HARVARD UNIVERSITY Library of the Museum of Comparative Zoology BREVIORA MUSEUM OF COMPARATIVE ZOOLOGY AT HARVABD COLLEGE, IN CAMBRIDGE Numbers 67-120 1957-1960 CAMBRIDGE, MASS., U.S.A. 1960 Edited By Nelda E. Wright V.^ .\ CONTENTS BREVIORA ^Museum of Comparative Zoology Numbers 67-120 1957 No. 67. Notes on Certain Species of Tetragnatha (Araneae, Ar- giopidae) in Central America and Mexico. By Arthur ]\[. Chickering. 4 pp. January 31. No. 68. The Genus Tetragnatha (Araneae, Argiopidae) in Ja- maica, B.W.I., and other Neighboring Islands. By Arthur M. Chickering. 15 pp. January 31. No. 69. A New Zodariid Spider from Panama. By Arthur M. Chickering. 7 \)\). January 31. No. 70. "Anguimorph" Tooth Replacement in Amphishaena alha Linnaeus, 1758, and A. fuliginosa Linnaeus, 1758 (Reptilia : Ampliisbaenidae). By Carl Gans. 12 pp. January 31. No. 71. Taxonomic Notes on the New World Forms of Troglo- dytes. By Raymond A. Paynter. Jr. 15 pp. March 29. No. 72. Is the Ant Genus Tetramorium Native in North Amer- ica? By W. L. Brown, Jr. 8 pp. March 29. No. 73. Additions to the Mammalian Fauna of Peru and Notes on Some Other Peruvian Mammals. By Oliver P. Pearson. 7 pp. March 29. No. 74. The Discovery of Cerapachyine Ants on New Caledonia, with the Description of New Species of Phyracace.s and Sphinctomyrmex. By E. 0. Wilson. 9 pp. May 1. No. 75. Oil a New Oetochaetine Earthworm Supposedly from Guatemala. By G. E. Gates. 8 pp. May 1. No. 7(). Two New Land and Freshwater Mollusks from New Guinea. By William J. Clench. 4 pp. June 18. No. 77. Dacetinopa, A New Ant Genus from New Guinea. By W. L. Brown, Jr. and E. 0. Wilson. 7 pp. June 21. No. 78. The Larva of the Ant Genus Dacetinops Brown and Wilson. By George C. Wheeler and Jeanette Wheel- er. 4 pp. Jvme 21. No. 79. Dasypeltis nicdici lamiiensis, A New Race of Egg-Eating Snake (Ophidia, Reptilia) from Coastal East Africa. Bj'- Carl Gans. 13 pp. August 9. No. 80. A Collection of Drawings of Fishes Ascribed to J. P. Kirtland (1793-1877), in the Library of Bowdoin College. By James M. Moulton. 4 pp., 2 pis. Sep-^ tember 30. No. 81. Contributions to a Revision of the Earthworm Family Lumbricidae. I. Allolobophora limicola. By G. E. Gates. 14 pp. September 30. 1958 No. 82. The Trunk ^lusculature of Sanzina and its Beai-ing on Certain Aspects of the ]\Iyological Evolution of Snakes. By Walter Auffenberg. 12 pp. January 31. Xo. 83. Thamnophis hovaUii Dunn Rediscovered (Reptilia, Ser- })entes). By Benjamin Shreve and Carl Gans. 8 pp. January 29. No. 84. Rediscovery of the Australian Chelid Genus Pseudemy- dura Siebeurock (Chelidae, Testudines). By Ernest E. Williams. 8 pp., 4 pis. January 30. No. 85. The Choanal Papillae of the Cheloniidae. By Thomas S. Parsons. 5 pp., 2 pis. January 31. No. 86. A New Sicistine Rodent from the ^Miocene of Wyoming. By Craig C. Black. 7 pp. May 29. No. 87. An Enibolonioro J aw from the JMid-Carboniferoiis of Nova Scotia. By Alfred Sherwood Romer. 7 pp., 1 pi. June 20. No. 88. A New Species of the Genus iirotheca (Serpentes: Colu- bridae) from Venezuela. B}' J. A. Roze. 5 pp. June 30. No. 89. Remarks on Some Forms of Cinclus (Aves). By James C. Greenway. Jr. and Charles Vaurie. 10 pp. July 15. No. 90. A Fossil X'ampire Bat from Cuba. By Karl F. Koop- man. 4 pp., 1 pi. July 30. No. 91. Contribution to a Revision of the Earthworm Family Lumbricidae. II. Indian Species. By G. E. Gates. 16 pp. August 13. No. 92. A New Genus of Erethizontid Rodents from the Col- huehuapian of Patagonia. By Bryan Patterson. 4 pp. September 17. No. 93. A New Barylambdid Pantodont from the Late Paleo- cene. Bj^ Bryan Patterson and Elwyn L. Simons. 8 pp. September 18. No. 94. Affinities of the Patagonian Fossil Mammal Necrolestes. By Bryan Patterson. 14 pp. September 18. No. 95. A New Bolivian Land Snail of the Genus Dryniaeus. By Juan Jose Parodiz. 3 pp. September 19. No. 9(j. A New Dichobunid Artiodactyl from the Uinta Eocene. By C. Lewis Gazin. (i pp. September 19. No. 97. Fusion of Cervical Vertebrae in the Erethizontidae and Dinomyidae. By Clayton E. Ray. 11 pp., 2 pis. October 27. No. 98. Two New Species of Bathylagus from the Western North Atlantic with Notes on Other Species. By Daniel M. Cohen. 9 pp. December 12. No. 99. A New Subspecies of Chamadeo jacksoni Boulenger and a Key to the Species of Three-llorned Chamaeleons. By A. Stanley Rand. 8 pp. December 19. No. TOO. On the Pineal Organ of the Tuna, Thijnnus thynnus L. By Uno Hohngren. 5 pp., 2 pis. December 23. 1959 No. 101. Cervical Ribs in Turtles. By Ernest E. Williams. 12 pp., 1 pi. March 2. No. 102. A New Jamaican Galliwasp (Sauria, Anguidae). By Garth Underwood. 13 pp. April 9. No. 108. Two New Species of Eleutheroductyius from Puerto Rico. By Juan A. Rivero. 6 pp., 1 pi. April 10. No. 104. Studies on Fishes of the Family Ophidiidae. III. A New Species of Lepophidium from Barbados. By C. Richard Robins. 7 pp. April 13. No. 105. Bufo (jmidlachi, A New Species of Cuban Toad. By Rodolfo Ruibal. 14 pp. April 14. No. 106. The Occipito-Vertebral Joint in the Burrowing Snakes of the Family Uropeltidae. By Ernest E. Williams. 10 pp. April 28. No. 107. A Revision of the Dacetine Ant Genus Neostruma. By William L. Brown, Jr. 13 pp. May 6. No. 108. Some New Species of Dacetine Ants. By William L. Brown, Jr. 11 pp. May 7. No. 109. On the Pineal Area and Adjacent Structures of the Brain of the Dipnoan Fish, Protopterus annectens (Owen). By Uno Holmgren. 7 pp., 2 pis. May 8. No. 110. The Spider Genus Coleosoma (Araneae, Theridiidae). By Herbert W. Levi. 8 pp.. 1 pi. June 16. No. 111. On the Caudal Neurosecretory System of the Teleost Fish, Fundulus heterodiUis L. By Uno Holmgren. 13 pp., 2 pis. June 17. No. 112. A Mounted Skeleton of the Giant Plesiosaur Kronosau- rus. By Alfred Sherwood Romer and Arnold D. Lewis. 14 i^p.. 1 pi. October 15. No. 113. A New Phyllomedusa from Bolivia (Salientia, Hyli- dae). By Benjamin Shreve. I^ pp., 1 })1. November 2. No. 114. Anomalophis bolccnsis (Massalonj^o), A New Genus of Fossil Snake from the Italian Eocene. By Walter Auffenberji'. 16 pp. November 23. No. 115. The Lemon-Colored Plexaurids from the West Indies and Brazil. By Elisabeth Deiehmann and P. M. Bayer. 12 pp., 5 pis. Noveml)er 25. 1960 No. 116. Insectivores of the Middle Miocene Split Rock Local Fanna, Wyoming-. By Katherine Milmine Reed. 11 pp., 2 pis. January 6. Xo. 117. Notes on Hispaniolan Herpetology. 1. Anolis christo- phci. New Species, from the Citadel of King Christo- phe, Haiti. By Ernest E. Williams. 7 pp. January 20. No. 118. A Survey of the Leptodactylid Frogs, Genus Eupso- phus, in Chile. By Jose M. Cei. 13 pp. February 24. No. 119. Arctic Archibenthal and Abyssal Mollusks from Drift- ing Station Alpha. By Arthur H. Clarke, Jr. 17 pp., 1 pi. March 8. No. 120. Two Species of Tortoises in Northern South America. By Ernest E. Williams. 13 pp., 3 pis. March 9. INDEX OF AUTHORS BREVIORA MUSEUM OP COMPARATIVE ZOOLOGY Numbers 67-120 1957-60 No. AUFFENBERG WALTER 82, 114 Bayer, F. M. and Elisabeth Deichmann 115 Black, Craig C 86 Brown, William L., Jk 72, 107, 108 Brown, William L., Jr. and E. 0. Wilson 77 Cei, Jose M 118 Chickering, Arthur M 67, 68, 69 Clarke, Arthur H., Jr 119 Clench, William J 76 Cohen, Daniel M 98 Deichmann, Elisabeth and F. M. Bayer 115 Gans, Carl 70, 79 Gans, Carl and Benjamin Shreve 83 Gates, G. E 75, 81, 91 Gazin, C. Lewis 96 Greenway, James C, Jr. and Charles Vaurie 89 Holmgren, Uno 100, 109, 111 KooPMAN, Karl F 90 Levi, Herbert W 110 Lewis, Arnold D. and Alfred Sherwood Romp:r 112 MouLTON, James M. 80 Parodiz, Juan Jose 95 Parsons, Thomas S 85 Patterson, Bryan 92, 94 Patterson, Bryan and Elwyn L. Simons 93 Paynter, Raymond A.. Jr 71 Pearson, Oliver P 73 Rand, A. Stanley 99 Ray, Clayton E 97 Reed, Katherine Milmine 116 RivERO, Juan A. 103 Robins, C. Richard 104 RoMER, Alfred Sherwood 87 RoMER, Alfred Sherwood and Arnold D. Lewis 112 Roze, J. a 88 RuiBAL, Rodolfo 105 Shreve, Benjamin 113 Shreve, Benjamin and Carl Cans 83 Simons, Elwyn L. and Bryan Patterson 93 Underwood, Garth 102 Vauree, Charles and James C. Green way, Jr 89 Wheeler, George C. and Jeanette Wheeler 78 Wheeler, Jeanette and George C. Wheeler 78 Williams, Ernest E. 84, 101, 106, 117, 120 Wilson, E. 0 74 Wilson, E. 0. and William L. Brown, Jr 77 BREVIORA Moseiiim of Comparative Zoology Cambridge, Mass. Januiiry 31, 1957 Number 67 NOTES ON CERTAIN SPECIES OP TETRAGNATHA (ARANEAE, ARGIOPIDAE) IN CENTRAL AMERICA AND MEXICO By Arthur M. Chickering Albion CollpRp, Albion. Michigan In connection with my study of the ^enns Tefrngnafha Latreille, 1804 in Panama and the West Indies I have also had occasion to examine a number of species from parts of Central America north of Panama, and also from Mexico, types of which are in the Museum of Comparative Zoology at Harvard College. The following notes are offered as a contribution to the further clarification of the genus. Tetragnatha versicolor Walckenaer, 1841 T. convexa Banks, 1898 T. convexa Petrunkevitch, 1911 T. convexa Eoewer, 1942 A vial labelled T. convexa Banks and marked "cotypes" now contains one male and three females all from San Jose del Cabo, Baja California. The chelicerae, palp, and other characters make it certain that the male belongs to T. versicolor Walckenaer. The females are always more difficult to place with accuracy but I feel certain that these belong in the same species with the male. There is another male in the collection from Sierra Laguna, Baja California, originally identified as a T. convexa Banks l)ut this- is also clearly a T. versicolor Walckenaer. Tetragnatha guatemalensis 0. P. Cambridge, 1889 T. fraterna Banks, 1898 T. mnndihulata Banks, 1898 BREVIORA NO. 67 T. fraterna Petrunkevitch, 1911 T. fraterna Eoewer, 1942 The characteristics of the male palp and the male chelicerae definiteh" identify the males collected at San Jose del Cabo, Baja California and described as T. fraterna, and I feel confident that we may be certain of the correct placement of the females as well. Banks identified both sexes collected at Tepic and San Jose del Cabo as specimens of T. mnndihiilata Walckenaer, 1841, known at present only from Ceylon, India, Australia, and Poly- nesia. Re-examination of these shows clearly that they belong to T. guatemalensis 0. P. Cambridge. Tetragnatha tristani Banks, 1909 (Figures 1-5) There is only a single specimen to represent this species so far as I have been able to determine. This is the holotype very briefly described by its author. I have carefully searched through my collections of Tetragnatha from the regions of Panama closely contiguous to Costa Rica, from which country the original was collected, without discovering any additional specimens. I have also compared the holotype with all other species known to me, with the result that I am compelled to regard it as a valid species. • ^ 'y / e D kA i * ' *■« External Anatomy of Tetragnatha tristani Banks Figures 1, 2. Cheliceral teeth along the fang groove; promarginal nnd retromarginal teeth, respectively. Figures 3, 4. Distal ends of cymbiiun, embolus, and eonduetor from two different views. Figure .t. Paracymbium. 1957 TETRAGNATHA FROM CENTRAL AMERICA 6 Since the orig:inal description was so brief I have thoup:ht it desirable to furnish a detailed treatment in accord with my usual procedure. Male holofypc. Total leng-th includinor the chelicerae 4.55 mm. ; exclusive of the chelicerae 4.16 mm. Carapace 1.495 mm. long, .97 mm. wide opposite second coxae where it is widest ; other features as usual in the genus. Eyes. Eight in two rows as usual in the genus; viewed from above, both rows moderately recurved ; viewed from in front, both rows slightly procurved, both measured by centers. Central ocular quadrangle wider behind than in front in ratio of about 14 : 11, slightly wider behind than long. Ratio of eyes AME : ALE : PME : PLE = 7 : 5.5 : 8 : 6.5. AME separated from one another by 1.5 times their diameter, from ALE by a little more than twice their diameter. PME separated from one another by 1.7 times their diameter, from PLE by slightly more than this distance. Laterals separated from one another by the diameter of PLE; AME separated from PME by nearly 1.5 times as far. Height of clypeus equal to a little more than the diameter of AME. Chelicerae. Moderately well developed, porrect and divergent; basal segment .78 mm. long and, therefore, about half as long as the carapace ; the prolateral spur is simple, rather poorly devel- oped, and not distally bifid ; the fang is moderately sinuous and distinctly bent posteriorly in distal half; the promargin of the fang groove has seven teeth but the last three are minute den- ticles and would probably be subject to much variation in a large population; the retromargin also has seven teeth with the last two very small (Pigs. 1-2). Maxillae. Nearly parallel ; somewhat concave along outer border ; about three times as long as wide at narrowest level. Lip. Wider at base than long in ratio of about 15 : 11 ; less than one-half as long as maxillae ; sternal suture gently pro- curved ; sternal tubercles well developed and robust at ends of sternal suture. Sternum. As usual in the genus ; with fourth coxae separated by about one-half their width. Legs. 1243. Width of first patella at "knee" .162 mm., tibial index of first leg 4. Width of fourth patella at "knee" .152 mm., tibial index of fourth leg 7. BREVIORA NTO . 67 Femora Patellae Tibiae Metatarsi Tarsi Totals (All measurements in millimeters) 2.990 MO 3.250 .3.380 .975 11.245 2. 2.080 .552 1.755 1.885 .700 6.972 3. 1.105 .390 .617 .845 .325 3.282 4. 2.210 .390 1.625 1.950 .520 6.695 Palp .660 .154 .176 .970 1.960 All le^s with moderately coarse spines and the normal eoatinp: of hair. Palp. Both tibia and patella are short with the former only slightly longer than the latter; the paracymbium is very trans- parent but appears to have the shape shown in Figure 5; the cyrabium is long and slender; the conductor and embolus are also long, slender, and gently curved (Figs. 3-4). Abdomen. Not prolonged posterior to spinnerets; only slightly swollen in anterior half; slightly notched dorsally at base; 2.73 mm. long; .95 mm. wide about one-third of its length from liase where it is widest. Color in alcohol. Color apparently well preserved. Legs, mouth parts and cephalothorax with various shades of yellowish to light reddish brown. Abdomen with a well defined folium and many closely placed yellowish white silvery spangles on dorsum and along lateral sides. Venter plain yellowish. There is only one specimen known at present and that was col- lected by Prof. J. Fid. Tristan of San Jose, Costa Rica, in his home city with no date recorded. REFERENCES Banks, Nathan 1898. Arachnida from Baja California and other parts of Mexico. Proc. California Acad. Sei., Ser. 3, Zoology. 1, (7): 205-309, 5 pis. 1909. Arachnida from Costa Rica. I'roc. Acad. Nat. Sci. Philadelphia, April, 1909: 194-234, 2 pis. BREVIORA Mmseiuiirii of Compsirative Zoology Cambridge, Mass. January 31, 1957 Number 68 THE GENUS TETRAGNATHA (ARANEAE, ARGIOPIDAE) IN JAMAICA, B.W.I., AND OTHER NEIGHBORING ISLANDS Arthur M. Chickering Albion College, Albion, Michigan For several years before her death in 1953 Miss Elizabeth B. Bryant, Museum of Comparative Zoology at Harvard College, had been engaged in a comprehensive study of a collection of spiders from Jamaica, B. W. I. This collection had come from several sources but it had been assembled largely through the interest of Mr. C. Bernard Lewis, Director and Curator, Sci- ence Museum, Institute of Jamaica, Kingston, Jamaica. After Miss Bryant's death this collection was placed in my possession for continued study. On my way to Panama in June, 1954, I was able to stop in Jamaica for a reconnaissance of the island preparatory to what may be a more or less extensive study of the spiders of that country. As an extension of my study of the genus Tetragnatha Latreille, 1804 in Panama, I have been much interested in examining the genus in Jamaica and in comparing the species found there with the tetragnathids in several of the larger islands of that general region. This paper is a result of that study, and types of the new species named here are deposited in the Mu- seum of Comparative Zoology. It is again a pleasure to acknowledge my indebtedness to the following persons for their continued encouragement in the pur- suit of my studies : Dr. A. S. Romer and Dr. P. J. Darlington, Jr., Director, and Curator of Insects, respectively, in the Mu- seum of Comparative Zoology at Harvard College, and Miss Nelda E. Wright, Editor of Publications in the same institu- tion. Without the privileges which have been extended to me 2 BREVIORA NO. 68 for many years in this museum the continued progress of my studies would have been much more difficult. Genus TetraGNATHA Latreille, 1804 The genus has been well defined by Seeley (1928) and is, in general, well understood by araneologists. There are, however, certain characteristics of the genus which have in the past made it difficult to identify the species correctly and numerous errors must eventually be eliminated. It is also my opinion that suffi- cient attention has not usually been given to the question of varia- tion within species in respect to several of the most important structural features used by taxonomists for identification. F. P. Cambridge (1897-1905) emphasized the value of the character- istics of the male palp such as the form of the paracymbium, shape and course of the conductor and embolus as well as the features of the eyes, chelicerae, and legs. Petrunkevitch (1930) and Wiehle (1939) were the first to appreciate the value of the genital area, which lacks an epigynum, in identifying females which are often exceedingly difficult to place with certainty. Color has been shown to be extremely variable and nearly worth- less as a means of identification. Cheliceral teeth are often quite variable in number, degree of development, and relative posi- tion. Size, when mature, is also subject to great variation in sev- eral species. I have tried to take account of all of these salient features in making my determinations. THE GENTJS IN JAMAICA Only the bibliographical references considered essential are given in this paper. Extensive bibliographies may be found in several sources. Tetragnatha antillana Simon, 1897 T. antillana Petrunkevitch, 1930 T. antillana Bryant, 1940 T. antillana Bryant, 1942 T. festina Bryant, 1945 (male only) T. haitiensis Bryant, 1945 1956 TETRAGNATHA IN JAMAICA 3 This species appears to be common in Jamaica. It was found abundant at Mavis Bank over water by R. P. Bengry. Collection records: One male from the Blue Mts., southwest side of Main Range, between 3000-4000 ft. elevation, August, 1934 (P. J. Dar- lington, Jr.) ; both sexes from Mavis Bank, over water, March, 1953 (R. P. Bengry) ; one female from Rio Cobre, June, 1954, Tetragnatha caudata Emebton, 1884 Eucta caudata Petrunkeviteh, 1911 T. caudata Seeley, 1928 T. caudata Bryant, 1940 Miss Bryant had a single female from Cuba. The species ap- pears but once in the collection from Jamaica placed at my dis- posal ; Papine, five miles north of Kingston, April, 1937. Tetragnatha exigua sp. nov. (Figures 1-5) Male holotijpe. Total length including chelicerae 2.795 mm., without chelicerae 2.34 mm. Carapace 1.04 mm. long; .67 mm. wide opposite second coxae where it is widest ; with the usual gen- eral form of the genus ; .209 mm. tall at about the middle ; nearly level from PE to posterior declivity; median depression very shallow, opposite interval between second and third coxae. Eyes. Eight in two rows as usual; lateral ocular tubercles rather prominent ; viewed from above, both rows moderately re- curved ; viewed from in front, anterior row slightly recurved and posterior row slightly procurved, both measured by centers ; cen- tral ocular quadrangle wider behind than in front in ratio of about 4 : 3, wider behind than long in about the same ratio. Ratio of eyes AME : ALE : PME : PLE = 5.5 : 4 : 5 : 4. AME separated from one another by about 1.2 times their diameter, from ALE by about the same distance. PME sep- arated from one another by a little less than twice their diam- eter, from PLE by about two thirds as far. Laterals separated from one another by about 1.25 times their diameter. AME sep- arated from PME by a little more than the diameter of AME, hence further from one another than laterals are from one another in ratio of about 6 : 5. Height of clypeus equal to nearly 1.5 times the diameter of AME. BREVIORA NO. 68 Chelicerae. Well developed, moderately porrect, quite diver- gent in distal two thirds, somewhat swollen in middle ; prolateral spur a simple spine not bifid distally; fang- slender, slightly sinuate, with a blunt tubercle on inner margin about one-fifth of its length from base ; promargin of fang groove with four teeth, retromargin with four smaller teeth; with no "large tooth" on the promargin (Fig. 1). Maxillae. Nearly parallel ; slightly concave in middle of lateral border ; somewhat more than twice as long as lip ; three times as long as wide in middle. r n External Anatomy of Tetragnatha exigua sp. nov. Fig. 1. Chelicerae of male from in front. Fig. 2. Paracymbium of male palp. Fig. 3. Distal end of cymbium, conductor, and embolus. Fig. 4. Cheliceral teeth of female. Fig. 5. Genital fold of female. Liip. Much widened in basal third where it is wider than long in ratio of 22 : 14; sternal suture only slightly procurved; with the usual sternal tubercles well developed at ends of sternal suture. Sternum. Quite convex ; surface finely pitted and granulated ; with the usual form ; continued laterally and posteriorly between all coxae ; only a little longer than wide ; posterior coxae sepa- 1956 TETRAGNATHA IN JAMAICA 5 rated by a little more than their width. Legs. 1243. Width of first patella at "knee" .1083 mm., tibial index of first leg 4. Width of fourth patella at "knee" .0758 ram., tibial index of fourth leg 5. Femora Patellae Tibiae Metatarsi Tarsi Totals (All measurements in millimeters) 1. 2.275 .390 2.275 1.755 .718 7.413 2. 1.625 .325 1.430 1.380 .580 5.340 3. .910 .198 .445 .550 .308 2.411 4. 1.430 .260 1.170 1.235 .455 4.550 Palp .440 .120 .176 .396 1.132 Spines. True spines appear to be entirely lacking in this species (a very unusual feature) ; hair and bristles are sparsely present. Trichobotliria are present but have not been accurately observed. Palp. Both tibia and patella are short with tibia longer than patella in ratio of about 3 : 2. The paracymbium is unusually broad. The conductor and embolus are shaped and related essen- tially as shown in Figures 2 and 3. Abdomen. Slender ; broadest near base and gradually tapered to a blunt point posteriorly; bluntly truncated at base which is not notched; 1.495 mm. long; longer than wide in ratio of about 23 : 9 ; not continued posterior to spinnerets. Other features as usual in the genus. Color in alcohol. First and second femora yellowish ; all other segments of legs a dusky yellowish. Palps light yellowish except the reddish brown tarsi. Chelicerae : basal segment a deep red- dish brown; fang yellowish. Lip a deep reddish brown, lighter along distal border. Maxillae yellowish in medial third and brown elsewhere. Carapace a deep reddish brown, darker along the margins; median region with a narrow dark stripe posteri- orly and widening at the median depression and extending to PLE ; all eyes except AME surrounded by black pigment. Ster- num : a deep reddish brown. Abdomen : nearly white dorsally with a few silvery spangles ; in the posterior third there are very poorly outlined median gray spots with a series of very narrow black transverse lines; a fairly broad gray stripe extends along 6 BREVIORA NO. 68 each lateral side ; the venter is generally white with a little gray around the genital area and spinnerets. Female allotype. Total length including nearly vertical cheli- cerae 3.12 mm. Carapace 1.28 mm. long; .715 mm. wide opposite second coxae where it is widest ; otherwise essentially as in male. Eyes. Central ocular quadrangle wider behind than in front in ratio of 5 : 4, wider behind than long in ratio of 5 : 4. Ratio of eyes AME : ALE : PME : PLE = 6 : 4.5 : 5.5 : 5. AME separated from one another by five-sixths of their diameter, from ALE by 1.5 times their diameter. PME separated from one another by slightly more than 1.6 times their diameter, from PLE by the same distance. Laterals separated from one another by the diameter of PLE. AME separated from PME by the diameter of PLE, hence as far from one another as the laterals are from one another. Height of clypeus equal to about tw^o- thirds of the diameter of AME. Chelicerae. Moderately well developed ; nearly vertical and parallel ; basal segment .454 mm. long and, therefore, about one- third as long as cephalothorax ; fang slender and evenly curved ; promargin of fang groove with four well-developed teeth fairly evenly spaced ; retromargin with four smaller and fairly evenly spaced teeth (delicacy of the specimen makes it difficult to ob- serve teeth accurately). Maxillae, Lip, and Sternum. Essentially as in male. Legs. 1243. Width of first patella at "knee" .119 mm., tibial index of first leg 5. Width of fourth patella at "knee" .097 mm., tibial index of fourth leg 7. Femora Patellae Tibiae Metatarsi Tarsi Totals (All measurements in millimeters) 1. 2.210 .378 2.015 2.015 .716 7.334 2. 1.625 .330 1.170 1.430 .585 5.140 3. .845 .200 .520 .550 .396 2.511 4. 1.495 .265 1.105 1.170 .401 4.436 Spines, hairs, and trichobothria essentially as in male. Abdomen. 1.95 mm. long; broadest near middle where it is .910 mm. wide ; slightly notched at base ; genital area essentially as shown in Figure 5. Otherwise essentially as in male. 1956 TETRAGNATHA IN JAMAICA 7 Color in alcohol. Abdomen : tlorsally the cardiac area is nearly colorless; there are numerous silvery spangles and a vaguely outlined folium; the venter has a central slightly gray- ish stripe with a stripe on each side outlined by silvery spangles. Otherwise essentially as in male. Type locality. Ilolotype male, allotype female, and three para- type males from Hanover, Askenish, Trail to Dolphin Head, Jamaica, June 24, 1954. Tetragnatha PAiiLESCENS F. P. Cambridge, 1903 Eugnatha pallcscens Petrunkevitch, 1911 r. pallescens Petrunkevitch, 1930 T. pallescons Bryant, 1940 T. pallescens Bryant, 1945 Collection records: A male and a female from Ocho Rios, January 1929 (W. S. Brooks) ; several of both sexes from St. Catherine, Port Henderson, Salina, November, 1949 (Bengry, Lewis, Wiles) ; both sexes from St. Thomas, Lysson, June, 1954. Tetragnatha tenuissima O. P. Cambridge, 1889 T. tenuissima Petrunkevitch, 1930 T. tenuissima Bryant, 1940 T. tenuissima Bryant, 1945 Only one specimen, a male, has appeared in the collection available to me; St. Elizabeth, Magotty, May, 1953 (G. R. Proc- tor). Tetragnatha visenda sp. nov. (Figures 6-9) 3Iale holotype. Total length including chelicerae 8.58 mm. ; without chelicerae total length 7.475 mm. Carapace 2.60 mm. long; 1.495 mm. wide opposite second coxae where it is widest; with the usual general form of the genus ; .66 mm. tall opposite third coxae just anterior to posterior declivity. Eyes. Eight in two rows as usual ; lateral ocular tubercles only moderately prominent; viewed from above, posterior row mod- erately recurved, anterior row strongly recurved; viewed from 8 BREVIORA NO. 68 in front, anterior row moderately recurved, posterior row slightly procurved, all measured by centers; central ocular quadrangle wider behind than in front in ratio of 6 : 5, wider behind than long in ratio of 9 : 8. Katio of eyes AME : ALE : PME : PLE = 11 : 5.5 : 8 : 7.5. AME separated from one another by slightly more than their diameter, from ALE by r <^\ External Anatomy of Tetragnatha Fig. 6. T. visenda sp. nov. ; clieliceral teeth of male from below. Fig. 7. Idem; the prolateral spur of male. Fig. 8. Idem; the male paracymbium. Fig. 9. Idem; distal end of male tarsus. Figs. 10-11. T. versicolor Walck.; distal ends of conductors and emboli from Cuba and Michigan, respectively. Figs. 12-13. T. parva Bryant ; distal end of male tarsus and paracymbium, respectively. nearly twice their diameter. PME separated from one another by 2.5 times their diameter, from PLE by slightly more than this. Laterals separated from one another by slightly more than the diameter of PLE. AME separated from PME by slightly 1956 TETKAGNATHA IN JAMAICA 9 more tlian the diameter of AME, thus are farther from one another than laterals are from one another in ratio of about 12 : 7.5. Height of clypeus equal to a little more than two-thirds the diameter of AME. Chelicerae. Well developed; moderately porrect; quite diver- gent in distal two-thirds of basal segment ; somewhat swollen in distal half; prolateral spur well developed and clearly bifid with the larger lobe directed inward (Fig. 6); the fang is long, slender, only slightly sinuate; the fang groove has the so-called "large tooth" with eight others on the promargin and eight on the retromargin (Fig. 6). Maxillae. Slender; considerably divergent in distal halves; a little more than twice as long as lip ; longer than wide in middle in ratio of 4 : 1. Lip. Only slightly wider at base than long; sternal suture clearly procurved; with the usual sternal tubercles well de- veloped at ends of sternal suture. Sternum. Only slightly convex; somewhat swollen opposite second coxae ; with the usual general form ; longer than wide in ratio of 12 : 7 ; continued laterally and posteriorly between all coxae ; posterior coxae separated by about one-fourth their width. Legs. 1243. Width of first patella at "knee" .395 mm., tibial index of first leg 5. Width of fourth patella at "knee" .260 mm., tibial index of fourth leg 6. Femora Patellae Tibiae Metatarsi Tarsi Totals (All measurements in millimeters) 1. 6.500 1.105 6.890 7.475 1.560 23.530 2. 4.680 .910 4.225 4.420 .975 15.210 3. 2.340 .550 1.397 1.755 .600 6.642 4. 4.940 .715 3.835 4.420 .845 14.755 Palp 1..560 .370 .520 .850 3.300 Spines. All legs with spines of moderate size and length; a sparse coating of hair is also present. Trichobothria are present on femora and probably other segments of legs but their position has not been accurately observed. Palp. Tibia and patella both short with tibia longer than patella in ratio of about 3 : 2. The paracj^mbium is rather long 10 BREVIORA NO. 68 and slender with the chitinoiis knob about one-third of length of the structure from base (Fig. 8). The conductor terminates in a characteristic manner best shown in Figure 9. Abdomen. Slender; only slightly concave at base; widest near middle; 5.07 mm. long and about 1.43 mm. wide in broadest region ; not continued posterior to spinnerets ; other features as usual in the genus. Color in alcohol. All legs light yellowish brown, lighter below ; first and second somewhat dusky dorsally and dorsolaterally with occasional grayish patches ; third and fourth mostly lacking the dusky coloring and grayish patches. Chelicerae reddish brown, grayish along lateral surfaces. Lip dark brown with yellowish distal border. Maxillae yellowish in medial halves, darker along lateral halves. Carapace reddish brown with darker radiating streaks and an irregular granular border. Sternum reddish brown Avith darker irregularly grouped fine dots. Abdo- men : dorsum light yellowish because of presence of numerous irregular subchitinous yellowish white deposits making this re- gion very granular in appearance ; there are also many short irregular grayish lines which become more longitudinal in posi- tion and prominent along the lateral sides; the venter has a median grayish stripe with a broader granular j^ellowish stripe on each side together with a white spot just lateral to each an- terior spinneret and a smaller white spot just dorsal to the larger one. Type locality. The male holotype was taken at St. Catherine, Port Henderson, June 20, 1954. One male paratype is in the collection from a house in St. Andrew, August, 1955 (G. R. Proctor). THE GENUS IN CUBA Tetragnatha antillana Simon, 1897 There is but one specimen, a male, in the collection in the Museum of Comparative Zoology but the species has been re- corded from several localities. Tetragnatha caudata Emerton, 1884 Two females are in the collection in the Museum of Compara- tive Zoology, both taken at different times in Soledad gardens. 1956 TETRAGNATHA IN JAMAICA 11 Tetragnatha elongata Walckenaer, 1805 This species is well represented by both sexes in the collection and appears to be the most common of all of the eight species recorded from the island. Tetragnatha guatemalensis 0. P. Cambridge, 1889 T. banhsi McCook, 1893 T. seneca Seeley, 1928 T. banlcsi Levi and Field, 1954 Drs. Gertsch and Levi have apparently agreed that T. seneca Seeley is the same as T. hanksl MeCook. I have made careful comparisons of the specimens from Cuba identified as T. seneca Seeley with my numerous specimens of T. guatemalensis 0. P. Cambridge from Panama and other parts of Central America with the result that I am convinced that here we have another case of synonymy. The characteristics of eyes, several features of the male palps such as vermiform distal end of the paracym- bium and course and shape of both conductor and embolus to- gether with the general characters of the chelicerae all point toward this conclusion. Some may object that I am taking too much liberty with the cheliceral teeth because the "large tooth" is absent in T. seneca Seeley. This "large tooth" is not well developed in T. guatemalc7isis 0. P. Cambridge and could very well have been reduced to the condition found in T. seneca Seeley. The specimens in the Museum of Comparative Zoology identified as T. ha^iksi McCook also agree well with T. seneca Seeley as concluded by Levi and Field (1954). Tetragnatha orizaba Banks, 1898 There are several specimens of both sexes from several locali- ties in Cuba. Also recorded from Hispaniola but from no other place in the West Indies so far as know^n to the author of this paper. Tetragnatha pallescens F. P. Cambridge, 1903 Both sexes are represented in the collection from Havana and Soledad. Tetragnatha tenuissima 0. P. Cambridge, 1889 Both sexes have been recorded from several localities. 12 BREVIORA NO. 68 Tetragnatha versicolor Walckenaer, 1841 T. extensa Emerton, 1884 T. d&ntigera F. P. Cambridge, 1903 T. extensa Seeley, 1928 The specimens from Soledad identified as T. dentigera F. P. Cambridge are, I believe, correctly recognized. These are espe- cially interesting because of the fact that I have been forced to the conclusion that T. dentigera F. P. Cambridge is a synonym for T. versicolor Walck. I have examined a large number of specimens assigned to the latter species and as many as possible of the former. The cheliceral teeth, several features of the male palps (paracymbium, conductor, embolus) and the eyes all point rather decisively toward the synonymy which I have indicated. The tip of the conductor is like nothing else in the genus so far as I have seen. I have provided a drawing of the tip of the con- ductor from a specimen collected in Cuba and another from a specimen of T. versicolor taken in Michigan. There are slight differences but the basic plan is the same and the differences are well within the normal variation of a species (Figs. 10, 11). THE GENUS IN HISPANIOLA Tetragnatha antillana Simon,, 1897 T. festina Bryant, 1945 (male only) T. haitiensis Bryant, 1945 (females) The males of T. festina Bryant have the bifid paracymbium, other features of the male palpi, eyes, and general cheliceral characters associated with T. antillana Simon. 2\ haitiensis Bryant has the general form, cheliceral characters, and genital area characteristic of females of T. antillana. The small differ- ences noted by Miss Bryant and myself are all, I believe, within the normal variation for a species ranging over a Avide area. Tetragnatha confraterna Banks, 1909 T. eloufjata Bryant, 1945 One female from Puerta Plata, Dominican Republic, was identified as T. elongata Walck., 1805. I have examined this specimen very carefully and I am convinced that it has been 1956 TETRAGNATHA IN" JAMAICA 13 misidentified. Its slightly extended abdomen, cheliceral char- acters, eyes, and f^enital area seem to place it in the species 7'. confraterna Banks where it is provisionally left. Tetragnatha ORIZABA Banks, 1898 r. Orizaba Bryant, 1945 The specimens of both sexes from the Dominican Republic seem to agree well with our current understanding of this species. Tetragnatha pallescens F. P. Cambridge, 1903 r. pallescens Bryant, 1945 Numerous specimens of both sexes from Haiti seem to indicate that this may be the most common species in Hispaniola. Tetragnatha tenuissima 0. P. Cambridge, 1889 T. tenuissima Bryant, 1945 T. festina Bryant, 1945 (females only) The cheliceral teeth, eyes, general form, lack of leg spines, and the genital area all indicate that T. festina females belong with T. tenuissima 0. P. Cambridge. THE GENUS IN PORTO RICO I have had very little opportunity to study the species of Tetragnatha from Porto Rico. Petrunkevitch (1930) listed the following species from this island : T. antillana Simon ; T. elymiquensis Petrunkevitch; T. lahoriosa Hentz; T. pallescens F. P. Cambridge ; T. piscatoria Simon ; T. suhextensa Petrunke- vitch; T. tenuissima 0. P. Cambridge; T. vicina Simon. There appears to be considerable doubt about the occurrence of T. vicina in Porto Rico. In 1947 Miss Bryant described T. parva from the Luquillo Mountains thus bringing the total number of recorded species in this island to nine. Two figures of the tip of the conductor and closely related structures have been prepared to supplement those provided by the author of the species (Figs. 12,13). 14 BREVIORA NO. 68 THE GENUS IN THE VIRGIN ISLANDS Very little opportunity has been afforded me to study the spiders of these small islands. I have carefully examined all specimens, however, now in the collection of the Museum of Com- parative Zoology and am prepared to summarize my findings as follows: The vial labelled T. antillana Simon contains specimens belonging to this species but it also contains two females which I am tentatively assigning to T. confraterna Banks. The same vial contains a male palp which I believe was derived from this same species. Perhaps we may at least tentatively regard this species as being in the West Indies. The single male assigned to T. piscatoria Simon is, in my judgment, not this species but a specimen belonging to T. visenda sp. nov. GENERAL SUMMARY At the present time it seems that either fifteen or sixteen spe- cies of Tetragnatha may be considered to be known from the West Indies under consideration here. T. elijunquensis Petrun- kevitch, T. exigua sp. nov., T. parva Bryant, T. suhextensa Petrunkevitch, and T. visenda sp. nov. are known only from the islands from which they were described. T. antillana Simon has been recorded from all islands mentioned here. T. caudata Emerton has been recorded from Cuba and Jamaica. T. con- fraterna Banks is probably to be regarded as present in both Hispaniola and the Virgin Islands. T. elongata Walck. is only certainly known from Cuba. T. guatenialensis 0. P. Cambridge is now known only from Cuba. T. lahoriosa Hentz is, apparently, present only in Porto Rico. T. orizaha Banks appears to be in Cuba and Hispaniola. T. pallescens F. P. Cambridge has been recorded from all islands considered here except the Virgin Islands. T. piscatoria Simon is known only from Porto Rico among the islands considered here but it was described from St. Vincent. T. tenuissima 0. P. Cambridge is known from all of these islands except the Virgin Islands. T. vicina Simon has been recorded from Porto Rico but its presence there appears doubtful. 1956 TETRAGNATHA IN JAMAICA 15 BIBLIOGEAPHY / Banks, Nathan 1898. Aiaclmida from Baja California and other parts of Mexico. Proc. California Acad. Sci., Ser. 3, Zoology, 1. (7): 205-309, 5 pis. 1909. Arachnida from Costa Rica. Proc. Acad. Nat. Sci. Philadelphia, April, 1909: 194-234. Bryant, Elizabeth B. 1940. Cuban spiders in the Museum of Comparative Zoology. Bull. Mus. Comp. Zool. at Harvard College, 86 (7): 249-532, 22 pis. 1942. Notes on the spiders of the Virgin Islands. Bull. Mus. Comp. Zool. at Harvard College, 89 (7): 319-363, 3 pis. 1945. The Argiopidae of Hispajiiola. Bull. Mus. Comp. Zool. at Har- vard College, 95 (4) : 359-418, 4 pis. Cambridge, O. P. and F. P. Cambridge 1889-1905. Arachnida-Araneida. Vols. I-II. In: Biologia Centrali- Amerieana. Dulau & Co., London. Levi, Herbert W. and Hoavard M. Field 1954. The spiders of Wisconsin. Amer. Midland Natur., 51 (2): 440- 467, 113 figs. Petrunkevitch, Alexander 1930. The spiders of Porto Eico. Pt. 2. Trans. Connecticut Acad. Arts and Sciences, 30: 159-355, 240 figs. Seeley, E. M. 1928. Eevision of the spider genus Tetragnatha. Bull. New York State Mus., 278: 99-150. WiEHLE, H. 1939. Die einheimischen Tetragnatha-Arten. Nova Acta Leopoldina (Halle), N.F. 6 (41) : 363-386. BREVIORA Mmseiiinni of Cooipsirative Zoology Cambridge, Mass. January 31, 1957 Number 69 A NEW ZODARIID SPIDER FROM PANAMA By Arthur M. Chickering Albion College, Albion, Michigan Until the discovery of the species described in this paper only one zodariid spider was known from Panama, althouoh a few others had been reported from Guatemala and Mexico. The single species hitherto known from Panama is Store7ia harroana (Chamberlin), 1925. During the summer of 1954, while sifting decaying hay along the roadside near Corozal, Canal Zone, I discovered one im- mature and two mature males together with one mature and three immature females all belonging to what I regard as a new species referred to the genus Leprolochus Simon, 1892. So far as I know, only two species in this genus have been previously reported. The genotype of this rather remarkable genus, Lepro- lochus spinifrons Simon, Avas from Brazil and L. parahyhae Mello-Leitao, 1917 was from Venezuela. The latter is known only from females. The description of this new species of Leprolochus is given in this brief paper in accord with my usual descriptive formula, and the types have been deposited in the Museum of Comparative Zoology. Genus LEPROLOCHUS Simon, 1892 Leprolochus decoratus sp. nov. (Figures 1-6) Male holotype. Total length 2.21 mm., including spines pro- jecting from head region. Carapace 1.235 mm. long, .770 mm. wide opposite second coxae where it is widest; bluntly rounded in front and with seven robust and bluntly pointed black spines External Anatomy of Leprolochtis decoratua Fig. 1. Anterior cephalic region of male from above. Pig. 2. Chelicerae, eyes, and special spines of male from in front. Figs. 3-4. Tarsus and tibia of male palp; ventral and retrolateral views, respectively. Fig. 5. Anterior cephalic region of female from above. Fig. 6. Ventral view of epigj'num. 1957 NEW ZODARIID SPIDER FROM PANAMA 3 projecting forward somewhat below AME (Figs. 1-2) ; the para- type male has nine of these spines; with a short and shallow median longitudinal thoracic groove opposite the third coxae ; finely granulated, especially over pars thoracica; gently arched from PME to posterior border with highest part shortly behind these eyes. Eyes. Eight in two rows; posterior row only slightly longer than anterior row; anterior medians dark, all others light. Viewed from above, posterior row strongly procurved, anterior row moderately so. Viewed from in front, both rows strongly procurved. Ratio of eyes AME : ALE : PME : PLE = 5.5 : 5.25 : 5 : 5. AME separated from one another by about 4/5 of their diameter, from ALE by about 8/5 of their diameter. PME separated from one another by about 9/5 of their diameter, from PLE by nearly twice their diameter. Laterals separated from one another by about 7/10 of the diameter of PLE. Central ocular quadrangle wider behind than in front in ratio of about 9 : 7, only slightly longer than wide behind. Height of clypeus equal to somewhat more than six times the diameter of AME. Chelicerae. Stout, vertical, parallel ; basal segment .352 mm. long ; with moderately distinct basal boss and fine striations anterior to boss ; the fang is weak, short, and apparently evenly curved ; fang groove obscure and with no teeth observed ; the pro-margin has a fleshy ridge with a row of stiff bristles. Maxillae. Distinctly convergent; tapered distally and nearly meeting beyond lip. Lip. Nearly triangular; wider at base than long in ratio of about 3:2; reaches about 2/5 of the length of the maxillae. Ster- nal suture straight. Sternum. Cordiform ; widest between first and second coxae where it is wider than long in ratio of 25 : 23 ; anterior border nearly twice as wide as base of lip ; very convex ; sharply pointed at posterior end but not extended between fourth coxae which are separated by 4/5 of their greatest width. All coxae moder- ately elongated. Legs. 4312. Width of first patella at "knee" .108 mm., tibial index of first leg 15. Width of fourth patella at "knee" .130 mm., tibial index of fourth leg 15. 4 BREVIORA NO. 69 Femora Patellae Tibiae (All measurements in Metatarsi millimeters) Tarsi Totals 1. .660 .264 .462 .660 .528 2.574 2. .638 .250 .440 .660 .506 2.494 3. .638 .275 .450 .748 .484 2.595 4. .811 .275 .660 .990 .690 3.426 Palp .330 .110 *.443 .374 1.257 * Including much extended apophysis. Spines. All legs with inconspicuous spines difficult in some instances to distinguish from coarse bristles. First leg: femur dorsal 1-1-1-2, prolateral 1 distal, retrolateral and ventral 1 weak proximal ; patella with numerous coarse bristles but probably only prolateral 0-1-1-0 spines; tibia dorsal 1-1-1 (all weak), prolateral 0-1-0-1, retrolateral 0, ventral only 2 distal; meta- tarsus 0. Second leg : essentially as in first except metatarsus which has a tuft of ventral bristles at distal end together with four spines on ventral and lateral sides. Third leg : femur apparently only dorsal 1-0-1-2, patella dorsal 1-1, prolateral 0 1-1-0, retrolateral 0-1-0, tibia dorsal 1-1-1, prolateral appar- ently 0-1-0-0 and wdth a group of distal bristles forming a comb- like structure, retrolateral 1 distal, ventral only distal 2 ; meta- tarsus only with distal spines essentially as in second. Fourth leg : essentially as in third. Trichohothria. Observed as follows : tarsi with a row of two or three with the most distal very long ; metatarsi the same ; tibiae apparently with two rows of two each in dorsolateral positions ; palpal tibia with two dorsal in a row with distal one very long. Claws. Two claws throughout, each pectinate in a single row ; with no claw tufts. Palp. Patella very short ; tibia with a short body but with a lateral apophysis which extends nearly the full length of the tarsus ; also with a pair of retrolateral spines as shown in Figures 3 and 4. Abdomen. Regularly ovoid ; longer than wide in ratio of 17 : 13; overlaps cephalothorax only slightly; widest slightly behind 1957 NEW ZODARIID SPIDER FROM PANAMA 5 the middle ; provided with a moderately well developed scutum ; well supplied wdth short appressed and recurved stiff bristles both black and transparent ; postabdomen and anal tubercle project posterior to scutum ; venter well chitinized ; with a short, recurved lip just anterior to genital groove ; tracheal spiracle obscure but apparently just anterior to base of spinnerets and appears to be marked by a row of relatively long stiff bristles; position of probable vestigial colulus obscure. With six pairs of spinnerets partly obscured by a group of bristles. Color in alcohol. Carapace : cephalic parts a light brown ; re- mainder a medium brown with darker streaks. Chelicerae light brown with other mouth parts and sternum yellowish brown. Legs : all coxae nearl}^ white ; femora brown, lighter beneath ; other segments light brown. Abdomen : dorsally light brown with three irregularly dark reddish brown cross bars in the posterior two thirds and an irregular central stripe reaching to base from first cross bar ; the posterior end of the abdomen behind the scutum and the postabdomen are white; the venter is yellow with varying shades. Female allotype. Total length 3.90 mm. Carapace 1.625 mm. long; 1.105 mm. wide opposite interval between second and third coxae where it is widest ; gently arched from PME to median, short, longitudinal thoracic groove from w^liich it descends rather gradually to posterior border ; cephalic portion well separated from thoracic part ; with eleven robust black spines in the posi- tion of the seven similar spines in the male (Fig. 5). Eyes. Ratio of eyes AME : ALE : PME : PLE = 6.5 : 5.3 : 6 : 6. AME separated from one another by about their diam- eter, from ALE by twice their diameter. PME separated from one another by 13/6 of their diameter, from PLE by 17/6 of their diameter. Laterals separated from one another by nearly the diameter of ALE. Central ocular quadrangle wider behind than in front in ratio of 4 : 3, longer than wide behind in ratio of 7 : 6. Height of clypeus equal to about eight times the diam- eter of AME. Chelicerae, Maxillae, Lip, and Sternum. Essentially as in male. Legs. 4321. Width of first patella at "knee" .16245 mm., tibial index of first leg 15. Width of fourth patella at "knee" .2166 mm., tibial index of fourth leg 15. BREVIOBA NO, 69 Femora Patellae Tibiae Metatarsi Tarsi Totals (All measurements in millimeters) 1. .975 .390 .682 .715 .650 3.412 2. .975 .390 .650 .975 .650 3.640 3. 1.040 .455 .780 1.235 .780 4.290 4. 1.170 .455 1.000 1.625 1.040 5.290 Spines. First leg : femur dorsal 1-0-1-2, probably prolateral and retrolateral only distal 1 or 2, ventral Ir (weak) near middle ; patella only prolateral 0-1-1-0 ; tibia dorsal 1-1-1, prolateral 1-1-1, retrolateral 0-1-0-1, ventral O-lr-0-2 ; metatarsus apparently with only two distal but with many spiniform bristles. Second leg : femur probably as in first; patella dorsal 1-1, prolateral 0-1-1-0, retrolateral 0-1-0 ; tibia probably as in first ; metatarsus as in first except with a distal ventral brush of bristles and probably a group of 3 or 4 weak spines. Third leg: femur, patella, and tibia essentially as in second ; metatarsus prolateral 0-1-0, retro- lateral 0-1-0-1, ventral with brush of bristles as in second and with five clear distal spines on ventral and both lateral surfaces. Fourth leg: essentially as in third except brush is absent but the spines are retained. Palp : with numerous spines on femur, patella, tibia, and tarsus ; tarsal claw finely pectinate with about ten slender teeth ; claw is opposed to a chitinous tubercle. Trich- obothria essentially as in male so far as observed. Abdomen. Well rounded, oviform; 2.405 mm. long, 1.95 mm. wide near middle; without scutum such as that possessed by male; with six spinnerets, the anterior pair quite robust and somewhat the longest ; without definite colulus ; tracheal spiracle close to base of the anterior spinnerets. Epigynum. Simple but well developed. There is a central shallow depression bordered anteriorly by a strongly chitinized recurved border. Two internal tubules occur at the posterior boundary and on each side there is a relatively large spermatheca (Fig. 6). Color in alcohol. Carapace essentially as in male except that at the base of pars cephalica there is a dorsal, somewhat oval brown spot on each side from which a thin irregular line extends forward to PLE. The dorsal abdominal region is also very sim- ilar to that of male except that the reddish brown markings are 1957 NEW ZODARIID SPIDER FROM PANAMA 7 broader and more united. Otherwise essentially as in male. Type locality. The male holotype, female allotype, one mature male paratype together with four immature specimens from Corozal, C. Z., July 10, 1954. BIBLIOGRAPHY Banks, Nathan 1929. Spiders from Panama. Bull. Mus. Comp. Zool. at Harvard Col- lege, 69:53-96, 4 pis. Cambridge, O. P. and F. P. Cambridge 1889- Arachnida-Araneida. In: Biologia Central- Americana. Dulau 1905. & Co., London. Chamberlin, R. V. 1925. Diagnoses of New American Arachnida. Bull. Mus. Comp. Zool. at Harvard College, 67:211-248. Chickering, Arthur M. 1947. The Male Allotype and Female Hypotype of Storena Barroana (Chamberlin). Papers Michigan Acad. Sci., Arts, Letters, 31 133-140. ROEWER, C. Fr. 1942. Katalog der Araneae. Vol. 1. Bremen. Simon, Eugene 1892- Histoire Naturelle des Araignees. Deuxi^me Edition. 2 Vols. 1903. Librairie Encyclopedique de Roret, Paris. BREVIORA Museiiam of Comparative Zoology Cambridge, Mass. January 31, 1957 Number 70 "ANGUIMORPH" TOOTH REPLACEMENT IN AMPIIISBAENA ALBA LINNAEUS, 1758, AND .1. FULIGINOSA LINNAEUS, 1758 (REPTILIA: AMPHISBAENIDAE) By Carl Gans In the course of an investigation into the status of the acro- (lont amphisbaenids it proved necessary to prepare the skulls of some related forms. When cleaning the mandible of a female specimen of Amphishoenn alha Linnaeus, 1758, it was noted that six to eight teeth were in the process of being replaced, while two or three others had only recently moved into position, and were as yet but imperfectly fused to the dentaries. The replace- ment teeth appeared to lie interdentally, between, rather than below^, their predecessors. McDowell and Bogert, in their recent revision of the angui- morph lizards (1954, pp. 102, 104, tig. 30), have stated that "alternate" tooth replacement was restricted to this group, all other lizards possessing "vertical" replacement. The amphis- baenids are generally considered to be Scincomorpha (Camp, 1923, p. 296) and if the distinction between an anguimorph and iiou-anguimorph pattern is as clearcut as McDowell and Bogert suppose, the occurrence of a pattern descriptively alternate would thus be unexpected here. For this reason and because certain other aspects of dental replacement in these forms seem worthy of special attention, it appears desirable to describe the female specimen mentioned above, as well as two specimens of A. fuligmosa Linnaeus, 1758, in which a similar replacement pattern was noted, in the hope of stimulating further research into these matters. BREVIOKA NO. 70 The mandibular ilentition of Amphishaena alba Figures 1 to 5 show lal)ial and lingual views of the mandil)h> of an adult (body length 440 mm.; length of mandibular ramus 15 mm.i) female specimen (MCZ 54299) from "Brazil." This had been cleaned by dissection following controlled applications of full strength commercial bleach to selected portions of the soft parts. Fig. 1. (Upper) Amphisbaena alba. Lingual view of right mandibular ramus of MCZ 54299. Fig. 2. (Lower) AiiiiiliLsb(n no alba. Labial view of same niamliliular ramus as Figure \. 1 The length of the mamllbulur riinius has been selected as a convenient indi- cator of the total length of the specimen and Is hence abbreviated as Imr. 1957 AMl'lllSBAENID TOOTH REPLACEMENT The first pa it of this description covers the general aspect of the tlentitiou and ap])lies equally well to the female cited and to larger si)eeinieus listed below. The description of the replacement l)attern and of individual or possibly ontogenetic variation fol- lows upon this. The dentition is weakly pleurodont, the height of the dorsal margin above the lingual shelf on which the teeth rest barel}^ Fig. 3. AmpMshaena alba. (MGZ 54299) Sketch of lingual view of right mandibular ramus showing location of the replacement teeth. equaling the width of the tooth base. There are eight teeth on each side ; the third is largest, the fourth, fifth, sixth, and seventh are slightly smaller, while the second, eighth, and first are pro- gressively smaller in that order. All of the teeth are curved, this being most noticeable near the tip, and the curvature is directed medially and slightly caudad. There is a rotation in the tooth alignment so that the planes of curvature of the individual teeth do not lie parallel to one another. The teeth are of oval cross- section w4th the long axis lying in the plane of curvature. The base of the fully formed tooth is hollow, the pulp cavity extending two-thirds of the total height of the tooth. About midway up the tooth the diameter of the cavity contracts so as to continue upwards as a thin cylindrical tube. There is no tooth-bearing shelf as in a typical pleurodont den- tition; instead, a ridge of bone (Fig. 6) rises to the projecting dorsal margin of the dentary between each two adjacent teeth (where replacement teeth are present this ridge is hollowed out to afford lodgement for the pit of the new tooth). Consequently UREVIORA NO. 70 tilt' base of eaeh tootli is almost entirely surrouiulcd by bone, to ^vhicll it is ankylosed by a ring of cement. The lowest exposed ])oint lies on the lingual aspect and careful, but thorouph, prepa- ration shows here a single round foramen.' The foramen leads into the ptdp cavity and ])resumably carries its vascular and nerve siipi)ly. A line extending to the level of the tip of the interior cavity is visible on the lingual side. Inspection along the lingual aspect of even the smallest cap-shaped tooth germs re- veals a slight scalloping so that this line may be formed during Fig. 4. AinpJtLshacna alhn. Lingual view of the left niaiulilnilar ramus of MCZ 54299. tooth development by the fusion of the anterior and posterior portions. Several of the skulls showed longitudinal cleavage of the teeth along this line, which coincides with the long axis of the oval tooth cross-section. \'arious stages of tooth replacement are shown in MCZ 54299. In the subsequent description of it and other specimens, L and R will stand for left and right mandibular ramus, and the number following this for the particular tooth or alveolus counting from front to back. At L-7, R-3 and R-4 the replacement tooth is but a small hollow conical shell, thicker at the top than at the sides, its lower edge somewhat excavated on the lingual side. It lies in a small and deep depression in the deutary in line with the posterior edge 1 These foramina, as shown in Figure 6, are of a larger specimen, which could be cleaned completely without danger of dislodging the firmly fused teeth. [9i )* AMPHISBAENID TOOTH KKPI.ACEMENT of the tooth it is replaciiij?. As a result, this pit lies on the ascendino; ridoe directly between the tooth to be replaced and the one posterior to it. Even at this stage there is already some erosion into the base of the tooth beinj^- replaced. L-4 and L-5 show a slightly more advanced stage with pro- gressive erosion into the precursor and an increase in size of the enamel cap of the new tooth. In R-8 the replacement tooth has reached approximately one- third of its final height, though its tip is still below the level of the projecting dorsal margin of the dentary. It has, however, destroved almost half of the base of the tooth Iving above it. 8 7 6 5 4 3 Fig. 5. Amphishacna alba. (MCZ 54299) Sketch of lingual view of left iiiniidilmlar ramus showing location of the replacement teeth. At R-6 the tooth has just moved into its final position, its predecessor having been pushed out at some prior stage. It is still only loosely held in place by struts of cement, and appears slightly smaller, and thinner-walled than its neighbors. Its dark appearance is due to the contents of the large, soft-tissue-filled pulp cavity being visible through the translucent walls. This may indicate that the buihl up of the internal dentine layers is not completed until after the tooth is finally cemented into place. The heavier enamel and more opaque, shell-like appearance of L-3 may indicate a more advanced stage of development. It is, however, still darker than its neighbors, and its base is still far from fused to the dentary. Its slightly cocked position seems to indicate that it is not fully aligned as yet. All of the replacement teeth except the last two lie freely in the soft tissues and show no fusion to the dentaries. 6 BREVIORA NO. 70 Another interesting point is that the alveolus formation ap- pears to start in the interdental ridge and only begins to extend into the base of the precursor tooth as growth takes place. No pits or alveoli were found at the base of any tooth that Avas not undergoing replacement. Though the replacement pattern is interdental it differs from that described by Camp (1923, p. 329, fig. H) for Gerrhonotus s. scincicauda (Skilton), 1849 (and from that seen in a skull of the same form in the MCZ). In this specimen there is no trace of cavity or alveolus formation and the replacenumt tooth appears Fig. 6. Ampliishaena alha. (MCZ 32257) Dorsal view of tip of mandible, to show interdental bone ridges, basal foramina, and fusion lines. to lie always in the layer of tissue next to the bone. Gerrhonotun .s. scincicauda is also definitely pleuroe alveoli in each of the four Fig. 7. (Upper) Amphishaena fnliginosa. (MCZ 2154) Lingual view of loft iii,niidit)nl;ir ramus. Fig. 8. (Lower) Amphisbaena fuliginosa. (MCZ 7799) Lingual view of liglit ninnrliliular ramus. \'Xu AMI'IUSBAENID TOOTH REPLACEMENT iiiaiHliliular i-aiui. Ilowcvcr, tlu' broken tooth spacinji' ot" the soiiu'wliat luacoratod nuuidil)l(' of one of tlic spccinicns (i\ICZ 779!)) (Figures 8. 10) leads one to suspect that Ave are dealinj; with a row of ei o lo " o o o CO CO o - aj » o in I cc eg o-i ^ S 5 E t~ m t~ a C ^ •-' w a ■ ■ • £; S " O O o t^_i S cq to I o "^ O o 3 =" 1/ CO o ; O M - o o lO o m o uo o d C-J 1 o 0—20. .5—19 .0—20, f-H ,0—19. 19.0 17.5 .0—18, 00 f-H 1— ( 00 t- t~ ^^ l~t r^ ui 00 t^ IM « 00 o CO a> ■* eg o -1 00 1 1 Tj< o o to lO <-H t^ t^ CO iO ^ C-J o ■■=: 1 1 "". o:' 00 00 oo' oo' oo' 00 1 1 t~ w ^ a ^ oi El001Cg00<*>''5o to CO rH to £ S O ; T-H O CO to OJ 00 ^ ' d o> o' t-' S SS c' CO oq CO (N "^ ^ CO o o o" to CO CO o in a> ■*' eg CO o o o o o o <:5 , ei o' oi to r; =£ d — . SeO lA m -Tj* up -^j* .^ > d = ,ii _ o o >q o lo -"ji oo" -^ ■* ei t^ r stoootoej^^cooo >Q^ejioco'^eJeQ o t^ ^ O 00 00 o .„CO»COiCOeO»Ou5 •*dtod»ot*wo6 ■^ •<**■«* Tj< -^ -^ Tt< lO CO eQ CO to' co' •^ in CO o oi , CO o eg o o d eg in CO d CO CO ;• in o 00 o 00 03 to Cl CO ■* ci h n to eg eg o to irj 00 CO ■* in i i o d LO f-H o o CO i-H lO *oo+*oo+*oo+*o CH*o ot*o o U3 o o in o o in CO CO CO CO CO CO in CO cfi ^7 to o o 1 o i 1 in 1 in 1 o i 1 o_ CO CO CO d CO d eg CO t-' eg d CO CO o in in r-i in •4" CO in to 1 1 I-H f-H f-I O lO o 00 o o o eg 00 o o CO 1 1 eg CO f-f CO CfO CO d CO eg' CO f-H CO in CO 1 i oo in t^ 00 t» OJ in - - >o o in in f-H o o in in '? f-i in m in m o o 1 in 1 in 1 in i i 1 o 1 in eo' in in m eg d in d f-H m oo' Cij m CO to f-^ o "*. to CO o eg to eg 1 1 o to eg to o in eg o in 1 1 ^' d lO Ci^ m d CO in d in in m 1 1 »o o *o Of S- ■g ■ ts ■ :2 ftt •2 .^- 4 BREVIORA No. 71 dorsally it is, almost constantly, slightly darker than birds of the northern race collected in the same season. A shorter wing and smaller bill were also attributed to pullus. However, 21 adult males collected between late May and early August in New- foundland, New Brunswick, Quebec, Alberta, Maine, New Hamp- shire, New York, and Michigan had a mean wing length of 49.38±:.28 mm., and 20 of the same series had a mean culmen length of 14.25±:.15 mm., dimensions which do not differ sig- nificantly from those of breeding males from the Api)alachians (Table 1). The race breeds in the mountains from Virginia to Georgia. It is absent from its breeding grounds during the winter and presumably mingles with hiemalis in the southeastern United States. Differentiation of the two races in the winter is uncertain. The post-nuptial plumage of the species is considerably darker than the breeding plumage and although one might assume that in this dress pullus remains slightly darker than hiemalis, there is no evidence that this is true. Troglodytes troglodytes pacipicus Baird Troglodytes hiemalis var. pacificiis Baird, 1864, Rev. Anier. Birds, 1, (18(i-t- 1873), p. 14.5 — Simiahmoo, Puget Sound, Washington. The western continental race is much more richly colored throughout than hiemalis and pullus and somewhat more fre- (piently lacks dorsal barring. Ridgway (1904) stated that the bill of pacificus is more slender and less curved, but I cannot confirm this. The mean wing length of males of pacificus is substantially shorter than that of the males of the eastern races (Table 1), although the overlap in measurements is almost complete, rendering the character of little taxonomic value. Females of pacificus also have shorter wings than their eastern counterparts, but the difference is less pronounced than in the males. The breeding range is from southeastern Alaska and the southern Yukon southward through the mountains to Idaho and central California. In the winter the race moves to lower elevations and occurs casually south to Arizona and New Mexico. 1957 NEW WORLD FORMS OF TROGLODYTES 5 Troglodytes troglodytes iielleri (Osgood) Anorthura hiemalis hcUeri Osgood, 1!)01, Aiik. 18. \k ]S] — Englisli B;iy, Kodiak IsLand, Alaska. Kodiak and Afognak Islands are occupied by a nonmigratory race which is most similar to pad fie us, but distinguished by its less rich coloration, reduced abdominal vermiculations. and slightly larger average size. Troglodytes troglodytes semidiensis (Brooks) Xanniis IiicmaJis semidiensis Brooks, 1915, Bull. Mus. Comp. Zool., 59. p. 400 — Chowiet Island, Semidi Islands, Alaska. It is with reluctance that this form is accepted. It appears to be distinguished from kiskcnsis only by the greater average length of the bill of the males (19. 66 ±.19 mm. vs. 18. 57 ±.14 mm. ; 0.05>P>0.02). However, there are only three males in the sample from the Semidi Islands, and their bill measurements (19.0, 20.0 and 20.0 mm.) fall within the upper range of males of kiskensis (17.0 to 20.5 mm.). Under many conditions one might ascribe the observed difference to chance and with little hesitancy consider the two populations unworthy of separate designation. However, it is obvious that the birds are markedly different from helleri of the Kodiak region, Avhich is about 100 miles away, and resemble the Aleutian form, from which they are isolated by roughly 400 miles. The extent of isolation makes it appear probable that the observed differences are indicative of two populations which are morphologically distinct. It is for this reason that it .seems best to accept the data at face value and maintain semidiensis. Specimens have been taken on Chowiet, Aghiyuk, and North Semidi Islands. Troglodytes troglodytes kiskensis (Oberholser) Kannus troglodytes kiskensis Olierholser, 1919, Proc. U.S. Nat. Mus., 55, p. 228 — Kiska Harbor, Kiska Island, Aleutian Islands, Alaska. Xa7inus troglodytes tanagensis Oberholser, 1919, Proc. U.S. Nat, Mus., 55, p. 230 — Tanaga Bay, Tanaga Island, Aleutian Islands, Alaska. Kannus troglodytes petrophilus Oberholser, 1919, Proc. U.S. Nat. Mus., 55. p. 232 — Unalaska, Unalaska Lsland, Aleutian Islands, Alaska. 6 BREVIORA No. 71 Nannus troglodytes stevensoni Obeiholser, 1930, Proc. Biol. Soc. Washington, 43, p. 151 — Amak Island, Alaska. Troglodytes troglodytes seguamensis Gabrielson and Lincoln, 1951, Proc. Biol. Soe. Washington, 64, p. 73 — Seguam Island, Aleutian Islands, Alaska. The Avreiis of the Aleutian Islands present a number of prob- lems, the most difficult of which is the condition of the available specimens. Almost all collecting has been done during the late spring and the summer, when the birds are in their most worn plumage, or are immature. Of 110 Aleutian specimens examined, only 15 were taken between the months of October and April. The condition of wear is frequently variable ; even birds from the same island, collected on the same day, maj^ exhibit marked dif- ferences. The poor condition of the specimens seems to have been the main reason for the description of six races from the region, although only two stand up under critical examination. On the basis of five adult males and ten young birds, the population of Amak and Amagat Islands was named "steven- soni." It was defined as differing from "petrophilus," of the Fox Islands, in being more gray in both adult and juvenal plumages, in having fewer vermiculations below in adult plum- age, and in having a slightly shorter bill in the adult. I cannot appreciate any of the supposed color characters. The difference between the mean lengtji of the bills of three males from Amak and Amagat Islands (17.50±.14 mm.) and that of nine males from the Fox Islands (18.16±.14 mm.) is without question insignificant. (P>.10). In the original description, " petrophiliis" was compared with alascensis, of the Pribilofs, even though the describer (Ober- holser, 1919) admitted that his new race was nearest to "tana- runneicollis and solstitialis, I do think that Hellmaja- (oj). cit.) was correct in believing it to be an offshoot of solstitialis. The only character which distinguishes it from all of the forms of solstitialis is its large size. Being con- fined to a small area on the top of a mountain, it is analogous to an insular population, in^ which situation a race often is larger than its congeners. It would seem best to consider monticola as merely another race of solstitialis. LITERATURE CITED Armstrong, Edward A. 1955. The Wren. Collins, London, viii + 312 pp. Bond, James 1956. Check-List of Birds of the West Indies. Acad. Nat. Sci. Phila- delphia, vi + 214 pp. Chapman, Frank M. and Ludlow Griscom 1924. The House Wrens of the Genus Troglodytes. Bull. Amer. Mus. Nat. Hist., 50: 279-304. 1957 NEW WORLD FORMS OF TROGLODYTES 15 DiCKFA-, Donald K. and A. J. van Rossem 1938. The Birds of El Salvador. Field Mus. Nat. Hist., Zool. Ser., 23. 609 pp. Hkllmavk, Charles E. 1934. Catalogue of Birds of the Americas and the Adjacent Islands. Field Mus. Nat. Hist., Zool. Ser., 13. pt. 7, vi + .531 pp. Marshall, Joe T. 19.")6. Summer Birds of the Rincon Mountains, Saguaro National Monument, Arizona. Condor, 58: 81-97. Nelson, E. W. 1901. Descriptions of a New Genus and Eleven New Species and Subspecies of Birds from Mexico. Proc. Biol. Soc. Washington, 14: 169-175. Oberholser, Harry C. 1919. Notes on the Wrens of the Genus Nannus Billberg. Proc. U.S. Nat. Mus., 55: 223-236. 193-1. A Revision of North American House Wrens. Ohio Jour. Sci., 34: 86-96. Paynter, Raymond A., Jr. 1955. The Ornithogeography of the Yucatan Peninsula. Bull. Peabody Mus. Nat. Hist., Yale Univer., 9. 347 pp. RiDGWAY, Robert 1904. The Birds of North and Middle America. Bull. U.S. Nat. Mus., 50. pt. 3, XX 4- 801 pp. Sutton, George Miksch 1951. Mexican Birds. First Impressions. U. of Oklahoma Press, Nor- man, XV -|- 282 pp. ^^Ui) BREVIORA Meseeei of Cooipsirative Zoology Cambridge, Mass. March 29, 1957 Number 72 IS THE ANT GENUS TETRAMORIUM NATIVE IN NORTH AMERICA? By W. L. Brown, Jr. For some years, a controversy has continued concerning the distribution in North America of the ant genus Tetramorium Mayr. On opposite sides in this argument ha.ve been Dr. M. R. Smith (1943), who believes that all five of the Tetramorium species reported from North America have been introduced by man from overseas, and Dr. W. S. Creighton (1950), who thinks that the evidence points to prehistoric endemicity in the continent for at least two of the species mentioned ; T. cacspitnm (Linnaeus) and T. riigiventris M. R. Smith. Apparently there is no serious disagreement about the origin abroad of 2\ guineense (Fabri- cius), T. simillimum (F. Smith) and T. pacificum Mayr. The first two of these are very likely African in origin, since they occur in wild parts of Africa and have their closest relatives among the species of that continent. T. pacificum is apparently from the Indo-Australian area, although its exact source has never been seriously tracked down. Certainly, it is Old World in origin. The two species caespitiim and rugiventris are therefore the central elements in the discussion, and it seems appropriate at this time to review the important facts in connection with these two forms and to add whatever significant observations are now available. Tetramorium caespitum This species is widespread in the Palaearctic region and in Africa, where it is extremely variable, as attested by an almost endless list of infraspecific variants. In terms of modern tax- 2 BREVIORA NO. 72 onomic practice, many of these entities would be considered good sibling species, as indeed some of them already have been treated by various European authors. Many others are mere synonyms of caespitum — individual or nest variants that do not represent natural, self-maintaining populations. No one has challenged the placement of the North American populations with the more nearly "typical" caespitum variants. Creighton's assertion that caespitum is a native Nearctic ant rests on two principal pieces of evidence : first, the species was known at a very early date, i.e., 1895, from states as remote as Tennessee and Nebraska ; second, there is present in North America, at least in the eastern states, a workerless parasite of caespitum, the aberrant species Anergates atraiulus (Schenek), the transport and establishment of Avhich would seem to present special difficulties. The difficulty with the first piece of evidence is that the exact situation of the collections made in Tennessee, Nebraska and elsewhere in the "interior" of North America was never specified. All of these collections may have been, and probably were, made in or near "culture areas," that is, regions strongly disturbed by the presence of man. The experience of several practiced myrmecological field workers, including that of Dr. Creighton (personal communication) and myself, indicates that so far as known, caespitum in North America is known only from rather heavily disturbed localities, such as cities, towns, road- sides, farmyards, picnic areas and the like. I have been able to gather no records at all to show the existence of the species in places remote from the works of man in North America. This situation contrasts with that holding in the Old World, at least so far as my own personal experience goes, and judging also from what I have been able to glean from various publica- tions dealing with the species, and from personal communications with European myrmecologists. To sum up this information, it can be stated that T. caespitum in Europe and (although given various infraspecific names) in China is often abundant in and around human habitations, just as in North America. However, it is also to be found, often in abundance, in localities that show little or no trace of human disturbance, and that are far from the nearest humanly-occupied places. The contrast is very marked if one collects, as I have, at intermediate altitudes in 1957 TETRAMORIUM IN NORTH AMERICA 3 West China, in pine-oaJc forest, and then compares the collec- tions of Tetramorium obtained with the results of a deliberate search for Tetramorimn in similar vegetational zones in Pennsvl- vania, New Jersey or Massachusetts. In West China, T. caespi- ium tends to occur uniformly throughout the pine-oak forest, regardless of roads, villages, etc., which are very sparsely dis- tributed in the areas I am recalling. In ecologically equivalent areas in the eastern United States, I have found the same species established only on or near the sites of more or less actively maintained human works. Clearly, the density of nests and individuals seen in North America is in large part proportional to the degree of urbanization of the area occupied, although the real extremes of urbanization, where almost all space is covered by concrete or asphalt, are certainly not favorable locales for colonization by this or any other ant species that lives largely in the open. In my opinion, the local, detailed distribution of T. caespitum in North America is that expected of an historically introduced, not a native ant. The second point of evidence, that concerning the presence of tlie workerless, and therefore obligatory parasite, Anergates atra- iulus, certainly seems on the face of it a real sign of long occupancy of North America by both host and parasite. Creigh- ton emphasizes the difficulties facing trial colonists of the parasite species : first, the species seems to be relatively rare in Europe ; second, it is unlikely that the parasite female could survive a long trip; third, a female arriving in North America would be hard put to find a suitable nest of the host species to enter. To take these difficulties one at a time, we should first recognize that, while Aner gates is not the commonest of ants in collections, it is nevertheless likely to be locally very common in restricted localities. Even in Europe, host populations are normally concentrated in gardens and waste places within towns and cities, including seaports, and these are accordingly \evy likely places for Anergates to occur unnoticed by primarily country-searching myrmecologists. In the United States, Aner- gates is known chiefly from East Coast localities in urban areas near the sea — exactly the kind of place from w'hicli colonists might be expected to be exported most easily. It should be added that Anergates females may be produced in very large numbers 4 BREVIORA NO. 72 from a single nest, so that a given locality may be heavily saturated with them during the period of nuptial flights. The difficulty of transport of live Aner gates propagules is real, but far from insurmountable. Females can be carried either as individuals carrying the necessary sperm, or as estab- lished inquilines in a Tetramorium colony. There is no reason why such a voyage might not be successfully made by a parasite queen, especially when one considers the evidence of Lindroth (ms., personal communication) for transport of faunal frag- ments to North America in ballast originating in Europe. Fur- thermore, there is no reason to believe that Tetramorium nests, with or without Anergates, cannot flourish on shipboard for at least the normal span of these species as colonies, a span which seems to be sufficient even for a long voyage under sail. The third objection is the least difficult one, for there ap- parently has been no shortage of suitable host nests at close proximity to the waterfront in at least some of the major western Atlantic ports, perhaps as far back as colonial times. If a colony of the host parasitized by Anergates arrived in ballast or other- wise stowed-away, it had only to release its flight of fertile female imagines on the new shore to create a high probability of successful establishment. A similar series of events may have led to the establishment of the workerless parasite Xenometra moriilicornis Emery in the West Indies, together with or following the establishment of its host, Cardiocondyla cmeryi Forel. A Xenometra of the same or a very closely allied species lives with C. clegans Emery in Italy ; Menozzi (1919) thought this was the male of elegans, but specimens from his collection indicate instead its affinity with X. monilicontis. Cardiocondyla is a primarily littoral and riparian genus from the warmer parts of the Old World ; records of several species from the New World seem to indicate rather clearly that it has been introduced by man on many separate occasions (M. R. Smith, 1944). To conclude the discussion of the bearing of parasites on the distribution of Tetramorium, 1 think we may safely consider that introduction of an obligatory parasite, while less probable than the establishment of the host, is nonetheless entirely possible if the opportunities exist for a long enough time, and if a dense host population is available to the immigrant parasite. 1957 TETKAMOHir.AI IX NORTH AMERICA 5 TeTKAMORIUM RUGIYENTRIS The type series and onl.y recorded sample oC T. rugiventris was obtained from an upland ponderosa jiine stand about ten miles south of Prescott, Arizona, and about one mile off the liijihAvay. In empliasiziny his difficulty in accepting Smith's hypothesis, namely, that the ant was introduced with camel food or stores at the time Avhen camels were imported from North Africa during the last century, Creighton wrote : ' ' Ento- mologists frequently strain at gnats but it is seldom that they are asked to swallow a camel." I have checked with care a syntype of T. rugiventris kindly sent by Dr. Smith. As a result of this examination, I can agree with Dr. Creighton that the ant in cpiestion is almost unques- tionably endemic to the locality where it Avas found. However, I cannot agree with either Smith or Creighton that the species rKgivcufris belongs to the genus Tctrcnnorium as it is now con- stituted. Instead, the type I have examined seems to me to be a cleareut, if somewhat aberrant, member of the genus Myrmica, closely allied to M. striolagaster Cole. M. striolagaster is re- corded from several localities in Arizona and New Mexico, and I have specimens collected by E. 0. Wilson at or near the type localit}- of rugiventris, in the vicinity of Prescott. Although the two species are separated hy the extent and strength of the gastric sculpturing and by other characters as well, it seems clear that they are congeneric, and also that they run rather close to the M. pnnctiventris group of Myrmica. The rugiventris type actually possesses minute barbulation on the posterior tibial spurs, as can be seen at magnifications of 90 X and better, so that in this character, the species would key to Myrmica, rather than Tetramorium, in the standard keys to tribes and genera of Formicidae. Actually, however, this spur barbulation is not worth much as a tribal character within the IMyrmicinae, despite the faith that key-makers have placed in it. The false distinction between Hylomyrma Forel (tribe Myrmicini) and Lundella (tribe Tetramoriini) appears to have been based partly on this character (Brown, 1953), and the same may hold true of the supposed difference between Crato- myrmex Santschi, purportedly a member of tribe Myrmicini, and Messor Forel, of tribe Pheidolini. 6 BREVIOBA NO. 72 Dr. Smith is fortified in his opinion that rugivcntris belongs to Tetramorium by the presence in his species of raised carini- form lateral wings of the clj^peus that border the antennal fossae in front, as in Tetramorium; but this character is shown with varying degrees of clearness in other, undoubted Myrmka species, both in North America and in southern Asia, so that it cannot be used as a point of separation between the two genera. The question comes to mind, of course, as to whether Myrmica and Tetramorium. really are separate genera after all, and this is precisely the kind of question that most needs asking in ant taxonomy these days. In deciding this particular question, fur- ther study must l)e made of the males, since tetramoriine males (with the exception of a couple of African forms that require closer study) have several funicular segments fused in such a way as to reduce the numl)er of antennal segments to ten in this sex. Unfortunately, the males of rugiventris remain unknown at present, so that it is not known whether they meet the strong criteria of this caste. From the habitus and lesser details of the worker, however, I consider the relationship with Myrmica is close enough to call for a new combination: Myrmica rugiventris. Whether or not this combination finally i)roves to be the valid one, it at least helps to establish strong doubts as to the pre- Columbian existence of Tetramorium in North America. In fact, were it not for Tetramorium lucayanum Forel and the Xiphomyrmex spinosus complex, the New World could h( considered free of endemic members of tribe Tetramoriini. With the synonymy of Lundella under Hylomyrma (Brown, 1953), the New World lost its one endemic tetramoriine genus. T. lucayanum presents no special difficulty, because its distribu- tion (Bahamas, Puerto Rico, etc.) is highly suggestive of intro- duced status. In its morphological characteristics, lucayannm seems closest to an African group of species, but the species itself has not l^een identified with any particular continental African population. However, our knowledge of African Tet- ramorium is in a very imperfect state, and it seems to me likely that lucayanum must have come from the Dark Continent, even though it may be rare there. A parallel case involving Strumi- genys rogeri Emery has turned out to follow exactly this pattern (BroAvn, 1954). 1957 TETRAMOKIUM IN NORTH AMERICA 7 The Xiphomyrmex sinnosus complex (which may represent a single variable species) is the one example that cannot be ex- l^lained away, and it is all the more remarkable, considering its isolated position in southwestern United States and through much of Mexico, far away from the remainder of the generic range, which is entirel}' Old World tropical and warm-temperate. Pre- limina.ry examination reveals no reason to consider this complex as other than bona fide Xiphomyrmex, and its distribution is almost certainly that of a long-established endemic group of populations, surely pre-Columbian in North America. Xipho- inyrmex is separated from Tetramorium by a very minor charac- ter, 11 antennal segments in place of the 12 of Tetramorium. Future revisers could well fail to be impressed by the soundness of the generic split based on this difference, so that we may eventually see a systematic technicality bring back Tetramorium as a native American genus. SUMMARY Of the five species of Tetramorium so far reported as occurring in North America, only two are under dispute as possibly having existed on this continent prior to the advent of European colon- ists ; these are the species heretofore known as T. caespitum and T. rugiventris. Evidence is presented to show that T. caespitum almost certainly was introduced by man from Europe, this evi- dence consisting primarily of the demonstration that T. caespi- tum in North America, unlike the Eurasian populations, is distributed exclusively in the manner of a man-accompanying "tramp" species. The species rugiventris, on the other hand, is removed from Tetramorium, where it does not fit well, and is transferred to Myrmica. The number of Tetramorium species occurring in North America is thus reduced to four, all of them likel}' introductions from the Old "World w^ithin historical times. The only member of tribe Tetramoriini that can safely be con- sidered as endemic to the New World at the present time is the Xiphomyrmex sj^inosus complex, widespread in southwestern U. S. and Mexico. 8 15IJEVIOKA NO. 72 KEFEKENCES CITED r>iio\v.\, \V. L., Jk. 19.").'}. C'haraetois aiul aynoiiymics among the genera of ants. Part II. Brcvioia, Mus. Comp. Zool., 18: 1-8. 1!)'>4. The ant genus Strumigcnys Fred. Smith in the Ethiopian and Malagasy regions. Bull. Mus. Comp. Zool., 112: 1-34, cf. pp. 4-7. CUEIGIITON, W. S. 1950. The ants of North America. Bull. Mus. Comp. Zool., 104: cf, pp. 241-245, 286-294. MiNCZZi (Mexozzi), C. 1919 (1918). Primo contribute alhi couoscenza della fauna mirmecologica del Modenese. Atti Soc. Nat. Mat. Modena, (5) 4: 81-88, cf. p]). 83-84. S.\iiTii, ^r. E. 1943. Ants of the genus Tetramorium in the United States with the description of a new si)eeies. Proc. Ent. Soc. Washington, 45: 1-5. 1944. Ants of the genus Cardiocondyla Emery in the United States. Proc. Ent. Soc. Washington, 46: 30-41. BREVIORA Mmseiuiini of Comparsitive Zoology Cambridge, Mass. Mak( h 29, 1957 Number 78 ADDITIONS TO THE MAMMALIAN FAUNA OF PERU AND NOTES ON SOME OTHER PERUVIAN MAMMALS By Oliver P. Pearson Museum of Vortoljiate Zoology, Berkeley, California All earlier report discussed the mammals of the highlands, or altiplaiio, of southern Peru (Pearson, 1951). Subsequent col- lecting has revealed important additions to the fauna of this area as well as considerable extensions of the range of certain species. 1 am indebted to Dr. Carl P). Koford, who collected several of the specimens mentioned, for permission to report upon them and to draw upon information in his field notes. All specimens are in the Museum of Vertebrate Zoology, Berkeley. Hesperomys sorella (Thomas). This species was not listed in the previous report on mammals of the altiplano. Eleven speci- mens have now been taken in the Department of Puno in bunch- grass terrain at 3 mi. NE Arapa, 12,600 ft., 5 mi. S Asillo. 13,000 ft. and at Hacienda Calacala, 13,000 ft., which is 7 mi. SW Putina. They are longer-tailed, longer-eared, and tawnier than H. lepidus ducilla, which also lives in this region, and have shorter, more slanting zygomatic plates. The range of measure- ments (in mm.) is: total length, 128-147; tail, 60-71; hind foot, 17-19 ; ear from notch, 18-19 ; and greatest length of skull, 21.5- 23.2. They do not match the type of sorella perfectly but seem to be more closely related to that form than to coJlosus, carillus, or frida. Since publishing the earlier report in which I used the name Hesperomys ducilla for the short-tailed species in southern Peru, I have examined the types of H. lepidus (Thomas) and H. ducilla (Thomas) and find that despite great difference in age of the in- 2 BREVIORA NO. 73 (li\i(liials iliey are (luite .siiiiilai- and can easily l)e included in the same species, characterized by short tails and tall zygomatic plates. Saiiliorn (1950^ also considered (liirtlJa to lie a subspecies of Jepidus. Klifjmodoiil l(( iJiKriilus piicruhis (Philippi). Four specimens of this desert mouse have l)een reported from Peru (Pearson, 1951), all from Santa Rosa de -luli, Department of Puno. and were listed as E. p. hirtipcs, the type locality of ^vhich is Lake Poopo in Bolivia. Our recent collections included 11 specimens taken in Peru between 13,000 and 15,300 ft. in the Departments of Moquegua, Puno, and Tacna. These specimens have now been compared Avith topotypes, in the Chicago Natural History Museum, of E. p. puendus from northern Chile and are iudis- tinguishal)le. E. p. iarnpacensis Mann has also been described from northern Chile, but the new material reveals that the shape of the zygomatic ])late and the bicolored nature of the tail, features on which farapacensis was based primarily, are un- reliable. The Peruvian Elifpiindoiil ia therefore should l)e E. p. pKcridus. FhyUotis [Galoionnjs) garlcppi Thomas. Two specimens of this rare mouse were taken in southern Peru, one at Pichupichuni. 12,600 ft., 5 mi. NW Iluacullani, and another at Pampa de Anco- marca, 13,700 ft., 76 mi. S Have, both in the Department of Puno. These are the first records of this species from Peru. Since no field measurements have been i)ublished and the type specimen has a smashed skull, we record measurements (in mm.) of our two specimens and photographs of one of the skulls (Fig. 1). Total length Tail Hind foot Ear Skull greatest length Zygomatic breadth Width of braincase jMaxillary tootlnow Compared with the type and another Bolivian specimen in the British Museum of Natural History, and one in the Chicago Natural History Museum, the Peruvian specimens are smaller, MVZ 115903 MVZ 11590 129 132 32 30 23 24 19 19 26.7 27.0 lo.O 15.0 12.8 13.0 5.3 5.2 li);")? PERUVIAN MAMMALS 3 have iimcli sliorlcf I'lir, color iniicli less l)ril)ruary 1). Tlie Peruvian s])ecinieiis auree with the Fis. 1 (;iliove;. I'liylloti.s (Galcnoviys) (jarleppi ; I'.o x. Fig. L' (below ). Fr'li'< jdcohlfd ; ().(> x. tyj)e in these i(bli'}iiis. Using Pearson's (1951) key to the rodents of the altiplano west of Lake Titicaca, Galeno- mys keys out as either Phyllotis osilae, Ph. darivini, or Akodon aruocinis. All of these have much longer tails than does garlcppi. Punomys lew minus Osgood. The range of this rare mouse has been extended considerably by its capture 55 mi. ENE Arequipa, 15,300 ft., Department of Are(|uipa, and 12 mi. NE Tarata. 14,600 ft., Department of Tacna. These new specimens agree well with those from Caccachara (Pearson, 1951) and with the lype. In addition, nine specimens were taken 8 mi. 8SW Limbani, 15,000 ft., Department of Puno. This is 130 miles north of the type locality and separated from it by the grassland of the Titicaca basin. A circuitous strip of more favorable habitat may connect the two regions by way of the mountains to the west and northwest, but no specimens have been taken there. The specimens from near Limbani, like so many forms from this more humid region of the Andes, are distinctly and consistently darker than those from farther south or west. Compared wiih the type and with the sj^ecimens mentioned in the preceding paragraph, the Limbani Punomys are greyer and darker on the back, the feet and hands duskier al)ove, ears darker, tail less distinctly bicolor, and belly considerably darker grey with a distinct bult'y wash. The Limbani specimens, as well as those from near Are(iuipa and Tarata, were captured in barren, broken I'ock areas and, as at Caccaciuira, were near fleshy-leaved, luuigent Soiccio plants or piles of Scnecio cuttings. l!);")? PERU VI AX MAMMALS 5 Covin tscJnuh'i osgoodi Sanborn. Two s])CM'i('s of guinea piprs live on the altiplano of Peru and at some localities ])robably oeciir together. Cavia (Galea) mustcloicles is diurnal (Pearson, 1951) and prefers i-ather open habitat with jiood visibility. At Ilaeienda Calaeala the mueh darker colored Cavia (Cavia) iscliudii lived in thick grass where it made distinct nuiways and was crepuscular. 8teel traps set in the runways failed to catch any during the day, but numerous individuals were seen and collected in the evening after sunset and in the early morning. Lagidinui poKanuDt Eleven. In Peru mountain viscachas have been considered to live only at high altitude. It was a surjirise, therefore, to find a small population living in the fog belt, or loma.-;, at only 2200 ft. at Nana, Department of Lima. The hilltops at Nana support scattered clumps of fog-nourished TiUandsia (Bromeliaceae). On the lower slopes there is no vege- tation, but thei-e are irrigated fields on the floor of the valle3^ Among rocks on one of the hilltops were many viscacha scats, which probably accumidate for years in the absence of rain, a few viscacha bones, and at least one living viscacha. This individual was seen at a distance of oidy 6 feet and appeared to be similar to the mountain viscacha,-; of southern Peru but with more than average buffy color. A maxillary toothrow picked up nearby agrees well with specimens of Lagidiuni pcruanum. This population living more than 6000 ft. lower than any other viscacha population known to me in Peru is isolated by several thousand feet of brushy and weedy terrain unsuitable for viscachas. Mustehi fytnaia Lichtenstein. Weasels have long been thought to live on the altiplano of Peru but have been inadequately rep- i-esented by specimens. A mounted skin is now available, taken in the spring of 1951 at Hacienda CaJ.acala, 13,000 ft., 7 mi. ^S^Y Putina, Department of Pnno. Orison (Gri^onella) cuja (Molina). The presence of this nuistelid on the altiplano of Peru has heretofore been inade- (juately documented. We now have the skin and skull of an adult male from Ilaeienda Pairumani, 13.000 ft., 24 mi. 8 Have, De- partment of Puno. JNIeasurements (in mm.) are: total length, 570; tail, 145; foot, 70; weight, 1700 grams. This specimen was shot at 4 p.m. with its stomach crammed with at least three mice and a lizard, indicating diurnal feeding. 6 BRFA'IORA XO. 78 Felis jacobita Cornalia. The complex taxoiiomie history of this wildcat (see Osgood, 1943) has been based on a half-dozen skins, many of them without adecinate locality information, and on drawings of a single skull (Pliilii)pi, 1873). Xo previous specimens are known from Peru. AVe now have a skin and skull of a male, trapped March 30, 1952, among rock outcrops at 15,500 ft., 57 mi. ENE Arequipa, Department of Arequipa. This is a barren region of rocks and bare ground with scattered c]um])s of buncligrass [Fcsiuca orihophxiUti) and small tola bushes. Other steel ti*aps nearby caught a. fox [Diisicifov ciilpdCHs) and a mountain caracara (PlialcohooiKs uicgaJopterus) . A mountain lion passed close by several times. Probable prey items in addi- tion to small rodents and small birds were mountain viscachas {Lagidiuni), tinamous (Tinamotis), and seed snipe {Attagis. Thinocorus) . Vicunas, the young of which might be killed by jncohito, were abundant. The skin matches well the vai-ious published descriptions, but several features of the skull, some of them previously considered to be diagnostic, do not matcii the illustrations in Philippi's rei)ort. The audital bullae of our specimen are not bisected by a deep sulcus; in fact, scarcely a trace of the sulcus shows. Com- pared to four Felis pajeros from southern Peru and northern Chile, the nasals of our jacobita are only slightly larger and the incisors not more jn-oodont. The skidl of our jacobita is slightly flatter and more i-obust than in pajvros, and has lai-ger teeth, but skulls of the two s])ecies are not as different in ai)pearance as are those shown in Philipi)i's illustrations. A possible diag- nostic cliaracter is the orientation of the upper premolars, in pajeros these two teeth on each side, when vie\\t'(l from l)elow. lie in a straight line. In jacobita the anterior premolar toes in abruptly in front. Since no other skulls seem to bi' available and since our specimen differs considerably from the one seen by Philippi, two photographs are presented in Figure 2. Measui-c- ments (in mm.) are: total length, 990; tail, 413; foot. 133; ear, (i3 ; Wright, 4.0 kg. ; condylobasal length, 9(i ; greatest length of skull, 100; zygomatic lu-eadth, (i!>.5 : greatest length of upi)er carnassial, 14.0. lO;")? I'KRirVIAX MAMMAliS 7 'I'lu' ]»t'la;:t' is Hut'tici- and jLii-cyci- than that of /•'. pujcros and is without a spinal crest of loiij^- hairs; tlic ears are rounded without tufts, and the tail is non-taperino-, much louf>er, anOI>. \V. il. 1943. The iiwmminls of Cliilc. Fiold Mus. Nat. Hist., Zool. Ser., 30: 1-268. I'EARSOX, U. P. li>.il. ^[;llnln;lls in tin- liigliliiTxls of southern Pern. Bull. Mus. Comp. Zool., 106: 117-174. rmi.ippi, R. A. 1S7.S. Ueber Felis Ciuina Molina and iiber die iScliJidelbildung bei Felis Pa.ieios und Felis Oolocolo. Areh. Xaturg., 39 (1): 8-15, pis. 2-3. Sanhokx. 0. C. 1 !».■)(•. Small rodents from Peru and Bolivia. Pul). Mus. Hist. Nat. "Javier Prado," Ser. A. Zool., No. 5: 1-16. BREVIORA MeseiJim of Comparative Zoology Cambridge, Mass. May 1, 1957 Number 74 THE DISCOVERY OF CERAPACHYINE ANTS ON NEW CALEDONIA, WITH THE DESCRIPTION OF NEW SPECIES OF PHYRACACES AND 8PHINCT0MYRMEX By E. 0. Wilson Biological Laboratories, Harvard University During December, 1954, and January, 1955, the author was fortunate in being able to spend an uninterrupted five-week period on New Caledonia, studying the ant fauna of the island.^ During this time three species of Phyracaces and Sphinctomyr- mex were collected, all undescribed and representing the first cerapachyine ants ever found on New Caledonia. Their presence is of some zoogeographic importance, in that they appear to ally the ant fauna of New Caledonia more closely with that of eastern Australia, as opposed to the remainder of Melanesia. Phyracaces and Sphinctomynnex are strongly developed both in species and in individual numbers in Australia, but are known from only three rather rare species (two Phyracaces and one Sphincto- myrmex) on New Guinea, and are unknown from the rest of northern and central Melanesia. In an obverse relationship, Cerapachys {s. sir.), is the predominant cerapachyine genus on New Guinea and the Fijis, but has never been collected in either Australia or New Caledonia. Finally, the New Caledonian Phyra- caces and Sphinctomyrmex are most closely allied to eastern Australian species, as indicated in the descriptions to follow. Phyracaces cohici Wilson, new species Diagnusis. A shining, black, medium-sized species closely re- lated to the turneri group of species of eastern Australia. P. 1 Field research was supported by grants from the Society of Fellows, Harvard University, and the Museum of Comparative Zoology. 2 BREVIORA No. 74 cohici can be easily distinguished from the latter group, which includes turneri Forel, adamus Forel, and larvatus Wheeler, by its more obtuse and rounded dorsal propodeal corners. In cohici these corners form an angle of more than 110° when viewed from the side, while in the turneri group of species they form an angle of 90° or less. P. cohici also bears a fair resemblance to the Australian P. senescens Wheeler, but can be separated from this species by its longer, flatter petiolar node and straighter and more horizontally aligned posterior petiolar teeth. Holotype worker. Head width ^ 0.99 mm, head length 1.05 mm, scape length 0.63 mm, cephalic index 94, scape index 64, exposed length of mandibles 0.21 mm, eye length 0.26 mm, pronotal width 0.79 mm, petiole width 0.83 mm, petiole length (measured from the midpoint of the anterior border of the node to the midpoint of the posterior border of the posterior peduncle) 0.63 mm, postpetiole width 0.83 mm, postpetiole length 0.67 mm, width of next gastric segment 1.00 mm. Occipital border very feebly convex in full-face view. Alitrunk viewed from above moderately constricted medially, marginate along the entire dorsolateral border except for an interval of about 0.20 mm in the region of the mesothorax. Petiolar node viewed from directly above with strongly concave anterior border, and evenly convex lateral borders, its widest point being about in the middle (see accompanying figure). In the same view the posterolateral teeth extend well beyond the posteriormost point of the posterior node border. Seen from directly above, the postpetiole is widest in the anterior half and is laterally marginate only in the anterior half. Entire body covered by scattered piligerous punctures spaced on an average of 0.03 to 0.06 mm apart, the interspaces com- pletely smooth and shining. On the sides of the alitrunk the punctures are unusually small, being barely visible at 40X magnification. 1 Head width and other body measurements were made in the standard fashion employed by W. L. I'.rown in his published serial studies of the daeetine ants and by myself in my recent revision of the ant genus Lasiug (1955, Bull. Mus. Comp. Zool. Ilarv., 113: 1-200). In the past, Brown and I have differed slightly in our cleflnitiuus of head length, but in the present paper I have decided for the purposes of uniformity to follow Brown's original definition, which was given as follows : "critical maximum length of the head, measuring from a transverse through the posteriormost point or points along the occipital border to a trans- verse through the anteriormost point or points on the anterior clypeal border" (1953, Amer. Midi. Nat., 50: 11). With this change I believe that our standard- ized measurements now all agree in every detail. 1957 CERAPACllYINE ANTS OF NEW CALEDONIA 3 Entire body deep blackish browu to jet black, except for the mandibles and gastric apex, which are dark reddish brown. Ap- })endag'es variably dark reddish brown. Worker variation. Maximum head width of tj^pe series (all shown by a single colony, ace. no. 190) : 0.88-1.01 mm. The worker tj^pe series shows very little variation in other external characters. MaJe. Head width (across and including compound eyes) 1.10-1.14 mm. Antenna 13-segmented. Mandible well developed, narrowly subtriangular, its masticatory border bearing a large, blunt apical tooth followed basally by an indeterminate num- ber of serial denticulae ; the masticatory and basal borders join- ing through an even, convex curve. Clypeus narrow, its dorsal surface gently convex, its anterior border seen from directly above moderately and evenly convex. Parapsidal furrows well developed, parallel with each other, and extending anteriorly for about half the length of the scutum. Notaulices absent. Median notal suture present and extending posteriorly for slightly more than one-third the length of the scutum. Wing venation generalized, essentially similar to that of Cerapacliys (see Brown and Nutting, 1950, Trans. Amer. Ent. Soc, 75: 132, pi. VIII), differing primarily as follows: Rsf2'3 is lacking, and the second radial crossvein and Bsf5 form a single unit coming off the stigma. The crossvein " x" indicated by Brown and Nutting in the basal cubital-anal region of Cerapacliys manni is missing in Fhyracaces cohici. Petiole completeh^ lacking the lateral margination that char- acterizes the Fhyracaces worker caste. Seen from above, the node is widest in its anterior half, and its anterior and posterior corners are gently rounded ; seen from the side, the node forms a single, even, strong convexity only weakly demarcated from the peduncles. Genitalia completely retractile. Subgenital plate relatively large (exposed length about 0.35 mm), sclerotized, tapering posteriorly to end in a pair of sharp, upward curving hooks each about 0.06 mm in straight length. Parameres short, broadly rounded apically. Pilosity, sculpturing, and color essentially as in the worker. Types. Described from a long series of workers and four males collected by the author at Ciu, near Canala, at 300 meters, and from two kilometers southwest of Ciu, at 500 meters, New Cale- 4 BREVIORA No. 74 donia. The following accessions are included : 190, 246, 263 (holotype nest series), 267, 275, 278, 298. Each represents a separate nest series with the exception of numbers 246 and 275, which are from the same nest. All of the collections were made in the period from December 21, 1954, to January 2, 1955. This species is named for Mr. Francois Cohic, of the Institut Frangais d 'Oceanic, an able and enthusiastic student of New Caledonian entomology. Ecological notes. All of the collections were made in the rich subtropical evergreen forest clothing the hills that extend from the south bank of the Canala River in the vicinity of the Ciu Falls. The holotype colony was found under a rock in a densely shaded part of the forest, and was occupying a single cavity and adjacent short vertical gallery in the soil. The males were very active and attempted to fly when the nest was exposed. Another colony (ace. no. 246-275) was nesting in open soil at the side of one of the forest trails. The nest entrance consisted of a single five-millimeter-wide opening surrounded by a low, indistinct turret of excavated earth. Lateral excavation revealed three or four galleries leading down from the entrance hole and into soil packed between several buried rocks. At about ten centimeters down two small adjacent chambers had been excavated in the soil against the vertical face of one of the rocks. In these were massed all of the brood and most of the workers. The following rough population estimate was made at the time of collection : 80-100 workers, 40 pupae (in cocoons), 30-40 half -grown to fully grown larvae, and 30 eggs. The reproductive of this colony was unfortunately not found. Workers belonging to colony no. 246-275 and other colonies were encountered on several occasions foraging during the day, always in file, over the surface of the ground, and on one occa- sion (no. 190) workers were discovered in the upper layers of a moist rotting log. Twice, workers were observed in the act of raiding colonies of the ant genus Pheidole. The foraging and raiding behavior of this species will be described in greater detail in a later paper on the general subject of cerapachyine behavior. 1957 CERAPACIIYINE ANTS OF NEW CALEDONIA 5 Phyracaces dumbletoni Wilson, new species Diagnosis. Distinguished from all other species of the genns known to me by the following combination of characters : mod- erately large size (head width of type series 1.16-1.25 mm), antennal scapes unusually long, sides of alitrunk non-marginate, body surface smooth and shining to shagreened and subopaque, body color deep blackish brown to jet black. The only other Indo- Australian Phyracaces lacking margina- tions along the sides of the alitrunk are aherrans Clark and pygmaeus Clark of northern Queensland and hewitti "Wheeler of Borneo. These species are easily distinguished from dumhletoni by their much shorter scapes, which reach only to about the posterior margins of the compound eyes when the head is viewed in full face — in dumbletoni the scapes extend beyond the eyes by about their own maximum width. Of the three species, dumbletoni most resembles aberrans Clark, being very close in both size and sculpturing. Holotype ivorker. Head width 1.17 mm, head length 1.30 mm, scape length 1.08 mm, cephalic index 90, scape index 92, exposed length of mandibles 0.29 mm, eye length 0.33 mm, pronotal width 1.03 mm, petiole width 0.95 mm, petiole length 0.76 mm, post- petiole width 1.09 mm, width of next gastric segment 1.26 mm. Occipital border in full-face view very feebly convex. Alitrunk viewed from above only feebly constricted medially, its dorso- lateral area evenly rounded and lacking any trace of margina- tion. Petiolar node seen from directly above with moderately concave anterior border and evenly convex lateral borders, its greatest width located in the anterior half. Puncturation as described for P. cohici, except that anterior to the postpetiole the punctures are more scattered, the majority being 0.09 to 0.12 mm apart. On the postpetiole and anterior to this segment the interspaces are for the most part "shag- reened," the shagreening in this case in reality consisting of regular, minute, contiguous foveolae each about 0.01 mm in diameter. The foveolae are deepest on the dorsal surface of the head and alitrunk, and render most of the cuticular surface there subopaque. They are absent posterior to the postpetiole. Color as described for P. cohici. BREVIORA No. 74 Worker variation. Maximum head width range, Chapeaii Gendarme and Mt. Mou series: 1.16-1.25 mm (all shown by ace. no. 65, the holotype nest series). Maximum head width range, Ciu series: 1.11-1.20 mm (all shown by ace. no. 245). The Ciu workers differ significantly from the Chapeau Gendarme and Mt. Mou workers in the intensity of the f oveolar ' ' shagreening. ' ' The latter approach the condition described in the holotype, with little deviation. The Ciu specimens have the same basic form of sculpturing, but on the head and alitruncal dorsum the foveolae are much shallower, so that the surface is strongly shining under ordinary reflected light. cohici dumbletoni Dorsal view of the petioles of the two New Caledonian species of Phyracaccs ; outlines drawn to scale from the holotyjoes. Ergatoyyne. Head width 1.14 mm, head length 1.26 mm, scape length 0.76 mm, cephalic index 90, scape index 67, exposed length of mandibles 0.22 mm, eye length 0.30 mm, pronotal width 1.00 mm, petiole width 0.95 mm, petiole length 0.62 mm, postpetiole width 1.32 mm, width of next gastric segment 1.55 mm. Ocelli lacking. More similar to the worker caste than to the queen caste of other Phyracaces species, differing primarily in the following external features. (1) The alitrunk is very worker-like, apparently differing only in the somewhat stronger pleural suturation. The posterior metapleural suture, which is absent in the dumbletoni Avorker, is present although feebly de- veloped in the ergatogyne. (2) The petiole is relatively shorter 1957 CERAPACHYINE ANTS OF NEW CALEDONIA 7 in the ergatogyne. (3) The postpetiole and gaster are much more massive and more poorly demarcated from each other than are the same structures in the worker. Types. Described from a long series of workers from Chapeau Gendarme (Yahoue), Mt. Mou, and Ciu, and a single ergatogyne from Ciu. The following accessions are included, each repre- senting a separate nest series: Chapeau Gendarme, no. 65 (holo- tj'pe nest series); Mt. Mou, single worker; Ciu, no. 245 and ' ' observation colony. ' ' This species is named after Mr. L. J. Dumbleton, formerly economic entomologist for the South Pacific Commission, and an expert student of the New Caledonian insect fauna. Ecological notes. P. dumhletoni was collected at Chapeau Gendarme and Mt. Mou in dry, semi-deciduous "valley-pocket" forest and at Ciu in moist broadleaf evergreen forest. In all three localities it was limited primarily to the least disturbed portions of the forest, and was never encountered in the adjacent open Melaleuca woodland. At Chapeau Gendarme a large colony (no. 65) was found nesting in several spacious galleries and chambers in the upper layers of a large, moist, fern-covered log. It contained at least 200 workers, a single ergatogyne (later lost), over 100 cocoons and larger larvae, and an undetermined number of eggs. Among the brood were found the hollowed-out propodeum of a worker of an undescribed species of Lordomyrma and the mangled remains of the larva of an undetermined ant genus. These insects ap- peared to be the prey of the Phyracaces, but of course this can- not be proven. When first disturbed, many of the Phyracaces workers sallied out with a display of aggressiveness unusual for cerapachyines, and one succeeded in stinging me on the forearm. I think it is worth mentioning that this is the only time I have ever been stung by a cerapachyine ant, despite the fact that I have ex- cavated many nests of Phyracaces and other genera without mak- ing any effort to protect myself from the workers. The sting caused a prominent welt about six millimeters wide surrounded by an erythema about twenty-five millimeters wide. There was a persistent, dull, throbbing pain of the sort commonly resulting from the stings of ponerine ants. 8 BREVIOBA No. 74 Sphinctomyrmex caledonicus Wilson, new species Diagnosis. Closely resembling S. steinheili Forel of eastern Australia, from which it can be distinguished by the following two characters. (1) Sculpturing, which is dominated by punc- turation, is overall denser in caledonicus. In steinheili, the inter- space distances over the genal surface are mainly as great as the adjacent puncture diameters or greater, while in caledonicus the genal punctures are contiguous and their borders form an even rugoreticulum. In steinheili, the lateral surfaces of the alitrunk are in large part feebly shining, while in caledonicus these surfaces are entirely opaque. (2) In side view, the dorsal posterior corners of the propodeum form a distinct angle of about 110° in steinheili, but are evenly rounded in caledonicus. Holotype worker. Head width 0.56 mm, head length 0.71 mm, scape length 0.41 mm, cephalic index 79, scape index 73, exposed length of mandibles 0.06 mm, pronotal width 0.42 mm, petiole width 0.35 mm, petiole length (including peduncles) 0.38 mm, postpetiole width 0.47 mm, w^idth of next gastric segment 0.60 mm, length of gaster posterior to postpetiole (measured in a straight line) 1.35 mm. Worker variation. Maximum head width range (internidal) 0.50-0.57 mm; (intranidal; aec. no. 195) 0.50-0.56 mm. In size as well as other external characters the worker type series is remarkably uniform. Ergatogyne. Head width 0.62 mm, head length 0.76 mm, scape length 0.42 mm, cephalic index 82, scape index 68, exposed length of mandibles 0.09 mm, pronotal width 0.44 mm, petiole width 0.42 mm, petiole length 0.41 mm, postpetiole width 0.60 mm, length of gaster posterior to postpetiole 1.96 mm. This specimen is very worker-like, and can be distinguished only by its slightly larger size, proportionately shorter head and scapes, broader petiole, and larger postpetiole and gaster. In addition the postpetiolar-gastric constriction is somewhat weaker than in the Avorker. Compound eyes and ocelli are completely lacking, as in the worker, and the structure of the alitrunk is essentiallj- the same. Types. Described from a long series of workers and a single ergatogyne from Ciu, 300 meters, New Caledonia. The following two accessions, representing separate nest series, are included : 1957 CERAPACHYINE ANTS OF NEW CALEDONIA 9 no. 195, no. 225. The holotype was chosen from no. 225. Ecological notes. The two type colonies were found in a small, isolated patch of broadleaf evergreen forest on the farm of Mr. D. Fere, about half a kilometer northwest of the Ciu Falls. This little woodlot did not exceed two acres in extent, and its floor had been badly disturbed by cattle. Most of the ants present, including the Sphinctomyrmex, were found underneath rocks embedded in the soil. It is curious that the Sphinctomyr- mex was not found in the undisturbed forest on the south bank of the Canala River a short distance away, despite intensive collecting there. Both colonies were quite large, one containing over 500 work- ers and the other over a thousand. That the colonies may have been in migration is suggested by the fact that they occupied poorly defined galleries in the soil which bore no sign of lengthy occupation. The brood of colony no. 195, collected on December 21, consisted of large numbers of mature larvae ; two days later, about three-quarters of a sample of these larvae kept alive in a bottle had spun cocoons. The brood of colony no. 229, collected on December 31, consisted of large numbers of cocoons, all of which contained prepupae of indeterminate caste. These data suggest a high degree of synchronization of brood development, a condition usually associated in ants with a nomadic way of life. An account of the behavior of the w^orkers of this species will be given in a later paper. BREVIORA Museiuoi of Comparsitive Zoology Cambridge, Mass. May 1, 1957 Number 75 ON A NEW OCTOCHAETINE EARTHWORM SUPPOSEDLY FROM GUATEMALA By G. E. Gates INTRODUCTION Tweuty-four species of earthworms have beeu reported from Guatemala. One of them is widely distributed in North America where it presumably originated. Another, "pantropical," may have come from the south. Four are obviously exotic, probably introduced accidentally by man since 1500 A.D. The eighteen supposed endemics are known only from the original material. As a result of one of those accidents that occasionally advance our knowledge of an infrequently collected group of animals, another presumed endemic now can be added to the list of species known only from a single specimen. The author's thanks are extended to Dr. P. W. Oman for providing the present specimen and for assistance rendered in connection with problems arising from the study of various other species. Family MEGASCOLECIDAE Subfamily OCTOCHAETINAE Genus RamiELLA Stephenson 1921 RaMIELLA AMERICANA n. Sp. Guatemala. In soil with Ceiba pentandra in cargo on plane arriving at San Pedro, November 25, 1955, 0-0-1. U. S. Bureau of Plant Quarantine. 2 BREVioRA No. 75 External characteristics. Length, 133 mm. Diameter, 8 mm. (through clitellum). Segments, 217 (+? possibly a recent posterior amputee). Body slightly compressed dorsoventrally and transversely elliptical in cross section behind clitellum. Un- pigmented (formalin preservation). Prostomium, prolobous. First two segments with numerous fine longitudinal grooves in both ventrum and dorsum. Secondary annulation, one presetal furrow each on iv-xii, one postsetal each on vi-xii, first presetal furrows deep but from ix less marked and like the postsetal; secondary furrows lacking behind the clitellum. Setae small, closely paired, only very slightly or not at all protuberant from epidermis, unrecognizable (lacking?) anteriorlj^; ventral couples first visible on vi, the lateral on xi, CD<^AB, J-Aonsisting of narrow lateral liars (of which at least the first liirce to eight fuse to form narrow Vs on the nape), extends the entire length of snake from the head on to the tail; ventrals more than 229 in males, more than 236 in females D. m. medici. 2. Pattern, if present, restricted to top of head (see Fig. 1) ; ventrals 229 or less in males, less than 233 in females D. m. lamucnsis. Dasypeltis MEDICI MEDICI (Bianconl), 1859 Dip.su.s Medici Bianeoni, 1859, p. 277. No locality designated, luit Mozambique b.v inference. Type in Bologna Museum. Dasiipeltis scaher var. fasciolata Peters, 1868, p. 451. Type (ZMU 5737) from " Angeblich aus Zanzibar". Dasypeltis elongata Moequard, 1888, p. 131. Type locality: Zanzibar. Type in Museum d 'Histoire Naturelle, Paris. Diagnosis. Three to eight narrow Vs commencino- on the nape, followed by a series of narrow, lateral, dark red-brown bars. These bars wall encircle a pink to silvery-white vertebral dot, if they coalesce with their fellows from the opposite flank. In some specimens the posterior ])ands are situated betw^een more or less clearly expressed oval dorsal saddles, while various intermediate patterns (see Fig-. 2) have also been observed. The dorsum is a light reddish-brown, which shows a considerable amount of mottling under the binocular microscope. The apical scale pits, as well as those of the head region, are a dark brow-n, this pig- mentation being found on all scales and very sharply set off. The ventrum is a clear pink, more or less regularly stippled with grey. Beneath the tail of many specimens, particularly^ in males, this stippling is arranged in two to four lines. In other individuals, hoAvever, the distribution of the stippling appears to be haphazard. Three lateral scale rows are definitely reduced and inclined, and they, as well as some of the adjacent rows, generally have strongly serrated keels. The frontals shoAv pitting or pustulosity over their entire surface as do most of the other cephalic scales (see Fig. 1). The inter-prefrontal suture is not BREVIORA NO. 79 depressed. Veiitrals in males 235-253, in females 237-259 ; eau- dals 82-109 and 71-80, respectively. Range. Coastal East Africa. Extreme southern Kenya, Tan- ganyika, northern and central INIozambiqnc ; inland to Xyasaland. Figure 2. Dasypeltis m. medici. Dorsal views at midbody showing color pattern variants of this race. Top. BM 12-1-30-6 from Matemo, Mozam- bique. Bottom. BM 97-6-9106 from between Nkata Bay and Ruarwe, Xyasaland. Note lighter middorsal spots between the saddles, the regular pigmentation of the apical scale pits, and the general speckling of the ground color. 1 ;).■)( DASYFELTIS MEDICI LAMUENS18 5 Disfribidio)! records. (Maj) 1 shows Die relation of these localities to eaeh other. Records from the literature are starred.) KENYA: Takannon. Between Voi and Ndi, Taita. TANGAN- YIKA: Zanzibar Coast. Kibonp:oto' Kilimanjaro ( Intergrade"? specimen^. Amani, Usamljara Mts. Usambara. Makindn Kiver. Morogoro. Nyange, lUugurn Mts. Northern Ukutu (Kuthu) Steppe. Tendaguru. Mikindani. Nchingidi. Liwale. ZANZI- BAR. JMAFIA. MOZAMBIQUE: Matemo. Ribaue, Nyassa Prov. NYA8ALAND: Between Nkata Bay and Rnarwe. Zomba. Cholo Mtn. DaSYPELTIS MEDICI LAMUENSIS Subsp. IIOV. Uasyiniiis pahnarum I'etors (not of Loach), 1878, p. 20G. 1 ex. Taita, Kenya (J. M. Hildebrandt) : ZMU. DasypcUis scahra var. F (part), Boulenger, 189-4, p. 356. 1 ex. Mt. Kili- manjaro, Tanganyika (F. J. Jackson) : BM. Dasyprltis scahrr Uthnioller (part), 1934, p. 113. 1 ex. nr. Goniberi, Kili- manjaro, Tanganyika (Uthmoller) : Zool. Staatssanimlung, Miinchen. DasypcUis scahcr Loveridge (part), 1936, p. 256. 1 ex. Mt. Mbololo; 2 ex. Lamu Island, Kenya (Loveridge) : MCZ. DasypcUis scahcr Scortecci (part), 1939, p. 276. 1 ex. Belet Aniin, Ital. Somaliland = Somalia (S. Patrizi) : Genoa Civ. Mus. Stated to be nnifonn grey dorsally. DasypcUis scahcr mcdici Loveridge (part), 1942, p. 283. Mention of uni- formly colored northeastern material : MCZ. Diagnosis. Uniform reddish-brown dorsally, fading to butf in some specimens after preservation (Loveridge). Pink ventrally with a fine speckle of a slightly darker pinkish-brown denser on the sides (Fig. 4, bottom). Some specimens are a uniform olive grey, fading to a plumbeous grey after preservation. These have a light grey ventrum, minutely flecked with white. A vague or distinct mottling of the ground color may be seen under the binocular microscope (see Fig. 4, top). Apical pits of body and marginal pits of head scales distinctly pigmented with a darker reddish-brown (see Fig. 1). With proper illumination faint to clear V-shaped markings can be discerned on the parietal scales of some specimens. Three lateral scale rows are distinctly reduced and inclined, and they, as well as several of the adjacent rows, 1 The status of this specimen is discussed helow and its data are not included in the above ranges tor tliis subspecies. BREVIORA NO. 79 Figure 3 1957 i\\s\i'i;i/ris medici hAMUENSis 7 generally have strongly serrated keels. Each frontal scale shows pnstulosity or pitting aroniul tlie ])(M'ipluM'v and in the center. The suture lietwcen the pi'efroutals is not depressed. V^nitrals in males 226-22!), in females 22G-232; caudals 84-94 and 72-84, respectively. Holoti/pc. jMuseuni of Comparative Zoology No. 40582, an adult male from Lanui Island, Ken\a, collected by Arthur Lovei-- idge 12 May 1934. Allotype. Museum of Comparative Zoology No. 40588, an adult female collected with the type. Paratypes. British Museum (Natural History) Nos. 51-1-3-72 and 51-1-3-73 from Kilifi, north of Mombasa, Kenya. Also Brit- ish Museum (Natural History) No. 98-1-8-15 from Maungu, near Voi, Kenya ; iMuseum of Comparative Zoology No. 40580 from Mt. Mbololo, Kenya ; Zoologisches Museum der Universitat (Berlin) No. 9244 from Taita (region), Kenya; and British Museum (Natural History) No. 87-11-3-31 from Mt. Kilimanjaro, Tanganyika. Description. Both the Lamu holotype and allotype are a uni- form brownish-red dorsall}^ and a speckled pink on the venter. Under proper illumination a very faint, posteriorly-directed V may be discerned on the parietal scales (see Fig. 3, top). Scale pits on head and body with sharply defined darker brown pig- mentation, though this is only slightly darker than the dorsal color (Fig. 3, bottom) and not as clearly visible as on lighter specimens (Fig. 4, middle). The frontals are pitted around their margins and pustules are also present in their centers, with the other head shields showing a similar pattern. The suture between the prefrontals is not depressed, the ocular-temporal formula is l+2-(-2-|-3, and the upper labials and eye contact are 7(34). Counts of ventrals are 226 for both type and allotype, caudals are 94 and 84, dorsals 23 and 24 at midbody respectively. The third, fourth and fifth rows of dorsal scales from each side are reduced, inclined and serrated, and the keels of the dorsal Figure 3. Dasypellis m. hnniiensis. Views of the holotype (MCZ 40582). Top. Dorsal view of head. Middle. Lateral view of head. Bottom. Dorso- lateral view of specimen in the midbody region. Note the very faint pattern on the parietals, and the fact that the prefrontals, oculars, anterior labials, etc. are covered with pigmented pits. 8 BREVIORA NO. 79 Figiiie 4 1957 DASYPELTIS MEDICI LAMUENSIS 9 scales are serrated around the body in the anal ref^ion of the type, hut not of the pai-atype. Body and tail lengths are 498+ 132 nun. and 570-|-128 nnn. respectively . Range. Coastal East Africa from Somalia soutli to the Kenya- Tanganyika frontier. Disiribufion records. SOMALIA: Belet Amin (Scortecci, hm). KENYA: Lamu Island (Loveridge, WSd) ; MCZ 40582, 40583. Kilifi, north of Mombasa. BM 51-1-3-72, 51-1-3-73. Mauno-u near Voi. BM 98-1-8-15. Mt. Mbololo (Loveridge, 1936); MCZ 40580. Taita (Peters, 1878); ZMU 9244. TAN- GANYIKA: Mt. Kilimanjaro (Bouleno-er, 1894) ; BM 87-11-3-31. Near Gomberi, Mt. Kilimanjaro (Uthmoller, 1934). Discussio)t. The map shows the relative position of collectino' localities for both patterned and unicolored forms of D. medici. The species appears to be restricted to the coastal regions below 1000 meters, generally characterized by their reddish laterite soils (Loveridge, 1942) . As may be seen from the map, all but two of the Kenya specimens are definitely unicolored. Pattern is present in the Voi-Ndi juvenile, but the Takaungu individual is an almost completely faded adult, which lacks even the pigmenta- tion of the apical pits so characteristic of the species. Two of the three specimens from the vicinity of Mt. Kilimanjaro (seen, or reported on in the literature) were unicolored, while the dorsal pattern of the third is faded out posteriorly. All other speci- mens and records from Tanganyika have, or are said to have, a color pattern. Of the other characters examined only ventral counts show elinal variation. Figure 5 demonstrates these counts for male and female specimens of the two color phases, showing the sharply Figure 4. Dasypeltis ?h. lamucnsis. Views of the midbody region of different specimens to demonstrate certain characteristics of the pigmenta- tion. Top. BM 51-1-3-73 from Kilifi, Kenya. Lateral view of dark grey specimen, showing the extensive speckling of the ground color, which almost masks the pigmentation of the apical pits. Middle. BM 98-1 8-15 from Maungu, Kenya. Lateral view of light colored specimen which clearly shows both pigmentation of apical pits and lateral scale arrangement. Note that the ground color is still speckled. Bottom. BM 51-1-3-72 from Kilifi, Kenya. Ventrolateral view of specimen showing color invasion and speckling of light colored ventrum. 10 BREVIORA xo. 7!) LAMU ISLAND MT. KILIMANJARO Q^ CY^ <^ I ® Okilifi VaitaO Y"^ka"ngu USAMBARA ^ AMANI MAKLNDU^ MOROGORO 0 NYANGE0 UKUTU* ZANZIBAR ISLAND MAFIA ISLAND KILWA ( TENDAGURU( LI WALE # NCHINGIDI V •MIKINDANI Jjasi/pt'ltis medici. Map showing loc-alities for all specimens acliiaUy examine.! as well as for the literature record from Belet Amin. Solid dots stand for records of m. medici, open circles for records of m. lamuoi.si.s. Divided circles refer to the paiis of immediately adjacent records discussed in the text. 195'; DASYPELTIS MEDICI LAMUENSIS 11 distinct ranfjcs. The only exception is furnished by the faintly patterned specimen from Kibonjroto' Kilimanjaro. This has a ventral count of 218, i.e. 8 less than the lowest count for a unicolored specimen and 19 less than the lowest count for a patterned specimen. In view of the fact that Ihe two populations replace each other jreographically, that they ditfer in two characters which break in the same zone, and that there exists a possibly intermediate speci- men from a geographically intermediate region, they are here considered to be subspecifically distinct. This decision has been MALES lamuensis me FEMALES lamuensis medic 220 230 240 260 V E N T R A L S Figure 5. Dasypcltis vtcdici. Graph showing ventral counts of all speci- mens of the two races actually examined. Note the anomalous count of the Kibongoto specimen. 12 BREVIORA NO. 79 consciously influenced by a desire to focus attention upon this population in the hope that this will permit re-examination of the matter on the basis of more nearly adequate collections. It has been suf>'0'ested that this situation represents two adja- cent species and that the patterned Kilimanjaro specimen is a hybrid. While the material is insufficient to decide the point, there is some evidence against this idea. Thus both of these color patterns feature a sharply defined pigmentation of the apical scale pits and are the only forms within the genus that have this characteristic. Besides which, 45 out of 46 specimens of the combined sample have frontals that are entirely, rather than marginally, pustulated or pitted, a frequency not approached by any other population, with the exception of Dasypcltis scahra from South Africa. In view of these facts and in the absence of more detailed data, the two forms are considered to belong to a single polytypic species. LITERATURE CITED BiANCONi, J. Joseph 1850- Specimma Zoologica Mosambicana. Bononiae, fasc. 1-1(5, ii 1862. +406(?) pp., (fasc. 14 (1859), pp. 277-278, tab. 14). First published in parts in Mem. R. Accad. Sci. Inst. Bologna. (Vol. 10, pp. 501-2, pi. 26.) BOULENGEK, GEORGE ALBERT 1894. Catalogue of the snakes in the British Museum (Natural His- tory). Volume II, containing the conclusion of the Colubridae Aglyphae. London, xi -[-382 pp. (353-357). LovERiDGE, Arthur 1936. Scientific results of an expedition to rain forest regions in Eastern Africa. V. Reptiles. Bull. Mus. Conip. Zool., vol. 79, no. 5, pp. 209-337 (256-57, pi. 4, fig. 1). 1942. Scientific results of a fourth expedition to forested areas in East and Central Africa. IV. Reptiles. P>ull. Mus. Com]). Zool., vol. 91, no. 4, pp. 237-373 (282-86). MocQu.vRD, Francois 1888. Sur une collection de reptiles et de batraciens rapportes des Pays Comalis et de Zanzibar par M. G. Revoil. Mem. pui)!. Soc. Philomat. Occas. Centcnnaire Foundation, i)p. 109-134 (131, plate XII, figs. 2, 2a, 2b, 2e). 1957 DASYPELTIS MEDICI LAMUENSIS 13 Peters, Wilhelm Carl Hartwig 1868. liber eine neuc Nagergattung, Chiropoilowi/.s prninllatus, so wie iiber ciiiigo neuc odcr weniger bekannto Ainphibien und Fisehe. Monatsbcr. kgl. Akad. Wiss. Berlin, pp. 448-461 (451). 1878. Ubcr die von Hrn. J. M. Hildebrandt wiihrend seiner letzten ostafrikaniscluMi Reise gesanimelten Saugetliiere und Aniphibien. Mouatsber. konig. Preuss. Akad. Wiss. Berlin 1878, pp. 194-209 (206). ScoKTEoci, Giuseppe 1939. Spedizione zoologica del Marchese Saverio Patrizi nel Basso Giuba e nell 'Oltregiuba. Giugno-Agosto 1934 XII. Eettili. Ofidi. Ann. Mus. Civ. Storia Nat. Genova, vol. 58, pp. 263-291 (276). Uthmoller, Wolfgang 1934. Beitrag zur Kenntniss der Schlangenfauna des Kilimandscharo. (Tanganyika Territory, ehenialiges Deutsch-Ostafrika). Teni niinckia, Leiden, vol. 2, pp. 97-134 (113-115). BREVIORA Mmseiuiinti of Comnpsirative Zoology Cambridge, Mass. September 30, 1957 Number 80 A COLLECTION OF DRAWINGS OF FISHES ASCRIBED TO J. P. KIRTLAND (1793-1877), IN THE LIBRARY OF BOWDOIN COLLEGE By James M. Moulton Department of Biology Bowdoin College, Brunswick, Maine During- the latter part of 1956, a pamphlet box removed from the shelves of the Bowdoin College Library was found to contain 61 uncatalogued drawings of fishes, nearly all bearing the sig- nature 'M. P. Kirtland, del." Of the drawing^s, all of which are in good condition, 57 appear to be originals of various of the illustrations accompanying a series of articles by Jared Potter Kirtland on the fishes of Ohio, which appeared in the Boston Journal of Natural History from 1840 to 1846. Most of the draw- ings are in pencil ; a few are finished in watereolors ; all appear to have been drawn by the same hand. Of the 69 figures accom- panying Kirtland 's articles, corresponding: drawings are lacking for but 11 which constituted 4 of the 26 plates and 2 of the 3 figures of a fifth plate. Various bits of evidence indicate that these drawings were executed b}^ Kirtland for his article. How might they have come to Bowdoin College? Probably through Dr. D. H.' Storer. David Humphreys Storer (1804-1891) and Jaretl Potter Kirt- land had much in common : New Englanders by birth, educated for the medical profession, eminent teachers and practitioners of medicine, among the founders of medical schools and active civic leaders, prolific authors in diverse fields, both had a lifelong in- terest in natural history and especially in ichthyology. Storer was a native of ]\Iaine and a graduate in 1822 of Bowdoin College ; 2 BREVIORA No. 80 he moved to Boston to begin his medical education at Harvard I^niversity where he later taught for many rears. Kirtland moved to Ohio in 1823 from New Haven where he had graduated from Yale Medical School in 1815. Gehr's (1952) biography of Kirtland states that he was a member of the first class to matriculate in the medical school of Yale University, and that he pursued a long career as a physician and teacher of medicine. He was a founder of, and a proponent of legislative action which established the medical school of AVestern Reserve ITniversity. He was for most of his life a con- tributing student of natural history — of mollusks, insects, fishes and birds. He was the author of man}' j)ublications and articles on subjects as diverse as medicine and horticulture. He was a correspondent of Professor Louis Agassiz and collected Ohio fishes for him. He Avas a frequent lecturer in many areas of knowledge, and included the duties of an editorship among his other responsibilities. He was elected a member of the American Philos()i)hical Society at the age of 82. Legislator, teacher, physi- cian, biologist, noted naturalist, Jared Potter Kirtland led a full and most productive life. Indicative of their stature as naturalists is the fact that Kirt- land and Storer were each appointed in 18.'57 by the government (jf their respective states (Storer by iMassachusetts and Kirtland by Ohio) to assume an important role in natural history surveys of those states. Storer 's report on the fishes and reptiles of Mas- sachusetts was published in 1839 by the Commissioners on the Zoological and Botanical Survey of the State. Of his report, Storer wrote to Nehemiah Cleaveland of Bowdoin College prob- abl}' sometime between 1855 and 1867, in a letter preservetl by the College, "In 1837, Commissioners were appointed l\v the Governor and Council to prosecute the Botanical and Zoological Survey of the Commonwealth. The departments of ichthyology and herpetology were entrusted to me. I concluded my report in 1839 ... In 1842, Dr. DeKay published his 'Zoology of the State of NeAV York. ' You will excuse me for copying from that report the following: 'In 1839, Dr. D. II. Storer, who had previously furnished several valuable papers on ichthyology in tbe Boston •Journal of Natural History, published a nuisterly report on the fishes of Massachusetts. In this rejioi-t, the author has enumerated one hundred and nine species, of whicli one liundi-ed and four 1!I57 KIRTI.AXn's DKAWING!^ OF FISHES 3 arc accompanied by oi-igiiial aiici carefully drawn up descriptions. This report is an invaluable document to the American ichthyolo- irist and in every way worthy of its eminent author." The Ohio survey was abandoned in 1839, after submission of the second annual report to the legislature. This report, which contained Kirt land's descriptions of the fishes of Ohio, was not l)ublished by the legislature. Storer had been an active member of the Boston Society of Natural History for nearly a decade when the Journal of that Society began in 1840 the publication of Kirtland's (1840, 1841, 1842a, b, 1843. 184rm, b, 1846) descriptions of Ohio fishes. It was perhaps during the period of publication from 1840 to 1846 that Kirtland and Storer became acquainted ; their friendship is indi- cated by Kirtland's (1845a, p. 82) description of Lenciscus Storerianus: "It atfords me pleasure to dedicate it to my friend, D. Humphreys Storer, M. D., to whom I am indebted for essen- tial aid, and many important suggestions, in arranging and de- scribing our western fishes." Storer 's connection witii the Boston Society of Natural His- tory, and his own interest in fishes, would certainly have made him cognizant of Kirtland's work on the fishes of Ohio, and pos- sibly it was through an earlier acquaintance with Storer that Kirtland decided to publish in the Journal of the Society. Kirt- land's figures would presumably have accompanied his manu- scripts to Boston, where Storer might well have received them, after publication, as a gift from Dr. Kirtland. (ielir (1952, p. 17) indicates that Kirtland's figures for the Ohio report were drawn by himself, and indeed that he learned to draw for the express purpose of illustrating his work on fishes. Dr. Storer maintained a lifelong interest in Bowdoin College. As late as the year of his death in 1891 he contributed money and pamphlets to the College, and at various times submitted infor- mation requested for alumni records of the College. After his death in 1891, his daughter sent to the College several items in the form of books, journals and miscellaneous pamphlets which had been in her father's library. The box in which the drawings w^ere found contained several pamphlets inscribed to Dr. Storer by various authors. It is not improbable that the drawings executed by Kirtland came to Bowdoin College with material donated after Dr. Storer 's death. 4 BREVIORA No. 80 111 view of their historical significance, I liave felt it worth- while to inform ichthyologists of the location of drawings which I ascribe to Dr. Jared Potter Kirtland, early describer of the fishes of Ohio. REFEBENCES CITED Gehr, a. R. 1952. Jared Potter Kirtland. The Explorer, 2. No. 7: 1-33. Kirtland, J. P. 1840. Descriptions of the fishes of the Ohio River and its tributaries. Boston Jour. Nat. Hist., 3: 338-352. 1841. [Same title.] Boston Jour. Nat. Hist., 3: 469-482. 1842a. [Same title.] Boston Jour. Nat. Hist., 4: 16-26. 1842b. Descriptions of the fishes of Lake Erie, the Ohio River, and their tributaries. Boston Jour. Nat. Hist., 4: 231-240. 1843. Descriptions of the fishes of the Ohio River and its tributaries. Boston Jour. Nat. Hist., 4: 303-308. 1845a. [Same title.] Boston Jour. Nat. Hist., 5: 21-32. 1845b. Descriptions of the fishes of Lake Erie, the Ohio River, and their tributaries. Boston Jour. Nat. Hist., 5: 265-276. 1846. [Same title.] Boston Jour. Nat. Hist., 5: 330-344. FIGURES Figure 1. Probably Acipenser fulvescens Eafinesque, named A. nudus by Kirtland due to loss of the bony plates with age. The pencilled signature on this figure, lacking in the Boston Society article, was retouched by the l)hotographer. .'^?HHBt?^Tr^fT^Tr^.'^'*.-Arr*<; '.•-■rjr-:-.'i^-f*^:^c^';^-''>Tr^r'?W.V/" <»■ v.-;^s'w ■>"!r-u'~.--';'-rar'ici>*">r"> ' //».,., ,/ / HraosKal Figure 2. A second figure of Acipenser fuJvescens Eafinesque, designated by Kirtland Accipenser mncrostonius Eafinesque. This figure, also unpub- lished in the Boston Society article, was signed in ink. w ■■-''''^"-'^4^1 \ ^ ^*<*«;' ipft»i„y iUlm '■' //■! r''r/ /■/ Figure 3. Kirtland 's drawing of Pomoxis annularis Eafinesque, appar- ently intended to be named after Storer, CichJa Sioreria. The signature is printed, unlike other signatures in the collection. The drawing was not in- cluded in the article of the Boston Society of Natural History. BREVIORA MuseiLiioi of Comparative Zoology Cambridge, Mass. September 30, 1957 Number 81 CONTKIBUTIONS TO A REVISION OF THE EARTH- WORM FAMILY LUMBRICIDAE. I. ALLOLOBOPHORA LIMICOLA By G. E. Gates INTRODUCTION External characteristics and the internal anatomy of many earthworm species are known only from the original descriptions, all too often based on a single specimen. Hence little information as to variation, even in common forms, is available. Such records as have been provided usually are for the genital system. More necessary at present are data for somatic systems that are now being found to be phylogenetically more conservative. This contribution is the first of a series that will provide, inso- far as circumstances permit and from lumbricid material pre- served in the field without benefit of special techniques, informa- tion as to characters used in the obsolescent taxonomy of the past as well as those that may be employed in the more natural classification of the future. The author's thanks are extended to Harold Davies for pro- viding three of his New Jersey specimens. Family LUMBRICIDAE Genus AlLOLOBOPHORA Eisen 1874 Allolobophora LIMICOLA Michaclseu, 1890 Allolobophora limicola Michaelsen, 1890, Jahrb. Hamburg. Wiss. Anst., 7, p. 10 (Type locality, Eolfshagen, Hamburg, Germany. Types probably in the Hamburg Museum.) 2 BREVIORA XO. 81 Allolobophora {Allolobophora) limicola, Bosa, 1893, Mem. Ac. Sci. Torino, 43, pp. 424, 450. Allolohoplwra limicola, Beddard, 1895, Monograph, p. 716. Allolobophora limicola, Bretscher, 1896, Rev. Suisse Zool., 3. p. 52."^. (Swit- zerland) Helodrilus (Allolobophora) limicola, Michaelsen, 1900, Das Tierreich, 10, p. 484. Ilelodrilus (Allolobophora) limicola, Piguet & Bretscher, 1913, Cat. Inv. Suisse, 7, p. 177. Allolobophora limicola, Ude, 1929, Tierwelt Deutschlands, 15 (1), p. 120. Allolobophora limicola, Michaelsen, 1931, Bull. Mus. Hist. Nat. Belg., 7 (1), p. 6. (Belgium?) Allolobophora limicola, Cernosvitov, 1936, Bull. Mus. Hist. Nat. Belg., 12 (21),pp. 5, 12, figs. 8-9. (Belgium) Allolobophora limicola, Tetry, 1940, Bull. Mus. Hist. Nat. Belg., 16 (31), p. 21. (Belgium) Allolobophora limicola, Wilcke, 1949, Senckenbergiana, 30, p. 178. Allolobophora limicola. Gates, 1953, Bull. Mus. Comp. Zool. Harvard, 107, p. 518 ; Breviora, 15, p. 2. (Massachusetts) Allolobophora limicola, Graff, 1953, Zool. Anz., 151, p. 26. Regenwiirmer Deutschlands, pp. 26, 60, fig. p. 61. (Germany) Allolobophora limicola, Davies, 1954, Breviora, 26, p. 3. (New Jersey) Allolobophora limicola, Satchell, 1955, Ann. Mag. Nat. Hist., (12) 8, p. 224. (England) Material examined Massachusetts Boston. Arnold Arboretum, peat bog. May 18, 1956, 1-10-62-1. New Jersey (Three specimens supplied by Harold Davies. Tide, Davies, 1954, p. 3.) External characteristics. Segments (c/. Table 1), of worms that are not obviously recent amputees, 83-136. Setae begin on ii where all usually are present. \'entral setae larger, slightly- more protuberant than lateral ones, more closely paired on xv. Apertures of follicles are very slightly farther apart than elsewhere in the genital tumescences. Inter- setal intervals varj^ in absolute width from one region of the body to another as well as on the same segment in individuals of dif- ferent sizes. Relative widths of the intervals are as previously indicated (Gates, 1953). 1957 ALLOLOBOPHORA LIMICOLA Table 1 Segment number in A . limicola Number af specimens from Segment Mass. Number 1956 1952 (b) N. J. (a) Totali 83 1 1 95 1 1 98 1 1 104 2 2 105 1 1 107 1 1 108 3» 3 110 1 1 111 1 1 112 1 1 2 113 2 1 1 4 114 3 3 115 2* 2 117 2 118 0 2 4 119 1* 1 120 3 2 5 121 4 o 6 122 4 1 1 6 123 8 5 o 15 124 8 3 1 12 12.- 6 1 i 126 2 1 3 127 1 1 2 128 4 4 129 1 1 2 136 1 1 (a) Cited from Da vies (1954) (b) Cited from Gates (1953) Nephropores usually closed and not certainly recognizable even after softening, treatment with picric acid, or use of other techniques, but probably so situated as to permit the following characterization. Nephropores alternating irregularly, with fre- quent asymmetry, between two levels, one well above D, the other somewhat above B, pores of xiv usually in DD, of xv-xvi usually at the lower level. * One of the worms, in each case, may be a posterior amputee. 4 BREVIORA NO. 81 Clefts in male tumescences, though obvious, go down only slightly, if at all, below general level of outer surface of the body. The clitellum always covers xxx-xxxv, occasionally extending into xxix (sometimes to the equator but never in front of it), and/or into xxxvi or even to 36/37. The clitellum reaches slightly farther posteriorly, in variant specimens, on that side which has the longer tuberculum. Tubercula pubertatis have a distinct margin, after appearance of clitellum, only on the me- dian side and occasionally are unrecognizable even under optimal illumination. The median margin is concave or incised at 33/34 (also at 34/35, 35/36, in variant specimens). Tubercula usually are in xxxiii-xxxiv (57 specimens) but on some of the variant worms are also in xxxv as follows .- on the left side (5), and right side (2), on both sides (8). Locations in the remaining specimens are : on left sides of xxxiii-xxxv and right sides of xxxiii-xxxvi (1), on left sides of xxxiv-xxxvi and right sides of xxxiii-xxxv (1), in xxxiv-xxxvi on both sides (2). Incidence of worms with variant tubercula in previous lots (of 129 aclitellate and clitel- late specimens), 14.7 per cent; in present lots (of 73 specimens), 26 per cent. Genital tumescences (cf. Table 2) may be present on any of .segments ix-xiii, xvi-xvii, xxvi-xxxii, xxxv-xxxviii. Locations, be- hind XV, on the 19 variant specimens (as regat"ds tubercula), are as follows: on xxix-xxxii, xxxv-xxxvi (1), on xxix, xxxi-xxxii, xxxv-xxxvii (1), xxx-xxxii, xxxv (3), xxx-xxxii, xxxv-xxxvi (2), xxx-xxxii, xxxv-xxxvii (2), xxx-xxxii, xxxvi (2), xxx, xxxii. xxxv (2), xxxi-xxxii, xxxv (2), xxxi-xxxii, xxxv-xxxvi (2). xxxii, xxxv (2). Internal anatomy. Septa 6/7-9/10 muscular, 5/6 and 10/11- 11/12 or 12/13 slightly muscular. Pigment unrecognizable in longitudinal muscle strip at niD and in sections of body wall, also in circular muscles after stripping off longitudinal layer. Epidermis thickened two to three times in male tumescences, less so in genital tumescences. Oesophagus widest in x, narrowing gradually through xiv, constricted fairly deeply at in.sertions of 10/11-11/12 so that portions in xi and xii appear subspheroidal. Additional but slighter constrictions, at 12/13-13/14, occasionally are recogniz- able. Lumen verticallv slit-like in xi-xii or widened in middle of 1957 ALIiOLOBOPIIORA LIMICOLA xii (internal constriction of gut wall). An internal constriction in xii usually is confined to one side but in xiii usually is on both sides. Oesophageal wall much thinner in xiii-xiv than in xi-xii, calciferous lamellae recognizable only in xii-xi and the posterior half of x. In x the dorsal blood vessel at first appears to be within a groove in dorsal face of gut but this appearance results from presence of two rounded protuberances (77 specimens) which, with a similar pair on the ventral face, provide the only evi- dence, in these strongly contracted worms, for existence of calci- Fig. 1. Oesophagus in segments x-xiv, dorsal view. Fig. 2. Oesophagus in segments x-xiv, horizontal section. ferous sacs. Intestinal origin in xv (77). Gizzard in xvii (77), the layer with brilliant muscular sheen in horizontal section abruptly narroAving and ending shortly behind insertion of 17/18. Posteriorly the gut wall is much thinner, weaker, breaking most easily in both wet and dry material just behind 17/18. Gut narrowing slightly through xix, rather valvular in region of insertion of 19/20. The typhlosole, shortly after its beginning. BREVIORA NO. 81 is nearly circular in transverse section and deeply grooved on each side clear to level of the small median ridge on the ventral face. The grooves appear to be exactly opposite each other (not alternate) and produce, on each side, a series of laterally di- rected, haU'moon-shaped leaflets. Superficially, as previously suggested, the appearance suggests a stack of coins with a cord running along it. Behind the clitellum the lateral grooving gradually disappears and the typhlosole becomes more nearly thickly laraelliform. The grooving is typical, characterizing each of the 77 worms of the 1956 series. The typhlosole ends in front of the 119th segment as shown in Table 3. Dorsal and ventral blood vessels have not been traceable an- terior to 5/6, probably because they are empty. Extra-oesopha- geals, median to the hearts and in contact with the gut only in mftmmim^fimmifammm^mwm Fig. 3. Typhlosole. Anterior end from ventral side. Fig. 4. Typhlosole. Horizontal section slightly posterior to level shown in Figure 3. regions of some septal insertions, pass up to the dorsal trunk in xii (77). Hearts present in vi-xi (35), usually slenderer in vi and there often empty or nearly so. Inability to see hearts in vi of the other 42 worms proljably not due to absence but to difficulty of distinguishing the vessels, when empty, from nephridial and other tissues with which they are associated. Last hearts in xi (77). Nephridia, behind xiv and except in cli- tellar region, large, back from xix reaching up to dorsal blood vessel. A bladder is present but its condition (not improved by various attempts to render nephropores visible) obviates accurate characterization, possibly XT-shaped. The supposed duet, slender and very fragile, always seems to pass into parietes close to B 1957 ALLOLOBOPHORA LIMICOLA 7 Brain, left by a transverse section exactly alonj? 3/4, in iii. Usually all of each circumpharyngeal commissure also is left in iii, at most only very short stubs remain in iv. Holandric (77). Seminal vesicles, four pairs, in ix-xii (77). Male deferent ducts, just behind funnel septum, variously looped, the loops (epididymis) or all but one of them bound together in a disc (77) often of about the same size as a testis. Ducts of a side come into contact posteriorly in xii but may not unite until into xiv. They disappear into longitudinal musculature about at eq/xv. Ovaries, each with a single distal egg-string containing one to three ova, erect discs with shape approximating to vertically oblong. Oviducts pass under a muscle inserted on h follicle of xiv and then into parietes just lateral to B. Ovisacs always present (77) and in the usual positions on posterior face of 13/14. Spermathecae, two pairs, always in ix-x (77) and open- ing to exterior posteriorly. Atrial, TP and supraparietal GS glands are lacking in the coelom (77) and no special glands are recognizable (in unstained free-hand sections) Mdthin the muscular layers. Follicles of ventral setae, from xv into postclitellar region at least, with two exceptions, much larger and more protuberant into the coelom than those of the lateral pair in the same segment. Exceptions are provided by segments xxxiii-xxxiv where the ventral follicles usually appear to be little if any larger than the lateral ones. Ventral follicles of xv usually appear to be little if any larger than in xvi. Follicles opening through genital tumescences al- ways (77 specimens) are especially enlarged and unusually pro- tuberant into coelomic cavities. Setae in all examined follicles have been genital (GS) and may be 1+ mm. long. Reproduction. Spermatophores were not found on any of the 1956 series. Iridescence of male funnels more or less brilliant in each worm except the juvenile and one of the aclitellates, localized in a few small spots in several aclitellates, more general in others and most l)rilliant in those worms with clitellar tumes- cence nearer the maximum (which probably had not been reached by any of the present lot) . Iridescence in the contents of some or all of the spermathecae (both spermathecae of a side empty in several worms) of 26 clitellate specimens shows that sperm are present there too and hence that copulation had taken 8 BREVIORA NO. 81 place. As sperm had been, or were being, matured so generally, and as they are exchanged during copulation, reproduction can be assumed to be biparental. Regeneration. A tail regenerate at 78/79 has ca. seven seg- ments but the metamerism is quite abnormal. Ahnormality. No. 1. Extra male tumescences, with equatorial cleft, present on left side of xvi. No pore is recognizable in that cleft and both male gonoducts of the left side apparently pass down through the muscular layers in xv. No. 2. Left anterior spermatheca bifid down to parietes. Both ampullae preseptal, filled with sperm. No. 3. Eight segments in front of anal region are involved in a complicated spiral metamerism. Nos. 4-6. One helicometamere, in the postclitellar region of each of two worms, was observed while counting segments. Two spirals were noticed in the same region of another worm. Table 2 Genital tumescences in Massachusetts collections of A. limicola Number of specimens in In segment 9 previous lots 100 present lot 74 Totals 174 10 5 5 11 97 68 165 12 112 74 186 13 3 3 16 2 2 17 1 1 26 1 1 27 2 2 28 6 6 29 31 23 54 30 88 58 141 31 81 51 132 32 118 72 190 35 98 71 169 36 39 25 64 37 18 6 24 38 1 1 I'sent lot is from the very same spot at which a majoritj- of the previous specimens were obtained. 1957 ALLOLOBOPHORA LIMICOLA Table 3 Typhlosole termination and segment number in American collections of A. limicola. 1- »«*«»* ^7-28 1 » « « * s » « 90 1 ******** A ********* A ****** 1 »»«»»*» 1 »*«-»*«* 9 ****** 1 »♦»«*««* 2 «»»»»♦**« 2 » •» * » 2 »»«**■» J »«♦»«»* 27 ««***•»■»» 5 »*«*»*««• ^ *«*»»» 2 *»*»•♦>■ 2 «»»•■»* 2 »»*«»»* 2 * « « > « g * * « ^ « * » * 13, 17, 27, 28 ******* 23 28 » »•» **** TO » » » » * oo ♦ Genital tumescences present In the segment thus indicated, but disregarding asymmetry, i.e., absence on one side. 1957 ALLOLOBOPHORA LIMICOLA 11 liemarks. A dark coloration with a bluish metallic lustre in the body wall of the New Jersey worms is an artifact that results, in appropriate conditions, when formalin around the worms has remained for some time in contact with cork. Precis. Clitellum, saddle-shaped, on 5i^-8i/2 segments, xxviii/2, xxix, xxix/n, xxx, xxx/2-xxxv, xxxvi/n, xxxvi. Tubercula puber- tatis with median margin concave or incised at 33/34, in xxxiii- xxxiv (xxxv, xxxvi), rarely lacking on xxxiii. Genital tumes- cences around apertures of A'entral follicles only, usually in ix, xi, xii, xxx-xxxii, xxxv, occasionally in one or more of x, xiii, xvi-xvii, xxvi-xxix, xxxvi-xxxviii. Male pores equatorial on xv. in BC, within transverse clefts in tumescences that obliterate 14/15-15/16. Female pores just lateral to B, equatorial on xiv. Quadrithecal, pores at C, on 9/10-10/11. Setae, closely paired, more so laterally, ventral setae larger ( ?), CD < AB, BC g t> ."S 2 rt >H ° '^ 0 oj o 53 53 o f^,^ „ PI > OJ O ' -^ .;.- ° •" I ^ g — . ts "2 tS -^ ,i; 'tS S 0 c3 SQ ---5 3 TS •'^ g o Tri O .2 a ■^^^-<^%^ " 1^ - "^ .. - ^ — 1$ 1=1 !h ^ -^ jj .f-i 03 0 '-^ .^ •" ■ ^ =2 ^ •" 53 •- 2 :g O 2 > S — " o fi -^ O 'H ''-' "-^ 03 2 ~ O !-. .S (u " "^ oj >-. 3 oS h«^ -g 5-2 S =H fl O C ■ o '^ fe fl ^Ks-~42t^ ft <» _! 3 rt <" > 05 ^ - - CI tj w» PI ei! CO C CU 5 m "^ --; S -S '^ -2 '^ 2 . in > on -tf d, 05 -k-J O O 50 Cj SH i ■•-= S M l-s >" +^ Ol S -5 £ ^ ■£ § ^ .2 •- « J, ci rs fn t-i > Oj ti » a .. . 3 S f- > o C ^ _ O O -*^ OJ ?= o S 2 --^ c3 o m ■ « iti. The two cotypes of the latter were statetl to have come from Orizaba, Vera Cruz, Mexico, but Smith thinks it more likely that they were collected on the Atlantic slope of the Isthmus of Tehuantepee. SpociniPii Goteborg A Goteborg B MCZ 54974 AMMI 12439 Sf.x Vcntrals -f Cauda Is 140 + 70 1-14-f 71 148 + 79 14S + .\- Dorsals 21-19-17 21-19-17 21-19-17 21-19-17 TABLE Dfirsals Pro- Post- rpiirr Liiwcr liod.v ilrcip to 19 ocular m-nlni- Labials I,abial^^ + Tail at ventral Length iniml)er in mm. 11 ir.-ll 7 S-9 1 ■•4 8(4,5) 10 1 so + RO % 8(4,5) 10 180 + 58 3 8(4,5) 10 485 + 1(57 3 8(4,5) 10 543 + x As Thamnophis bovallii Dunn is known only from southern Nicaragua and northern Costa Rica, it would be surprising if it BREVIORA No. 83 Figure 3. Thamnophis bovallii. Dorsal, lati'inl and ventral views of the head of ^I("Z 54974. Xoto the very irrcsular Ijlaek fleeks in tlie eentor of the parietals where the other specimens have the light, black-edged spot. The color difference between left and right sides is due to the shedding of the scales of the left side. 1957 THAMNOPHIS BOVALLll 5 were quite identical with a form from eastern Mexico, the more so since two races (later three, Smith, Nixon, and Smith, 1950, p. 579) of the same species were recognized by Smith from inter- vening Guatemala, and since there appear to be no records at all for this group of garter snakes between Guatemala and southern Nicaragua. W. W. Milstead (1953) has treated this group of Mexican and Central American Thamnophis very differently than did Smith but has followed him in synonymizing hovallii, again with- out discussion. Milstead has made Smith's species, sumkhrasti, a race of cyrtopsis {equcs of authors), and has contributed to the understanding of the variational range of scale counts and color of the populations he treats. However, he is so far from giving serious attention to hovallii that after citing it in the synonymy of T. cyrtopsis sumichrasti he forgets to mention Nicaragua in the range of the subspecies in the section on ' ' Sub- specific Variation" {loc. cit., p. 359), although he does include it later under the formal presentation of the range of sumichrasti {loc. cit., p. 361). Several characters persuade us not only to regard hovallii as valid, but even provisionally to retain it as a full species until the taxonomic uncertainties surrounding this section of Thamno- phis are further clarified. In scale characters the most significant point is the presence of 21 rather than 19 scale rows at the nape, while in coloration the most striking feature is the dark anterior border (narrowed medially) of all the ventrals. Both these fea- tures serve to separate T. bo vail ii from all T. cyrtopsis as con- ceived by Milstead. The shortness of the lateral bars (in our two examples), and the presence always of two rows of spots in the vertebral region (even if partly coalesced) are additional char- acters. ECOLOGY The San Jorge specimen was collected by two boys under lakeshore debris, early on the morning of July 1, 1956. Just above the water line there is a windrow of vegetable debris, wood, coco- nut husks, etc. This and the small puddles of this zone were inhabited at dusk by large numbers of small toads and frogs. Series of Leptodactylus melavonotus (Hallowell), and Bufo mari- bREVIORA No. 83 nus (Linne), were collected during the visit. During the daytime the frogs appeared to hide away from the beach, but large num- bers of lizards were found along the bluff beyond the beach {Anolis achilles Taylor?, Sceloporus rariabilis oUoporus Smith. and CnemidopJiorous d. deppii Wiegmann were collected). In addition to the specimen of horallii the beach also yielded a speci- men of Thamnophis saurita chalceus (Cope) and one of Lepto- deira rhombifera Giinther. NICARAGUA GRANADA SAN JORGE COSTA RICA Map. Thamnophis bovallii. Sketch map showing the relation of the thioe localities from which this species has been recorded. Granada, the type locality of bovallii, also lies on Lake Nicara- gua (elev. 105 ft.) and the two specimens listed l)y Dunn (loc. cit.) probably came from a locality similar to that described. "Colorado Bar," on the other hand, refers to the mouth of the Colorado river (Boca do Rio Colorado on Costa Kican maps). While this lies on the Caribbean coast, it is part of the delta of the San Juan River which drains Lake Nicaragua and which has 1957 THAMNOPHIS BOVALLII 7 such a low gradient that it was used for navigation in historic times. Thus all three localities at which T. bovallii has been taken are lowland. This is in strong contrast with Milstead's descrip- tion of the situation in T. cyrtopsis which he regards as a species that does not descend below 1000 feet in altitude. He admits one exception, the Edwards Plateau of Texas (minimum altitude 700 feet), but ascribes this to the temperate, subhumid climate of the region. He doubts the records of cyrtopsis from British Hon- duras, Yucatan (Mexico), and Escuintla, Guatemala, which would bring the form into the tropical zone. Here again is evi- dence of his lack of attention to hovallii since he fails to mention it or Nicaragua at all in this discussion. ACKNOWLEDGMENTS The jvuiior author welcomes this opportunity to thank those friends, whose unselfish and valuable assistance made this trip and the visit to Rivas and San Jorge both possible and successful. I should like to express my particular appreciation to Messrs. Julio Cesar Moya, Inspector Porestal, and Gustavo Hernandez, Agronomo, both of Rivas as well as to the municipal officials of this city, and in addition, to my companions Gerardo Budowski of the Instituto Interamericano de Ciencias Agricolas of Tur- rialba, and Phil Noyce. We also wish to thank Arthur Loveridge for comments, and Ernest E. Williams for additions to and revision of the manu- script, and the staff of the Department of Amphibians and Rep- tiles of The American Museum of Natural History for loan of the Colorado-bar specimen. LITEEATURE CITED Dunn, Emmett Reid 1940. Notes on some American lizards and snakes in the Museum at Goteborg. Herpetologica, vol. 1, pp. 189-94. MiLSTEAD, William W. 1953. Geographic variation in the garter snake, Thamnophis cyrtopsis. Texas Jour. Sci., vol. o, no. 3, pp. 348-379. 8 BREVIORA No. 83 Smith, Hobart M. 1942. The synonymy of the garter snakes (Thavmophis) , with notes on Mexican and Central American species. Zoologica (New York), vol. 27, pts. 3-4, pp. 97-123. Smith, Hobart M., C. William Nixon, and Phillip W. Smith 1950. Mexican and Central American garter snakes (Thamnophis) in the British Museum (Natural History). Jour. Linn. Soc, vol. 41, no. 282, pp. 571-584. BREVIORA MiuseuiirM of Comparative Zoology Cambridge, Mass. January 30, 1958 Number 84 REDISCOVERY OF THE AUSTRALIAN CHELID GENUS PSEUDEMYDURA SIEBENROCK (CHELIDAE. TESTUDINES) By Ernest E. Williams In 1954, L. Glauert described a new species — Emydura iv- spectata — from western Australia, comparing his new form only with species of Emydura. A g:lance at his photograph revealed that the new form was congeneric with, and perhaps conspecific with, Pseudemydura umhrina described by Siebenrock from an unique type specimen in the Vienna Museum with no more pre- cise locality than "Australia." and which had come into the possession of that museum in 1839. Pseudemydura umhrina was named by Siebenrock in a pre- liminary note in 1901, but it was not figured nor fully described until 1907. The long interval between these two papers may, I tliink, be easily explained : The specimen was unique and without adequate locality, yet it was made the type of a genus. Only Sie- benrock's wide knowledge of the order as a whole justified such a procedure, and even for him it was natural and desirable to wait for some years in the hope of procuring additional specimens the better to document his case. Siebenrock never obtained any further specimens, but in default of them his 1907 paper compared Pseudemydura carefully and in a key with every other genus in the family. Pseudemydura belongs to the group of short-necked chelids, and is peculiar among these in having a large intergular plastral scute that not only separates gulars and humerals but penetrates some distance between pectorals. A somewhat similar condition occurs in the long-necked chelid genus Chelodina, in which, how- ever, the gulars usually meet in front of the intergular, separat- 2 BREVIORA NO. 84 ing the latter from the anterior ])lastral margin. Only in Chclo- dina interc/idaris (Fry, 1917) among Reeent turtles is there any elose parallel to the condition found in Pscudcrnydura. Among known fossil forms, however, a very similar intergular pattern occurs in some members of the related family Pelomedusidae — KlochcJ]!^ perfect a Nopcsa of the Cretaceous of Europe, and prob- ably in some Taphrosphys (Cretaceous and Tertiary of Xoith and South America). The relationship of the intergular to the other anterior plastral scutes in Pseudemydura nmhrina (clearly seen in Glauert's speci- mens) is thus striking enough among Recent members of the Chelidae and in particular tlu^ short-necked chelids to be l)y itself a feature permitting recognition at the generic level. However, the judgment that the specimens from Warbrook, 24 miles north of Perth, West Australia, described by Glauert as Emydura Inspcciata belong not to Emydnra but to Pseudemydura does not rest only on this feature but also uj^on a similarity that in crea- tures so variable as turtles is surprisingly complete. Neither the Vu una tj'pe nor any of Glauert 's material has been available to im\ but the similarity of the two described forms can be placed beyond all question merely by comparing in parallel columns a translation of Siebenrock's 1907 description and Glau- ('i-t"s 19;14 account. I have given Siebenrock's description in full and rearranged Glauert's shorter description to make the paral- lels and ditferences more obvious : Pseudemydura umbrina Emydura inspectata Carapace length: 106 mm Carapace length (two siiecimens) : 133 mm, 98 mm ('.■II apace breadth: 81 mm Carapace breadth (two specimens): 103, 83 mm Carapace height: 33 mm Carapace very strongly flattened with Carapace slightly depressed with a a definite vertebral furrow, most distinct sulcus. di.stinct on tlie middle three verte- brals. U)5b KEDISCOVERY OF PSEUDEMYDURA I'fsciidriii iidina uinhi'ina '1'Ir' postiM-ior carapace rim only sliglitly expanded so that the greatest breadth (between the eighth marginals) only slightly ex- ceeds the breadth at the middle of the shell. Anterior carapace rim truncate and only insignificantly incurved. Posterior carapace rim projecting medially into a keel produced by the arching of the supracaudals and adjoining eleventh marginals. The supracaudals also forming ventrally an acute-angled notch, which, with the posterior end of the plastron, also angularly notched, forms a rhomboidal open- ing for the tail. First vertebral longest, broader an- teriorly than posteriorly, narrower than the second and third, as broad as the fourth and fifth. First costal as broad as, the remain ing three costals narrower than, the corresponding vertebrals. Of these the second costal broadest, almost twice as broad as the fourth. All costals broader at the lateral margins than medially, and broader than the lateral margins of the corresponding vertebrals. Carapacial shields leathery, finely wrinkled. Nuchal rather large, trapezoidal, broader in front than behind. Kmydura inspcctata Slightly narrower in front (96 mm as compared with 103 mm) ; mar- ginals 5-8 with a distinct flange. First vertebral largest, as wide as long, obtusely pointed in front, the second and third much wider than long, fourth and fifth slightly wider than long, longer than the second and third. Second and third vertebrals shorter than the adjacent costals. Nuchal very small (9 x .5 mm) en- tirely free below. BREVIORA NO. 84 Pseudemydura umbrina Lateral marginals very narrow; the sixth narrowest, only about one- third as wide as the ninth. The form of the lateral marginals is correlated with the great size, es- pecially breadth, of the second costal. Plastron flat, about as large as the shell opening, with a distinct angle at the bridge. Anterior lobe broader than posterior lobe, truncate. Posterior lol)e with an angular notch. The two corners of the anal notch are bent upward to make contact with the posterior rim of the shell, thus forming the rhomboidal open- ing l)efore mentioneil. Breadth of bridge one-third the length of plastron and one-half the width of the anterior lobe. Intergular very large, heart-shaped, not much longer than broad, its broad anterior margin straight, weakly dentated, the pointed pos- terior end inserted between the pectorals, separating these in their anterior third. (iulars very small, forming equilat- eral triangles, their medial sides scarcely one-third the adjacent lat- eral rim of the intergular. numerals small, widely separated from one another, in form and size showing great similarity with those of Chelodina. Emydura inspectata Fifth, sixth, seventh and eighth marginals with distinct flange, fourth to seventh narrowest. Plastron (124 x 95^2 nim, measured to the edge of the bridge) flat, large, almost as wide as the cara- pace, bridge rounded. Plastron narrower behind, semi-cir- cular in front, posteriorly with a broad, straight-sided, anal notch. Bridge one-third the length of the plastron measured along the mid- line. Intergular very large (29.5 x 25.5 mm). Widely in contact with the pectorals. Gulars very small, widely separated. Humerals triangular slightly nar- rower than the intergular. I 1958 KEDlt^COVERY OP PSEUDEMYDURA Fseudernydu ra ii in hrtit a Anal middle suture longer than the pectoral and siK"iticaiitly longer than the femoral. Intergular longer than the pectoral middle suture. Head broad and tiat, the upper sur- face finely wrinkled. Parietals strongly expanded, cover- ing the whole width of the head, their posterior margin not ])ointed but somewhat incurved. Snout short. Interorbital space broad and con- cave, its breadth exceeding the transverse diameter of tlie orbits. Both jaws narrow; the l)readth of the lower jaw at the symphysis not quite two-thirds the transverse di- ameter of the orbit. Two small chin barbels, widely sep- arated. Back of neck with numerous large erect conical tubercles. Limbs with rather large flat scales. Transverse lamellae such as occur on the anterior surface of the loiver arm in Emydvra entirely lacMng. Webbing well developed, extending to the claws. Fifth toe of hind foot clawless. Tail short, reaching scarcely to the posterior rim of the shell. (■'.nijidura tiinptrtalu Pectorals slightly shorter than the femorals which are shorter than Ihe anals, the l;itter forming a broad shallow notch behind. Head broad and flat, rugulose above. Snout projecting. Interorbital space twice the width of the symphysis, the latter equal to the vertical diameter of the orbit. Two small barbels. Neck above and laterally with pro- nounced conical tubercles ; tem- poral region, chin and throat reticulate. Forelimbs with three series of trans- verse lamellae and a flap of three or four enlarged scales. Tail very short, hardly projecting. 6 BREVIORA NO. 84 Pseudemydura umbrina Emydura inspectata Carapace and upper surfaces of head Color of carapace bone brown, umbra brown. Young: brighter brown. Head above dark olive, sides of face, lower lip, chin and throat mar- guerite yellow, a few dark mark- ings on the throat. Plastron dirty yellowish green. All Plastron buffy olive, over the sutures sutures of the shell brown. buff. Young: plastral plates edged with brown. Limbs and back of neck dark brown, Forelimbs olive black, hind limbs the jaws horn-colored. darker. Siebenrock "s type was intermediate in size between Glauert's type and the latter 's young specimen, but closer to the young specimen. Careful comparison of the two descriptions discovers few differences that are not merely verbal, while there is agree- ment in all essentials even in color — the latter remarkable when it is realized that CTlauert's fresh specimens are being compared with a specimen that had been in the Vienna Museum 62 years before it was first described. Only one of the few described differences would seem possibly significant. Siebenrock makes a point of the absence in umhrina of transverse lamellae such as exist in the genus Emydura. Glau- ert specifies in inspectata "three series of transverse lamellae and a flap of three or four enlarged scales." Glauert's photographs accompanying the type description of inspectata, however, while excellent for other aspects of the ani- mal do not show this area. I have therefore obtained from Dr. Glauert two photographs of the forelimb in his type. These are reproduced in Plates 3 and 4 and sliow the flap or spur very well and in the ventral view exhibit also the three series of scales de- scribed as "transverse lamellae" Ijy Glauert. Siebenrock 's 1907 figure (Plates 1 and 2) definitely does not show any flap or sjiur on this forelimb on the type of Pseudemys )())ibri)ia but does show in the ventral view several series of scales apparently less regular than those of Glauert's type but easily comparal)le to these. These scales are individually rather squar- ish, though arranged in transverse rows. In this respect the dif- 15)58 REDISCOVERY OP PSEUDEMYDURA 7 ference between Glauert's and Siebenroek's descriptions would seem to be primarily verl)Hl. However, the absence of the flap would still remain as a serious difference between Pseudemyduro }(*iibriua and IJini/diira iitspectata. To ehet'k this final point 1 reqvu^sted Dr. Eiselt of the Vienna Museum to re-examine Siebenroek's specimen. He reports: "The type of Pseudeniydura nmhrina is a stuffed and lacquered speci- men. In correspondence with Siebenroek's figures it showed no spurs whatsoever. Reing somewhat of a sceptic, I had the lac removed from the front leg's and feet, and then soaked them in water. Tlie effect was surju'ising : both lower front legs now show on their outer (hind) edges a series of three triangular scales which had, up to now, been glued firmly to the lower surface of the leg. Also the scales on the outer edge of the fifth toe (finger) show a tendency to form a flat and serrated lamella. As a whole the scalation of the legs shown in Siebenroek's figures is rather inaccurate and needs redescription or refiguring. " Dr. Eiselt has also provided a sketch of the flap, which appears to be not significantly different from that described and photo- graphed by Glauert. With this evidence the synonymizing of rseudemiidura uni- brina and Emydura iiispectata must be considered established. Glauert's species name must thus disappear from the list, but his discovery is still important. A genus known from a single specimen without adequate provenance is an unliappy problem. The Glauert specimens provide an exact locality, confirmation of the characters described by Siebenrock, and knowledge of the peculiar spur-flap, a feature missed by Siebenrock and which, though approached in other chelids by a flap of scales along the whole external range of the forelimb, is in its peculiar spur-like and that the German collector Dr. L. Preiss collected there in the late 1830 's and early 1840 's. This accords very well with the date of arrival of the Vienna specimen at that museum, but Dr. Eiselt could supply me with no information about the collector. Acknowledgonoifs. 'My indebtedness to Dr. Eiselt and Dr. Glauert has been conspicuous throughout this paper. I want to 8 BREVIORA NO. 84 make, however, a very personal expression of thanks i'or their kindness and courtesy in answering my questions. The photo- graphs of the forefoot of E my dura inspcctata were furnished by Dr. Glauert ; the reproduction of Siebenrock 's figure is the work of Mr. Frank White. EEFERENCES CITED Fry, D. B. 1917. Herpetological notes III. On a new Chelodinu from Australia witli a key to the genus. Proc. Eoy. Soe. Queensland, vol. 27, pp. 88-90. Glauert, L. 1954. Herpetological Miscellanea. IV. A new swamp tortoise from the Swan River District. The Western Australian Naturalist, vol. 4, no. 6, pp. 125-127. Siebenrock, F. 1901. Beschreibung einer neuen Schildkrotengattung aus der Familie Chelydidae von Australien. Sitzber. Akad. Wiss. Wien, math, nat. Kl., Jahrg. 1901, pp. 248-250. 1907. Beschreibung und Abbildung von Psevdemydina umbriiKi Siebenr. und ilue systematische Stellung in der Familie Chelyd- idae. Sitzber. Akad. Wiss. Wien, math. nat. Kl., vol. 116, pp. 1205-1211. Plate 1. Type of Fseudemydura umbrina, dorsal view. From Siebenrock (1907). Plate 2. Type of Pseudemydura umbrina, ventral view. From Siebeurock (1907). Plate 3. Dorsal view of forefoot of type of Emydura inspectata Glauert. f?J--^^ Plate 4. Ventral view of forefoot of type Emydura inspectata Glaueit. BREVIORA Museiiimi of Comparsitive Zoology Cajibridge, Mass. January 31, 1958 Number 85 THE CHOANAL PAPILLAE OF THE CHELONIIDAE By Thomas S. Parsons Harvard Biological Laboratories Cambridge, Mass. INTRODUCTION There appear to be no published descriptions of the papillae whirh are found along the lateral margin of the choanae or in- ternal nares in the members of the Cheloniidae. They are men- tioned in two papers, Derauiyagala (1939) and Carr (1942), but in each case only the numl)er present is noted and given as a generic character. The following note is based primarily on the study of two heads of Chelonia my das which were obtained from the Carolina Biological Supply Company. Supplementary observations were made on specimens in the collections of the Museum of Com- |)arative Zoology. Histological investigations were confined to Chelonia mydas. The specimens had been preserved in formalin and were stained with hematoxylin and eosin and by the Biel- schowsky silver method. A third method, Mallory's triple con- nective tissue stain, was tried, l)ut did not prove successful. 1 am indebted to Dr. E. E. AVilliams for allowing me to examine the specimens in the Museum of Comparative Zoology, and for his advice throughout the course of this study. Thanks are also due to Dr. A. B. Dawson and Mrs. T. S. Parsons for advice and assistance in making the histological preparations. CHELONIA MYDAS ■■■'■.1. , in Clielo7iia my das the papillae form a row along the entire lateral margin of the choana. They are long and narrow conical 2 BREVIORA No. 85 pi-ojectioiis of irregular length, so that tlie row of papillae ap- pears quite uneven. In the anterior two-thirds of the choana, they lend to be longer and more prominent, Avith ai)proximatply half of the width of the ehoana occupied by the papillae; farther pos- teriorly they do not extend out into the choana any great dis- tance. Deraniyagala (1939) reports not only a lateral roAV. but also a single anteromedial pa{)illary vestige. However no such vestige was seen in any of the specimens studied in the present investigation. The number of papillae in the lateral row is not constant. Carr (1942) states only that they are numerous, and Deraniya- gala (1989) that there are eight. Unfortunately the latter authoi- does not mention the number of specimens which he examiucd. In the two adults dissected for the present study there were eleven to fourteen pai)illae on either side (one had thirteen on each side, the other eleven on the left and fourteen on the right i. The only other adult specimen which was examined (MCZ 7150) possessed ten papillae on the right side and eleven on the left. Thus there are probably ten or more pai)illae along the latei-al margin of each choana in most fully mature Chdonia nijidos: however this is not true of juveniles. In three specimens under six inches in total length (MCZ 1413, 1414, and 29594) there was only a single, rather weakly developed papilla near the anterior end of the lateral margin of each choana. A larger juvenile (MCZ 3454) possessed two papillae on the right side and three on the left. In section, each papilla can be seen to consist of two major histologic elements. First there is a thick, cornified, stratified squamous epithelial covering, and secondly a mass of dense con- nective tis.sue forming a central core. Plate II shows transverse sections of papillae. The upper section, stained with hematoxylin and eosin, was tjiken near the center of the i)apilla. while the lower one, stained by the Bielschowsky silver method, was taken near the base. The epithelium closply resembles normal body epidermis, lu some regions all four strata — germinativum, granulosum, hu-i- dum, and corneum — are Avell developed. The first and last are l)resent throughout, but the granidosum and lucidum are not always recognizable. There is a gradual increase in the thickness 1958 CHOANAL PAPILLAE OP CHELONIIDAE 3 of the stratum eorneum from a minimum at the l)ase of the pap- illa to a maxinuim at its tip. The external surface of the epi- thelium is relatively even at the distal end of the papilla, hut hasally it becomes in-e^ular with numerous small projections and. more rarely, indentations. As in typical epidermis, the boundary betAveen the epithelium and the underlying connective tissue is generally a rather wavy line, although near the tip of the papilla it becomes far more even. No basement membrane eould be seen ; in fact the tibers of the connective tissue appear to continue into the basal epitlielial cells. However in human epidermis, which a])pears sinular when viewed through a light microscope, a thin l)asement membrane can be demonstrated by electron microscopy (Maximow and Bloom, 1957). The core of the papilla is a mass of typical dense connective tissue. Irregularly arranged bundles of collagenous fibers occupy most of the space, but there are numerous included cells and small relatively clear areas. The cellular constituents of the con- nective tissue were not identified in the present study. Within the connective tissue tlicre are (piite numerous very small blood ves.sels; however only a single arteriole and venule could be found at the base of the papilla, and the tissue is not very heavily vascularized. No nerve fibers or endings could be recognized. However nerve fibers would resemble most closely the fibers of the connective tissue, and thus be very difficult to demonstrate. Presumably there are some nerves present; certainly the lining of the oral cavity is well innervated in most forms. OTHER CHELONIIDS Plate I, lower figure shows a specimen of Cantta carcita. In this specimen (MCZ 53183) there is a single choanal papilla located near tlie anterior end of the lateral margin of the choana. This is in agreement with the findings of both Deraniyagala (1939) and Carr (1942). The form of the papilla is essentially similar to that found in Chelonia mydas, but it appears to be slightlv shorter and broader in Carefta. The only specimen of Chelonia depressa which was examined (MCZ 54712) is a juvenile, and possesses a single choanal papilla exactly like that of Chelonia my das juveniles. 4 BREVIORA No. 85 In Eretmochelys imhricata, Deraniyagala (1939) found a single choanal papilla, and Carr (1942) reports one or two papillae. All seven specimens examined for the present study (MCZ 1138, 1141, 1416, 1418, 4049, 49406, and 49407) possess a single papilla exactly like that of Caretta. No specimens of LcpidocJuiys were seen by me. Both Derani- yagala (1939) and Carr (1942) report a single choanal papilla like that found in Caretta. FUNCTION OF THE PAPILLAE In adult specimens of Chelonia mydas the form of the papillae suggests the possibility that their function is to prevent pieces of food from entering the nasal cavities. However the single choanal papilla of the other cheloniid genera and of juveniles of Chelonia mydas could not be very effective in this regard. According to Deraniyagala (1939), all the cheloniids are omni- vorous, preferring an animal diet but often subsisting on purely plant food. Chelonia in particular is primarily hei'bivorous. Its eating habits also differ slightly from the other members of the family in that its "stomach contents are never in a finely divided condition but appear to be 'bolted' entire" (Deraniyagala, 1939. p. 234) ; the other cheloniids tend to break up their food more. However it does not appear to me that an unchewed piece of alga would cause any more injury than a broken piece of mollusc shell if lodged in the choana or nasal cavity. Thus food habits do not seem to ex])lain the differences between the genera. Another possible function of the papillae is that they might bear some special sensory receptors. However the liistological observations provide no evidence for such a theory, although they are not absolutely conclusive. As was noted above, no nerves or sensory endings were observed in the papillae of Chelonia. Since the material was not ideally preserved for histologic study and since nerve fibers would not be clearly differentiated from connective tissue fibers by the stains employed, their ab- sence cannot be considered as proven ; in fact, it appears highly probable that there is some sensory innervation of the papillae. However the absence of any conspicuous nerve trunk entering the base of the papilla and of any distinguishable receptors renders it most unlikely that the papillae function primarily in 1958 (CHOANAL PAPILLAE OF CUELONIIDAl': 5 a sensory capacity. Furthermore, the thickness and high degree of cornification of the epithelium would be surprising, though not impossible, in a sensory structure. Thus I am unable to suggest any function for the single papilla in Caretta, Eretmochelys, or Lepidocluhjs. In Chdonia (adults) they must serve to prevent the contents of the mouth from entering the nasal cavities ; whether or not this is their primary function cannot be stated. BIBLIOGRAPHY Cakr, a. F., .Ik. 1942. Notes on sea turtles. Proc. New England Zool. Chili, vol. 21. pp. 1-16. Deraniyagala, p. E. P. 1939. The tetrapod reptiles of Ceylon. Volume 1. Testiidinates and erocodilians. Columbo. 412 pp. .Ma.xi.mow, a. a., and W. Bloom 1957. A textbook of histology, 7th ed. Philadelphia. 628 pp. I'LATE I Upper, Chelonia mi/da.s. \'iew of part of the palate showini; ilic cho'inal papillae. Lower, Caretla r(ir(thi. \\v\s of tlii' loft side of llu- palate slidwiiiy the single choanal papilla. ^' '^ IM.A'I'K II i'pper, Cluloiiia iiii/da.-^. Transverse section of a choanal papilla taken near its center and stained with hematoxylin and eosin. 250x. Lower, Chelonia mydas. Transverse section of a choanal papilla taken near its base and stained by the Bielschowskv silver method. 2o0x. BREVIORA MiuiseiLim of Comparative Zoology Cambridge, Mass. May 29, 1958 Number 86 A NEW SICISTINE RODENT FROM THE MIOCENE OF WYOMING By Craig C. Black ^ During the course of a study of the Split Rock local fauna of Middle Miocene age, several isolated teeth, two partial maxil- laries, and one fairly complete lower jaw of a new sicistine rodent were discovered. Because of the rather limited occurrence of this group in North America and the paucity of previous records, it is felt that this form should be described before the study of the fauna as a whole can be completed. For the loan of material described below I am indebted to Drs. P. 0. McGrew, Loris Russell and A. E. Wood and to the Chicago Natural History Museum. I would like to acknowledge the helpful assistance and suggestions of Professor A. E. Wood and Professor Bryan Patterson. The following abbreviations are used throughout : A.C. — Amherst College C.N.H.M. — Chicago Natural History Museum E.O.M. — Royal Ontario Museum of Zoology and Paleontology U.W. — University of Wyoming Family ZAPODIDAE Subfamily SICISTINAE SCHAUBEUMYS SABRAE,^ U. Sp. Holotype. U. W. No. 1405, partial left mandible with the incisor. Mi, M2, and the anterior roots of M3. Referred Specimens. Partial maxillaries with P'^-M^ A.C. 1 This study was made while the author was Rufus B. Kellogg Fellow from Amherst College. 2 The specific name is given in honor of my wife. 2 BREVIORA No. 86 No. 10490, R.O.M. No. 1950; Mi A.C. Nos. 10489, 10491, 10493, 10494, 11293, C.N.H.M. Nos. PM 2101, 2102, U.W. No. 1406; Ji C.N.H.M. Nos. PM 2105, 2106; Mi A.C. Nos. 10492, 11294; Mo C.N.H.M. No. PM 2103 ; M3 C.N.H.M. No. PM 2104. Diagnosis. Intermediate in size between Plesiosminthus? clivosus (Galbreath, 1953, p. 99) and Shauheumys grangeri (Wood, 193ob, p. 1) ; no anteroeonid on Mi ; mesostylid closely applied to metaconid on M2 ; mesolophid connected to metaconid on M2 ; hypoconulid well-developed on M1-M2. Horizon and Locality. Middle Miocene. Vicinity of Brachy- crus quarry "Exposure No. 2" (Schultz and Falkenbach, 1940, p. 251), seven miles by road northwest of Muddy Gap filling station, in a draw about 14 i^ils south of U.S. Highway 280, Fremont County, Wyoming. Description. The jaw (Fig. 1 H) is broken through the middle of the alveolus of M3 and down the outer half of the body so that the incisor is revealed beneath Mo. The ridge marking the lower border of the masseteric scar is thick and prominent and extends forward to below the anterior root of Mi. The mental foramen lies almost immediately below the center of the diastema. The upper teeth are associated in several cases with fragments of palate, the most complete of which is shown in Figure 1 A. The posterior boundary of the anterior palatine foramen is on a line with the anterior end of P*. A small part of the zygomatic arch is preserved, arising somewhat anterior to P"^, with the ventral border of the infraorbital foramen preserved. However, only that portion of the foramen which transmits the infra- orbital nerve is present so that no estimate of the size of the opening for the masseter can be made. There are a number of isolated upper incisors (Fig. 1 F, G) which agree so closely with the description given by Schaub (1930, p. 621, Fig. 4) for those of the European sicistine, Plesiosminthus schauhi Viret, that they are here referred to the North American form, Schauheumys sahrae. They are prom- inently grooved with the enamel extending 0.2 to 0.3 mm. over the medial side of the tooth. The pulp cavity is Y-shaped, the median arm of the Y being the longer. The upper fourth premolars (Fig. 1 A, E) are bud-like in ap- pearance with two small buccal cusps, a deep central pit, and a slight lingual swelling. The upper first molars (Fig. 1 A-E) 1958 NEW MIOCENE SICISTINE RODENT are almost square. The paracone and metacone are higher than the other cusps. The auterocone and anterior cingulum are well developed with the latter sloping steeply buccally to join the Figure 1. Teeth of Schaubewmys sairae, all X 10 except for F and G X 5. A, R.O.M. No. 1D50, part of palate with RP^-Mi; B, A.C. No. 10491, LMi; C, A.C. No. 10489, LMi ; D, A.C. No. 10494, LM^; E, A.C. No. 10490, RP4-M1; F, C.N.H.M. No. PM 2105, Ii, lateral side to right; G, the same, lateral view; H, U. W. No. 1405, partial left mandible with Ii, Mi, and M2 (type) ; /, U.W. No. 1405, LM1-M2; J, U.W. No. 1405, Ii, lateral side to right; E, A.C. No. 11294, LMi ; L, C.N.H.M. No. PM 2103, LMo ; M, C.N.H.M. No. PM 2104, RM3. Anterior end to the left except for A, E, G, and M where it is to the right, and F and J where it is at the top. paracone at its base. The paracone and protocone are usually connected through what is probably the protolophule II, either directly, or via the mesocone. In one specimen (Fig. 1 B), 4 BREVIORA No. 86 however, there is no trace of the protoloph. The mesoloph extends to the buccal margin of the teeth and is occasionally connected to the paracone, metacone, or both, by low lateral ridges. The posterior cingulum lies below the level of the meta- loph. The lower incisors (Fig. 1 J) are triangular in outline with a flat anterior face and a small elliptical pulp cavity on the wear surface. The enamel extends slightly over the medial side of the tooth. There is no anteroeonid on the first lower molars (Fig. 1 I, K), but on A.C. No. 11293 (Fig. 1 K) there is a small distinct cuspule in the deep basin between the protoconid and the meta- conid. This cuspule is not on the anterior cingulum, however, and there is no trace of such a cuspule on either of the other two first low'er molars. In all the specimens there is a strong ridge or arm rising steeply from the anterior cingulum to the protoconid while lingually the cingulum abuts against the base of the metaconid Avith no indication of such a cingular arm. In two instances the mesolophid is w^ell-developed in contrast to the distinct mesostylid and mesoconid of U.W. No. 1405 (Fig. II). The hypoconulid is distinct in all cases and closely con- nected to the hypoconid. The posterior cingulum descends steeply from the hypoconulid and passes to the lingual edge of the tooth terminating well below the entoconid. The second lower molars (Fig. 1 I, L) are in general similar to the first molars with the one exception that in the former the mesolophid has shifted anteriorly abutting against the meta- conid. The mesostylid is thus distinct, lying half way down the steep metaconid slope. The protoconid and metaconid are con- nected, probably through the metalophulid II. The mesolophid and metalophulid II enclose a small pit between the protoconid and metaconid. The third lower molar (Fig. 1 M) is proportionately smaller than any of the first or second molars so far encountered, but since so few specimens are available no dependable ranges for tooth size can be determined. The anteroeonid is not distinct, but blends into the two cingulum arms, one of which passes to the lingual margin while the other slopes downwards to the base of the protoconid. The mesolophid passes anteriorly to unite 1958 NEW MIOCENE SICISTINE RODENT 5 with the metaconid at its base, forming a deep pit between the metalophid and mesolophid. The mesoeonid and entoconid are not distinguishable within the crests. The hypoconid is large and is connected to the entoconid region through the posterior cinguluni and the anterior hypoconid arm, the two isolating a posterior pit. There is also a central pit between the mesolophid and hypolophid. Discussion. Schau^eumys sdhrae is the third mid-Tertiary species of sicistiue to be reported from North America. The species previously described are Schaubeumys grangeri (Wood, 193ob) from the Lower Rosebud, and Plesiosminthusf clivosus (Galbreath, 1953) from Quarry A in NE Colorado, of Middle Miocene age. I have assigned the present population to the North American genus, Schaubeumys, rather than to the Euro- pean genus, Plesiosnmithus, because I do not feel that at present there is enough evidence available to warrant considering the North American and European forms congeneric. Galbreath (1953, p. 99) was of the same opinion when, in his discussion of P.? clivosus, he stated, "It seems unlikely that this North American zapodid from the middle Miocene is congeneric with the lower Miocene zapodid of western Europe." I also feel that the three North American species are congeneric and that PJ clivosus should be referred to Schaubeumys. The North American forms can be distinguished on the basis of size: S. clivosus is the smallest; S. sabrae is intermediate; and 8. grangeri is the largest. S. sabrae is strikingly different from either of the others in the absence of the anteroconid on M^ and in the fact that the mesolophid abuts against the metaconid and fails to reach the mesostylid on M2. 8. clivosus appears to have no distinct hypoconulid on any of the lower teeth whereas this cusp is relatively large in the other two species. It is of interest to note that all three species occur in a rela- tively limited time span, no sicistines yet being known in North America from any deposits earlier than Miocene (I am excluding from the Zapodidae the peculiar Simimys from the Sespe Eocene), or later than Middle Miocene. This brings up the problem of where the North American sicistines came from. No definite answer is possible but some guesses may be hazarded. 6 BREVIORA No. 86 All the North American species show a great many similarities to contemporaneous forms in Europe, and to forms known from the Oligocene of Europe and Mongolia. In fact, if it were not for the presence of grooved upper incisors, *S'. sabrae might be referred to the Mongolian genus Parasminthus described by Bohlin in 1946. Parasminthus in turn is equally close to the European Plesiosminthus (Bohlin, 1946, pp. 16-17), from which it is separated only on the basis of the non-grooved incisors. Wood (1935a, pp. 225-226) has pointed out that it is quite pos- sible that the presence or absence of a groove on rodent incisors may be controlled by a single gene. He goes on to state that if this is the case such grooving could be lost and regained within one lineage practically at will. In the case of the sicistines this would mean that the North American forms could be derived from the Mongolian complex and that the European forms could also be so derived, or vice versa. Bohlin (1946, p. 43) points out that the structural differences observed between species of Para- sminthus are of the same kind as those found between the species of Plesiosminthus. This is also true when the North American complex is compared with either Parasminthus or Plesiosmiyithus. These complexes show the presence of the anteroconid in some species and its absence in others. The degree of development of the mesolophid and the position of the mesostylid vary in a similar manner in each group. The way in which the lophs are formed also shows the same variation, and, finally, the same size ranges occur in each group. This great morphological similarity would seem to indicate that these three complexes are basically similar genetically. (No attempt has been made here to determine generic limits within the Miocene sicistines.) TABLE OF MEASUREMENTS ^ 11 Number of Specimens Me.in a c.v anteroposterior 8 1.20 .10 8.3 transverse 8 .74 .05 6.8 P4 anteroposterior 2 .55 transverse 2 .65 1 All measurements in nun. o = Standard Deviation, C.V. = Coefficient of Variation. 1958 NEW MIOCENE SICISTINE RODENT Ml anteroposterior 10 1.32 ,08 6.0 transverse Ii anteroposterior 10 1.29 .10 7.8 1 .85 transverse 1 .60 ^h anteroposterior 3 1.32 transverse 3 1.02 Mo anteroposterior 2 1.28 transverse 2 1.11 Ms anteroposterior 1 1.03 transverse 1 .91 EEFEEENCES BOHLIN, B. 1946. The Fossil Mammals from the Tertiary Deposit of Taben=buluk, Western Kansu. Part II: Simplicidentata, Carnivora, Artio- dactyla, Perissodactyla, and Primates. Palaeont. Sinica, N. S. C, 8b:l-259. Galbreath, E. G. 1953. A Contribution to the Tertiary Geology and Paleontology of Northeastern Colorado. Univ. Kansas Publ., Contr. Paleont., Yertebrata, 4:1-119. SCHAUB, S. 1930. Fossile .Sieistinae. Ecologae Geologicae Helvetiae, 33:616-637. SCHULTZ, C. B. AND C. H. FaLKENBACH 1940. Merycochoerinae, a New Subfamily of Oreodonts. Bull. Amer. Mus. Nat. Hist., 77:213-306. Wood, A. E. 1935a. Evolution and Eelationships of the Heteromyid Rodents with New Forms from the Tertiary of Western North America. Ann. Carnegie Mus., 24:73-262. 193ob. Two New Genera of Cricetid Rodents from the Miocene of Western United States. Amer. Mus. Novit., 789:1-3. BREVIORA MuiseiLinti of Comparsitive Zoology Cambridge, Mass. June 20, 1958 Number 87 AX EMBOLOMERE JAW FROM THE MID-CARBONIFEROUS OF NOVA SCOTIA By Alfred Sherwood Romer It is not always realized that our knowledge of Carboniferous tetrapods is almost entirely confined to fornLs of relatively late age. Apart from a meager fauna from Scotland (ef. Watson 1929) and a single specimen from England, all described Car- boniferous amphibians are of Westphalian or later date — i.e., from Middle to Upper Pennsylvanian strata (Romer 1947, p. 322). The jaw here figured is the first specimen to he described from any American pre-Westphalian formation.^ in 1956 the Museum of Comparative Zoology undertook a sum- mer's exploration of the Carboniferous deposits of Nova Scotia in search of vertebrate fossils. Some excellent material of West- phalian amphibians was obtained, and will be described later. The present specimen is the only pre-Westphalian amphibian identified in our collections. It was derived from the Point Edwards formation (Hyde 1913, p. 252), considered by Bell (1938; 1944, pp. 11-12. 16) to belong to the Canso group. It is definitely earlier than the Riversdale series, with a flora of Westphalian A type, and younger than the Windsor series of the Mississippian. It thus appears to l)e e([uivalent to some horizon in the Namurian of European nomenclature and is currently considered as late Mississippian (Weller et al. 1948, p. 173, chart 6). Exposures of the Point Edwards formation occur only in limited areas adjacent to Sydney Harbor in Cape Breton. They ai-e well shown in the series of maps illustrating the geology of 1 I hoiii' to describe at a later time remains from the Mississippian of West Virginia (ef. Romer 1941). 2 BREVIORA Xo. 87 the Sydney Coalfield which were published by the Canadian Bureau of Geology and Topography in 1938. The best exposures are along- the shores of Point Edward. A Gyracmithus spine was collected there many years ago. Our group in 1956 discovered a beach exposure containing dipnoan and other fish bones about one-half mile southwest of the point. The present specimen was found on the shore about 100 yards southeast of the tip of the ])oint, on a detached slab of rock. An incomplete Gyracanthu.s spine and fragmentary fish remains were found nearby in the ledge from which the jaw slab was derived. On another loose slab was a crushed bone, about 50 mm. long and 29 mm. broad at one end, probably a Sagenodus quadrate. Intensive search of the area failed to joroduce further amphibian material. The specimen consists of a slab of calciferous siltstone which, with its missing counter])art, had contained the greater part of the right lower jaw of a large laljyrinthodont amphiljian. Wave action had removed all the bone, leaving, however, an almost j^er- fect mold of the outer surface, from which excellent casts have been made. The portion preserved extends from the symphysis to a diagonal line running down and forward from a point a short distance l)ehind the termination of the tooth row ; the great- est length as preserved, measured in a direct line, from the symphysis dorsally is 235 nun ; the length of the tooth row is 191 mm. The shape of the jaw, as preserved, is that common to many early labyrinthodonts, particularly embolomeres ; super- position of our fragment on an outline of such embolomeres as Archeria and Ptcroplax suggests a total length of jaw — and hence, approximately, of skull — of about 365 mm. This indi- cates an animal of considerable size, this measurement being within 10-15 ])er cent of that of typical skulls of the familiar Permian labyrinthodont Eryops. If outlines of the jaws of such familiar embolomeres as Fteroplax and Archeria be superposed on that of our specimen by using as a common measurement the length of the tooth row, it will be found that the jaw ramus is intermediate in proportions between the long- and slender Archeria jaw and the relatively deep and short ramus of Ptero- plax. There are few non-embolomere jaws of Carboniferous age with which comparison can be made, and little reliance can be l)laced on propoi-tions of the jaw ramus in attemi)ts to determine 1!)5S EMBULOMEKK JAW FKOAl NOVA SCOTIA 3 relationships. If, howevor, a scries of represontative jaws of labyrinthodoiits bo exainiiicd (ef., for example, Komer 1J)47. figs. 9, 10), it will be seen that in general there is little tendency for the dcveloinnent of an "ascending rannis" in tiie snrangnlar region in teninospondyls, but some trend in this direction in seymonriamorphs, and in embolomeres a strong development of depth posteriorly, with a notable nji-swing of the dorsal margin of the jaw posteriorly. This last trend is obvions in the specimen as preserved, and the edge of the jaw back of the end of the tooth row shows a thin broken margin, indicating that in life the jaw curved upwaril posteriorly still more strongly. This posterior depth strongly suggests that the jaw is that of an anthracosaurian (as I have used that term) and ])ro) ably an embolomere. It is regrettable that the inner surface, Avliere posi- tive embolomere features are to be found is not available. When the cast is viewed dorsally, it is seen that although the mold is slightly incomplete in the symphysial region, the jaw was not widely bowed outward as in such a broad-snouted type as Eryops, but turns sharply backward to run a straight course for almost the entire length of the portion preserved. The surface has an ornamentation of pits anteriorly and ex- tending back along dorsal and ventral margins; tlorsally, the pits give way posteriorly to deeply incised striations. Between dorsal and ventral margins the surface shows a lighter series of antero- posterior striations. At the posterior end of the ventral margin there is part of a deep groove carrying the mandibular lateral- line canal ; farther forward the mold does not extend quite to the ventromedial line of the ramus and hence the anterior pro- longation of the canal is not visible on the cast. The jaw was crushed before fossilization. As is frequently the ease in labyrinthodont specimens (and well shown in an Archeria jaw before me at the moment), the outer surface was crushed downward (i.e. medially) in the area between the thick- ened dorsal part of the jaw carrying the tooth sockets and the stout ventral margin. Sutures are, for the most part, not ap- parent and since we are dealing with a mold only, further devel- opment is impossible. It seems certain that the uncrushed upper margin is entirely, or almost entirely, included in the dentary, and it mav be that the longitudinal line of breakage below this 4 BREVIORA No. 87 reofion is on or close to the suture between the dentary and the infra-dentary series. An area below this line of breakage, in which parallel striae extend forward and slightly downward, appears to be part of the surangular, which thus extends re- markably far forward (as in Kotlassia) . Below this region, posteriorly, is one in which the striae extend diagonally an- terodorsally ; this seems surely part of the angular. The rugose ventral margin was surely occupied by postsplenial and splenial. The pertinence of the wedge-shaped anterior part of the finely striated area is uncertain. A sympliysial tooth pair is present, with both members de- veloped and striated in labyrinthodont fashion. Much of the lateral tooth row is preserved. The most anterior teeth seen on the cast, lying close to the symphysis, appear to be somewhat crowded and irregularly placed. These are conical teeth of modest size, striated basally and slightly recurved at the tips. One complete tooth extends 8 mm. above the jaw margins; an adjacent incomplete tooth appears to have been several milli- meters higher. For 5 cm. back of this region the teeth are l)roken oft', but the alveolar surface of the dentary bone is pre- served. Here there is seen a typical labyrinthodont arrangement of alternate teeth (broken) and empty alveoli; the interval be- tAveen successive elements of the row is about 4 mm. The tooth bases suggest that the teeth are comparable in size to those seen at the anterior end, without the development of any large "fangs." For the next 10 cm. the mold does not extend inward beyond the outer rim of the dentary, so that the alveolar region is not visible and there are impressions of only 5 teeth in all, some of them obviously loose from their sockets. The most anterior of these teetli appears to be comparable to those seen near the symphysis; the others give the impression of greater slenderness, although this may be due to incompleteness of the mold. For the last 3 cm. of the series the alveolar surface is again visible, and 5 of the 6 most posterior teeth are present, tilthough none is complete. Naturally these are smaller than the more anterior teeth, and the interval between successive elements is reduced to about 8 mm. There is little indication that the teetli were markedly recurved or had the chisel-like tijis seen in Archcria. Calculation suggests that the total marginal row con- sisted of about ')() teetli, which were rather closely si)aced and of modest size. 1958 EMBOLOMERE JAW PROM NOVA SCOTIA 5 AVhat arc tlic relationships of tliis specimen? Thei-e are no described American i)re-Westphalian amphibians witli which cominii-ison can be made. It was at one time 1 elieved that the older teti-apod faunas of Europe and North America were as radically distinct as are tlie reptile and amphil)ian faunas today, and tliat trans-Atlantic comparisons were invalid. However, such recent work as my study of the Kounova fauna (1945) and Baird's current (unpublished) studies of the Linton fauna indi- cate that eastern North America and western Europe were ex- tremely similar in faunas in the Carboniferous, and that the supposed contrasts in genei'a present in the two areas are due in "Teat measure to false assumptions as to faunal contrasts and lack of comparison of American and European materials. But even so. the number of pre-Westphalian specimens from Europe, with which this jaw can be lefjitimately compared, are few (Watson lf>29). There are several loxommids from the late Mississippian and early Pennsylvanian of Scotland, but this jaw is (^uite different from that of known loxommids. in which the jaw is shallow posteriorly, and in which the dentition con- sists of a smaller number of much more rol)ust teeth. The straigrht ramus of the jaw indicates an animal with a more or less pointed snout (altliouji'h the skull may have broadened posteriorly, with a triangular sha]^e). This appears to eliminate from considera- tion Otocrat'd and Palaeogyrinus and perhaps Crassigyrinns. The massive upper teeth of Antliracosaurus (in the proper sense) make it improliable that our jaw is pertinent. Two Scot- tish forms remain — Pholidogastcr from the Mississippian (Gil- merton Ironstone), and a skull attributed by Watson to Pholidrrpeton from the Lanarkian of Airdrie. Both have a pointeil snout and, appropriately, a marginal dentition includ- ing a long series of small teeth. Of Pholidogastcr, the type skele- ton (Huxley 1862; Watson 1929, pp. 230-233, tigs. 7-9) is that of an animal much smaller than the present one ; a skull from tlie same beds as the type, and which may l)e a specimen of the same form, is of somewhat larger size, with a length, measured to the (piadrate of about 156 nun. The lower jaws are present of the type specimen ; they do not appear to show the increasing depth posteriorly which is characteristic of this jaw, as of em- holonieres (which PJwIidogaster is not). Of Pholiderpeion, the 6 BREVIORA Xo. 87 type (Huxley 186!)) is of Westplialiaii age, but Watson (1929, 223-224, pi. I, fig'. 2) assigned to this form a skull of earlier date. The latter skull appears to have had a length of approximately 405 mm., a length comparable to that estimated for our speci- men. I have not seen the Manchester Mviseum slab containing part of this specimen but from Watson's description and the Museum of Practical Geology specimen it -would appear that the jaw (definitely of embolomere type) is comparable to that described here. Plence, although the matter lies at two removes from certainty, our specimen may be provisionally assigned to PhoJidcrpcton. It may be reasonably questioned whether a frag- ment of this sort should receive a specific name. iSince, however, it is of considerable interest from stratigraphic and topographic points of view, 1 will, nevertheless, describe it as Pholiderpe- louC^.) hrrtonensis, sp. nov. A specific diagnosis is, of course, valueless, but since this is required by the ''regies," I will men- tion as a supposed sjiecific character the observable gently striated nature of the extei-nal jaw scul]iture between the coarser sculpture bands of dorsal and ventral borders. Tlie holotvpe is M.C.Z. Xo. 2772. Althorgli it now seems probable that many of the Carbon- iferous labyrinthodonts once thought to be embolomeres are members of other groups — rhachitomes, ichthyostegalians, or pre-seymouriamorph anthracosaurians — there nevertheless re- mains a "hard core" of typical embolomeres which carry the story of the group downward through much of the Carboniferous. Archcria is the last survivor; Pteropla.v {'' Eoyyrinus") is a typical Westphalian representative; the skull assigned to PhoJi- derpeton by Watson and the present jaw carry the story down to or across the Pennsylvanian-Mississippian boundary, and undescribed materials from West Virginia show the presence of a true emlolomere in Upper ^lississippian times. The trend within the group appears to 1 e the development of persistently a(|uatic fish-eating ampliibians with an elongated body and tail and short limbs, and, as in piscivorous vertebrates of many groups, a trend toward development of a long slendei- snout and a long row of relatively small marginal teeth. In this first of several projected papers on our Nova Scotia materials, 1 wish to express our thanks to the X'^ational Science Foundatiou, Avhose grant for fiekl expenses made the summer's 1908 EMHOLO.MKIiK .lAW I'I{()M NOVA SCOTIA 7 work {)(>ssil)le, and. further, our gratitude to others who aided and facilitated our work. Dr. II. D. Smith, of tlie Nova Scotia Research Foundation, was most helpful, as was Mr. John Ilaslam of New Glasji'ow. in charge of mining operations at Stellarton and Springhill. Dr. W. A. Bell gave us the benefit of his expert knowledge of the Nova Scotia Carboniferous. To the Nova Scotia Bureau of Mines and the Massachusetts Institute of Technology we owe thanks for generous hospitality at the summer geological camp at Crystal Clitfs, conducted by Dr. Walter 0. Whitehead. LITERATURE CITED Bell, W. A. 1938. Fossil rtoia of Sydney Coalfield, Nova Scotia. Canada, Dept. -Mines, Mem. Geol. Surv., 215: 1-115. 1944;. Carboniferous rocks and fossil floras of northern Nova Scotia. Canada, Dept. Mines, Mem. Geol. Surv., 238: 1-119. IlUXLEY, T. II. 1862. On new labyrinthodonts from the Edinburgh Coal-field. Quart. •Tour. Geol. See. London, 18: 291-296, pi. 11. 1869. On a new labyrinthodont, from Bradford. Quart. Jour. Geol. Soc. London, 25: 309-310, pi. 11. Hyde, J. E. 1913. The Carbonifei'ous sections on Sydney Harbour ; Xlllth Internat. Geol. Cong., Excursion in Eastern Quebec and the Maritime Provinces, Guide Book No. 1, pt. 2, pp. 251-262. ROME», A. S. 1941. Earliest land vertebrates of this continent. Science, 94: 279. 1945. The late Carboniferous vertebrate fauna of Kounova (Bohemia") compared with that of the Texas redbeds. Amer. Jour. Sci., 243: 417-442. 1947. Review of the Labyrinthodontia. Bull. Mus. Conip. Zool., 99: 1-368. Watson, D. M. S. 1029. The Carboniferous Amphibia of Scotland. Palaeont. Hungarica, 1:221-252. Wkller, J. M. (Chairman) and others 1948. Correlations of the Mississippian fornuitions of North America. Bull. Geol. Soc. Amer., 59: 91-196. I h- BREVIORA Miiseimi of Comnparative Zoology Cambridge, Mass. June 30, 1958 Number 88 A NEW SPECIES OF THE GENUS UROTHECA (SERPENTES: COLUBRIDAE) PROM VENEZUELA By J. A. RozE Escuela de Biologia Universidad Central de Venezuela While touring several U. S. museums in 1957 I discovered among the Venezuelan herpetological material in the Museum of Comparative Zoology, Harvard University, a very peculiar undescribed species of the genus Urotheca. Later I found that the U. S. National Museum (USNM), the Carnegie Museum, Pittsburgh (CM), and Museo de Biologia, Universidad Central de Venezuela (MBUCV) also have one specimen each of this form. I take the opportunity to describe this new species. I express my appreciation to Dr. E. E. Williams, Dr. D. M. Cochran and Mr. N. D. Richmond, not only for permitting me to examine the collections at their museums, but also for other kindnesses I received during my stay in the United States. I would also express due thanks to the Fundacion Creole, Caracas, Venezuela, and to the Council Research Fund, American Museum of Natural History, whose help enabled me to take this study tour. Urotheca williamsi n. sp. Leimadophis bimaculatus opisthotaenia, Marcuzzi, 1950, Nov. Cient. La Salle, Ser. Zool., No. 3, p. 8. Holotype. 9 MCZ No. 51329, collected by G. Marcuzzi in October, 1949. 2 BREVIORA No. 88 Type locality. El Junquito, Distrito Federal, Venezuela, 1900 meters. Paratypes. $ USNIVI No. 121206 collected in Colonia Tovar, estado Aragua, Venezuela, 1800 meters, by E. G. Holt on March 30, 1929; $ CM No. 7393, collected in Colonia Tovar, estado Aragua, Venezuela, 1800 meters, by E. G. Holt on April 30, 1929 ; 9 MBUCV No. 3044, collected in Rancho Grande, estado Aragua, Venezuela, 1200 meters, by J. Racenis on August 23, 1949. Distribution. The Central Cordillera de la Costa in northern Venezuela, between 1000 and 2000 meters. This zone is covered by tj'-pical cloud forest (high rain forest). Diagnosis. Urotheca with one scale pit on the dorsals ; 17-18+2 maxillary teeth ; 146 to 158 ventrals ; 53 to 61 subcaudals ; 17-17 dorsals, without any reduction ; 1 or 2 preoculars, and a Leima- dophis-like coloration : dark longitudinal lateral lines beginning on the anterior part of the body and extending to the tip of the tail. Description of the holotype. Maxillary teeth 17+2, the last two somewhat enlarged and separated from the rest by a gap. Mandibular teeth subequal. Head squamation that of Urotheca, rostral scarcely visible from above, nasals divided, the nostrils situated on the dividing suture. A pair of internasals and pre- frontals, the latter a little broader than long and longer than the internasals. The frontal shield twice as long as the prefrontals and approximately as long as the parietals. A loreal shield, deeply penetrating the preocular, and reaching almost to the orbit. One preocular, trapezoidal, twice as wide above as below. Two postoeulars, the upper much larger than the lower. Tem- porals 1+2 on both sides. Seven upper labials, the third and fourth entering the orbit. Nine infralabials, the first four in contact with the anterior pair of chin shields, of which the first pair is longer than the second. Ventrals 150, anal divided, and 53 pairs of subcaudals. The dorsal formula 17-17, without any reduction. Coloration. Head and upper part of the body grey (brown in life?). Upper lip black, crossed by a white line, which begins on the rostral shield and disappears on the neck. Infralabials and other ventral shields white, heavily mottled with black. This mottling is present also on the ventrals, where it concentrates 1958 NEW SPECIES OF UROTHECA 3 mainly along the borders of the ventral scales, but fades gradu- ally posteriorly on body. On mid-line of the subcaudals the black coloration appears again as an irregular ill-defined dark line. The black supralabial coloration continues on the neck as a series of black blotches, which on the anterior third of the body form a lateral black line, at first occupying the whole of the scales on the fourth dorsal row, but more posteriorly shifting to the lower part of the scales of the fourth and upper part of the scales of the third dorsal row. Anteriorly on the mid-dorsum there are irregular black spots, which gradually unite to form two black lines on the eighth scale row on each side. These lines are well defined on the pos- terior part of the body, the zone between them being darker than the general coloration. On the tail the two middorsal lines fuse, so that there are only three black lines (one middorsal and two lateral) on the tail. Total length : 420 mm., tail 85 mm. Notes on the paratypes. The general characteristics are those of the holotype. The total variation of the scale counts in all specimens, including the holotype, is as follows : 146 to 158 ventrals {$ $ 155-158; 9 5 : 146-150), 53 to 61 subcaudals { S $ : 58-61; 9 9 : 53-59), anal always divided, dorsal formula 17-17. Supralabials 7 (3, 4), infralabials are 9 (4) or 10 (5). There is one preocular on both sides in one specimen, two pre- oculars on both sides in two specimens, and one specimen has one on the left and two on the right side. There are 1+2 or l+l-|-2 temporals. The coloration of the paratypes is similar to that of the holo- type, with exception of the MBUCV No. 3044, which seems to be a somewhat more melanistic and aberrant specimen. It has the middorsal black line broader, covering the superior part of the seventh and the whole of the eighth and ninth rows, and the zone between the middorsal and lateral black lines is lighter (pale brown). The MBUCV specimen is darker in general: the mottling on the underside of the head being almost black and the dark streak on the middle line of the subcaudals being darker. 4 BREVIORA No. 88 The CM specimen has the following hemipeneal characters : The sulcus spermaticus is bifurcated, the bifurcation taking place rather close to the base. At the base are several large spines or hooks. About 12 longitudinal rows of larger spines run from the base to the place where the bifurcation of the hemipenis occurs, each row containing 5 to 7 larger spines ; among them are placed several smaller spines, which eventually may form addi- tional longitudinal rows, between the big rows. At the apex, where the sulcus ends, there is a calyculated zone, formed by several larger and smaller calyces. Generic position. This is one of the most complicated cases in the already sufficiently complicated Leimadophis-Liophis-TJ roth- eca, etc. group of genera. This species has the following charac- teristics of the genus Urotheca: 1. Absence of any reduction in the dorsal squamation. 2. Calyculated and simple bifurcated sulcus of the hemi- penis, as well as the presence of large spines or hooks at the base. 3. Dark lip coloration, with a white line crossing it. It has, on the other hand, these characteristics of the genus Leimadophis: 1. The presence of one scale pit on dorsals. 2. The general coloration, i.e., a longitudinal dark lateral and dorsal line beginning to form at midbody, occupying all the posterior part of the body and the tail. The maxillary dentition is in general intermediate between the two, although with a slight predominance of the Urotheca-like characteristics, i.e., there is a gap, but it is rather small, and also the two last teeth following the gap are somewhat smaller than usually is the case in Leimadophis. The hemipenis and the dorsal scale formula were the charac- teristics which caused me to put this new species in the genus Urotheca, although the presence of one dorsal pit has been con- sidered as a very decided Leimado phis-like character. However, on several occasions I have observed the absence of pits in speci- mens of several species of Leimadophis (L. mclanotus, L. regiyiae), and I have found scale pits occasionally present in specimens of Lygophis and — more important — in Liophis. It seems that the presence or absence of scale pits in this group of genera and occasionally in others (cf. Amaral's genus Barhourina!) can ]f)58 NEW SPECIES OF UHOTIIECA 5 be used only as an auxiliary generic character. Moreover, scale pits are present in practically all reptiles, although usually so small that they are not seen except under high magnification. The problem thus arises : when are they sufficiently visible to be useful as a character? As to the question of Vrotheca vs. Rhadinaea, there can be little doubt, as pointed out by Dunn, 1957, that the earlier name should be used. Vrotheca dumerilii, the genotype, apparently is not a Cuban form, as originally presumed, but most probably a species from Central America (the description suggests this very strongly), described later as a species of Rhadinaea {R. pachyura fnlviceps). On this view, the genus Rhadinaea Cope, 1863, is a synonym of Vrotheca Bibron, 1843. To use Vrotheca will certainly simplify the taxonomy greatly, since in the past various authors have distributed species to one or the other genus, without any clear concept of the differences between them. The holotype and the MBUCV specimen have been reported by Marcuzzi, 1950, as Leimadophis himaculatus opisthotaenia, but this is an error, since the latter species is a form from the Andes of Merida, Venezuela, and is a true Leimadophis (I have seen several specimens of that form), and very distinct indeed from V. ivilliamsi. I take pleasure in naming this species for my friend and col- league. Dr. Ernest E. AVilliams, from whom I received much kind help and cooperation, while staying at the Museum of Comparative Zoology. BIBLIOGEAPHY BOULENGER, G. A. 1908. Descriptions of new South American reptiles, Ann. Mag. Nat. Hist., (8) 1: 111-115. Dunn, E. R. 1957. Neotropical frog genera: Prostherapis versus Hyloxalu.s with remarks on Phyllobates. Copeia, 1957 (2) : 11-18. .Marcuzzi, G. 1950. Ofidios existentes en las colecciones de los museos de Caracas (Venezuela). Nov. Cient. La Salle, Ser. Zool., 3: 1-20. I BREVIORA Mimseiuiinii of Comparative Zoology Cambridge, Mass. July 15, 1958 Number 89 REMARKS ON SOME FORMS OF CINCLUS (AVES) By James C. Greenway, Jr. and Charles Vaurie The genus Cinclus was studied by us independently. Green- way studied the collections in the Museum of Comparative Zool- ogy and in the American Museum of Natural History in prep- aration of a part of the Check-list of the Birds of the World. Vaurie studied the collection of the American Museum in preparation of a cheek-list of the Palearctic avifauna (1958, in press), and published some notes on this genus on two occa- sions (1951 and 1955). Greenway deferred publication. Later he sent the draft of his paper to Vaurie, who in 1957 examined the material in several European museums. We have decided to present our conclusions jointly. Cinclus cinclus Two main types are represented on the European continent: one which has a blackish breast, and the other in which the breast is several shades of brown. Some populations are con- stant or relatively constant, as in Scandinavia and northern Europe where the birds (nominate cinclus) are blackish but other populations vary individually. The latter may consist of individuals that are blackish and very similar, if not identical, with nominate cinclus or of individuals in which the color of the breast varies from dark and dull brown to rather bright rufous. In some regions, as in the Alps and to some extent in Spain, the birds inhabiting a watershed may be blackish, whereas in a contiguous watershed they are brown and these blackish, or 2 BREVIORA No. 89 brown, populations may resemble other populations from which they are widely separated. For instance, blackish birds inhabit Scandinavia and northwestern Spain and brown birds inhabit Germany and southern Spain. On the other hand, the popula- tions may grade more or less smoothly into one another, as from southeastern Germany southeastward through central Europe. These complicating factors render a division for nomencla- tural purposes somewhat arbitrary and Vaurie was, at first, in- clined not to recognize any subspecies. However, all the birds of Scandinavia and northern Europe are blackish and identifiable at a glance from the brown-breasted ones from farther south, and the geographical variation, though not constant in central and southern Europe, shows several trends. In the end, Vaurie (1955) recognized five subspecies. These five subspecies are also widely recognized by conservative authors and are as follows : nominate cinclus Linnaeus, 1758, type locality, Sweden ; aquati- cus Bechstein, 1803, type locality, Germany; pyrenaicus Dresser, 1892, type locality, Pyrenees ; meridionalis Brehm, 1856, type locality, Carinthia, Austria; and orientalis Stresemann, 1919, type locality, Macedonia. There is, however, little agreement about the limits of their ranges. Hartert (1910) stated that the range of nominate cinclus was Scandinavia and western Russia south to East Prussia ; Witherby (1922, p. 341) added the Cantabrian Mountains of northwestern Spain, and in 1928 (pp. 618-620) the mountains of central Spain; while Vaurie (1955) referred all the birds of the Iberian Peninsula to pyrenaicus which Hartert had re- stricted only to the Pyrenees. Hartert (1910) stated that the birds of France were aquaticus (except in the Pyrenees and in the southwest where they are replaced by meridionalis) but Mayaud (1953, p. 49) includes Brittany and the Massif Central of France in the distribution of pyrenaicus, stating that in the southeast the populations show a tendency toward meridionalis. Hartert (1922) recognized orientalis for southeastern Europe, a region which he had formerly included in the range of meridion- alis, while Rokitansky (1939) extended the range of orien- talis westward to Moravia in central Europe. Finally, von Burg (1924) and Troller (1935) described six new forms from the 1958 REMARKS ON CINCLUS 3 Alps, and Floericke (1926, Mitt. Vogelw., p. 78. Non vid.) one from northern Portugal which he named atroventer. The latter, the type locality of which is Serra do Gerez, is a very black form. It is evident from the foregoing that students of this group have often disagreed (especially regarding the classification and distribution of the populations of France, the Alps, Balkans, and Iberian Peninsula), and Greenway came to the conclusion that it would be best to recognize only two subspecies : a blackish one (nominate cinclus), and a brown one (aquaticus) . He was aware that the range of the black form is discontinuous, and also that a more realistic treatment might include but a single subspecies. Black and brown forms are, however, separable at a glance. Vaurie, after studying the material in the European collec- tions, now shares the opinion of Greenway. "We consider that it is best to synonymize pyrenaicus, sapswortJii, atroventer, and amphitryon with nominate cinclus, and to synonymize meridion- alis, orientalis, olympicus, and the forms described by von Burg and Troller, with aquaticus. SapswortJii was described by Arri- goni in 1902 from Corsica but we find that good series in New York and Europe from Corsica and Sardinia are not separable from specimens from the Pyrenees. Olympicus Madarasz, 1903, type locality, Cyprus, and amphitryon Neumann and Paludan, 1937, type locality, Lasistan, northeastern Turkey, will be dis- cussed below. The populations of Austria and Italy (meridionalis) , or from the Balkans (orientalis), vary too much individually to warrant nomenclatural recognition as about one third to one half or more of the specimens examined cannot be distinguished from typical aquaticus of Germany. Professor Stresemann, with whom Vaurie has discussed this question, is now also of the opinion that orientalis is not sufficiently constant and is best synonymized with aquaticus. The subspecific status of the birds of the Iberian Peninsula and Pyrenees is the least clear and has been the subject of most com- ment. The result of Vaurie 's examination of relevant material in European collections is set forth below ; that in New York was discussed in 1955. Eleven specimens from the Pyrenees support the statement made by Vaurie in 1955 after he had examined 25 birds that are 4 BREVIORA No. 89 virtual topotypes of pyrenaicus. As stated then, this popula- tion varies individually, but "in series . . . differs from nom- inate cinclus by being not so black, browner above and below including the crown and nape, which are distinctly paler, and in fresh plumage by having the gray edges of the feathers of the back better developed." Whistler and Harrison (1930, p. 467) stated that their three specimens from the Pyrenees ''cannot be separated from Swedish topotypes, ' ' but much larger series show that Witherby (1928) is correct when he states that the birds of the Pyrenees though "very much like" nominate cinclus are "not quite so black." Witherby recognized pyrenaicus but we believe it is best to synonymize it with nominate cinclus, as it is impossible to separate the birds of the Pyrenees from specimens which, in series, are even more similar to nominate cinclus. Nine specimens from the province of Santander are very similar above to the birds of the Pyrenees, including the prom- inent gray edges, but in series average a little more blackish, especially below. Three specimens from Riano, northern Leon, are still blacker. These three match four from the Serra do Gerez in northern Portugal (topotypes of atroventer), one from Serra da Estrela, also in northern Portugal, three from Candeleda, Sierra de Gredos, Avila, and twelve from Lagunilla, Be jar, Sala- manca. All these are very blackish below and similar to nominate cinclus but not quite identical. They average deeper black below but above they are more similar to the birds from Santander and the Pyrenees, the crown and nape averaging slightly more rufous and the gray edges of the feathers being, as a rule, better de- veloped than in nominate cinclus. However, individuals are found in northwestern or central Spain that are not similar to nominate cinclus. One specimen from Santo Domingo de Silos, Burgos, and also one from Can- deleda are browner. Of three specimens from San Ildefonso la Granja on the northern side of the Sierra de Guadarrama, one is similar to the black birds from Riaiio, though very slightly browner, and the other two are a little browner still. One from Cercedilla, on the southern side of the Sierra de Guadarrama, is distinctly browner than the foregoing and Witherby (1928) 1958 REMARKS ON CINCLUS 5 remarked that it is intermediate between nominate cinclus and aquaticus, adding; that one from nearby Escorial "is like aquatimt^." This last specimen is now in the collection of the American Museum of Natural History and was examined by Vaurie in 1955 tog-ether with two from the Sierra Nevada in southern Spain. These three are identical and not separable from aquati- cus. Five other skins, examined in 1957, from the Sierra Nevada are also very brown and identical with aquaticus. In short, it appears that the population of southern Spain is constant and not separable from aquaticus. The birds of central Spain and the Pyrenees vary individually but, taken as a series, these populations are closer to nominate cinclus than they are to aquaticus, while in northwestern Spain and northern Portugal the populations seem constant again, and, though not quite iden- tical, are best called nominate cinclus as they are so similar to it. Two of the additional specimens from the Sierra Nevada were collected in 1955 by Dr. G. Niethammer and are in the collection of the Bonn Museum. This museum possesses also the type and paratypes of amphitryon. This series is identical with nominate cinclus in coloration, as stated by Neumann and Paludan (1937), who separated it from the latter on the basis that it had a ' ' some- what shorter wing and an almost imperceptibly thinner bill" (trans.). The bill shows a tendency to be more slender, but the difference is extremely slight and not of taxonomic importance, and the wing measurements show too much overlap to warrant the recognition of amphitryon. According to Vaurie, the wing length measures 93, 93, 94, 94 (+, molting), 95, 98, 98 (95) in the males from northeastern Turkey, as against 92, 94, 95, 95, 95, 96, 96, 97, 98, 98, 99, 100 (96.3) in males from Scandinavia. Two other forms {olympicus from Cyprus, and uralensis Serebrovski, 1927, type locality, Urals) require brief comment though we are handicapped by the lack of adequate material. The material that we have seen from Cyprus consists of three rather worn and not too well prepared specimens collected in the spring of 1906. Vaurie (1955) stated he could match them with specimens of orientalis, but in Greenway's opinion they resemble caucasicus rather than Macedonian birds. Hartert (1910), who had the same specimens, stated that he "was not 6 BREVIORA No. 89 able" to separate them from caucasicus but cautioned that more and better prepared material was necessary from Cyprus. Never- theless, he sjaionymized olynipicus with caucasicus, though with a query. The name olympicus has remained a synonym of cau- casicus ever since but, unfortunately, olympicus was described in January 1903 as against December of the same year for caucasicus; it should replace the latter if indeed the populations of Cyprus and the Caucasus are not separable. Fresher and more abundant material from Cyprus is necessary, however, to establish beyond dispute that the two forms are not separable. We therefore recognize provisionally the validity of olympicus. Serebrovski (1927) has separated the population of the Urals as uralensis, stating that it differs from nominate cinclus by being paler and more brownish above and below, but is darker above and less rufous below than aquaticus. The validity of uralensis has been recognized by Hartert and Steinbacher (1935), Dementiev (1935), and Portenko (1937), but Sudilovskaya (1954) considers that uralensis is a synonym of nominate cin- clus. Her opinion, however, may be extreme as she also synony- mizes caucasicus (which in our opinion is perfectly valid) with nominate cinclus. The only specimen of uralensis that we have seen is not fully adult but shows the differences mentioned by Serebrovski when compared to specimens in the same plumage. The population of the Urals is very widely isolated from all the other populations of the species and may well be distinct. Cinclus pallasii As Vaurie (1951) has remarked, there are four specimens of the Brown Dipper, in the collection of the American Museum of Natural History, having gray rumps contrasting with brown backs. These were taken in Amurland and near Mukden. It would perhaps be possible to separate on the basis of such material the populations of extreme eastern Siberia and Japan, on the one hand, from those of China and Formosa, on the other. The probability is strong that this one difference in color of rump is due to a single gene, however, and a theory that a single popu- lation inhabits eastern Asia, Japan, Formosa, and northern Indochina is probably the most correct one. To be sure, there is 1958 REMARKS ON CINCLUS 7 a small size difference, as is shown in Table 1, below. That this is significant may be doubted. Of a series of 15 kinds, 6 are intermediate. Application of a coefficient of difference test shows a 20 per cent overlap. Table 1 WING (in nun.) ■i III aria 11(1 China Fiinnii.sd 94.0 1 104.0 1 96.0 1 102.0 1 105.0 1 102.0 1 103.0 1 106.0 1 103.0 1 106.0 1 107.0 108.0 1 1 105.0 1 109.0 9 Efx=: 406.0 110.0 3 n = 4 111.0 1 m = 101.50 114.0 1 = 3.87 120.0 1 121.0 1 Japanese populations are intermediate to some extent. Of a series of 13 specimens from Japan and Amurland, 5 are impos- sible to diagnose as belonging- to one population or the other. In other words, only 62 per cent can be recognized on the character of the gray versus brown rump. Certain Japanese specimens have a somewhat more reddish tinge than those from the Asiatic mainland, but only 65 per cent of the series at hand can be distinguished by this character. Vaurie (1955), recognized only two subspecies: nominate pallasii Temminck, 1820, type locality, "Crimea," but error for Okhota River, eastern Siberia, as shown by Stresemann (1948, pp. 115, 126) ; and tenuirostris Bonaparte, 1850, type locality, central Asia. Among the forms synonymized with nominate pallasii were wilderi La Touche, 1925, type locality, Hopeh, and dorjei Kinnear, 1937, type locality, eastern Bhutan. 8 BREVIORA No. 89 The two co-types of wilderi from the Eastern Tombs near Peking have brown rumps and do not appear to differ in any way from birds of western and southern China (see above). The affinities of dorjei were not made too clear by Kinnear (1937, p. 263), and Professor Stresemann kindly pointed out to Vaurie that the latter was probably wrong in considering that dorjei was not separable from nominate paUasii. In Stresemann 's opinion dorjei was probably a valid race, or, if not, a synonym of tenuirostris rather than of nominate pallasii. Examination of the original series shows that Stresemann is correct and that dorjei is valid. Its diagnosis and range, as given by Vaurie (1958) are: "Darker than tenuirostris in all plumages but paler than nom- inate pallasii. This race, though intermediate in coloration, is closer to tenuirostris. Range : Eastern Himalayas, east of ten- uirostris [which ranges eastward about as far as northern Sikkim where it grades into dorjei']., hills of Assam south of the Brah- maputra, and higher hills and mountains of Burma south to the Chin Hills in the west and the Shan States in the east, to north- ern Siam and probably to neighboring northern Yunnan though there are no records." CiNCLUS SCHULTZr Cincliis schultzi Cabanis of the mountains of northwestern Argentina has been treated as a distinct species by all authors. In view of the intraspecific variation in other parts of the world it would appear to reflect the probable biological truth better to treat it as a subspecies of leucocephaliis. Compare Cinclus cinclus leucogaster of Afghanistan with caucasicus and cash- meriensis, for example. The brown breast of schultzi replaces the white of the more northern form (leucocephalus) . The w^hite wing bar of schultzi is not consistently of greater extent. It is a grayer bird and a very distinct subspecies. 1958 REMARKS ON CINCLUS 9 ACKNOWLEDGMENTS Vaurie received much friendly cooperation during his recent visit to Europe and takes pleasure in thanking the curators of the various museums and their staffs for much help. Professor Stresemann kindly discussed several of the forms mentioned in this paper, giving him the benefit of his great experience, and Mr. Macdonald of the British Museum sent to New York the specimens of dorjei for further study. LITERATURE CITED Anonymous [= von Burg] 1924. Die schweizerischen Wasseramseln. "Der Weidmann" (Biilach- Ziirich), 6(24): 7. Dembntiev, Georges P. 1935. Systema A\dum Rossicarum. L'Oiseau, p. 456. Hartert, Ernst 1910. Die Vogel der palaarktischen Fauna. 1(6): 788-795. 1922. Op. cit., 3 -.2172. Hartert, Ernst and Friedrich Steinbacher 1935. Die Vogel der palaarktischen Fauna. Erganzungsband, p. 344. Kinnear, N. B. 1937. In Ludlow and Kinnear, The birds of Bhutan and adjacent territories of Sikkim and Tibet, pt. 2. Ibis, 249-293. Mayaud, Noel 1953. Liste des oiseaux de France. Alauda, 21(l):l-63. Neumann, Oscar, and Knud Paludan 1937. Zwei neue geographische Rassen aus Klein-Asien. Orn. Monatsb., 45:15-16. Portenko, L. a. 1937. Fauna Ptits Vnepoliar. Chasti Severnogo Urals. Moscow-Lenin- grad. Akad. Nauk SSSR, p. 164. Rokitansky, G. F. von 1939. Zur Verbreitung der Wasseramseln, spezieU der Rasse Cinclus cinclus orientalis Stres. Ann. Naturhist. Mus, Wien, 49:282-294. 10 BREVIORA No. 89 Serebbovski, p. 1927. New races of Palearctic birds. Doklady Akad. Nauk SSSE, 1927 (A):325-326. Stresemann, Erwin 1948. Dr. 0. H. Mercka ornithologischc Aufzeichnungen wahrend der Billingsschen Expedition von Oehotsk naeh Alaska (1787-1791). Zool. Jahrb., Abt. fiir Syst., 78(1) : 97-132. SUDILOVSKAYA, A. M. 1954. In Dementiev and Gladkov (editors) ; Ptitsy Sovietskogo Soiuza. Mosco-w, 6:672. Troller, Julius 1935. Die Wasseramsel, Cinclus cinclus L., und ihre Unterarten in der Schweiz. Areh. Suisses d'Ornith., 1(6) :193-224. Vaubie, Charles 1951. Notes on tlie wrens and dippers of western Asia and India. Amer. Mus. Novitates, 1485:9-19. 1955. Systematic notes on Palearctic birds, no. 16. Troglodytinae, Cinclidae, and Prunellidae. Amer. Mus. Novitates, 1751 : 12-17. 1958 (in press). The birds of the Palearctic Fauna, part 1. Passeri- formes. London, H. F. and G. Witherby. Whistler, Hugh, and James M. Harrison 1930. Some auturnn observations of the avifauna of the western and central Pyrenees. Ibis, 453-470. Witherby, H. F. 1922. Kesults of a collecting trip in the Cantabrian Mountains, north- ern Spain. Ibis, pp. 323-345. 1928. On the birds of central Spain with some notes on those of south- east Spain. Ibis, pp. 385-436, 587-663. BREVIORA Miuiseimm of Connparsitive Zoology Cambridge, Mass. July 30, 1958 Number 90 A FOSSIL VAMPIRE BAT FROM CUBA By Karl F. Koopman Academy of Natural Sciences of Philadelphia, Philadelphia, Pa. Of the nine families of bats in the Western Hemisphere, all but three (Desmodoutidae, Fnripteridae, Thyropteridae) have been known from the West Indies for some time, althongh the Em- ballonuridae reach the West Indies only in Grenada. The precise ecology and food habits of the Thyropteridae and Fnripteridae are poorly known so that no explanation for tiieir absence othei than the presence of water barriers is apparent. The Desmodon- tidae, on the other hand, are the vampire bats, Avhose blood diet is well known. The absence of this family from the West Indies might, therefore, be explained by the paneity of native, non- flying, land mammals of which in the Greater Antilles, only five genera are known, none larger than a cat. Chicken-sized or larger ground birds are also absent. Thus it is possible that at the time of arrival of man in the West Indies, there were no animals present which could support a population of vampire bats. For a number of years, however, it has been known that during the late Pleistocene, several genera of large and medium-sized mammals existed in the Greater Antilles. Four genera of ground sloths are known from Cuba, while elsewhere in the Greater An- tilles, both ground sloths and large rodents are know^n. It there- fore appears that at this time a suitable habitat for vampire bats did exist. Until now, however, none of the fossil deposits in which these larger forms are found, at least in Cuba, has yielded bat remains. 2 BREVIORA No. 90 During: the past IS years, the Speleological Society of Cuba has made extensive excavatious in numerous caves, chiefly in western and central Cuba, uncovering abundant mammal re- mains, including a great deal of ground sloth material (Ar- redoudo 1955). One of the richest series of fossil deposits was in the L'lmas cave, a few miles west of Havana. The following statement was prepared for me by Sr. Oscar Arredondo of the Sociedad Espeleologica de Cuba, Havana, and Sr. Gilberto Silva Taboada, a student of recent bats, also of Havana. "The Lamas Cave is merely a rocky .shelter located in a land slope to the southwest of Santa Fe Beach, on the northern coast of Habana province. On January 21, 1951, an excavation was made at this cave by Oscar Arredondo, Director of the Paleonto- logical Section of the Speleological Society of Cuba, together with Dr. Manuel Rivero de la Calle, from Las Villas Central University. "According to the stratigraphical profile, the flrst level was a layer of very dark earth. 60 centimeters in depth, containing jilenty of shells and human bones pertaining to the oldest Indian culture of Cuba (the Cuanahatabeyes), together with skeletal remains of Capronnjs, Gcocaproiiiys, and Boroniy.s offella. The second level, extending about 40 centimeters in depth, was com- posed of a red earth completely lacking any bones or areheological remains. The third level was a 30 centimeter deep layer of a light yellow colored earth, containing bones of small extinct rodents, like Geocapromys. "The fourth level, having a depth of 1.40 meters and com- posed of a humid earth, reddish-yellow in color, Avas the richest fossil-bearing layer of the trench. Thousands of scattered frag- mentary bones and skulls, representing almost all the known extinct Cuban fauna, were unearthed from this level. Here was found a fragmentary bat skull in association with the extinct rodents Gcocapromxjs columbianus and Boromys torrei, the in- sectivore Nesophontes micrus, the ground sloths Megalocnus rodens and Mesocnus iorrei, an unknown bird allied to the ex- tinct Phororacos'^ of Argentina, and also Testudo cuhensis and the living Crocodylus rhomhifer." 1 This iilentiflcation may be erroneous. l})r)8 FOSSIL VAMPIRE HAT KKOM CUBA 3 AVe are licre eoneernod with the richest, deepest, and therefore oldest, of tile four layei's. This, it may l)e noted, consists entirely of extinct species, at least as far as the inanunals ai-e concerned. I can verify the identity of at least the Ncsuphoiitrs, Boroinys, Geocapronujs, and "Testudo" {Geochelone), since Sr. Arredondo has very generously turned ovei- to nie a small sample of the fossils from this layer. The assemblaoe probably indicates a truly Pleistocene level and tlie presence of <>round sloths [Megolocuus, Mesocnus) and of large ground birds (whatever their at^nities may be) indicate a very ditferent habitat, in terms of ability to support vampire bats, from that which existed in late pre-Columbian times. As mentioned above, among the numerous fossil bones of this deposit, there was found the rostral portion of a bat skull, which Sr. Gilberto Silva Taboada kindly presented to me. It is now No. 4737 in the Vertebrate Paleontology collection of the Mu- seum of Comparative Zoology. In this specimen, the very char- acteristic vampire dentition is immediately apparent, togethei- with the peculiar rhombic shape of the palate characteristic of this group of bats. Comparison with the three living species of the family Desmodontidae shows that the fossil is clearly refer- able to Desmodus rofiDidus, the commonest mainland species. Unfortunately the condition of the skull permits few measure- ments to l)e made on it (Plate). The following are the only standard ones I was able to make. In each case, the figures in parentheses refer to a series of seven specimens of Disuwdus rotKiidiis i)ntri)U(s from El Pachon, Tamaulipas. Measurements are in millimeters. Least interorbital width — 5.1 (4.5-5.5) Length from anterior edge of incisor to posterior edge of canine — 4.9 (4.6-4.9) Width across base of incisors — 3.9 (3.3-3.7) Except for the slightly greater width across the incisors, the Cuban fossil appears to be indistinguishal)le from the Middle American subspecies, DcsDiodus rofuiidits niuruius. We may assume that Desmodus reached Cuba after prey w^as available in the form of ground sloths and large ground birds, and that when these forms became extinct and thus the ecological 4 BREVIORA No. 90 niche for vampires disappeared, the latter also died out in Cuba. The introduction of domesticated animals in post-Columbian times has presumably reopened the niche, but, at least so far, vampires have failed to ^et acro.ss the fairly wide water bar- riers and hence do not form part of the living bat fauna of Cuba. In conclusion, I wish to acknowledge the help very graciously given me by the two above-mentioned Cuban naturalists, 8r. Oscar Arredondo and Sr. Gilberto Silva Taboada. LITERATURE CITED Arredondo, O. 1955. Resumen de las Ac-tividades paleontologk-as realizada.s por la Seecion de Geologia y Paleontologia. Bol. Soc. Espeleologica dc Cuba. Alio I (Deeiembrc), pp. 3-31. tt Palatal views of J)csnw(hii-: skulls. Top, the Cuban fossil, MCZ (Verteliiato Paleontology) 4737. Bntfom, D. rotundus $, MCZ (Mammalogy) 10445 from Bogaba, Panama. BREVIORA MmseiJiitTii of Coiraipsirative Zoology Cambridge, Mass. August 13, 1958 Number 91 CONTRIBUTION TO A REVISION OF THE EARTH- WORM FAMILY LUMBRICIDAE. II. INDIAN SPECIES By G. E. Gates INTRODUCTION The domain of the ' ' recently evolved and dominant ' ' Inmbricid family, a group which has "almost completely conquered the Palaearctic" (Stephenson, 1930, p. 668), long has been thought to reach well into India. The southern boundary of that domain, on recent distributional maps (for example those of Tetry, 1938, and Julin, 1949), runs from the Mediterranean so as to include the Punjab portion of the Sind Valley and all of the Gangetic Valley to Calcutta from whence it is extrapolated, with question marks or predictions, across northern Burma and through China to the Pacific. Two routes from European centers of lumbricid evolution to the vicinity of Calcutta could have been available to the advancing lumbricids : through the Himalayas in an easterly direction and then down to the plains, or south along the Indus River and thence easterly through the Gangetic Valley. In either case, lumbricids could be expected throughout an eastern portion of the Gangetic plain as well as in the western part of the Himalayas. A survey of the earthworm fauna of the Allahabad sector (Gates, 1945, 1947) and collecting in such Himalayan sites (Gates, 1951) as could be visited during wartime, provided data that made a review of the taxonomic status of the supposedly endemic species advisable. The results of that study are presented herewith. The text of the present contribution, except for this paragraph, has been excised from a chapter of an unpublished manuscript written several years ago during tenure of a John Simon Guggenheim fellowship. 2 BREVIORA No. 91 SYSTEMATICS Genus AlLOLOBOPHORA Eisen, 1874 " Allolobophora caliginosa (Savigny, 1826) " A complex of four or more morphologically distinguishable species long has gone under Savigny 's name. The data aA'ailable in the literature rarely permit a decision as to which species was present in a particular collection. Indian localities from which "caliginosa" has been recorded are: Fern Hill and Ootacamund (at elevations of 6-8,000 feet in the Nilgiri Hills, South India), Mt. Abu (Rajputana), Lahore, Ferozepur, Kotla, Murree, Peshawar, Mardan, Chitral (Punjab and Northwestern Frontier Province), Gilgit, Gurez, Gundarbal, Anchar Lake (Kashmir), Almora, Mussoorie, Naini Tal, Simla, Kufri, Junga, Kasauli (western Himalayas). All except the Nilgiri localities are well within the temperate zone. None of the complex ever was found at Darjiling where exotic lumbricids now appear to be dominant (Gates, 1951). There are no records for Ceylon, Burma, the Andaman and Nicobar Islands, the Malay Peninsula, and other parts of southeast Asia, not even for hill stations where intro- duction might have been expected. The characterization "cosmopolitan" so frequently applied to "caliginosa" obviously implies more than has been warranted by the facts, even if the complex had been but a single species. AlLOLOBOPHORA jassyensis Michaclsen 1891 This little known species has been reported from a single Indian locality well north of the thirty-second parallel in the Murree subdivision of the Punjab. Individuals of the caliginosa complex sometimes have been mistaken for jassyensis. Presence of the latter in India, if confirmed, presumably is due to importation by man. AlLOLOBOPHORA TRAPEZOTDES (DugCS, 1828) Helodrilus {Helodriliis) mariensis Stephenson 1917, Rec. Indian Mus., Cal- cutta, 13, p. 414, fig. 6. (Type locality, Murree. Types, several, pre- sumably in the Indian Museum.) Allolohophara (Eophila) mariensis, Stephenson, 1923, Oligochaeta, in Fauna of British India, Ceylon and Burma, London, p. 504. 1958 REVISION OP INDIAN LUMBRICIDS 3 Stephenson's species is known only from the original descrip- tion. Subsequent collections at the type locality provided only specimens that have been referred to the caliginosa complex. The types probably were brittle and strongly contracted, as female and spermathecal pores were unrecognizable. A number of differences, of a minor sort, apparently were thought to ob- viate consideration of caliginosa. The greenish coloration may have been an artifact. Shape of the posterior portion of the body (shown in fig. 6) may have resulted from unusually strong contraction at preservation and from dehydration in too strong alcohol. A similar condition of the posterior end has been seen in a number of specimens of trapezoides recently examined. Location of the first dorsal pore at 4/5 is of little importance taxonomically. A functional pore has been found in trapezoides as far forward as 6/7. Just which ones of a series of more or less pore-like markings at mD really are patent apertures through the body wall into the coelomic cavity often is undiscoverable in field-preserved material. Absence of mention of tubercula pu- bertatis in Stephenson's description presumably means that those structures (as often is the case in browned or other alcoholic museum material) were quite unrecognizable. The crypts or yellowish projections from the oesophagus in x, that were men- tioned by Stephenson, are, of course, calciferous sacs. The supposed hearts of xii are only the vertical portions of the extra- oesophageal trunlcs (cf. 0. cyaneum below). Spermatophores have not been found externally nor sperma- tozoa in the spermathecae of any of a considerable number of specimens of trapezoides that have been available from various parts of the world. The male funnels usually have shown no indications of presence of mature sperm. Male sterility requires that reproduction be parthenogenetic and that permits reduction in number as well as in size of the seminal vesicles. Most liable to early elimination after reproduction becomes asexual are the vesicles of x and ix. Absence of those vesicles certainly is no justification for assigning a species to Relodrilus or to Eophila, whatever those two genera may prove to be. In Allolohophora, where the species undoubtedly belongs, there is no good reason for specific distinction from trapezoides. 4 BREVIORA No. 91 Allolobophora tuberculata Eisen 1874 Among specimens from Simla that were originally (Gates, 1951) referred to caliginosa are a number that can now be placed in Eisen 's species. Synonyms are: similis Friend 1911, arnoldi Gates 1952, and caliginosa (various authors). Elsewhere, the species now can be recorded from the United States (northeast- ern portion), Canada, England, Chile, Denmark, Norway, Sweden and probably Germany. The distribution, even as now known, must have resulted in part from transportation. Obviously there can be no imputation of endemicity in India. Genus BiMASTOS Moore, 1893 BiMASTOS PARVUS (Eiscu, 1874) Jalla, Patna, Bihar, March, 0-0-1, E. C. Lacy. Rawalpindi, Punjab, May, 0-0-6. F. G. Dickason. Forty -four specimens of this species were collected at Allahabad during a period of four years. More than half were found in dirt around roots of potted plants. The others were obtained from a single small area (sansiharicus habitat) by a drainage canal that was sampled at least once a month through most of the period (Gates, 1945, 1951). The sansiharicus habitats, in the region around Jubbulpore to the south of Allahabad were in- vestigated (Gates, 1956) on several occasions during that same period but yielded no specimens of parvus. Indian localities where parvus has been found are : Kodaikanal (Palni Hills, South India), Darjiling (eastern Himalayas), Patna (Bihar), Allahabad and Saharanpur (United Provinces), Partabgarh (Rajputana), Lahore, Lyallpur, Ferozepur, Pesha- war, Mardan, Chitral (Punjab and Northwestern Frontier Province), Gorai, Srinagar (Kashmir), Kasauli, Barogh, Naini Tal (western Himalayas). All are in the temperate zone except Kodaikanal which is a summer resort where elevations are 6000- 7000 feet. The species was found in Burma, during twenty years of intensive collecting, only at elevations of more than 4000 feet, in and around four hill stations where European cultivated plants had been introduced. A hill station also provided the only record for the Malay Peninsula. B. parvus obviously is more widely distributed in southeast Asia than the supposedly cosmopolitan 1958 REVISION OP INDIAN LUMBRICIDS 5 "caliginosa." The species usually is assumed to have originated in America from whence, or indirectly through Europe, it was brought by man to India. The plants along with which parvus was brought to the orient were not taken only to the hill stations of Burma or to the Indian localities just mentioned. Such plants, potted or other- wise, are common throughout the Allahabad sector of the Gangetic Valley and also in the Jubbulpore region. The plants undoubt- edly have been taken on innumerable occasions into the tropical lowlands of India, Burma and the Malay Peninsula. Absence in those lowlands after a hundred years and more of such introduc- tions must have resulted from inherent inability of the species to adapt itself to a tropical climate. BiMASTOS EiSENi (Levinscn, 1883) This peregrine species, which probably does not belong in Bimastos, is known in India only from the region of Naini Tal in the western Himalayas. Genus DeNDEOBAENA Eisen 1874 Dendrobaena octaedra (Savigny, 1826) This species, hitherto unknown from India, is now recorded from Kodaikanal (Palni Hills, South India) where it appeared to be quite common. This is another form to which the charac- terization of cosmopolitan has been erroneously applied. South of Siberia, the species has been found only at the Kodaikanal summer resort. The species is also absent from almost all of South America. D. octaedra is not usually geophagous and, accordingly, would not have been expected to be carried around the world as much as the earth-eating species. Nevertheless, specimens of this species have been intercepted at American ports during the last few years, in shipments of plants from Spain, Belgium, Holland, Germany, Denmark, Norway, Sweden, Italy, Czechoslovakia and Portugal. Importations into North America in days when quarantine restrictions were unenforced or lacking must have been much more numerous. The species obviously is exotic in India as well as in South and North America. 6 BREVIORA No. 91 Dendrobaena rubida (Savigny, 1826) Athecal Indian specimens that have been identified as Bimastos constrictus or Bimastos tenuis belong here. D. suhrubicunda also is a synonym. Indian localities from which the species is now known are : Kodaikanal, Fern Hill and Ootaeamund (Palni and Nilgiri Hills of South India, at elevations of 6000-8000 feet), Darjiling (east- ern Himalayas), Almora, Naini Tal, Simla and vicinity (western Himalayas). Genus ElSENIA Malm 1877 EisENiA foetida (Savigny, 1826) Indian localities from which the species has been recorded are : Ponmudi, Travancore ; Kodaikanal and vicinitj^ (Palni Hills) where it is very common; Fern Hill, Coonoor (Nilgiri Hills) ; Calcutta; Darjiling and vicinity (eastern Himalayas) where it is also very common; Simla, Kasauli, Dharmsala (western Hima- layas). Absence of records for the Punjab, where Stephenson and his students worked, is noteworthy. The Travancore and Calcutta records never have been con- firmed. E. foetida was deliberately introduced, by the hundreds or thousands, on various occasions within the last twenty years, to lowland tropical areas of the Malayan and Indian peninsulas. All such introductions have been failures. Many more times the species must have been brought into the plains near the hill stations. Some such importation, most probably from Darjiling, may well have been responsible for the Calcutta record. EiSENiA HORTENSis (Michaelscu, 1890) Eisenia veneta f. hortensis, Gates, 1951, Proc. Nat. Acad. Sci. India, B, vol. 21, pp. 19, 21. This peregrine species has been found, in India, only at two Himalayan localities, Darjiling and Simla. Eisenia rosea (Savigny, 1826) Helodrilus (Bimastus) indicus Michaelsen 1907, Mitt. Nat. Miis. Hamburg, vol. 24, p. 188. (Type locality, Calcutta. Types, five, presumably in 1958 REVISION OF INDIAN LUMBRICIDS 7 the Indian Museum but probably valueless as they already were ' ' weak- ened" from maceration in 1907.) Helodrilus (Allolobophora) prashadi Stephenson 1922, Rec. Indian Mus., Calcutta, vol. 24, p. 440. (Type locality, Gundarbal, Kashmir. Types, several, in the British Museum.) Allolobophora (Allolobophora) prashadi + Allolobophora (Bimastus) indica, Stephenson, 1923, Oligochaeta, in Fauna of British India, Ceylon and Burma, London, pp. 501 and 506. Stephenson's species is known only from the original descrip- tion. Insofar as can be determined from that description the types differed from Savigny's rosea only by the absence of spermathecae. As various strains of 7'osea are now known to be athecal there is no longer any reason for recognition of prashadi. Michaelsen 's species is known only from the types the descrip- tions of which contain no indications that external characteristics or internal anatomy warrant distinction from athecal, quadri- vesiculate strains of E. rosea. Tubercula pubertatis, not men- tioned, may have been unrecognizable because of poor condition or they may have been lacking — they are disappearing in some of the athecal strains. The diameter, 6 mm., is 2 mm. greater than the maximum usually mentioned for rosea but the types may have been bloated (as are some recently examined specimens where thickness in the clitellar region is about 7 mm.) . The types of indicus could have been collected at the Botanical Gardens in Sibpur where there were secured types of at least one other earthworm species erroneously thought to be endemic in Bengal. Indeed, in absence of subsequent records to confirm domicile in the area, it now seems possible that the specimens in question may have been secured directly from earth around roots of imported plants. Certainly, there is now not the slightest reason for believing that any lumbricid is endemic in Bengal. All individuals of athecal xgtrains of rosea that have been studied by the writer w^ere male sterile. Tw^o spermatophores, each containing "two oval sperm masses" were, however, present on one of the types of indicus. Acquisition of ability to reproduce asexually often precedes male sterilization. As soon as reproduc- tion becomes parthenogenetic, important organs such as the spermathecae and male terminalia can be eliminated. In the genus Pheretima, for example, several instances are know^n in which extrusion to the exterior of sperm has become impossible though 8 BREVIORA No. 91 maturation still is profuse. Too little is known as yet about the consequences of becoming parthenogenetic, in the Lumbricidae, to permit attribution of much taxonomic value to the presence of spermatophores even if containing sperm. Savigny's species obviously does not belong in the genus Eisenia but may as well remain there until lumbricid genera can be defined by reference to structures that are less liable to rapid evolutionary change than are the organs of the reproductive system. E. rosea has been reported from Kashmir, Murree Subdivision of the Punjab, Chitral in the Northwestern Frontier Province, Simla (western Himalayas), Darjiling (eastern Himalayas), and to the south only from Kodaikanal (Palni Hills). The new synonymies require no additions to the distribution as domicile in the region of Calcutta is improbable. Genus LUMBEICUS L. L. rubellus Hoffmeister 1845 has been recorded from the Nicobar Islands. No confirmation has been provided in the last sixty years. Subsequent collections from those islands have con- tained no lumbricids and domicile is improbable. The specimens may have been found in imported earth. L. terrestris L. This species has been found only once in India and then at Simla in the western Himalayas. Genus OCTOLASIUM Orley 1885 OcTOLAsiUM CYANEUM (Saviguy, 1826) Helodrilus (Dendrobaena) Tcempi Stephenson 1922, Rec. Indian Mus., Cal- cutta, vol. 24, p. 441. (Type locality, Kufri, Simla Hill states. Types, two, presumably in the Indian Museum.) Allolobopliora (Dendrobaena) kem-pi, Stephenson, 1923, Oligochaeta, in Fauna of British India, Ceylon and Burma, London, p. 502. Octolasium oyaneum, Cernosvitov, 1937, Rec. Indian Mus., Calcutta, vol. 39, p. 111. (Murree, Punjab.) Dendrobaena kempi, Gates, 1939, Vest. Cesk. Spol. Nauk, Praze, 1938-1939, p. 151. (Ootacamund, Nilgiri Hills, South India, at elevations of 6700-8000 feet.) Octolasium cyaneum, Gates, 1951, Proc. Nat. Acad. Sci. India, B, vol. 21, p. 19. (Simla.) 1958 REVISION OF INDIAN LUMBRICIDS 9 Stephenson's species was distinguished from other Indian lumbricids, except mariensis, by presence of hearts in xii. The Ootacamiind specimens also appeared to have an extra pair. Junc- tions with the ventral blood vessel, which lateral hearts must have, were not found in specimens secured at Simla not far from the type locality of kenipi. However, the supposed hearts are only dorsally directed continuations of the extra-oesophageal trunks. These continuations, unfortunately, do look very much like hearts in a dissection from the dorsal side. There remain then no taxonomically important differences from cyaneiim. Reproduction probably is not biparental as no sperm have been found in the spermathecae (nor any spermatophores ex- ternally) of any of the specimens that have been available from various parts of the world. Uniparental reproduction, whether sexual or asexual, presumably would predispose to successful colonization after introduction. The species would appear to have been much more common at Ootacamund than elsewhere in India or indeed at any of the localities from which the writer has collected it. OcTOLAsiUM LACTEUM (Orley, 1881) Eophila liimalayana Cernosvitov 1937, Eec. Indiaai Mus., Calcutta, vol. 39, p. 109. (Type locality, Simla. Types, two, presumably in the Indian Museum.) Cernosvitov 's species is known only from his description. Sev- eral days' search at the type locality in 1944 by a good collector yielded a number of specimens of lacteum as well as representa- tives of two other peregrine forms that were previously unknown in India. Absence of seminal vesicles in segments ix and x is the only character available to distinguish liimalayana from lacteum. Undoubtedly the quadrivesiculate condition was the reason, and indeed the only one, for referring types of liimalayana to the genus Eophila. Unfortunately the condition of the material sub- mitted to Cernosvitov for study was poor, as the writer saw before it was mailed. Most of the specimens were more or less softened or brittle and alcohol-browned. In macerated individuals the smaller vesicles of the ix-x, especially when vestigial, may be indistinguishable from adjacent tissues. In brittle material the 10 BREVIORA No. 91 vesicles are apt to be broken off during dissection so as to leave no macroscopically recognizable traces of their former presence. Reproduction in lacteum probably is not biparental (though copulating couples have been seen, Gates, 1953), spermatophores never having been found externally nor sperm seen in the spermathecae of any of the specimens that have been available from various parts of the world. Some strains probably are male sterile but in others maturation of sperm appears to be so sparse as to warrant anticipation of parthenogenesis. With reproduc- tion now asexual, seminal vesicles no longer are necessary and can be eliminated without harmful results to the individual or the strain. Seminal vesicles of ix-x have been quite small to vestigial in some strains recently examined and apparently have been lacking in others. The pair in ix alone or that in x alone has been quite small, vestigial, or unrecognized in other strains. Occa- sionally even the vesicles of xi have been vestigial when those of ix or ix and x were better developed. As these reductions or eliminations take place within the limits of what must otherwise be considered a species, there is now no good reason for recog- nition of kimalayana. Indian localities at which lacteum has been found are Darjiling, Mussoorie, Simla (all Himalayan), Murree subdivision (Punjab). All are well within the temperate zone. DISCUSSION Fourteen lumbricid species, if the unallotted residue of the caliginosa complex is counted as one, are now known to be pres- ent in the Indian subcontinent. All of them obviously are exotic there. No evidence is available to indicate any endemicity in adjacent regions of Baluchistan, Afghanistan and Tibet, below the Hindu Kush and Karakorum ranges. Lumbricids never were found in northern Burma and no endemics were discovered by Chen (1933, 1946) in the Yangtze Valley and the interior prov- ince of Szechuan in China. There is now no reason to anticipate lumbricid endemicity south of the Tian Shan and Altai moun- tains, Mongolia and Manchuria. Korea and Japan appear to have one or more native forms but none is known from Man- churia. Twelve of the Indian lumbricids are present in the British Isles where no members of the family now seem to be endemic. 1958 REVISION OF INDIAN LUMBRICIDS 11 Twelve are widely distributed in North America. Many of them are present in South America, Africa, Australia, New Zealand as well as in various oceanic islands. One species, Bimastos par- vus, usually is assumed to have had an American origin. The others certainly are all from Europe, from whence the worms were ferried to other continents and to the oceanic islands. The only way earthworms are known to be transported for such distances is in ballast or in earth with those plants that are cultivated by man. Transportation of the plants is unquestion- able. The data now available (in MS) show that even a handful of the fertile, unsterilized earth surrounding the plant roots is liable to have contained cocoons or small juveniles and that larger samples often must have contained one or more adults. From the earliest transoceanic centers of successful colonization the exotic worms may have been transported by flood waters as well as by anglers, and, certainly, were distributed, on a geometri- callj' increasing scale, by florists and horticulturalists as well as by innumerable householders. The alien species accordingly are properly called peregrine, the original meaning of which is foreign or imported. In India, lumbricids are common in the Punjab and in the Northwestern Frontier Province. In the Himalayas, lumbricids never have been reported from Sikkim, Bhutan and Assam where there are no summer resorts, nor from Nepal where until very recently Europeans we/e excluded. Only in the region around Darjiling and in the sector west of Nepal are these worms known to be present. In the regions around the hill stations of both sectors lumbricids not only are common but also, in some part of the year at least, dominant. In the Gangetic Valley, which is well within the temperate zone, lumbricids are rare. Only 44 individuals, all of one species {B. parvus), were included among the 30,000 earthworms identified during the course of a four-year survey of the fauna of the Allahabad sector. No lumbricids were present among some 15,000 thousand earthworms collected in a Jubbulpore sector of the Deccan below Allahabad. Farther down the peninsula lumbricids have not been found until well into the far south and then at elevations of 5000-8000 feet but only in and around summer resorts. Exotic plants are grown in pots as well as in the ground at all Indian, Burmese and Malayan localities where lumbricids are 12 BREVIORA No. 91 present. These plants were taken to the hills since 1800, as sum- mer resorts were developed by Europeans, but could have been introduced into many lowland areas of southeast Asia during the previous three centuries. Much earlier than 1500, during the period between Alexander's invasion and the downfall of the Graeco-Bactrian kingdoms, some cultivated plants may have been taken from Europe to the Punjab and Kashmir. The exotic plants with which European earthworms were taken to the hill resorts of southeast Asia are common in the western portion of the Gangetic plains, in the Jubbulpore region of the Deccan as well as in Ceylon where there are no lumbricids. The same plants also have been taken many times from the hills to the lowlands, probably most frequently to larger municipalities quickly reached from hill resorts, e.g., from Darjiling to Cal- cutta. Absence of all lumbricids in most of such plant trans- portations and presence of only B. parvus in the remainder is improbable. Deliberate mass introductions of lumbricids to cer- tain sections of India and the Malay Peninsula in the last twenty years have been no more successful. Failure of the very lumbricids that have successfully colonized in the Americas, Africa, Australia and New Zealand to establish themselves in so many parts of southeast Asia to which they Avere introduced does not now appear to be attributable to the rigor of any competition provided by native species. An American glossoscolecid earthworm, Pontoscolex corethrurus (Miiller, 1856), possibly introduced more recently and much less fre- quently, is widely domiciled throughout southeast Asia as well as Malaysia and in some localities appears to have become the dominant species. An exotic ocnerodrilid and a peregrine megascolecid, as well as a south Indian species of Lampito, now are common throughout the Gangetic Valley. A score or more of other exotic earthworms are established in Ceylon. Moreover, the very numerous unsuccessful introductions that there must have been do not now seem to be attributable, as a rule, to absence of appropriate habitats. Most of the peregrine lumbricids seem to be tolerant of considerable variation in pH, moisture content and basic chemical constitution of the soil as well as in the amount of digestible organic matter. Nor has the reversal of seasons that exists below the equator prevented acqui- sition of domicile. 1958 REVISION OF INDIAN LUMBRICIDS 13 A first approximation to an explanation of the unsuccessful introductions probably could have been given long ago if distri- butional data had been particularized instead of being camou- flaged by geographical generalizations and erroneous adjectives such as cosmopolitan. The Lumbricidae, prior to human inter- vention in the situation, were confined to the temperate zone of the Northern Hemisphere and, possibly, to some adjacent Arctic portions of Eurasia. Transported to everj^ part of the world where Europeans have lived, lumbricids have acquired perma- nent domicile in the temperate zones and the tropics but in the latter only on the highlands. The elevations at which worms were collected usually are not stated in the literature but the tropical sites appear to be above the 5000 foot level. An inherent inability to survive in the less rigorous and more equable climate of the tropical lowlands seems to have been demonstrated. All lumbricids are able to withstand the rigors of a temperate zone winter. Forms such as Z>. octaedra, L. ruhellus and A. chlorotica (Savigny, 1826), that now live respectively in Nova Zemlya, northern Siberia and Greenland, must be able to with- stand freezing temperatures (and being frozen?) for long pe- riods. D. octaedra also has been able to colonize around Mexico City and Bogota (Colombia) as well as on the Palni Hills of south India. The other two species appear to be somewhat less adaptable as they are not known to be present anywhere in the tropics though more liable to transportation. A. trapezoides appears to be even less adaptable as it is certainly known only from areas with a climate more like that of the Mediterranean lands from which it presumably was taken. The "endemic" lumbricids occasionally must have been transported but without being able to take advantage of the opportunities for extending their range. Furthermore, the distributional records published in the last fifty years seem to hint that even in their own areas the natives are being replaced by peregrine forms that may have been introduced there too by man. Accordingly, the family, as a whole, seems now to be definitely limited in ability to ''ad- vance," by climatic factors (presumably thermal), and seems to lack the youthful characteristics implied by such words as ' ' powerful " or " conquering. ' ' Even in the area that the lumbri- cids have "conquered" without human assistance, dominion is shared with two other families, the Hormogastridae and Criodrili- dae. 14 BREVIORA No. 91 Characteristics of youthful vigor and great evolutionary adaptability, it now appears, were too hastily imputed to the Lumbricidae as a whole because of the distributions of less than a score of peregrine forms. Reproduction in some of these forms is parthenogenetic which has been thought to predispose to successful colonization. Peregrine forms with obligatory bi- parental reproduction, however, are much more common in con- siderable areas of North America. More important then than parthenogenesis now seems to be an ability to tolerate human interference in the earthworm environment. Those species, the haemerophiles, which are able to endure the disturbances of the soil involved in agricultural and horticultural practices are pre- disposed, as it were, to transoceanic and subsequent intraregional carriage as well as to survival in new areas that are also culti- vated or otherwise influenced by man. One of the peregrine species, Eisenia foetida, has been thought to be haemerobiontic, i.e., dependent on culture, though that characterization, at least in the vicinity of Darjiling, seems exaggerated. Even D. oc- taedra, sometimes said to be haemerophobic, certainly lives so close to sites of human activity as to be frequently transported. Introduction of lumbricids, according to Stephenson (1930, p. 905), "frequently causes the disappearance of the endemic earthworm fauna." The evidence usually cited for that con- clusion is absence of native species in and around municipal areas of Chile and Australia that were investigated by Michaelsen during the present century. Whether endemics previously had been common at those sites is of course unknown, and the earth- worm habitats therein had been long subject to modification by man. Such little information as is now available hints that indigenes are haemerophobic. Culture, then, could have left areas formerly inhabited by earthworms open to rapid coloniza- tion by any of the preadapted species. These, of course, were most likely to be from Europe, as were the settlers. As the supposed competitive virility of the Lumbricidae also seems to be evidenced by little more than the predisposition of a few species to colonization in vacant but climatically restricted areas, there remains no reason for believing the family to be the "youngest" of the earthworms. Although India and China now must be excluded from the lumbricid domain, there still seems to be little reason to question a former range from the Mississippi 1958 REVISION OF INDIAN LUMBRICIDS • 15 River in North America across Europe and through Siberia to Japan. Most of the earthworm families of Michaelsen 's last classi- fication (1928), such as Sparganophilidae, Criodrilidae, Hormo- gastridae, Syngenodrilidae, Eudrilidae, Moniligastridae, are much more restricted geographically. If area is an indicator of family age (and there is at present little else from which to judge), the Lumbricidae are much older than has been thought, possibly as ancient as any of the four remaining families in Michaelsen 's system. Prior to the Pleistocene ice ages there may well have been, in the vast area between the Mississippi and Japan, manj'' more lumbricid species than are extant. Advance of ice sheets resulted in disappearance of earthworms from Greenland, Iceland, Can- ada, the northeastern part of the United States, northern Europe and perhaps also the British Isles. In North America a few spe- cies of two lumbricid genera survived but to this day they are largely confined to infrequently investigated habitats in the Appalachians. In Europe, the endemics, with a few exceptions that need further consideration, also have not advanced into the glaciated area. In that region the common species are those which seem most likely to have been living, during the glacial periods, just south of the ice cap. How closely they were able to follow the northward retreating ice remains to be learned. As none is known to be haemerophobic and many are haemerophilic the influence of man on their distribution in northern Europe must be determined first. That influence, hitherto underestimated for much of the world, enabled peregrine lumbricids, mainly from northern Europe, to acquire domicile throughout most of the settled regions of non-tropical North America in less than 500 years. SUMMARY Soil around roots of plants has carried lumbricids to every part of the world where Europeans have lived. A high degree of tolerance for human disturbance of their habitats predisposes a score of species to transoceanic as well as subsequent intra- regional carriage and also colonization in vacant habitats where haemerophobic endemics have been eliminated by agriculture. 16 BREVIORA No. 91 Acquisition of domicile in the tropics only at elevations above 5000 feet by otherwise successful colonizers shows an inherent inability to adjust to the milder and more equable climate of the lowlands. The size of the lumbricid domain, though India and China are excluded, indicates that the family is old. The Lumbri- cidae are conservative except as parthenogenesis permits more rapid accumulation of mutations. REFERENCES " Chen, Y. 19.33. A preliminary survey of the earthworms of the lower Yangtze valley. Contr. Biol. Lab. Sci. Soc. China, Zool., vol. 9, pp. 177- 296. 1946. On the terrestrial Oligochaeta from Szechwan. III. Jour. West China Border Ees. Soc, B, vol. 16, pp. 83-141. Gates, G. E. 1945. The earthworms of Allahabad. Proc. Nat. Aead. Sci. India, vol. 15, pp. 44-56. 1947. Earthworms of the Allahabad sector of the Gangetic Plain. Ibid, vol. 17, pp. 117-128. 1951. On the earthworms of Saharanpur, Dehra Dun and some Himalayan hill stations. Ibid., vol. 21, pp. 16-22. 1953. On the earthworms of the Arnold Arboretum, Boston. Bull Mus. Comp. Zool., Harvard College, vol. 107, pp. 500-534. 1956. Earthworms of the region south of the Allahabad sector of the Gangetic Plain. Proc. Nat. Acad. Sci. India, vol. 26, pp. 145-152. JULIN, E. 1949. De Svenska Daggmaskarterna. Ark. Zool., vol. 42 (17), pp. 1-58. MiOHAELSEN, W. 1928. Oligochaeta. In Kiikenthal and Krumbach, Handbuch der Zoologie, vol. II, fasc. 2, part 8, pp. 1-108, 103 text figs. Berlin and Leipzig. STia'HENSON, J. 1930. The Oligochaeta. Oxford. 930 pp. Tetky, a. 1938. Contribution k 1 'etude de la fauna de 1 'est de la France. Nancy. BREVIORA MiuseuntTi of Comparsitive Zoology Cambridge, Mass. September 17, 1958 Number 92 A NEW GENUS OF ERETHIZONTID RODENTS FROM THE COLHUEHUAPIAN OF PATAGONIA By Bryan Patterson Completion of a monograph of the Deseadan rodents by Pro- fessor Albert E. Wood and myself, which includes incidental discussion of Colhuehuapian rodents, necessitates publication of this note. A full account of the new form will be given in a revision of the Colhuehuapian mammals now in preparation. Material of this rodent is contained in the collections of the Museo Argentino de Ciencias Naturales (MACN), the Museo Municipal de Mar del Plata, the Museum Nationale d'Histoire Naturelle, The American Museum of Natural History and the Chicago Natural History Museum. For the privilege of examina- tion and description, I am obliged to the authorities of these institutions, and in particular to Dr. J. -P. Lehman who was so kind as to forward specimens from Paris to Cambridge. For the opportunity of working in Buenos Aires and in Paris my thanks are due, respectively, to the John Simon Guggenheim Memorial Foundation and to the Carnegie Corporation. Seven species of erethizontids have been described from the Colhuehuapian : Eosteiromys homogenidens Ameghino, Steiro- mys pseudonectus Bordas, Steiromys segregatus Ameghino, Parasteiromys uniformis Ameghino, Steiromys axiculus Ame- ghino, Steiromys tahulatus Ameghino and Steiromys nectus Ameghino. The first three of these seem to be synonymous. Eosteiromys is a valid genus with a non-molariform Pi-, distinct postorbital processes, and nasals very strongly convex both anteroposteriorly and transversely. "Parasteiromys" uniformis. as Wood and I point out, appears to be referable to Steiromys, and is the only representative of this genus in the fauna. The BREVIORA No. 92 last three species are indistinguishable, and S. tabulatus and S. nectiis are hereby placed in the synonymy of S. axiculus. This species, which is clearly not referable to Steiromys, is the type of the new genus. B Figure 1. Hypsosteiromys axiculus Ameghiiio. A, RP± — MJ.; B, LP-r- M: X4. Hypsosteiromys 1 gen. nov. Type species: — Steiromys axiculus Ameghino 1902. Distribution ;— Colhuehuapian, Patagonia. Diagnosis: — Cheek teeth with higher, more compressed lophs and lophids and deeper fossettes and fossettids than in other forms; Pi_not molariform, lacking metaloph; neoloph absent on M^^^, present on M-, rudimentary on P-i ; lophs and lophids showing marked tendency to irregular, very shallow notching; 1 v^'^., on high, plus Steiromys, in allusion to the relative height of the molar crowns compared to those of other erethizontlds. 1958 NEW ERETIIIZONTID RODENT lower 1 small, short, extending posteriorly to end of M3 ; con- spicuous depression in maxillary anterior to tooth row and medial to root of zygomia. Hypsosteiromys axiculus Ameghino Steiromxjs axiculus Ameghino 1902, p. 109. Steiromys nectus Ameghino 1902, p. 109. Stciromys tahuJatus Ameghino 1902, pp. 109-110. Stichomys regius? Bordas (non Ameghino) 1939, p. 428 fig. 9. Type: — MACN no. A 52-171, fragment of right ramus with root of I, Mtij (Mj not fully erupted) . Type of 8. nectus: — MACN no. A 52-177, fragment of right ramus with root of I, dmj, Mp2 (^2 ^^^ fuHy erupted). Lectotype of S. tahulatus :—MAC^ no. A 52-172, fragment of left ramus with Fj — M3, crushed alveolus of I. Two other syntypes, an upper and a lower incisor from different indi- viduals (MACN no. A 52-180), are of very dubious reference. Horizons and localities: — All specimens but one are from the ( 'olhue-Huapi formation in the Barranca south of Lake Col- hue-Huapi, Province of Chubut ; the exception (Bordas 1939, MACN no. 11247) is from the Trelew formation in the South Barranca of the Chubut Valley at Caiman, Province of Chubut. Diagnosis: — As for the genus. I append measurements of Ameghino 's types, and drawings by Mrs. Dorothy Marsh of two specimens in the Museum Nationale d'Histoire Naturelle collection. Measurements in millimeters. IT dm^" P4 MT M2 M3 ap. tr. ap. tr. ap. tr. ap. tr. ap. tr. ap. tr. MACN no. A 52-171 — 3.0+ 5.6 4.4 5.7 4.8 MACN A 52-177 • — - 2.4 7.3 3.5 5.0 4.3 MACN A 52-172 7.1 5.2 5.3 4.9 5.2 5.2 5.7 4.6 4 BREVIORA No. 92 REFERENCES Ameghino, ¥ 1902. Premiere contribution a la connaissance de la faune mam- malogique des couches a Colpodon. Bol. Acad. Nac. Cienc. Cordoba, 17:71-138. BOBDAS, A. F. 1939. Diagnosis sobre algunos mamiferos de las capas con Colpodon. Physis, 14:413-433. Wood, A. E. and B. Patterson (In press.) The rodents of the Deseadan Oligocene of Patagonia and the beginnings of South American rodent evolution. Bull. Mus. Comp. Zool. BREVIORA MiiseiiMti of Comparative Zoology Cambridge, Mass. September 18, 1958 Number 98 A NEW BARYLAMBDID PANT(3D0NT FROM THE LATE PALEOCENE By Bryan Patterson and Elwyn L. Simons A quarry discovered during the field season of 1939 by a Chi- cago Natural History Museum party working in the late Paleo- cene beds of western Colorado yielded numerous disarticulated remains of a large barylambdid pantodont. It was at once (^-ident, following preparation, that this animal was distinct generically from the large Barijlarnhda faberi, and certainly distinct specifically from the smaller Haplolamhda quinni (Pat- terson, 1939), both from the same horizon. The nature of the type and only specimen of H. quiinii (the anterior half of a skeleton, much of it pieced together and restored from weathered- out fragments) left some doubt as to the degree of relationship between this form and the new one, and description of the latter Avas accordingly postponed. During recent years much panto- dont material has been obtained from the Paleocene of the Polecat Bench formation of Wyoming by Princeton University parties under Professor Glenn L. Jepsen, supported by the William Berryman Scott Research Fund. This has now been prepared and is under study by Simons. Comparison of the Chicago and Princeton collections shows that Haplolamhda and the new form are represented in both and that they are beyond doubt generi- cally distinct. A specimen from Colorado in the Carnegie IMuseum also proves referable to the new form. Edinger (1950, 1956) has published some notes on the frontal sinus and endocranial cast of Colorado specimens of the new form under the name of Haplolamhda? sp. "We wish to express our sincere thanks to Professor Jepsen and to Dr. J. LeRoy Kay for the opportunity of studying specimens in the collections under their respective charges. BRF.VIORA No. 93 PANTODONTA BARYLAMBDTDAE Leptolambda^ '. Postcervical verte- brae with lower nein-al spines than in Barylaiiihda and Haplo- lamhda. Caudals without (or at most with small, anterior) haemapophyses ; caudal centra longer in i)roportion to width and height than in Barylamhda. Scapula not as broad antero- posteriorly, particularly in postscapular portion, as in Bary- lamhda, broader than in Haplolamhda ; spine proportionately narrower than in Baryhnitbda, lacking tuberosity. Clavicle more slender than in Barylauihda. Bones of forelimb shorter relative to hind than in Baryhnnhda; deltopectoral crest of humerus not as fiat, less projecting medially ; shaft of ulna less broad from front to back, strongly retrofiexed, olecranon relatively longer than in Barylamhda and Haplolamhda. Metacarpal I and tra]')ezium not fused as in HapJoJamhda. Metacarpal V much shorter than in Barylamhda, relatively shorter than in Haplo- la)iihda, with large lateral expansion in proximal half. Proximal and mesial phalanges differing from Barylamhda in much greater compression (or foreshortening) along axis of digit. Bones of hind limb less robust than in Barylamhda; femur with more slender, less flattened shaft, third trochanter nearly vesti- gial ; tibia with more slender sliaft, no pronounced cnemial crest. distal extremity proportionately less broad; proximal articula- tion for fii)ula less projecting laterally; tuber calcis of calcaneum very much shorter. / Leptolambda schmidti- sp. nov. Figs. 1-3 Type. CNHM^ No. P26075, incomplete skull, mandibles, numer- ous vertebrae and ribs, incomplete scapula and pelvis, various leg and foot bones. 2 Named for Dr. Robert G. Schiiiiilt. who, with Dr. James H. Quinn, discovered the quarry from wliich nearly all the Plateau Valley material of this species was obtained. 3 Ahl)reviations are as follows : CXIIM. Chicago Natural History Museum ; CM, Carnejrie Museum : PU, Princeton University. BREVIOBA No. 93 Hypodigm. Type and CNHM Nos. P26076-7, incomplete sknll, various vertebrae and ribs, incomplete scapula, various leg and foot bones ; P15558, P15571, incomplete dentitions ; CM No. 11353, facial region of skull with dentition ; PU Nos. 14680, 14879, and 14996, partial skeletons; PU Nos. 14681, 14990, 14992. mandibles. Horizon. DeBeque formation, Plateau Valley local fauna. Mesa Co., Colorado ; lower levels of the Silver Coulee beds. Pole- cat Bench formation. Park and Bighorn counties, Wyoming. B Figure 'l. A, ulna of Ihuylarnhda faheri Patterson, CNHM No. P26110, and B, ulna of Lcptolamhda .scJiiitiiIti gen. et sp. nov., PU No. 14879. x %. Localities. CNllM Xos. P26075-7 from one ([uarry apjiroxi- mately four miles 8SE. of DeBeque, Mesa Co., Colorado (due to complete lack of association in this quarry there is uncer- tainty as to the ])recise allocation of the individual bones to these numbers) ; Ci\I No. 11353, 2 to 3 miles west of DeBeque, Mesa Co., Colorado; VV Nos. 14680 and 14681 from the south side of Polecat Bench, Park Co., Wyoming; PU Nos. 14879. 14990, 14992, nm\ 14996 from separate localities ah)ng a NU-SW 1958 NEW BARYLAMBDin PANTODONT line between the towns of Lovell and Greybnll in T 54 and 55 N, R 95 and !)(i AV, Bijiiiorn Connty, Wyoming. I)i.a(/H()sis. As for tlie genns ; size approximately as in Bary- lambda faberi or somewhat smaller, hut proportions clearly very different. Sjieeimens from Colorado and from Wyoming fall B D Figure .'5. Femora and til)iae of A, C, Barylamhda faheri Patterson, CNHM No. ri494-4, and B, I), of Leptolamhda schmidii gen. et sp. nov., CXHM Xos. P26075, type and P2(i077. x %. 6 BREVIORA No. 93 iuto two rather sharply defined size groups, one 20-30 per cent smaller than the other. The distinction is most apparent in the region of the canine, within the dentition, and in the postcranial skeleton. Differences of approximately this order of magnitude also occur in Bafyhniihda fahcri. We interpret them as indicat- ing a sexual difference in size. Table 1 MEASUREMENTS OF THE DENTITION, IN MILLIMETERS si l-O ^•5 c 11 ap tr 6.0 5.2 11 ap tr 8.0 6.0 8.8 6.2 8.0 6.2 12 ap tr 7.7 6.3 12 ap tr 10.0 6.2 11.0 6.8 10.5 7.0 I-"' np 8.8 9.1 I-'^ ap 11.5 12.7 10.5 tr 9.0 6.8 tr 7.0 7.7 7.0 c ap 1-1.5 13.4 10.9 C ap 12.8 15.7 12.3 tr 15.0 13.5 8.8 tr 8.8 11.4 7.2 pi ap 15.0 14.6 13.8 Pl ap 13.5 16.0 tr 14.0 8.4 8.0 tr 8.6 8.6 p- ap 18.0 16.0 18.2 P-" ap 20.5 21.5 19.0 tr 25.0 26.0 25.5 tr 12.3 15.7 12.6 p3 ap 20.5 17.0 17.0 17.8 p8 ap 22.5 20.5 20.5 19.0 tr 27.2 29.5 25.0 27.0 tr 18.0 16.0 18.3 15.0 P-l ap 18.0 I'-i ap 22.0 21.0 21.6 20.2 tr 32.0 tr 20.5 18.0 19.5 17.0 Ml ap 26.0 23.0 24.5 Ml ap 2.1.0 22.5 23.2 21.8 tr 36.0 31.5 31.6 tr trig 19..1 17.0 18.0 18.2 M^ ap 26.0 27.0 23.5 22.8 25.3 tr tal 19.5 17.8 18.3 16.0 tr 36.0 39.0 32.5 34.1 35.3 Mo ap 26.7 23.0 24.5 23.3 M3 ap 21.0 21.11 17.4 18.1 19.8 tr trig 21.0 17.0 18.2 17.2 tr 34.5 30.0 31.2 33.4 tr tal 17.0 17.0 18.3 15.4 4 The a nteroposterior ■ diameters of the M;: ap 31.5 26.0 30.4 29.0 upper cheek teeth of this sp lecimen ap- tr trig 19.0 16.3 18.2 17.5 pea; r to h; five bee: 11 somewhat redueed h\ tr tal 13.5 13.4 14.4 12.5 wear. 1958 NEW BARYLAMBDII) PANTODONT 7 RcDKirhs. The Carnegie Museum specimen furnishes an in- teresting footnote to the history of Paleoeene exploration. The Plateau Valley fauna did not come to seientifie attention until the early 19;30's, following Edwin B. Faber's discovery of the tyiie and other specimens of Banjlambda faheri. CM No. 11353, h()W(>\('i-, was collected by Earl Douglass about 1909, was tenta- tively identified as Pantolamhda? sp., and the horizon recognized as Paleoeene. Having made and correctly interpreted this re- markabl(> find, Douglass went on to the discovery of the great Jensen dinosaur quarry, exploitation of which .so fully occupied the rest of his working life that he was never able to follow up the lead provided by this first identified Paleoeene mammal from western Colorado. As an additional item of interest, his specimen provides our only adequate knowledge of the facial region of Lcpfolamhda schmidti, a part otherwise represented only by isolated maxillaries in the Chicago Natural Historj- Museum material and by a much crushed skull in the Princeton collec- tions. The field evidence in Colorado suggests that Leptolamhdo schmidti differed from other Plateau Valley pantodonts as re- gards habitat. The previously known forms were all found in mud- or siltstones, whereas every specimen identifiable as L. schmidti was collected from the fillings of stream channels. The quarry that yielded the great bulk of the material w^as in such a situation, the disarticulated bones occurring at the junction of a channel fill with the underlying mudstone but clearly deposited in the channel. Leptotamhda presumably inhabited higher ground marginal to the basin in which DeBeque sedimentation was going on, wuth fragments and occasionally whole or partial carcasses being carried down into the area of deposition. The specimens of Leptolamhda collected in the Bighorn Basin of Wyoming neither confirm nor deny this conjecture. At least three specimens, PU Nos. 1-1680, 14681, and 14879 were collected from mudstones, whereas PIT Nos. 14990, 14992, and 14996 were recovered from sandstones. Nevertheless, the possibility that Barylamhda and Leptolamhda were adapted to rather different environments is suggested, in Wyoming, by the fact that the former genus has not been recovered from the Bighorn Basin, although Leptolamhda is comparatively common there for a 8 BREVIORA No. 93 pantodont, being represented by more than twenty-tive separate finds. Since these two genera occur at the same stratigraphic levels in Colorado it is clear they were co-existent in time and therefore the absence of Barylamhda in the Polecat Bench for- mation suggests that the environmental conditions under which it flourished did not exist in the region of the Bighorn Basin during the deposition of the Silver Coulee beds. REFEEENCES Edinger, T. 1950. Frontal sinus evolution (particularly in the Equidae). Bull. Mus. Comp. Zool., 103:-409-496. ]9.j(). (^l)jets et resultats de la Paleoneurologie. Ann. Palaeont., 42:97-116. Patterson, B. 1939. Xew Pantodonta and Diiiocerata from the upper Paleocene of western Colorado. Geol. Ser. Field Mus. Nat. Hist., 6:351-384. (And earlier references there cited.) Simons, E. L. MS. The Paleocene Pantodonta and their allies. Doctoral disserta- tion, Department of Geology, Princeton University (to be pub- lished). BREVIORA Miuseunra of Coimparative Zoology Cambridge, Mass. September 18, 1958 Number 94 AFFINITIES OF THE PATAGONIAN FOSSIL MAMMAL NECROLESTES By Bryan Patterson Since its description by Ameghino in 1891, the curious little fossorial Necrolestes from the Santa Cruz, beds has never rested securely in the system. Ameghino himself always regarded it as an insectivore allied to the African chrysoehlorids. Scott (1905), in his description of excellent material collected by the Princeton Patagonian Expeditions, compared it with Chrysochloris and Notoryctes and concluded that it was an insectivore related to the African form : ' ' What does seem probable is that Necrolestes is an offshoot from some early member of the family which was at the same time the ancestor of Chrysochloris." Gregory (1910, p. 259) and Schlosser and Woodward in various editions of Zittel accepted Necrolestes as an insectivore, but others have queried or denied such a relationship. Leche pointed out that Necrolestes resembled Notoryctes about as much as it did Chrysochloris, that as a marsupial it would raise fewer zoogeographical difficulties than as an insectivore, and that its affinities were far from decided : " . . . die Beziehungen des Necrolestes zu den lebenden Saugern durch das bisher dargelegten material keineswegs als festgestellt betrachtet werden konnen" (1907, p. 143). Abel (wi Weber 1928, pp. 128-129), while discussing the genus under the Insectivora, expressed similar reservations. In current reference works, Necrolestei^ has been questionably included either among the marsupials (Homer 1945, p. 611) or among the insectivores (Simpson 1945, p. 53). Saban (1954, p. 422) has made the passing suggestion that it may be a palaeanodont edentate. One author who came out unequivocally for marsupial affinities was Winge (1941, pp. 79-80, 217-218), who regarded the incisor 2 BREVIORA No. 94 formula, reported by Scott to be;*, as proof of such relationship. He believed Necrolestes to be a didelphoid derivative, classifying it in a separate division, Necrolestini, of his family Cladosictididae ( =Borhyaenidae ) . No first hand account of this little creature has appeared since Scott's study. His work was perforce done without benefit of the binocular microscope, and his conclusions as to affinities suffered from the fact that he evidently compared Necrolestes only with the equally highly specialized Notoryctes and the chrysochlorids, not at all with more generalized insectivores and marsupials. It had always seemed to me that a form represented by such excellent material • — with most of the skeleton preserved in one or another of the Princeton specimens it is actually among the better known fossil mammals — should reveal enough in the way of structure to permit at least its infraclass affinities to be firmly established. Recent examination and further preparation of the Princeton material has shown this to be indeed the case. Necrolestes is beyond any doubt a marsupial and, as regards its position within the order, Winge's suggestion of a remote rela- tionship with the borhyaenids is perhaps as reasonable a one as can now be offered. Scott's account is in general very good indeed. The brief description that follows is supplementary to his and corrects it in a few points. I wish to express my sincere thanks to Professor Glenn L. Jepsen and to Dr. Donald F. Baird, who very kindly took the photographs from which the drawings prepared by Mrs. Dorothy Marsh were made. MARSUPIALIA BORHYAENOIDEA NECROLESTIDAE Necrolestes patagonensis Ameghino Knowledge of the species, the only one of the family thus far discovered, rests essentially on three specimens in the Princeton collections : 15065. Skull, crushed dorsoventrally and lacking parts of occiput and basicranium, mandible (now missing), various vertebrae and ribs, proximal end of humerus. Ten miles South of Coy Inlet, Santa Cruz; collected by 0. A. Peterson. 1958 AFFINITIES OF NECROLESTES 3 15384. Occiput and part of basicranium, mandible, some iso- lated upper teeth, numerous vertebrae from all parts of the column, pelvis, humerus, radius, ulna and various bones of the fore foot. Five miles South of Coy Inlet, Santa Cruz; collected by 0. A. Peterson. 15699. Skull, incomplete posteriorly, mandible, right humerus, radius and ulna, left radius, incomplete ulna, femur, tibia, fibula, calcaneum, astragalus, cuboid. Killik Aike, Santa Cruz ; collected by J. B. Hatcher. DENTITION Dental formula. It can be definitely stated that this is I ^, Cf, postcanines -| (probably P|, M-|)- Recogni- tion of the correct number of upper incisors at this late date, more than half a century after the full account of the material, comes as somewhat of a surprise. The evidence is unmistakable, however (Fig. 1 A, B). In 15699, I^^ are preserved on the left side and the alveoli of these four teeth, together with I i, on the right. No. 15065 retains right and left I—, right I-^nd the alveoli of all other members of the series. Curiously enough, van Iterson, the artist who drew Scott's figures, indicated the correct number (Scott 1905, pi. 64, figs. 1, 2a). For 15699 he showed five in- cisors in right side view, reversing li^ from the left side, and for 15065 the right and left I- , right I— and the alveoli of right I^^ (the alveoli of left I-^ were at that time concealed by matrix). The figure of 15065 has right 1-5- erroneously labeled as I-^. Determination of the postcanine formula is un- certain. Believing that he was dealing with an insectivore, Scott gave it as Pf , Mg . Since we have a marsupial on our hands, either this formula could be correct, with M-f missing, or the first premolar could be the missing tooth, which would give P|, Mj. The latter possibility seems to me the more likely of the two. The last four cheek teeth above and below are uniform in structure, which suggests, although of course does not prove, that all are molars, while the second pair of teeth, especially the lower, although largelj^ molariform, do show some resemblances to the anterior pair. 4 BREVIORA No. 94 Upper teeth. The first four incisors are of similar struc- ture, with the crowns rectangular, compressed anteroposter- iorly, the labial face slightly convex in the dorsoventral diameter and bearing a faint median groove. I— is the largest, I— slightly smaller and I-^of equal size and slightly smaller than 1J-. I — Is the smallest of the series. It is a bluntly pointed tooth, worn obliquely on the posterior face by the anterior surface of the lower canine. Relative to the sizes of the other teeth, the single-rooted, tri- hedral canines, both upper and lower, are enormous. The upper is somewhat variable, that of 15699 being stouter, more triangular and with a larger anterior face than that of 15065. The first upper cheek tooth, P^as here interpreted, is double- rooted, implanted obliquely, and consists of a high, conical paracone, an anterior basal cusp and a shearing edge running from the apex of the paracone to the large but non-cuspidate metastylar area. The remaining teeth of the postcanine series are unilaterally hypsodont, curving outward toward their bases and tapering to small, bluntly pointed single roots (Fig. IC). The enamel extends nearly to the end of the root of the lingual surface and for only a short distance up the crown on the labial. M^^ are essentially narrow isosceles triangles as re- gards crown contour; Pyhas a convex anterior and M— a convex posterior border. ]\I— are subequal, M— is slightly smaller and M— reduced in size. The internal cusps are presumably paracones, being in line with the paracone of P—, and the greater part of the crowns is formed by the stylar area. There is no vestige of a protocone. These teeth are thus descriptively zalambdodont, but no structural details are apparent in the worn material available. Lower teeth. The incisors resemble those of the upper series and are not, as Scott stated, laterally compressed. ly is much larger than the others, which are subequal. The canine is narrower and less trihedral than the upper, with a smaller anterior face. Pyis double-rooted, elongate anteroposteriorly and bears three cusps, a high, acute protoconid and smaller, subequal an- terior and posterior basal cusps. The molars are alike in structure and are mirror images of the uppers, with no traces of talonids. The second postcanine, P-j- as I believe, differs from them in being less narrowly triangular and in the pos- 1958 AFFINITIES OF NECROLESTES 5 session of a vestigial anterior basal cusp serially homologous with that occurring on P ^ . The less worn of the lower molars available have heart-shaped crown contours caused by a shallow, ephemeral indentation between the bases of the paraconids and metaconids. No doubt P-3--M— are mirror images of the cor- responding upper teeth as regards unilateral hypsodonty, enamel distribution and root structure but no loose examples are available. Wear on all molars is limited to the crowns, where the abrasion surfaces are nearly plane. It is quite evident that in active occlusion the trigonids of the lowers worked across the trigons of the uppers. SKULL As is the case in other small fossorial forms, no suture is apparent in any specimen of Nccrolesfes. The chief peculiarity of the skull is of course the extraordinary upturned snout, which extends well above and in front of the incisors (Fig. 1 A, B). Scott was in some doubt as to whether this feature might not have been exaggerated by distortion in 15699, particularly as there is little trace of it in 15065. It appears to be natural, however. No. 15699 has undergone slight lateral compression, in the course of which the maxillaries and premaxillaries have been broken away from and slightly pinched in relative to the nasals, but this does not appear to have affected the dorsal profile of the facial region. No. 15065 has been badly crushed dorso- ventrally in this region and the upturned portion has been w^eathered off. The anterior narial openings (Fig. 1 A, B) extend forward to the tips of the nasals and do not end short of them, as Scott's figure shows them as doing. The median spine de- scribed by Scott as extending upward and forward from the mid-point of the premaxillaries evidently continued to the tips of the nasals as an ossification in the septum. It is partly broken away but remnants of it are still present. Between I- and the ca- nines, the premaxillaries are very deeply excavated for the reception of the lower canines, perhaps the most extreme develop- ment of this character to be seen in the Marsupialia. The upwardly arching z^^gomatic arch is well developed for so small an animal (Winge described it as slender, but this was in comparison with the arch of the borhyaenids) and quite simi- BREVIORA No. 94 lar to that of small didelphids such as Marmosa. Extension of the jugal to the glenoid articulation is a marsupial character. Despite the obliteration of the sutures, there can be little doubt that the jugal of Necrolestes was so constructed. A prominent Fig. 1. Necrole.stcfi patagoneiisis Ameghino. A, B, riglit anr1 left sides of snout region of skull, X3 approx. (LI, left incisor) ; C, left M2, anterior view, X6 ; D, left innominate bone, lateral view, X2 approx. j)reglenoid process is present. The glenoid surface extends for- ward to this buttress and is in consequence deeply concave anteroposteriorl}'. Precisely similar conditions exist in other marsupials, in which the buttress is formed by the jugal. A post- 1958 AFFINITIES OF NECROLESTES 7 glenoid process is lacking;. The extent of the lachrymal cannot, of course, be seen. The lachrymal foramen, however, is so very similar to that of various didelphids as regards both position and size (far more so than to that of tlie chrysochlorids, with which Scott compared it) as to suggest that the bone itself was likewise comparable. Within the orbito-temporal region, the optic foramen appears to be confluent with the sphenorbital fissure. The long, slitlike incisive foramina are very didelphid-like, while the complete absence of palatal vacuities is a point of resemblance to the borhyaenids. The posterior border of the palate is moderatelj- thickened and the posterior nares are di- vided by a median bony septum. For a marsupial this is very unusual ; it may be an adaptation for providing an increased rigidity that would be useful to an animal using its upturned snout as an earth mover. Specimens 15065 and 15384 reveal a good deal of the structure of the basicranium and auditory region, much of which had hitherto been concealed by matrix. The basicranium is very short. The greater part of the right side of the basisphenoid is preserved in 15065 and this portion is perforated, at the level of the glenoid cavity, by a large foramen, which appears to be the carotid. The bone is almost flat and there is no indication of a tympanic wing. The basioccipital, of which only a small part is preserved in 15384, was clearly short and w4de. As noted by Scott, the condyles are relatively very large and wide ; as in Marmosa there is a conspicuous median notch between them that opens forward from the foramen magnum. The latter opening is large and faces posteriorly. There is no paroccipital process, the "broad and distinct" process so identified by Scott being the mastoid. That part of the alisphenoid external to the foramen ovale, of which the lateral rim is preserved, is present in 15065; it shows no trace of any development of an alisphenoid bulla. It is perforated by a small foramen, presumably vascular, situated midw^ay between the foramen ovale and the glenoid cavity. The periotic is short and robust, with a well rounded promontorium, which I take to have been the basis for Scott's incorrect statement that the "tympanic bulla" was "ossified and moderately inflated." Anteriorly, the bone terminates bluntly and, posteriorly, there is no trace of a tympanic wing. 8 BREVIORA No. 94 The fossa subarcuata is very large, and the anterior vertical semicircular canal is free from surrounding bone. The mastoid process appears to be large. , Immediately posterior to the glenoid cavity is a more or less triangular pocket, whose medial border is formed by the petrous portion of the periotic, the lateral by the squamosal and mastoid, which descend below the level of the glenoid and are concave on their medial surfaces, while the squamosal appears to be the main contributor to the roof, with the other elements par- ticipating to a lesser extent. A low transverse ridge crosses the roof of this pocket, partially dividing it into a large anterior portion, which is open below, and a smaller posterior portion, which is floored by the mastoid process. The fenestra ovalis and the apertura tympanici canalis facialis open into the pocket and the groove for the facial nerve is plainly visible in the lateral side of the periotic. This pocket, then, is the tympanic cavity, and the tympanic itself, of which no part remains, obviously provided a partial floor to the ventrally open anterior portion. There is no emargination of the squamosal marking the passage of the auditory meatus, which must have passed laterally below and behind the glenoid cavity. Within the pocket, in 15065, there is a minute foramen immediately behind the glenoid and a larger one just posterior to it. These I take to be the postzygo- matic and postglenoid foramina. The tympanohyal, descending from the crista facialis, is fused with a small medial projection from the mastoid process in 15065 but stands free in 15384. The "foramen" stylomastoideum primitivum is entirely open, merely the termination of the facial groove, as in various didelphids. Posterior to the tympanohyal is the opening of a short tubular recess, covered laterally by the mastoid and exoccipital, that runs back to the condyle. An almost identical recess occurs in small didelphids, e.g. Marmosa, and presumably houses M. stapedius. Of the preserved foramina not heretofore mentioned, the lacerum posterius is large and the condylar, due to the shortness of the basicranium, is very close to it. The latter, so far as I can detect, is single. There appears to be no subsquamosal foramen. The posterior portion of the mandible is greatly reduced and the angle consequently very small. Notwithstanding, this jn'ocess bears an internal ledge and tliis is inflected and concave dorsallv. 1958 AFFINITIES OP NECROLESTES 9 The fossorial specialization has resulted in a highly modified skeleton that shows numerous resemblances to other, similarly adapted forms, as Scott pointed out. Making due allowance for this, there is nothing that argues against a didelphoid ancestry for the Necrolestidae. The lack of an anticlinal vertebra, a point emphasized by Scott as a resemblance to the chrysochlorids is also true for the thoracic series of some of the smaller didelphids, and in these the neural spines of the lumbars do not incline forward to any appreciable extent. There is clear evidence of a well-developed tail, in agreement with didelphids and bor- hyaenids. The proportioiLs of the pelvis closely resemble those of the smaller didelphids and, as Scott pointed out, there w^as surely a well-developed pubic symphysis. The ilium is more rodlike than in the Didelphidae, a diiference possibly related to the fossorial adaptation of Nccrolestcs. In very close agreement with didelphids, there is a conspicuous pectineal prominence that stands well out from the anterior margin of the pubis. The rela- tively large epipubis of the smaller opossums extends up nearly to the pectineal prominence, and I suspect that the same w^as true of Necrolestes. On the right side of 15384 there is a small fragment of bone attached to this area of the pelvis by matrix, and it is just conceivable that this could be a remnant of the epipubis itself. DISCUSSION Winge was correct in his opinion that Necrolestes was a marsupial. The incisor formula is diagnostic for a large seg- ment of this order. There is nothing of a conflicting nature in the structure of the skull or postcranial skeleton ; on the contrary numerous features described above point so strongly in the same direction that there is no further need to elaborate on the ordinal position of the genus. Saban was quite right in seeking to exclude the Necrolestidae from the Insectivora, and his tentative sug- gestion of palaeanodont affinities was a reasonable one, given the information in Scott's account. The new facts do not support it, however. The Palaeanodonta resemble the Xenarthra in the possession of an ossified bulla composed of ecto- and entotympanic elements (Patterson and Segall, MS.) which Necrolestes does not 10 BREVIORA No. 94 have; the dentition and the pelvis are quite different, and so on. There is, of course, a general resemblance between Epoico- therium and Necrolestes but, like the resemblance between the latter and the chrysochlorids, this is due to convergence.^ There remains for discussion the problem of where, within the Marsupialia, Necrolestes should be placed. The isolated posi- tion of the genus — its lack of any ancestors combined with its great specialization — prevents any satisfactory conclusion, al- though some approximation is possible. The facts of zoogeography would seem to limit the field to the known South American groups. One of these, the Caenolestoidea, may at once be ex- cluded from consideration. No known caenolestoid approaches Necrolestes in any significant way ; the dentitions of the earliest forms are specialized in completely different directions. The remaining two groups, the Didelphoidea and the Borhyaenoidea, on the contrarj^ do show resemblances of a basic sort to the Necro- lestidae. Some of these have been mentioned in the description. As regards the dentition, the incisors of didelphids — and pre- sumably of the earliest borhyaenids — could readily have given rise to those of Necrolestes; the canines are large in the two, and the lower is received into a socket in the upper jaw ; the zalamb- dodont molar type of the necrolestids could readily have evolved from tribosphenic molars of the sort that were no doubt present in the early members of both groups. The skull and postcranial skeleton of Necrolestes reveal, beneath the facade of fossorial specialization, unmistakable borhyaenid and didelphid resem- blances. If, as seems virtually assured, the ancestry of the Necrolestidae is to be sought near the base of the didelphoid-borhyaenoid com- plex, then the question arises as to whether it is nearer to one or to the other of these superfamilies, or is equally distinct from both. Necrolestes is one of the most specialized of American marsupials, specialized in a manner totally unlike anything seen in the Didelphoidea or in the Borhyaenoidea. To seize upon this as an excuse for erecting a superfamily for the reception of one 1 Since this was written, McDowell (Bull. Amer. Mus. Nat. Hist., 115, 173-4, l'J58) lias asserted that A'ccro/c.y/e.s- is a member of the suborder Xeuarthra. In addition to the characters of the auflitory region just cited that oppose reference to the suborder I'alaeanodonta, the presence of rooted, enamel-covered teeth and the at)sence of the diainiostic xenarthran vertebral and pelvic structures prevent acceptance of this opinion. 1958 AFFINITIES OF NECROLESTES 11 genus whose phylogeny is totally unknown, would be the height of irresponsibility, however. Furthermore, there are one or two characters that hint at a more acceptable possibility. As will be evident from the descriptive notes, there are more resemblances to the didelphids than to the borhyaenids in the anatomy of Necrolestes. This is not surprising : the didelphids include the least specialized of marsupials and no doubt retain a good deal of the ancestral structure relatively unmodified. The adequately known borhyaenids are relatively much more specialized and would therefore be less likely to display basic resemblances to Necrolestes. However, in such characters as the absence of palatal fenestrae and the lack of an alisphenoid bulla, Necro- lestes is closer to all or some of the borhyaenids, which conceiv- ably may indicate that the family arose from the borhyaenoid stock shortly after its divergence from the didelphoid. Such evi- dence, of course, is of the slenderest yet I nevertheless clutch at these straws and, in accord with Winge, very tentatively place the Necrolestidae in the Borhyaenoidea.^ As a marsupial, Necro- lestes adds one more example to the long series of convergences and parallelisms for which the extinct South American mam- malian fauna is justly famous: one of convergence toward an- other order (Insectivora, Chrysoehloridae), one of parallelism within the order (Notorijctes). AVith the removal of Necrolestes from the Insectivora, an order of mammals — one widespread elsewhere at an early date — has definitely been eliminated from the early Tertiary South Ameri- can scene. As I have briefly stated (1957, p. 57), it is becoming evident that the earliest South American mammalian fauna must have been an unbalanced one, of the sort encountered on islands. If, as has usually been supposed in recent years, the Americas were united by land in the late Cretaceous and early Paleocene it is very difficult to understand the absence from the southern continent of any traces of multituberculates, insectivores, taenio- donts, ereodonts, fissipeds and pantodonts, to name only the groups that are known, or may with confidence be inferred, to ^I prefer Simpson's classilicatlon of the marsupials into six groups (1945) to tliose classilit-ations that unite Didelphiilae, Borhyaeniclae and Dasyuridae into one. Admittedly, these families are closer to each other morpholnKically than any one of thi'iu is to other marsupial .i:,'r(Hips, l)ut keepinjc them separate (or ar least keeping the Australian and South American groups separate) accords better with the zoogeosraphical t as to the reality of absence, and I therefore omit them from the list, which is quite Imposing enough without them. 1958 AFFINITIES OF NECROLESTES 13 pendent evolution. From the haphazard, unbalanced beginning thus visualized, the South American mammalian fauna in time achieved balance of a sort peculiarly its own. The early immigrants had by chance drifted to an area whose size, physical diversity and multitude of available niches provided the space essential for such a de- velopment. REFERENCES Ameghino, F. 1891. Nuevos restos de mamiferos fosiles descubiertos por Carlos Ameghino en el Eoceno inferior de la Patagonia austral — especies nuevas, adiciones y eorreciones. Rev. Argent. Hist. Nat., 1: 289-328. Darlington, P. J. 1957. Zoogeography: the geographical distribution of animals. New York: John Wiley and Sons, i-xiii, 1-675. Gregory, W. K. 1910. The orders of manmials. Bull. Amer. Mus. Nat. Hist., 27: 1-524. Leche, W. 1907. Zur Entwicklungsgeschichte des Zahn-Systems der Saiigetiere, zugleich ein Beitrag zur Stanimesgeschiohte dieser Tiergruppe. Zweite Teil : Phylogenie. Zweites Heft : Die Familien der Centetidae, Solenodontidae und Chrysocliloridae. Zoologica, 20 (49): 1-158. Patterson, B. 1957. Mammalian phylogeny. In Premier symposium sur la speciticite parasitaire des parasites des vertebres. Neuchatel: Paul At- tinger, 15-48. Romer, a. S. 1945. Vertebrate paleontology. Chicago : The University of Chicago Press, i-ix, 1-687. Saban, R. 1954. Phylogenie des insectivores. Bull. Mus. Nat. d'Hist. Nat. Paris, (2) 26: 419-432. 14 BREVIORA No. 94 Scott, W. B. 1905. Palaeontology. Part II. Insectivora. Repts. Princeton Univ. Expeds. Patagonia, 5: 365-383. Simpson, G. G. 1945. The principles of classification and a classification of mammals. Bull. Amer. Mus. Nat. Hist., 85: i-xvi, 1-350. Weber, M. 1928. Die Saiigeticre. Einfiihrung in die Anatomie und Systematik der recenten und fossilen Mammalia. Band II. Systematischer Teil. Jena: Gustav Fischer: i-xxiv, 1-898. WiNGE, H. 1941. The interrelationships of the mammalian genera. Volume 1. Monotremata, Marsupialia, Insectivora, Chiroptera, Edentata. ILerbenhavn: C. A. Reitzel: i-xii, 1-412. BREVIORA Mnaseiainti of Comparative Zoology Cambridge, Mass. September 19, 1958 Number 95 A NEW BOLIVIAN LAND SNAIL OF THE GENUS DRYMAEUS By Juan Jose Parodiz Acting Curator of Invertebrates, Carnegie Museum I'ittsliurgh, Pa. Drymaeus pereirai sp. nov. Description. Shell subcorneal, elongated. Whorls, seven, in- creasing gradually, the last being approximately % of the total length with its major diameter scarcely larger than one half of the same length. Suture ^vhite, marginated below and somewhat shouldered above, a character which is more evident at the base of the third post-embryonic whorl. Surface sulcate with very fine and irregular axial riblets, more distinct on the middle por- tion of the shell. The interspaces between these riblets are variable in width and there are 3, 2 or occasionally only 1 riblet per mm., with .some microscopical spiral liratioiis, more noticeable on the zone of the last whorl opposite the aperture. The riblets start on the suture, forming a very distinct marginated area. Color (in the holotype) white, with traces of brown bands broken into points and divided by three white zones. This spotted coloration is limited to the intercostal spaces and is not present on the high part of the riblets which remains white, giving a diffused silky aspect. Aperture oval, with rounded base, shiny white, enameled inside and somewhat iridescent; the superior part of the columellar lip a very pale pink ; peristome thin and not reflexed. The protocouch has the same pale pink-lilac shade of the peristome. BREVIORA No. 95 Measurements in mm. : Shell Length Width Ilolotvpe 30 12.5 10.5 Paratype 10 9 Aperture Leno-th Width 13.5 12 i.O 6.5 Last Whorl Length 19.5 16.5 Drymaeus iwreirai, sp. nov. M.C.Z. no. 168843, El Carmen, Prov. Chi- quitos, Depto. Santa Cruz, Eepubliea de Bolivia. Hno. F. S. Pereira and Carl Gans, c-ollectors. (Holotype X "IV-i.) Locality. Estacion El Carmen (between Pahnito and Santa Ana) on the railroad Corumba-Santa Cruz, Prov. Chiquitos, Depto. Santa Cruz in the oriental plains of southeastern Bolivia. Holotype No. 168843 in Museum of Comparative Zoology and paratype in Carnegie Museum Aces. 16,073. Collectors : Hno. F. S. Vereira and Carl Gans, 1953. The paratype (which has the early whorls decollated) is evi- dently a younger specimen as shown not only by its smaller dimensions but by some traces of carination in the last whorl, and 1958 DRYMAEUS PEREIRAI X. SP. 3 fresher colors. The very pale spotted coloration in the paratype deserves an explanation, since it is very possible that other speci- mens with intermediate patterns will be found. The three brownish bands are more distinct, with the dividing white zones also very conspicuous ; the medium, broader band shows darker or clearer triangular areas ; the umbilical zone is completely white ; the lower spiral white zone continues into the aperture and the upper one, on the suture ; the dark bands are visible inside the aperture ; as in the paratype, the oblique axial riblets remain white. The marginated suture, the irregular axial-oblique riblets, and the very pale pink-lilac tint on the columellar lip, are the more important features of this new species. Two other species of Drymacus are known which have a roughened surface of pseudo- riblets: />. sulcosus (Pfeilt'er) from Mexico, which is a larger and stronger species with a thick and refiexed peristome, and some young specimens of the Central American D. jonasi (Pfeif- fer) — such as those figured by Pilsbry in 1899 (Manual of Conchology, (2) 13, p. 54, pi. 10, fig. 64) — which is different in many other characters such as the angulate l)ase of aperture and its broader shape. The region where this new species lias been found should have an extensive exploration from the malacological point of view. We know better a number of characteristic species from the western side of Santa Cruz, on the subandean zone and oriental Cordillera of Bolivia, as well as many others along the Paraguay River, from Corumba to the south. The locality El Carmen is in an intermediate region, more than 200 miles wide, between the Llanos of Izozo and the Paraguay River, being a northern ex- tension of the Boreal Chaco. The species is named for Hno. F. S. Pereira who obtained the specimens while collecting with Dr. Carl Gans along the zone Corumba-Santa Cruz in 1953. I wish to express my appreciation to Dr. J. Bequaert and Dr. W. J. Clench of the Museum of Comparative Zoology, for giving me the opportunity to report upon this new species. BREVIORA Miaseiiiimi of Comparsitive Zoology Cambridge, Mass. September 19, 1958 Number 9(i A NEW DICHOBUNID ARTIODACTYL FROM TPIE UINTA EOCENEi By C. Lewis Gazin Smithsonian Institution INTRODUCTION Recently, Mr. Bryan Patterson turned over to me for study a lower jaw belonging to an Antiacodon-\ike dichobunid from the Uinta upper Eocene. The specimen was discovered by him among some materials secured by the Museum of Comparative Zoology in an exchange with the Cleveland Museum of Natural History. The label accompanying the specimen reveals that the jaw was collected by Mr. D. C. Roberts near the entrance to Myton Pocket, a well-known collecting locality for upper Eocene mammals in the Uinta Basin of northeastern Utah. The specimen is particularly significant in showing that the Antiacodon line did not become extinct in Bridger middle Eocene time but survived at least until the Uinta C horizon of the upper Eocene. The form represented, while clearly related to Antia- codon, shows certain distinctive modifications rather beyond the limits that may be ascribed to that genus. Moreover, the ex- tended range in time of this line, together with the very distinc- tive and divergent characteristics brought to light (separating it from the better-known Homacodontinae), weakens my earlier objection (1952, p. 24) to applying separate subfamily recogni- tion to the antiacodonts. The name Antiacodontinae might well designate this phyletic sequence. The first antiacodont to be placed on record was the Bridger form, described by Cope (1872, p. 2) in his first " Palaeontological 1 Published by permission of the Secretary of the Smithsonian Institution. 2 BREVIORA No. 96 Bulletin" as " Lophiotherium" pygmaeum. The same form was described scarcely a month later by Marsh (1872, p. 210) as Antiacodon venustus. Cope early recognized Marsh's genus but later felt that it was not sufficiently well characterized and in 1875 (p. 256) proposed the name Sarcolemur for the species he had named Antiacodon fiircatiis. These I have shown (1955, p. 23) are synonymous, so that the genotypic Bridger species is Antiacodon pygmaeus. The first recognizable illustration of the lower jaw of Antiacodon (as Sarcolemur) was furnished b}' Osborn (1902) but, following Cope, he regarded it as a primate. Matthew in 1909 was evidently the first to realize that An- tiacodon was an artiodactyl, although earlier in the same year, while including "Sarcolemur" in the Artiodactyla, he believed Antiacodon to be a leptictid. Evidently the earliest form showing a relationship to the antiacodonts is the Wasatchian genus Hexa- codus (Gazin, 1952, p. 73), a form that might be regarded as an antiacodont, but that also includes within its range of character- istics certain indications of an ancestr}^ for Microsus, a typical homacodont. DICHOBUNIDAE Gill ANTIACODONTINAE, new subfamily AUXONTODON", new genus Type. Auxontodon pattersoni, new species. Generic characters. Small incisor-like canine. Much enlarged and probably caniniform Pi. Second to fourth premolars antero- posteriorly elongate. Marked diastema between Pi and P2. Little or no diastema between P2 and Ps. Cheek teeth resembling those of Antiacodon. Inferior margin of lower jaw strongly convex in anteroposterior profile. Comparison. The three lower incisors of Auxontodon, as indi- cated by the root portions, were small, about equal and decidedly procumbent. The root of the lower canine, which is also pro- cumbent and closely parallel to those of the incisors, is only slightly larger. The root of Pi, on the other hand, is much larger - From Greek, Auxontos, iucreasing : and odon, tooth ; in allusion to the in- creased size of the first lower premolar. 1958 NEW MCHOBUNID ARTIODACTYL 3 than that of the canine and appears to be more accurate, so that Avhile the root appears to extend well back in the jaw, the crown may have been more nearly erect. There is a marked diastema betw^een the alveolus for Pi and those for P2, and evidently only a very slight diastema separating P2 from P3. The diastema be- tween P2 and P3 in Antiacodon is variable and evidently some- what greater than in Auxontodon. In A.M.N.H. No. 12697 it is about 2.5 mm. whereas in U.S.N.M. No. 18000 it is 3.9 mm. P2 is not preserved in the Auxontodon specimen, but Pa is seen to be an elongate tooth with a distinctly high primary cusp or protoconid and a very well-defined metaconid almost posterior to it. There would appear to be no talonid on Ps but there is a low, distinct paraconid. P4 has a more distinctly developed talonid with a conspicuous entoconid, better developed and more widely spaced from the crest extending down the posterior slope of the protoconid than in Antiacodon. The metaconid is pro- minent as in Antiacodon but the anterior extremity of this tooth, though damaged, seems more forward-protruding than in the earlier genus. Mi shows a characteristic antiacodont form with a high paraconid and with the crista obliqua extending to the apex of the more worn metaconid. The hypoconulid is posterior to the crest between the hypoconid and entoconid, but appears less conical than in Antiacodon — more as a prominent upward arching of the posterior cingulum. The jaw of Auxontodon shows a decidedly convex lower mar- gin, viewed from the side. It is more nearly straight in Antiaco- don. A mental foramen of moderate size is noted beneath the posterior root of P3. It is in this position or slightly more forward in Antiacodon. However, in Auxontodon there is a larger fora- men below a point just posterior to the alveolus for the Pi, open- ing lateral to the root and extending backward parallel to it. In A.M.N.H. No. 12697 of Antiacodon pygniaeus the larger and more anterior mental foramen is just anterior to the posterior root of P2. Discussion. Perhaps the most significant feature observed in Auxontodon is the much-enlarged root of Pi. In homacodonts in general, although details in this respect are not known in all genera, there is no marked tendency toward differentiation of a stronglv caniniform tooth. Nevertheless, it was noted that in 4 BREVIORA No. 96 such more advanced liomacodonts as Mytononieryx and Pentace- mylus the root portion of Pi is distinctly larger than that of the canine. Moreover, it is more robust than in either of the roots for P2. However, in these genera the crown of Pi is premolar-like and forms with the incisors, canine and following premolars a nearly gradational series in size. In the more-selenodont Eocene artiodactyls, such as the agriochoerids, leptotragulids and hyper- tragulids. Pi early becomes caniniform. In none of the known material of Antiacodon pygmaeus is the anterior portion of the lower jaw sufficiently preserved to indi- cate with any certainty the extent to which Pi was developed. In A.M.N.H. No. 12697 of this species the jaw is preserved to about 1.7 mm. in advance of the roots for P2 but this portion of the lower jaw is relatively slender and there is no evidence of a root for Pi extending back in the jaw. Nevertheless, Antia- codon and Auxontodon show a rather strong tendency toward selenodonty and it is not surprising to find this line within the Dichobunidae showing tendencies paralleling the more highly selenodont families. AUXONTODOX PATTERSONI^, new species Figure la Type. Left ramus of mandible with Ps-Mi and part of M;:, M.C.Z. No. 9316. Horizon and locality. Uinta C, near entrance to Myton Pocket, about 7 miles east of Myton, Duchesne County, Utah. Specific characters. Size of lower jaw much larger than in Antiacodon pyg^naeus. Other characteristics of the species have not been distinguished from those cited for the genus. Discussion. The lower jaw, M.C.Z. No. 9316, is the only spec- imen known of Auxontodon pattersoni. Evidently this species was a very rare element of the fauna, as represented in the Uinta Basin, particularly considering the extensive collecting that has been carried on in the Uinta area and the relative abundance of artiodactyls encountered, both bunodont and selenodont. :i Named f(ir Mr. r.ryaii r;itlrrsi;ii. who first recoffiiiz(Ml the true relatioiishiiis of Auwontodon. 1958 NEW DICHOBUNID ABTIODACTYL 5 Measurements in millimeters of lower jaw and dentition in type specimen of Auxontodon pattersoni, M.C.Z. No. 9316 Deptli of lower jaw at diastema between Pi and \'-i 8.0 Depth of lower jaw beneath M^, lingually 12.1 Ijengtli of cheek tooth series from anterior margin of alveolus for Pi to posterior margin of Ms 46.5 Length of diastema between Pi and Pa 5.0* Length of c-heek tooth series from anterior margin of first alveolus of P2 to posterior margin of Ms 37.2 Length of lower molar series, Mi-Ms, inclusive 19.2 Pi, anteroposterior diameter of root portion at alveolus 4.0" Pi, transvers-e diameter of root ])orti()n at alv(»olus 2. .5 Ps, length: greatest width 6.3^:3.0 P4, length: greatest width 6.8'':3.6 Ml, length: greatest width 6.1:4.4 ''Av>proxiniat'' References Cope, Edward D. 1872. Description of some new Vertebrata from the Bridger group of the Eocene. Palaeont. Bull. No. 1, pp. 1-6 (also, Proc. Anier. Philos. Soc, vol. 12, pp. 460-465, 1873). 1875. On fossil lemurs and dogs. Proc. Acad. Nat. Sei., Philadelphia, pp. 255-256. Gazin, C. Lewis 1952. The lower Eocene Knight formation of western Wyoming and its mammalian faunas. Smithsonian Misc. Coll., vol. 117, No. 18, pp. 1-82, figs. 1-6, pis. 1-11. 1955. A review of the upper Eocene Artiodactyla of North America. Smithsonian Misc. Coll., vol. 128, No. 8, pp. 1-98, pis. 1-18, charts 1-2. Marsh, Othniel C. 1872. Preliminary description of new Tertiary mammals, Parts 1-lV. Amer. Journ. Sci. and Arts, vol. 4, pp. 122-128, 202-224. Matthew, William D. 1909. The Carnivora and Insectivora of the Bridger Basin, middle Eocene. Mem. Amer. Mus. Nat. Hist., vol. 9, pt. 6, pp. 291-567, figs. 1-118, pis. 42-52. OsBORN, Henry F. 1902. American Eocene primates, and the supposed family Mixodec- tidae. Bull. Amer. Mus. Nat. Hist., vol. 16, art. 17, pp. 169-214, figs. 1-40. BREVIORA No. 96 Figure 1. (a) Auxontodon pattersoni, new genus and species, occlusal and lateral view of left ramus of mandible, M.C.Z. No. 9316, with (b) lateral view of left ramus of mandible of Antiacodon pygmaeus, U.S.N.M. No. 18000. All natural size. Drawing prepared by Lawrence B. Isham. L-/ BREVIORA Musceiuim of Coeiparative Zoology Cambridge, Mass. October 27, 1958 Number 97 FUSION OF CERVICAL VERTEBRAE IN THE ERETHIZONTIDAE AND DINOMYIDAE BY Clayton E. Ray Comparative postcraiiial osteology has been a curiously neg- lected source of evidence in the study of relationships among caviomorph and hystricomorph rodents though Landry, 1957, pp. 5-30, has made some use of postcranial osteology on the subordinal level. It seems useful, therefore, to call attention to a peculiarity observed in the cervical vertebrae of the living erethi- zontine (Fig. 1) and dinomyid (Fig. 2) caviomorphs. The second (axis) and third cervicals are firmly fused together in Erethizon, Coendou, Echinoprocta (the postcranial skeleton of Chaetomys is apparently unknown), and Dinomys. This condition was noted by Giebel (1877, p. 273) for Coendou {=Cercolahes) only and has apparently not been recorded in the literature since, having been overlooked, for example, by Scott (1905, p. 422) in con- trasting the skeletons of Sciamys and Erethizon, by Anthony (1926, pp. 124, 126) in his detailed comparison of the skeletons of Elastnodontomys, Erethizon, and Dinomys, and by Swena and Ashley in their recent (1956, pp. 8-10) osteology of Erethizon. The first four cervicals are missing from the type (in part) skeleton of Dinomys figured by Peters (1873, PI. 3, fig. 4). Lan- dry (1957, p. 46) and Fields (1957, pp. 337, 341, 384) have mentioned but not described the tendency toward fusion of cervi- cal vertebrae in Dinomys and in Olenopsis.'^ Cervical vertebrae of the following caviomorph and hystrico- morph genera have been available for examination or are ade- 1 I'atterson (personal conamuiicatioii) (juestions Fields' (1957, p. o23) assiem- ment of his Colombian material to the genus Olenopsis. Thus, even though the animal describert is a ilinomyifl, it is not necessarily true that Olenopsis is a dinomyid. 2 BRKVTOKA No. 97 (juately described in the literature : Acanthion, Thecurus, Hysfri.r, Athcriirus, Triclnjs, Erethizon, Coendou, Echinoprocta, Kocardia (Scott, 1905, p. 465), Schistomys (Scott, 1905, p. 482), Cavia, Kcrodon, Galea, PrndoUchotis (Fields, 1957, p. 373), Dolchotu, Tlydrochoerus, T>in()my>i, " Ohnopsis'" (the material from the La Veiita of Colombia, not the Santa Vvwz of Pata- li'onia), I'Jlastnodonfomys, Amhlyrhiza, Cunicnlus, Dasyprocta, Myoprocta, Periniys (Scott, 1905, p. 437), Lagostomus, Lagidium, ChinchiUa, Ncoreomys (Scott, 1905, p. 395), Capromys, Geocap- vomys, Plagiodontia, Myoeasior, Sciamys (Scott, 1905, p. 422). Ctenomys, Ahrocoma, Isolohodon, ProecJwnys, Hoplomys, Iso- thrix, Echimys, Dactylomys, Thryonomys, and Bathyergus. Included in this list is at least one representative from every family of caviomorph and hystricomorph rodents (as given by Wood, 1955) excepting the Petromuridae. It is not likely that the condition of the cervicals in this monotypic x^.frican family would alter the general picture in any way. In none of the genera listed, save the three erethizontids and Dinomys, was cervical fusion noted. ERETHIZONTINAE The fusion in Erethizon is typical of the .Erethizontinae and will serve as the basis for description. Some two dozen individuals were examined. The neural spine of the third cervical is here relatively higher than in mammals generally and is enveloped on either side by the broad lateral lamellae of the neural spine of the axis. The two neural spines become more broadly fused to one another with increasing age (though in part distinguishable in all specimens seen) . In immature individuals the lateral lamellae of the axis are closely applied to the neural spine of the third cervical but are not fused to it, and in one individual (MCZ- 811 "i diverge posterodorsally to form a bifid process. The area of most profound fusion is between the postzygapophyses of the axis and the prezygapophyses of the third cervical, the trace of the articu- lar surface becoming obliterated with age. Although the centra become tightly fused, particularly near the ventral midline, their 2 Throufrhout tliP discussion MCZ reCcis lo Muscuiii of < '(iiii|)iir;iti\ i' Zoolo^'y. AMXII to .\inoritaii .Museum of Natural History, CXIIM to Cliicaffo Xatural His lory Musouni, ami I'CMl* to Univorsity of California Museuni of I'alcoutoloKy. 190(S FUSION Ol' ('K1J\I(AI>S IN ( AViOAiOKl'l IS 3 union is in part (liscciniblc in all individuals. The pleurapophyses of the axis approach those of the third eervical quite closely, but attain contact in only one individual (MCZ 81!)), in which the distal extremity of the rifjfht plenrapophysis (only) of the axis touches and is narrowly fused to tiie dorsolateral surface of the correspond in and by Shadle and Po-Chedley, l!t4'.l. p. 17."., fig. 1). whereas I am able tn find such barbs in none of the five genera of Ilystricidae (Loweg, 1900, p. 85.3 notes their absence in Hystrix crintdtn and none are described liy Lochte. 19.t7 in his detailed study of the quills in Hifi^trir hirKiitiroxtris) . 8 BREVIORA No. 97 pointed out that the serological, parasitological, and (Deseadan) paleontological information is unknown for dinomyids, which therefore cannot on these grounds be summarily lumped with "other caviomorphs" as opposed to erethizontids. Aside from the similarity in cervical fusion demonstrated here, Pocock (1926, p. 228) has noted a "tolerably close resemblance" be- tween the feet of Dinomys and of Erethizon and Coendou, and Fields (1957, p. 348) has pointed out the similarity of the audi- tory ossicles in Erethizon, Dinomys, and Olenopsis acquatorialis. Attractive as this evidence is, it furnishes an inadequate basis for asserting any special relationship between erethizontids and dinomyids. Indeed contrary evidence from dentition and cranial osteology seems at present to be more compelling. In this con- nection the relationship between the axis and third cervical in Chaetomys and in the Deseadan erethizontids will, when known, be of great interest in helping to determine, respectively, the affinities of Chaetomys, and the antiquity of fusion in the erethi- zontids. If fusion had already taken place in the Deseadan (Oligocene) erethizontids then it might reasonably be concluded that the fusion in dinomyids is independent since it apparently had not occurred by late Miocene time. Furthermore, one would expect fusion to occur in Chaetomys if it is truly an erethizontid."'' In conclusion, present evidence indicates that the constant and exclusive fusion of cervical vertebrae two and three, together with the morphological features described above, furnishes a valid taxonomic character for the living Erethizontinae. Ap- praisal of the possible broader value of the character awaits evidence from Chaetomys, normal adult Dinomys, and fossil erethizontids (and additional fossil dinomyids). 5 A cautionary, if confusing, example is provided by tlie Pedetidae, in whicli cervical fusion may prove to be irrej^ularly distributed. Hatt (1932, pp. 640, 715) found cervicals two and tliree to be fused in Pedetes surdastcr but not in P. cajir (as Hatt was well aware, this character should be checked on more specimens 1958 FUSION OF CERVICALS IN CAVIOMORPHS d ACKNOWLEDGMENTS 1 wish to thank the following people for assistance during the preparation of tliis paper : Professor Bryan Patterson for read- ing an early stage of the manuscript and for examining certain caviomorphs and hystricomorphs in the U.S. National Museum, Dr. Ernest E. Williams for reading the manuscript and for encouraging the undertaking of the study, Professor Albert E. Wood for information on Platypittamys, Dr. D. Dwight Davis for observations on Ahrocoma in the Chicago Natural History Museum and for the loan of Dinomys skeletons, Dr. Donald E. Savage for the loan of the cervical vertebrae of Olenopsis aequa- torialis in the University of California Museum of Paleontology, Drs. Stuart 0. Landry, Jr., and Robert W. Fields for useful information via correspondence, and Miss Barbara Lawrence and Dr. Charles P. Lyman for free access to the collections in their care at the Museum of Comparative Zoology. The illustrations were made by Mrs. Patricia Chaudhuri. The work was carried out while on the tenure of a National Science Foundation pre- doctoral fellowship. LITERATURE CITED Anthony, H. E. 192(i. Mammals of Porto Rico, living and extinct — Rodentia and Eden- tata. New York Acad. Sci., Scientific Survey of Porto Rico and the Virgin Islands, vol. 9, pt. 2, pp. 97-241. Fields, R. W. 1957. Hystricomorph rodents from the late Miocene of Colombia, South America. California Univ. Geol. Sci. Pub., vol. 32, no. 5, pp. 273-404. GlEBEL, C. G. 1877. 771 H. G. Bronn's Klassen und Ordnungen des Thier-reichs, vol. 6, sect. .T, nos. 15, 16, pp. 257-304. Hatt, R. T. 1932. The vertebral columns of ricochetal rodents. Bull. Am. Mus. Nat. Hist., vol. 63, art. 6, pp. 599-738. Landry, S. O., Jr. 1957. The interrelationships of the New and Old World hystricomorph rodents. California Univ. Zool. Pub., vol. 56, no. 1, pp. 1-118. 10 BREVIORA No. 97 LOCHTE, T. 1957. Die Haare, Borsten und Stacheln des Stachelschweines (Hystri.r Jiirsutirostris) . Der Zoologisehe Garten, Leipzig, vol. 23, no. 1/3, pp. 145-162. LOWEG, T. 11)00. Studien iiber das Integument des Erethizon dorsatus {Erethizon dorsatum Cuvier). Jenaische Zeitschrift, vol. 34, no. 4, pp. 833- 866 (n.s., vol. 27). Macinnes, D. G. 1957. A new Miocene rodent from East Africa. Brit. Mus. (Nat. Hist.). Fossil Mammals of Africa, no. 12, pp. 1-35. Moody, P. A. and D. E. Doniger 1956. Serological light on porcupine relationships. Evolution, vol. 10, no. 1, pp. 47-55. Parsons, F. G. 1894. On the myology of the sciuromorphine and hystricomorphinc rodents. Proc. Zool. Soc. London, pp. 251-296. Peters, W. C. H. 1873. Ueber Dinomys, eine merkwiirdige neue Gattung von Nagethieren aus Peru. Festschrift zur feier des Hundert-Jiihrigen Bestehens der Gesellschaft Naturforschender Freunde zu Berlin, pp. 227- 234. POCOCK, R. 1. 1922. On the external characters of some hystricomorph rodents. Proc. Zool. Soe. London, 1922, pp. 365-427. 1926. The external characters of a young female Dinomys hranickii exhibited in the Society's gardens. Proc. Zool. Soc. London, 1926, pp. 221-230. Scott, W. B. 1905. Mammalia of the Santa Cruz beds, III. Glires. Rept. Princeton Univ. Exped. Patagonia, 1896-1899, vol. 5, pp. 384-489. Shadle, A. R. and D. Po-Chedley 1949. Rate of penetration of a porcupine spine. Jour. Mammalogy, vol. 30, no. 2, pp. 172-173. Simpson, G. G. 1945. The principles of classification and a classification of mniumnls. Bull. Am. Mus. Nat. Hist., vol. S5, pp. 1-349. 1958 FUSION OF CERVICALS IN CAVIOMORPHS 11 Stromer, E. 1926. Restf Land- und Siisswasser-bewohnendcr Wirbelticre aus den Diamantenfeldern Deutsch-Siidwestafrikas. Chap. XXI, pp. 107- 153 in Kaiser, E., 1926, Die Diamantenwiiste Siidwestafrikas, II: Berlin, 535 pp. SwENA, B. and L. M. Ashley 1956. Osteology of the common porcupine. Dept. Biol. Sci. and Biol. Station, Walla Walla College, Washington, no. 18, pp. 1-26. TULLBERG, T. 1899. IJber das System der Nagethiere, eine phylogenetische Studie. Nova Acta Reg. Soc. Sci. Upsaliensis, vol. 18, sect. 2, no. 1, pp. 1-514. Vanzolini, p. E. and L. R. Guimaraes 19.')5. Lice and the history of South American land mammals. Rev. Brasileira de Entomologia, vol. 3, p]). 13-46. Wood, A. E. 1950. Porcupines, paleogeography and parallelism. Evolution, vol. 4, no. 1, pp. 87-98. lit.!;"). A revised classification of the rodents. Jour. Mammalogy, vol. 36, no. 2, pp. 165-187. Wood, A. E. and R. R. White, III 1950. The myology of the chinchilla. Jour. Morphology, vol. SO, no. 3, pp. 547-598. 'V ,^ cd .-""N h pq a. -*j »« b a> ^^ -fj CIS c« Si J > »> •-^ « << tl >•— ^ TS ^ h 1— ( '.a -*j - • Ed (M OJ -a x S •=! X O 03 ^ 1— 1 to £b GO <4^ ^*^ •q « 03 C 'U ^-H X ■< 3 '■*-! ID ?d JH 00 ■«-> 1— ( o in 0) > ^ M o O 'iH :§' rig o ^ ft 1 O fQ ^^ so 3 s C3 "3 f$ -*j =t-i fl o Qj > '^^ ^ m <» ^ tl '3 f» t< m rt t. hJ '? t. ^,-^ a> <1 v , r^ C-l •^ X ci ^ !>-. ■4-» ^3 03 D O § s m O 'S ^ c« a 1^ CAA-f BREVIORA Museiiinn of Comparative Zoology Cambridge, Mass. December 12, 1958 Number 98 TWO NEW SPECIES OF BATHYLAGUS FROM THE WESTERN NORTH ATLANTIC WITH NOTES ON OTHER SPECIES By Daniel M. Cohen Ichthyological Laboratory, U.S. Fish and Wildlife Service U.S. National Museum, Washington, D.C. While reviewing the Atlantic bathylagids for the forthcoming volume of "Fishes of the Western North Atlantic," two appar- ently unnamed species were encountered and are herein described as new. In addition, notes on several related bathylagids are presented. Bathylagus compsus, sp. nov. Figure 1 Bathylagus glacialis Beebe {-non Re^an), 1933, j). 114. Holotype. U.S. National Museum 171755, 44.4 mm. in stand- ard length; collected by Richard H. Backus on board the "Blue Dolphin," haul RHB 457, July 17-18, 1953, Isaacs-Kidd mid- water trawl. Trawl down at 2255 hours, up at 0210 hours. Depth 69.5 m. Locality, within a circle 12 miles in diameter with the center at 39° 45' N., 71° 08' W. Paratype: USNM 171754 (1), 38 mm. in standard length ; data as for the holotype. Diagnosis. A Bathylagus with an extensive gill opening, reach- ing almost halfway up the side of the body; an anal fin base longer than the length of the caudal peduncle ; and dark pigment present on the margins of the scale pockets. Counts and measurements. Measurements given as per cent of standard length, holotype first, followed by the paratype in parenthesis. Dorsal 11 (10), anal 19 (20), pectoral 9 (9), ven- tral 9 (10), branchiostegals 2 ( — ) ; preanal 75.9 (76.8), preven- BREV'IORA No. 98 tral 52.5 (52.1), predorsal 47.3 (50.0), prepectoral 27.0 (— ). head length 25.0 (26.3), snout 5.6 (5.3), eye 10.0 (9.2), greatest body depth 16.9 (15.3), least depth of caudal peduncle 6.3 (6.6), dorsal fin base 8.8 (6.8). anal fin base 16.9 (15.0), interorbital 9.0(— ). Description. A graceful, elongate fish with the greatest depth a short distance behind the head. The dorsal profile of the head descends from the nape to the interorbital in almost a straight line. It curves downward near the anterior edge of the orbit and Figure 1. USNM 171755. Holotype of Baihylagus compsus. descends to a point above the nares where another inflection occurs causing the profile to drop more steeply to the upper lip. The dorsal rim of the orbit does not enter the dorsal profile in the larger specimen but does in the smaller. The ventral profile of the head rises more gently than the dorsal profile descends. The interorbital area is broadly concave between the supraorbital canals. The maxillary extends back to a vertical from a point slightly in front of the anterior edge of the orbit. Minute papillae are present on the ventral margins of the maxillary and premax- illary. The jaws are rounded and subequal. The eyes are directed laterally and lack adipose eyelids. A crescent of white tissue is present over the iris of the posterior half of the eye. The height of the gill opening is slightly less than half the greatest height of the body and extends beyond the dorsal edge of the pectoral fin base by a distance equal to two and one-half times the length of the pectoral fin base. In his description of post-larvae of this species, Beebe (1933, p. 119) stated ''gill openings characteristically small"; however, he gave no quanti- tative data, nor did he discuss this character in larger specimens. 1958 NEW SPECIES OF BATHYLAGUS 3 The pectoral fin inserts a short distance behind the head on the ventrolateral contours of the body. The dorsal fin orip:inates close to the midpoint of tlie body and has its base raised above the body. The ventral fins insert on the ventral surface of the body under the posterior part of the dorsal fin base. The anal fin base is set off from the body and is longer than the dorsal fin base and longer than the length of the caudal peduncle. The adipose fin is placed over the posterior end of the anal fin base. The fins are all broken off short. The pigmentation of specimens in alcohol is striking and serves as an important diagnostic character, at least in adolescent specimens. The ground color is a light yellow-brown. The muzzle, the opercular apparatus and a ring around the orbit are dark brown. The opercle is overlaid with an iridescent sheen and is probably silvery in life. The iris is iridescent gun-metal blue and may also be silvery in life. The dark lining of the peritoneal cavity shows through the body wall and outlines the body cavity. A single row of large, dark chromatophores runs the entire length of the body slightly below the lateral line. A second and less distinct line parallels the first line a short distance above the lat- eral line. The right side of the paratype shows a striking resem- blance to Beebe's (1933, p. 118) figure of an adolescent specimen. The left side of the paratype and both sides of the holotype vary in that the anterior portion of the upper line is lost in a scatter- ing of smaller, lighter chromatophores. The venter is lightly pep- pered with small, dark chromatophores which are more densely distributed on the dorsum. Brown pigment cells are also present on the dorsal surface of the head behind and before the interor- bital area and on the cheeks. The caudal, dorsal and adipose fins have scattered brown chromatophores, while the anal, ventral and pectoral fins are immaculate or bear only a very few pigment cells. In both specimens the posterior half of the body has the ragged remnants of scale pockets colored by large, dark chromatophores (not shown in Fig. 1). Relationships. Bathylagus compsus appears most closely re- lated to B. nigrigenys and B. longirostris, the only other known species of Bathylagus with extensive gill openings and elongate anal fin bases. The types of B. compsus differ from specimens of B. nigrigenys of comparable size in having more anal rays, 19 4 BREVIORA No. 98 to 20 in conipsus, 18 to 17 in nigrigenys; a shorter preventral distance, 52.1 to 52.5 in conipsus, 56.7 to 61.4 in nigrigenys; in having- a less deep body, 15.3 to 16.9 in compsiis, 18.9 to 22.6 in nigrigenys. In addition, adolescent nigrigenys apparently lack the striking- pigment pattern found in compsus. Other less trenchant differences are also present. I have examined neither adolescent specimens of B. longiros- tris nor adult specimens of B. cumpsus and so hesitate to com- pare proportional measurements ; however, B. Inngirostris differs markedly from conipsus in its complete lack of pigment on the scale pockets. Discussion. In his studies on Bermuda Bathylagus, Beebe (1933) identified one of the forms upon which he reported as Bathylagus glacialis Regan, a species originally described from the Antarctic. More recent studies by Norman (1937) and Cohen (MS.) strongly suggest that B. glacialis Regan is a syno- nym of B. antarcticus. I have examined a series of B. anfarcticus from southern waters, and I find they differ from Beebe 's de- scriptions of his Bermuda material. Unfortunately, I have not been able to obtain Bermuda specimens for direct comparison ; however, the two specimens at hand differ so markedly from B. antarcticus (which is probably restricted to far southern seas) and agree so well with Beebe 's description that I feel no qualms at referring Beebe 's glacialis to Bathylagus compsus. Beebe (1933) also presented an osteological study of this species, and there seem to be a number of discrepancies between his findings and those of Chapman (19-43), who presented an osteological study of Bathylagus pacificus. The differences ap- pear so marked that they could be used to separate genera or even higher taxa. I was fortunate in having at my disposal the identical cleared and stained specimens upon which Beebe based his descriptions (Stanford University 45302, from Bermuda) and I was able to note several points which require comment. These are briefly noted since they are of some importance in the classification of the bathylagids. 1. The frontals : Beebe states that the frontals are fused and he also figures them as such. I find that the frontals are not fused but instead are slightly sep- arated in the midline. The underlying cartilage has taken up some of the stain and the edges of the bone are difficult to see 1958 NEW SPECIES OF BATHYLAGUS 5 unless they are gently lifted. 2. The sphenotie : The bone which Beebe labels the pterotic in his dorsal view of the skull (Fig. 38) is actually the splienotic and is so labeled in the lateral view (Fig. 3G). ■I. Supraniaxillary : Beebe shows supraniaxillaries in Figure 36, although he does not mention them in the text. I find no supraniaxillaries, but instead a more heavily stained ridge which may have been mistaken for a joint. 4. Branchio- stegals : Although he shows two branehiostegals in all of his figures and I find two branehiostegals in his specimen, Beebe gives three or four branehiostegals as a generic character. I know of no species of recent Bathylagus with other than two branchiostegal rays. Hubbs (1919) notes a single branchiostegal ray in Bathylagus pacijicus; however, I have found two in this species. Chapman (1943) also reports two. With regard to general shape and placement of bones, I find Bathylagus compsus in good agreement with Chapman's account of B. pacificus. Since Beebe did not dissect his specimen, his account of the deeper bones of the skull is necessarily sketchy. The name compsus is from the Greek kompsos, elegant or pretty, and refers to the general appearance of this attractive sjjecies. Bathylagus geeyae, sp. nov. Figure 2 Holotype. Chicago Natural History Museum 49730, 71 mm. in standard length; collected by the "Caryn," haul 38, August (5, 1948, 35 foot otter trawl. Trawd in at 2245 hours, out at 0200 hours. Depth of haul 500-550 m., 1500 m. wire out, bottom depth 2000 to 2560 m. Locality, Bermuda, 32° 13.5' N., 64° 32.5' W. Paratypc. Stanford University 50974 (1), 35.5 mm. in stand- ard length, collected by the Bermuda Oceanographic Expedi- tions of the New York Zoological Society, 1929-1930, net 35, April 24, 1929. Net down at 9 :54 A.M., duration of haul 3 hours, 36 minutes. Depth of haul 1646 m. to surface. Locality, Ber- muda, an 8 mile circle with its center at 32° 12' N., 64° 36' AV. Diagnosis. A Bathylagus with an extensive gill opening, reach- ing more than halfway up the side of the body ; an anal fin base equal to or shorter in length than the length of the caudal peduncle ; and dark pigment present on the margins of the scale pockets. 6 BREVIORA No. 98 Counts and Measurements. Measurements given as per cent of standard length, holotype first, followed by the paratype in parenthesis. Dorsal 11 (13), Anal 13 (13), pectoral 12 (13), ventral 10 (11), branchiostegals 2 ( — ), gill rakers on lower arm of first arch in holotype 16, preanal 81.0 (79.1), preventral 58.4 (61.1), predorsal 48.2 (50.7), prepectoral 28.9 (32.7), head length 25.3 (29.6), snout 5.9 (7.0), eye 9.4 (see description for comments on eye of paratype), greatest body depth 13.4 (16.9), least depth of caudal peduncle 8.4 (8.4), dorsal fin base 12.0 (11.8), anal fin base 9.7 (9.9), lateral line scales probably be- tween 45 and 50 on both specimens. Description. An elongate and much laterally compressed fish with the greatest depth behind the head, tapering very little to the caudal peduncle. The dorsal profile of the head slopes gently downward in almost a straight line from a point behind the eye to a point near the nostrils, from whence it descends more steeply. The dorsal rim of the orbit projects into the dorsal pro- file of the head. The interorbital area between the supraorbital canals is broadly rounded. A well developed pineal apparatus Figure 2. CNHM 49730. Holotype of Bathylagus greyac (this figure is a reconstruction of the somewhat distorted holotype and should not be used for measurements). Lower figure is of the opercle of Bathylagus ochoien.'^i.^. is present and is visible through a clear window in the deeper pigmentation of the braincase. A broad ring of dense pigment about the window gives the appearance of an iris. The posterior edge of the maxillary does not reach a vertical from the anterior edge of the orbit. Miimte papillae are present on the epidermal covering of the maxillary, premaxillary and dentary. The jaws are rounded and equal. 1958 NEW SPECIES OF BATHYLAGUS 7 The mouth of the holotype is damaged, making it difficult to discern any details of dentition; however, the single row of upper teeth are conical and apparently restricted to the head of the vomer. The compressed dentary teeth are smaller than the vomerine teeth with the exception of a pair of small, spike-like teeth which project at the symphysis of the dentaries. The eyes are directed laterally and lack adipose eyelids. A crescent of white material is present over portions of the iris. The left eye of the paratype presents certain peculiarities. It has a horizontal diameter of 3.5 mm. compared to 4.2 mm. for the right eye. In addition, the left eye has two small lenses visible through the pupil. The height of the gill opening is more than half the greatest height of the body and extends beyond the dorsal edge of the pectoral fin base by a distance equal to about two and one-half times the length of the pectoral fin base in the holotype and about one and one-half times in the paratype. The pectoral fin inserts a short distance behind the head on the ventro-lateral contours of the body. The dorsal fin originates close to the mid-point of the body. Both the dorsal and anal fin bases are set off from the body. The dorsal fin base is slightly longer than the anal fin base. The anal fin base is equal in length to the length of the caudal peduncle. The ventral fins insert on the ventral surface of the body under the posterior part of the dorsal fin base. The dorsal adipose fin is placed over the center of the anal fin base. A weak ventral adipose fin is present directly anterior to the vent of the paratype. The ventral half of the holotype has a light, straw-colored ground color. The dorsal half of the fish is brown, while the operele and the snout are dusky. Small, dark-brown ehromato- phores are scattered at random over the sides of the fish on a field of still smaller, more densely distributed, light-brown chromatophores. The sides of the head and all of the fins are lightly peppered with dark-brown pigment cells. Dark pigment is also thickly- distributed on the margins of the scale pockets. Some of the scale pockets bear remnants of a purplish, trans- lucent material which is common to a number of Bathylagus species. Tlie paratype has faded badly, and little in the way of pigmentation remains. 8 BREVIORA No. 98 Relationships. B. greyae is closest to Bathylagus ocJiotensis (which has B. nakazaivai as a synonym, Cohen, MS.). It differs in having a greater preanal distance, 81.0 in the type of greyae, 73.7 and 77.6 in two specimens of ochotensis of comparable size; in having a longer preventral distance, 58.4 to 61.1 in greyae, 51.3 to 57.1 in 17 specimens of ochotensis; in having the anal fin base slightly shorter than the dorsal fin base, while in ocJiotensis the anal fin base is longer than the dorsal fin base. In addition, the postero-ventral margin of the opercle is smooth in greyae, while in ochotensis it is drawn out into five, finger-like projec- tions (Fig. 2). This species is named for Mrs. Marion Grey of the Chicago Natural History Museum in recognition of her contributions to the study of deep-sea fishes. Bathylagus longirostris Maul Until the present time this species has been known only from the type, apparently taken at Madeira. I have examined a single specimen from the Gulf of Mexico, south of Pensacola, Florida, collected by the U.S. Fish and Wildlife Service vessel "Oregon" and deposited in the Chicago Natural History Museum. In addi- tion, two specimens collected in the Central Pacific (30° 16.5' N., 179° 54' E. and 29° 54' N., 168° 32' W.) by the Pacific Oceanic Fishery Investigations (POFI) are apparently referable to B. longirostris. Descriptions of the above specimens will be included in a projected revision of the Bathylagidae. Bathylagus bericoides (Borodin) This species has been previously recorded three times, each from the Western North Atlantic. I have examined an unre- corded specimen from Bermuda (SU 42715) and a specimen from "Oregon" st. 1028 (Chicago Natural History Museum). The latter was provisionally identified as Bathylagus micro- cepha.lus by Springer and Bullis (1956). B. bericoides is also present in POFI collections from the central Pacific (39° 31' N., 178° 54' W.; 39° 08' N., 164° 51' W., and 42° 16' N, 179° 52' W.). These specimens will be treated at greater length in the future. 1958 NEW SPECIES OF BATHYLAGUS 9 ACKNOWLEDGMENTS 1 am indebted to Dr. Richard Backus, Mrs. Marion Grey, Dr. George Myers, Miss Margaret Storey, Dr. Donald Strasburg, Dr. Denys Tucker, and Mr. Loren Woods for the loan of speci- mens. I thank Mr. G. E. Maul for information on the type of B. longirostris and Miss Esther Google for preparing the figures. LITERATURE CITED Beebe, William 1933. Deep-sea fishes of the Bermuda Oceanographic Expeditions. Family Argentinidae. Zoologica (New York), 16: 97-147. Chapman, Wilbert M. 1943. The osteology and relationships of the l)athypelagie fishes of the genus Bailiylagus Gunther with notes on the systematic position of Leuroglossus stilbius Gilbert and Therobromus caUorhinus Lucas. Journ. Washington Acad. Sci., 33: 147-160. HuBBS, Carl L. li)19. A comparative study of the bones forming the opercular series of fishes. Journ. Morph., 33: 66-71. Norman, John R. 1937. Fishes. Rep. Br. Aust. N.Z. Antarct. Res. Expedn. 1929-31, Adelaide, (B) 1 (2): 49-88. Springer, Stewart and Harvey Bullis 19.36. Collections by the Oregon in the Gulf of Mexico. Dept. of the Interior. U.S. Fish and Wildlife Service. Spec. Scient. Repts. — Fisheries, 196: 1-134. Taaning, Aaqe V. 1931. Postlarval stages of Bathylagus from the North Atlantic. Vi- densk. Medd. fra Dansk Naturh. Foren., 92: 269-274. BREVIORA Miasemm of Compsirsitive Zoology Cambridge, Mass. December 19, 195S Number 99 A NEW SUBSPECIES OF CHAMAELEO JACKSONI BOULENGER AND A KEY TO THE SPECIES OF THREE-IIORNED CIIAMAELEONS By a. Stanley Rand Biological Laboratories, Harvard University A small collection of reptiles collected on Mt. Meru, Tangan- yika Territory, in August 1956, by Lt. Col. J. Minnery was sent to the Museum of Comparative Zoology by Mr. C. J. P. lonides. It included three male chamaeleons that Mr. Loveridge believed to represent a new form but did not describe since he desired to confirm the characters on a large series. Fourteen additional specimens were obtained through Mr. lonides in 1957 and seven in 1958 after Mr. Loveridge had retired as Curator of Herpetol- ogy. Five of these last have been sent to the British Museum (Natural History). These twenty-four specimens, together with an additional one already in the collection of the Museum of Com- parative Zoology, form the basis on which the following form is described. (The name has been suggested hy Mr. Loveridge.) ChAMAELEO JACKSONI MERUMONTANA SUbsp. nov. Type. A male, Museum of Comparative Zoology No. 56152 (collector's No. 7651), apparently adult, collected at Laikinoi, near Arusha, Mt. Meru, Arusha District, Northern Province, Tanganyika Territory, at 7500 ft. alt., August 1957 by Lt. Col. J. Minnery. Paratypes. Twenty-four. MCZ 44327, male, Mt. Meru, east at 9000 ft. alt., 1938, Cooper collector ; MCZ 54790-92, males, same locality as type, August 1956, Lt. Col. J. Minnery collector ; MCZ 56153-59, males and 56151, 56160-64 females, same data as type; 2 BREVIORA No. 99 MCZ 56989 male and 56990 female, same locality as type, Janu- ary 1958, Lt. Col. J. Minnery, collector; BM (5 specimens), same locality as type, January 1958, Lt. Col. J. Minnery, collector. Diagnosis. A chamaeleon with three annulated horns in the male, closest to Chamaeleo jacksoni Boulenger, but differing from it in smaller size, in having the occipital and postorbital areas covered with smooth or weakly convex scales (rather than strongly convex scales) and in having fewer and larger scales between the preocular horns. Description. Three annulated horns, one rostral and two pre- ocular in position and subequal in length. (In males only, females with only a rostral horn, proportionately shorter and with two sharp horn scales instead of the preocular horns.) Preocular horns arising from the anterior end of the raised supraorbital ridges and separated by a mosaic of enlarged flattened scales, a few of them with antero-posteriorly oriented ridges. (This ridg- ing is more marked in some of the paratypes than in the type.) The minimum number of scales counted between the preocular horns (males) or horn scales (females), 7 in the type, 6 or 7 in the paratypes (see table below). The supraorbital ridges continued posteriorly as low ridges to the posterior border of the head, separating the posterior dorsal and lateral head scales into postorbital and occipital areas. These areas are covered with enlarged scales, some of which are flat and others weakly convex. (The number of completely flat scales in these regions varies in the paratypes but never are there many strongly convex scales present.) A moderately low curved profile to the casque ; a parietal crest present, two or three scales high, forked anteriorly to form low ridges that extend antero Jaterally toward the supraocular ridges, but do not meet them. These branches of the parietal crest sep- arate the large flat occipital scales from an area of small strongly convex scales that are distinct from both the occipital scales and the interorbital scales anterior and lateral to them. No occipital lobes; a very low ridge marking the posterior border of the casque and separating the head scales from the smaller nape scales. Small, soft, triangular tubercles on the upper portion of the eyelid. 1958 NEW SUBSPECIES OF CHAMAELEON 3 A median dorsal crest of enlarged scales extending from the shoulder region nearly to the base of the tail, these trianguloid scales largest and most pointed anteriorly, becoming smaller and more rounded posteriorly, separated by one to four scales. In some cases the small scale immediately in front of the large one is distinctly larger than the others but it never approaches the largest in size or shape. No crest on the tail. The scales on the rest of the dorsal and lateral surfaces markedly heterogeneous in size on both body and tail, with an irregular arrangement of large flat scales among small granular ones and intermediates in size and convexity. The dorsal surfaces of the limbs similarly scaled but with large plates predominating, particularly on the forelegs. The ventral surface covered by uniform granules with some very tiny ones between them, except for the sides of the throat where there are elongate patches of the very small granules. Xo gular or ventral crest evident. Color (in alcohol) : The horns, the interorbital and snout scales, the supraorbital ridges, and the anterior enlarged dorsal crest scales light brown. The rest of the animal blue, lighter dorsally and on the head, limbs, and tail and darker on the lips, sides of throat and sides and venter of the body. (All the male paratypes show this distribution of light yellow brown but vary in the intensity of blue on the rest of the body. The females are blue with the rostral horn dark brownish.) Measurements: See Tables 2 and 3. Discussion. The Mt. Meru specimens described above are most similar to Chamaeleo jacksoni Boulenger which was described from Nairobi. However, they differ in a number of respects. The comparison below is based on eighteen adult specimens of the typical race from various localities in Kenya Colony. The most striking difference is the small size of the Mt. Meru series. The largest male measures 91 mm. in snout-vent length, the largest female 86 mm. In the Kenya series of eighteen, eleven measure 97 mm. or over, snout-vent length. The largest male is 167 mm., the largest female 140 mm. In merumontana the occipital and postorbital scales are flat or a few of them weakly convex, giving a smooth appearance to the occipital and postorbital areas. In the Kenya specimens many 4 BREVIORA No. 99 of these scales are strongly convex giving a bumpy appearance to these areas. In the Mt. Meru series, the scales between the preocular horns (interorbital scales) are larger and fewer in number than in the Kenya specimens. The table below shows the specimens arranged according to the least number of scales counted betw^een the horns. Table 1 SCALES BETWEEN PEEOCULAE HOENS OE HOBN SCALES "lumbers of scales 6 7 8 9 10 11 12 s'uinljc'i- of JNIt. i\Ieiu S 17 0 0 0 0 0 specimens with each scale count Kenya 0 1 3 8 5 0 1 In the Mt. Meru series the large fiat interorbital scales (some of them sometimes having a median keel) differ sharply from the mure posterior small strongly convex scales between them and the branches of the parietal crest. In the Kenya specimens the inter- orbital scales do not differ markedly from the scales behind them. In the Kenya specimens the dorsal crest is formed by groups of two scales, the first of which is smaller than the second, but both are larger than the scales between groups, and both are pointed — at least in part. The enlargement of the anterior scale is much less evident in some of the smaller specimens. In the Mt. Meru specimens the anterior scale in each group is rela- tively smaller. It is usually but little larger than the surrounding scales, so that the crest seems to be composed of single, isolated, enlarged scales. The ridges on the head of the Kenya form are more distinct than those of the Mt. Meru series. This is particularly true of the ridge along the posterior border of the casque and the anterior branches of the parietal crest. The more marked parietal crest gives a noticeably higher profile to the casques of the large males from Kenya. 1958 NEW SUBSPECIES OK CHAM AELEON 5 The body scales of all of the specimens are markedly hetero- geneous. The two forms differ only in that the large scales tend to be flatter in tlie Mt. Mern specimens than in the Kenya speci- mens. In seven of the eight Mt. Meru females the rostral horn is well developed, although shorter than that of a male of equivalent size. The eighth female has a very small rostral horn, and none has preocular horns. In the six large Kenya females, one has both rostral and preocular horns developed. In the other five the preocular horns are 2 nnn. or less. Three of these have a rostral horn and two have only a sharp rostral scale. Loveridge, 1957, gives the range of Chamacleo jacksoni as "Highlands of Kenya Colony and Tanganyika Territory." Ap- parently his inclusion of Tanganyika was based on the MCZ specimen No. 44327 from Mt. Meru, listed here as a paratype of C. j. merumoniana. No other specimens or records of C. jack- soni from Tanganyika have been located. Therefore, the range of C. j. jacksoni can be restricted to the highlands of Kenya Colony, and that of C. j. mcrumontana can be given as Mt. Meru, between 7500 and 9000 ft. alt. Of the type and those paratypes collected by Lt. Col. J. Min- nery, C. J. P. lonides writes: "They are found on bushes and in low small trees. Laikinoi is a farm on the very edge of the rain forest." (Letter dated 5 December 1957.) Three-horned chamacleo7is. In the course of describing this form, specimens of the five other species of three-horned chamael- eons were examined. A few notes and a key for the identification of these species are included below. The six species characterized by the possession of three annu- lated horns in the male do not seem particularly closely related. The other morphological characters that they have in common are negative rather than positive : lack of spurs in either sex ; lack of axillary j^ockets. Among them, C. jacksoni seems closest to C. johnstoni Boulen- ger, again primarily on negative characters : absence of occipital lobes, absence of a gular crest, absence of a dorsal fin, but also in the possession of heterogeneous body scales. C. wcrneri Tornier and C. fullehorni Tornier seem quite closely related to eacli other, though not to either C. jacksoni or C. 6 BREVIORA No. 99 johnstoni. The only other close relationship involving these species seems to be between C. johnstoni and C. ituriensis Schmidt, the latter a hornless chamaeleon of the eastern Belgian Congo forest. C. oweni Gray is widespread throughout the lowland rain forest in West Africa extending into eastern Belgian Congo. All the others are mountain forest forms in East Africa : C. fulle- horni from the Ngosi Volcano, Poroto Mountains, Ukonde, Tan- ganyika Territory; C werneri from the Uluguru and Uzungwe Mountains, Tanganyika Territory; C. johnstoni from the virgin forest of Uganda, Belgian Ruanda-IJrundi and the adjacent Belgian Congo ; C. deremensis Matschie from the Usambara and Uluguru Mountains, Tanganyika Territory; C. jacksoni from the highlands of Kenya Colony and Mt. Meru, Tanganyika Terri- tory. It is noteworthy that except in the Uluguru Mountains no two species of three-horned chamaeleons are known to occur together. This is evidence supporting the suggestion that the three horns serve these chamaeleons as a means of species recognition. The case of the two species occurring together on the Uluguru Moun- tains supports rather than contradicts this view. These two chamaeleons are very different in appearance. C deremensis is large, laterally compressed and with a high dorsal fin ; C. werneri is smaller and stocky with no dorsal fin. Even with the presence of three horns in both there could be little opportunity for confusion. In eastern Belgian Congo where a close relative of the three-horned C. johnstoni, C. ituriensis, occurs with the three- horned C. oweni, both species are similar in size and proportions, but C. ituriensis is hornless. Key to the three-horned chamaeleons 1. High sail like dorsal fin C. deremensis No high sail like dorsal fin 2. 2. Body scales equal in size or nearly so C. oweni Body scales differing markedly in size 3. 3. Well developed occipital lobes 4. No occipital lobes 5. 4. Occipital lobes continuous across back of casque as a flap C. iverneri Occipital lobes separated by a gap in the middorsal region C. fulleborni 1958 NEW SUBSPECIES OF CHAMAELEON 5. Dorsal crest of much enlarged scales in groups of one or two 6. No dorsal crest of much enlarged scales 7. G. Occipital and postorbital scales strongly convex, scales between preocular horns smaller and more numer- ous (7-12), size larger C. j. jacksoni Occipital and postorbital scales flatter, scales between preocular horns larger and fewer (6-7), size smaller C. j. merumontana 7. * Dorsal profile of body smooth or undulating C. j. jo'hnstoni Dorsal profile of body crenulated C. j. crenulata * C. johnstoni crenulata not seen, this section of the key taken from Laurent, 1951. Table 2 MEASUREMENTS OF MT. MERU SPECIMENS w a a — a u a ^ c ** ® -• o . It' tc -—^ 4-1 = t. > ^ Sr S"" = -^^ s= >< 5= i-r ^^ 2§" «S Sz 32 x^:; 5J^j CL(^ S^o wa MCZ 44327 male 68 13 11 22 7 54790 '' 88 19 18 28 6 54791 " 76 16 14 24 7 54792 <' 60 7 6 19 7 *56152 " 91 19 18 29 7 56153 " 73 18 18 24 6 56154 " 78 15 13 26 6 56155 " 75 16 15 25 7 56156 " 73 17 16 24 7 56157 " 90 20 16 29 7 56158 " 71 11 10 24 6 56159 " 64 10 8 20 7 56989 " 86 23 22 26 6 BM " 78 17 15 26 7 " 82 16 15 24 7 "57 9 7 19 7 " 79 14 13 25 7 56151 female 74 7 — 23 7 56160 " 77 7 — 23 6 56161 "80 8 — 22 6 56162 "81 8 — 25 7 56163 "74 8 — 23 7 56164 " 86 6 — 26 7 56990 "78 8 — 24 6 BM "79 2 — 24 7 * Type of C. j. merumontana 8 B[iEVIORA No. 99 Table 3 MEASUREMENTS OF KENYA SPECIMENS u i) s N a 5a a o .a C3 - o a o s Head length (snout to end of casque) it: a a, o "3 ? II Embu Dist lict 7842 male 110 15 15 35 9 7842 female 140 — — 41 12 Meru 11467 male 167 35 33 52 9 29454 I < 118 22 22 41 10 Lukenya 34995 female 99 7 — 31 10 34996 male 111 16 16 36 9 Parklands, Nairobi 18365 male 98 15 13 30 9 18366 female 123 13 6 37 9 18367 It 97 5 2 29 8 31365 t i 103 3 2 32 9 31366 ( ( 105 — _i_ 31 8 LITERATUEE CITED BOULENGER, G. A. 1896. Description of a new chamaeleon from Uganda. Ann. Mag. Nat. Hist., (6)17: 376. Laurent, R. 1951. Deux reptiles et onze Imtraciens nouveaux d'Afrique central. Rev. Zool. Bot. Afr., 44: 360-381. LOVERIDGE, A. 1957. Checklist of the reptiles and amphibians of East Africa. Bull. Mus. Comp. Zool., 117: (2): 153-362, i-xxxvi. BREVIORA MmseiLiini of Comparative Zoology Cambridgk, Mass. December 23, 1958 Number 100 OX THE PINEAL ORGAN OF THE TUNA, THYNNUS THYNNUS L. By Uno Holmgren ZooJogical Institute, Uppsala, Sweden and Biological Laboratories, Harvard University, Cambridge A conspicuous piueal spot has been found in many seombrid fishes, e.g-. Thynnus, Germo, Parathynnus and Neothynnus and the related genera Auxis, Katsuivonus, Enihynnus and Sarda. Underlying- the pineal spot a fontanelle or "pineal window" (Rivas, 1952) is present in the last four mentioned genera and also in Scomber and Scomheromorus. Rivas investigated the pineal apparatus of the tuna, Thynnus thynnus L. and claimed a photoreceptory role of the pineal area, but he "failed to reveal the presence of a pineal organ, which is apparently absent or greatly reduced" (Rivas, 1952, p. 174). My researches show, however, that a jiineal organ is present in the tuna. A supply of the bluefin tuna, Thynnus thynnus L. was obtained from an expedition on board M/S "Delaware," belonging to U.S. Fish and AVildlife Service. Five fishes were dissected, and the pineal organ of each was sectioned and stained, using Bodian's protargol method (Romeis, 1948, p. 421) to determine the presence of nerve elements, and Heidenhain's azan stain (Romeis, 1948, p. 344) for the general histology. The sections were cut at 8 microns, using paraffin technique. Dissections of the Bouin-preserved material of the tuna re- vealed, as had previously been pointed out by Kishinuya (1923), Gregory (1933), Godsil and Byera (1944).^ and Rivas (1952), the presence of a fontanelle, underlying a conspicuous white pineal spot 15 x 30 mm. in size, in a specimen weighing 350 lbs. (Fig. 2). The spot is situated slightly caudally from an assumed 2 I3REVIORA Xo. lUO line eoniiecting tlu^ centers of the eyes. The dermal layer over the pineal fontanelle was found to be transhieent, transmittininous tyi)e of tissue. A tubular excavation around the pineal organ was formed below the pineal fontanelle, leading down to the brain at an angle of 60 or 70 degrees (Fig. 1). The excavation had a diameter of 10 mm. The end-vesicle of the pineal organ was situated immedi- ately below a dome-shaped cartilage (Figs. 1, 4). The vesicle was approximately 10 mm. in diametci- and connected to the diencephalon by a stalk. \'entrally the stalk had a white cord which distally seemed divided into two branches ; in the specimen photographed the stalk was 30-35 unn. in length (Fig. 1). The histological examination of the end-vesicle showed a struc- ture similar to that of most other teleost fishes. The lumen, hoAvever, was larger and the walls were thinner, consisting of only one or two layers of cells (Figs. 4, 5). Because the pineal cells were secretory and had cytoplasmic processes extending into the lumen of the vesicle, the cells of the end -vesicle and the stalk (Figs. 4, 5) were similar to the primary sensory cells in teleosts described by N. Holmgren (1920) and *Friedric'h-Freksa (1932). There were also present su])porting cells ("8tiitz/,ellen" Fried- rich-Freksa. 1932) which had more hyperchromatic nuclei and more indistinct cytoplasm than the sensory cells. A cross-section of the pineal stalk revealed, especially in the distal part, folded walls. Connective tissue, stained blue with lleidenhain's stain, formed a thin layer inside and outside the wall of the stalk. Silver impregnation did not give any evidence of axons eitlier in the pineal vesicle or in the stalk. A tractus pinealis is ])resent, at least in the ])i'()ximal |)ortion of tlic ])ineal stalk, in many teleost fishes. Apparentlv the tuna is one of the exceptions. 1958 I'INEAL ORGAN OK THE TINA 3 Tlic pineal structures in the tuna were t'ound to i)e very vuscil- lar, osj)e('ially below tiie eiul-vesiele of tiie pineal organ ( Fir)(i. Vohiiiu'tiic jind rytologic vaiiat icDi in the piiR'al IkkIv of Pcro- myscu^ leucopu,^ (Rodentia) with respeit to sex, captivity and day-length. Jour. Morph.. 98: 471-498. R.ASQUIN, P. 1958. Studies in the control of pigment cells and light reactions in recent teleost fishes. Bull. Amer. Mus. Nat. Hist., 115: 1-68. RiVAS, L. R. 19r)2. The pineal apparatus of tunas and related sconibrid fishes as a possible light receptor controlling phototactic movements. Bull. Mar. Sci. Gulf and Caribbean, 3: 169-180. ROMEIS, W. 1948. Mikroskopische Technik. l.'tth ed. Munich. SCHARRER, E. 1928. Die Lichtempfindlichkeit blinder Elritzen. (Untersuchungen iiber das Zwisehenhirn der Fisehe). Zeitschr. vergleich. Physiol., 7: 1-38. Young, J. Z. 1935. The photoreceptors of lampreys: II. The functions of the pineal complex. Jour. Exper. Biol., 12: 254-270. Fig. 1. ]Jissec-ted .spei-iir.eii showing tlic end -vesicle of tlie jjineal organ and the i)ineal stalk, ending in tlie diencephalon. Tliis pieture also shows tlie tulielike excavation extendintj- from tlie iipi>er edges of the frontal lione to the brain. Photo 3.") x. Fig. 2. The pineal spot of Thynnus iliynnus. Specimen of 350 lbs. Natural size. Fig. ?y. The pineal fontanelle in the frontal bone. A jelly-like material is present in the fontanelle. Fig-. 4. Part of the pineal organ situated below the cartilaginous tissue mentioned above. Note the conspicuous lumen of the organ and the presence of blood vessels below the end-vesicle. Fixation Bouin 's solution. Heiden- hain 's azan stain. 150 x. Fig. 5. Section of the wall of the end-vesicle. Cytoplasmic processes are extending into the lumen of the pineal organ (arrow). Fixation Bouin 's solution. Heidenhain's azan stain. 1200 x. Fig. 6. Cross-section through the distal part of tlie pineal stalk, showing its folded walls. Fixation Bouin 's solution, Heidenhain's azan stain. 300 x. BREVIORA Museiiim of Comparative Zoology Cambrid(;e, Mass. March 2, 1959 Number 101 CERVICAL RIBS IN TURTLES By Ernest E. Williams The existence of cervical ribs in Recent turtles is frequently denied. In fact, however, cervical ribs in turtles have been dis- covered and rediscovered several times over. They are rudi- mentary in all post-Triassic forms and the nature of these highly peculiar rudiments is evident only in embryos or young speci- mens. Tlie only description with any pretensions to fullness is that of C. K. Hoffman (1879, repeated in Bronn's Thierreich 1890). There is brief mention of their occurrence in embryos by Ruckes (1929), Naef (1929) and Emeliaiiov (1937). None of these previous observations has been sufficiently detailed to make the nature and relationships of the rib elements quite clear, and, except for ratber crude and incomplete sketches in Hoffman, no figures have been given. I have myself seen cervical ribs and rib rudiments in botb adults and embryos (Emydinae) or in adults only (Testudininae, Chelydridae, Dermochelyidae, Chelyidae and Pelomedusidae). Embryos of Emys orhicidaris given me by Dr. B. AV. Kunkel and of Pseudemys sp., loaned me by Dr. H. Ruckes, and of Chrysemys )iiorgi)iafa and Chelydra .serpentina in the Minot Collection, Harvard University, have provided me with the opportunity to study the embryonic condition and developmental history of the cervical rib rudiments, while I have been able to examine adult conditions from specimens in the osteological collections of the Museum of Comparative Zoology (MCZ), the American Museum of Natural History (AMNH), the United States National Mu- seum fUSNM), the British Museum (Natural History) (BM), BREVIORA No. 101 the Institut des ►Scieiiees Naturelles in Brussels, the Senekenberii' Miiseuni in Frankfort, the Zoolop'ische Staatssammlung- in Munich and the Laboratoire d'Anatomie Comparee of the Museum National d'Histoire Naturelle de Paris (LAC). In tiie course of tlie present study a wide variety of conditions lias been observed, iiu'hulinannna ossicles are typically present in the Emydinae. Indeed, ill one instance they are probal)ly invariably present: the antero- Fig. 3. Tcnapene caroUiui. A. Ventral view of joint between 5th and Gth cei'vical centra showing separate beta and gamma rib elements. B. Anterior view of 6th vertebra in the same specimen. Compare Fig. 5. ventrolateral processes lyinw alongside the hypapophysis of the eijjrlith centrum in emydines are apparently fused gannna ossicles. Similar fusion of g'amma ossicles occurs rather frequently at the anterior end of the seventh centrum. The beta element, on the other hand, is unknown in adults except on the middle cervieals of a few individuals of certain emydine species. Here a^rain, however, the probability of loss of these very small and loosely-attached elements must be taken into account as a possilile explanation of the apparent rarity of these elements in adults. 6 BREVIORA No. 101 All instance in whidi hotli ventral rib elements are well de- veloped is seen in Figure 3A. The joint between centra 5 and 6 of a specimen of Terrapenc Carolina from my own collection is shown in ventral view. The g'amma elements are relatively large flattened ossicles attached to the ventral side of the centrnm (centrum 5) just below the paired, convex, posterior central articular surfaces. These ossicles extend posteriorly and some- what laterally to lie ])artly underneath the anterolateral diminu- tive transverse processes (alpha rib elements) of the posterior vertebra (vertebra 6). (lamma and alpha rib elements are con- nected on each side by ligaments in which are embedded small rounded ossicles which are evidently the persisting beta elements. Figure 8B shows the same elements in anterolateral view. (Corn- Fig. 4. Pseutlcmy.i sp. cml)iyo. Seetion tlnouKli liglit side of n middle cervical vertebra in early cartilage stages showing all three lil) elements. Comiiaie Fig. 3B. pare Fig. 4.) The only element of the morphological chain which we saw in the eml)ryo that is missing in tliis case is the median intercentrum ; this is consistently absent on the posterior cervicals of turtles. Museum specimens in wliicii 1 have seen comparable condi- tions (i.e. both beta and gamma ossicles ))resent) are: Clenintys muhlcnhcrgi U.S.N.M. 2922S, Clemrnys insciilpta F.S.N.M. 6308!) and rhrysrmys picta F.S.X.M. 29238, as well as Emjfs urbicularis l\M. 1!)2()-1 -20-2248 and other specimens of this ]!)5!) Clil{Vl(Ah i;iBS L\ TUHTLES 7 sperics at the British Muscuin. I should emphasize tliat tlicsc specimens are not the only ones by any means in wliieli 1 have observed beta and ji'anima ossicles together; these are only certain readily accessible specimens upon which my observations may be confirmed. Museum specimens are not likely to have preserved these elements; careful preparation of specimens is required to reveal them. As mentioned above, on anterior cervicals the gamma ossicles may be fused with the true median intercentra (or with the anterior of the two centra between which they lie), but occasional si)ecimens show them separate from and lying alongside the intercentra, presenting in such cases the appearance of tripartite intercentra. 1 have seen this condition in Chrysemys picta U.S.N.M. 29233, Pseudemys floridana U.S.N.M. 60492, and Orlitia horneensis U.S.N.M. 37788 — in the first case only at the odon- toid-second centrum articulation, in the two latter also at the joint between centra 2 and 3. I have seen such apparent triple intercentra also in Chehjdra serpentina (a specimen at Lafayette College, A.M.N.H. 28942, and specimens 18308 and 18361 in the Natural History Museum of the University of Kansas). Siebenrock (1907) mentions gamma ossicles in Plotysfernon megacephoIi())i, regarding them, howevcn-, as rudimentary ventral arches. He says : ' ' Rudimentare unteren Bogen, Hypapophysen, kommen so wie bei Macrodemmys Gray and bei vielen Emydidae am ersten bis f iinften Halswirbel vor ; sie f ehlen dagegen bei Chelydra Schw., wie ich mich an drei darauf untersuchten Skelet- ten iiberzeugen konnte." As Siebenrock states, gamma ossicles appear to be usually absent in Chelydra except at the odontoid-second centrum articu- lation, but one specimen of this genus at the Senckenberg Museum has gamma ossicles on centra 3 to 5, as does another specimen in the Staatssammlung at Munich. 1 have not myself seen any gamma ossicles in Maeroclemys, though Siebenrock mentions them above, nor in Staurotypus or in Kinosiernon. Beta ossicles have not been seen in any of the Chelydridae sensK lato. I have seen neither of the ventral rib ossicles in Dermatemys, and I have seen gamma ossicles only rarely in the Testudininae (e.g. at the odontoid-second centrum joint in Gopher us herlan- dieri A.M.N.H. 91916). l)ut Siebenrock (1906) records a series 8 15REVI0RA No. 101 of gamma ossicles in Acinixys planicauda, and Thomson (1932) reports them on the fourth, fifth and sixth vertebrae of Toiindo ihera and T. graeca. In the Cheloniidae and Dermoehelyidae one ventral rib element is usually present on most or all the eervieals. It is cartilaginous in Dermochelys (e.g. U.S.N.M. 29492, see also the figure in Volker 1913), bony in Caretta caretta (e.g. U.S.N.M. 29372) (Fig. 5) and Eretmochelys imhricafa (e.g. U.S.N.M. 59866). A special Fig. 5. Caretta caretta. A. Anterior view of 6th vertebra. B. Ventral view of joint l)etween 6th and 7th centra in the same specimen. Fig. 6. Oiclonia mydas. A. Anterior view of 7th cervical vertelira. B. Anterior view of right side of the 7th vertelira of another s])ecimen. condition is seen in Chclonia jiit/das (A.M.N.H. 46837) (Pig. 6A) in which on vertebra 7 the rib rudiment is seen to be fused to the right anterior side of the vertebra in such a way as to enclose a small "vertebrarteriar" canal. The " vertebrarterial" canal is incomplete laterally on the left side of the same vertebra. In another specimen of the same species (A.M.N.H. r)912) (Pig. 6B) the " vertebrarteriar " canal is completed lat(M-aIly liy car- tilage on both sides of the seventh vci'tebra. 1!);")!) CERVICAL KIIJS IN TURTLES 9 I have s(MMi ossified fjamma ossit'les also in the Eocene sea turtle Eospliargis (jiuas in the magnificently preserved specimen in the Rruskels Museum. Beta ossicles are unknown in the ("heloniidae, and they may in fact be merged with the rather larnrer apparent e(}uivalents of the jijunma ossicles of emydines. 1 have seen no trace of rib rudiments except the transverse ])rocesses in adult trionychids or in Careftochelys, but gamma ossicle ecpiivalents exist in at least ChcJus and Er]imnoehely!< of the Pleurodira. These peculiar conditions of the rib elements in modern turtles are fortunately traceable to a more primitive condition in Triassic turtles. Jaekel (1916) has described and figured the situation in Trinssochelys dux {-=Proganochelys qucnstedfi), and 1 can confirm his description in all essentials, having examined the magnificent skeletons of Profjanochelya in the possession of the Stuttsi'art Museum, which are to be described bv Dr. K. Staesche. Fig. 7. Proganochelys quenstedti (=Triassochelys dux). Anterior view of 7tli ccrvieal vertebra showing liieipital lili. (After Jaekel.) Figure 7 presents an anterior view of the seventh cervical vertebra in Proganochelys (after Jaekel). It will be seen that these are ribs of a normal bicipital type, the tuberculum lying across the neurocentral suture, the capituhnn articulating more ventrally with the side of the centrnm. Ribs of this type were present on cervicals 2 to 7. The eighth rib mucii resembled the abnormal vertebrae of Pseudcniys, Pyjris, etc. described above. It is single-headed rather than bicipital, elongate rather than .10 BREVTORA No. 101 short. It curves downward distally to a degree appropriate to its leugrth. Its single head lies across the neuroeentral suture in the position of the tuberculum. Probably the eighth and the ribs behind it retain only the tuberculum of the tAvo original rib heads. The congruence of the bicipital ribs in Proganochelys with the ribs that would be expected if the embryonic rib elements of Recent turtles were fully developed is very complete. With this evidence before us, the interpretation of the otherwise very singu- lar, even anomalous rudimental rib ossicles of Recent turtles can hardly be questioned. I have seen similar small ossicles that are apparently rib rudi- ments in large lizards, e.g. Iguana iguana A.M.N. H. 1269 (Com- parative Anatomy collection). The phenomenon of rib frag- mentation is thus not confined to the neck of turtles. Indeed, there is a considerable literature on anomalous rib conditions in Amniota (Baur 1886, Howes and Swinnerton 1!)()1 — Sphcno- ilon; Xoble 1921 — Kphaerodactylus; Schinz and Zangerl 19371) — Gallus; Virchow 1929 — Ornithorhyncus; Albrecht 1882, Hayek 1928, de Burlet 1917, Federow 1910, Abel 1909 — various mannnals.) These special conditions have, I feel, implications of greater morphological moment than appears on the surface. They may require a careful re-examination of the origin and morphology of ribs throughout the Amniota. Rib fragmentation in tetrapods occurs typically in regions of the column in which special movements are taking place (neck of turtles, lizards, etc., tail of amniotes generally), and in which, tlicrefore, the thoracic rib pattern is not useful, or in acjuatic forms in which the thoracic pattern itself may prove unstable. It is a phenomenon, therefore, of degeneration — of breakdown of the two-headed ribs })rimitively present in tetrapods, ])re- sumahly throughout the cohnnn. However, it is precisely the origin and evolution of the two-headed ribs of tetrapods that is still a puzzle — a puzzle to which, 1 believe, these special condi- tions along with the ontogeny of less modified ribs have something to contribute. Sjiecifically, the (piestion of the possil)le origin of two-headed tetrapod ribs from a union of the dorsal and ventral I'ilis of fishes still needs a definitive ves or no answer. ]})5!) CKKA'ICAL HIBS I X 'ITHTLKS H LITKRATURE (MTKD Abel, O. 1909. Cotaceenstudioii. I Mitt(>ilnn{j. Dns Sk(>lc'tt von Kiiriiinodrlphis coclieteuxi aiis den 01)0iniioziin voii Aiitweipcii. SitzlitM-. Akiid. Wiss. Wien, Math.-Nat. Kl., 118 (1): 241-253. Al.HKlX'llT, I'. 1882. Note sur uno sixienie costoide cervical clicz lui jcuiu' Hippopota- iiiiis tinipliihiiis. Bull. .\rus. Roy. Hist. Nat. linuxellcs, 1: 197-202. B.\uR, G. 1886. The ril).s of Hphenodnn (Hofferia) . Anier. Nat., 20: 979-202. i)F Ht'RLBn', H. M. I!n7. BeitrJige ziii- Eiitwieklungsgeschiclitc dcr Wiilielsiiulc der Ceta- ceen. :\rori>h. .Talnl)., 50: .",73-402. I'^jrKLIANOV, S. M. 1937. Die Morpliologie der T('tiaitndciiii])iieii. Zool. .T.-ihib. (A))t. Anat.), 62: 173-274. Fkderow, Y. 1910. Zwei FJille der .selteiieii Bildiing vou Qiierfortsjitzen des ersten Brustwirl)els. Anat. Anz., 36: 556-500. Hayek, H. 1928. Uber Querfortsatze und Rippenrudiinente in der Hals luid J.en- densegmenten. ]Mori)h. .Tahrb., 60: 371-416. Hoffman, C. K. 1879. Over het voorkonien van lialriblien by d(> s(diild])adden. Yersl. Medel. kon. Akad. Wet., 14: .12-67. 1890. Seliildkroteu. In II. G. Bronn — Klassen und Oi-dnungen des Thierreichs. Leipzig, 442 pp. Howes, G. B. and H. H. Bwinnerton 1901. On the development of the skeleton of the tuatara, Sphcnodon pimctdtus, with remarks on the egg, on the hatching and on hatched yonng. Trans. Zool. Soe. London, 16: l-7(i. Jaekel, O. 1916. Die \\'irbeltiere aus dem Kenper von Hall)erstadt. Serie II. Testndinata. Palaeont. Zeitsehr., 2: 88-214. 12 BREVIORA No. 101 Naef, a. 19'2':K Notizen zur Moiphologie unci Stamniesgesehiehte der Wirbeltiere. 15. Dreissig Thesen ii))cr Wii-lielsiiiile und Rippen iiisbesondeiT bei den Tetrapoden. Zool. .Tahil). (Abt. Anat.), 50: 581-6n0. Noble, G. K. 1921. TIk' bony structure and ]ih.\ietic relations of Sphaerodactylus and allied lacertilian genera with a desci'iption of a new genus. Anier. ]\[us. Novitates, 4: 1-1 (i. RUCKES, H. 1929. Studies in elielonian osteology. Part II. The morphological rela- tionships lietween the girdles, ribs and carapace. Ann. New York Acad. Sci., 31: 81-120. HcHixz, 11. R. AXi) R. Zangerl 1937a. Uber die Osteogene des Bkelettes beini Haushulin, liei der Haustaube und beim llaubensteissfuss. ^forph. .Tahrb., 80: ()20-628. 19371;. Beitriige zur (Jsteogenese des Knochensystenis beini Haushuhn, bei der Haustaube, und l)eini llaubensteissfuss. Eine ver- gleichende osteologische Studie. Denkschr. Schweiz. Naturf. Ges., 72 (Al)h. 2): 117-16.-1. SlEBENROCK, F. 190(i. Schildkroten von Ostafrika und Madagascar. In Voeltzkow, Heise in Ostafrika in den Jahren 1903-1'90.5, 2: 1 -lU. 1907. Uber einige zum Teil selteue Schildkroten aus Siidchina. Sitzber. Akad. Wiss. Wien, Math-Nat. Kb, 116 (1): 1741-1775. Thomson, J. S. 1932. The anatomy of the tortoise. Sci. Proc. Roy. Dublin Soc, 20: 359-461. VlRCHOW, II. 1929. Die tiefen Riickenmuskeln des ()niilli' the species. With more material it will probably be possible to simplify and improve the description. I do not follow Burt and Burt (1931) and Taylor (1956) in maintaining i\ distinction between the genera Diploglossus and Celestus. After examining material of 22 different species I cannot see good grounds for supposing that there are two stocks of galliwasps which merit generic rank, and I therefore use the older name Diploglossus. As the generic name is compounded from Greek roots, I form the specific name likewise. To draw attention to what I believe to be its most striking single feature I call it : Diploglossus microblepharis sp. nov. Type and only known specimen. M.C.Z. 55764; collected by R. P. Bengry and G. R. Proctor, 3rd. May 1952, Boscobel, St. Mary, Jamaica, West Indies, 18°24'N, 76°58'W. Diagnosis. A plump, short-limbed galliwasp with small palpe- bral aperture, nasal shield excluded from rostral, frontal broader than long, very pronounced median keels on proximal caudal scales, few scale-organs on dorsal scales (less than one per scale). Description. I believe this form to be most closely related to D. delasagra of Cuba and D. pleeii of Puerto Rico ; the de- scription has been prepared with specimens of these species before me. A number of the features of the head squamation appear to be related to the reduction of the size of the palpebral aperture; these are nuirked with an asterisk; in some cases modi- fication of the squamation is asymmetrical. To make the descrip- tion easier to follow T have italicised those characters in respect of which the new species differs from both of the related species. Where a count can be made on two sides T have separated the left hand count from the right hand count by a comma. Plead broad; width: length = 0.7!) :1. Rostral twice as broad as high, in contact with supranasals and 1st labial. Three supranasals (asymmetrical), right meets frontonasal, 1st loreal, nasal, rostral and in midline left supranasal (also in j)Jccii) ; the two left supranasals together apparently represent a divided scale similar to the right supranasal ; together the two scales make the same contacts as the right supranasal. Nasal meets 1st and 2nd labials, NEW JAMAICAN UALLIWASI' ;j 1st loreal and suprauasal. Prefrontal about as long as fronto- nasals, contact with 2nd loreal and 1st supraorbital, ah art contact with 3rd loreal. \'ery broad frontal, width greater than lengtii, broad contact between frontal and parietals (also in pleeii). Four supraorbitals* , 1st much the largest. Third loreal meets 1st supraorbital, not separated by intercalary scale*. Preocular Fig. 1. Dorsal view, bond of Biplofflnssnn microhlppJiaris type. meets 1st supraciliary, 1st supraorliital, ^rd loreal, 5th supra- labial and subocular*. Subocular meets 5th and 6th labials (also in pleeii). Three supraciliaries. Postorhitals 3,2*. Postocular ap- pears as last member of partly concealed row below supra- ciliaries. Three temporals between f ronto-parietal and labials : upper temporal separated from parietal ; lower temporal meets labials 6 and 7 and nearly enters margin of orbit between sub- ocular and lower postorbital*. Single nasal between supranasal, loreal 1 and labials 1 and 2. Three loreals; 1st low, elongated, between supranasal, nasal, labials 2 and 3 and loreal 2 (also in pleeii) ; 2nd loreal scpiarish, between frontonasal, prefrontal, 4 BREVIORA loreals 1 and 3 and labials 3 and 4; 3rd loreal squarish, between 2nd loreal, prefrontal, supraorbital 1*, preocular and labials 4 and 5. No scale between 3rd loreal and prefrontal*. Eight supra- labials (also in pleeii). Small mental, half width of rostral; 7 infralabials, prominent postmental in contact with labials 1 and 2, followed by 4 pairs of chinshields (also in pleeii). 1st pair meeting in midline. Fig. '1. A. Scales aiuund left eye of Diploglo.s,su.s inicroblepharin type. B. Scales around right eye of Diploglossus pleeii. Head length (snout to posterior margin of earhole) 14 mm., width 11 mm. Palpebral aperture about 1 mm. long, ear-opening about i/> mm. in diameter. Palpebral marginal scales reduced, lower palpebral scales very thin. Pupil circular. Middorsal scales with 16-20 parallel striae and a j^ronounced central keel, caudal end of most striae broken. Accentuation of central keel fades out and interruption of sti-iae increases on lateral doi-sal scales; striae fade out altogetlier on Hanks, the vcnti-al scales being smooth and fiat. Towards shouhlers central NEW JAMAICAN GALLIWASP 5 keels fade, and on neek striae fade out; towards rump central keels become very pronounced, and keels become so pronounced on proximal balf of tail as to fj:ive it, dorsally, the appearance of a polygonal prism. Towards tip of tail keels fade out. \'entral caudal scales smooth and flat. Scale organs on head extend back dorsally about as far as occipital scale, ventrally about as far as 15th postmental scale ; scale organs very few on back (less than 1 per scale), absent on flank save before groin, absent on belly ; a few scale-organs on outer face of thigh; scale-organs present on back and sides of proximal tapering portion of tail, aggregated on the mesial (dorsal) side of the central keel of each scale. Fourteen scales from a middorsal row gave the following scale-organ counts : 011010010 3 1100 (mean 0.64) ; 4 groin scales gave the counts : 2 8 1 1. Scale-rows at midbody 43, mental to vent 109 scales, occipital to point above vent 105, from anal to chest (on line joining axillae) 72, some union of dorsal scale-rows over rump, at level of ankles 20 scale-rows round tail. Third digit of forelimb slightly longer than 4th ; lamellae, 3rd digit 7, 8, 4th digit 7, 7. Scales on outer face of upper arm with faint striae (no central keels), scales of forearm quite smooth, scales of inner face of forelimb with striae. Ilindlimb : length of 4th toe to length of 3rd toe as 1.4 to 1 ; 4th toe lamellae 11, 12 ; scales of outer face of thigh with striae ; scales of shank also with central keels; scales of inner face of hindlimb with faint striae but no keels. Cieneral colouration medium brown. Top of head uniform brown ; upper labials white with brown spots, lower labials with a few spots only ; throat white ; neck and back speckled wdtli brown, speckles on back loosely arranged in about 20 backwardly pointing chevrons ; speckling fades out fairly abruptl.y on sides, above ear-hole, on flanks and above groin ; speckling less regular on tail ; upper side of forelimb and outer face of forearm and hand closely speckled ; upper side of liindlimb and outer face of shank and foot speckled, inner face of thigh darker brown, some brown onto sole of foot. Dimensions. Snout -vent 87 mm., tail (tii) regenerated) 83 mm., forelimb 13 mm., hindlimb 18 mm., axilla-groin 58 ram. b BBEVIORA Discussion. Burt and Biirt (19;U) make an explicit distinetion between Diploglossus and Celestus. Their discussion begs the (juestion as to wliether or not there are two natural groups which might properly be regarded as genera; they seek only a formal basis for their discrimination. They overlook the fact that two West Indian species of '' Celestus" have claw sheaths and there- fore belong to Diploglossus as they define it. This mistake ap- pears to be due to the fact that Boulenger (1885) did not have material of delasagra or plceii. He implied that they do not have Fig. 3. A. Lateral view of 4111 toe of Dijyloglo.ssus nticrohlcphuiit: type B. Lateral view of ith toe of Diploglossus plceii. (Claw is coin-enlcd liy sheath.) C. Lateral view of 4th toe of Diploglossus Jicicarrlii. slieathed claws by ])lacing them with the other island forms in section 11 of his key : "No large ungual sheath ; frontal anteriorly in contact with a single shield." 1 am very doubtful of the state- ment by Burt and Burt that the claws of the forms with a sheatii are "semi-retractile to retractile"; I see no indication of mobil- ity. 1 checked a series of species in respect of : claw sheaths, number of prefrontal shields, nasal-rostral contact and scale- NEW JAMAICAN (iAI.Ll VVASl' 7 organs on tlie dorsal scales. The results are summarised in the accompanying table; I have included the Ilispaniolan endemic genera Saurcsia and Wetmorena. We see that there are four Antillean forms which have sheathed claws and conversely two mainland forms with naked claws. As is recognised by Burt and Burt and by Taylor the number of prefrontals is not correlated with the claw-sheath condition. The character of three prefrontals is however confined to the mainland (and Malpelo) species, with the interesting ex- ception of darlingtoni from the mountains of Hispaniola. Of the three M.C.Z. specimens, one, the type, has 3 prefrontals, the other two have only one (Cochran, 1941, states that the type has only one). Contact between nasal and rostral scales is shown by all the mainland forms but also by two Antillean forms {delasagra and pleeii). The scale-organs of the trunk call for more detailed comment. These organs do not appear to have been noticed in the Angui- dae; McDowell and Bogert (1954) state that they are not pres- ent; Scortecci (1940) did not find them. I have, however, found them, at least on the head, on every anguid examined. They were checked by prying off the horny scales with a needle and mount- ing in polyvinyl alcohol. The preparations are in the Museum of Comparative Zoology. I am not yet in a position to offer a critical appraisal of this character. A cursory comparison of the scale organs on various parts of the body of several species indi- cated some striking differences in respect of their occurrence on the dorsal scales ; representatives of the available species were therefore checked for this character. The scale-organs are visible as thin refracting spots in the horny scales. On the trunk and tail scales they are localised in two areas on either side of the midline and nearer to the posterior margin. Commonly there are scale-organs on both sides of the midline of a scale. Counting the scale-organs separately for each side of a longitudinal row there is sometimes a significant asymmetry ; the higher number of scale-organs is in such cases found on the side nearer to the mid-dorsal line, i.e. the mesial side. The scales of the flanks and the sides of the tail usually have all the scale-organs on their dorsal sides, i.e. the sides nearer the dorsal midline. I have not found scale-organs on belly scales nor on ventral tail scales. 8 BREVIORA Most of the dorsal scale-organ counts were made on a row of about ten dorsal scales taken from near the dorsal midline. To form an idea of the reliability of such counts I mounted pieces of dorsal slough from tAvo Jamaican D. criisculus. Each lateral Fig. 4. Scales and scale-organs in Diploglossus microblepharis. A. Mid- dorsal scale to show interrupted striae. B. Scale from right side of tail near base to show size and position of scale-organs (reversed under microscope). C. Right fronto-nasal to show distribution of scale-organs. dorsal scale-row is clearly indicated in this species by a break in the colour pattern. Including the lateral dorsal rows one slough had a complete block of 196 scales (14 rows of 14) and the other of 403 scales (13 rows of 31). Counts of scale-organs were NEW JAMAICAN GALLIWASP 9 made on all of these. On tlic outer three rows of scales 97.8 per cent of the scale-organs lay on the inner side. The scale- organs were less asymmetrically distrilmted on the inner rows of scales; there were however more on the inner side of each scale- row than on the outer side. The reversal of the asymmetry did not take place about the midline, however, but between the 2nd and 3rd rows to the left of the midline. The lateral dorsal scale- rows had rather high scale-organ counts ; the other rows differed from one another in apparently random fashion, but a difference between two rows would be maintained throughout their lengths (not scale by scale but for groups of scales large enough to smooth out random variations). One specimen had the highest count (3.0 organs per scale) for the 5th row, right of centre, and tiie lowest counts (2.2 organs/scale) for the 2nd, 3rd and 4th rows, right of centre. The other s})ecimen had the highest count (5.6 organs/scale) for the centre row and the lowest (4.2 organs/scale) for the 3rd row, right of centre. Discounting the lateral dorsal scale-rows it was possible to make a count of 10 consecutive scales in 60 different ways for one specimen ; the lowest total was 20, the highest 33, mean 26.25, standard varia- tion 11.0 per cent of mean. There were some irregularities in the squamation of the other specimen; avoiding these, 222 such counts were possible; the lowest total w^as 31, the highest 67, the mean 51.49, standard variation 15.7 per cent of mean. 1 tried some counts of 10 scales in a transverse row and 10 scales diagonally but the results did not suggest that these would be more reliable. 1 have not eliminated the possibility that the counts may be correlated with size or with sex. All of the mainland species lacked dorsal scale-organs save monolyopis (23 scales from two specimens of the latter averaged 4.] organs/scale). A group of Antillean species (dclasagra, pleeii and darlingto7ii) likewise lacks dorsal scale-organs; micro- blepharis is intermediate in that many scales are without organs. It is noteworthy that SSauresia and W etynorena both are in agree- ment with the majority of insular forms in the possession of dorsal scale -organs. Preliminary examination of a number of forms suggests that these scale-organs may have value as a systematic character at the species level. There may be differences in the number of organs 10 BREVIORA per dorsal scale ; when the organs are absent from the back they may or may not also be absent from tlie flanks. The number of organs on the scales of the groin may also be useful. For example, the type of Panolopus cost at us Cope has a dorsal average of 2.2 organs/scale (10 scales), the type of Diploglossus stenurus Cope (listed as a synonym of costatus by Cochran) has a dorsal average of 5.0 organs/scale (16 scales) ; to this is added a mean of 7.4 organs/groin scale for costatus (5 scales) and 17.4 organs/groin scale for stenurus (5 scales). When it is noted also that the adpressed limbs of costatus fail to meet by about 15 scales whereas those of stenurus overlap by about 5 scales, there seems to be little question that stenurus is a distinct form. D. eymeagrammus is noteworthy for the fact that each groin scale bears a single very large organ. Differences were also noted in the distribution of the scale-organs on the occipital scale. For example, the Jamaican D. hewardi has scale-organs scattered around the jiosterior margin of the occipital whereas D. crusculus and D. harbour i have the scale-organs generally scattered over the occipital. Purveying the distribution of variant characters amongst the different forms, 1 cannot arrive at any firm ideas concerning the species groups. However, to provide a basis for discussion I put forward a scheme which accords reasonably well with the evi- dence. I propose to regard monotropis as primitive in respect of : the presence of dorsal scale-organs ; the presence of sheathed claws; the contact between nasal and i-ostral scales; the presence of 3 prefrontal scales and the stoutness of the limbs. The general occurrence of scale-organs on the trunk of lizards of various families suggests that this is a i)rimitive condition. As with the teeth, however, we cannot assume that the onlj^ changes in phylogeny have been reductions in number, so we must be pre- pared to allow re-extension of scale-organs onto the back if other evidence requires it. Conditions in lizards generally suggest that many small scales are more primitive than few large scales ; on these grounds I suppose three prefrontals to be more primitive than one. In the case of claw-sheaths we have either to suppose that they have arisen within the genus several times or that they have been lost several times; as a working hypothesis 1 prefer the second proposition. As regards the pi'oposition that the NEW JAMAICAN GALLIWASP 11 nasal-rostral oontact is primitive, 1 can only ai-no)i()t)-opi,s but has the claw-sheath reduced as well as lacking- dorsal scale-organs. In northern Central America enneagrammus and vijatiochloris superficially resemble such Antillean forms as crusadus and costatus; however tliev retain a nasal-rostral contact and lack dorsal scale-organs. The three species delasagra, plecii and microhlepharis do appear to represent a natural group ; even here we have some complexity since uiicroMepharis, presumably primitive in letaining a few dorsal scale-organs, has lost the nasal-rostral contact and partly lost the claw-sheath retained l)y the other two species. It seems clear therefore that microhlepharis cannot be derived from delasagra or pleeii, nor can the converse view be entertained. The delasagra-pleeii and microhlepharis stocks may be separately derived from a Central American ancestral stock. On the evi- dence of the distribution of mammals, Darlington (lf)57, p. 510) suggests that Cuba and Jamaica represent two ports of entry for Central American stocks in the Am'illes. This group must be surely represented by an undiscovered torm in Ilispaniola. The remaining Antillean species pei-hajis form a natural group : darlingtoni is the most isolated form. On the grounds that they retain claw-sheaths I suppose Sauresia and Wetmorena to be early offshoots of the Antillean stock. [ cannot see good grounds for recognising the genus SaiD'esia as distinct from Diploglossus. The disappearance of one digit, small in a number of other forms, is a trivial feature Avhich cannot be of much adaptive sig- nificance; the species is not set otf in other respects. Wetinorena, on the other hand, is clearly set apart by the loss of the ear- opening, a feature of manifest adaptive significance in a burrow- ing animal. The forms so far discussed all belong to McDowell and Bogert's snbfamily Diploglossinae. All of them have numerous scale, organs on the head although they ma}- be i-educed on the trunk. A few representatives of their subfamily Gerrhonotinae were 12 BREVIORA cursorily examined. No scale-organs at all were found on the trunk and only a fe^v on the anterior part of head. A loreal and 3rd supralabial scale -were exaniijied of one specimen each of : Gerrhonotus liocephahis, Bariasia imhricutu, Elgaria mtdtica- rinata, Ahronia deppii and Ophisaurus apus. Elgaria most closely approached the Diploglossinae in the partially sheathed claws, short nasal-rostral contact and the possession of more loreal scale-organs than the others. The three prefrontals of Ahronia resemble those of DiploglossKs. The loreal and labial of Gerrhonotus had only one scale-organ each. The Barissia had none on the loreal and four on the labial. Ophisaurus resembled the others in the absence of scale-organs from the trunk and posterior part of head and in the small number on the anterior part of head. Angnis fragilis is similar to Ophisaurus in the possession of a few scale-organs on the anterior head only. I am indebted to Mr. C. Bernard Lewis, Director of the Insti- tute of Jamaica, who generously turned this specimen over to me for study. I am grateful to Dr. E. B. Williams for access to the collection of the Museum of Comparative Zoology and for the benefit of discussion. Dr. R. Zweifel kindly allowed me to study some additional specimens in the American Museum of Natural History. LITERATURE CITED BOULENGEK, E. G. 1885. Catalogue of the lizards in the British Museum. London. Burt, C. E. and M. D. Burt 1931. South American lizards in the collection of the American Museum of Natural History. Bull. Amer. ]\Ius. Nat. Hist., 61: 227-395. Cochran, D. M. 1941. The herpetology of Ilispaniola. Bull. U. S. Nat. Mus., 177:1-398. CousENS, p. N. 1956. Notes on Jamaican and Cayman Island lizards of the genus Celestiis. Breviora, Mus. Comp. Zool., No. 56:1-6. Darlington, P. J. 1957. The geographical distribution of animals. New York, 675 pp. NEW JAMAICAN GALLIWASP 13 Grant, C. 1040. The reptiles in Lynn nnd Grant: Tlie licrpetology of .T.-iiiiiiica. Bull. Inst. Jamaica, Sci. Ser., 1:1-6"). .McDowell, S. B. and C. M. Bogert 1954. The systematic })osition of Ldntlionotus and tlic alUiiitics of the anguinomorphan lizards. Bull. Anier. Mus. Nat. Tlist., 105:1-142. SCORTECCI, G. 1940. Eecettori dcKli iKuauidi o di altri Sauri. Atti Roc. Ital. Sci. Nat., 79:1-10. Taylor, E. II. 1956. A review of the lizards of Costa Rica. Sci. Bull. Univ. Kansas, 38:1-322. Table Claw-sheath Prefrontal Xasal-rostral Scale-organs contact on back DipJoglossus fasciatus + 3 + — resplendens + 3 + — monotropis + 3 + + millepunctatus % 3 + — cyanochloris — + — enneagrammus — + — microhle pilaris ¥2 — few delasagra + + — pleeii + + — darlingtani — 1 or 3 — — hewardi — — + ocdduu^ — — many duquesnayi — — + crusculus — — + barbouri — — + hadius — — + costat-u.s — — + rugosus — — + ciirtissi — — + stenurus — — + Sauresia sepsoides + — + W'ctmorena haetiana + — + BREVIORA MiLaseniim of Compsirsitive Zoology Camhridgk, Mass. April 10, 1959 Ximber 103 TWO NEW SPECIP]S OF ELEUTHERODACTYLUS FROM PFERTO RICO By Juan A. Rivero Institute of Marine Biology and Department of Biology University of Puerto Eico Mayag'iiez, Puerto Eico Some years ago Dr. J. A. Ramos of the Biology Department of the University of Pnerto Rico collected, at the Oamhalache Forests, a small frog that differed strikingly from the Elcnthcro- dactylus crampioni collected at the same locality and date, by the very short length of the fingers and relatively short, thick liiiid limbs. The specimen was said to have been collected in bromeliads l)nt no voice record was taken, so that the description of the apparently new species was postponed nntil more para- types could be obtained and the voice recorded. So far, four more trips have been made to the same locality, at night and during the daytime, but all of them have been unsuccessful, although a strange voice that could have been of frog or insect was heard on several occasions. The Cambalache Forests are of the semiarid, deciduous type, and spiny trees and bushes are common, so that catching a calling frog at night is not an easy matter. More recently, another apparently new frog was collected at Dona Juana Forests and, although this first collection consisted of 40 individuals, the animal was not heard to produce any voice in spite of the fact that all the specimens were caught at night and enough time was given each one to call. The description of this form was likewise postponed and, in the month following the first collection, 16 other individuals of the same species were collected at El Yunque, on the other end of the island. No voice 2 BREVIORA No. 103 was heard to come from any of these animals, so that it was tenta- tively concluded that the species is nuite. Since well developed gonads are fonnd in both sexes, the lack of voice cannot be at- tributed to immaturity, especially if it is considered that frogs of the genus Eleuthcrodactylus seem to call throughout the year in Puerto Rico. Descriptions of both species have been prepared in the belief that they should be recorded and that their jniblication may be of help in determining distribution and habits. The author wishes to express his appreciation to Mr. Jorge Rivera Lopez, to Dr. Ruth Turner, and to his wife, Mrs. Eneida B. Rivero, who helped him collect in the field. Eleutherodactylus ramosi sp. n. Type. Museum of Comparative Zoology No. 30428, a 9 from Cambalache Forests, n. Puerto Rico, Coll. J. A. Ramos, Aug. 1952. Diagnosis. A small Eleutherodactylus with uniform colora- tion, very short fingers and short limbs, the heel and elbow of which do not meet when adpressed along the side of the body. Description. Head broader than long ; snout subovoid ; tongue large, % free and nicked behind ; vomerine odontoids in two strong oblique series behind and between the choanae ; eyes of moderate size, their diameter equal to distance between eye and nostril ; interorbital space broader than an upper eyelid ; canthus rounded ; loreal little inclined, not concave ; tympanum moder- ately distinct, covered above by a supratympanic fold, V2 the eye diameter ; fingers strikingly short, the longest not longer than the orbital diameter ; no definite palmar tubercles ; larger disks smaller than the tympanum ; no apparent metatarsal tubercles ; toes free, the first well separated from the others ; disks of toes smaller than those of fingers ; heel of the adpressed hind limb extends to the posterior margin of the tympanum ; heels overlap when legs are placed at right angles to the body; heel and elbow do not meet when adpressed along the flank. Skin above, smooth. Ventral surfaces smooth except on the posterior part of the belly, where slight rugosities or granula- tions can be seen under a lens ; posteroventral aspect of thighs, granular. VJdU NEW FROGS FROM PUERTO RICO Color. Above, unifonn brown, with iniimte dots (melano- phores) iindor a lens; no ai)parent. eantlial or supratympanic streaks. Below, infuseated white, the nielan()i)liores being scat- tered and sparser than on the npper surfaces. MiasurcDunts. (nun.) $ snout-vent 22; head length 7.3; head breadth 8.3 ; femur 8.2 ; tibia 9.6. Fig. 1. Eleutherodactyhis ramosi Type, MCZ 30428 (M. Estey del). Remarks. E. ramosi is most closely related to E. cramptoni with which it occurs at the Cambalache Forests. It differs from that species in its verj- short fingers, smaller disks, broader head and shorter and thicker hind limbs, the heel of which does not even approximate the elbow when both anterior and posterior limbs are adpressed alongside the flank. It was the very short fingers and thick hind limbs that made the author describe this species on the basis of only one specimen. 4 BREVIOKA No. 103 BlEUTHERODACTYLUS ENEIDAE Sp. 11. Type. Museum of Comparative Zoology Xo. 30-129, a 9 from Dona Juaiia Forests, \'illalba, Puerto Kico. Coll. J. A. and E. B. Rivero, August 1956. Diagnosis. A medium sized Eleutherodactylus with rounded canthus rostralis, interorbital space as broad as an upper eyelid, tubercular dorsum, barely overlapping heels, mottled and variegated dorsal coloration and immaculate ventral surfaces. Description. Head slightly l)roader than long; snout short subovoid ; tongue oval, ^ ^o free and nicked behind ; vomerine odontoids in two oblicpie series behind the small, oval choanae, their external extremities commencing at level with the vertical of the center of the latter ; canthus rounded ; loreal sloping ; inter- orbital space as broad as an upper eyelid ; eye diameter greater than distance between eye and nostril, almost as long as the snout ; tympanum small, not too distinct, ^ the eye diameter ; first finger shorter than second, which is longer than last ; three metacarpal tubercles and two smaller ones in line behind the wrist, just in back of the central metacarpal tubercle ; disks fan shaped, broader than the t^'iiipanum ; two small metatarsal tubercles ; subarticular tubercles distinct ; toes free, their disks slightly smaller than those of the fingers : heel of the adpressed hind limb extends to the posterior corner of the eye ; heels barely overlap when tibiae are placed at right angles to the body. Skin above, studded with small, evenly distributed tubercles; eyelids tubercular ; a small tubercle at the heel. Below, granular on the belly and thighs, very slightly granular on the throat; a ventral discoidal fold. Male with a slight fold on each side of the throat. Color. Above, light gray with blotches and mottles of darker gray or l)rownish gray ; eyelids darker than the body color; loreal area whitish gray, witli small darker spots along the upper lip; limbs light gray, with l)i-owiiish gray striations and mottling; loins with a golden brown stain ; posterior aspect of the thighs with a diffuse, irregularly margined brownish marking on the distal end. ^"entral surfaces immaculate, powdered with brown, except for a thin gray rim along the lower margin of the li]). Measurements, (mm.) 9 snout-vent 26.2; head length 10.5; head breadth 10.5; femur 11.9; til)ia 12.6. 1IJ5!> NEW FROtJS FKOM I'l'EKIO RRO 5 Re)it(n-hs. Paratypos include M!) specimens from Dona . I nana cloud forests, with tiie same data as the type (IMMv. (Jll 12. 614-51) and 16 from El Yiuuiiie (U.P.K. 5!):)-6l(). El Ynn(iue, sixth i-ain shelter to Pinnacles, 1 Sept. Ifi56). All were collected at iiig:ht on the floor or on the leaves or trunks of busiies, palms or tree ferns at not more than 5 or 6 feet from the ground. The largest of the 54 paratypes is only 2 nnn. larger than the average sized type, an indication of a very uniform size for fully grown individuals. This is contrary to the situation existing in Elcuthcrodacff/l us porioriccuxis where considerable variabil- ity is found in any collection. The following variations are found among the paratypes of FA( >iih( rodaciyhiH encidac : 1. In poorly preserved specimens, a slight canthus may be apparent, but ordinarily the snout is rounded in cross-section and there is no angle at the canthal margin. This character easily differentiates this species from E. portoricensis, its closest Puerto Rican ally. 2. The heel of the adpressed hind limb may extend to any point on the eye (posterior corner, middle or anterior corner). 3. The golden brown stain of tlie loins is present in most but not all individuals. 4. A small, black spot or short streak is generally found above the tympanum. 5. The dorsum is always mottled and variegated but the ground color may be lighter or darker than in the type. Eleuihcrodaciylus cneidae is most similar to E. portoricensis from which it can be distinguished by its more uniform, adult size, rounded canthus not bordered by a white line (as in most E. portoricensis) , absence in most specimens of a well defined interorbital bar ; smaller tympanum ; dark eyelids ; narrower interorbital space (much broader than the upper eyelids in E. portoricensis), more tubercular dorsum, different coloration and ])r()l able absence of voice. Some examples of E. porioricensis are also mottled above, but the mottling is usually brown, and the well defined canthus, broader interorbital space and persistent interorbital bar make the species easily distinguishable. It can also be said that if a specimen is more than 80 nun. snout-vent length, it is not E. eneidae. 6 BREVIORA No. 103 EleutkerodactyUis cneidae is quite common in the cloud forests of Doiia Juana and El Yunque and it is very strange that the species had not been named before. It appears possible, however, that collections of E. portoricensis may include this species. This may be the case, for example, with Schmidt's (1928:46) Figure lie, but this figure may just as well represent a color phase of E. portoricensis. E. eneidae is somewhat similar in coloration to E. flavescens from the Dominican Eepublic, but it differs from that species in many important details. It also looks somewhat like Schmidt's (op. cit.-.od) figure of E. locustus but Dr. R. G. Zweifel of the American Museum compared it with the type of that species and found that three specimens of E. cneidae "are very different in size and general appearance from locustus.'" Furthermore, Zweifel found very little difference between cramptoni and locnstus, and suggested that they might be synonyms. Males of E. crampto)ii have vocal sacs (described as absent by Schmidt) and the proportion of tibia and femur (heels cross in locustus, meet in cramptoni) , mentioned as a difference between the two species, does not, according to Zweifel, seem to exist. On the other hand, Schmidt described the voice of E. locustus as the "most distinctive of any noted in Puerto Rico," so that it is })erhaps better to keep the two forms separate until more field studies can be made. EEFEEENCES CITED Schmidt, K. P. 1920. Contriljutions to the herpetology of Porto Kico. Ann. Xcw York Acad. Sei., vol. 28, pp. 167-200. 1927. A new Eleutherodacti/lus from Porto Rico. Auier. Mus. Novitate.s, No. 279, pp. 1-3. 1928. Amphibians and land reptiles of Porto Rico, with a list of those reported from the Virgin Islands. Sci. Survey Porto Rico and the Virgin Islands. (New York Acad. Sei.) Vol. 10, part 1, pp. 1-160. "^d^ FJeutherodactylus eneidae , Type, MCZ 30429 (J. A. Eivero phot.). BREVIORA MiaseiLinn of Comparative Zoology ('A:\inKiDGE, i\lAss. April 13, 1959 Number 104 STUDIES OX FISHES OF THE FAMILY OPHIDIIDAE. III. A XEW SPECIES OF LIJPOPHIDIUM FROM BARBADOS^ By C. Richard Robins The Marine Laboratory, University of Miami Anionof the eolleetions of The Museum of Comparative Zoology, Harvard University, are two eiisk eels collected off Barbados, British AVest Indies, during the operations of the Steamer "Blake." They represent a new species of Lepophidium which is described below. Lepophidium kallion, sp. nov. (Figure 1) Holotype. Museum of Comparative Zoology 28057 A, a female, 117 mm. in standard length; collected in 200 fathoms "off' Bar- bados," "Blake" station 291, presumably in February or March, 1879 (see Pierce and Patterson. 1879:15). Paratype. MCZ 28057B (1), a male, 162 mm. in standard length ; data as for holotype. Diagnosis. Dorsal rays 133, anal rays 108; pectoral rays 23 or 24, gill rakers 10, vertebrae 74 (15 precaudal and 59 caudal vertebrae), pyloric caeca 6 (1 dorsally, 5 ventrally). Tapper sides with dark brown markings in two series, the upper form- ing a longitudinal stripe on either side of the dorsal-fin base. Snout tip fringed with numerous fleshy projections. Description. A healed injury at the caudal end of the para- type has resulted in the loss of about eight caudal vertebrae including the hypural plate. Counts of vertebrae and of rays in the vertical fins, listed in the diagnosis, are therefore based iCuntiilnitioii Xn. 221 from The Marine Laboratory. University of Miami. 2 BREVIORA No. 104 solely on the liolotype. Pectoral rays number 23-23 (holotype) or 23-24 (paratype), the caudal rays 9. Three rudimentary gill rakers are on the upper arm of the tirst arch of botli spi^'imens and 6 or 7 developed rakers are on the lower arm. The holotype has an additional rudimentary raker on the lower arm ; both specimens liavc a total of 10 gill rakers on the first arch. Branchiostegal rays number 7. The pyloric caeca are blunt, and number 6 (1 dorsally and 5 ventrally). L. kallion possesses the general characteristics of Lcpophidiinn (see Robins, 1959:361). Morphometric data follow (in each instance the figure in parenthesis refers to the paratype) : standard length, mm., ]]7 (162-damaged) ; head length 23.0(34.6) ; tip of snout to origin of dorsal fin, 27.8 (39.3) ; tip of snout to origin of anal fin, 42.8 (67.2) ; tij) of snout to occiput, 15.5(22.5) ; tip of snout to pos- terior end of maxillary, 11.1 (15.8) ; depth of body, at occiput. 13.7 (23.2); at origin of dorsal tin, 13.0 (24.6), at origin of anal fin, 13.1 (20.7); length of pectoral fin. 11.4 (14.5); length of pelvic rays, 8.6 and 6.4 (12.4, 9.5) ; length of caudal fin, 7.1 (damaged) ; tip of snout to posterior end of lateral line, 112 (150) ; horizontal diameter of bony orbit, 3.5 (5.1) ; post-orbital length of head, 12.7 (19.5) ; snout length (to margin of orbit), 4.2 (6.5) ; bony interorbital, 3.5 (5.1). Despite their age the distinctive color pattern is readily seen on the type specimens. The dorsal, caudal and anal fins are pale with scattered melanophores and with a well defined dark mar- gin. The pectoral fin is clear with some melanophores on its basal third. The body is almost straw-colored in preservative although the scattered melanophores lend a dusky appearance to the belly. Two series of dark brown markings on the upper sides dominate the body coloration. The uppermost series starts as a collar just behind the occiput, runs laterally to a point above the opercle and then extends as a dark stripe, just above the lateral line, most of the distance to the caudal-fin base (Fig. 1). It is broken posteriorly into a series of dashes. In the holotype the stripe is less continuous than in the paraty])e. The lower series of dark brown markings extends posteriorly from the shoulder as a series of elongate blotches, just below the lateral line. Caudally, the lower series crosses the lateral line and merges with the dorsal 1 !*'")!» NEW SPECIES OF LEPOPIIIDIUM 3 series. Two dark spots occur on the nape between tlic dorsal- fin origin and the occipital coHar. Tlic head is nfenerally dusky. A sliort dark bar extends from the upper end of the gill slit, reai'ward along the dorsal niai-gin of the operculum. The snout is not spotted but a few dai-k si)ots may have been present on top of the head, behind the interorbital region. As seen from above, the dark margin of the dorsal fin appears to be a mid- (h)rsal stripe. The floor of the mouth and the gill bars are pale and the gill chamber silvery with numerous melanophores. The roof of the mouth especially toward the pharynx is dusky but not black. The gastrointestinal ti-act is sooty black except for the stomach. Fisu:e 1. Laternl view of the paratype of Lepophidiiim l-allio-n, MCZ i;S0o7B, 162 nun. in standard length, eollected off Barbados, "Blake" station 291. As in other species of Lepophidium, the swimbladder is similar in both sexes, a rather thin-walled sac, with no posterior opening or tube. It occupies the dorsal portion of the body cavity. Numerous short, pointed, non-depressible teeth are present on the premaxillary, dentary, vomer, palatines, and pharyngo- branchials. Those on the outer row of the premaxillary and dentary are test developed. The dentigerous areas are papillose, especially the lips. A prominent spine is easily detected under the skin on the snout and on the upper posterior margin of the opercle. Scalation is typical of the genus. The scales are cycloid, imbri- cate and in regular rows on the body. The postorbit^l portion of the head is also scaled. In L. kallion the scales are imbedded BREVIORA No. 104 and the skin must be scraped away to determine the number of scale rows. About 8 and 25 rows are present between the lateral line and the dorsal- and anal-fin origins, respeotively. Some 165-170 rows cross the lateral line between the posterior edge of tile hypural plate and the point of opercular attachment. So many other characters are available to the systematist that the labor involved in approximate enumeration of scale rows in ophidiids seems unjustified. 2mm lOmm d / 0/ Figure 2. (a, b). Diagram of lateralis system on head of Lepopliidiuni kallion (see text for explanation), (e). Development of fleshy tal)S on tiji of snout of L. kalluin. (d). Tip of snout of L. mdrmnratum (Goode and Bean). The lower rim of the snout is incised anterior to the nostrils. Four fleshy tabs project from this margin (Fig. 2 c). The actual shajic of these tabs apparently varies, foi- tlie ])aratype is less 1!)")!) M;\V SPECIES OF LEPOPHIDIIM 5 (Iccoi-atcd tliaii the Iiolotype. Tavo rows of ficsliy tabs i-uii dor- sally from ill fi-oiit of the aiitoi"ioi' nostril an.l join just aiitcrioi' to the rostral spine { Fi<>'. 2 v) . IJohlke and Kobins (in j)ress ) have demons! i-atcd specific dif- ferences in the lateralis system on the head. In L. hallioti this system may l)e divided into the lateral, supratemporal, supra- orbital, infraorbital and })reopereulomandibnlar canals. Ilubbs and Camion (1935, pi. 2) provided illustrations of these canals in several darters and Robins and Miller (1957:216-217) simi- larly treated Cuttus. A sehematic presentation of the lateralis system of L. kallion is given in Figure 2 (a, b). In the lateral view the eanals are >hown in black. A single pore occurs on the lateral canal just anterior to the attachment of the gill membrane. Three pores (one median) are on short caudally-projecting side canals ex- tending from the supratemporal canal. Each supraorbital canal consists of six pores (one above and slightly behind the eye, one median coronal pore, two over the anterodorsal margin of the eye, one behind and above the anterior nostril and one opening caudally into the snbnasal depression. Each infraorbital canal consists of eight pores (one behind the eye, one behind and below the eye, one at the posterior end of the suborbital rim, four on the sul)orbital rim above the jaw% and one opening anteriorly into the subnasal depression. The preopereulomandibular canal is more difficult to trace on the type material of L. kallion, espe- cially on the sides of the head. Apparently it opens through 6 pores. The four pores along the ramus of the lower jaw" are large and easily seen (Fig. 2 b). One tiny pore opens oft" a side canal just below the junction of the preopereulomandibular canal with the lateral canal. Another pore open., oft' a long side canal on the lower third of the operculum. In at least one other species of Lepophidium a second pore occurs below this one but it could not be found in L. kallion. The posterior nostril is nearly on the anterior rim of the orbit and is small and circular. The larger anterior nostril opens through a short tube, above the subnasal depression, near the anterior end of the snout. 6 BREVIORA No. 104 The name is from the CTi-eelv haJlion, meaning- more beautiful, and refers to the distinctive pattern of this speeies in contrast to the more drab species of Lcpophidiuni. Relatio)} ships. Serious consideration need be accorded only /.. nwrworaium (Goode and Bean). Other Atlantic species were l)riefly mentioned by Robins (1959:366-367). L. )nan))()rat}(iii differs from L. kail ion in many features. 1) The upper sides are spotted instead of striped. 2) The upper half of the head and the snout are spotted. 3) It possesses more gill rakers, usually 11, occasionally 12 or 13. 4) The longer division of the pelvic ray is considerably longer than the pectoral fin. 5) Probably, there are fewer dorsal rays ( 121-129) and, 6) fewer pyloric caeca (3-5). 7) The snout is smooth, its lower rim entire (Fig. 2 d) without any of the numerous fleshy tabs of hallion. 8) A sub- nasal pore is present on the infraorbital canal just below the posterior nostril. L. kaJlion and L. marmoraium are intimate species within the framework of the genus. The color patterns of the two al- though quite different are of a similar nature, quite unlike that of L. staurophor Robins, J., jcannae Fowler or L. ccn'iiiuni (Goode and Bean). The pigmentation of the gut, peritoneum, and oral and branchial chambers are the same, a distinctive combination. The gill rakers are similar, those on the lower arm diminishing gradually in size instead of shifting from equally developed rakers to rudiments. More detailed consideration of morynordinin is beyond the scope of this paper. ACKNOWLEDGMENTS Thanks are due to the Museum of Comparative Zoology for the loan of specimens and to Dr. Henry B. Bigelow, Mr. William C. Schrocder and Mrs. Myvanwy Dick for courtesies extended to the writer during his visit to the Museum. Dr. (Jiles W. Mead of the Tnited States P^ish and AVildlife Service kindly provided x-rays of the type specimens of L. kail ion. Mr. Walter R. Courte- nay photographed the i)aratype. This study has been supported by a grant-in-aid from the National Science Foundation (NSF- CJ-7116) which 1 gi-atefnlly acknowledge. I!*.")!) XKW ST'KCIFS OF T.Ki'oiM nniT'^r 7 LITERATURE CITED BoHLKE, James E. and C. RiciiARn Robins (in ]iress). Studies on fishes of the family Ophidiidae. II. Three new species from the Bahamas. Proc. Acad. Nat. Sci. Philadelphia. HuBBS, Carl L. and Mott Dwight Cannon 1935. The darters of the genera Eololcpis and Villora. Misc. Pnhl. Univ. Mich. Mus. Zool., 30:1-93, 3 pis. Pierce, B. and C. P. Patterson 1879. List of dredging stations occupied by the United States Coast Survey Steamers "Bibb," "Hassler," and "Blake," from 1867 to 1879. Bull. Mus. Comp. Zool., 6(1):1-15. Robins, C. Richard. 1959. Studies on fishes of the family Ophidiidae. I. A new species of Lepophidium from the Caribbean Sea. Bull. Mar. Sci. Gulf and Carib., 8(4) :360-3G8, 2 figs. (1958). Robins, C. Richard and Robert R. Miller 1957. Classification, variation, and distribution of the sculpins, genus Coitus, inhabiting Pacific slope waters in California and south- ern Oregon, with a key to the species. California Fish and Game, 43(3) :213-233, 3 figs. ^~ <- o^^^u BREVIORA MuseiiintTi of Comparative Zoology Cambridge, Mass. April 14, 1959 XrMHKK 105 BVFO GIJNDLACHI, A NEW SPECIES OF CUBAN TOAD By Rodolfo Ruibal Division of Life Sciences University of California, Riverside INTRODUCTION The discovery of a new species of Cuban Bufo is surprising- considering the amount of herpetological colleetino- that has previously been done in Cuba, the apparent abundance and widespread distribution of the new species, and the fact that Cuba already possesses three species of Bufo. Bufo peltacephalus Tschudi, B. cmpusus Cope and B. longinasus Stejneger have Ion": l)een known from the island, the last species having- been described in 1905. In contrast, the other Greater Antillean islands, Hispaniola and Puerto Rico, possess but a single species each of Bufo: B. lemur in Puerto Rico and B. guniheri on His- paniola. The new species is named in honor of Juan Gundlach, the 19th century Cuban naturalist and collector. Gundlach in his ''Con- tribucion a la er])etologia cubana,"' published in 1880, actually recognized the existence of the species here described as new. lie mentions collecting a small species of toad 33 millimeters in length, near H(d)ana. The toads had a distinctive call and some were in amplexus. (4undlach sent some specimens of this toad to Wilhelm Peters at the Berlin Museum who declined to recog- nize the specimens as a new species because he thought that they might represent juveniles of Bufo peltacephalus. Gundlach justly chided Peters for failing to realize that a juvenile would not be calling, in amplexus and laying eggs. It has taken nearly 80 yeai-s to prove Gundlach correct. 2 BREVTORA No. 105 BUFO GUNDLACHI, sp. IIOV. Typr. M.C.Z. W^nil, male, collected about 14 kilometers N.E. of the city of Cama<>iiey, Provincia de Camaguey, Cuba, by Kamon Molina and Kodolfo Kuibal on August 22, 1!)57. Paratypes. M.C.Z. 30552-81, all with the same data as the type. Diagnosis. A small toad differing from Bufo longinasns, the only other known small toad of Cuba, by the following charac- ters: shorter hind limbs (see Table 1); high supraorbital and supratympanic crests that are lacking in B. longinasxis; the skin Figure 1. Dorsal view of the head of B. gundlachi, MCZ 305.")2. The pis- ment patterns have been excluded from the drawing. of the males more spinous than in B. longinasus males; and in Iiaving the parotid gland small and latei-al rather than large and dorsolateral as in B. Jonginasus (see Figures 1-4 for compari- sons). Readily distinguished from juvenile specimens of B. pcltacephalus by the structure of the suprii()rl)ital and supra- tympanic crests as well as by the i)resence of a thin white mid-dorsal stripe in B. giiiiure 3), covers most of the parotid jrland (except for the ventral edge of the gUiiid which is white), and extends along the ventrolateral margin of the l)ody. TTsually there is evident a darker stripe anterior to the insertion of the forelimb and in some cases a short thin dark stripe is present on the midline of the chest. Otherwise (with the exception of the dark throat of males) most specimens have a spotless ventral surface, though a few specimens show various degrees of dark mottling on the throat, chest and/or belly. Other spccime)is. Specimens of the new species have also been collected in Pinar del Rio, Habana, Isla de Pinos, Las Villas, and at other localities in Camague3\ Most of these specimens were collected bv Dr. Albert Schwartz of Albright College, who Figure 4. Lateral view of the head of B. Innginasus dnnni. AMNH 00802. very generously made his material and notes available to me. Morphologically, all of the populations are similar. The only specimens that may be sufficiently distinct from the tyjies to I'epresent a ditt'erent subspecies are from Isla de Pinos. These specimens aj)pear to be small, immature individuals; however, they differ from other specimens in being much less rugose and spinovis and in having a more shari)ly contrasting color pattern. Since only two specimens are Hvailable from Isla de Pinos it would appear best to reserve any subs])eciiic designation of the forms until more specimens are available. The only other noticeable differences observed were that the Contramaestre, C'amaguey, specimens had swollen and enlarged parotids as compared to other populations. Also the specimens 6 BREVIORA No. 105 from near llerradura. ]'inar del Kio, were characterized by the ("act that males were more spinous than the females. List of other specinicus examined. Pinar del Rio: 2.8-2.8 ini. E. of La Mulata, AMXII 60821-28; 2.9 mi. E. of Isabel Kul)i() (formerly Mendoza), AMNH 60829: 2 mi. S.E. of llerradura. AMNII 60880-56; near llerradura. MCZ 2833 (3 spec.) ; 5.5 mi. E. of Candelaria, AMNII 60890-918. Ilabana : Prov. of Habana, MCZ 19394; Isla de Pinos, AMNH 61222-23. Las Villas: 28 km. E. of Trinidad. AMNII 60884-89. Cama«-uey : Embarca- dero de Moron, AMNII 60857; 9.4 mi. S. of Contramaestre, AMNH 60858-83. Ecological data. The type and paratypes were collected at two sites in the open serpentine savanna between Camaguey and the Sierra de Cubitas. The toads were found in m<^dium size clioruses in the afternoon after a heavy and intense rainstorm. All the low, poorly-drained areas of the savanna were flooded and the congregations of toads were iu these shallow, grass- covered temporary pools. The toads were in the water when calling and were wary, submerging at the approach of the col- lectors. The call was loud for such a small animal and might be reproduced as CKEE-CKEE-CKEE. Also calling at this time was a small species of EleutJierodactt/lus. A single specimen of B. empusus was found near the pools. No egg-laying had yet occurred when the specimens were collected. The many specimens collected by Albert Schwartz were col- lected in similar shallow ponds in open fields. In this respect the new species differs from B. loii- the difference in the lengtli of the hind limb of B. (jiindlachi as compared to the other species of West Indian toads is provided. In the case of B. oxpusiis and B. pdtaccphahis it was felt to be more meaniiio'fid to measnre only jnvenile specimens since only the small s])('cimeiis of these two species conld be confu.sed with B. (jundlachi. The measnremeiit of the body lent>th was done Avith Vernier calipers and actnally is the measurement from the tip of the snout to tlie posterior tip of the urostyle. When measnrin<»- the specimen, it was held so as to flatten the body. The len<>t]i of the hind limb is considered to ho the distance from Figure 6. Lateral view of the head of B. empu.sns, MCZ 3733. the vent to the tarso-metatarsal joint measured with a ruler (see Figure 5). The data on Table 1 do not demonstrate any apparent ditference between the various populations of B. gnnd- lachi in regard to hind limb length. The data do demonstrate a significant difference between B. (jundlachi and B. longinasns dunni (and probably the other subspecies of B. longinasiia as well) in this re^jard. The dnta for the juvenile specimens of B. peltacephaliis overlaj) the lower range of B. gundlavhi while the juvenile specimens of B. eurpusits over-lap the upper range of the new .species. 11).')!) NEW yi'LClES OF CUBAN TOAD A comparison of th(> external characters of llic new species with those of the otlier species of Cuhan toads indicates that B. (inndlachi resembles B. ()iipiisiis ;iiid B. peUdccpJinJus more than it does B. lo)ighiasus. />. iiiiiju.siis, B. pcltavcphdUis and B. (jundlaclii 1. I'aiotid gl.'iiui short and lateral in position. '1. Various degrees of supratyni- panie. and sii])raorl)ital cranial crests. /.'. loiKjinana.s 1. I'arotid t;land elong:ate and ex- ] landed dorsally. 2. Absence of well developed cranial crests. Figure 7. Lateral view of the head of B. pcltacepholus, MCZ 292S7. As a consequence of the i)eciiliarly ossified tympanic re<>ion, B. empusHs would appear morphologically to stand out from B. (jundlaclii and B. pcJtacepJialus. HoAvever, all three species ap- pear to be more closely related to the otlier West Indian species, B. lemur and B. guniheri, than to the Cuban B. longinasus. In addition to the characters listed above, B. longinasus is dis- tinctive in the pointed snout of adult specimens, a feature which is, however, less obvious in the youn. /. longinasus and B. I. ramsdeni are available it will lie impossible to critically determine the Figure 9. Lateral view of the head of B. g)tiii]icri, MCZ 3104. validity of the three named forms. Also not included in the key is tlie introduced B. nun'iinis. This form has been success- fully introduced into the Lesser Antilles and Puerto Rico, Ilis- paniola, and Jamaica. It has been introduced into Cuba at least once (Scaramozza, 1947) but interestinoly enouprh there is no evidence that it has established itself on Cuba. It is an attractive hypothesis to attribute the failure of B. marinu:< in Cuba to the existence of B. pcltacephalus. This native species is almost as large as B. mariiius and successfully inhabits tlie cultivated areas of Cuba and thus represents a form that has adapted itself to a situation in which much of the original habitat has been destroyed. 12 BREVIORA No. 105 Kfi) to flic )iafirc \V(.st Indian species of BuFO ]. Tyinpaniun not evident, covered by i)ustulous skin (rigure 4); parotid elongate, the length greater than the width (Figure 2) B. lonfjina.su.s Tynipaniini not covered by pustulous skin; parotid 7iot elongate '2 '2. Tympanum small, completelj' encircled by bone (Figure 6); sharply in- clined profile to snout (Figure (i) . B. empu.sus Tympanum not encircled liy bone; profile not sharply inclined ;^ 'A. Hind limbs short, the body length/hind limb ratio ranging from 1.57 to 1.93; adult size small (maximum body length, 36 mm.); a thin median dorsal white line (Figure 5) B. gundlachi Hind limbs longer, the body length/hind limb ratio ranging from 1.27 to 1.6o; adult size larger (body length as great as 170 mm.) ; no median dorsal white stripe 4 Figure 10. Lateral view of the head of /.'. li-m>ir, M("Z 30615. 4. .\ cauthal knob (Figure 7 ) ; orbital crests indistinct B. pcliaccpltahix Xo canthal knob; well developed supraorbital and supratympanii- crests 5 5. Median supraorbital crest aud posterior supraorbital crest forming a right angle (Figure 8) ; preorbital crest e(|ual to su])raorl)ital crest in thickness B. gunthrri Supraorbital crests confluent and forming an aic; preorbital cresi thiii- ucr aud less jirououuced than su|)ra()rbital cicst (Figure 10) U. liiiiiir 1 f)r)f) XEW SPECIES OK CUBAN TOAD 13 1 -,-> •O >n CO ■* CO o o 00 i-H to o "O CO It •:i f I- >o CO CD a F— I O ■2 o CO CO CD CO o Oi C-5 r^ ^ f§ I— ( !—? ^' ^^ ^H I—I ^H rH rH l-H 1-? iH to i a i-H ■--. o 1 ■to -=> -^ °2 &3 •ra 00 t- H J , > o • i-l > CJ o PM .S ^ ■^ o o o S o 1 -^ o > X b Sh o ew 1 W p o (li PW Oh O s PLh o O O Ah •PH 03 (K ti _f3 ^^ OS K^ 1-^ o P-»^ _K & P-l "3 0^' _'2 s c3 J3 > >i-4 3 c c3 > c3 "to M ci 1-^ ci O cc! o .a a 3 3 PM .^ »3 OS ec hH CB (U 'h fl »; DC 3 '5 rt o > 3 cc ^4 .« 5tl 3 W3 oc ^ "3 sa p— ^ yi w S 0! Ch tH a^ o c^ c3 C3 ^ d '•^ « ^ bo .3 o S o 3 2 s 3 g o o o o O o .3 o o ^ =4-1 «n =*-» «*H !»H ^^ ^+-1 t- S M pq 3 pi pq « M P2 ^^ 14 BREVIORA No. 105 Table 2 Maxiiuum body length of native "West Indian spreios of Biifo Maximum Body Speciefs Scir Leriffth (in mm.) Kimdlachi $ 36 longinasus 9 34 empusus 9 74 peltacephalus $ 170 5?untheri 9 80 lemur 9 105 ACKNOWLEDGMENTS The new species of Ciil)an toad -was collected while doing field work during the summer of 1957 in Camaguey with the finan- cial assistance of a grant from the Penrose Fund of the American Philosophical Society. T am grateful to the Society for this assistance. I Avould also like to thank Miss Dereth Bogert for her excellent drawings, and Dr. Doris Cochran for measuring the type of B. longinasus longinasus for me. LITERATURE CITED B.\RBOUR, T. 1914. A contribution to the zoogeography of the West Indies. Mem. Mils. Comp. Zoo]., 44 (2^: 209-346. 1917. Notes on the hcrpetology of the Virgin Islainls. Pioc. Biol. Soc. Washington, 30: 97-104. 192(i. Xew amphibia. Oct-. Papers Boston So.-. Xnt. Tli.st., 5: 19M94. B.\RBOUR, T. .\ND C. T. RaMSDEN 1919. Ilerpetology of Cuba. Mem. Mus. Comp. Zooi., 47 (2): 71-21:;. (irXDLACII, J. 1880. Contrilnu-ion a la erjietologia cubann. C. Montiel y ('in., ll;ili;ni;i, pp. 1-99. SCARAMOZZA, L. C. 1947. Correspondi'iu-i.-i. Mem. Soc. Cubanu IJist. X;il., 19 (li: 112 113. Schmidt, K. V. 1928. Amphibians and huid reptiles of Porto Hied, with a list of those reported from the Virgin Islands. Sci. Survey Porto Rico and Virgin Isl., 10: 1-1 (id. Stkj.nkgek, L. 1904. Herpetology of Poito liicu. Ann. liept. l'. S. .\;it. Mus. fi.i- 1902, pp. .349-724. 190."). r)escii]»tion of a new tiiad from Cuba. Proc. T'. S. Nat. .Mus., 28 (1406): 765-767. ( jsJbOU BREVIORA Mmseiim of Coimiparative Zoology Cambridge, Mass. April 28, 1959 Number 106 THE OCCIPITO-VERTEBRAL JOINT IN THE BURROW- ING SNAKES OF THE FAMILY UROPELTIDAE By Ernest E. Williams In 1939, R. Hoffstetter described an amazing condition of the head joint in the peculiar burrowing snakes of South India and Ceylon — the Uropeltidae. Quite in contrast with the condition found in all other amniotes, the first vertebral centrum was said to be fused with the occipital condyle, while the second vertebral centrum was unmodified and precocious like those behind it. According to Hoffstetter, also, the atlas neural arches, united ventrally by a ligament, embraced the neck of the occipital con- dyle. The species he personally examined was Rhinophis hlythii. The head joint of Rhinophis planiceps (^ R. philippinus), as Hoffstetter was well aware, had previously been described by Baumeister (1908, pp. 499-502, fig. 22). He, like Hoffstetter, had found the first vertebral centrum (= odontoid process) to be absent as a distinct structure. His account differed, however, from that given more than thirty years later by Hoffstetter in one very important detail. He did not find the atlas embracing the neck of the occipital condyle ; instead it was in its normal place posterior to and articulating with the head of the condyle. Baumeister also interpreted the situation very differently, be- lieving the first vertebral centrum to have been reduced and for the most part fused with and forming the major element in the anterior articular surface of the second vertebral centrum. Whether as described by Hoffstetter or by Baumeister, it is sufficiently evident that the head joint in the uropeltids exam- ined by them is profoundly different from the head joint in any other amniotes. In the course, therefore, of a general study of the head joint in reptiles I have found it necessary to give 2 BREVIORA No. 106 special attention to this feature in this family and in related groups. The simple anatomy of the articulation needed to be carefully restudied, its taxonomic incidence established, and its morphological interpretation reassessed. Since a detailed discussion of these points would be a lengthy digression in any more general paper, I publish this account separately. Skeletons of the following members of the family Uropeltidae have been specifically prepared for this study. In each case the head joint was examined at the moment of separation of head and vertebral column and before further preparation. Melanophidium wyandense Plat y 2)1 efrurus madurensis Plectrurus perroteti PseudotypJdops philippinus Rhinophis hlythii philippinus sanguineus Teretrurus rhodogaster Uropeltis ceylanicus ocellatus ptdneyensis ruhrolineatiis This list includes at least one member of every genus currently recognized in the family and several species of two of these genera, including both the species for which the head joint has previously been described. In this suite of specimens differences of a specific or generic character are indeed observable, but they involve only minor details. The general pattern is remarkably uniform. In every case the second vertebra articulates directly and by a simple concavity with the ball -like convexity of the occipital condyle. There is no intervening odontoid process. The condylar articular surface is never notched dorsally or indented medially. An atlas neural arch is always present. It is never placed over the neck of the condyle, as reported by Hoffstetter (Fig. IB), but instead is, as Baumeister stated, always in articulation posteriorly with the neural arch of the second vertebra, and ventroanteriorly with the ball-like convexity of the occipital 1959 UROPELTID HEAD JOINT condyle. In its articulation with the occipital condyle it fur- nishes a dorsolateral supplementary rim to the cup-like anterior concavity of the centrum of the second vertebra. An atlas ventral arch may or may not be present. It is present as a distinct element only in Vropeltis pulneyensis (Figs. 3A and B). In some of the forms examined it may have been fused Fig. 1 A. Head joint of Typhlops steinhmisi (after Hoffstetter) . B. Head joint of Ehinophis Uythii (after Hoffstetter). C. Real condition of the head joint of Ehinophis blythii. Abbreviations: Ci = centrum 1, the odontoid ; C2 = centrum 2, the axis centrum. lo = second intercentrum ; I3 = third intercentrum. O = occiput. into the ventral margin of the anterior concavity of the second centrum. The presence of a slight anteroventral projection — absent in U. pulneyensis — is suggestive of this (Figs. 2B and D ; Fig. 3D). The contribution of the pedicels of the atlas neural arch- halves or of the ventral arch (when present) to the concavity which receives the occipital condyle, is in all cases minor. The BREVIORA No. 106 major articular surface is always provided by the second cen- trum. The uropeltid head joint ma^' be usefully contrasted with that of a more primitive snake, Cylindrnphis rufus, which is quite typical of the more usual squamate condition. Fig. 2 A. Anterior view, atlas and axis of Bhinophis sanguineus. B. Side view, atlas and axis of Ehinophis sanguineus. C. Anterior view, atlas and axis of Platyplectrurus madurensis. D. Side view, atlas and axis of Platyplectrurus madurensis. l!)5i) ITROPKLTID IlKAD .JOINT 5 111 Cyiindi'ophis rnfiis the occipital condyle bears definite traces, in well-marked grooves, of its origin by the union of three bones and is not hemispherical but distinctly indented on its dorsal surface. This indentation is the fovea dent is (Gaupp, 1908, p. 518), and is the place of attachment of the ligament urn apicis dcntis. In many forms this ligament chondrifies or ossifies in ■whole or in part as a small conical projection from the odon- toid process which it is then convenient to call an apex dentis (new term). In C. ruftis the apex dentis is a small cartilage cone. The occipital condyle is received into a concavity formed by the ventral portions of the atlas ring, the lateral margins of which are formed by large articular surfaces on the pedicels of the atlas neural arch-halves and the articular surface of the ventral arch. The dorsal margin of the concavity is furnished by the liga- nientum transversum atlantis, while the outer portion of the bottom of the cup is the septum interarticulare (Gaupp, 1908, p. 528). Beneath the ligamentum transversum atlantis and per- forating the sept inn interarticulare by means of the foramen septi the apex dentis projects forward into the fovea dentis of the condyle. The contact of the occipital condyle in C. rufus is thus pri- marily with the atlas ring and its associated ligaments, only to a minor degree with the odontoid process, mostly by its apex dentis, and not at all with the centrum of the second vertebra. The atlas ring in C. rufus, as in other typical squamates, forms posteriorly a cup for the reception of the odontoid process, en- tirely similar to that described for the occipital condyle. The same parts participate in a substantially similar way. The odontoid process in C. rufus projects forward from the second vertebra. It is not a conical, distinctly tooth-shaped proc- ess, as in most mammals. In C. rufus it is an essentially wedge- shaped element, flat dorsally, posteriorly with a flat surface that is suturally united with the true centrum of the second vertebra. Its anterior surface slopes sharply backward. Medially, just below the point of juncture of anterior and dorsal surfaces the small cartilaginous cone that is the apex dentis arises so abruptly as to seem a separate element; it has frequently, though prob- ably incorrectly, been regarded as representing the vestigial centrum of a proatl antic vertebra. 6 BREVIORA No. 106 Ventral and somewhat posterior to the odontoid process is the second or axis intercentrum, an element much larger than the odontoid process; it is suturally united to the ventroanterior surface of the true centrum of the second vertebra and projects rig. 3 A. Anterior view, atlas and axis of Uropeltis pulneyensis. B. Side view, atlas and axis of U. pulneyensis. C. Anterior view, atlas and axis of Uropeltis ocellatus. D. Side view, atlas and axis of V . ocellatus. 1959 IIROPELTID HEAD JOINT 7 obliquely backward as the first of two ventral spines on the axis vertebra. Anteriorly this element articulates with the ventral arch of the atlas. With minor differences in details of shape and size this de- scription would serve for the head joint of any typical squamate or of Sphenodon (cf. Typhlops, Fig. lA), and with no serious modifications except in regard to relative size of the odontoid process and second intercentrum, it will serve for any crocodile or most turtles. The essential differences between the uropeltid head joint and that typified by Cylindrophis rufus are : 1. The condyle of the uropeltids lacks a fovea dentis. 2. There is neither a ligamentum apicis dentis nor an apex dentis. 3. The articular surfaces on the atlas neural arch pedicels are reduced. 4. The atlas ventral arch is reduced as a whole or more fre- quently has disappeared altogether as a distinct element. 5. There is no odontoid process. Somewhat aberrant among squamates is the head joint of Cylindrophis maculatus of Ceylon (Fig. 4). The occipital con- dyle is not notched dorsally, nor indented mediall3^ The grooves indicating the place of union of the three participating bones, are, however, present. There was apparently no ligamentum apicis dentis or apex dentis. The basal portions of the atlas ring — including the neural arch pedicels, but especiallj^ the ventral arch — are highly developed. The two neural arch pedicels almost meet mid-dorsally, reducing the ligamentum transversum atlantis to an extremely short connection between the two. The area which could have been filled by a septum interarticulare is very small ; perhaps there was none. The odontoid process itself seems somewhat reduced and truncate in front. In certain ways in which uropeltids differ from typical squa- mates, C. maculatus appears to be modified in a direction oppo- site to that seen in the uropeltids. The basal parts of the atlas are enlarged in C. maculatus, reduced in uropeltids. In certain other ways C. macidatus approaches the uropel- tids: in the absence of the fovea dentis and of the correlated ligamentum apicis dentis or apex dentis. 8 BRE\aORA No. 106 Even the enlargement of the atlas basal elements can, how- ever, be interpreted as tending in the direction of the nropeltids. The typical squamate and other reptilian condition is complex, essentially two joints, one within the other. On the outside there is a major condylar convexity fitting into an atlantal cup ; inside this is a joint oriented in the opposite direction — a con- vexity arising from the odontoid (atlas centrum), fitting into a small concavity within the condjde. See diagram Figure 5A. The nropeltids and C. maculatus have simplified this arrange- ment by abolishing the internal odontoid joint and perfecting a Fig. 4. Cyli)uh-op]tis tiianilatits. A. Anterior view of atlas. B. Side view of axis. ball and socket joint entirely comparable to the joints between vertebrae, posteriorly. (Pig. 5B.) If this be the correct interpretation of the results achieved by the uropeltid - C. maculatus modification, it may reasonably be wondered why the extreme uropeltid condition was found adaptively necessary. C. maculatus has achieved a reasonably good ball and socket joint without losing the odontoid process entirely. A hypothesis may be offered here, for future mechani- cal and functional analysis to test. The axis of C. maculatus even with some reduction of the odontoid is much longer than 1!)59 ITROPELTID HEAD JOINT 9 the vertebrae behind it. The uropeltid axis on the contrary is ff.s short or shorter than the vertebrae behind it. We may assume that in tlie still unanalyzed functioning of the uropeltid head joint this shortening- of the axis^ has been mechanically and thus selectively valuable. That this shortening has been accomplished by reduction of the odontoid to essential nullity we may believe to have been an accident of evolution in this specialized group — or we may see in it the culmination of a trend begun in an ancestral form which must have been similar to C. maculatus. It must be explained that although C. mamdatiis occurs in the same area as the uropeltids, shows in several respects special morphological resemblances to them, and may w^ell have had a long history independent of other members of the group, part of the resemblances between maculatus and the uropeltids may well B. Fig. 5. Diagrammatic horizontal section of the head joint region. A. in most reptiles. B. in uropeltids. be parallelism. At least, as Bellairs and Underwood (1951, p. 231) have indicated, the presence of two ventral transverse scale rows to each body segment in uropeltids rather than one as in Cylindrophis is possibly primitive and prompts the inference that the uropeltids have not been directly derived from Cylin- drophis or its immediate relatives. Thus, while it is interesting to suggest that the uropeltid head joint may in the course of its evolution have gone through a stage directly comparable to that in C. r)iaculatus — a stage in which the basal parts of the atlas were enlarged rather than reduced — there is no compelling reason derived from any demonstrated direct phyletic relation of C. maculatus to the uropeltids to ac- cept such a stage as a necessary intermediate. It would appear 1 Most uropeltids have also elongated the pedicel of the occipital condyle, but this is not true in M elan o phi dium which seems in skull and scutellatlon the most primitive member of the family. 10 BREVIORA No. 106 simpler to suppose tliat in the iiropeltids the modification of the head joint was accomplished from the first by strong reduction of the ventral parts of the atlantal ring and of the odontoid simultaneously. The question is one. however, on which critical evidence is unavailable. If the course of evolution of the uropeltid head joint is likely to remain for some time unknowii, more hope may be offered that its ontogeny can be examined. The uropeltids are ovo- viviparous ; it will therefore be worthwhile to examine the col- lections already in existence for pregnant females. Though this would be unlikely to provide a stage by stage view of the on- togeny of the head joint region, in the present state of our knowledge any glimpse of the development (as of the function or the evolution) of this region in uropeltids would put us sub- stantially farther ahead. LITEEATURE CITED Baumeister, K. L. 1908. Beitrage zur Anatoinie und Physiologie der Ehinophiden. Integument, Driisen der Mundhohle, Augen und Skeletsystem. Zool. Jahrb. (Abt. Anat.), 26: 423-526. Bellairs, A. d'A. and G. Underwood 1951. The origin of snakes. Biol. Review, 26: 193-237. Gatjpp, E. 1908. Ueber Entwicklung und Bau der beiden ersten Wirbel und der Kopfgelenke von Echidna aculeata nebst allgemeinen Bemer- kungen liber die Kopfgelenke der Amnioten. Jenaisciie Denk- schriften, 6: 483-538. HOFFSTBTTER, R. 1939. Sur 1 'articulation oceipito-vertebrale des Uropeltidae (ophidiens fouisseurs). Bull. Mus. Hist. Nat. Paris, (2) 11: 426-433. BREVIORA MitasemiM of Connparsitive Zoology Cambridge, Mass. May 6, 1959 Number 107 A REVISION OF THE DACETINE ANT GENUS NEOSTRUMA By William L. Brown, Jr. Museum of Comparative Zoology, Harvard University INTRODUCTION The genus Neostruma Brown includes six known neotropical species forming a compact and distinctive group in subtribe Strumigeniti of tribe Dacetini. This genus evidently was derived within the New World tropics from the gundlachi group ("sub- genus Pyramica") of the large tropicopolitan genus Strumi- genys Fr. Smith. The >S^. gundlachi group is also neotropical in distribution. It contains a graded series of species, linked to the more "typical" Strumigenys through the S. connectens group (Kempf, 1958, Rev. Brasil. Ent., 8:59-68), that foreshadow Neostruma in head shape, development of serial denticulation of the mandibles, lengthening of labral lobes, reduction of pilositA" and spongiform appendages, and other characters. Neostruma is set off from these by its very long labral lobes with their short trigger hairs, and by the distinctive form and armament of the mandibles, featuring very short apical fork teeth and a series of denticles on each side of a submedian (pre- apical) tooth (Fig. 5). With the exception of N. myllorhapha new species (Fig. 4), a "countercurrent" specialist with sec- ondarily elongate mandibles, the Neostruma species show a co- adaptive complex of characters converging toward those of Smithistruma Brown, but not reaching the same degree of com- pleteness (see Brown and Wilson, in press). Among the charac- ters involved in this complex are the following: 1. Shortening of mandibles. 2. Reduction of apical fork. BREVIORA No. 107 3. Development of serial mandibular deuticulation. 4. Transformation of basal tooth into a perpendicular "basal lamella." 5. Slight lengthening, depression and narrowing of head, with some lessening in prominence of occipital lobes and occipital excision. 6. Shortening and incrassation of antennal scapes. 7. Elongation of labral lobes. Figs. 1-4. Neostruma spp., dorsal views of heads, mandibles and antennae. Fig. 1. N. zeteTci n. sp., paratype worker. Fig. 2. N. metopia n. sp., holotype female, funiculus omitted. Fig 3. N. crassicornis (Mayr), left antennal scape of worker. Fig. 4. N. myllorhapha n. sp., holotype worker, drawn to same scale. All figures 8. Reduction in length of labral trigger hairs. 9. Inability to open mandibles widely (limit is much less than 180°). 10. Hunting behavior is of "strike -hold -sting" type (as based on observation of N. mustelina; see below). 11. Foraging less active and less frequent {N. mustelina) than in several Strumigenys species studied, but much as in Sntithistruma spp. 1959 REVISION OF ANT GENUS NEOSTRUMA 3 The species form small colonies, chiefly in the leaf litter of I'ain forest or tropical ever