about 1.5 mm), and has an acute tooth on each inferior pronotal angle. The third known Melanesian species, A. celata Mann (Solomon Islands), is much smaller (3 syntype workers in the MCZ collection have HL 0.69-0.71 mm; HW 0.60-0.63 mm) with minute genal teeth (max- imally only about as large as the median clypeal denticles) and with the head evenly and rather finely shagreened. The other luzonica group species have reduced genal teeth, as in celata, and most specimens (MCZ collection) are somewhat smaller than the noonadan types : 3 silvestrii syntypes have HL 0.87-0.92 mm; HW 0.76-0.81 mm; the luzonica female holo- type from Los Banos (Luzon) has HL 0.82 mm, HW 0.74 mm; and six luzonica workers from Dumaguete (Negros) have HL 0.85-0.90 mm, HW 0.74-0.80 mm. Three williamsi syntype workers from Baguio (Luzon) are somewhat larger with HL 1.03-1.09 mm, HW 0.91-0.97 mm, and Karawajew (1935) gave similar measurements for his amhlyops type — HL 1.05 mm, HW 0.98 mm. These "species" differ among them- selves in the conformation of the clypeal denticles and in the sculpturation (Brown, 1960, p. 196), but they may ultimately prove to be geographical variants of a single species, especially considering the extensive variation known in other members of the genus. Notwithstanding, there can be little doubt that noona- dan and celata are good "species. 5 J EEFEEENCES Arnold, G. 1915. A monograph of the Formicidae of South Africa (Ponerinae; D.orylinae). Ann. S. Afr. Mus., 14 (1): 1-159. Brown, W. L., Jr. 1954. Eemarks on the internal ph.ylogeny and subfamily classification of the family Formicidae. Insectes Soc, 1: 22-31. 1965 MELANESIAN ANTS 11 1960. Contributions toward a reclassification of the Formicidae. III. Tribe Amblyoponini (Hymenoptera). Bull. Mus. Comp. Zool., 112 (4): 145-230. 1962. A new ant of the genus Amblyoponc from Panama. Psyche, Cam- bridge, 69 (2): 73-76. Karawajew, W. 1935. Neue Ameisen aus dem Indo-Australischen Gebiet. Treubia, 15 (1): 57-117. Menozzi, C. 1926. Nuove formiche delle isole Filippine e di Singapore. Atti Soe. Nat. Mat. Modena, 6 (4): 92-103 (1925). Wilson, E. O. 1958. Observations on the behavior of the cerapachyine ants. Inseetes Soc, 5: 129-140. ^v ,iU BREVIORA MniseMiii of Compsirative Zoology Cambridge, Mass. May 28, 19G5 Number 222 THE GENUS LEPTOTYPHLOPS IN THE WEST INDIES WITH DESCRIPTION OF A NEW SPECIES FROM HIS- PANIOLA (SERPENTES, LEPTOTYPHLOPIDAE) By Richard Thomas 10,000 SW 84tli St., Miami, Florida 33143 Three specimens of Leptotyphlops were collected in the summer of 1964 in the vicinity of the town of Pedernales, Dominican Republic, by Mr. David C. Leber and myself. These snakes ap- pear most closely related to L. hilineata Schlegel which they re- semble in the failure of the ocular to reach the labial border. No snakes of the genus Leptotyphlops have previously been definitely noted from the island of Hispaniola. Boulenger (1893), however, recorded a specimen of Lepto- typhlos alhifrons Wagler from Santo Domingo de Guzman col- lected by Dr. A. C. Buller. It so happens that the name of the capital city of the Dominican Republic, rarely referred to in its entirety, is Santo Domingo de Guzman. The question of the provenance and relationships of the Buller specimen might therefore be important. Thanks to Miss A. G. C. Grandison, I have been able to examine this specimen (BM 90.10.10.73) and can confirm its affinities with the alhifrons group of the genus. It appears to agree most closely with L. phenops hakewelli Oliver although the rostral-prefrontal fusion is apparently lacking. (Dunn and Saxe, 1950, regard phenops as a race of alhifrons; but I here follow more recent authors such as Peters, 1954, and Duellman, 1961, in giving phenops specific rank.) Miss A. G. C. Grandison writes (in litt.) that "Dr. Audley C. Buller . . . made quite extensive collections in Mexico in 1891 and 1892, travelling from L. [Lago] Chapala and Guadalajara ... to Bolafios and back to Ixtlan and later ... to an area west of Guadalajara." Examination of a recent map of Mexico shows that roughly 200 km to the soutli of the city of Guadalajara in the state of Jalisco is a Ciudad Guziiuni, whicli may well he another case of the un- wieldy Santo Domingo de Guzman liaving been sliortened. Smith 2 BREVIORA No. 222 and Taylor (1945) list L. phenops hakewclli from the states of Colima, Guerrero, Jalisco, Michoacaii, and Oaxaea. That identity, collector, locality and time should all approximately coincide seems too remarkable for mere coincidence. This specimen (BM 90.10.10.73)1 is here considered to be close to L. p. hakewclli from what is now Ciudad Guzman, Jalisco, Mexico, and hence irrelevant to the Hispaniolan problem. The specimens from Pedernales, however, are distinct from other species and apparently represent an endemic form, here named in allusion to the type locality - : Leptotyphlops pyrites new species Holotypc: MCZ 77239, collected at the southern outskirts of the town of Pedernales, approx. 1 km from the center of town, Pedernales Province, Dominican Republic, 3 July 1964, by Richard Thomas. Paratypes: USNM 152452, same locality as type, 26 June 1964, Richard Thomas; ASFS V2601, 9 km n' Pedernales, Pedernales Province, Dominican Republic, 26 June 1964, David C. Leber. Diagnosis: A species of Leptotyphlops of closest affinities to hilincata in that the second and third upper labial scales exclude the ocular from the labial border. It is further characterized by considerable attenuation, a high number of middorsal scales (from rostral to tail spine), 12 scale rows around the base of the tail, and 15 to 16 subcaudals, a trilineate dorsal pattern and unicolor, dark sides and venter. Distributio7i: Known presently only from the northwestern lowlands of the Barahona peninsular region of Hispaniola. Description of holotypc (Fig. 1) : Head rounded, of same width as neck. Rostral at widest point slightly less than width of head at eyes, broadly truncate posteriorly at contact with prefrontal. Nasals separated by a transverse suture proceeding from first labial diagonally upward across naris to rostral ; dorsal half of nasal also in contact with rostral, prefrontal, supraocular, ocular. IBM 90.10.10.7.1 : Total length ca. ITjI luin ; tiiil 7.(> iiiui : middorsal scales al)out 24() : siibcaiKhils ] ■«s> ^ rO (B S ^J 1 o =4-1 locality was published by Rensch in 1934. Solmopina macfarlanei (Cox) PI. 1, figs. 1-2 Helix (Geotrochus) macfarlanei Cox 1873, Proc. Zool. Soc. London, p. 567 (Solomon Islands). Helix (Papuina) macfarlanei Cox. Pilsbry 1891, Manual of Conchology (2) 7:13 (Solomon Islands). Papuina macfarlanei (Cox). Eeiisch 1934, Amer. Miis. Novit. No. 763:7 (Kiota, Bougainville, Solomon Islands). BREVIORA No. 224 Height Grejiter dianiete nun mm 23 25 23 24.5 21 22 Measurements near Buin, Bougainville near Buin, Bougainville near Buin, Bougainville Description : Shell troeliiforni, imperforate, smooth, and reach- ing 25 mm in greater diameter. Color a pale ivory with a dark brown band at the whorl periphery and a broad circular area of the same color on the base of the shell. In addition, there are numerous irregular spots of brown both above and below the periphery. These brown bands and spots are translucent. The somewhat lighter brown spots are a little more translucent than the bands. Whorls 51/0 and convex. Spire somewhat extended and produced at an angle of about 70°. Aperture subquadrate with the li]) slightly reflected and cast at an angle of about 42° from the base. Columella at an angle from the umbilical area to the base. Suture indented. Sculpture consisting of tine and irregular growth lines. Protocunch of IV^ whorls, white and smooth. Remarks: This species differs from S. coxiana (Angas) by being larger, having more convex whorls, and in having the peripheral band of brown. The shell of /S*. coxiana is more deli- cate in structure and the brown coloration is considerably lighter in color. The brown spots in coxiana are also translucent. 1 do not agree with Renscli that these two species may be only geo- graphic races. They appear quite distinct. In 101 specimens examined only a single specimen lacked the peripheral band. Rensch has recorded this species from Kieta, Bougainville. Spccinifns ( xaniini d : BorGAixvn.LE : Buin (MCZ). SOLMOPINA COXIANA (Augas) PI. l.tig. 3 (ieotioiliu.s coxiana Angas 18(37, Pioe. Zool. Soi-. Lonchui, ji. SSii, pi. 4.'>, figs. 7-8 (Ysabel Island, Solomon Islands). Helix (Papuina) coxiana (Angas). Pilslny isyl, Manual of ("oncholog.v (2) 7:13, pi. 3, figs. 36-37. Papuina coxiana (Angas). Renseh 1934, Amer. Mus. Xovit. No. 763:8. Measurements Height Greater Diameter mm mill 17 20 Ysabel Island 1965 LAM) MOLU'SKS FROM THE SOLOMONS 3 Dcscripiion: Slu^ll trochifonn, imporforatc, smooth and r(»a('liiii<>- 20 niiu in <>-i'c'at('i" diaiiu'tcr. Color a j)al(' ivory with a broad ciiH-ular area of Ii»i'ht brown on tlic liase of tlie shell, and, in addition, a few brownish spots irre<;ularly disjiosed over the shell. Both the basal band and the spots are translucent. Whorls 5, slightly eonvex. Spire somewhat extended and produced at an augle of about 70°. Aperture subcpiadrate with the li{) slightly reflected, colored brown and cast at an angle of about 40° from the base. Columella white and at an angle from the umbilical area to the base. Suture slightly indented. Sculpture consisting of very fine and irregular growth lines. Protoconch of IVij whorls, white and smooth. Remarks: See RcDiarks under S. mdcfarlanei. Spcci)ncns exa)}ii.)ic(1 : Ysabel. l^pecimcns recorded: Choiseul: Luti and Tauro (Kensch). ARIOPHANTIDAE Trociiomorpiia vanderrieti new species PI. 2, figs. 1-2 Holoiiipc: MCZ 251176, from Ataa, Malaita Island, British Solomon Islands. Rev. J. Vander Riet, collector, 1964. Paratypes: Figured paratype, MCZ 251177; additional para- types, MCZ 247959, all from the same locality as the holotype. Greater Lesser Heiglit (lianu'tei- diameter mm mm mm 11 26.5 22 10 27 22 11.5 26 20.5 11.5 25 20.5 10 28 21 Holotype Paratype MCZ 251177 (figured) Paratype MCZ 251177 Paratype MCZ 251177 Paratype MCZ 247959 (figured) Description: Shell reaching 27 mm in width, subdepressed, finely carinated and umbilicated. Color a dark, yellowish brown, uuicolorous below the thread-like carina and flecked with nu- merous narrow, straw-colored patches in axial arrangement above the carina. The brown coloration is in the shell; the straw coloration is invested in the periostracum. Whorls 5 to 5iX>, con- vex, and the last whorl with a tine, thread-like carina. Spire depressed, dome-shaped and forming an angle of about 1:^5°. Aperture subovate, with the outer lip slightly depressed near its juncture with the whorl above. Outer liji narrowly reflected 4 BREVIORA No. 224 along- the basal area. Umbilicus rather narrow and very deep, and is Yq tlie width of the shell in greater diameter. Suture slightly indented and well defined. Sculpture consisting of nu- merous oblique and fine costae above the whorl periphery and nearly smooth below. Re))iarks: In relationship, this species is nearest to Trocho- morpha aukiensis Cla]ip, also from Malaita. Trochomorpha van- derricti differs from H. aukknsis by being: larger, not being spirally banded and in having a narrower umbilicus, being % of the greater diameter, while in aukiensis the umbilicus is I/4 of the greater diameter and there are no straw-yellow flecks above the whorl periphery. Including- T. vandcrrieti, there are seven species of Trocho- morpha now known from Malaita Island. These are : Trochomorpha aukiensis C'lapp 1923, Bull. Mus. Comp. ZooL, 65:361. j)l. 2, figs. 10-15 (Auki, Malaita). [Holotype, MCZ 32535]. TrochomorpJia^ hclmorci (Cox) 1871, Proc. Zool. Soc. Lon- don, J). 647, pi. 52, fig. 12 (Solomon Islands). Trochomorpha concava Clapp 1923, Bull. Mus. Comp. Zool., 65:363, pi. 3, figs. 1-3 (Auki, Malaita). [Holotype, MCZ 32523]. Trochomorpha crust idio)) (Cox) 1873, Proc. Zool. Soc. Lon- don, p. 15U (Solomon Islands). Trochomorpha flava Clapp 1923, Bull. Mus. Comi). Zool., 65:366, pi. 3, figs. 4-6 (Auki, Malaita Island). [Holotype, MCZ 32521]. Trochomorpha mclcaiii Clench 1959, Natural History of Rennell Island, British Solomon Islands, I"^niv. Copenhagen, Denmark, Vol. 2:179, pi. 17, fig. 6 (10 miles inhuul from Sun, Malaita Island). [Holotype, Amer. Mus. Nat. Hist. 7!)016]. ASSIMINEIDAE Setaepoma Clench Setaepoma Clench 19.")0, Nautilus, 63:1. U (fy]K' species, Japonid ( .' ) hcdi fieri T. Miul R. Renscli). Setaepoma iioodi new species PI. 2, fig. 3 Jlolotffpr: MCZ 25130!), from the west side of the Teiiarn River, about one-half mile above the Catholic Mission, Guadal- canal, Solomon Islands. James R. Hood, collector, 194i). 1965 LAND MOLLISKS FROM TllK SOLOMONS 5 Measurcinciits Height Width mm mm H:2 7.!) TTololypo Description: Shell tiii'binatc, thin, umbilicato and having' nu- merous, spiral rows of bristle-like processes of periostraeuiu. Color a o-oldeu brown. Whorls 6 and tubular. Spire moderately extended and produced at an angle of about 60 \ Aperture nearly circular, holostomatous and with a thin, simple lip. No columella ; umbilicus narrow and deep. Suture deeply indented. Periostracum thin, shining', and with numerous, spiral and axial ridges which support a bristle-like process at each intersection. Protoconch of about 2 whorls which are smooth and shining. Operculum calcareous, multispiral and dished. Inner surface with a papilliform central nucleus, smooth, shining, and with faint indications of the outer sculpture. Remarks: This species is distantly related to Sietoepoma hccligeri I. and B. Renscli (Bougainville Id.) and S. )iw}iri Clench (Ysabel and Choiseul Ids.). It differs from both by being more attenuated, having fewer spiral ridges and a much nar- rower umbilicus. This present new species is the third known from the Solomon Islands. The other two are : Sciacpoma Judifjcri (I. and B. Rensch) 1986, Revue Suisse de Zoologie, 43:678 (Bougainville Id., Solomon Ids.). SetacpoDia maijri Clench 1959, Natural History of Rennell Island, British Solomon Islands, Univ. Copenhagen, Denmark, Vol. 2:3 68, jil. 17, tig. 5, text fig. 1 (Fulakora, Ysabel Id., and the Wurulata River. Choiseul Id., Solomon Ids.). CYCLOPHORIDAE Palaina (Palaina) brazieri (Cox) PL 2, fig. 4 Diplommaiina haizieri Cox 1870, Proe. Zool. Soc. London, p. 84 (Wanga, San Christoval Id., Solomon Islands). This species has not been figured previously. The specimen we figure came from Kira Kira, San Christoval Id., collected by the Whitney Expedition in 1929. 6 BEEVIORA No. 224 Palaixa (Palaina) delli new species PI. 2, fig-. 5 Holotype: MCZ 258017, from one-half mile above the Seventh Day Adventist Mission Station on the west side of the Lunga River, Guadalcanal Island, Solomon Islands. James R. Hood, collector, li)4y. Paratypcs: Five paratypes, MCZ 2"'.8018, from the same lo- cality as the holotype. Measurements Height Width iiini mm 4.5 2 Holotype 4.7 2.2 Paratype 4.5 2 Paratype 4.5 2.1 Paratype 4.3 1.8 Paratype Description: Shell reaching 4.7 mm in height, dextral, imper- forate and sculptured. Whorls Gi/o and convex. Color a dull grayish brown. Spire attenuate and straight. Aperture circular, with a I)roadly reflected lip, and attached at the parietal area. No columellar tooth. Suture impressed. Sculpture consisting of numerous, diagonal riblets which are very irregular as to number on the different whorls and between different specimens. These ril)lets may be singk' or in groups of 2 to 5 and varying from whorl to whorl. Protoconch of 2 whorls, smooth and rather dull. Operculum with a vertical, spiral lamella. Remarks: This species is nearest in relationship to P. hrazieri (Cox) from San Christoval. It diff'ers by being more than twice as large and in having more regular and more numerous axial riblets. Both are dextral and neither has a columellar tooth. Named for Dr. Richard K. Dell, Dominion Museum, Welling- ton, Xcw Zealand. DU'LOMMATINA AERARI Dell DiploiiDiKil inn (i 20% snout- vent length). 2. Legs short (80 - 120% head length). 3. Middorsal and flank scales smaller than ventrals, granular, subequal in size or very slightly and gradually enlarged in the middorsal line. 4. Ventral scales hexagonal to square, subimbricate or juxta- posed, smooth or weakly keeled but never mucronate. 5. Mental wider than long, 4 scales + 2 sublabials in contact with infralabials. Smallest anterior gular scales not smaller than 1 /6 first sublabial and usually larger. 6. Dewlap small or vestigial. 7. Lamellae under fourth toe always more than 20. 8. Tail round or trigonal, never distinctly compressed; 4-5 dorsal scales per verticil. 2 BREVIORA No. 227 SPECIES DIFFERENCES The green anoles of Hispaniola divide into two subgroups, each a superspecies with two species. The species of the first subgroup {Anolis chloroajanus Dumeril and Bibron 1837, and A. coelestinus Cope 1862) are widespread and very common in the lowlands and up to at least 1500 meters; those of the second subgroup (Anolis aliniger Mertens 1939, and A. singularis n. sp.) occur only at elevations well above sea level and appear always to be scarce. The species of subgroup 1 occur sympatrically with those of subgroup 2, but the species within each subgroup are primarily allopatric. The two wide-ranging species of the first subgroup show enough geographic variation to permit description of subspecies; the subspecies will not, however, be discussed in this paper, which is concerned solely with species differences. In a majority of the characters in which anoles are distinguished by taxonomists, there is conspicuous overlap between these two subgroups and even more overlap between the species of a single subgroup. It is significant that, lacking the distributional evi- dence that is now available, Mertens (1939) described aliniger (of the second subgroup of my terminology) as a subspecies of chlorocyanus (of the first subgroup) and was not very sure (p. 62) that coelestinus was a species distinct from chlorocyanus. Table 1 shows the extent of this overlap very strikingly. Though in the better represented species of the first subgroup there are evident differences in the means of the several numerical charac- ters, it is quite as evident that there are individuals it would not be possible to place on these characters alone. Tables 2 and 3 show the very few qualitative characters that seem useful in discriminating the species of Hispaniolan green anoles. Size also differs. A. chlorocyanus and A. coelestinus may exceed 70 mm in snout-vent length, A. aliniger and A. singularis do not reach much over 50 mm. SUBGROUP 1: A. CHLOROCYANUS - A. COELESTINUS These two species overlap very little geographically. One has a distribution north, the other south of the Cul de Sac trench (still partly below sea level) which formerly divided Hispaniola into northern and southern islands. At the southern edge of this trench, now dry and very arid land, there is contact and sporadic real sympatry, always it appears, with one species or the other predominating. The actual zone of contact has never been mapped and is known at the moment only from occasional tran- sects. A. coelestinus occurs in Port-au-Prince and extends to Damien, but A. chlorocyanus occurs in the area too. 1965 HISPANIOLAN GREEN ANGLES Even in tliia area of contact there is no indication of reinforce- ment of species differences in body squamation. There are only- average differences in the size of the head scales which are reflected in counts of loreal rows, scales across snout, scales between inter- parietal and supraorbital semicircles, etc. However, the best differences between the two species are in body color (the presence in codestinus of the labial white streak which is continued above the shoulder, and its absence in chlorocyanus) , and in dewlap color and squamation (very large dewlap scales and black dew- lap skin in chloroajanus, small dewlap scales and relatively unpig- mented skin in coelestinus) (Fig. 1). Thus, in color characters there is some evidence of reinforcement of species difference because of secondary contact. Fig. 1. Scales along edge of dewlap. Above: Anolis coelestinus, MCZ 64883, Ca-ira near Leogane, Haiti; below: Anolis chlorocyanus, MCZ 80719, Nan Palmiste, Gonave Island, Haiti. Drawn to the same scale; the speci- mens have the same snout-vent length. Both body color and dewlap color differences, while they hold over most of the species ranges, are lost or weakened in popula- tions of these species remote from the zone of contact. Thus the Isle Vache population of coelestinus lacks the white labial to shoulder streak of other coelestinus populations. At least pre- served specimens of coelestinus remote from the Cul de Sac tend to show darker dewlap, occasionally almost to a degree that 4 BREVIORA No. 227 would be confusing were not the dewlap scales consistently small. In chlorocyanus also, geographic variation slightly weakens the diagnostic differences. The population described by Mertens from the vicinity of Santo Domingo City as A. chlorocyanus cyanostictus has reduced the black of the dewlap and exhibits a cadmium yellow basal spot. However, both the Isle Vache population of coelestinus and the Santo Domingo City population of chlorocyanus are very small segments of the total range of these species. The non-distinctive segment of coelestinus is isolated on an island far to the west; the exact range of chlorocyanus cyanostictus is very limited and on present evidence very uncertain: typical chlorocyanus have been collected in Santo Domingo itself. Both species are characteristic inhabitants of mesic lowland forests and appear not to extend to the highest peaks. They are, for example, unreported in the Foret des Pins, Massif de la Selle in Haiti, or at Valle Nuevo in the Cordillera Central in the Dominican Republic. Both these localities have been visited by several investigators and the apparent absences are quite probably real. SUBGROUP 2: A. ALINIGER-A. SINGULARIS N. SP. This pair of species is not known to overlap at all, but they both occur on the jMassif de la Selle. The previously described member of this pair, A. alinigcr, was until recently known only by the unique type, and was regarded by its describer jMertens as a subspecies of chlorocyanus. A. aliniger is, however, widely sympatric with chlorocyanus. A specimen of A. aliniger collected by Clayton Ray and A. S. Rand at 7 km N Car- pintero Prov., San Juan, Dominican Republic, was collected on a tree ca. 45 meters from a specimen of ^4. chlorocyanus collected on a pole fence. The head scales of these two specimens are shown in Figure 2. In the Constanza area. La Vega Province, Dominican Republic, natives collecting for Juan Rivero brought in A. aliniger and A. chlorocyanus at the same time. A. aliniger is one of the most peculiar of anoles, not indeed in general habitus which is that of a small and somewhat stockier version of chlorocyanus or coelestinus, but in the singular feature which gives it its name. This is the strange coloration of the axilla, bright orange or yellow followed by a larger or smaller spot of in- tense black. Thi.s is present in both sexes and very conspicuous in freshly preserved specimens, but the yellow or orange is, of course, (juickly bleached out by alcohol. There is, however, an 1965 HISPANIOLAN GREEN ANGLES JBC Fig. 2. Head squamation. Above: A. aliniger, ]^ICZ 57403; below: A. chlorociianus, MCZ 57473. Both specimens from 7 km N Carpiiitero, Prov. San Juan, Dominican Republic. area of scaleless skin which represents the area formerly occupied by the orange spot (Fig. 3). Since the dewlap is hardly developed, almost non-functional, it is very possible that the orange, made more conspicuous by the black behind it, is a flash pattern used in some fashion in high intensity agonistic behavior. This is at present a mere suspicion ; the only specimens of aliniger seen have been merely captured and preserved, or only very briefly ob- served. 6 BREVIORA No. 227 The best report of the ecology and behavior of A. aliniger is that by James Lazell (field notes, December, 1963) at Paraje la Palma near Constanza: "Up in largish trees along the edge of the woods by the stream. Since they retreat upwards, collect- ing them is merely a problem of having a long enough pole. Just like coelestinus-chlorocyanus, therefore, in habits — except for the vertical flattening in display. In display the whole body is ver- tically compressed — showing much of the venter. Extended, the throat fan is quite small." Lazell's observations were unfor- tunately terminated by rain which prevented him from seeing any further specimens during the remainder of his stay. The vertical compression, according to Lazell's sketch from life, emphasizes the yellow color of the belly. Lazell saw, however, no instance of exposure of the black and orange axillary pattern. A number of descriptions of color in life are available for aliniger. They appear to indicate not only that the green of the light phase is different in tone from that of the species of subgroup 1 but also that there is a greater play of patterns and tones in the darker state. (See also ^4. singularis below.) In view of our very inade- quate knowledge of the species, I quote these color descriptions in full. Mertens (1939, translated): "In life this Anolis when caught was a uniform chocolate-brown with a large pitch black spot in each axillary region. In the bag in which it was transported the lizard turned blue-green, the black axillary spot remaining unal- tered. The dewlap was bright green, the tip of the tail black." Rand (field notes, 1958). Female — 7 km N of Carpintero: "Gray green above, head grayer, upper lip white, rear of thigh with a dark line, axilla bright yellow with smaller black spot be- hind it." Lazell (field notes, 1963): "Duller and bluer green than coe- lestinus-chlorocyanus. Venter, throat fan and frosted spots on sides (especially shoulder region) bright saffron yellow. Orbital skin butter yellow. Axillary spot plain black. "Changes, when unhapjjy, to lichenate frosted grey-brown with white. A pattern of large dark bilaterally arranged spots emerges with transverse bands — especially posteriorly. Loreals and ir- regular stripe through eye emerge slate blackish." Schwartz (field notes, 1964). ASFS V 1625— 12 km S of Ca- brera de Loma: "Dorsal ground color blackish brown to olive with green sacral 'butterfly' marking and tail banded olive and cream. A white subocular mark. Ventral ground color dull greenish." 1965 HISPANIOLAN GREEN ANGLES Fig. 3. Axillary squamation. Above: A. chlorocyanus, MCZ 57473, 7 km N Carpintero, Prov. San Juan, Dominican Republic; below: A. aliniger, MCZ 79341, Paraje La Palma, Municipio Constanza, Prov. La Vega, Dominican Republic. 8 BREVIORA No. 227 The type locality of Anolis chlorocyanus aliniger Mertens was Paso Bajito on the northern rim of Valle Constanza in La Vega Province, Dominican Republic. Most of the recent specimens have been taken within that valley itself. However, a female has been taken in San Juan Province and a male in Dajabon Province. All specimens are from elevations near or above 600 meters. These Dominican Republic localities appeared to place A. aliniger as an anole of the Cordillera Central or its immediate foothills, present at moderate elevations — within the pine zone — and widely sympatric with A . chlorocyanus. It was with some surprise, therefore, that two specimens were found in a collection of .4. coelestinus made by the Whiteman brothers at Furcy in Haiti. Furcy is south of the Cul de Sac trough that separates Hispaniola into northern and southern faunal areas — formerly separate islands. It is not, of course, especially surprising that a form common in the northern faunal area should sometimes penetrate some dis- tance or even extensively into the southern area. This seems a frequent phenomenon. What is surprising about the discovery of aliniger at Furcy, Haiti at ca. 1500 meters at the north margin of the Massif de la Selle is that this is not a lowland species, which could without special difficulty cross the dry and hot Cul de Sac Plain, but a creature of higher elevations, cooler temperatures and of quite different floral associations from those which are now seen in either the desertic or mesic habitats of the lowlands. Essen- tially, the discovery of aliniger at Furcy is analogous to finding a species of one island present on one cape of a neighboring island. The ecology of aliniger appears to require that at some past time the Cul de Sac gap must have been climatically and floristi- cally passable for the species: the passage from one montane island to the other must have been made at a time of much cooler temperatures — presumably the last cool-wet period of the Pleistocene. The specimens from Furcy remain the only record of the species in Haiti. Doubtless in part this absence of record is an artifact of collecting. A spur of the Cordillera Central enters the northeast of Haiti, but little or no collecting has been done there. At present the known localities for aliniger are: Dominican Republic. La Vega Province. Below Paso Bajito at about 900 m (type locality), Senckenberg 25825. Valle Constanza, MCZ 56912. Tireo near Constanza, MCZ 56913-15. Paraje La Palma, east of Constanza, MCZ 75140-41, 79341-43. El Rio, AMNH 44852-53. San Juan Province. Seven kilometers north of Carpintero, 1965 HISPAOTOLAN GREEN ANGLES 9 MCZ 57463. Dajabon Province. Twelve kilometers south of Loma de Cabrera, ASFS V 1625. Haiti. Departement de VOuesl. Furcy, MCZ 63444-45. Over most of the southern island of Hispaniola — south of the Cul de Sac trough in both Haiti and Hispaniola — the absence of alinigcr is plausibly accounted for by the presence of a closely related species — essentially alinigcr without the axillary light and dark spots. This peculiar species is known from very few specimens, ever3^one of which is from a different locality. I have, therefore, called this new species by the Latin adjective which means "one at a time." Anolis singularis new species Holotypc: MCZ 72043, adult male, Pourcine, Massif de la Hotte, Haiti, collected by Frangois Vuilleumier, 31 December 1962. Parafypes: MCZ 13778, La Gonave Island, Haiti, G. AL Allen coll. 1919; YPAI 3229, Nan Cafe, La Gonave Island, P. Humphrey and Sarita Van Vleck coll., March 26, 1959; YPM 3194, Foret des Pins, Massif de la Selle, Haiti, P. Humphrey and Sarita Van Vleck coll., February 19, 1959; AMNH 51728, Valle de Polo, Dominican Republic, W. G. Hassler coll., September 14, 1932; ASFS V 2608, 5 km NE Los Arroyos, 5800 feet {ca. 1750 meters) elevation, Pedernales, Dominican Republic, D. C. Leber coll., June 27, 1984; ASFS V 2985, 30 km N of Pedernales, 2680 feet {ca. 810 meters), Pedernales, Dominican Republic, hatched from an egg collected under a limestone rock in Cajetal, July 3, 1964, by R. Thomas; hatched July 16, 1964. Diagnosis: An anole most closely allied to ,4. alinigcr Mertens but differing in the absence of a scaleless, highly pigmented axillary area. Like A. alinigcr, the new species differs from A. chlorocyanus and A. coelestinus in the greater enlargement of scales around the interparietal (Fig. 4), and in reduction of the dewlap in males. Description: Head scales rather large, 6-8 scales across snout between second canthals. Frontal depression feeble, scales within it not smaller than those surrounding it. Five to seven scales border rostral posteriorly. Anterior nasal scale in contact with rostral. Three to four scales between supranasals. Supraorbital semicircles separated by 1-2 scales, separated by one row of scales from supraocular disk. Seven to eleven smooth scales in the poorly defined supraocular disk which is separated from supraciliary by 2-3 rows of granular or sub- granular scales. One rather short supraciliary, flanked medially 10 BREVIORA No. 227 by smaller scales and continued posteriorly by granules. Six to eight canthals, second and third largest, decreasing thence for- ward, anteriormost under rostral. Three to five loreal rows. Fig. 4. Head squamation. Above: A. singularis Holotype, MCZ 72043; below: A. coelestinus, MCZ 74708. Both specimens from Pourcine, Massif de la Hotte, Haiti. Temporal scales subgranular, smallest in center. A poorly defined intertemporal line. Supratcmporals larger than temporals, increasing in size toward interparietal. Interparietal larger or slightly smaller than ear, separated from semicircles by 2-3 scales. Scales surrounding interparietal largest laterally but a very distinct zone of 5-6 rows of enlarged scales posterior to inter- parietal. 1965 HISPANIOLAN GREEN ANGLES 11 Suboculars broadly in contact with supralabials, anteriorly sepa- rated from canthal ridge by two scales, posteriorly merging into temporals. Seven to eight supralabials to center of eye. Mentals broader than long, in contact posteriorly with 6-7 scales between infralabials (2 sublabials and 4-5 smaller scales) ; 5-7 sub- labials in contact with infralabials. Central throat scales small, swollen, smooth. Dewlap: Hardly differentiated, only indicated as a longitudi- nal fold; scales larger than throat scales, smaller than ventrals. Trunk: Middorsals hardly enlarged, grading very gradually into flank granules. Axilla with normal granular scales. Ven- trals larger, smooth, quadrate, juxtaposed, transverse. Postanals enlarged in male. Scales posterior to vent smooth. Limbs: Largest forelimb scales smaller, largest hindlimb scales larger or smaller than ventrals, both weakly unicarinate. Supra- digital scales smooth; 21-23 lamellae (27 in one hatchling) under phalanges ii and iii of fourth toe. Tail : Verticils with four scales above, three below. Scales sub- equal. Color in life: YPM 3194, Foret des Pins, Haiti: Emerald green labials and limbs. Yellow eyelids, oUve head and back, yellow green venter, last quarter of tail black with yellow tip. YPM 3229, Nan Cafe, La Gonave Island, Haiti: Head silvery grey mottled with turquoise and brown, the latter extending to the shoulder. A brown spot just behind shoulder. Back silver gray mottled with turquoise. Tail and limbs very pale buff, tail banded with sky blue. Underparts pale turquoise becoming lemon on femora and at vent. ASFS V 2608. 5 km NE Los Arroyos, Dominican Republic: Dorsal ground color gray to brown, a pattern of dark chevron- shaped middorsal blotches and smaller, roughly triangular lateral blotches. Light lateral stripes. Venter faint rust, speckled with greenish. ASFS V 2985 (hatchhng). 30 km N Pedernales, Dominican Republic: Dorsal ground color green with longitudinal darker stripes, especially middorsally. Species status. On present evidence, A. aliniger and A. singu- laris are wholly allopatric. The question of species or subspecies status does therefore arise. My decision to describe singularis as a species is based upon two considerations: (1) a high valuation placed upon the axillary differences in pigmentation and squama- tion, which I assume to imply behavioral differences; (2) the presence of typical aliniger at Furcy appears to imply a capacity 12 BREVIORA No. 227 to invade the territory of an allied taxon — a feature more prob- able for a species than a subspecies. Both these points require further elaboration. (1) It is frequent in Anolis for the most conspicuous and taxo- nomically useful differences between very closely allied species to be in structures such as the dewlap that are involved in species recognition and intraspecies agonistic behavior. Thus, in sub- group 1 of the present paper, the southern island form A. coe- lestinus has a dewlap with pale skin and small scales while the northern island species A. chlorocyanus has a dewlap with pig- mented skin and greatly enlarged scales. A. semilineatus and A. olssoni differ in an exactly similar way, A. semilineatus having the dewlap skin pale, the dewlap scales small, A. olssoni having pigmented dewlap skin and greatly enlarged dewlap scales. (In the latter case there are other strong differences.) Dewlap color, however, does vary intraspecifically in Anolis (e.g. in A. distichus, A. hrevirostris, A. cybotes) and sometimes individually, as, of course, body pattern may do also. To what extent such color variation disturbs species recognition is quite un- known, and equally unknown are any differences in display be- havior between allied full species such as .4. coelestinus and .4. chlorocyanus or A. semilineatus and A. olssoni. In the present case, I have chosen to infer. that the orange axillary spot of aliniger, attended as it is by loss of normal squa- mation in that area and contrasted with the unmodified axilla of singularis, is comparable to the species recognition marks that distinguish species pairs and not to the simpler color variations that may occur intraspecifically. (2) The species pair A. coelestinus- A. chlorocyanus may be the only instance in which the geographic boundary between species is exactly that area of the island which at one time, as an open seaway, divided Hispaniola into northern and southern parts. It is certain that this seaway has been profoundly important for origin of many widel}^ distributed Hispaniolan species; Mertens (1939) cari^^ recognized its importance. (See also Williams, 1961.) But it is in fact very unusual for allied species which have origin- ated north and south of this important zoogcographic boundary to be precisely limited by it now that the seaway has become dry land. The situation which I described in the species pair .4. srmi- linvatus-A. olssoni is commoner: one or the other or perhaps ijoth members of the species pair interpenetrate the range of the other to a greater or lesser extent. Usually, ii would appear, there is, 1965 HISPANIOLAN GREEN ANGLES 13 as in .1. scmilincatus-A. olssoni, sufficient ecological difference be- tween the members of a species pair to permit this, but in the case of A. coelesfiniis-A . chlorocyanus the ecologies are too similar to permit anything but a stand-off, with a very narrow, perhaps fluctuating zone of sympatry. The case of A. aUnigcr-A. singnlaris, which is surely another example of a northern island-southern island pair, differs from that of ^4. coelestinus-A. chlorocyanus in that these are inhabitants of cooler upland areas and should now be quite separated by the whole extent of the hot Cul de Sac Plain and by much of the mesic woodlands on either side. That any A. aliniger occurs on the south side of the Cul de Sac Plain indicates, as I have suggested above, the former existence of a zone of passage right across the Cul de Sac for species now characteristic of cool upland climates. At such a period A. aliniger invaded the southern island and may well have met A. singnlaris. If contact occurred, certainly no merging of populations resulted. The Furcy aliniger are quite un- modified, as typical as any from Valle Constanza. Perhaps there was a slight ecological difference with A. singnlaris living at higher elevations than A. aliniger. (The known true south island records of A. singnlaris would fit this pattern but the Gonave records would not.) Or perhaps ^4. aliniger displaced A. singnlaris in part of the latter 's range. Too little is known about either taxon to permit any but the most provisional hypothesis. It does, however, seem plausible to regard the invasion of the southern island by aliniger, without any indication of compromise of its characters, as evidence that it and singnlaris have indeed achieved species status and to infer that A. aliniger and A. singnlaris are, like A. coelestinus-A . chlorocyanus, too similar ecologically to occur together. ACKNOWLEDGMENTS I am indebted to Dr. Albert Schwartz (Albert Schwartz Field Series, ASFS), Mr. C. M. Bogert, American ]Museum of Natural History (A]\1NH), and Dr. Charles Reed, Yale Peabody Museum (YP]\1) for the privilege of examining specimens under their care. This paper, as part of a general study of West Indian anoles, has been supported by National Science Foundation grants NSF- G 16066 and NSF-GB-2444. The illustrations were prepared by Mr. Joshua Clark. I have had the advantage of utilizing counts and tabulations made by Dr. A. S. Rand. 14 BREVIORA No. 227 in c: a> 3 C O >. o o o a c a a a en w -1.3 a o CO Q 6 ai c c \/ D ^ 1965 HISPANIOLAN GREEN ANGLES 15 REFERENCES COCHKAN, D. M. 1941. The herpetology of Hispaniola. Bull. U. S. Nat. Mus., 177: 1-398. Mertens, R. 1939. Herpetologische Ergobnisse einer Roise nach der Insel Hispaniola, Westiudien. Abhandl. Senckenberg. Naturf. Ges., 449: 1-S4. Williams, E. E. 1961. Notes on Hispaniolan herpetology. 3. The evolution and rela- tionships of the Anolis seniilinealus group. Breviora, Mus. Comp. Zool., No. 136: 1-S. (Received 10 May, 1965.) Table 1. Scale characters of the Hispaniolan green anoles coelestinus chlorocijanus aliniger singularis scales across snout 6-10 4-7 6-9 6-8 loreal rows 4-7 3-5 3-4 3-5 scales between supraorbital semicircles 1-3 0-2 0-1 1-2 scales between inter- parietal and semicircles 2-6 1-5 2-4 2-3 lamellae 23-32 27-34 21-24 21-23' 27 in one hatchling, ASFS V 2985 Scale charact coelestinus scales behind interparietal grad- ing rather gradu- ally into dorsal scales Table 2. ers differentiating the Hispaniolan green anoles chlorocyanus as in coelestinus aliniger Scales behind interparietal not grading into dorsals but several rows abruptly larger singularis as in aliniger size of scales of dewlap ca. = ventrals size of scales of dewlap > > ventrals dewlap vestigial, size of scales of dewlap area < ventrals as in aliniger supradigital scales multicarinate as in coelestinus supradigital scales as in aliniger smooth axillary area with normal granular squamation as in coelestinus axillary area with- as in coelestinus out granular scales 16 BREVIORA No. 227 Table 3. Color characters differentiating the Hispaniolan green anoles coelestin us a white line on supralabials pro- duced posteriorly to above the shoulders (absent in He Vache popu- lation) chlorocyanus aliniger never any as in chlorocijanus white markings on side of head or neck singularis as in chlorocyanus axillary area with vertical black and light bars or unpatterned as in coelestinus axillary area with a light spot (orange in life) followed by an ink-black blotch of larger or smaller extent — both concealed by normal position of arm as in coelestinus dewlap scales blue, skin olive-gray dewlap scales white to blue, skin with black or (cyanostictus) cadmium yellow basal spot dewlap skin and scales greenish as in aliniger BREVIORA Mesemm of Coimpsirative Zoology Cambridge, Mass. September 15, 1965 Number 228 RELATIONSHIPS AMONG INDO-AUSTRALIAN ZOSTEROPIDAE (AVES) By Ernst Mayr During- preparation of the maniiscriiDt of the Zosteropidae for Peters' Checklist of Birds of the World, I reviewed once more the relationships of the Indo-Anstralian -white-eyes. The first attempt to arrange these species in natural groups was made by Stresemann in a pioneering work in 1931. The revisionary studies of the ensuing 25 years were summarized by G. F. Mees in a very thorough systematic review of the Indo-Australian Zosteropidae (1957, 1961). Mees' work is based not only on an exhaustive analysis of the literature, but also on an examination of most of the available museum material. His fundamental monograph will remain for many decades the basis of all taxo- nomie research in Indo-Australian white-eyes. In a few cases my own interpretation of relationship differs from his, resulting in a somewhat different sequence of species. This paper may serve to explain the reason why, in these few instances, I have adopted a sequence different from that chosen by Mees in his systematic review. I am deeply indebted to Dr. Mees for much valuable information on Zosteropidae. Absolute criteria of relationship in the Zosteropidae do not exist at the present time, and neither Mees nor I can prove that the sequence adopted by one of us is "more correct" than the sequence adopted by the other. White-eyes have characteristic songs and call notes, and perhaps analysis of these and other behavioral characters may lead to a better understanding of rela- tionships. I know of no other group of birds in which close relatives, for example the subspecies of Zostcrops atrifrons or the semispecies of the superspecies griseotiitcta, may differ more from each other than do distantly related species. Indeed some 2 BREVIORA No. 228 Oriental species are almost indistinguishable from African forms, from which they must have been isolated since remote times. ZosTEROPS CEYLONENSis Holdsworth This species is often considered closely related to Z. palpcbrosa, and Mees (1957, p. 22) lists it immediately preceding palpchrosa. Yet, he states correctly, "morphologically Z. ceylonensis is un- doubtedly rather closer to several other species than to palpe- hrosa" (ibid., p. 26). For this reason I have placed it after palpchrosa and closer to some of the Indonesian species from which it might possibly have been derived. ZoSTEROPS ERYTHROPLEURA Swinhoe This species, with its rufous flanks, does not resemble any of the species (pajpehrosa and japonica) with which it is usually placed. Being also the only white-eye restricted to the Palearc- tic region, I prefer to indicate its distinctness by placing it first in the sequence, as had been previously done by Stresemann (1931, p. 206). ZoSTEROPS CONSPICILLATA Kittlitz Mees (1957) lists this Micronesian bird as the last species of the genus. To me this species does not seem nearly as aberrant as Z. cinerea. Indeed, in spite of its paleness, a frequent char- acter in island birds, conspiciUata resembles in some ways the japonica-palpehrosa assemblage. Since several Micronesian birds were derived from the west (e.g. Acroccphalus) , relationship of conspiciUata with japonica is a distinct possibility. It is hoped that placing the species earlier in the sequence will bring it closer to its real relatives. ZoSTEROPS WALLACEi Finsch As Mees has stated rightly, this is an old and peculiar en- demic. It seems distantly related to the western group of species (atricapilla, everetti, nigrorum, and others) and I have there- fore placed it earlier in the sequence. ZoSTEROPS FLAVA-CHLORIS-LUTEA grOUp I have adopted Mees' sequence for the sake of uniformity. I still feel, how^ever, that these species are more closely related 1965 RELATIONSHIPS AMONO ZOSTEROPIDAE 3 to each other than believed by Mees. To separate lutea from chloris by 14 other species does not seem to be the best possible arrangement. ZosTEROPS CONSOBRINORUM Meyer This species is so similar to Z. chloris citrmclla that in any other genus one M'ould consider them conspecific. Even though I have retained Z. consohrinorum as a full species, I have placed it next to chloris, while I now treat the very distinct peripheral forms grayi and uropygialis as full species (see also Mees, 1953, 1961). ZoSTEROPS ATRIFRONS grOUp I agree, on the whole, with Mees' (1961) arrangement. How- ever, Z. atriceps is best listed before the atrifrons-minor-deli- catula series, because the latter is close to the forms on the islands east and northeast of New Guinea {meeki, hypoxantha) and should not be separated from them by atriceps. In this group of species close relatives may appear rather different (belly, yellow or white ; forehead, black or olive ; eye- ring, absent or broad; throat, orange, yellow or whitish). The delimitation of the species is therefore a difficult task. Two of the most distinct forms, minor and delicaUila of New Guinea are connected by the intermediate forms chrysolaema and rothschildi. Mees (1961) quite rightly combines atrifrons and minor in a single species, but if one goes that far one must also include meeki (close to delicatula) and hypoxantha. I have maintained Z. mysorensis as a separate species, because in its combination of characters (no yellow on throat, heavy gray wash on breast and flanks, olive forehead, absence of eye-ring, blackish loral region, and pale yellow under tail coverts) it resembles some other species (e.g. ugiensis) more closely than atrifrons. ZoSTEROPS NATALIS Listcr Mees (1957) notwithstanding, there is much to indicate that this species is closer to one of the east Indonesian or Australian species {chloris, lutea, etc.) than to any of the Malaysian species. I agree in this with Stresemann and Chasen. Mees (1957) makes the peculiar comment that my (Mayr, 1944) 4 BREVIORA No. 228 association of this species with lutca, instead of with citrinella, is an ' ' unfortunate choice, ' ' overlooking^ that I include citrinella in lutea in the cited paper. ZosTEROPS RENDOVAE Tristram The description of rendovac was based on a Rendova Island specimen misidentified as Tephras olivaceus Ramsay, but newly named rendovae since the name olivaceus Ramsay (nee Certhia olivacea Linnaeus 1766) was considered unavailable. That the original author, Tristram, considered rendovae a name for the Rendova bird far more than a replacement name for olivaceus is evident from his subsequent statement (1894, p. 30) : "I give a figure (PI. Ill, fig. 2) of Zosterops rendovac of Rendova Island . . . which I described in the Ibis for 1882, p. 135." Galbraith (1957) has well stated the reasons for retaining the name rendovae for the Rendova White-Eye to which it had been applied universally from 1882-1955, including general books in ornithology and evolutionary biology. Zosterops lateralis Latham The arrangement of the Australian races is largely based on unpublished research kindly made available to me by Drs. A. Keast and G. F. Mees. The resulting sequence of species of Indo-Australian Zos- terops is as follows : erythropleura uropygialis japonica anomala palpehrosa (ttric(ps ccylonensis atrifrons conspicillata mysorensis salvadorii fuscicapilla atricapilla h urucnsis everetti kuehni nigrorum novaeguincae montana metcalfii wallacei natalis flava lutea chloris griseotinctu consohrinorum renneUinna grayi vellaJarrlla 1965 RELATIONSHIPS AMONG ZOSTEROPIDAE 5 lutcirostris minuia rendovae xanthochroa murphyi lateralis ugiensis strenua stresemanni tenuirostris sanctaecrucis alhogularis samocnsis inornata cxplorator cincrea flavifrons The ABERRANT GENERA OF InDO- AUSTRALIA Some white-eyes differ from the normal structure or colora- tion of the genus Zosterops, as represented by a species like palpchrosa or lateralis, to such an extent that they have been separated generieally. If all the larger white-eyes, with some- what aberrant coloration and a longer or heavier bill, could be placed in a single genus, no one would mind. The fact of the matter is, how^ever, that 13 genera have been proposed to accommodate 18 species. Excluding Lophozosterops (with 6 species), there are 11 generic names for 13 species. Some further simplification is possible by combining Sanfordia with Woodfordia, as well as the Micronesian Kuharyum, Megazos- terops, and Cinnyrorhyncha with Rukia. In the "Oreosterops group" of authors, Mees (1953, pp. 57-66) recognizes six genera, reduced in 1957 to five by combining Apoia with Lopho- zosterops. Of these five genera, three {Madanga, Tephrozos- terops, and Oculocincta) are monotypic, while Heleia has two species in one superspecies. The five genera seem to form a natural group and a renewed analysis may result in further lumping, perhaps of all five genera into Heleia. The sequence chosen by Mees (1953) does not seem quite natural. By starting with the species that is most like Zosterops and also keeping the pattern of geographic distribution in mind, we arrive at the following sequence: Tephrozosterops (staJhcri), Madanga (ruficollis), Lophozosterops {pinaiae, goodfellowi, squamiceps, javanica, superciliaris, dohertyi), Oculocincta (squamifrons) , and Heleia {miielleri, crassirostris) . I entirely agree with Mees {in litt.) that Hypocryptadius Hartert does not appear to be a white-eye. Fresh material is needed to determine whether it should go to the Monarcliini, Sylviinae, Timaliinae, or some other group. 6 BREVIORA No. 228 LITEBATURE CITED Galbraith, I. C. J. 1957. On the application of the name Zosterops rendovae Tristram, 1882. Bull. Brit. Orn. Club, 77:10-16. Mayr, Ernst 1944. The birds of Timor and Sumba. Bull. Amer. Mus. Nat. Hist., 83:123-194. Mees, G. F. 1953. An attempt at a natural classification of certain Zosteropidae of the Indo-Australian archipelago. Zool. Meded., 32:57-68. 1957. A systematic review of the Indo-Australian Zosteropidae (part I). Zool. Verh., No. 35: 1-204. 1961. A systematic review of the Indo-Australian Zosteropidae (part II). Zool. Verh., No. 50: 1-168. Stresemann, E. 1931. Die Zosteropiden der indo-australischen Region. Mitt. Zool. Mus. Berlin, 17:201-238. Tristram, H. B. 1894. On some birds from Bugotu, Solomon Islands, and Santa Cruz. Ibis, (sixth series) 6:28-31. (Received 5 June, 1965.) JL-^t/ BREVIORA MmseiuiiM of Coniiparative Zoology Cambridge, Mass. September 15, 1965 Number 229 THE GENUS DARLINGTONIA (SERPENTES) IN HISPANIOLA, INCLUDING A NEW SUBSPECIES FROM THE DOMINICAN REPUBLIC By Albert Schwartz AND Richard Thomas 10,000 SW 84th St., Miami, Florida 33143 The West Indian island of Hispaniola is outstanding for the rela- tively large number of endemic snake genera which inhabit it. These include laltris Cope (with two species), Uromacer Dumeril and Bibron (with apparently two or three species), Hypsirhynchus Giinther, and Darlingtonia Cochran. The latter is a small snake (snout-vent length to about 300 mm) which was described (Cochran, 1935) on the basis of a single Haitian individual col- lected by P. J. Darlington from Roche Croix, about 5000 feet (1515 meters) altitude, near Pic Macaya in the Massif de la Hotte. Since the genus was first described, the snake has been found to be not uncommon in Haiti in the vicinity of Furcy and Kenscoff on the Montague Noire. Additional specimens from the Massif de la Hotte remain rare, and we have seen only one other snake (in addition to the type) from Camp Perrin in that region. Considering that there is a distinct faunal resemblance between the Massif de la Hotte, Massif de la Selle, and Sierra de Baoruco, all serially arranged from west to east along the south island of Hispaniola, we expected that ultimately Darlingtonia might be taken in the last named range. Such indeed was the case when in the summer of 1964 the junior author and David C. Leber succeeded in taking two females at a sawmill in the Baoruco. It was later learned that a single Darlingtonia had been collected in the Repiiblica Domini- cana by W. G. Hassler in 1935. In an attempt to compare these eastern specimens with the more western populations, we have borrowed Darlingtonia from the following collections: American Museum of Natural History BREVIORA No. 229 (AMNH) ; Museum of Comparative Zoology, Harvard University (MCZ); United States National Museum (USNM); INIuseum of Zoology, University of Michigan (UMMZ). We wish to thank Charles M. Bogert, Grace M. Tilger, Ernest E. Williams, Doris M. Cochran, James A. Peters, Charles F. Walker, and George R. Zug for the loan of these small serpents; Mr. Leber and Ronald F. Klinikowski aided us in assembling material in the field (designated as ASFS), and Mr. Klinikowski has likewise made some of the illustrations for the present paper. We have examined 30 speci- mens in all, including the holotype (MCZ 38252) of the species, with all but five of these originating in the Furcy area. We are thus hampered in comparing these Furcy snakes with topotypical material from the east in the La Hotte, but for the moment there is no choice but to accept the agreement of the Montague Noire material with that from the La Hotte (admittedly an unlikely possibility, vide infra). The snakes from the Sierra de Baoruco, which are distinct from the more western populations, may be named : Fig. 1. Darlinglonia haedana haetiana, pattern at midbody, ASFS X20S0, Peneau, Bassin Bleu, 5000 feet, D^pt. de I'Ouest, Haiti. 1965 SNAKE GENUS DARLINGTONIA Darlingtonia haetiana perfector^ new subspecies Holotijpe: MCZ 77217, a female, from 24 km SW Barahona, 3700 feet (1221 meters), Barahoua Province, Repuhlica Domini- cana, taken by David C. Leber on 6 July 19G4. Original number ASFS V 2897. Paratypes: ASFS V 2898, same data as type, but collected by Richard Thomas; AIMNH 49738, near Polo, 3000 feet (910 meters), Barahona Province, Republica Dominicana, W. G. Hassler, 19 August 1935. Distribution: Known only from the Sierra de Baoruco in the Republica Dominicana, and probably occurring throughout moderate to higher elevations in that range. Diagnosis: A subspecies of Darlingtonia haetiana distinguished from the nominate race by a combination of a wide buffy nuchal collar (Fig. 3), a bolder and more contrasting longitudinally lined and dotted dorsal pattern (Fig. 2), and by less ventral and sub- caudal scales, giving a total underbody scale count from 174 to 178 in contrast to 181 to 193 (both sexes) in D. h. haetiana. Fig. 2. Darlingtonia haetiana perfector, pattern at midbody, MCZ 77217, holotype, 24 km SW Barahona, .3700 feet, Barahona Province, Republica Dominicana. 'From the Latin for "conipleter, finisher" in aUusion to the fact that with the specimens from Barahona the genus is known from the entire south island, whereas previously it had been known only from Morne La Selle and Massif de la Hotte. 4 BREVIORA No. 229 Description of holotype: A female with the following measure- ments and scale counts: snout-vent length, 190 mm, tail, 47 mm; ventral scales 133 (counted as suggested by Bowling, 1951), sub- caudal scales in 41 pairs; anal single; supralabials 7/7; infralabials 8/8; loreal absent; preoculars 1/1; postoculars 2/2; temporals 1 + 1 on right side, 1+2 on left side; dorsal scale rows 19-19-17. Coloration and pattern in life: dorsolateral area light brown with a middorsal dark reddish brown zone including a black middorsal stripe; a narrow dark brown conspicuous lateral stripe on scale rows 4 and 5 on each side; mid ventral region with a wide band al- most covering all of ventral scales including within it darker black blotches, the entire band with a blue iridescence; lowermost scale row on each side, and central part of second scale row and some- times third scale row and lateralmost ends of ventrals brick red; 20 and 22 lateral dots on sixth to eighth scale rows on each side, each individual dot involving one, two, or three scales on adjacent rows, and yellowish orange with black margins in life. Tail brown above, black below. Head without a definitive pattern, but gen- erally blackish to reddish brown or orange on parietals, the brighter and paler color delimiting a vague, dark, T-shaped figure with its bar across the supraoculars and frontal, and the stem along the parietal suture, joining on the neck the middorsal black longi- tudinal band; a pale bar across the anterior half of the frontal, the prefrontals and supranasals variously marbled with brown and paler. A buffy collar, edged posteriorly with black, and two scales in width, goes from the angle of the jaws on either side, and is interrupted middorsally by the median black band. Variation: The female paratype (ASFS \'2898) has a snout- vent length of 180 mm and a tail length of 44 mm; there are 136 ventral scales and 40 pairs of subcaudal scales. The supra- and infralabials, pre- and postoculars are as in the type. The temporals are 0 + 2 on both sides, the single anterior temporal being fused with the parietal. The scale rows are 19-19-17, and the loreal is absent. In color and pattern the paratype is very like the tj^pe, except that, due to a smaller amount of pale color on the anterior half of the frontal, the transverse dark bar of the T-shaped cephalic figure is wider and more conspicuous. The buffy collar is present and readily visible. The male paratype (AAINH 49738) has a snout-vent length of 216 mm and a tail length of 63 mm; there are 132 ventral scales and 46 pairs of subcaudal scales. The supralabials, pre- and postoculars, and temporals are like those of the type, except that there are 1 + 2 temporals on each side. The infralabials are 7 on each side. The scale rows are 19-19-17, and the loreal is absent. 1965 SNAKE GENUS DARLINGTONIA Although tlio male paratype is darker (possibly due to length of preservation) than the two smaller female pcrfector, the longitudi- nal zonation is still visible. The nuchal collar is bold and promi- nent. The head pattern is like that described for the type, except that the amount of pale coloration across the anterior portion of the frontal is less, thus making the transverse dark bar of the T-shaped figure broader. The ventral scales are entirely dark, with the lateralmost tips irregularly stippled light and dark. Fig. 3. Dorsal view of heads of Dariingionia haetiana: left, D. h. haeiiana, ASFS X2080; right, D. h. perfector, MCZ 77217. Comparisons and discussion of variation: Since there are more specimens available from the Furcy area than elsewhere, it is most profitable to first discuss the variation in snakes from this region, and then compare them both with the two western and three eastern serpents. At Furcy, the scalation shows the following variation: ventrals in males (13 specimens) range from 137 to 144 (mean 139.7), ventrals in females (12 specimens) having the same range (mean 139.9). Subcaudals are 46 to 51 (mean 48.3) in males, and 41 to 50 (mean 45.3) in females. Ventrals + subcaudals are 185 to 193 (mean 188.0) in males, and 183 to 191 (mean 185.4) in females. Females have slightly less total underbody scales than do males. The largest male has a snout- vent length of 281 mm with a tail of 82 mm, whereas the largest female measures 305 in snout-vent length with a tail length of 83 mm. All specimens lack a loreal, and have 1 /I preoculars and 2/2 postoculars. The labials are normally 7/7 above and 8/8 below, although one snake has 7/8 supralabials, three snakes have 7/7 infralabials, and four have 7 /8 infralabials. The temporals are either 1 + 2 or 1 + 1 ; four 6 BREVIORA No. 229 snakes have 1 + 1 on each side, four have 1 + 1 and 1+2, and the balance have 1 + 2 on each side. The niimljer of lateral dots varies from 0 (a single snake has no lateral dotting visible) and 4 (one dot on one side and three on the other), to a maximum of 54 (27 dots on each side) ; usually the number of dots on the two sides is not identical. The scale row formula is 19-19-17 with four exceptions; three of these are 20-19-17, and the other 21-19-17. The Furcy material shows an interesting ontogenetic pattern change. A series of six young snakes ranging in snout-vent lengths from 104 mm to 166 mm shows the development of the ventral dark band. In the smallest of these snakes (which still has evi- dence of the umbilicus at ventrals 115 to 117) the entire venter has a pale ground color with scattered dark brown or black blotches. The two snakes next in size (108 and 118 mm snout-vent) show the same condition, although the larger of these two shows the beginning of deposition of dark pigment ventrally. The next largest snake Avith a snout- vent length of 153 mm shows the obliterative effect of additional dark pigment so that the individual dark ventral blotches have become somewhat obscvn-cd. Finally, in two snakes with snout- vent lengths of 165 and 166 mm, the adult condition of a black or dark venter, with the original ju venal dark blotches barely discernible, is attained. In these young snakes the collar is pale and conspicuous; with increasing age the collar becomes fainter and less obvious so that in large adults it is much reduced and may be seen only as a small restricted pale nuchal area with some dark pigment deposited about it. Only two snakes from Furcy are as pale as are the two perfedor', and these two individuals have the lateral lines and dots as conspicuous as do the Baoruco snakes. In general, the Furcy snakes are dark brown: description in life of a series of D. h. haetiana from Bassin Bleu, Peneau, Dept. de FOuest, Haiti, noted the coloration as dark brown with a middorsal dark brown longitudinal band on a slightly lighter brown color (Fig. 1). The ventral ground color was dark brown with occasional whitish edges to the ventra's. The lateral margins of ventrals and first scale row on each side were brick red to orange-red, the extent of this color greatest anteriorly. The dorsal blotches were buffy-tan, outlined in brown. The collar (Fig. 3) was very much reduced or absent in this series which ranged in size from 159 to 245 mm in snout-vent length. Note also the lack of mention of the lateral stripe which was not obvious because of the dark lateral coloration. From the abo\e details in both scalation and coloration, it i.s apparent that 1). h. pcrfcclor differs not only in having fewer 1965 SNAKE GENUS DARLINGTONIA 7 uiiderbody scales than docs D. h. haetiana, but also in having a prominent nuchal collar and prominent lateral lines. There are two specimens of Darlimjtonia from the JNIassif de la Hotte, the type and another from Camp Perrin. These two indi- viduals are both females, the type having a snout-vent length of 294 mm and the second specimen a snout-vent length of 247 mm. The tail of the type (88 mm) is unusually long, longer than any specimen of either sex, regardless of snout- vent length, from the Furc}^ area. The ventrals in these two snakes are 186 and 137, and the subcaudals 49 and 51, with total underbody scales 185 and 188. There are no other scale differences, although the Camp Perrin specimen has 1 /2 preoculars and the type has 2+1 and 2 + 2 temporals. Both are pallid snakes with conspicuous lateral lines and without collars; the amount of ventral darkening is variable, involving almost the entire width of the ventrals in the type, and with a clear reddish area on the lateral ends of the ven- trals in the second specimen. Using ventrals minus subcaudals as an index, the two western snakes both (with 86) lie just outside the range of the series of females from Furcy (with a range of 87 to 102). Whether this hints at a basic difference between these two populations can be determined only by the acquisition of more material from the Massif de la Hotte. The extremely long tail of the type and, in the two La Hotte females, the relatively high number of subcaudals (49 and 51 versus 41 to 50 in Furcy females), and the tail /snout-vent ratio of 29.9 and 31.2 (versus 23.7 to 27.2 in Furcy females) are all likewise suggestive of differ- entiation. Altitudinally, DarUngtonia haetiana ranges from 1000 feet (303 meters) at Camp Perrin to a know^i maximum of 5000 feet (1515 meters) at Roche Croix. Specimens from the Furcy area, from localities which can be mapped, show an altitudinal range of from 5000 feet (1515 meters) at Kenscoff and Peneau to 5600 feet (1697 meters) at Furcy. The narrow altitudinal range is doubtless an artifact of collection, since the wider altitudinal limits to the west indicate that the snake may occur much lower. Once again, the extreme deforestation of the accessible mountains near Port-au-Prince may have been a crucial factor in limiting this snake to more favored higher elevations. The elevations for D. h. perfector (3000 and 3700 feet = 910 and 1221 meters) are intermediate and within the known limits of the jNIassif de la Hotte distribution of DarUngtonia. 8 BREVIORA No. 229 The series of five specimens from Peneau was collected in a montane ravine with dense growth of bamboo; the specimen from Kenscoff was secured under a pile of drying vetiver. The Camp Perrin snake was received from a native and there are no habitat data available for it. Of the two recent specimens of D. h. per- fector, the type was taken under a rock in a small weedy clearing surrounded by broad-leaf forest and cafetales, and the paratype was taken within a mat of cut vines in the same clearing. The absence of records of Darlingtonia from the area between Furcy and the eastern end of the Sierra de Baoruco is puzzling. If Darlingtonia has the altitudinal limits indicated by the La Hotte specimens, it is strange that it has not been taken at such localities as Foret des Pins in e.xtreme eastern Haiti or along the Dominico-Haitian border. It may be significant, however, that the junior author and Mr. Leber visited the type locality of D. h. perfedor in the summer of 1963, and although they secured a long series of Wetmorena, no Darlingtonia were taken. The two specimens were secured at the same locality a year later. One would expect that Darlingtonia occurs throughout much of the south island montane massifs, but, on the other hand, it is possible that the populations are in reality disjunct. Only further collect- ing may pro^"e which is the case. Hemipenis The everted hemipenis of Darlingtonia haetiana is relatively small (extending to the level of the 6th or 8th sul)caudal), and bilobed; the sulcus spermaticus is deeply forked: and the sulcate side (we depart from the usage of Dowling and Savage, 1960, and use sulcate and non-sulcate instead of medial and lateral for the surfaces of the everted organ) is strongly differentiated from the non-sulcate side in being covered with papillae from the region of the sulcus spermaticus to the apices. The sulcus spermaticus proceeds through the papiilate region to the apices. This papillate zone is sharply set off from the non-sulcate side which is uni- formly adorned with very small spines. The or-gan has the ap- pearance of having an elongate cordate shield (papillate zone) affixed to the sulcate side (Fig. 4). A row of enlarged but pro- gressively smaller spines begins on each edge of the non-sulcate side about midway the length of the organ (!e\-el of tiiird sub- caudal) and proceeds basally and diagonally, becoming lost in the profusion of smaller spines on the basal sulcate side. Papillate calyces cover the papillate sides of each lobe; the papillae become 1965 SNAKE GENUS DARLIxNGTONIA Fig. 4. Hemipenis of Darlingtonia haetiana; left, sulcate surface; right, non-sulcate surface; ]\ICZ 65100, near Palmiste, Furcy, Dept. de I'Ouest, Haiti. smaller and more sparse on the more basal portions of the hemi- penis. Several larger, isolated papillae are present on the non- sulcate surfaces of the lobes or in the crotch between them. One specimen, ASFS X2078, appears abnormal in having more sparse and less prominent calyces, and accessory sulci partially di\dding the papillate zone, making the organ almost capitate. Specimens Examined Darlingtonia haetiana haetiana: Haiti, Dept. du SucI, Roche Croix, northeastern foothills. Massif de la Hotte (= Pic Macaya), about 5000 feet (1515 meters) altitude, MCZ 38252 (type): Camp Perrin, 1000 feet (303 meters), ASFS X3058; Dept. de VOuest, Peneau, Bassin Bleu, 5000 feet (1515 meters), ASFS X2077-81 ; Furcy, AICZ 60060-61, MCZ 66996-98, UMMZ 123097 (4 speci- mens), USNM 123803; Vendome near Furcy, MCZ 65104-07; near Palmiste, Furcy, MCZ 65098-101 ; Morne Bourette, USNM 117286; Kenscoff, 5000 feet (1515 meters), ASFS X2254. Darlingtonia haetiana perfector: Republica Dominicana, Bara- hona Prov., 24 km SW Barahona, 3700 feet (1221 meters), INICZ 77217, ASFS V2898 (type and paratype); AMNH 49738, near Polo, 3000 feet (910 meters) (paratype). 10 BREVIORA No. 229 LITERATURE CITED Cochran, Doris M. 1935. New reptiles and amphibians collected in Haiti by P. J. Darlington. Proc. Boston Soc. Nat. Hist., Vol. 40, no. 6, pp. 367-376. DowLiNG, Herndon G. 1951. A proposed standard system of counting ventrals in snakes. Brit. Jour. Herpetology, Vol. 1, no. 5, pp. 97-99, 1 fig. DowLiNG, Herndon G., and Jay M. Savage 1960. A guide to the snake hemipenis: a survey of basic structure and systematic characteristics. Zoologica, Vol. 45, pt. 1, pp. 17-28. (Received 2 June, 1965.) BREVIORA MiiseiiitM of Comparsitive Zoology Cambuiuge, Mass. Skptkmhek 15, 19(55 Number 2:^0 NOTES ON SOME NON-PASSERTNE BIRDS FROM EASTERN ECUADOR By David W. Norton A new collection of birds from eastern Ecuador contains a number of significant specimens. The Ecuadorian segment of the upper Amazon basin has been generally neglected by orni- thologists since Chapman's study of 1926, while the adjacent areas of Colombia and Peru have been more recently investi- gated. The latter studies have raised questions about the continuity of populations along the eastern base of the Andes and this new collection answers some of these questions by documenting the presence of certain forms in eastern Ecuador. In 1963 and 1964, during the months June-September, I made collecting trips to eastern Ecuador, working at altitudes be- tween 300 and 1,500 meters. In 1963, I spent a month and a half at Limoncocha, and a month on Mount Sumaco. Accom- panied by Richard D. Chandler in 1964, I returned to collect on the slopes of Sumaco for two months. A collection of 1,900 specimens from these two expeditions is deposited in the Mu- seum of Comparative Zoology. The present paper contains the most noteworthy records from among the 145 forms of non- passerines represented in this collection. COLLECTING LOCALITIES CoTAPiNO (often labeled on maps as ''Concepcion") : 0°45'S, 77°25'W; alt. 700 m^ 25 June-12 July 1964; 250 specimens. At the junction of Rio Pucuno and Rio Cotapino, this hacienda serves as a departure point for Sumaco trips. Besides an airstrip 1 Altitudes are based on readings taken with a pocket altimeter at each locality. 2 - BREVIORA No. 230 and some 60 acres of cultivated land, Cotapino offers shelter and a small number of Quechuan Indian laborers. Most of the speci- mens were collected here with mist nets. . . CuYUJA: 0°25'S. 78°08'W; alt. 2,400 m; 19 June 1904; 9 specimens. Chandler and I visited this town 10 km ea.st of Papallaeta, hoping to find conditions comparable to those on the slopes of Sumaco. Collecting here was incidental, but promising, because the deforestation characterizing Papallaeta has not yet reached Cuyuja. EuGENio: 0°46'S, 77°24'AY; alt. 700 m; 16-22 July 1964; 111 specimens. A day's foot travel northwest toward Sumaco from Cotapino, the locality takes its name from the lone Quechuan inhabitant of this last settlement on the way to Sumaco. LiMONCOCHA : P°25'S, 7d°38'W ; alt. 300 m ; 25 June-2 August 1963 ; 300 sj^ecimens. Limoncocha is the site of a large missionary base camp, in- cluding an airstrip, housing for United States missionary fami- lies, and a Quechuan village, all of which has been established since 1955 on the shores of a lake in the midst of tropical forest. The large lake whose lemonade-colored water inspired the name, Limoncocha, lies two kilometers inland from the mouth of Rio Jivino on the Rio Napo. There are few such lakes in eastern Ecuador, and none as unspoiled as Limoncocha. Consequently, the lake and its outlet have yielded some unique specimens, sev- eral forms new to Ecuador, and at least one new race. Lower Rio Pucuno: 0°46'S, 77°28'W; alt. 500 m; 19-31 Au- gust 1963 ; 135 specimens. Halfway between Cotapino and Eugenio, a traveller must ford the Rio Pucuno afoot. In 1963, high water made the river im- passable for a few days, during which delay 1 collected along the banks above and below the ford. MoNTALVO: 02°05'S, 76°57'W; alt. 250 m; 5 specimens. I purchased several specimens of apparent interest from this locality from R. Olalla, who worked here as an independent collector in 1964. Rio Negro: 01°25'S, 78°03'W; alt. 1,200 m; 12-14 September 1964; 35 specimens. Along 'the road near this village, a few specimens were col- lected for comparison with those from similar altitiules on Sunii^po. 1965 BIRDS FROAl EASTERN ECUADOR 3 Mount Sumaco : Mount Siimaco, an isolated massif rising to about 4,000 ni, lies about 40 km southeast of Baeza, and about 100 km southeast of Quito. Pour collecting cami)s were estab- lished on the southeast slojies. I have arbitrarily prefixed the name Sumaeo to all localities above Eugenio, the highest point settled by Quechuan Indians. Sumaco, Guaticocha: 0°45'S, 77°24'W; alt. 750 m; 16-24 August 1964; 123 specimens. Only a few hours due west of Eugenio lies this tiny, perfectly round, and very deep lake, where Chandler and I established a camp. Sumaco, head of Kio Guataraco : 0°40'S, 77°35'W; alt. 1,350 m ; 24 July-5 August 1964 ; 350 specimens. The junction of several brooks in a series of waterfalls a few hundred meters below this camp, marks the beginning of the Rio Guataraco. Sumaco, Palm Peak (translated from local Quechuan hunt- ers' designation, Ramus-Urcu) -. 0°39'S, 77°36'W; alt. 1,500 m; 6-14 August 1964 ; 209 specimens. Palm Peak is the rim of the altiplano which slopes gradually upward for two days' travel to the final steep slope of the moun- tain peak. Most of the collecting was done on the altiplano. Sumaco, Upper Rio Pucuno: 0°36'S, 77°35'W; alt. 1,200 m; 11-16 August 1963; 135 specimens. The highest camp in 1963 was on the narrow ridge separating the two principal rivers of this face of Sumaco, the Pucuno and the Guataraco. SYSTEMATIC NOTES Nycticorax pileatus (Boddaert) Limoncocha, 1 £ . This wide-ranging form has not been recorded before from Ecuador, where its occurrence was to be expected. This indi- vidual was shot in the cow pasture at Limoncocha, a habitat somewhat unusual in eastern Ecuador, as cattle in the country are confined mostly to the highlands. Ixobrychus exilis limoncochae subsp. nov. Type: Adult male, No. 285,860, Museum of Comparative Zoology, collected at Limoncocha, alt. 300 m, Rio Napo, eastern Ecuador, by D. W. Norton, on 6 July 1963. BREVIORA No. 230 Diagnosis: Similar to nominate exilis, but cheeks and auricu- lars shading to chestnut, ventrum less streaked. Similar to erythromelas in color of cheeks, but wing and bill somewhat longer, ventrum less richly colored, throat less streaked ; dorsum of female grayish brown, not reddish brown as in erythromelas. Distinguished from hogotensis by chestnut rather than ochra- ceous cheeks, by lighter ventrum, with less streaking on throat, and by a somewhat shorter bill. Generally smaller than pe- ruvianus, with ventrum, head, and wing coverts tawny rather than ochraceous. Range: Known only from the type locality. Measurements Wing Culmen (exposed) limoncochae (1$ - -type) 120 mm 45 mm (19) 120 46 exilis (45) 112-121 (117.8 ± 1.7) 44-49 (46.3 ± .09) (1$) 112 43 erythromelas (35) 106-113 (109.3 ± 1.7) 42-44 (42.7 rt 0.5) (2$) 109-113 (111.0 ± 1.4) 40.0 hogotensis (2 5) 116-122 (119.0 ± 2.1) 40.0 (5$) 113-125 (120.8 ± 2.1) 37-40 (39.0 ± 0.6) peruvianus (15) 125 53 Remarks: The two specimens of limoncochae were collected with a single shot and, therefore, are presumed to have been mated, although their gonads were not enlarged. While examining comparative material, I found that a female (AMNH No. 151,639) from Antioquia, Colombia, is referable to hogotensis, which extends the known range of that race con- siderably to the northwest. I tentatively identified as peruvianus a mounted specimen in adult male plumage at the Colegio San Bolivar in Ambato, Ecuador. The wing measures 127 mm and the culmen 50 mm. Although there were no data, the curator of the collection, who is the widow of the collector, claimed that her husband shot all his specimens in Ecuador. Corroboration of this doubtful record would extend the known range of peruvianus north from coastal Peru (Dept. Libertad) to western Ecuador. The origins of the races in northern South America are deserv- ing of speculation. Significantly, I believe, the southern limits of migration of the two North American races are just to the north (eastern Panama). Ixohrychus exilis may originally have been a temperate and entirely migratory species. In this case. 1965 BIRDS FROM EASTERN ECUADOR 5 the South American races have arisen at the southern limits of the wintering populations, and may represent descendants of a few birds which gave up the migratory habit to breed near the equator. The South American forms must then be more recent developments than the splitting of the migratory population into eastern and western subspecies, because the four southern forms are neatly divisible morphologically into eastern and west- ern types. The easternmost form, erythromelas, seemingly shows the extreme in characteristics of nominate exilis of eastern United States, in its small size and generally reddish brown coloring. The westernmost, peruvianus, is similarly related to hesperis of the western United States, as it shows the extreme of large size and grayish brown coloring. In between, limon- cochac is a less extreme eastern type, and hogotensis a less extreme western type. Specimens examined: limoncochae : Ecuador, Limoncocha, 1 (? , 1 9 ; erythromelas : Surinam, Paramaibo, Z $ , 1 9 ; hogoten- sis: Colombia, Savana de Bogota, 2$, 49 ; Antioquia, l9 ; peruvianus: Peru, Dept. Lima, Vegueta, 1^,1 imm. ; "Ecua- dor", 1[5 ]. Aramides calopterus Sclater and Salvin Sumaco, Guaticocha, 1 i ; Montalvo, 1 $ . A specimen from Montalvo, collected by R. Olalla in April 1 964, is significantly lighter than any in a series from Rio Suno and Sumaco. However, I believe Olalla dries skins in direct sun- light, which Avould cause fading, and explain this disparity. PoRPHYRULA MARTiNiCA (Liunaeus) Limoncocha, 1 6 , 1 9 . Although it is recorded from western Ecuador, I find no pre- vious record of this wide-ranging species in eastern Ecuador, where its occurrence was to be expected. PoRPHYRULA FLAViROSTRis (Boddaert) Limoncocha, 1 $ . The species, new to Ecuador, is apparently uncommon in up- per Amazonia, for de Schauensee (1949: 432) includes it in the faunal list of Colombia, also on the basis of a single specimen (Florencia). Individuals occur regularly at Limoncocha, but I have not seen any elsewhere in eastern Ecuador. 6 BREVIORA No. 230 Jacana spinosa intermedia (Sclater) Limoncocha, 2 5 . Specimens from Limoncocha are clearly referable to inter- media. This first record of the species from eastern Ecuador extends the known range of intermedia south from eastern Co- lombia to at least the Rio Napo in Ecuador. Hellmayr and Conover (1948 :9) questioned the validity of peruviana, of north- eastern Peru but comparison of 14 specimens from northeastern Peru with a series of intermedia from Venezuela, Colombia, and Ecuador, shows Peruvian specimens to be much the largest and darkest in the group, and easily separable from the reasonably uniform intermedia specimens. Vanellus resplendens (Tschudi) Limoncocha, 2 i ; Mt. Cotopaxi, Laguna de Limpio, 2 9 . The Andean lapwing seems never to have been recorded from below 2,000 m anywhere in its range (Ecuador to northern Chile). The Limoncocha specimens were observed daily (per- sonal communication) on the grassy airfield of the mission sta- tion (alt. 300 m) from February to June 1963, before I collected them there in July. This unusual pair, frequenting an artificial habitat cleared only recently of tropical forest, proved to be two males coming into adult plumage and having small gonads. These birds were much tamer than individuals observed and collected on Mt. Cotopaxi. The unique record probably repre- sents young strays lost during the seasonal altitudinal migrations of this species in the Andes. Aside from the airstrip, the only other sizable unforested areas near Limoncocha are the gravel bars of the nearby Rio Napo. This pair was indeed traced twice to the gravel bars during the daylight hours. Significantly, the Quechuan Indians of Limoncoclia could give no local name for this bird, whereas all regularly occurring species on the Napo receive specific and descriptive Quechuan names. Genus Eutoxeres Eiitoxeres condamini condamini (Bourcicr) : Cotapino, 3,5, 59. Eutoxeres aquila aquila (Boureier) : Eugenio, 2 c^ ; Sumaco, head of Rio Guataraeo, 1 S ; Sumaco, Guaticocha, 1 9 . The ranges of these sibling si)eeies overlap in eastern Ecuador and adjacent areas of Colombia and Peru. Since Chandler and I 1965 BIRDS FROM EASTERN ECUADOR Fig. 1. Distribution of Euioxeres aquila and Eutoxcres condamini. never found both species in the same locality, it is i)ossible that there is a difference in the habitat preferences of the two species. It seems that E. condamini occurs in open cultivated areas, at low elevations, whereas E. aquila occurs in dense forest at any altitude up to at least 1,500 m. 8 BREVIORA No. 230 The distribution of the forms of Eutoxeres may be related to the distribution of a principal food source. Greenewalt (1960: legend, PL 34) observed E. aquila feeding on "platanillos, " or plantains of the genus Hcliconia, using its highly specialized, downcurved bill to draw nectar from the deep, upright bracts of the flowers. Chandler and I had best results netting both species of Eutoxeres when the nets were placed close to Hcliconia plants, which further suggests the dependence of Eutoxeres upon the plantains. The genus Eutoxeres occurs in northwestern South America, roughly where the greatest concentrations of the 35-odd species of Heliconia also occur. The flowering periods of the various species of Heliconia doubtless span the year, prob- ably providing a steady supph^ of food for the hummingbirds. Observations made in eastern Ecuador suggest that the critical factor in determining local distribution of the sibling species is the abundance of Heliconia plants. When land in eastern Ecua- dor is cleared, and particularly when bananas are planted, many "platanillos" invade the clearing. Apparentlj^ E. c. condamini establishes itself in the midst of this abundance, while E. a. aquila retreats to the forest. Although both forms have been recorded sympatrically at La Morelia, Colombia (de Schauensee, 1949 : 541), and at the mouth of Rio Curaray (Zimmer, 1950 : 1), I suspect that condamini is found in the settlements while aquila is found farther afield. These records reflect either inexactness in recording the locality, or possibly the passage of time be- tween collectors' visits, during which a locality supporting aquila was cultivated, thereby attracting coiulamini . The origin of these sibling species merits some speculations. The simplest explanation of the present distribution of the spe- cies of Eutoxeres is that an original South American population was split by the Andean uplift, giving rise to condamini to the east and aquila to the Avest of the cordillera. Subsequently, aquila has colonized the isthmus of Panama, and has spilled east- ward over the Andes to invade the range of condamini. Genus Topaza Topaza pyra (Gould) : Cotapino, 1 S ; Sumaco, Guaticocha, 19. These specimens are the first definitely to extend the known range of pyra so far up the Rio Napo, contradicting Oberholser's ( 1902 : 322) assertion that the species does not occur above Coca on the upper Napo. Of greater interest is the proof that T. pyra 1965 BIRDS FROM EASTERN ECUADOR 9 occurs west of T. pella pamprepta, an endemic form known only from Boca Suno on the Napo in eastern Ecuador. The popula- tion of T. p. pamprepta seems to be surrounded by the morjilio- logically similar T. pijra, isolated in Boca Suno, 2,000 km from the other races of pella, which occur in the Guianas and in Para, Brazil. The validity of pamprepta, and the present range exten- sion of pyra make it impossible to accept Peters' (1945: 92) suggestion that pella and pyra might be conspecific. I have examined the type of pamprepta, finding Oberholser's (1902: 322) description and the data given by the collectors, both accurate. The type is similar in pattern to nominate pella, but it is distinguished by a much longer tail and somewhat shorter bill. One can conclude that the curiously spotty distri- butions of the taxa of Topaza are artifacts due to the rarity of these equatorial hummingbirds. Future records of pella may come from northern Brazil, between Ecuador and the Guianas. Trogon rufus sulphureus Spix Sumaco, Guaticocha, 1 $ ; Sumaco, Guaticocha, 1 9 (preserved as a skeleton). Although this subspecies was to be expected in eastern Ecua- dor, Peters (1945: 157) omits both eastern Ecuador and eastern Colombia from its range. Apparently, the only previous records were specimens labeled with doubtful accuracy as being from Coca, Rio Napo, and "Equateur" (Zimmer, 1948: 29). The present record shows sulphureus to occur almost to the foot of the Andes and up to nearly 800 m in this part of its range. CaPITO NIGER PUNCTATUS (LcSSOn) Limoncocha, 3 S , 4 $ ; Lower Rio Pucuno, 1 S ; Cotapino, 1$, 1$. The subspecific status of this species in upper Amazonia is in confusion. Several authors have disputed Brodkorb's (1939) two races, macintyrei of eastern Ecuador and conjunctus of northeastern Peru. Examining a large series from Colombia, Ecuador, and Peru, I fail to find any of the consistent geo- graphic variations mentioned by Brodkorb. Furthermore, the 12 birds from Limoncocha show all extremes in pileum color and streaking of the ventrum. Bond (1954: 49) synonymized conjunctus with macintyrei, but distinguished macintyrei from pnnctatus on the basis of the amount of ventral streaking in one female from Villavicencio, Colombia. Two of four Limoncocha 10 BREVIORA No. 230 females have the ventrum as lightly streaked as this female from Villavicencio. I therefore agree with Peters (1948: 25) and ascribe to punctatus birds of upper Amazonia from Colombia to northeastern Peru. PicuMNUs RUFR'^NTRis RUFiVENTRis (Bonaparte) Limoncocha, 1 £ ; Cotapino, 1 S ; Eugenio, 1 S ; Sumaco, Palm Peak, 1$ , 19. Chandler and I found these birds quite common up to at least 1,500 m on Mt. Sumaco, although the species is usually consid- ered an inhabitant of the Tropical Zone (de Schauensee, 1949: 641). At Palm Peak, specimens were taken simultaneously with such typically subtropical forms as Cyanocorax yncas. ACKNOWLEDGMENTS I thank Dean Amadou of the American Museum of Natural History, James Bond of the Philadelphia Academy of Natural Sciences, and George E. Watson of the U.S. National Museum, all of whom lent comparative material for this study. Richard D. Chandler was a valuable companion and collector on the trip in 1964. I am especially grateful to Raymond A. Paynter, Jr. for his guidance during the preparation of this paper. LITERATUEE CITED Bond, James 1954. Notes on Peruvian Pieiformes. Proc. Acad. Nat. Sci. Phila- delphia, 106: 45-61. Bbodkorb, Pierce 1939. Two undeseribed South American barbets. Proc. Biol. Soc. Washington, 52: 135-136. Chapman, F. M. 1926. Distribution of bird-life in Ecuador. Bull. Am. Mus. Nat. Hist., 55: xiii -f 784 pp. DE Schauensee, R. M. 1949. The birds of the Republic of Colombia. Caldasia, 5 (23j: 381- 644. Grebinewalt, H. C. 1960. Hummingbirds. New York, Doubleday. 250 pp. Hellmayr, C. E. and B. Conover 1948. Catalogue of l)irds of the Americas. Field Mus. Nat. Hist., Zool. Ser., 13: ]it. 1, no. 3, vi -f 383 pp. 1965 BIRDS KHOM EASTERN ECUADOR 11 Oberholsek, H. C. 1902. Catalogue of a collection of huniniinsliinls from Ecuador and Colombia. Proc. V. S. Nat. Mus., 24: .309-342. Peters, J. L. 1945. Check-li.st of birds of the world. Vol. 5. Cambridge, Mass., Harvard Univ. Press, xi + 2.38 pp. 1948. Check-list of birds of the Avorld. Vol. (5. Cambridge, Mass., Harvard Univ. Press, xi + 306 pp. ZiMMER, J. T . 1948. The family Trogonidae. Am. Mus. Novit., No. 1380, 56 pp. 1950. The genera Eutoxeres, Campylopterus, Eupetomcna, and Flnri- .suga. Am. Mus. Novit., No. 1450, 14 pp. (Received 15 June 1965.) Harvard MCZ Library 3 2044 066 302 738 w^jgt^wif* iLi,^^-r» >«"rr4/**ill ATC